Proceedings

of
National Conference on Advanced
Materials, Manufacturing and
Metrology
(NCAMMM – 2018)

Editors
Dr. Samik Dutta
Dr. Shitanshu Shekhar Chakraborty

CSIR- Central Mechanical Engineering

Research Institute
(CSIR-CMERI)
M.G. Avenue, Durgapur- 713 209
[February 16-17, 2018]

ISBN: 978-93-87480-56-8
 Preface

Whenever a product is to be developed, the main aspects that have to be taken care
of, are Materials, Manufacturing Processes and Metrology relevant to the product.
Therefore, this event, named National Conference on Advanced Materials,
Manufacturing and Metrology, NCAMMM-2018 was hosted for fruitful interaction
among researchers working in these areas.

The conference provided a great opportunity to the industry professionals and

researchers in the fields of Materials, Manufacturing and Metrology to get
benefitted from their interactions and deliberations to enhance the capability of
industry oriented research. An opportunity of Industry – Academia research
collaboration can be enhanced through this kind of events. Hence, in today’s
backdrop of enhanced efforts towards indigenous technology development, in order
to achieve self reliance, the need for such platforms are felt. The topics for this
event, chosen carefully encompassing important aspects, are as follows:

 Light metals and alloys

 Ceramics and composites
 Materials for energy storage
 Mechanical metallurgy
 Casting and Powder Metallurgy
 Conventional and Non-conventional machining
 Forming, Welding and Additive Manufacturing
 Monitoring and Control in Manufacturing
 Machine Tool Metrology
 Statistical Quality Control and Optimization
 Signal processing and Machine learning

The content papers from the young researchers, the industry professionals and the
experienced speakers in the above fields will surely provide valuable impetus to the
future research direction in the theme-areas. We gratefully acknowledge the
contribution of the organizers of the conference, reviewers and contributors of the
papers, without which this proceedings would not see the daylight.

Editors
NCAMMM - 2018

I
 Advisory Committee
Prof. Kamanio Chattopadhyay, Former Professor, IISc Bangalore

Prof. Pradip Dutta, Professor, IISc Bangalore

Prof. Ashish Kumar Nath, Professor, IIT Kharagpur

Prof. Soumitra Paul, Professor, IIT Kharagpur

Prof. Surjya Kanta Pal, Professor, IIT Kharagpur

Dr. Shyam S Rao, Sr. Vice President, CUMI

Mr. Subhro Pratim Dutta, GM, Philips, Noida

Mr. Anand Dayal, GM, IDTR, Jamshedpur

Mr. Sabyasachi Roy, Director, ANTS Ceramics

Mr. Pradeep Nair, BDM, National Instruments

Organizing Committee
Chief Patron - Prof. (Dr.) Harish Hirani

Chairman - Dr. Ranjan Sen

Convener - Dr. Nilrudra Mandal

Organizing secretary - Dr. Samik Dutta

Joint Org. secretary - Dr. Prosenjit Das

Treasurer - Dr. Shitanshu. S. Chakraborty

Joint Treasurer - Dr. Himadri Roy

Accommodation committee - Dr. Ranajit Ghosh, Dr. Swarup Kr. Laha

Food and Local Hospitality - Mr. S. K. Naskar, Mr. P. Chowdhury

Transportation - Mr. S. Chatterjee, Mr. R. S. Mondal, Mr. A. K. Singsardar

Media & Sponsorship - Mr. A. K. Roy & Mr. Swapan Barman

II
 Editorial Team
Dr. Ranjan Sen
Dr. Nilrudra Mandal
Dr. Himadri Roy
Dr. Prosenjit Das
Dr. Ranajit Ghosh
Mr. Swapan Barman
Dr. Swarup Kr. Laha
Dr. Sivaprakash Sundaresan
Mr. Rajesh Prasad Barnwal

III
 List of Reviewers
Dr. Manas Mondal Dr. Kaustav Barat Prof. Santanu Das
NIT Durgapur CSIR-NAL Kalyani Govt. Engg. College

Dr. Shailesh Singh Dr. Probir Saha Dr. Amritendu Roy

CSIR-CMERI IIT Patna IIT Bhubaneswar

Dr. Nirmal B Hui Dr. Rashmi R Sahoo Dr. Prosenjit Saha

NIT Durgapur CSIR-CMERI IIT Kharagpur

Dr. Pranab Samanta Sk Tanbir Islam Mr. Abhijit Mondal

CSIR-CMERI CSIR-CMERI CSIR-CMERI

Dr. Swati Ghosh Acharyaa Mr. Bikash Bhunia Dr. Bijay Show
University of Hyderabad JIS College of Engg. NIT Durgapur

Dr. Debashish Ghosh Dr. Rahul Jain Dr. Suvradip Mullick

CSIR-CMERI IIT Kharagpur IIT Bhubaneswar

Dr. Yuvraj Madhukar Dr. Manas Das Mr. Abhijit Sadhu

IIT Indore IIT Guwahati IIT Kharagpur

Mr. Debashish Mishra Mr. L. Gopinath Dr. A. B. Puri

IIT Kharagpur CSIR-NAL NIT Durgapur

Dr. B. B. Ghosh Dr. Ranjib Biswas Dr. Karali Patra

CSIR-CMERI MCKVIE, Liluah IIT Patna

Dr. Harshadeep S Joshi Mr. Sumanta Mukherjee Dr. Kuntal Maji

BATU, Lonere IGIT, Sarang NIT Patna

Mr. Debapriya Patra

Dr. A. K. Lohar Dr. Saurav Datta
Karmakar
CSIR-CMERI NIT Rourkela
IIT Kharagpur

Dr. A. K. Jalan Mr. A. Srinivasan Dr. Ranajit Ghosh

BITS Pilani CSIR-CMERI CSIR-CMERI

Dr. Samik Dutta Dr. S. S. Chakraborty Dr. Prosenjit Das

CSIR-CMERI CSIR-CMERI CSIR-CMERI

Dr. Nilrudra Mandal Dr. Swarup Kr Laha Mr. Swapan Barman

CSIR-CMERI CSIR-CMERI CSIR-CMERI

IV
 Erudite Lectures
1. Chief Guest Lecture on
Composites for High Temperatures: Does Intermetallic has a Role?
by Prof. Kamanio Chattopadhyay, Indian Institute of Science, Bangalore 560012

2. Lecture by Guest of Honour on

Alternative Energy Storage Opportunities using New Materials and Manufacturing
Innovations
by Dr. R. N. Das, General Manager-FCR,CTI, Projects & HSE, BHEL, Corp R&D

3. Lecture by Guest of Honour on

Quality Assurance in Engineering
by Mr. U. Thanu, Director General, National Test House

4. Keynote Lecture
Laser Additive Manufacturing: Scientific &Technological Challenges
by Prof. Ashish Kumar Nath, Department of Mechanical Engineering, Indian Institute of
Technology Kharagpur
5. Keynote Lecture
Industry 4.0 in Friction Stir Welding
by Prof. Surjya K Pal, Professor, Department of Mechanical Engineering, Chairman, Steel
Technology Centre
6. Keynote Lecture
Improving Grindability of Titanium Alloys through Different Fluid Delivery Techniques
by Prof. Santanu Das, Professor, Department of Mechanical Engineering, Kalyani Government
Engineering College
7. Keynote Lecture
Additive Manufacturing in Macro, Micro and Nano Scale: Current Status, Challenges and
Future
by Dr. Shibendu Shekhar Roy, Associate Professor, Department of Mechanical Engineering,
National Institute of Technology Durgapur
8. Keynote Lecture
Graphene-Based Materials for Energy Storage Applications
by Dr. Tapas Kuila, Surface Engineering and Tribology, CSIR-Central Mechanical Engineering
Research Institute, Durgapur

V
 Invited Lectures

1. Laser beam welded skin-stringer joints: comparative assessment of processes and

parameters
by Dr. Kaustav Barat, Scientist, CSIR-National Aerospace Laboratory, Bangalore

2. Investigation on Mechanical and wear properties of novel in-situ 6351Al-Al4SiC4

composites
by Dr. Manas Kumar Mondal, Assistant Professor, Department of Metallurgical and Materials
Engineering, National Institute of Technology, Durgapur

3. A study on the microstructure and property of additively manufactured NiTi based shape
memory alloys
by Dr. Indrani Sen, Assistant Professor, Department of Metallurgical and Materials
Engineering, Indian Institute of Technology Kharagpur

4. Processing-Microstructure-Property Relationship in Additive Manufactured (Laser

Engineered Net Shaping, LENS) Ti-6Al-4V Alloy
by Dr. Shibayan Roy, Assistant Professor, Materials Science Center, Indian Institute of
Technology Kharagpur

VI
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Contents

Preface I

Committee II

List of Reviewers IV

Erudite Lectures V

Invited Lectures VI

Theme: Materials Development and Characterizations

Sub-theme: Light Metals and Alloys

1. The microstructure and oxidation behavior of arc- melted Nb-Si-Ti-Mo based alloy 1
-- Sayantani Santra, Kasturi Sala, Rahul Mitra

2. Corrosion performance behaviour of Mg and Al reinforced nano hybrid-metallic composites 6

-- Anand Pandey and Dalip Singh

3. Simulation study of solid bumper beam under frontal impact load 12

-- AK Rana, S. Chakraborty, S. Datta

4. Optimization of CVD parameters for synthesis of carbon nanofibers on Inconel substrate 17

-- Amit Thakur , Alakesh Manna, Sushant Samir

5. Microstructural modifications of cast Al-17Si-5Cu alloy through isothermal heat treatment 24

-- B Hazra, P Baranwal, S Bera and B K Show

6. Study the effects of Bi addition on microstructure and hardness of hypereutectic Al-17.6Si Alloy 30
-- Kona Durgaprasadu, Prosanta Biswas and Manas Kumar Mondal

7. Effect of space holder size on microstructure, deformation and corrosion response of Ti4Al4Co 36
alloy foam
-- Pradeep Singh, Hemant Jain, Prashant Nair, Anup Khare, I.B. Singh, D.P. Mondal

8. Closed cell aluminium composite foam for crashworthiness applications 42

-- D. P. Mondal, A.N.Ch. Venkat, Sanjeev Saxena

9. Studies of dendritic arm growth during T6 heat-treatment on directionally solidified Al-4.5Cu- 46

SiC composite
-- S. Debnath, G.R.K Sastry R. N. Rai
Contents

Sub-theme: Ceramics and Composites

10. Evaluation of metallurgical and mechanical properties of CeO2 reinforced zirconia toughened 52
alumina
-- B K Singh, Nitai Chandra Adak, S.S.Roy, S.S. Chakraborty, Nilrudra Mandal

11. Effect of MWCNTs addition on the wear and compressive deformation behavior of LM13-SiC- 56
MWCNTs hybrid composites
-- Bishnu Nand Yadav, Gaurav Kumar Verma, Dilip Muchhala, D.P. Mondal

12. Cyclic oxidation behavior of ZrB2-SiC based ultra high temperature ceramic composite 63
-- Sayandip Sarkar and Manab Mallik

13. Effect of CNT-Ni-P composite coating on tribological behaviour for brake pad system 68
-- Atul Kumar Harmukh, Santosh Kumar, Sushma Bharti, Subrata Kumar Ghosh

14. Effect of interphase and dispersion of CNTs on the elastic properties of CNT-Polyethylene 74
nanocomposite
-- Ashish K. Srivastava, Vimal K. Pathak, Mithilesh K. Dikshit

15. Tribological behaviour of coated mild steel with nano Al2O3-Ni-P composite material 80
-- Sushma Bharti, Santosh Kumar, Atul Kumar Harmukh, Subrata Kumar Ghosh

16. Influence of the delamination geometry on the shear behaviour of glass/epoxy composites 86
-- Subhankar Roy, Tanmoy Bose, Kishore Debnath

17. Static and dynamic mechanical properties of glass/carbon fiber reinforced epoxy composite 92
-- Nitai Chandra Adak, Suman Chhetri, Bipin Kumar Singh, Saikat Bolar, Naresh Chandra
Murmu, Pranab Samanta, Tapas Kuila

18. Effect of nanostructured 8% partially yttria stabilized Zirconia (YSZ) coating in oxidation 98
behaviour of Inconel alloy
-- D. Ghosh, S Das, H Roy

Sub-theme: Mechanical Metallurgy

19. Effect of Heat Treatment Parameters on the Carbide Spheroidization of 0.48% Carbon Steels 106
-- Nandita Gupta, S.K.Sen

20. Modelling tertiary creep of in-core pure aluminium using hyperbolic function 113
-- K. Vinay, A. Syed, M. K. samal, A. arya

21. Structural analysis and measurement of mechanical properties of sputter deposited tungsten films 117
-- Satyajit Sarkar, Shubhra Bajpai, Ankeet Pati

22. Study the effect of basicity and FeO contents of slag on dephosphorisation of Steel 124
-- Sujata Devi, Niladri Sen and Basudev Mishra

VIII
 Contents

23. The effect of Al-5Ti-1B master alloy addition on the microstructure, hardness and mechanical 131
properties of hypoeutectic Al-7.6Si alloy
-- Sourabh Gupta, Prosanta Biswas, M. K. Mondal, Rahul Bhandari, S. Pramanik

24. Influence of microwave heat treatment on natural iolite 139

-- Shubhashree Swain, Siddhartha Kumar Pradhan

25. Experimental investigation on tribological behaviour of interacting surfaces with microstructural 144
matrix
-- Santosh Kumar, Subrata Kumar Ghosh, A. Mukhopadhyay

26. Effect of boron modified microstructure on impression creep behaviour of simulated multi-pass 150
heat affected zone of P91 steel
-- Akhil Khajuria, Modassir Akhtar, Rajneesh Kumar, Jaganathan Swaminanthan, Raman Bedi,
Dinesh Kumar Shukla

Theme: Advanced Manufacturing

Sub-theme: Casting and Powder Metallurgy

27. Counter gravity casting- potentials and challenges -A review 158

-- Vineet Chak, Himadri Chattopadhyay, Md. Mahfooz Alam

28. Tribological properties of aluminium graphite composite journal bearing 162

-- S. Ansary, R. Shaikh, M. Sekh, R. Haque, Md. Kamaruzzaman, S. Haidar

29. Microstructure and mechanical properties of rheocast of ADC12 aluminium alloy 169
-- Sujeet Kumar Gautam, Himadri Roy, Aditya Kumar Lohar, Sudip Kumar Samanta, Goutam
Sutradhar

30. Feasibility study of forging of reduced pure Fe2O3 Briquettes 178

-- Ritwik Das, Manas Kumar Mondal, Susanta Pramanik

Sub-theme: Machining

31. Some experimental studies on relative effects of employing more eco-friendly and less 182
hazardous vegetable oil in drops on chip formation and cutting forces in high speed machining of
Inconel-718
-- Arijit Dasgupta, Aayush Dubey, Monojit Deb, Mrityunjoy Mondal and Asit Baran
Chattopadhyay

32. Topological surface of H.S.S and Titanium31 using micro electro discharge machining 188
-- Jush Kumar Siddani, C. Srinivas, N. Nagabhushana Ramesh

33. Experimental investigation of fiber laser cutting of alumina 192

-- Rahul Rakshit, Umar Arif, Shakti Kumar, Mukul Anand, Vikas Kumar, Alok Kumar Das

IX
Contents

34. Influence of different cooling conditions on machinability during turning of EN-24 steel 197
-- Archana Thakur, Alakesh Manna, Sushant Samir

35. Generation of various micropatterns by electrochemical micromachining 203

-- Sandip Kunar and B. Bhattacharyya

36. An empirical view on accuracy and machinability of TiNiCu shape memory alloys during wire 210
electro discharge machining
-- Abhinaba Roy, Narendranath S.

37. Determination of johnson-cook material model parameters for machining simulations using 216
inverse analysis: a review
-- Tarun Kumar S, Hemant Gandhi, Chithajalu Kiran Sagar, Amrita Priyadrashini

38. Effective study on MRR based on different process parameters in micro-electrical discharge 223
machining
-- Arjita Das, Sucharita Saha, Sourav Halder, Kalyan Chatterjee, Nagahanumaiah

Sub-theme: Forming, Welding and Additive Manufacturing

39. Simulation of deep drawing deformation behaviour under simple loading path: a study 229
-- A. K. Rana, A. K. Singh, A. Maitra, B. Mukherjee, A. Bhuiya, S. Datta

40. Thermal and mechanical response in FSSW of sandwich sheets at different dwell periods 235
-- Pritam Kumar Rana, R. Ganesh Narayanan, Satish V Kailas

41. A study on the influence of shielding gas on TiN decomposition in laser surface alloying 241
-- Muvvala Gopinath, G Sai Krishna, and Ashish Kumar Nath

42. Influence of heat input on shear strength and macro-hardness of 316 austenitic stainless steel 247
cladding onto E250 low alloy steel by GMAW process
-- Manas Kumar Saha, Lakshmi Narayan Dhara, Santanu Das

43. Effect of laser cladding parameters on clad-track uniformity 253

-- Debapriya Patra Karmakar, Gopinath Muvvala, Shams Perwez, Ashish Kumar Nath

44. Comparison of laser marking on aluminum, stainless steel and copper sheet Using Nd:YVO4 259
laser
-- A. Roy, N. Kumar, Santanu Das and A. Bandyopadhyay

Theme: Precision Engineering and Metrology

Sub-theme: Monitoring and Control in Manufacturing

45. Prediction of cutting forces in high speed ball-end milling considering inertial forces 264
-- Mithilesh K Dikshit, Vimal Pathak, K.J. Uke, A.B. Puri, Atanu Maity

X
 Contents

46. Electromechanical characterization of dielectric elastomer actuator based pump for optimum 272
volume flow rate
-- Amit Kumar, Dilshad Ahmad, Karali Patra

47. R&D activities for enhancing new product quality: a combined approach of analytic hierarchy 278
process and structural equation modeling approach
-- Sudeshna Roy, Nipu Modak , Pranab Kr. Dan

48. An artificial neural network approach for predicting flank wear, cutting force, and surface 285
roughness for turning operation using ceramic tool insert
-- Ananda Rabi Dhar, Bipin Kr. Singh, Nilrudra Mandal, Shibendu Shekhar Roy

49. Optimization of machining parameters during hard turning of AISI D3 steel using Fuzzy- 291
TOPSIS approach
-- Debabrata Rath, M. Priyadarshini, K. Pal, S.Panda

50. Online experimental characterization of micro-EDM dressing on Ti6Al7Nb biomedical material 299
-- M.S. Shah, Probir Saha

Sub-theme: Statistical Quality Control and Optimization

51. Parametric analysis of surface roughness of electroless Ni-Co-P coating using response surface 306
method
-- Subhasis Sarkar, Jhumpa De, Rajat Subhra Sen, Buddhadeb Oraon, Gautam Majumdar

52. Warpage and shrinkage minimization in an injection molded component using improved particle 313
swarm optimization algorithm
-- Vimal Kumar Pathak, Mithilesh Dixit, Ashish Shrivastava

53. Parametric optimization of aluminum metal matrix composite (AMC) with reinforcement of 322
coconut shell ash
-- G. Srinivasarao, Siva Sankara Raju, K. Vikash Kumar

54. Parametric investigation of E-Jet micro manufacturing process: Taguchi robust design approach 328
-- Raju Das, Amit Ball, Atul Priya, Shibendu Shekhar Roy, Naresh Chandra Murmu

55. Data-based modeling and continuous adjustment of CNC turning process: A case study 334
-- O M Vinod and P B Dhanish

56. An integrated entropy-combinative distance-based assessment (codas) method for aerospace 340
material selection
-- Anirban Roy, Prasenjit Chatterjee, Shankar Chakraborty, Suprakash Mondal

57. Regression model formulation for prediction of clad layer characteristics in an in-house built 347
coaxial nozzle based DMD system
-- Anirban Changdar, Piyush Pant, A.K.Lohar

XI
Contents

58. Optimization of micro hardness and facture toughness of Zirconia Toughened Alumina (ZTA) 355
under different compacting pressures and sintering temperatures using Response Surface
Methodology (RMS)
-- Subhrojyoti Mazumder, Kunal Ghosh, Himadri Roy, Nilrudra Mandal

Sub-theme: Signal Processing and Machine Learning

59. An expert system based approach for selection of wear resistant materials for steel plant 366
applications
-- K. K. Singh, C. Mandal, Santosh Kumar

60. Identification of diseased cells using image processing 373

-- Puja Mitra, Samik Dutta, Abhiram Hens, Nagahanumaiah

61. Brain MR image analysis using discrete wavelet transform with GLCM feature analysis 379
-- Srinivasan Aruchamy, Partha Bhattacharjee, Goutam Sanyal

XII
 THEME

Materials Development
and Characterization
 Light Metals and Alloys
 Ceramics and Composites
 Mechanical Metallurgy
Sub - theme

Light Metals and

Alloys
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

The Microstructure and Oxidation Behavior of Arc- melted Nb-Si-Ti-Mo based Alloy

Sayantani Santra1, Kasturi Sala2, Rahul Mitra2

1
Metallurgical and Materials Engineering Department, NIT Durgapur
2
Metallurgical and Materials Engineering Department ,IIT Kharagpur

Abstract: Nb-silicide based alloys exhibit surpassing high temperature strength as compared to the commercial
nickel based superalloys. Endeavor has been made to comprehend the idea of the eutectoid and eutectic reactions in
Niobium-rich segment of the Nb-Si binary system to enhance the low-temperature ductility through microstructural
control. The arc-melted Nb-Si-Ti-Mo based alloys consist of the constitutional phases Nbss and Nb 5 Si 3 and the
phases are affirmed using scanning electron microscopy (SEM) and X-ray diffraction (XRD). The Nb 5 Si 3 network of
dendrites of hypereutectic composition may go about as powerful boundary to oxidation resistance at high
temperatures. The top surface has been portrayed by SEM and XRD after oxidation, which was exposed at 1100°C
in air for 12 hrs. Addition of Mo and Ti upgrade oxidation resistance.
Keywords: Microstructure, XRD, SEM, Oxidation

1. Introduction
The jet engines efficiency firmly relies on the most extreme temperature in the engines i.e., inlet temperature of
high-pressure turbine [1]. In gas turbine engines, the hot-end segments comprised of nickel based super alloys have
◦C) and i t has come to or surpassed 85% of its softening
been near its most extreme temperature constrain (~1100
point. So, for higher temperature structural constituents of gas turbine engines, the improvement of the alternative
materials was required urgently. It has been appeared from late research that Nb-Si-based alloys demonstrate
extraordinary potential to defeat the working temperature obstruction of Ni super alloys and to enhance the
proficiency of jet engines [2,3]. Numerous materials analysts have been pulled in by Nb silicide alloys because of
their high liquefying point, comparatively lower density and great high-temperature strength. Niobium Silicide
system ultra-high temperature intermetallics are extremely encouraging for supplanting Ni based superalloys in the
scope of 1100~1400 ℃ application. [4]. Nb silicide in situ composites provides increased temperature ability and
reduced density. These alloys usually consist of ductile Nb solid solutions (Nb ss ) and stiffening Nb/Si silicides. In
case of these composites, the ductile phase of Nb solid solution (Nb ss ) can provide ambient temperature toughness
and hard-brittle intermetallic of Nb 5 Si 3 (and/or Nb 3 Si) is able to provide elevated temperature strength. Be that as it
may, because of the lacking harmony between elevated-temperature strength and low-temperature damage tolerance,
it is as yet one of the significant issue for commonsense reason. If there should arise an occurrence of Nb-Si based
compounds, a promising strategy for enhancing the mechanical properties is microstructure control. It includes two
kinds of phase reactions i.e. eutectic solidification (Eq.1) which is trailed by an eutectoid composition reaction (Eq.
2)
L → Nbss + Nb 3 Si (1)
Nb 3 Si→ Nbss + Nb 5 Si 3 (2)
The microstructure and oxidation behavior of arc- melted Nb-Si-Ti-Mo based alloy

The primary Nbss dendrite phase and the Nb 3 Si phase are produced by solidification (Eq. 1). As of late, researchers
have concentrated on mostly developing the ternary Nb-Ti-Si system based alloys and these are contemplated of
having great mix of properties. In any case, these are still extremely prone to oxidation at high temperatures.
Alloying is one of the most advantageous technique for optimization of integrated properties of the alloys. For
investigation the effect of alloying elements on the microstructure, phase formation and oxidation behavior of the
alloys, alloying Nb-Ti-Si based system with elements such as Cr, Mo, Ge and Sn, etc. has been taken under
consideration [5-6]. It has been reported that the appropriate content ranges additions of these elements can be
advantageous to hinder the oxidation in case of the Nb-Ti-Si based alloys. Alloying addition of Mo in the alloys
helps to straighten the materials by solid solution hardening, while inconvenient effect on the oxidation resistance
has been observed as a result of the development of porous scale and the evaporation of MoO 3 . Nb-based or Nb-Si–
based alloys are very prone to oxidation at high temperature and this is one of the major disadvantages [7]. Nb 5 Si 3
experiences accelerated pest disintegration in the temperature scope of 700◦C to 1000◦ C, producing Nb 2 O 5 [8].
Nb 5 Si 3 has performed complete disintegration on exposure at 1000◦C for 1 to 3 hours[7].To ameliorate the
oxidation resistance in case of the binary Nb-Si alloys, research has adopted the addition of various alloying
components, for example, Ti, Al, and Cr [9-11]. Rapid oxidation behavior is experienced by Arc-melted specimens
having large number of micro-cracks and then they fully get transformed into powder after 3 hrs exposure in the air
at 1023K. Grain boundary and pores can enhance oxidation reaction rate [12].

2. Experimental procedure
Nb silicide based alloys were prepared by adding 20 wt% Ti and 5 wt% Mo. Arc melting was carried out under
argon atmosphere and then samples were cut by electrical discharge machine (EDM). The specimens were polished
to mirror finish and then cleaned in acetone and alcohol consequently prior to observation. Microstructures of
samples were examined with the help of scanning electron microscope (SEM) and energy dispersive spectroscopy
(EDS) was used to do elemental analysis. Arc-melted sample was exposed at 1100 °C for 12hrs after polishing. X-
ray diffraction (XRD) was carried out to portray the constitutional phases which were present in the oxides scale and
afterward EDS was completed for elemental analysis.

3. Results and Discussion

3.1. Microstructure

Figure 1: Typical SEM (BSE) images of the hypoeutectic Nb-Ti-Si-Mo alloy at (a) lower and (b) higher
magnifications.
The Nb-Ti-Si-Mo alloy having Si concentration less than the concentration at the eutectic point is hypoeutectic.
Figures 1(a) and (b) depict the typical SEM (BSE) images of the hypoeutectic Nb-Ti-Si-Mo alloy. The

2
 Proceedings of NCAMMM - 2018

microstructure of Nb-Ti-Si alloy comprises eutectic mixture of Nb ss and Nb 5 Si 3 and intermetallic dendritic phase
Nb 5 Si 3 .Nb 5 Si 3 is the primary phase in the hypoeutectic alloy.

3.2. Oxidation behavior

The isothermal oxidation behavior of Nb- Silicide was evaluated at 11000C and characteristics of the oxide scales
are discussed.

3.3. Scale morphology

Figure (2) depicts the XRD pattern acquired from oxide scale of the investigated alloy delineating the peaks
representing Nb 2 O 5 , TiO 2 and SiO 2 .

Figure 2: XRD profile of the oxide scale produced over the Nb-Si-Ti-Mo based alloy after exposure at 1100◦C for
12 hrs

At the given temperature, for the corresponding alloy, the top surfaces and cross sections of the oxide scales relate to
higher mass pick up or oxidation and it is normally to a great degree rough with discontinuities and most presumably
created by spallation.

Figure 3(a) at lower and (b) at higher magnifications demonstrate the SEM images delineating the oxide scale top
surface created over the hypoeutectic alloy because of exposure at 1100◦C for 12 hrs. The oxide scale can be found
of having no hint of Mo.

The Nb 2 O 5 and SiO 2 produce a eutectic mixture of softening point 1449◦C, where the softening temparature of
Nb 2 O 5 is approximately 1550◦C . Regardless of whether the temperature of isothermal oxidation explore is bring
down in this examination, it is likely that the diffusivity of oxygen would be higher in the eutectic mixture.

3
The microstructure and oxidation behavior of arc- melted Nb-Si-Ti-Mo based alloy

Addition of Ti and Mo upgrade oxidation resistance. Ti enhances the oxidation resistance by the arrangement of
protective layer over the surface and Mo helps in oxidation resistance by increasing the sinterability of oxide scale
surface. Voids are showed up as imperfections.

Figure 3: SEM images of oxide scale produced over the Nb-Si-Ti-Mo based alloy after exposure at 1100◦C for 12
hrs.
4. Conclusions
Arc-melted specimen has been observed under SEM and found to be comprised of Nb 5 Si 3 and Nbss phases. The
hypoeutectic Nb Silicide based alloy, prepared by arc melting, is subjected to isothermal temperature at 11000C for
12hrs. Nb 2 O 5 , SiO 2 and TiO 2 phases have been found from XRD analysis and these protect the sample from further
oxidation.

Acknowledgements
The author is very much grateful to Prof. Rahul Mitra, Kasturi Sala (Research Scholar) and technicians of
Metallurgical and Materials Engg. Department, CRF at IIT Kharagpur.

References
[1] Yang C, Jia L, Zhou C, Zhang H, Sha J. Microstructural Evolution and Mechanical Behaviors of an Nb-16Si-
22Ti-2Al-2Hf Alloy with 2 and 17 at. pct Cr Additions at Room and/or High Temperatures. Metallurgical
and Materials Transactions A. 2014 Oct 1;45(11):4842-50. B.P. Bewlay, J.J. Lewandowksi, and M.R.
Jackson: JOM, 1997, vol. 49, pp. 44–45.
[2] Subramanian PR, Mendiratta MG, Dimiduk DM. The development of Nb-based advanced intermetallic
alloys for structural applications. JOM. 1996 Jan 1;48(1):33-8.
[3] Wu M, Wang Y, Li S, Jiang L, Han Y. Effect of Si on microstructure and fracture toughness of directionally
solidified Nb silicide alloys. International Journal of Modern Physics B. 2010 Jun 30;24(15n16):2964-9.
[4] Grammenos I, Tsakiropoulos P. Study of the role of Al, Cr and Ti additions in the microstructure of Nb–
18Si–5Hf base alloys. Intermetallics. 2010 Feb 1;18(2):242-53.
[5] Li Z, Tsakiropoulos P. Study of the effect of Cr and Ti additions in the microstructure of Nb–18Si–5Ge based
in-situ composites. Intermetallics. 2012 Jul 1;26:18-25.

4
 Proceedings of NCAMMM - 2018

[6] Jackson MR, Bewlay BP, Rowe RG, Skelly DW, Lipsitt HA. High-temperature refractory metal-intermetallic
composites. JOM. 1996 Jan 1;48(1):39-44.
[7] Chattopadhyay K, Mitra R, Ray KK. Nonisothermal and isothermal oxidation behavior of Nb-Si-Mo alloys.
Metallurgical and Materials Transactions A. 2008 Mar 1;39(3):577-92.
[8] P.R. Subramanian, M.G. Mendiratta, and D.M. Dimiduk: Mater. Res. Soc. Symp. Proc., 1994, vol. 322, pp.
491–502.
[9] P.R. Subramanian, M.G. Mendiratta, D.M. Dimiduk, and M.A. Stucke: Mater. Sci. Eng., A, 1997, vols.
A239–A240, pp. 1–13.
[10] P.R. Subramanian, M.G. Mendiratta, and D.M. Dimiduk: JOM, 1996, vol. 48, pp. 33–38.
[11] Zhang F, Zhang LT, Shan AD, Wu JS. Microstructural effect on oxidation kinetics of NbSi2 at 1023 K.
Journal of alloys and compounds. 2006 Sep 28;422(1-2):308-12.

5
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Corrosion Performance Behaviour of Mg and Al Reinforced Nano Hybrid-Metallic

Composites

Anand Pandey1, Dalip Singh2

1 2
Department of Mechanical Engineering, Department of Automobile Engineering, Manipal University Jaipur, India
1 2
Email: anand.pandey@jaipur.manipal.edu, dalip.shekhawat@jaipur.manipal.edu

Abstract: Nano-hybrid metallic composites-stir cast process based were the object of the study in respect to their
corrosion testing of aluminium alloy (Al-6061) and Magnesium alloy (Mg-AZ91D). The cast hybrid nano-metallic
composites were manufactured using nano size reinforcements viz. SiC, graphite and alumina of size 100 nm. The
comparative results finding of cast composites were done using potentiodynamic polarization tests. The results
reported out and best results achieved for aluminium reinforced graphite based composites with better corrosion
behaviour performances.
Keywords: Composites, Nano, Al, Mg

1. Introduction
Nano Hybrid Metallic composites (NHMC) are a new special case of advance materials category of materials, in
which the reinforcements in the range of nano-meter size are added using solid, liquid or gas processing techniques
[1-2.].NHMC’s have a great demand in aerospace, naval and nano electro mechanical components and parts.
[3].Aluminum alloys are promising materials in high technology fields owing to their excellent specific mechanical
properties. Both aluminum and Magnesium alloys have unique physical, chemical and mechanical properties, which
enables their application to be used as matrix material for manufacturing of metallic c composites [4-5]. NHMC’s
are metallic based composites which have a combination of two or more reinforcements [6-7]. The aim of the work
presented here is to investigate the corrosion behavior of stir casted nano-composites with addition of silicon
carbide, alumina and graphite of nano sizes.

2. Experimental Details
2.1 Matrix Materials
In the present investigation Aluminium and Magnesium alloy has been selected and used as matrix alloy. Al (6061)
and Mg (AZ91D) finds a wide applications in production of aerospace, automotive, turbine blades and nuclear
industry. Chemical composition of both matrix materials has been shown in Table 1 and Table 2.
Table 1. Chemical Composition of Mg Matrix alloy [AZ91D]
Al Cu Fe Mn Si Zn Ni Mg
8.4 0.03 0.005 0.14 0.10 1.0 0.001 Balance

Table 2. Chemical Composition of Al Matrix alloy [AA6061]

Cu Mn Si Zn Mg Cr Ni Fe Al
0.147 0.07 0.10 0 2.11 0.23 0.01 0.123 Balance

2.2 Reinforced Materials

The following reinforcement nano particles have been used to mix with matrix alloy through stir casting route:
Corrosion performance behaviour of Mg and Al reinforced nano hybrid-metallic composites

1. Graphite (100 nm) 2. SiC (100 nm)3. Al2O3 ( 100 nm)

3. Materials Processing
Al and Mg base metallic nano composites were produced using stir casting route, which is one cheapest and most
popular. Induction furnace of Max 1600 °C has been used to melt the raw alloy.The Stirrer blade was stainless steel
as shown in Fig.1 Matrix Materials (Al and Mg) were melted in a graphite crucible in induction furnace. They were
preheated at 300°c for 1-2 hours before melting and before mixing the nano particles (SiC/graphite/Al2o3) was
preheated at 300°c-500°C) for 1-2 hours.

Fig.1: Stirrer blade in induction furnace Fig.2 Stir Casting Setup

As shown in Fig. 2 the induction furnace temperature was raised above the liquidus temperature to melt the alloy
completely and then cooled down just below the liquidus temperature to keep the slurry in a semi solid state.At this
stage the preheated reinforcement nano particles were added and mixed manually. The composite slurry was
reheated to a fully liquid state and then automatic mechanical mixing was done for 30 minutes at a stirring rate of
100 rpm.

4. Hardness and Tensile Tests

The produced nano hybrid composites hardness has been measured using Brinell hardness tester. A load of 400 kg
was applied to the nano-composites samples through 10 mm steel ball for a 15 seconds. All the composites samples
has been investigated through tensile tests (ASTM E8 standards) using low strain rate testing (SSRT) as shown in
Fig.3. The results has been shown in Table 3. It has been found that maximum hardness has been reported with Mg
reinforced nano size particles in comparison to Al reinforced graphite. This may be due uniform distribution and
high hardness of graphite, SiC and alumina particles.

Fig. 3: Slow strain rate testing (SSRT) using Tensometer

7
 Proceedings of NCAMMM - 2018

Table 3. Mechanical Properties of samples

Composition Hardness Ultimate tensile
(BHN) strength (MPa)
Al (6061) 66 320
Al (6061)+graphite 74 311
Mg(AZ91D) +(graphite+SiC+Al2O3) 95 240

5. Results and Discussions

Corrosion testing performance behaviours of stir cast nano-hybrid composites has been tested using Potenstiostate
SP 150 as shown in Fig.4

Fig. 4: Testing Potenstiostate SP 150

5.1 Corrosion Tests for Al

Prior to EIS and potentio-dynamic polarization, OCP has to stabilize. In present experiments sample of 4 Al-samples
were kept in NaCl solution for 5hrs before performing OCP and then OCP were recorded as a function of time for
up to 1 hrs in order to achieve a stable value. Fig.5 indicate that the value of potential are not stable, it decreases
from 1.45 V 1.27 V and show instability throughout 1hrs OCP tests.

Fig. 5: Bode plots for Al -alloy

Fig. 6 show Bode plots for 4 Al-sample reinfoced with SiC. The relaxation process set at high-intermediate
frequencies along with the coating things at 800, indicates a slight capacitive performance with dielectric properties
viz. coatings have the ability to charge and avoids the ionic flow of the corroding solution and the impedance of 4
aluminum sample shows 3.92 Ohm. The results also shows the corrosion properties of the coating achieved in the
form of plots viz. Nyquist diagrams as shown in Fig. 7. Results also illustrated the good arc-like/weber behaviour
over the frequency range examined and indicate the decent corrosion behavior.

8
Corrosion performance behaviour of Mg and Al reinforced nano hybrid-metallic composites

Fig.6 Bode plots: Al -alloy reinforced SiC Fig.7 Bode plots: Al-alloy reinforced Alumina

5.2 Corrosion Tests for Al reinforced Graphite

Prior to EIS and potentiodynamic polarization, OCP has to stabilize. In present experiments sample keep in NaCl
solution for 5hrs before perform OCP. The open circuit voltage potential has been recorded for up to 2 hrs to achieve
a stable charge. Fig.8 (a) indicate that the value of potential are not stable it increases from -1.3 V to -8.5V and after
½ hrs it’s fixed with the stable value -8.5 V.

Fig. 8(a): Bode plots: Al alloy reinforced graphite

Fig. 8.(b) Show typical Bode plots for Al-graphite sample. The relaxation route at high-intermediate frequencies,
related with the coating properties has been shown as phase angle being 700, indicating slight capacitive
performance of dielectric properties viz. coatings ability to charge, avoiding the ionic flow of the corroding solution
and the impedance of aluminum graphite sample shows 4.25 Ohm. Corrosion properties of the Al-Gr. sample can
be achieved by plotting this data in the form of Nyquist diagrams shown in fig.8 (b). Al-graphite sample, shows
good arcperformance over the frequency range observed.

9
 Proceedings of NCAMMM - 2018

Fig. 8(b): Bode plots for Al alloy reinforced graphite

5.3 Corrosion Tests for Magnesium reinforced-Graphite/SiC/Al2O3

Experiments sample MG- Graphite - SiC - Al2O3 alloy keep in NaCl solution for 5hrs before perform OCP and then
open circuit voltage has been recorded up to 2 hrs in order to achieve a stable value. Fig.9 indicate that the value of
potential are not stable it increases from -1.697 V to -1.690 V and after one hrs it’s fixed with the stable value -1.690
V

Fig. 9: Value of potentials for Magnesium reinforced alumina, graphite and silicon carbide

Fig. 10. Indicating the bode plots forMg reinforced- Graphite/SiC/Al2O3 alloy. During the investigation at high-
intermediate frequencies, connected with coating properties has beenobtainable at phase angle 600, indicating a
poorer capacitive performance with dielectric properties viz. coatings have the ability to charge, avoiding the ionic
flow of corroding solution along with the impedance of Mg- Graphite/SiC/Al2O3 alloy. Sample shows 1.9 Ohm
which is indicate less corrosive resistance behaviour.A improved insight into the corrosion properties of the Mg-
Graphite/SiC/Al2O3 alloy can be achieved by plotting this data in the form of Nyquist diagrams shown in fig.10. The
experimental data for Mg- Graphite/ SiC/Al2O3 alloy show good arc-like performance over the frequency range
with lesss impedance.

10
Corrosion performance behaviour of Mg and Al reinforced nano hybrid-metallic composites

Fig. 10: Bode plots of Magnesium reinforced SiC, graphite and alumina

6. Conclusions
The nano hybrid metallic composites prepared through stir casting route with reinforcing graphite, silicon carbide
and aluminium oxide has increased the hardness and tensile value of the fabricated composites. The results reported
out and best results has been achieved for aluminium reinforced-graphite based composites in terms of corrosion
performance.

Acknowledgement
The authors are thankful to NICOP lab, Manipal University Jaipur for assisting in corrosion testing of fabricated
samples.

References

[1] Kumar D,Agnihotri G, Purohit R. A Review on Properties, Behaviour and Processing Methods for Al- Nano
Al 2 O 3 Composites. 2014. (6): 567–89.
[2] Pandey A,Bains HS,Manna A.Particulate reinforced Al-MMC:Oppurtunity and production. 2007 CPIE.22-24
March, NIT-Jalandhar,India
[3] Kannan C, Ramanujam R. Comparative Study on The Mechanical and Microstructural Characterisation of
AA7075 Nano and Hybrid Nanocomposites Produced by Stir and Squeeze Casting. Journal of Advanced
Research.2017 (8): 309-19.
[4] Mohanavel V, Rajan K, Senthil PV, Arul S. Mechanical Behaviour of Hybrid Composite (AA6351+Al2O3+Gr)
Fabricated by Stir Casting Method.Materials Today.2017(4): 3093–3101.
[5] Kandpal BC, kumar J, Singh H. Fabrication and Characterisation of Al2O3/Aluminium Alloy 6061 Composites
Fabricated by Stir Casting. Materials Today.2017 (4): 2783–92.
[6] Liu J, Li J, Xu C. Interaction of The Cutting Tools and The Ceramic-Reinforced Metal Matrix Composites
During Micro-Machining: A Review.CIRP Journal of Manufacturing Science and Technology. 2014 (7): 55–
70.
[7] Reddy AP, Krishna PV, Rao RN, Murthy NV.Silicon Carbide Reinforced Aluminium Metal Matrix Nano
Composites-A Review.Materials Today.2017(4):3959–71.

11
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

Simulation Study of Solid Bumper Beam under Frontal Impact Load

A. K. Ranaa, S. Chakrabortya, S. Dattaa,*

a
Mechanical Engineering Department, Academy of Technology, Aedconagar, Hooghly 712121, W.B, India *Email
of corresponding author: suchibrata.datta@aot.edu.in

Abstract: The car bumpers play an important role in absorption of impact energy. Based on the design and material
properties, the performance of its behaviour is judged. Application of advance high strength steel (AHSS) in car
manufacturing have been increasingly used due to their high strength and ductility, though it remains a strict
competitor to composites and aluminium but AHSS provides a good balance between strength, ductility and weight
reduction. Thus AHSS TRIP (Transformation-Induced Plasticity steel) 780 steel was considered for the analysis,
here 780 represents the ultimate strength of the material. Throughout the analysis the beam was considered
homogenous and isotropic, and the individual material properties of constituent phases and its volume fraction
effects was not considered, only the overall strength and ductility properties of TRIP steel was considered. The finite
element analysis of this low velocity impact was carried out in Abaqus software and the outcomes such as increase
in strain energy and kinetic energy absorption were analysed along with deformation behaviour. The time period of
the simulation was found by considering the natural time period of the beam structure and using it as a reference.
Keywords: Car bumper beam, Advance High Strength Steel, Finite Element analysis, Strain energy

1. Introduction
The automobile bumper main elements are its facia and its energy absorbing unit. The energy absorbing unit may
consist of a formed thermoplastic backing up bumper beam made of steel, fibre glass composite, plastic or
aluminium [1]. But the essential component i.e. the bumper beam is always present in any types of car. During a low
scale impact with the car running at 15mph, these car bumpers are suitable to absorb this impact energy and thus
preventing any passenger injury. This work focuses on application of AHSS TRIP steel as a suitable material for a
bumper beam. AHSS provides a number of advantages over aluminium and mild steel. Notably aluminium in order
to provide the same strength must be three times thicker than AHSS for beams used in car. Mild steel can provide
the same strength with double thickness i.e. AHSS provides a weight reduction of 50%. Hence it is advantageous to
use AHSS for safety components, structural parts of the car body and the chassis [2]. In this simulation, a continuum
consideration was used, and did not involve the effect of volume fraction of its phases, only its overall material
properties were considered. Researcher Thacker et.al [3] conducted crash-testing simulation study of a 1997 Honda
Accord. The vehicle was stripped down to its basic parts, and each component was analyzed considering different
material properties. A similar study was carried out by Yehia A. Abdel-Nasser [4] where he conducted crash
simulation on different types of lightning columns and predicted the variation of the columns thickness with the
energy absorption capacity thus modeled lightning columns instead of car bumper. Various research works were
done, and many experiments were conducted for determination of material in frontal car crash test. A method for
selection of individual types of steel based on deformation work and bio-mechanical limits of human organisms
during car crash [5] and various designs of bumper and their behavior have been studied [6], thus different varieties
of steels that are used in automotive industry, their design and mechanical properties [7] play a key role in
Simulation study of solid bumper beam under frontal impact load

absorption of impact energy. And changing the body thickness greatly affects the energy absorbed during impact
[8].

2. Design of Car Bumper Beam using Creo 3.0 Parametric Software

Modeling of car bumper (fig. 1) is done with help of

Creo CAD software. The design and dimension of the
car bumper is approximated from Honda Civic car’s
width and design [9], Dimensions of car bumper
beam is taken as follows: Length of bumper beam:
1800 mm, Width of bumper beam: 100 mm, Angle of
bumper beam: 36°, Thickness of bumper beam: 50
mm, Width and length of retention plate of bumper
beam: 120 mm, Thickness of retention Plate: 20 mm.
Fig. 1. Designed bumper beam
3. Results and Discussion
3.1 Finite Element simulation
Commercially available software Abaqus FEA was
used to perform this simulation. Abaqus does not have
any specific units of measurement, thus a consistent
unit of Kg, N, sec, mm was considered for the whole
analysis. Following assumptions were made to carry out
the simulation: i) the body of the bumper beam which is
considered as deformable structure is kept stationery for
analysis and ii) the analytical rigid wall that is
considered as rigid structure is movable. This was done
to carry out a free body analysis of the bumper beam in Fig.2. a) Actual bumper beam [10]
b) Meshed bumper beam for analysis
order to evaluate its crashworthiness. Thus only the
mass of the bumper beam was considered. The wall was designed in Abaqus with a length of 1000 mm and width of
100 mm. In this dynamic analysis the boundary conditions of the beam involved is constrainment at its ends such
that it is immovable, while the analytical wall is made to move with a certain velocity for a certain time period. The
retention plates of the beam were fully constrained, while for the wall, it was constrained in all directions except
along the axis of application of low velocity. The bumper beam was considered homogenous and isotropic and
overall strength and ductility properties of TRIP steel [11] was considered. The bumper beam was meshed with
C3D4 linear tetrahedral elements (fig. 2) with total number of elements 12002 and total number of nodes 2806 and
since wall is considered as analytical rigid thus it was not meshed. Since the wall is allowed to accelerate instead of
the beam, mass had to be assigned to it. To find the mass of the bumper beam, its density and its volume were
needed. Since the beam was designed in CREO, thus it was suitable to use the modeling software to find its volume.
Using CREO software, it is found that the volume of the bumper beam to be V=12388146 mm3and as per auto-steel

13
 Proceedings of NCAMMM - 2018

[11], the density(ρ) of TRIP 780 steel is 7.80 × 10-6 kg/mm3, Therefore the mass of the whole body is V × 𝜌𝜌= 96.62
Kg. The gap between the rigid wall and the bumper beam was kept as 5mm. The time period of the simulation was
found by considering the natural frequency of the beam structure and using it to find the time period of the
experiment. From the Abaqus simulation, the found natural frequency was 73.831 Hz. 𝑇𝑇 = 1/𝑓𝑓 in seconds, where 𝑓𝑓
is the frequency of response of the beam. Putting the found value of 𝑓𝑓 = 73.831 𝐻𝐻𝐻𝐻, thus T = 0.0135 sec.

The inputs that were taken for impact simulation were: density of steel = 7.80 × 10-6 kg/mm3, plastic stress and
plastic strain data obtained [11], Modulus of elasticity = 206824 MPa, Poisson ratio = 0.28, Impact velocity = 10000
mm/s, Time period of application = 13.5 milli-sec, Surface friction coefficient = 0.4. The deformation characteristic
obtained in the simulation has been illustrated in fig 3 and fig. 4.

Fig. 3. Distribution of equivalent plastic strain at different time frame

Fig. 4. Distribution of von Misses stress at different time frame

Figure 3 shows the plastic equivalent strain occurring in the beam under different time frame. At first 2ms, different
deformation zone at various areas had already occurred thus deformation initiated before 2ms and during the whole
time period deformation continues to grow and spreads up to actuating tubes. In 2ms, initial contact area is deformed
and whereas at 5ms some portion of the areas can be recovered, while the deformation spreads in other more areas.
At 12 ms and 13.5 ms, the front area shows slow deformation and the actuating tubes area starts deforming. If the
plastic equivalent strain is compared to von Misses stress distribution in the beam (fig.4), the plastic equivalent
strain occurred is less compared to the subjected stress. This is due to high flexural rigidity. Results show that

14
Simulation study of solid bumper beam under frontal impact load

initially the frontal area is mostly stressed. At 5ms it spreads to side areas of the beam which allows the initial
frontal part to recover. At 12ms and 13 ms the stress is transferred to its retention plates. Figure 5 shows the
variation of the strain energy with time and fig. 6 shows the variation of kinetic energy with time.

Fig. 5. Average strain energy variation in the beam

Fig. 6. Average kinetic energy variation in the beam

The initial increase in strain energy shows that the material that is initially in contact during this time period has no
molecular slip or any other form of energy dissipation, thus the material in contact starts absorbing the kinetic
energy. After an increase in strain energy, there is a sudden drop of strain energy, this indicates at that time period
the material no longer absorbs energy, the drop in this energy shows that there is molecular slip taking place
internally. The increase in strain energy again after the sudden drop shows that new unstrained molecules take part
in absorption of rest of the energy. As illustrated in fig. 6, during the whole period although strain energy fluctuates,
but the kinetic energy is steadily dissipated.

4. Conclusion
Frontal crash simulation of vehicle bumper beam was done to examine the deformation to the frontal area of the car
in order to reduce injury risk and potential of safety. It is observed that during the frontal impact to the solid bumper

15
 Proceedings of NCAMMM - 2018

beam, there will be sudden increase in strain energy absorption followed by reduced absorption. Different beam
segments have different energy absorption. Initially the contact area will have highest stress concentration.
Subsequently the adjacent areas will take up this stress promoting stress relief for previous affected areas and
deformation in the current stressed ones.

References
[1] What Are Car Bumpers Made of?. It Still Runs | Your Ultimate Older Auto Resource. https://itstillruns.com/car-
bumpers-made-of-6717239.html (retrieved on January 17, 2018).
[2] Baluch N, Udin ZM, Abdullah CS. Advanced high strength steel in auto industry: an overview. Engineering,
Technology & Applied Science Research. 2014 Jun 12;4(4):686-9.
[3] Thacker JG, Reagan SW, Pellettiere JA, Pilkey WD, Crandall JR, Sieveka EM. Experiences during
development of a dynamic crash response automobile model. Finite elements in analysis and design. 1998 Oct
15;30(4):279-95.
[4] Abdel-Nasser YA. Frontal crash simulation of vehicles against lighting columns using FEM. Alexandria
Engineering Journal. 2013 Sep 30;52(3):295-9.
[5] Fechová E, Kmec J, Vagaská A, Kozak D. Material Properties and Safety of Cars at Crash Tests. Procedia
Engineering. 2016 Dec 31;149:263-8.
[6] Ashtikar AS, Mali VP. Crash Analysis of Car Bumper Beam in Frontal Impact. International Engineering
Research Journal. 2016 Sep 27; 2(5):1856-1861.
[7] Huh H, Lim JH, Song JH, Lee KS, Lee YW, Han SS. Crashworthiness assessment of side impact of an auto-
body with 60TRIP steel for side members. Int. J. Automotive Technology. 2003 Sep 1;4(3):149-56.
[8] Takahashi M. High-strength steel sheets offering high impact energy-absorbing capacity. Nippon Steel
Technical Report. 2000 Jan;81(0).
[9] Honda The Power of Dreams. Honda; https://www.honda.com.my/model/specifications/civic (retrieved on
January 17, 2018).
[10] Car frame 3D model, https://www.cgtrader.com/3d-models/vehicle/part/car-frame (retrieved on January 17,
2018)
[11] AHSS Data Utilization – Autosteel, http://www.autosteel.org/research/ahss-data-utilization/trip780.aspx
(retrieved on January 17, 2018).

16
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

Optimization of CVD Parameters for Synthesis of Carbon Nanofibers on Inconel Substrate

Amit Thakur a,*, Alakesh Manna b, Sushant Samir c

a
Mechanical engineering Department, UIET, Panjab University, India
b
Mechanical engineering Department, Punjab Engineering College, India
c
Mechanical engineering Department, Punjab Engineering College, India

Abstract: The paper presents optimization of process parameters for direct synthesis of carbon nanofibers on
Inconel metal substrate on custom built external heating custom built chemical vapour deposition setup. The
selected carbon precursor and carrier gas for synthesis were acetylene and argon gas. Effect of variation of
temperature and time on morphology and yield i.e. growth of carbon nanofibers was found to be the highest. Low
synthesis temperature i.e. 650-700°C favors coiled growth of carbon nanofiber whereas high temperature favors
carbon nanofibers with straight morphology. The highest recorded yield of carbon nanofibers on optimized process
parameters was 385.4 mg after 45 min. of reaction time.

1. Introduction
Carbon and its allotropes such as carbon nanotubes (CNTs), carbon nano fibers (CNFs)and graphenehas been the
torch bearer in the field of nano technology. These nano structures are synthesized via different routes mainly arc
discharge, chemical vapour deposition (CVD) and laser ablation(1–3). The different selected methods have their
own advantages and disadvantages. Due to their inherent physical and chemical properties they have been reported
to increase mechanical and electrical properties many times over(4–6).The morphology of CNT/CNFs can be
controlled to certain extent by controlling the process parameters like temperature, carbon precursor, gas flow ratio
etc. The carbon nanofibers are ideal for application as reinforcement nano particles in nano composite as they are
more economical to be synthesized than carbon nanotubes in large quantities. Process of synthesis of CNT/CNF
utilize catalyst or chemically treated substrate for the favored growth (7,8).The CVD process involves introduction
of catalyst at the reaction site. This could be done by either of the following ways, mixing the catalyst with gaseous
or liquid carbon source, placing the catalyst in powdered form inside the rector, or a thin layer of catalyst is
deposited on the surface of substrate(9–11), which upon heating breaks down into small catalyst particles. All these
methods of CNFs synthesis requires purification of catalyst particles. This process of purification can damage the
morphology and surface of CNFs produced.To overcome this problem direct synthesis of CNTs/CNFs have been
investigated utilizing Inconel metal as substrate. Inconel being an alloy of transition metals favors growth of
CNTs/CNFs. The role of transition metals in synthesis of CNTs/CNFs is very important as they help in breakdown
of carbon precursor into hydrogen and carbon. The growth of CNTs/CNFs have been explained utilizing different
mechanisms such as Vapour-Liquid-Solid (VLS)(12),rate of dissolution (13), special velocity hodograph(14),
yarmulke mechanism (15).
From the review of the available literature, it is not identified that the optimization of the CVD process
parameters for synthesis of CNFs on metal substrate. Keeping in view, present work investigates the optimization of
CVD process parameters for high yield with homogenous growth of coiled CNFs on Inconel metal substrate.
Optimization of CVD parameters for synthesis of carbon nanofibers on Inconel substrate
2. Experimentation for Synthesis of Carbon Nano Fibers
Synthesis of CNFs was done utilizing an external heating custom build CVD setup. The schematic of developed
setup is shown in Fig. 1. Quartz tube with 50 mm internal diameter was used as reactor with single zone split type
external heater. Two mass flow controllers were used to measure and control the flow gases. Argon and acetylene
were used as carrier and precursor gas during experiments on developed CVD setup. The gas inlet pipe was heated
externally heated using coil heater to preheater the gases thus helping in dissociation of carbon precursor over metal
substrate surface. Flow of gases was directed on surface of substrate using 8 mm alumina tube is shown in Fig. 1.
Temperature was measured using four K type thermocouples and controlled using proportional–integral–derivative
controller (PID controllers).The substrate used was Inconel of dimension 10 mm x 13 mm x 1 mm. The substrate
was cut from 1mm sheet by wire electric discharge machine (WEDM). To remove oil and grease from the substrate
surface, substrate was sonicated in acetone for 45 min.

Fig. 1: Schematic diagram of fabricated external heating CVD

The metal substrate was aligned parallel to flow of carbon precursor and kept at a distance of 4-5 mm from outlet of
gas inside the rector is shown in Fig. 1.The growth of carbon on metal substrate was removed by scraping the
surface after cooling of substrate. The collected carbon growth was characterized using field emission microscope,
X-Ray Diffraction and Transmission electron microscopy. Before TEM, sample were ultra sonicated in ethanol for
two hours at 52oC. The parameters were selected based on review of literature and performance based on feasibility
experiments. The substrates were placed in the center quartz tube. The input process parameters considered are
represented in Table 1. The heating rate was kept constant at 9oC/min and surface of Inconel substrate was hand
polished with 100 mesh size sand paper. The L 18 (21 x 37) mixed orthogonal array was employed for experimental
planning and accordingly experiments were carried out. Table 2 represents the L 18 (21 x 37) mixed orthogonal array
with response i.e. yield of carbon nanofibers (CNFs, mg) and S/N ratio (dB). As the process has to be optimized for
larger yield, therefore larger the better relation has been used to calculate S/N ratio and study the effect of variation
of input process parameters on yield.

18
 Proceedings of NCAMMM - 2018

Table 1 Process Parameters and their levels used for Taguchi method based detail experiments

Levels
1 2 3
A:Temperature(oC) 650 700 750

B:Time (min) 15 30 45

C:Ratio of Ar: C 2 H 2 60:40 50:50 40:60

D:Gas Flow Rate (ml/min) 50 100 150

Table 2 L 18 (21x 37) mixed orthogonal array, parametric levels, experimental results and S/N ratio

Experiment Gas Flow rate S/N ratio

Temperature Time Gas Flow ratio Yield (mg)
No. (ml/min.) (dB)
1 1 1 1 1 31.432 29.9474
2 1 2 2 2 90.326 39.1163
3 1 3 3 3 140.82 42.9733
4 2 1 1 2 29.023 29.2548
5 2 2 2 3 265.174 48.4706
6 2 3 3 1 240.0862 47.6073
7 3 1 2 1 68.194 36.6749
8 3 2 3 2 395.593 51.9450
9 3 3 1 3 340.671 50.6467
10 1 1 3 3 45.673 33.1932
11 1 2 1 1 64.371 36.1738
12 1 3 2 2 102.831 40.2425
13 2 1 2 3 91.0032 39.1811
14 2 2 3 1 225.173 47.0503
15 2 3 1 2 270.142 48.6318
16 3 1 3 2 155.481 43.8335
17 3 2 1 3 370.0013 51.3641
18 3 3 2 1 174.531 44.8375
*where Yield is wt. of substrate before and after the experiment

3. Results and Discussion

Effect of input process parameters on growth i.e. yield of carbon nanofibers (mg) were plotted and are shown in Fig.
2. From S/N ratio graph, it is clear that the optimal parametric combination for higher growth of carbon nanofibers is
A 3 B 3 C 3 D 3.

19
Optimization of CVD parameters for synthesis of carbon nanofibers on Inconel substrate
48 48

46 46

Average S/N ratio (db)

44 44

42 42

40 40

S/N ratio for A

38 38
S/N ratio for B
S/N ratio for C
36 S/N ratio for D 36
Grand Mean of S/N ratios
34 34
A1 A2 A3 B1 B2 B3 C1 C2 C3 D1 D2 D3
Input Paramter Levels

Fig. 2: S/N Ratio (dB) for growth i.e. yield of carbon nanofibers

3.1 ANOVA for growth of carbon nanofibers

Table 3 represents the ANOVA for growth of carbon nanofibers. Table 3 also represents the ‘F-test’ value with %
contribution. From Table 5.6, it can be concluded that the time (B) is the most significant parameters with ‘F-test’
value 18.427 and 39.61% contribution. The temperature (parameter, A) is the significant parameter with F-test value
17.389 and 37.36% contribution. But other two parameters gas flow ratio and gas flow rate both cannot be ignored
as their ‘F-test’ value more than 2. However, error identified as 9.67 % that may be due to uncontrollable factor.

Table 3 ANOVA for growth of carbon nanofibers and ‘F-test’ with % contribution
Source DF SS Adj. MS F % Contribution

Temperature (A) 2 90136.18 45068.09 17.389 37.36

Time (B) 2 95520.92 47760.46 18.427 39.61

Gas Flow Ratio (C) 2 15361.54 7680.77 2.963 6.37

Gas Flow Rate (D) 2 16866.1 8433.052 3.253 6.99

Error 9 23325.65 2591.739 9.67

Total 17 241210.4

To study the morphology of synthesized CNFs, FESEM image analysis was done. Fig. 3 shows the images that
represents the growth i.e. yield of carbon nanofibers. Fig. 3 (a) shows the growth of carbon nanofibers when
experiment was carried out at 750oC temperature, 30 min time, 40:60 gas flow ratio and 100 ml/min gas flow rate.
Fig. 3 (b) shows the growth of carbon nanofibers when experiment was carried out at 750oC temperature, 30 min
time, 60:40 gas flow ratio and 150 ml/min gas flow rate. Fig. 3 (c) shows the growth of carbon nanofibers when
experiment was carried out at 750oC temperature, 45 min time, 60:40 gas flow ratio and 150 ml/min gas flow
rate.Fig. 3 (d) shows the growth of carbon nanofibers when experiment was carried out at 700 oC temperature, 45
min time, 60:40 gas flow ratio and 100 ml/min gas flow rate. Form Fig. 3, it is clear that all the growth is

20
 Proceedings of NCAMMM - 2018

homogenous with length of CNFs up to few tens of µm. The surface of fibers is smooth from outside with slightly
coiled structures inter twined with each other.

(a) (b)

(c) (d)
Fig. 3: FESEM images shows the growth of carbon nanofibers at different setting of parameter

The CNFs synthesized at high temperature i.e. 750oC are relatively straight but CNFs growth at temperature 700oC
is coiled in structure(16). Hence, it is concluded that the temperature (parameter, A) has great effect on morphology
of synthesized CNFs. A separate experiment at optimal parametric combination i.e. at A 3 B 3 C 3 D 3 with three
replications was carried out to identify the growth of carbon nanofibers. The yield i.e. growth was identified as
385.4 mg. Fig.4 shows the characterization of the carbon growth at optimum set of parameters i.e. A 3 B 3 C 3 D 3. From
Fig. 4 (a) the growth is straight with very few coiled nanofibers as the reaction temperature is 750oC. The Fig. 4 (b)
shows the TEM image for that sample. The Core of the fibers has parallel arrangement of graphite layers and are
more dense at the center. The conformation of CNFs in the sample was verified by carrying out an XRD of the
sample. The peak for CNFs only appears at 2θ = 25.7 degree corresponding to (0 0 2) plane and broad peak around
2θ= 43.8 degree corresponding to (1 0 0) plane with interlayer d-spacing of 3.47284 nm and 2.0622nm.

21
Optimization of CVD parameters for synthesis of carbon nanofibers on Inconel substrate

3500 3500
Pure CNF
3000 3000

2500 2500
Intensity

2000 2000

1500 1500

1000 1000

500 500

0 0
10 20 30 40 50 60
2 theta

Fig. 4: (a) FESEM image of CNFs, (b) TEM image of CNFs with internal structure and (c) XRD result

4. Conclusion
The growth of CNFs was successfully optimized for higher yield. The effect of temperature is highest on yield and
morphology of CNFs. Higher temperature favors CNFs with straight nanofibers and less of coiled CNFs, whereas
lower temperature i.e. 650oC and 700oC favors coiled nanofibers with fewer straight nanofibers. The contribution of
time and temperature were identified as 39.61% and 37.36 %, respectively.

Reference
[1] Zeng H, Zhu L, Hao G, Sheng R. Synthesis of various forms of carbon nanotubes by AC arc discharge.
Vol. 36, Carbon. 1998. p. 259–61.
[2] Kong J, Cassell AM, Dai H, Cassel AM, Dai H. Chemical vapor deposition of methane for single-walled
carbon nanotubes. Chem Phys Lett. 1998;292(August):567–74.
[3] Konishi H, Matsuoka H, Toyama N, Naitoh M, Nishigaki S, Kusunoki M. Growth control of carbon
nanotubes on silicon carbide surfaces using the laser irradiation effect. Thin Solid Films. 2004;464–
465:295–8.
[4] Lau K, Lu M, Liao K. Improved mechanical properties of coiled carbon nanotubes reinforced epoxy
nanocomposites. Compos Part A Appl Sci Manuf. 2006;37:1837–40.
[5] Verma P, Saini P, Malik RS, Choudhary V. Excellent electromagnetic interference shielding and
mechanical properties of high loading carbon-nanotubes/polymer composites designed using melt

22
 Proceedings of NCAMMM - 2018

recirculation equipped twin-screw extruder. Carbon N Y [Internet]. 2015;89:308–17. Available from:

http://linkinghub.elsevier.com/retrieve/pii/S0008622315002717
[6] Ramasubramaniam R, Chen J, Liu H. Homogeneous carbon nanotube/polymer composites for electrical
applications. Appl Phys Lett. 2003;83(14):2928–30.
[7] Yamada M, Kawana MA, Miyake M. Synthesis and diameter control of multi-walled carbon nanotubes
over gold nanoparticle catalysts. Appl Catal A Gen. 2006;302(2):201–7.
[8] Talapatra S, Kar S, Pal SK, Vajtai R, Ci L, Victor P, et al. Direct growth of aligned carbon nanotubes on
bulk metals: Nat Nanotechnol [Internet]. 2006;1(2):112–6. Available from:
http://www.nature.com/doifinder/10.1038/nnano.2006.56
[9] Angermann A, Topfer J. Synthesis of magnetite nanoparticles by thermal decomposition of ferrous oxalate
dihydrate. J Mater Sci. 2008;43:5123–30.
[10] Choi YC, Shin YM, Lim SC, Bae DJ, Lee YH, Lee BS, et al. Effect of surface morphology of Ni thin film
on the growth of aligned carbon nanotubes by microwave plasma-enhanced chemical vapor deposition. J
Appl Phys [Internet]. 2000; 88:4898. Available from:
http://link.aip.org/link/JAPIAU/v88/i8/p4898/s1&Agg=doi
[11] Suda Y, Maruyama K, Iida T, Takikawa H, Ue H, Shimizu K, et al. High-Yield Synthesis of Helical
Carbon Nanofibers Using Iron Oxide Fine Powder as a Catalyst. Crystals [Internet]. 2015;5(1):47–60.
Available from: http://www.mdpi.com/2073-4352/5/1/47/
[12] Wagner RS, Ellis WC. Vapor-liquid-solid mechanism of single crystal growth. Appl Phys Lett.
1964;4(5):89–90.
[13] Baker RTK, Barber MA, Harris PS, Feates FS, Waite RJ. Nucleation and Growth of Carbon Deposits from
the Nickel Catalyzed Decomposition of Acetylene. J Catal. 1972;26:51–62.
[14] Amelinckx, S., Zhang B. X., Bernaetrs D., Zhang F. X., Ivanov V. NBJ. A formation mechanism for
catalytically grown helix-shaped graphite nanotubes. Science. 1994;265:635–9.
[15] Dai H, Rinzler AG, Nikolaev P, Thess A, Colbert DT, Smalley RE. Single-wall nanotubes produced by
metal-catalyzed disproportionation of carbon monoxide. Chem Phys Lett. 1996;260(3–4):471–5.
[16] Thakur A, Manna A, Samir S. Direct growth of coiled carbon nanofibers without nanocatalyst. Diam Relat
Mater [Internet]. 2017 Apr;74:100–7. Available from:
http://linkinghub.elsevier.com/retrieve/pii/S0925963516305416

23
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

Microstructural Modifications of Cast Al-17Si-5Cu Alloy through Isothermal Heat

Treatment

B Hazra1, P Baranwal1, S Bera1 and B K Show1,2

1
Department of Metallurgical and Materials Engineering, National Institute of Technology Durgapur, West
Bengal,713209, INDIA; 2Corresponding author

Abstract: In the present investigation, isothermal semi-solid heat treatment is employed in an as-cast Al-17Si-5Cu
alloy (AR alloy) to modify the morphology of Si particles (particularly eutectic Si particles having a high aspect
ratio in as-cast condition). To establish the best heat treatment cycle for this alloy, different soaking times of 10, 15
and 20 minutes are used as isothermal holding times at 590°C temperature (as the solidus and liquidus
temperatures are 577 ° C and 648°C, respectively) followed by cooling to room temperature. Three different cooling
media (viz. furnace, air, water) are employed to study the effect of various cooling rates. Detailed microstructural
investigations involving size and shape of primary as well as eutectic Si particles, porosity measurements along with
hardness measurement are carried out to find the effect of different isothermal heat treatments. The study reveals
considerable improvement in microstructure in all the heat treated samples. Accordingly, hardness values in heat-
treated samples improve significantly as compared to AR alloy. However, isothermal heat treatment involving 15
minutes soaking time at 590°C temperature followed by water (at 60°C) quenching (WQ15 alloy) is found to
produce optimum combination of microstructure and properties.

1. Introduction
Al-Si alloys have widespread application in cast form in critical components of automobile application owing to
their excellent combination of properties[1]. It accounts for more than 90 % of the total Al castings produced.
However, recently, hypereutectic Al-Si alloys have drawn the attention of researchers due to their attractive
properties. Their properties are primarily dependent on the size, shape, and distribution of primary as well as
eutectic silicon particles in the alloy. Mechanical properties of these alloys can be improved by the combined effect
of refinement and modification of the primary and eutectic silicon. The as-cast hypereutectic Al-Si alloy exhibits a
microstructure consisting of primary Si particles along with α-Al grains and eutectic phase. However, the Si
particles in eutectic phase generally have long rod/acicular morphology with high aspect ratio.
Wear behaviour of cast Al-Si alloys needs to be evaluated for their successful use. Among different Al-Si alloys,
hypereutectic Al-Si alloys are found to be used in engine components due to their good wear resistance properties.
Hypereutectic A390 alloy which was first introduced at the AFS Casting Congress [2], used as an engine block
material in the linerless Chevrolet Vega 2300 engine. Thereafter, many commercial solutions were reported [3, 4]
and successfully employed. However, costly production method restricted their use in all vehicles. Si, in Al-Si alloys
is mainly responsible for goodwear resistance. Increasing the silicon content in Al-Si alloys increases not only the
wear resistance but also the strength of the alloy [5]. However, the improvement in strength and wear resistance
comes at the cost of machinability and castability.
Microstructural modifications of cast Al-17Si-5Cu alloy through isothermal heat treatment

Thus the hypereutectic alloys are generally treated with various modifier and grain refiners(like phosphorus) to
modify primary as well as eutectic silicon. Application of isothermal semi-solid heat treatment to modify Si particles
has not been reported by many. Naglaa Fathy [6] studied microstructural evolution of hyper-eutectic Al-18% Si
alloy during semi-Solid isothermal heat treatment. In this research the effect of semi-solid isothermal heat treatment
on the size and shape of primary Si and α-Al grain has been studied for a hyper-eutectic Al-18% Si alloy. However
the detailed study was not taken up. Therefore the objective of the present study is to carry out a systematic study on
the effect of various soaking times and cooling rates on the morphology, distribution, and size of primary as well as
eutectic Si and finally optimize heat treatment parameters. Accordingly, semi-solid heat treatment is employed in an
as-cast Al-17Si-5Cu alloy (AR alloy) to modify the morphology of Si particles. Detailed microstructural
investigations involving size and shape of primary as well as eutectic Si particles, porosity measurements along with
hardness measurement are carried out to find the effect of different semi-solid heat treatments involving varying
soaking times and cooling rates.

2. Experimental Procedure
2.1 Material
As-received cast hypereutectic Al-17Si-5Cu alloy (AR alloy) was used for the present study. Standard samples for
various tests were machined from the as-cast cylindrical block of approximately 25 mm diameter. Samples of size
15 mm X 15 mm were prepared for optical metallography and hardness measurement. One set of the sample was
retained for the further characterization in the as-received condition & others were subjected to different semisolid
heat treatments.
2.2 Heat Treatment
The liquids temperature of the alloy is 648ºC and the solidus temperature for the same is 577ºC as per standard
phase diagram of Al-Si alloy. Therefore, a temperature of 590ºC was chosen for semi-solid heat treatment (Figure
1). The specimens were heated at 590ºC for different soaking times of 10, 15 and 20 minutes in a muffle furnace
followed by cooling to room temperature. Three different cooling media were employed to study the effect of
different cooling rates. One set of samples were cooled in a furnace, while the other set was subjected to air cooling
and water quenching (warm water at 60°C) was carried out for the last set of samples (Figure 1). Accordingly, the
samples were designated as FC, AC, and WQ for furnace cooling, air cooling, and water quenching respectively.
The different heat treated alloys were thus designated as per Table 1.
2.3 Optical Metallography
Small samples were machined out for metallographic investigation.They were then polished using standard
procedure. The polished surfaces were rinsed in acetone, dried by the hot air blast and finally etched with Keller's
reagent. Both qualitative and quantitative studies of the microstructures were carried out using an optical microscope
(Leica DM 2500 M) with digital photomicrography and image analysis (Leica MW image analyzer). The grain size
was measured by linear intercept method as per ASTM E112 standard [7].Besides, the sizes of the primary particles
were determined(in terms of mean ± standard deviation) from the micrographs considering ten image frames. For
each particle, the size was considered as an average of the major axis and minor axis approximating the particle
shape to be an ellipse in two dimensions. Furthermore, the aspect ratio was also calculated from the same

25
 Proceedings of NCAMMM - 2018

measurement by taking the ratio of the major axis and minor axis for different particles (in terms of mean ± standard
deviation).
2.4 Hardness Test
The hardness of as-cast alloy and the heat treated specimens was measured in a Vicker’s hardness testing machine
(BV 250 (S), BIE, Miraj, India) using 20kgf load. For a particular cast alloy or heat treated alloy, three specimens
were tested (with five hardness readings per specimen) and the mean value along with standard deviation of the
hardness was reported.
Table 1 Different heat-treated alloy and their
designations
Soaking Time Alloy
(min.) Designation
Cooling Mode
Furnace cooling 10 FC 10
(FC) 15 FC 15

10 AC 10
Air cooling (AC) 15 AC 15
20 AC 20

10 WQ 10
Water Quenching 15 WQ 15
(WQ)
10 WQ 20
Figure 1: Semi-solid heat treatment schedule.

3. Result and Discussion

3.1Microstrucural characterization of as-cast alloy
The optical microstructure of as-cast alloy (Figure 2(a)) exhibits the presence of primary Si particles along with α-Al
grains and eutectic phase. The average size of the primary Si particles is 17±1.7 with an aspect ratio of 1.5±0.26
(Table 2). However, Si particles (having size 8.±1.2 µm) of eutectic phase are having long rod/acicular morphology
with an aspect ratio of 7±1.3 (Table 2), as opposed to the primary Si particles which are nearly equiaxed/spherical
(Figure 2 (a)).Eutectic Si particles are found to possess a variety of morphologies and are randomly distributed. The
α-Al grains in as-cast alloy possess variation in size with an average grain size of 26±6 µm.
3.2 Microstructural characterization of furnace cooled alloy
Figures 2(b) and (c) resemble the presence of very coarse primary Si particles (mean particle size of 32±9 and
32±7µm respectively for 10 min and 15 min soaking time) in FC10 and FC15 alloy.In addition, eutectic Si phase
also exhibits the presence of coarser Si particles having size 17±7 and 15±5 µmfor FC10 and FC15 alloy
respectively. Furthermore, the aspect ratio of eutectic Si was not reduced to a considerable extent (Table 2).
Moreover, the furnace cooled sample exhibit more numbers of merged particles (Figure 2 (b) and (c)) as compared
to air-cooled sample (to be discussed next). The slowest rate of cooling in the furnace is held responsible for
this.The %porosity contents are 10 % and 8% (Table 3) respectively for FC10 and FC15 alloys, which are high in
amount. This is attributed to volumetric contraction which resulted from solidification of liquid formed during the
semi-solid temperature. In case of furnace cooling, the last liquid to solidify will be in the central zone. Therefore

26
Microstructural modifications of cast Al-17Si-5Cu alloy through isothermal heat treatment

the major porosities were located at the central region of these samples. Hence soaking time of 20 min was not given
in furnace cooling.
3.3 Microstructural characterization of Air cooled alloy
The optical microstructures after isothermal semi-solid heat treatment of AR alloy at 590°C followed by air cooling
for three different soaking times are shown in Figures 2(d)-(f). It can be seen from Table 2 and Figures 2 (d) and (e)
that soaking times up to 15 min. result into lowering of aspect ratio for both primary as well as eutectic Si particles.
On the other hand, there is a marginal increase in sizes of both primary as well as eutectic Si particles (Table 2).
However, with a soaking time 20 min. A considerable decrease in aspect ratio for both primary as well as eutectic Si
particle is observed (Figure 2 (f)). Furthermore, the eutectic Si particle, as well as primary Si particle, is found to be
refined for AC20 alloy (Figure 2(f), and Table 2). On the other hand, for AC20 alloy, primary Si particles are found
to be merged into bigger particles at few locations as shown in Figure 2(f). Relatively slower cooling rate compared
to water quenching (to be discussed later) may lead to merging of primary Si particle due to more diffusion time.
3.4 Microstructural characterization of warm water quenched alloy
On the other hand, water quenching after isothermal holding results into considerable improvement in the size as
well as the aspect ratio for both primary and eutectic Si. Figures 2(g)-(i) depict the optical microstructures of alloys
quenched in water after soaking at 5900C temperature for 10, 15, 20 min. respectively. Eutectic Si starts to assume
equiaxed/spherical morphology (aspect ratio: 1.4±0.21) only at 15 min. soaking time in the WQ15 alloy. In contrast
to this, eutectic particle shape modification (AR 1.6±0.35) starts at 20 min. soaking time for AC20 alloy. Moreover,
the eutectic Si particles are subjected to considerable variation in size in case of AC20 alloy, though the particle size
is finer compared to WQ 15 sample. The optical microstructure of WQ15 sample (Figure 2(h)) clearly shows the
uniform distribution of particles (both primary as well as eutectic Si) in the α-Al. On further increasing the soaking
time to 20 min in WQ 20 sample, particles start becoming coarser (both primary as well as eutectic Si). Few
locations are also identified (Figure 2(i)), where particles are found to merge with each other. However, the merging
tendency is less compared to AC 20 sample.
3.5 Microstructural characterization-a comparative study
The morphology of Si particles has been modified on isothermal heat treatment, particularly for eutectic Si.
Hardness values are also (Table 3) improved on heat treatment except in FC10 alloy. The increase in hardness is
attributed to faster cooling rates in AC and WQ samples. The matrix retains more solute on faster cooling resulting
in solid solution strengthening. Table 3 also exhibits the %porosity content in AR and different heat treated alloys. It
is clear that furnace cooling results into a high amount of porosity since volumetric contraction on solidification is
not substituted by the surrounding liquid here as last liquid solidifies at the center (since cooling rate is slow here).
From Table 2 it can be seen that there is marginal improvement in size and shape of primary Si particles upon
isothermal heat treatment. On the other hand, there is considerable improvement in aspect ratio from a value of
7±1.3 in AR alloy to a value of 1.4±0.2 in WQ15 alloy and 1.6±0.4in AC20 alloy after isothermal heat treatment.

27
 Proceedings of NCAMMM - 2018

Figure 2: Optical microstructures: (a) AR alloy; (b) FC10 alloy; (c) FC15 alloy; (d) AC10 alloy; (e) AC15 alloy; (f)
AC20 alloy; (g) WQ10 alloy; (h) WQ15 alloy and (i) WQ20 alloy.
Table 2 Mean particle size and aspect ratio of primary and eutectic Si particles in as cast and different heat treated
alloys
Condition Si particle Soaking time(min) Average AR Mean particle size(µm)
As cast sample Primary - 1.5±0.3 17±1.7
Eutectic - 7±1.3 8±1.2
Air cooled Primary 10 1.3±0.24 19±6
sample 15 1.3±0.22 21±5
20 1.2±0.04 17±4
Eutectic 10 5±1.3 10±3
15 6±1.3 8±1.5
20 1.6±0.35 5±1.5
Water Quenched Primary 10 1.5±0.35 17±4
sample 15 1.4±0.2 15±8
20 1.4±0.24 16±6
Eutectic 10 3.9± 1.8 10±3
15 1.4± 0.21 7±1.0
20 1.5±0.3 8±2
Furnace cooling Primary 10 1.4±0.32 32±9
sample 15 1.4±0.30 32±7
Eutectic 10 3±1.6 17±7
15 3±1.7 15±5

In addition, there is a reduction in eutectic Si Table 3 Bulk hardness and %porosity values for AR and

28
Microstructural modifications of cast Al-17Si-5Cu alloy through isothermal heat treatment

particle size in both the cases (in case of AC20 different heat treated alloys.
particles are slightly finer than WQ15). However Alloy Hardness (HVN) % Porosity
α-Al grain size is found to increase with soaking AR alloy 111 ± 4.5 1.59
AC 10 120±4.2 3.43
time. Thus WQ 15 is selected as the optimum AC 15 154±7.1 2.45
heat treatment schedule with a soaking time of 15 AC 20 141±3.7 3.25
WQ 10 119±4.7 2.11
minutes followed by water quenching. WQ 15 134.6±9.8 3.30
WQ 20 120±9.2 4.77
FC 10 94.6±7.7 9.52
FC 15 114±7.7 7.52
4.Conclusion
As cast Al-17Si-5Cu alloy possesses microstructure consisting of primary Si particles (size = 17 ± 1.7 µm, AR= 1.5
± 0.26) and acicular/rod-like eutectic Si particles (size = 8 ± 1.2 µm, AR= 7 ± 1.3) in a matrix of α-Al grains.
Furnace cooling after isothermal holding results in particle coarsening, particle merging along with increased level
of porosity. Air cooling, on the other hand, provides considerable improvement in AC20 alloy. But, for AC20 alloy,
primary Si particles are found to be merged into bigger particles at few locations as shown in the micrograph due to
longer heating time. On the other hand, WQ15 alloy exhibits considerable microstructural modifications particularly
for eutectic Si particles (eutectic Si having AR= 1.4 ± 0.21 are obtained after heat treatment). Accordingly, hardness
values increase for all AC and WQ samples. However, WQ15 alloy can be considered to be optimum because of its
considerable modification in microstructure and improvement in hardness value.

Acknowledgment
Authors acknowledge DST-SERB grant, vide Project No. YSS/2014/000172 dated 17-08-2015.

References
[1] Vijeesh V, Prabhu KN. Review of microstructure evolution in hypereutectic Al–Si alloys and its effect on wear properties.
Transactions of the Indian Institute of Metals. 2014 Feb 1;67(1):1-8.
[2] Jorstad JL. The hypereutectic aluminum-silicon alloy used to cast the Vega engine block. Modern Casting. 1971
Oct;60(4):59-64.
[3] Jorstad JL. 09-152 Silver Anniversary Paper-The Progress of 390 Alloy: From Inception until Now. Transactions of the
American Foundrymen's Society. 2009;117:241.
[4] Kainer KU, editor. Metal matrix composites: custom-made materials for automotive and aerospace engineering. John Wiley
& Sons; 2006 Aug 21.
[5] E Erginer, The strengthening of aluminum due to its cast microstructure modified by silicon, Ph.D. Dissertation, Brown
University. (1969) 206.
[6]. Fathy N. Microstructural Evolution of Hyper-Eutectic Al-18% Si Alloy during Semi-Solid Isothermal Heat Treatment.
International Conference on Research in Science, Engineering and Technology (ICRSET’2013) Nov 2013 Nov (pp. 13-14).
[7] Standard AS. E112, 2010," Standard Test Methods for Determining Average Grain Size," ASTM International, West
Conshohocken, PA, 2010, DOI: 10.1520/E0112-10.

29
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

Study the Effects of Bi Addition on Microstructure and Hardness of Hypereutectic Al-

17.6Si Alloy

Kona Durgaprasadu, Prosanta Biswas and Manas Kumar Mondal*

Department of Metallurgical and Materials Engineering, National Institute of Technology, Durgapur-713209
Email: manas.nitdgp@gmail.com

Abstract: In the present research, hypereutectic Al-17.6Si alloy with and without bismuth (Bi) has been developed
through gravity casting method using commercially pure Al (99.71%), Si (99.28%) and Bi (99.90 %). Effects of Bi
concentration (0, 0.5, 1.0, 1.5 and 2.0 wt.%) on the microstructure and hardness of the Al-17.6Si alloy have been
investigated. The cast unmodified hypereutectic Al-17.6 Si alloy has dendritic irregular polygonal and plate like
shape coarse primary Si particles with very sharp corner and a needle and rod-like eutectic Si particles. Bi addition
effectively changed the shape of the primary Si into fine pentagonal shape particles with blunt corners and the
primary Si particles average size and aspect ratio and volume fraction are decreased with an increase in Bi
concentration in the Al-17.6Si alloys. The bulk hardness is increased with increase in Bi concentration in the Al-
17.6Si alloys as an effect of microstructural modification.
Keywords: Hypereutectic Al-Si alloy; Bismuth; Microstructure; Hardness

1. Introduction
Al-Si alloys possess outstanding properties such as high strength to weight ratio, excellent castability, weldability,
corrosion resistance and wear resistance, low thermal expansion coefficient, and better high-temperature strength.
Because of this, Al-Si alloys are extensively used in aerospace and automotive industries [1-3]. In recent time, the
hypereutectic Al-Si alloy is mainly used to manufacture piston because it has a potential to replace the cast iron [4].
Generally, the microstructure of the hypereutectic Al-Si alloy consists of α-Al, coarse primary Si (Si P ) particles and
eutectic Si (Si E ). However, the mechanical properties of the Al-Si alloys generally depend on the morphology and
volume fraction of the containing phases. Coarse, needle and plate-like morphologies of Si P and Si E deteriorate the
mechanical properties of that alloy. Therefore, fine Si P and Si E particles with uniform distribution are desirable to
achieve better mechanical properties. Mechanical properties of the Al-Si alloy can be improved by modification and
refinement of the Si P and Si E particles [5, 6]. Different methods have been employed to modify or refine the Si P and
Si E in the Al-Si alloys such as rapid solidification technique, semi-solid processing, electromagnetic stirring [7-9]
and chemical modification. Chemical modification is popular, very effective and simplest method and generally
execute through simple conventional casting. Previously, few researcher [10, 11] were investigated the effect of
various elements as modifier and refiner in Al-Si alloy. In a previous study, Weixim et al. [10] investigated the
effects of Nd on the primary Si particles of hypereutectic Al-15 wt.% Si alloy and it was reported that the pure Nd
addition strongly refine the Si P particles as an effect of this the ultimate tensile strength (UTS) increased about
32.6% and the elongation increased about 160%. In another recent study, Li et al.[11] studied the effect of rare earth
(Er) addition on the hypereutectic Al- 20 wt. % Si alloy and it was found that the Si P particles are refined
significantly and platelet-like and star-like polygonal shape transformed to blocky shape, whereas the Si E structure
Study the effects of Bi addition on microstructure and hardness of hypereutectic Al-17.6Si Alloy

is modified into the fine coral like fibrous from coarse platelet-like/needle-like structure. Because of Si P and Si E
modification, the UTS and elongation increased by 72.5% and 72%, respectively.
There are many works have been carried out on the modification and refinement of Al-Si alloy. But, still
many other elements are left to investigate. Among those, one of the element is bismuth (Bi). In the present
investigation, hypereutectic Al-17.6Si alloy has been developed through gravity casting method with and without Bi
addition to understanding the effects of Bi on the microstructural morphology and hardness properties of the Al-
17.6Si alloy. Finally, a correlation has been established between microstructural morphology and hardness of the
developed alloys.

2 Experimental Procedure
The hypereutectic Al-17.6Si-X wt.%Bi ( X=0.0,0.5,1.0,1.5 and 2.0) alloy was developed through gravity casting
method. Commercially pure Al (99.5%), Si (99.8%) and Bi (99.99%) were used as starting materials. Initially, the
commercially pure Al was melted at 760oC in a clay graphite crucible using an electrical resistance furnace. After
melting of Al, Si granules were added and held for 50-60 minutes for complete Si dissolution. Then, aluminium foil
wrapped Bi was added into the melt at 760oC as per requirement and the molten mixture was stirred manually for 2-
3 minutes using a graphite rod and held for 5 minutes. After that degassing was performed by adding
hexachloroethane (C 2 Cl 6 ). Then, slag was removed using a stainless steel spoon and molten metal was poured into a
preheated (160oC for 1hrs) permanent mould, designed as per BS1490 standard [12].
After complete solidification, optical metallography and hardness samples with an approximate dimension of 10 mm
× 10 mm × 10 mm were machined from the middle portion of the as-cast billets. The standard procedure and
guidelines were followed during metallography samples preparation and the polished samples were etched with
Keller’s reagent (1% HF, 1.5% HCl, 2.5% HNO 3 and distilled water). The Leica optical microscope was used for
optical microscopy and the hardness test was performed by the Vicker’s hardness testing machine with standard
procedure (load 10kg and dwell time 20 sec) and five different values were taken at different locations on each
specimen. The size of primary Si particles and volume percentage (serologically equal area percentage) of different
phases were measured from 10 optical images of every developed alloy using the ImageJ image analysis software.

3 Results and Discussions

3.1 Microstructure evaluation
Fig.1 shows the optical micrograph of the hypereutectic Al-17.6Si alloy with and without bismuth addition. It
reveals that microstructural morphology of the hypereutectic Al-17.6Si alloy is changed due to change in Bi
concentration. The optical microstructure of unmodified (0.0 wt.% Bi) Al-17.6Si alloy exhibits the presence of
coarse dendritic irregular polygonal and plate-like Si P particles with very sharp corner and the average Si P particles
size is about 49 µm and the Si E phase has both the needle and rod-like morphology (Fig.1(a)).
Figure 1(b) displays the optical microstructure of 0.5wt.% Bi added alloy and it contains mainly plate-like
and polygonal Si P particles with an average size of 44 µm and small rod-like eutectic Si.

31
 Proceedings of NCAMMM - 2018

Fig.1 Microstructure of hypereutectic Al-17.6Si-Xwt.%Bi alloy (a) 0.0 wt.% Bi (b) 0.5 wt.% Bi (c) 1.0 wt.% Bi (d)
1.5 wt.% Bi and (e)2.0 wt.% Bi addition
Therefore, 0.5 wt.% Bi addition in the alloy doesn’t have a significant impact, the dendritic nature of the Si P
particles got changed and the length of the particles increased and the Si E phase structure is more or less similar to
the unmodified Al-17.6Si alloy. The optical microstructure of 1.0 wt.% Bi modified alloy (Fig.1(c)) shows that most
of the sharp corners of Si P particles are changes to blunt corners and average size is reduced to 40 µm and length of
the Si P particles are less compare to the 0.5 wt.% Bi added alloy but the eutectic Si phase not changed significantly.
Fig.1(d) and Fig.1(e) shows the microstructures of 1.5 wt.% Bi and 2 wt.% Bi addition. The addition of 1.5 wt.%
Bi, the shape of primary Si particles is changed into fine pentagonal shape with effectively blunt corners, but the
eutectic Si remains unchanged and the average size of primary Si particles is decreased to 27 µm. This changed is
archived by the Bi addition. A growth twin will occur at the interface when the ratio of the atomic radius of
modification elements to that of Si is close to 1.646 [13]. For Bi, r Bi /r Si is 1.39 closed to the values 1.646. Therefore,
Bi may be absorbed by the growth steps of Si and promote multiple twining to cause structural modification.
Further, increase in Bi concentration to 2 wt.%, the eutectic Si is fragmented, but some of the primary Si particles
are started to over modify and the average size of the primary Si particles decreases to 24 µm.
Figure 2 shows that the primary Si particles size is significantly decreased with increase in Bi concentration
in in the Al-17.6Si alloy. Further, Fig. 3 depicts the typical curve of the aspect ratio of primary Si particles as a
function of Bi concentration in the Al-17.6Si alloy.

32
Study the effects of Bi addition on microstructure and hardness of hypereutectic Al-17.6Si Alloy

Fig. 2 Primary Si particles size as a function of Bi concentration in the Al-17.6Si alloy

It is found that the aspect ratio is initially increased at 0.5 wt.% of Bi addition and then decreases with
increase in Bi concentration in the alloys due to the transformation of irregular polygonal and plate-like primary Si
changes particles to fine pentagonal shape.

Fig. 3 The aspect ratio of the primary Si particles as a function of Bi concentration in the Al-17.6Si alloys

Fig. 4 The primary Si and eutectic Si phase volume percentage as a function of Bi concentration in the Al-17.6Si
alloys

Furthermore, Bi addition restrict the growth of Si by absorbing the surface energy of Si due to this the Si became
more compact and finer. Therefore, the volume fraction of both the primary and eutectic Si phase is changed with Bi
concentration in the alloy. The primary Si phase volume is decreased with increase in Bi concentration in the Al-

33
 Proceedings of NCAMMM - 2018

17.6Si alloy as the Si became more compact in modified alloys because Bi addition restricts the growth of Si. The
eutectic phase volume fraction in Bi modified alloys is also low compared to the unmodified one as shown in Fig. 4.

3.2 Hardness
The microstructure morphology and volume fraction of phases have a significant impact on the bulk hardness of the
Al-Si alloy. Figure 5 shows the bulk hardness (H) profile as a function of Bi concentration in the Al-17.6Si alloy,
which reveals that the bulk hardness is increased with increase in Bi concentration in the Al-17.6Si alloys. This
changed in bulk hardness is the cause of changes in microstructural morphology of the primary Si particles. As the
sharp corners of the primary Si particles are became blunt resistance to plastic deformation is increased. Also, the
compactness of the modified alloy Si particles has a significant impact on the increase in bulk hardness values.

Fig. 5 Bulk hardness of the Al-17.6Si alloy as a function of Bi concentration.

4. Conclusion
Hypereutectic Al-17.6Si alloy with and without Bi addition was successfully developed through gravity casting
method and effects of Bi on the microstructural morphology and bulk hardness has been studied and following
conclusion are drawn:
• The cast unmodified hypereutectic Al-17.6 Si alloy consisting of dendritic irregular polygonal and platelike
shape coarse primary Si particles with very sharp corner and a needle and rod-like eutectic Si particles.
• In addition to 1.5 wt% of Bi, the shape of the primary Si particles is changed into fine pentagonal shape
particles with effectively blunt corners.
• The primary Si particles average size and aspect ratio are decreased but with an increase in Bi
concentration in the Al-17.6Si alloys.
• The volume fraction of primary Si particles is decreased with an increase in Bi concentration in the Al-
17.6Si alloys. The eutectic Si phase volume fraction is lower in Bi modified Al-17.6Si alloys.
• The bulk hardness of the modified alloy is more as compared to the unmodified alloy and the bulk hardness
is increased with increase in Bi concentration in the Al-17.6Si alloys.

Acknowledgements
The authors are like to thanks, NIT, Durgapur RIG # 2 project for financial support and the Director of National
Institute of Technology Durgapur, India for his continuous encouragement.

34
 Study the effects of Bi addition on microstructure and hardness of hypereutectic Al-17.6Si Alloy

References
[1] Basavakumar KG, Mukunda PG, Chakraborty M. Influence of grain refinement and modification on
microstructure and mechanical properties of Al–7Si and Al–7Si–2.5 Cu cast alloys. Materials Characterization.
2008 Mar 31;59(3):283-9.
[2] Joy-Yii SL, Kurniawan D. Effect of rare earth addition on microstructure and mechanical properties of Al-Si
alloys: an overview. In Advanced Materials Research 2014 (Vol. 845, pp. 27-30). Trans Tech Publications.
[3] Raghukiran N, Kumar R. Effect of scandium addition on the microstructure, mechanical and wear properties of
the spray formed hypereutectic aluminium-silicon alloys. Materials Science and Engineering: A. 2015 Aug
12;641:138-47.
[4] Vijeesh V, Prabhu KN. Review of microstructure evolution in hypereutectic Al-Si alloys and its effect on wear
properties. Transactions of the Indian Institute of Metals. 2014 Feb 1;67(1):1-8.
[5] Liu G, Li G, Cai A, Chen Z. The influence of Strontium addition on wear properties of Al–20wt% Si alloys
under dry reciprocating sliding condition. Materials & Design. 2011 Jan 31;32(1):121-6.
[6] Li Q, Xia T, Lan Y, Zhao W, Fan L, Li P. Effect of rare earth cerium addition on the microstructure and tensile
properties of hypereutectic Al–20% Si alloy. Journal of Alloys and Compounds. 2013 Jun 15;562:25-32.
[7] Lu D, Jiang Y, Guan G, Zhou R, Li Z, Zhou R. Refinement of primary Si in hypereutectic Al-Si alloy by
electromagnetic stirring. Journal of Materials Processing Technology. 2007 Jul 6;189(1):13-8.
[8] Jung HK, Seo PK, Kang CG. Microstructural characteristics and mechanical properties of hypo-eutectic and
hyper-eutectic Al-Si alloys in the semi-solid forming process. Journal of materials processing Technology.
2001 Jun 15;113(1):568-73.
[9] Feng HK, Yu SR, Li YL, Gong LY. Effect of ultrasonic treatment on microstructures of hypereutectic Al-Si
alloy. Journal of materials processing technology. 2008 Nov 21;208(1):330-5.
[10] Weixi SH, Bo GA, Ganfeng TU, Shiwei LI, Yi HA, Fuxiao YU. Effect of neodymium on primary silicon and
mechanical properties of hypereutectic Al-15% Si alloy. Journal of rare earths. 2010 Dec 1;28:367-70.
[11] Li Q, Xia T, Lan Y, Li P, Fan L. Effects of rare earth Er addition on microstructure and mechanical properties
of hypereutectic Al–20% Si alloy. Materials Science and Engineering: A. 2013 Dec 20;588:97-102.
[12] Biswas P, Mondal MK, Roy H, Mandal D. Microstructural evolution and hardness property of in situ Al–Mg2Si
composites using one-step gravity casting method. Canadian Metallurgical Quarterly. 2017 Jul 3;56 (3):340-8.
[13] Lu SZ, Hellawell A. The mechanism of silicon modification in aluminium-silicon alloys: impurity induced
twinning. Metallurgical Transactions A. 1987 Oct 1;18(10):1721-33.

35
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

Effect of Space Holder Size on Microstructure, Deformation and Corrosion Response of

Ti4Al4Co Alloy Foam

Pradeep Singh, Hemant Jain, Prashant Nair, Anup Khare, I.B. Singh, D.P. Mondal
CSIR- Advanced Materials and Processes Research Institute

Abstract: Ti4Al4Co alloy foam was made using mechanically alloyed powder and elemental powder. The space
holder size is varied in three different average sizes: 80 µm, 175 µm and 246 µm. t is noted that the average pore
sizes are 54 µm , 128 µm and 196 µm respectively for space holder size of 80 µm, 175 µm and 246 µm respectively.
It further noted that the pore size reduced more when one used elementary powder. In case of elementary powder,
the pore size is noted to be 180 µm when one used space holder of 246 µm. Finer space holder exhibited greater
relative densities and less porosities. Because of this fact, foam made using finer space holders exhibited higher
plastic collapse stress or yield stress. The foam made with elementary powder exhibited higher strength because of
the same reason. The foam made with coarser space holder size or with coarser pore size exhibited less corrosion
rate as compare to that one with finer pore size or space holder size. The Foam made with unmilled powder
exhibited higher corrosion rate as compared to that one made with milled powder.

1. Introduction
Ti and its alloys are unique materials for bone replacement due to their favourable properties like light weight with
high strength, biocompatibility and excellent corrosion resistance[1, 2]. However, some problems remain persist
with the Ti alloy bone replica. High elastic modulus of Ti alloy in comparison to the natural bone is one of the major
problem that cause stress shielding effect, consequent bone resorption and failure of implant [3, 4]. Smooth surface
of Ti alloy implant also cause the problem in osteointegration with fibre like tissues of the bone resulting implant
loosening [5].Also, relative movement of implant against natural bone in the medium of body fluid produces
liberation of elemental ions and debris in the body that cause detrimental effect on the human health [6].Ti alloys
foam implant reduce the chance of stress shielding effect as well as improve the oseointegration between implant
and the natural bone due to anchoring the bone tissues in the pores of the implant. An excellent method to tailor the
density of the foam are space holder technique. Mechanical alloying is a noble method for the grain refinement,
uniform mixing and different phase formation at comparatively lower temperature [7]. The refinement of powder
particles is caused because of the fracturing, re-welding and re-fracturing processes due to impact among the powder
particles and balls[8].In the present study, milled Ti4wt%Al4wt%Co powder was used as the base material for the
manufacturing of foam. Ammonium bi carbonate of different particulate size was used as a space holder.

2. Experimental Procedure
.Ti powder (99.9% pure) of irregular shape with average particle size 28µm, 4 wt% circular shaped Al (99.8% pure
and particle size < 20µm) and 4 wt% Co (99.5% pure and average particle size 36 µm) were taken as initial
materials. The constituents materials were mixed with stainless steel balls (diameter = 6mm) in proportion of 15 Ball
to Powder Ratio (BPR) in hardened steel cylindrical container of 250 ml volume. Planetary ball Mill was operated at
200 rpm for 16 hrs. Three group of space holder were collected that have the size ranges<105 µm (S 1 ), 150-200 µm
 Proceedings of NCAMMM - 2018

(S 2 ), 200-300 µm (S 3 ).Milled powder was vigorously mixed with ammonium bi carbonate (NH 4 HCO 3 ) space
holder of different sizes in volume proportion of 1:1 and compacted in a die by applying 200 MPa compaction
pressure. Three types of samples (M/S 1 , M/S 2 and M/S 3 ) were prepared by using milled powder and the space
holder having three different sizes. One additional type of samples also compacted by using elemental powder
(Ti4wt%Al4wt%Co) and the space holder having size S 3 in the proportion of 1:1 (sample is assigned as
U/S 3 ).Compacted samples were dried in an oven at 1500C up to 4 hrs and then sintered at 8000C and 11000C for 60
and 90 minutes respectively.
Microstructures of milled and unmilled powders and prepared foams were investigated using Field
Emission Scanning Electron Microscope (FESEM) of Nova Nano SEM 430 model. Elements present in the milled
Ti4Al4Co was investigated by using EDX that is additionally attached to the FESEM. Compressive test was carried
out by using instron 8801 Universal Testing Machine (UTM) at strain rate of 0.001/s. Corrosion test was performed
by a potentiostat using three conventional electrodes. Platinum wire was used as counter electrode while saturated
calomel electrode as a reference electrode. Foam samples were employed as working electrode having exposed
surface area of 0.188 cm2 in the electrolyte. Simulated body fluid (SBF) was prepared by dissolving the different
components of appropriate concentration in distilled water. The composition of the SBF is taken from somewhere
[9].

4. Results and Discussion

4.1 Characterization of Space holders
Average particulate sizes for the three group of space holders (SH) were analysed as 75μm, 152μm and 220μm for
S 1 , S 2 and S 3 respectively. Average major and minor dimensions of S 1 were observed as 87μm and 60μm while
their value for S 2 were analysed as 152μm and 117μm respectively. Aspect ratio is an important factor that represent
the shape of the particles in the form of sphericity. It is measured by dividing the average major dimension of the
particles to the average minor dimension. The values of aspect ratio for S 1 , S 2 and S 3 were measured as 1.45, 1.54
and 1.73 respectively. It indicates that in this study, sphericity of the SH reduces with the increment of average size
of the particulates.

4.2. XRD Analysis of milled and unmilled powders

Milled powder (M) and unmilled powder mixture (U) were examined by XRD for identification of phase formation
and change in crystallographic structures of the material due to milling. The XRD diffraction patterns of U and M
were represented by the Fig. 1.The results show that U contains the peak of Ti, Al and Co as mixed. After milling,
the possible phases in M that were analysed are α-Ti, TiO 2 and α (BCC) iron. Although in normal atmosphere,
intense reaction of Ti and O to form TiO 2 takes place above 5000C whereas during milling, reaction occurs at very
low temperature (temperature of milling temperature could not exceed above 400C).It may be caused due to
generation of crystalline defects (grain boundary, dislocations, vacancy, stacking fault etc.) and new grain surface
formation during milling that provide fast diffusion path and increase the kinetics of reaction.Peaks of α (BCC) iron
also present in Mas impurity because of wear of iron particles from the stainless steel container and balls due to
impact, rolling and rubbing actions among them. It is clear that TiO 2 was formed due to milling. Fe is present as
impurity due to wear of the milling media. U has the peaks of as mixed elemental powder.

37
Effect of space holder size on microstructure, deformation and corrosion response of Ti4Al4Co alloy foam

4.3. Microstructural analysis of milled and unmilled powders

The micrographs of U and M obtained by FESEM were used for analyse the change in morphology, particle size
distribution and average particle size of powder samples. Particle size distribution curves and respective
microstructures of U and M were shown in Fig. 2(a) and 2(b) respectively. From the microstructure of Fig. 2(a), it is
clear that U has large size scattering ranges of particle with irregular shape. Particles have the size between 5 to
65μm in which about 50% particles have the size in the range of 20 to 30 μm. Standard deviation indicates the
scattering of the powder particles. For the sample U, it was measured as 12 that indicates the large scattering of the
particles. Average particle size and aspect ratio of U were measured as 31μm and 1.34 respectively.
The micrograph of M shows that the morphology of the powder particles got change due to milling. Range
of particle size distribution reduces significantly in the range of 5 to 30μm having standard deviation 5.36. Average
particle size and aspect ratio of the milled powder M were measured as 13μm and 1.56 respectively. Reduction in
particle size range and hence average particle size is caused due to fracturing of particles because of the impact force
imparts by the moving balls on the powder particles. Along with fracturing, particles got flattened that cause the
increment in aspect ratio.

a b

Fig. 2: (a) represents the frequency distribution curve for the powder U while, (b) shows the frequency distribution
curve for milled powder

4.4 Microstructures of the foam samples

Densities of the samples were measured as 2.05 g/cc, 1.89 g/cc and 1.82 g/cc for the sintered samples M/S1, M/S2
and M/S3 respectively. Microstructures of the M/S 1 , M/S 2 and M/S 3 are shown in Fig. 3(a), (b) and (c) respectively.
From the figures, it is clear that according to the descending order of number of pores per unit area accumulated by
the foam samples is M/S 1 , M/S 2 and M/S 3 . During formation of foam, the quantity of space holders used is same
whereas size is varying. So, space holder of smaller size particulate has higher number of particulate resulting
production of foam of higher number of pores.. The average size of the pores for the foam samples M/S 1 , M/S 2 and
M/S 3 are calculated as 65 μm, 128 μm and 196 μm respectively that is 16.87%, 15.58% and 11.34% smaller than the
average size of the space holders S 1 , S 2 and S 3 that were mixed with powder during foam making. It is caused due
to the shrinkage of the pores during sintering.

38
 Proceedings of NCAMMM - 2018

4.5 EDX Analysis of the foam

The foam samples were characterized by the use of EDX to investigate the elements present in the foam material.
Samples made with milled powder have impurities like iron, chromium, nickel and oxygen rather than used
materials (Ti, Al, and Co). Iron, chromium and nickel exist due to wear of milling container and balls that were
made of stainless steel material. Oxygen presents by the intensive reaction of titanium with air present in the
container during milling. It was analysed that the amounts of present impurities are almost same for the M/S 1 , M/S 2
and M/S 3 foams as given in table (1). Existence of iron is advantageous for the foam in mechanical properties
strengthening due to hindering the grain growth cause by the pinning action on the grain boundary during sintering
by the formation of Fe 2 Ti. Chromium and oxygen may help in improvement of corrosion resistance. Foam sample
U/S 3 has the oxygen as impurity. The amount of oxygen in U/S 3 was analysed as the 57% lower in its value than the
milled powder.

a b c

Fig. 3: (a), (b) and (c) represent the micrographs of the foam samples M/S 1 , M/S 2 and M/S 3 respectively. From the
figures, ascending order of number of pores is for M/S 3 , M/S 2 and M/S 1
Table 1: EDX analysis of the of milled and unmilled foam samples
S/N Foam Sample Detected Elements (wt%)
Name Ti Al Co O Fe Cr Ni
1 M/S 1 53.16 3.77 3.71 27.61 8.45 1.87 1.43
2 M/S 2 52.62 3.84 3.68 28.03 8.32 1.99 1.52
3 M/S 3 52.42 3.98 3.76 28.26 7.96 2.01 1.61
4 U/S 3 78.71 4.03 3.94 13.32 ----- ----- -----

4.6. XRD analysis of the foam

XRD analysis of the foam materials were performed to investigate the phases formed after the sintering. For U/S 3 ,
Any intermetallic or precipitate formed among α-Ti, Al and Co was not detected in the XRD pattern as shown in
Fig. 4.

Fig. 4: XRD analysis of the foam samples made by the use of milled and unmilled powder

39
Effect of space holder size on microstructure, deformation and corrosion response of Ti4Al4Co alloy foam

Formation of TiO 2 takes place due to the intense reaction of atmospheric oxygen with α-Ti during the any stage of
foam sythesization. It was noted that the peaks of Al and Co were vanished after sintering. It is caused due to the
solid solutionization of Al and Co in α-Ti lattice. Foam samples made by the use of milled powder (M/S 1 , M/S 2 and
M/S 3 ) have numerous phases of Ti and Fe. The formation of FeTi and Fe 2 Ti was caused due to the limited solubility
of iron in the α-Ti lattice. Some free α-Fe was also detected. Cr 2 O 3 , Ni 3 Ti and α-Fe 2 O 3 were also observed in small
amounts. Because of rutile TiO 2 was traced at the peak having higher intensity in the pattern, shows considerable
amount of TiO 2 formation after sintering.
4.7 Mechanical properties
It is evident from the Fig. 5(a) that plastic collapse stress of M/S 1 , M/S 2 and M/S 3 are 79 MPa, 51 MPa and 38 MPa
respectively. The value of elastic modulus is reported as 27 GPa, 21 GPa and 16 GPa for the respective M/S 1 , M/S 2
and M/S 3 samples. It indicates that mechanical properties of the foam is strong function of relative density. A
comparison was done between M/S 3 and U/S 3 to investigate the effect of milling on the mechanical properties of the
foam. It was observed that the foam with made with unmilled powder has higher plastic collapse stress, elastic
modulus, plateau stress and energy absorption capacity than the foam made with milled powder Fig. 5(b). It might
be cause due to improper sinterability of milled powder due to containing higher amount of TiO 2 .

a b

Fig. 5: (a) indicates the stress- strain graph for the milled foams having different pore size. (b) showsthe stress strain
graph for the milled and unmilled foam samples ( value of mechanical properties are shown in below table).

Sample Max. Collapse Stress Elastic Modulus Plateau Stress SEAC Densification
(MPa) (Gpa) (MPa) (MJ/m3) Strain
M/S1 79.64 26.38 19.41 11.4 0.61
M/S2 51.18 20.16 15.34 9.5 0.63
M/S3 38.22 15.79 11.62 7.1 0.67
U/S3 44.89 22.11 29.43 19.21 0.63

4.8 Corrosion behaviour

To evaluate the corrosion behaviour of the milled foam having different size of pores, Tafel plot was drawn by the
use of potentistat that was given in Fig. 6.

40
 Proceedings of NCAMMM - 2018

Fig. 6: Tafel plot for milled and unmilled foam samples

The rate of corrosion for the M/S 1 , M/S 2 and M/S 3 are 2.69 mm/y, 1.33 mm/y and 1.01 mm/y respectively. This
indicates that the foam of larger pore size has higher corrosion resistance than the smaller one. It is caused due to the
large number of pores available for the pitting and crevice corrosion. Also, M/S 3 has higher corrosion resistance
than U/S 3 . Because in U/S 3 , alloying elements are not uniformly distributed and due to difference in electrochemical
potential of each element, galvanic corrosion takes place while, in M/S 3 alloying elements are uniformly distributed
in the material due to milling cause less chance of galvanic corrosion.

5 Conclusions
 Microstructures of the foam samples indicate that average size of the pores is directly depend on the average
size of used space holder during foam making.
 Foam sample that contains smaller pores has higher plastic collapse stress, elastic modulus and energy
absorption capacity. Also, foam samples made with unmilled powder has higher mechanical properties than the
foam sample made with milled powder having same density and pore size.
 The foam of larger pore size has higher corrosion resistance. Also, the foam made with milled powder has
improved corrosion resistance than the foam made by the use of unmilled powder.

References
[1] Albrektsson T, Brånemark P, Hansson H-A, Kasemo B, Larsson K, Lundström I, et al. The interface zone of inorganic
implantsIn vivo: Titanium implants in bone. Annals of biomedical engineering. 1983;11:1-27.
[2] Chu C, Chung C, Zhou J, Pu Y, Lin P. Fabrication and characteristics of bioactive sodium titanate/titania graded film on
NiTi shape memory alloy. Journal of Biomedical Materials Research Part A. 2005;75:595-602.
[3] Long M, Rack H. Titanium alloys in total joint replacement—a materials science perspective. Biomaterials. 1998;19:1621-
39.
[4] Niinomi M, Hattori T, Morikawa K, Kasuga T, Suzuki A, Fukui H, et al. Development of low rigidity β-type titanium alloy
for biomedical applications. Materials Transactions. 2002;43:2970-7.
[5] Lewis G. Properties of open-cell porous metals and alloys for orthopaedic applications. Journal of Materials Science:
Materials in Medicine. 2013;24:2293-325.
[6] El Feninat F, Laroche G, Fiset M, Mantovani D. Shape memory materials for biomedical applications. Advanced
Engineering Materials. 2002;4:91-104.
[7] Santhanam PR, Dreizin EL. Predicting conditions for scaled-up manufacturing of materials prepared by ball milling.
Powder technology. 2012;221:403-11.
[8] Suryanarayana C. Recent developments in mechanical alloying. Rev Adv Mater Sci. 2008;18:203-11.
[9] Krząkała A, Służalska K, Widziołek M, Szade J, Winiarski A, Dercz G, et al. Formation of bioactive coatings on a Ti–6Al–
7Nb alloy by plasma electrolytic oxidation. Electrochimica Acta. 2013;104:407-24.

41
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

Closed Cell Aluminium Composite Foam for Crashworthiness Applications

D. P. Mondal, A.N.Ch. Venkat, Sanjeev Saxena

CSIR-Advanced Materials and Processes Research Institute, Bhopal-462026

Abstract: Closed cell aluminium foams with various densities have been made through stir casting technique. The
crucible temperature and melt temperature were controlled to control the foaming temperature for controlling foam
cell size and foam densities. The foam made were then characterized for their micro-architectural characteristics as
well as deformation responses. The crash of vehicle take place at very high speed and therefore, the foams were
tested both at quasi-static and dynamic conditions. It is noted that at dynamic conditions the foams exhibit much
higher strength and energy absorption. Then the foams were filled manually inside the commercially available
crash-box and tested using drop weight test methodology. The weight is varied upto 375 kg and speed is varied up to
55Km/hr. The deformation behavior of bare foam blocks, foam filled crash box and empty crash boxes were studied.
It is observed that the foam synthesized through this technique is excellent for crash energy absorption. It is also
noted that hardly any significant weight will be added into the vehicle. These closed cell foam has also the potential
to be used in other transport sectors.

1. Introduction
A lot of casualities and damage to the vehicles are being encountered in transportation sectors particularly in road
ways and railways. To avoid these, there is requirement of materials which can absorbed accidental impact energy or
blast energy without increasing the stress waves or impulse waves to the vehicles or to the passenger on boards. In
this connection ultralight weight materials with exceptionally high energy absorbing capacity at low stress or impulse
waves are being developed [1-4]. Aluminium foams are one such kind of materials. It can also be used as core of
sandwich panels for increasing strength and stiffness of the structure [5-6], vibration and sound attenuation [7-
8].CSIR-AMPRI, Bhopal has developed expertise, facilities and know- how for synthesis, characterization and
utilization of these materials under one umbrella. The aluminum foams have been developed and characterised in
detailed for their microachitectural characteristics as well as deformation response in sample level as well as in
component level.

2. Synthesis of Closed Cell Aluminium Foam and its Characterisation

Aluluminium composites based with different aluminium alloys and reinforcements are developed in large scale in
the laboratory. Different kinds of reinforcements like SiC, Al2O3, Fly ash, Zircon sands are successfully reinforced
uniformly within the matrix. CSIR-AMPRI has developed closed cell aluminium foams of different densities. High
strength low cost closed cell hydride aluminium fam is also developed. Presently, in the laboratory, the technology is
developed to make ~150 kg of foam per day ( A foam billet weight of 35 Kg). The relative density of foam varies
from 0.25 gm/cc to 0.90 gm/cc. The Foam billet, its cross-section and its typical digital microstructure are shown in
Fig.1(a), (b) and (c) respectively. The cells are quite uniformly distributed throughout its cross-section.
Closed cell aluminium composite foam for crashworthiness applications

The density of foam were measured from weight per unit volume using its mass and dimension. Regular dimension
of foam samples were considered. The density is measured from different locations of the foam billet. It is noted that
the bottom portion of foam has around 10% higher density than the top portion. The degitised photographs are taken
to measure the cell size and cell wall thickness. The density of foam is used to get the porosity content in foams. The
porosity of foam is defined as (1-relative density), where relative density is defined as the ration of foam to the ratio
of dense composite. Compressive deformation behavior of these foams were studied under quasi static condition (in
an INSTRON UTM) and under dynamic condition using speed Hopkinson pressure bar. In addition the bare foam
block, empty crash box and foam filled crash boxes were tested using Drop weight test facility at ARAI, Pune (Drop
weight ~375 Kg and speed 55 Km/hr).

Fig. 1 (a) foam billet , (b) Crossection of foam billet and (c) higher magnification digitized micrograph of foam
(Relative density =0.15)

3. Results and Discussion

The quasistatic compressive stress –strain curves of foam of varying relative density is shown in Fig.2. It is evident
from this figure that the foams deforms under zig-zag stress response upto densification strain after yielding. The
yield stress is defined as the plateau stress. It is noted that the plateau stress or flow stress increases with increase in
relative density. It is noted that at a relative density of 0.31, the plateau stress is around 24 MPa. The energy
absorption is calculated from the stress strain curve upto strain of 0.6 and it is noted that the maximum energy
absorption by the foam of RD=0.31 is 10 MJ/m3 up to 40% od deformation. When these foams are tested under
dynamic conditions in an speed Hopkinson pressure bar, the plateau stress and flow stress increased significantly
(Fig.3). For the foam with relative density of 0.31, the plateau stress increased to 45 MPa at a strain rate of 1000/s.
This signifies that its energy absorption increased to 16 MJ/m3 when foams deforms up to 40%. This also
demonstrate that the foam have greater capability to absorb more energy under dynamic condition. In view of these
the foams were filled into the crash box commercially available and tested using drop weight test facility at a speed
of 55 Km/hr. The force displacement curves under drop weight tests for empty crash box and the foam filled crash
boxes are shown in Fig.4 and Fig.5 respectively. It is evident from this figure that initially stress increases upto
certain limit and then there is a hump, followed by which the stress again increases and reach to the maximum. After
reaching to the maximum the stress again decreases. This behavior is for empty crash box. The initial hum is for
starting of folding of crash box during deformation. When the crash box is filled with foam, the clear hum in the
initial period is not existing. But the slop changes at certain displacement where the increment of stress with strain

43
 Proceedings of NCAMMM - 2018

decreases, indicating slower rate of load transfer which causes deceleration of impacting object. This is required for
any crashworthiness applications.

Stress- Strain plot

50 1500

40 1200

Stress (Mpa)

Strain Rate
30 900

20 600
Stress
Strain Rate
10 300

0 0
0.00 0.05 0.10 0.15 0.20 0.25 0.30
Strain

Fig. 2: Compressive stress strain curves of closed Fig. 3: Compressive stress strain curves under dynamic Under
quasistatic loading condition Loading condition

Fig. 4: Force displacement curves of Empty crash box Fig. 5: Force displacement curve of foam filled under
dynamic loading condition (drop weight tests) crash box under dynamic loading (drop weight test)

The energy absorption calculated for these tests were 9.5 KJ and 20.5 KJ fr empty crash box and foam filled crash
box. The foam block absorbs ~ 4 KJ. It is further noted that the base plate at which the crash boxes were fixed for
testing did not undergo any damage. In this case the crash box weight was 850 gm, the foam weight was 350 gm.
Thus by adding 350 gm of foam the energy absorption is improved more than 100%. This is even much higher than
that of sum of energy absorbed by bare foam and empty crash box. This clearly demonstrates that the bare foam
should not be used. For getting effective crash-energy absorption, foam filled crash boxes are better option. If one
considers for weight increment, it is hardly anything as compared to the weight of the car. As per as safety is
considered, the cost of the foam is hardly anything in comparison to the car or the life of the passenger. The foam

44
Closed cell aluminium composite foam for crashworthiness applications

required for the crash boxes would be costing hardly Rs 2000.00 to 3000.00/- depending on the weight of the car
and level of safety.

4. Conclusions
The following conclusion can be drawn from the present study:
• Closed cell aluminium foams in batch scale process has successfully synthesized. In a single heat a foam billet of
size up to 35 kg could be made.
• The foams have quite uniform cell size and its densities could be varied by varying the process parameters.
• The closed cell foams exhibited significanntly higher strength and flow stress under dynamic condition as
compared to that in quasi-static conditions.
• Closed cell aluminium foams could economically and technically be used for crashworthiness of vehicles.

References
[1] Stefano Rossi, Lorenzo Bergamo, Vigilio Fontanari, Fire resistance and mechanical properties of
enamelled aluminium foam, Materials & Design, Volume 132, pp.129-137, 2017,
[2] Zhifang Liu, Zhichao Huang, Qinghua Qin, Experimental and theoretical investigations on lateral crushing
of aluminum foam-filled circular tubes, Composite Structures Volume 175, Pages 19-27, 2017,
[3] S. Talebi, M. Sadighi, M.M. Aghdam, S.M.H. Mirbagheri, Micro–macro analysis of closed-
cell aluminum foam with crushing behavior subjected to dynamic loadings, Materials Today
Communications, Volume 13, Pages 170-177, 2017,
[4] Xin Pang, Hejun Du, Dynamic characteristics of aluminium foams under impact crushing, Composites Part B:
Engineering, Volume 112, Pages 265-277, 2017,
[5] Zhibin Li, Zhijun Zheng, Jilin Yu, Fangyun Lu, Deformation and perforation of sandwich panels
with aluminum-foam core at elevated temperatures, International Journal of Impact Engineering, Volume
109, Pages 366-377, 2017,
[6] Chengjun Liu, Y.X. Zhang, Jing Li, Impact responses of sandwich panels with fibre metal laminate skins
and aluminium foam core, Composite Structures, Volume 182, Pages 183-190, 2017,.
[7] Xingchuan Xia, Zan Zhang, Weimin Zhao, Chong Li,Yongchang Liu, Acoustic properties of closed-
cell aluminum foams with different macrostructures, Journal of Materials Science & Technology, Volume 33,
Pages 1227-1234, 2017.
[8] Dong Guan, Jiu Hui Wu, Jiangling Wu, Jing Li, Weitao Zhao, Acoustic performance of aluminum foams with
semiopen cells , Applied Acoustics, Volume 87, Pages 103-108, 2015.

45
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Studies of Dendritic Arm Growth during T6 Heat-Treatment on Directionally Solidified

Al-4.5Cu-Sic Composite

S. Debnatha*,G.R.K Sastryb and R. N. Raic

a,c
Department of Production Engineering, National Institute of Technology, Agartala, 799046
b
Department of Mechanical Engineering, NIT, Agartala, 799046, India
*
Corresponding Author

Abstract: In this work investigate the T6 heat treatment behaviour upon the directionally solidified Al-4.5%Cu-SiC
composite in terms of dendritic arm spacing (λ1, λ2). The directionally solidified (DS) casting was perform with
averticallyupward mold bottom cooled by water jet.A power law is derived with the cooling rate (𝑇𝑇̇) as a function of
position of the casting and it was also associated to dendritic arm spacing (λ1,λ2).The microstructural
characterizationperformed with opticaland scanning electron micrograph. It has been found in the primary (λ1) and
secondary dendritic arm (λ2) spacing no significant changewith thefunction of cooling rate.The Primary arm
spacings (λ1) found slightly more increase compared to secondary dentritic arm spacing(λ2) in 6 hr of solution time
at 520°C.
Keywords: Al-4.5Cu-SiC composite, Directional solidification, primary and secondary dendritic arm spacing.

1. Introduction
Nowadays engineering application demand for materials which have good strength to weight ratio and less
expensive suitable for automobile applications where fuel economy with improved engine performance are critical
aspects [1]. For many engineering application required abroad spectrum of properties which is difficult to achieve by
monolithic material system [2]. Al-Cu-SiC metal matrix composite has been noted to offer excellent mechanical
properties with good corrosion resistance and suitable weldability compare to Al-4.5Cu alloy [3].In the solidification
process during casting considerable metallurgical and mechanical variables involves which makes it a complex
phenomenon. This solidification process primarily driven by thermo physical properties related to solidifying alloy
and it vary with temperature and time of the process. The microstructure is depends on cooling rate (𝑇𝑇̇) up tolarge
extent which is directly related to the mechanical properties like yield strength, hardness, toughness etc.[4].
The one of the important study of solidification phenomena is investigate the morphology of the dendritic arm
pattern which primarily characterized by all three arm spacings like primary (λ1), secondary (λ2) and tertiary
(λ3).During solidification process micro segregation affects, the microstructure of the material which is directly
related to ultimate tensile strengths, toughness and ductility all the mechanical properties. The porosity distribution
as well as homogenization kinetics of solidified alloys also related to this microstructural features [5-7]. So that
during casting solidification processes it is important to observed the formation pattern of dendritic arm structures
changes, either columnar or equiaxed [8,9].Many complex features like thermal gradient, crystal structure,
anisotropic,solutal, thermodynamics and kinetics of diffusion regulated the formation of dendritic arm growth [10].
Some scientist Hunt et. al [11] and Bouchard et. al [12] proposed predictive growth model for the complex primary,
secondary and equiaxed dendritic arm growth in unsteady state condition. The present investigation focused on the
Studies of dendritic arm growth during T6 heat-treatment on directionally solidified Al-4.5Cu-SiC composite

study of primary and secondary dendritic arm spacings formation pattern with the function of cooling rate (𝑇𝑇̇).Also,
microstructural characterizations with scanning electron microscopy (SEM) are used to observe the arm profiles of
microsegregation in the Al-4.5Cu-SiC composite matrix.

2. Experimental Procedure
The Directional Solidified experiment was performed with binary Al-4.5wt-%Cu alloy mixed with 2% of SiC
particulates. The chemical composition of obtaining Al-obstructs that are utilized to develop the matrix alloys
are illustrated in Table 1.The chemical composition investigation was done under an Optical Emission
Spectrometer (ARL 3460, Switzerland).For the synthesis of Aluminium based MMC's, the pure aluminium alloy
with 4.5% is used as a matrix material the 25±6µm mesh size of SiC particle used in the melt.
0
The Al-Cu alloys were melted at the temperature around 650 C in the graphite crucible by using induction furnace.
Silicon carbide (SiC) particles are used as ex-situ reinforcements, where as scrap magnesium is added as a wetting
reagent in between particles and matrix. Before pouring the SiC powder into the melt, it kept for pre-heating at
8500C for 3 hours. Silicon carbide (SiC) in particles were added to the molten alloys with the help of a cone-shaped
funnel with feed rate 1.2 to 1.6 gm/s. During the powder, feeding the whole mixture was mixed continuously by a
mechanical stirrer coupled with a motor at a rotational speed around 350r.p.m.The other various process parameters
associated with the stirring casting process encapsulated in Table2.

Table1Elemental composition of Al-Cu alloy

Element Cu Mg Si Fe Zn Mn Ni Sn Al
wt% 1.5~3.5 0.30 9.6~12 0.90 1 0.50 0.50 0.20 Bal

Table2 Process parameter

Parameter Unit Value
In Fig. 1 (a) shows the stir casting setup used
Spindle speed rpm 350
in the development of MMCs. The
Stirring time Sec 600
0 schematic of vertical ascendant
Stirring temperature C 750-800
0 unidirectionally solidified casting set up is
Pre-heating temperature of SiC C 850
shown in Fig.1 (b).The experimental setup
Pre-heatingtime Min. 90
0 for casting is designed insuchaway that mold
Pre-heating temperature of mold C 300
is heated under the eutectic temperature of the
Powder feed rate gm/sec 1.2-1.6
alloys in a heat-insulated chamber with a
bottom opening. During the pouring of the melt intothemold, the cold water-jet is applied to the bottom of the
mold which promoting to vertical ascendant unidirectional solidification. A split cylindrical stainless steel mold was
used which have internal diameter of 30mm, height170 mm and thickness of 10mm. Insulating alumina paste is
used on the internal surface of the mold to minimize radial heat losses of the metal/mold system. During
solidification four k type, the thermocouple is used in different position 5, 10, 15 and 20 mm from the heat-
extracting surface along the length of the mold to measure the continuous temperature of casting.

47
 Proceedings of NCAMMM - 2018

(b) Melt
(a)

(d)

(c)

Fig.1 (a) Induction furnace (inset pic: crucible) (b)schematic of directional solidification set up (c)As cast specimen
(d) longitudinally sectioned specimen

In Fig 1(c), illustrate the cylindrical shape specimen developed by directionally solidified casting. Each directionally
Solidified casted specimen sectioned along longitudinally Fig. 1(d) from the four different position of the
thermocouple insertedinto the mold to record the cooling rate. After that cut the samples at different four positions
and polished using a standard metallographic technique like rough polishing in belt grinder, intermediate polishing
on different grades of SiC abrasive papers followed by final polishing with alumina and diamond paste.The
polished sample washed in acetone and etched them with an acid solution composition of H2O 5 ml, HCl 60 ml,
HNO3 30 ml and HF 5 ml to reveal dendritic arm boundary.An Optical microscope (Leica Qwin) attached with
image processing software was used to measure the primary dendritic arm spacing (DAS) (50 independent readings
for each selected position, with the average taken to be the local spacing) and their range of distribution. The
triangular method was applied to a transverse section (right angle to the solidification front) measure the primary
dendritic arm spacing [13]. The triangle is considered joining the three adjacent dendrite centers, the sides of the
formed triangle consider as primarydendritic spacing(λ1) shown in Fig. 2(a).The secondary dendritic arm (λ2)
spacing was measured by averaging the distance between adjacent side branches in longitudinal section of a primary
arm illustrate in Fig. 2(b). In this case, also50random λ2valuesare considered in each longitudinally sectioned of
casting.

(a)
λ1 (b)

48
Studies of dendritic arm growth during T6 heat-treatment on directionally solidified Al-4.5Cu-SiC composite

Fig. 2 Schematic of methods used for measurement (a) primary and (b) secondary dendritic arm spacings

The proposed T6 heat treatment, which is performed in this experiment, consists of following steps:
a) Solution heat treatment for 6hr at 520+20c;
b) Warm water Quenching (60+20C);
c) Subsequent Ageing for 5 h at 1600c;
d) Air cooling.
Muffle furnace is used to perform all the heat treatment operation. The time to gain the solution temperature for the
samples 20-30minutes, which is excluded from the written times.

3. Results and Discussion

3.1 Microstructure, cooling rate, and dendritic arm spacing
It can be observed from the graph Fig. 3(a) that near to bottom of the mold experience higher cooling rate and
profile is decreasing as going up along the length of the casting.

Fig. 3 (a) cooling rate vs position graph (b-e) microstructure changes along the length of the casting

In Fig. 3(b-e) represented the typical microstructure of the DS cast Al-4.5%Cu-SiC composite. From the
microstructures along the length of the casting (5, 10, 15, 25 mm) it has been observed that cooling rate employs less

49
 Proceedings of NCAMMM - 2018

effects on the microstructure. It is direct the casting range from which part of the casting samples was extracted for
carried out the T6 heat treatment process and studied the variation of primary and secondary arm spacing in a
different position. The length estimation was made of the primary(λ1) and secondary(λ2) dendritic arm spacings on
samples extracted from the all four position (5,10,15 and 25 mm) of the DS composite casting. It can be seen Fig.
4(a) that λ1 and λ2 profile slightly increased with the change in distance from the bottom of the casting. It is observed
that the coarsening process of dendritic arm initiated after 6hr solution time at 5200C.In the water-cooled bottom of
the mold due to higher thermal gradient solidification front get relatively less time to diffusion along the
matrixwhich result, small dendritic arm formation observed near the bottom and towards the top of the casting arm
length increasing.An Experimental power law was derived as a function of position with primary (λ1),and secondary
(λ2) arm length. The equation is as follows,

λ1 and λ2 =constant(P)0.54.
The other thermal parameter solidification-cooling rate (𝑇𝑇̇) also co-relate with the primary (λ1)and secondary (λ2)
dendritic arm spacings. The solidification cooling rate ( 𝑇𝑇̇ ) calculated from the data recorded in K-type
thermocouples, which are stealth in the mold contact surface of the liquidus front passage. The average values of λ1
and λ2plotted as a function of cooling rate (𝑇𝑇̇) in Fig.4(b) It has been observed that higher values of cooling rates
recorded near the bottom of the mold and continuously decreasing profile towards the casting head due to the
increasing of thermal resistance of the solidified metal. From the graph both primary and secondary dendritic arm
spacing developed experimental growth law by a power function relating λ1 and λ2to 𝑇𝑇̇ given by

λ1/λ2=41(𝑇𝑇̇ )-1/3.

This result is complete concurrence as reported by previous researcher Rocha et al. [14], Peres et al. [15] and costa
et al. [3] with this outcome.

Fig. 4 measured primary (λ1) and secondary dendrite arm spacing (λ2) as a function of (a) position of the casting and (b)
solidification cooling rate along the length of the casting.

4. Conclusion
The present experimental work deals with the response of Al-4.5%Cu-SiC composite during T6 heat treatment
process. In this work developed a power function law with both primary (λ1) and secondary (λ2) dendritic arm
spacings as a function of position and cooling rate both. The Solution time 6h at 5200C followed by 5hr ageing at

50
Studies of dendritic arm growth during T6 heat-treatment on directionally solidified Al-4.5Cu-SiC composite

1600C is adequate for dissolution of inter metallic in the Al-Cu matrix alloy. It has found that Primary arm spacings
(λ1) slightly more value compared to secondary dentritic arm spacing(λ2) and with the higher cooling rate (𝑇𝑇̇) both
primary and secondary dendritic arm spacing in decreasing slope.

Acknowledgement
The authors acknowledge the support provided by Production Engineering Department of National Institute of
Technology Agartala to perfrom the experiment.

Reference
[1] Tjong SC. Processing and deformation characteristics of metals reinforced with ceramic nanoparticles. In
Nanocrystalline Materials (Second Edition) 2014 (pp. 269-304).
[2] Rino JJ, Chandramohan D, Sucitharan KS, Jebin VD. An overview on development of aluminium metal
matrix composites with hybrid reinforcement. IJSR, India online ISSN. 2012 Dec:2319-7064.
[3] Costa TA, Moreira AL, Moutinho DJ, Dias M, Ferreira IL, Spinelli JE, Rocha OL, Garcia A. Growth
direction and Si alloying affecting directionally solidified structures of Al–Cu–Si alloys. Materials Science
and Technology. 2015 Jul 1;31(9):1103-12.
[4] Venkatesan A, Gopinath VM, Rajadurai A. Simulation of casting solidification and its grain structure
prediction using FEM. Journal of Materials Processing Technology. 2005 Sep 15;168(1):10-5.
[5] Yuan SN, Jia LN, Ma LM, Jiang H, Zhang H. Microstructure and room temperature mechanical properties of
hypereutectic Nb–Si based alloy processed by directional solidification. Materials Science and Technology.
2014 Jan 1;30(1):75-80.
[6] Goulart PR, Spinelli JE, Cheung N, Garcia A. The effects of cell spacing and distribution of intermetallic
fibers on the mechanical properties of hypoeutectic Al–Fe alloys. Materials Chemistry and Physics. 2010 Jan
15;119(1-2):272-8.
[7] Garcia LR, Osório WR, Peixoto LC, Garcia A. Mechanical properties of Sn–Zn lead-free solder alloys based
on the microstructure array. Materials Characterization. 2010 Feb 1;61(2):212-20.
[8] Zaïdat K, Mangelinck-Noël N, Moreau R. Control of melt convection by a travelling magnetic field during
the directional solidification of Al–Ni alloys. Comptes Rendus Mecanique. 2007 May 1;335(5-6):330-5.
[9] Kurz W, Bezencon C, Gäumann M. Columnar to equiaxed transition in solidification processing. Science and
technology of advanced materials. 2001 Mar 1;2(1):185-91.
[10] Salgado-Ordorica MA, Rappaz M. Twinned dendrite growth in binary aluminum alloys. Acta Materialia.
2008 Nov 1;56(19):5708-18.
[11] Hunt JD, Lu SZ. Numerical modeling of cellular/dendritic array growth: spacing and structure predictions.
Metallurgical and Materials Transactions A. 1996 Mar 1;27(3):611-23.
[12] Bouchard D, Kirkaldy JS. Prediction of dendrite arm spacings in unsteady-and steady-state heat flow of
unidirectionally solidified binary alloys. Metallurgical and Materials Transactions B. 1997 Aug 1;28(4):651-
63.
[13] Gündüz M, Çadırlı E. Directional solidification of aluminium–copper alloys. Materials Science and
Engineering: A. 2002 Apr 30;327(2):167-85.
[14] Rocha OL, Siqueira CA, Garcia A. Heat flow parameters affecting dendrite spacings during unsteady-state
solidification of Sn-Pb and Al-Cu alloys. Metallurgical and Materials Transactions A. 2003 Apr 1;34(4):995-
1006.4
[15] Peres MD, Siqueira CA, Garcia A. Macrostructural and microstructural development in Al–Si alloys
directionally solidified under unsteady-state conditions. Journal of Alloys and Compounds. 2004 Nov
3;381(1-2):168-81.

51
Sub - theme

Ceramics and
Composites
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Evaluation of Metallurgical and Mechanical Properties of CeO 2 Reinforced Zirconia

Toughened Alumina

B K Singha*, Nitai Chandra Adaka, S.S.Royb, S.S. Chakrabortya, Nilrudra Mandala

a
Centre for Advance Materials Processing, CSIR-Central Mechanical Engineering Research Institute,
Durgapur - 713209, INDIA; bMechanical Engineering Department, National Institute of Technology
Durgapur, INDIA; Email: bipinmech2008@gmail.com

Abstract: In the present work, the effect of different wt% CeO 2 addition (1-20 wt %) on the phase transition,
microstructure and mechanical properties of zirconia toughened alumina (ZTA) has been studied. The
stabilization of crystallographic and thermal phases has been characterized by field emission scanning
electron microscopy (FESEM) and x-ray diffraction (XRD). The samples are calcined at 800°C followed by
sintering at a temperature of 1600°C for 2 hours soaking time. Effect of microstructure on physical properties
i.e. density as well as mechanical properties viz. hardness and fracture toughness has been studied. The
morphology and microstructure have been observed by FESEM. Maximum value of hardness obtained is
around 15.39 GPa, at 5 wt % CeO 2 in ZTA matrix. The maximum fracture toughness of 5.83 MPa.m1/2 is found
at 20 wt % CeO 2 in ZTA matrix.
Keywords: partially stabilized zirconia; hardness; fracture toughness

1. Introduction
In current high productivity manufacturing scenario, industry wants to develop component with longer life at
low cost. In this context, ceramic is one of the good choices because of its properties like hot hardness,
chemical inertness etc. Low toughness and high brittleness properties restricts the use of ceramics in many
applications. Transformation toughening is the most important phenomenon which facilitates to increase the
toughness property of the ceramics [1]. Alumina or aluminium oxide (Al 2 O 3 ) in its different levels of purity is
used more often than any other advanced ceramic material due to its superior properties like high mechanical
strength. Zirconium oxide is also used with aluminium oxide to produce very strong and tough composite for
mechanical and structural applications. The controlled stress induced volume expansion due to tetragonal to
monoclinic inversion of zirconium oxide enhances the toughness of the material. The significant improvement
of mechanical properties is observed when zirconia and alumina ceramics are doped with some elements such
as Y 2 O 3 [2], CeO 2 [3], Cr 2 O 3 [4] etc. Maiti et al. [5] showed that when rare earth elements (Y, La) are doped
in Al 2 O 3 –ZrO 2 ceramic composites the fracture toughness of develop composite is significantly enhanced.
Nik et al. [6, 7] illustrates that cerium dioxide (CeO 2 ) doping improves hardness and fracture toughness of
ZTA–MgO composites and have a significant role in the toughening phenomenon. Tsukuma and Shimada [8]
have made an attempt to investigate the strength, hardness and fracture toughness of CeO 2 stabilized
tetragonal ZrO 2 polycrystals (TZP). It has been found that there is a direct effect of grain-size and amount of
CeO 2 on fracture toughness and hardness of the material. The work done by Hirano and Inada [9] has revealed
that there is a strong relation between hot isostatic pressing temperature with fracture toughness, strength and
Evaluation of metallurgical and mechanical properties of CeO2 reinforced zirconia toughened alumina

hardness of CeO 2 and Y 2 O 3 added ZrO 2 and Al 2 O 3 . The investigation has also revealed that the fracture
toughness of 6-7 MPa.m1/2 can be obtained for (4Y, 4Ce)-TZP/ 25 wt % Al 2 O 3 composite.
A little work has been carried out in respect of ceria mixed ZTA. In this investigation the authors tried
to develop different composites of ZTA with different concentrations of ceria, to observe the effect of ceria
content on properties such as hardness and fracture toughness. The developed powders were characterized by
FESEM, XRD and density measurement.

2. Materials Synthesis
In this work, zirconia toughened alumina composite is used as initial powder. The powder comprises of 90 wt
% alumina and 10 wt % yttria stabilised zirconia having 3mol% yttria (average particle size 1.0 μm, supplier
Zirox). The composite is further doped with different wt% (0 to 20%) ceria (average particle size 1.0 μm,
supplier Loba Chemie, Mumbai, India). The ceria mixed composites are ball milled with 0.8 wt % of
polyethylene glycol 1000 as a plasticizing agent for 24 hours. The properly mixed composites are placed in an
oven for 24 hours at 100ᴼC. The dried composites are crushed in a mortar pestle followed by calcination at
800ᴼC for 1 hour. The calcined specimens are uniaxially compacted to form into circular shapes. The prepared
specimen is placed in a slow temperature (2°C/min) rise furnace for 2 hours at 1600°C. The sintered specimens
are polished by honing process using silicon carbide powder (mesh size 400) and subsequently by bane
polisher using diamond paste (size 0.5-1 micron).

3. Results and Discussion

3.1 Microstructural analysis
Microstructure of the developed composite with different proportion of CeO 2 was observed through FESEM
(CARL-ZEISS-SMT-LTD, Germany, Model: SUPRA 40). It can be observed from Table1 that the overall
grain size increases with increasing percentage of ceria. The grain size varies from 1.36 micron to 1.88 micron.
The crystal sizes of samples also increases with increase in the percentage of ceria. Table 1 also shows that the
density of the composites decreases with increase in the percentage of ceria. These may be resulted by the
porous structure of ceria and the large grain size of the developed composite. The theoretical density of
developed composites has direct relationship with the unite cell size and the morphology of grains (i.e. XRD).
From Table 1 it can be illustrated that, the crystallite size (nm) of develop composites are increases, which can
be responsible for increase in volume of unite cell. Therefore, the value of theoretical density decreases with
increase in percentage of ceria. A typical FESEM image with point EDAX along with XRD of 5 wt% CeO 2 is
portrayed in Figure 1. The intensities of monoclinic phase (m-ZrO 2 ) as well as tetragonal phase (t-ZrO 2 ) of all
composites were determined from XRD plot. Table 1 clearly illustrate that when the percentage of ceria
increases, the tetragonal phase increases and the monoclinic phase decreases.
3.4 Mechanical properties
In this experiment, Vickers hardness testing machine Matsuzwa, MXT-70 is used to measure the hardness and
fracture toughness of all developed samples. A square-based pyramid indenter in which the opposite side faces

53
 Proceedings of NCAMMM - 2018

are fixed at 136 degree from one another is employed to calculate the hardness. The load applied by the
indenter on the surface ranging up to 2 Kgf approximately.

Table 1: Phase Composition, Density, Crystallite Size, and Grain Size of different wt % CeO 2 in ZTA

Density (% Bulk Crystallite Phase Avg grain size

Composition m-ZrO 2 t-ZrO 2
Theoretical) Density size (nm) (micron)
Ce-ZTA (0 wt % CeO 2 ) 94.15 4.32 25.51 60 40 1.36
Ce-ZTA (5 wt % CeO 2 ) 92.58 4.19 27.16 45 55 1.42
Ce-ZTA(7.5 wt % CeO 2 ) 90.35 3.98 29.72 38 62 1.58
Ce-ZTA (10 wt % CeO 2 ) 86.54 3.38 32.20 32 68 1.62
Ce-ZTA (15 wt % CeO 2 ) 83.69 3.29 34.02 28 72 1.72
Ce-ZTA (20 wt% CeO 2 ) 81.86 3.04 36.31 26 74 1.88

Figure 1: FESEM image and XRD plot for 5 wt% Ce-ZTA powder

The measured values of hardness and fracture toughness are listed in Table 2. The highest value of hardness is
observed for 5 wt% Ce-ZTA powder. Beyond 5 wt% CeO 2 again decreases the hardness of the composite.
High value of hardness at 5 wt% is attributed to the distribution of ZrO 2 grains as intergranular and
intragranular particles in the Al 2 O 3 grains, which work as an obstacle for plastic deformation and increases
hardness. Furthermore, the decrease in hardness for CeO 2 addition beyond 5 wt%, is due to higher grain size
of the composite, responsible for the formation of bigger voids and porosity.
Table 2: Hardness and Fracture Toughness of different wt %CeO 2 in ZTA

Composition Sintering Hardness (GPa) Fracture Toughness

Temperature (MPa.m1/2)
Ce-ZTA (0 wt % CeO 2 ) 1600 14.80 4.08
Ce-ZTA (5 wt % CeO 2 ) 1600 15.39 5.34
Ce-ZTA (7.5 wt % CeO 2 ) 1600 15.20 5.53
Ce-ZTA (10 wt % CeO 2 ) 1600 15.04 5.69
Ce-ZTA (15 wt % CeO 2 ) 1600 14.65 5.75
Ce-ZTA (20 wt % CeO 2 ) 1600 14.50 5.83

Table 2 also shows that the fracture toughness of the developed composite increases with CeO 2 content. This
happens mainly due to better retention of metastable tetragonal ZrO 2 for higher amount of CeO 2 content. The

54
Evaluation of metallurgical and mechanical properties of CeO2 reinforced zirconia toughened alumina

metastable tetragonal ZrO 2 expands in volume and restricts the crack propagation. This transformation
toughening phenomena increases the fracture toughness of the composites.

4. Conclusion
The different composition of ceria doped yttria stabilized zirconia toughened alumina has been developed to
study the hardness and fracture toughness. The results revel that for 5 wt% Ce-ZTA, sintering at a temperature
is 1600°C, have maximum hardness i.e. 15.39 GPa. The maximum fracture toughness is 5.83 MPa.m1/2.

Reference:
[1] Smuk B, Szutkowska M. Alumina ceramics with partially stabilized zirconia for cutting tools. J Mater Proc
2003;133:195–8.
[2] Nilrudra Mandal, B Doloi, and B Mondal. Machining Parameters Optimization of Developed Yttria Stabilized
Zirconia Toughened Alumina Ceramic Inserts While Machining AISI 4340 steel. IJMIE, 6, 159-169, (2012)
[3] B. Mondal, N. Mandal, and B. Doloi, Development of Ce/Y-PSZ toughened Alumina inserts for high speed
machining steel, Int J Appl Ceram Tec, 11 (2), 228-239, (2014).
[4] B K Singh, B Mondal, and Nilrudra Mandal Development of Cr 2 O 3 doped Zirconia Toughened Alumina
Ceramic Cutting Insert: Desirability Function Optimization of Turning Parameters for High Speed Machining,
Ceram Int. 2016; 42: 3338–335.
[5] Maiti K, Sil A. Microstructural relationship with fracture toughness of undoped and rare earths (Y, La) doped
Al 2 O 3 –ZrO 2 ceramic composites. Ceram Int 2011; 37: 2411–2421.
[6] Nik Akmar Rejab, Ahmad Zahirani Ahmad Azhar, Mani Maran Ratnam, Zainal Arifin Ahmad, The
relationship between microstructure and fracture toughness of zirconia toughened alumina (ZTA) added with
MgO and CeO 2 , IJRMHM, 41 (2013) 522–530.
[7] Nik Akmar Rejab, Ahmad Zahirani Ahmad Azhar, Mani Maran Ratnam And Zainal Arifin Ahmad, Structural
and Microstructure Relationship with Fracture Toughness of CeO 2 Addition into Zirconia Toughened Alumina
(ZTA) Ceramic Composites, Adv Mat Res. 2013; 620:252-256.
[8] Tsukuma, K. and Shimada, M., Strength, fracture toughness and Vickers hardness of CeO 2 -stabilized
tetragonal ZrO 2 polycrystals (Ce-TZP). J. Mater. Sci., 1985; 20: 1178-1184.
[9] M. Hirano, H. Inada, Fracture toughness, strength and Vickers hardness of yttria-ceria-doped tetragonal
zirconia/alumina composites fabricated by hot isostatic pressing, J Mater Sci, 1992; 27: 3511-3518.

55
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Effect of MWCNTs Addition on the Wear and Compressive Deformation Behavior of

LM13-SiC-MWCNTs Hybrid Composites

Bishnu Nand Yadav, Gaurav Kumar Verma, Dilip Muchhala, D.P. Mondal
CSIR-Advanced Material and Processes Research Institute, Bhopal-462026

Abstract: Aluminium-SiC-CNT nano-hybrid composite is prepared through stir casting technique. Here SiC and
aluminium alloy powder was used as secondary reinforcement for uniform distribution of CNT. The wear rate,
coefficient of friction and compressive strength of hybrid composite increase with increase in CNT concentration. But
when only CNT is used as reinforcement its strength, wear rate and microhardness values are found to be significantly
less than other composite materials. This is due to mechanical interlocking of SiC, greater strengthening of matrix due
to uniform CNT dispersion. When CNTs are used separately as reinforcement, it get agglomerated and thus the
strength and wear rate of composite decreased.

1. Introduction
The aluminum based metal matrix composites are potential materials for engineering as well as automotive
applications. Composites properties considerably increases when it converts in to hybrid composites, in which two
or more than two types of reinforcing particles used. Currently, researches are being carried out to evaluate the
effect of nanoparticles on tribological and mechanical behavior of hybrid composites. Nano particles are difficult to
disperse which have greater tendency of agglomeration. Hence, for uniform distribution of nano particles, a
secondary dispersion is required. The secondary dispersion would also improve the properties of the alloy. CNTs [1,
2] are attractive due to their light weight, corrosion resistance, flexibility, high aspect ratio, high conductivity and
high mechanical strength. In view these, the CNTs are found to be reinforcing agents in metal matrix for wide range
of applications. Effect of CNTs addition on the wear rate of aluminium alloy was investigated by some researchers
and it was reported that wear rate decreased significantly due addition of CNTs[3, 4]. On the other hand, SiC is the
ceramic particle having very attractive properties like high melting point, light weight, high abrasive resistance, high
thermal stability, high hardness and high thermal shock resistance. These are commonly used for making aluminium
matrix composite for improvement in strength, modulus and wear resistance [5, 6]. Zhou and coworkers [3] made
an attempt to synthesize aluminium composite reinforced with CNTs through presureless infiltration under N2
atmosphere at 800˚C. According to these investigations, CNTs particles get embedded in aluminium matrix quite
effectively. The wear rate and co-efficient of friction (COF) of composite decreased with increase in volume fraction
of CNTs. However, when these composites are made by these investigators through powder metallurgy route
following spark plasma sintering of aluminium alloy CNTs mixture, the mechanical properties and micro hardness
starts reducing when CNTs concentration increased above 1.5 wt%. It is reported by these investigators that this is
primarily due to increased tendency of CNT agglomeration [7]. Attempt was made by a group of investigators [8] to
improve CNTs dispersion through its acid treatment followed by ultra-sonication prior to mixing with aluminium
powder. Lin-zhi et al also reported in their work that above 1.5 wt. %, CNTs starts clustering to a greater extent
when its concentration in aluminium which deteriorate the mechanical properties of composites. They further
Effect of MWCNTs addition on the wear and compressive deformation behavior of LM13-SiC-MWCNTs hybrid composites
reported that the minimum average wear rate and coefficient of friction of the AlSi10Mg-CNTs composites is
obtained in case of 1.0 wt. % CNTs reinforcement. Aluminium alloy when reinforced with SiC particles, the wear
rate also decreased [9].The coarser SiC particles acts as a protuberance against the asperities over the ally surface
and reduce wear rate and coefficient of friction [10]. Thus, if MWCNT and SiC are added in combination, their
synergic effect could lead to further improvement in strength and wear resistance of alloy. However, to the best of
our knowledge, no attempts are made to examine synergic effect of SiC and MWCNTs in wear and compressive
behavior of aluminium composite. The present paper deals with the preparation of hybrid LM13-SiC-MWCNT
composite and to understand the effect of these particle on the dry sliding wear and compressive behavior.

2. Experimental
LM13 Al alloy (Table.1) is used as the matrix material, and SiC particle (size: 40-60 μm) and multiwall CNTs
(MWCNT) are used as reinforcement. SiC concentration varied in the range of 0-10wt% and whereas MWCNTs
concentration varies in the range of 0.5 to 1.5 wt% (Table.2). The length and diameter of MWCNTs are 25±5 μm
and 150±20 nm respectively. The as-recieved MWCNTs are in agglomerated forms. In order to get better
distribution, these MWCNTs are functionalized through acid treatment followed by ultra-sonication and finally
cleaning with water and acetone. The functionalization of MWCNTs was performed by adding 1 g of MWCNTs in
200 ml concentrated HNO3 [60% (v/v)] which is heated at 900C for 24 hrs followed by cooling. After cooling,
modified MWCNTs were collected by filtration and washed repeatedly with distilled water till neutralization.
Finally, it was washed with acetone to make sure that no water molecules are trapped inside the tube bundles. After
that, the entangled MWCNTs were stirred with ultrasonic stirrer in acetone (1gm MWCNTs in 1 lit. acetone) for 3h.
After sonication MWCNTs mixed with Al powder (1 g MWCNT for each 100 g of Al powder) using 5:1 ball to
powder ratio at 150 rpm. Milled Al-MWCNTs powders are dried in oven and made loose tablets with the help of Al-
foil and then added in to the melt. Microstructure of as received and functionalized MWCNTs are shown in Fig.1 (a)
and (b) respectively. The MWCNTs after milling with aluminium powder is shown in Fig.1(c).

Table-1: Chemical composition of LM13 alloy (in Wt. %)

Si Ni Fe Cu Mg Mn Ti Zn Al
11.5 1.5 1 1.2 1.1 0.5 0.2 0.5 82.5

Table-2: Reinforcement composition variations (in Wt. %)

No. 1 2 3 4 5
SiC 10% 9.5% 9% 8.5% 0%
MWCNT 0% 0.5% 1% 1.5% 1.5%

For synthesis of composites, firstly ingot of LM13 alloys was cleaned and put in electrical resistance furnace and
melted at 690°C. Simultaneously SiC particles were put in to another furnace for preheating at temperature of 640°C
for vaporizing the volatile material and minimized the temperature difference between melt and SiC particles during
addition. After melting the alloy, preheated SiC particles, and Aluminium-MWCNTs powders mixtures are mixed
and then added into the melt. Simultaneously, the melt is stirred mechanically at a speed of 750-780 rpm until the
reinforcing mixture are completely added into the melt. Stirring is continued for 2 minutes after reinforcing addition

57
 Proceedings of NCAMMM - 2018

is completed for better distribution. After well mixing, melt was poured in to preheated and graphite coated cast
iron finger dies having dimensions as 20mm diameter and 150 mm height for sample preparation (Fig.2). The
samples for metallographic characterization, wear tests and compression tests were taken from the cast fingers. For
metallographic characterization, samples were polished using standard methodology and finally etched in Keller’s
reagent. Prior to SEM (JEOL-5600 SEM) examinations, the samples were gold sputtered. The polished and etched
Samples were gold coated before SEM characterization. The Compression tests of the hybrid composite samples
were carried out with the help of UTM (INSTROM UTM Model no. 8801) at a strain rate of 0.01/s. The samples
dimensions were of 10mm in diameter and 15mm length. During testing, a solid lubricant, MoS 2, was used between
sample and machine ram surfaces to minimize the friction. The stress strain data are recorded in the form of raw data
as well as stress-strain curves within the system interfaced with the UTM.

Figure 1: Microstructure of (a) received MWCNTs, (b) functionalized MWCNTs, (c) after milled with Al powders. Figure 2:
Cast iron die images (a) semi assembled, (b) split, (c) assembled

Dry sliding wear tests were done through pin on disc method in which dimension of pin samples (hybrid
composites) were 10 mm in diameter and 30mm height. During this tests, the disc is made with AISI-4340 hardened
steel (hardness 640 HV). Both pin and disc were well cleaned with ethanol every time prior to and after the testing.
The tests were carried out at three different loads of 2kg, 4kg, and 6kg for a sliding distance up to 2km with at a
fixed track radius of 50 mm. The test was carried out at three different speeds of 382 rpm, 572 rpm and 764 rpm
corresponding to linear speed of 2 m/s, 3 m/s and 4 m/s respectively. Samples were weighted before and after every
testing using a weighing machine with an accuracy of 0.0001gm. Coefficient of friction were calculated on the basis
of recorded frictional forces using the equation μ= F/N, where F is the frictional force and N is applied normal load
during testing. Microhardness on polished and etched hybrid composites are examined at a load of 100g.

4. Result and discussion

4.1 Microstructure
The microstructure of base alloy is shown in Fig.3 (a) indicating dendtric structures of aluminium and eutectic
silicon at the interdendritic region. The microstructure of LM13 10 wt% SiC composite showed uniform distribution
of SiC particles with in matrix (Fig.3 (b)). LM13-SiC-MWCNTs hybrid composites with 8.5 wt% SiC and 1.5 wt%
MWCNT also exhibits uniform distribution of SiC particles (Fig.3 (c)). Relatively longer MWCNTs are noted
within the matrix (marked arrow). But, all the MWCNTs are not resolved. In order to examine the uniformity is
distribution of MWCNTs, microhardness values are taken in different locations and compared. It is noted that the

58
Effect of MWCNTs addition on the wear and compressive deformation behavior of LM13-SiC-MWCNTs hybrid composites
microharness values increased with increased in MWCNT addition and the scatter associated with the
microhardness values are less than 4% (Fig.4). However, in LM13 1.5 wt% MWCNT composites the error bars are
very high (15%) indicating nonuniform distribution of MWCNTs. However, it is noted clearly that the
microhardness of LM13-1.5 wt% MWCNTs are relatively higher than that of LM13-10 wt% SiC composite, but
significantly higher than the matrix alloy. This indicate that addition of both SiC and MWCNTs increase the matrix
strength. But Strengthening due to MWCNTs is higher than that obtained through SiC addition. But, at the same
time, there might be greater chance of agglomeration of MWCNTs, causing larger scatter in microhardess values,
This may be resulting in less strengthening when only 1.5 wt% MWCNTS are used as reinforcement separately as
reinforcement in LM13 alloy. The improvement in microhardness led to the fact that there must have improvement
in wear resistance and strength of alloy due to hybridization of SiC and MWCNTs as reinforcement.

Figure 3: Microstructure of (a) composite reinforced with 10% SiC and 0% MWCNTs, (b) hybrid composite reinforced with
8.5% SiC, 1.5% MWCNTs, (c) base alloy (LM-13)

Compressive Deformation behavior

Figure 5 represents the true stress strain curves of the investigated materials under compression loading at a strain
rate of 0.1/s. It is observed that the nature of all curves are similar irrespective of the materials. It is also found that
the young modulus increases marginally, while yield stress increases significantly due to reinforcement with both
SiC with MWCNTs together. It is noted that even if the SiC concentration is reduced and MWCNT contents
increases, the yield strength and flow stress increases monotonically. But when 1.5 wt. % MWCNTs are added
separately without adding any SiC, the yield strength, modulus and the flow stress reduced significantly. Even this
become less than that of LM13-10 wt% SiC composites. This figure thus demonstrates that addition of MWCNTs
along with SiC are advantageous for improvement in strength and modulus of composites. Significantly less strength
in LM13-1.5 wt% MWCNT composites is due to greater degree of agglomeration of MWCNTs. Where MWCNTs
are agglomerated, the microhardness values are very less. Again where there is no MWCNTs, the strength will gain
will be less. AS a results there is higher degree of scatter in the microhardness values of this composites.

5. Wear Behavior
The wear rate as function of applied load at fixed of 2m/s, 3m/s and 4m/s are shown in Fig. 6(a), (b) and (c)
respectively. It is clearly observed that the wear rate(mm3 /m) of hybrid composites (where both SiC and MWCNTs
are added as reinforcement) with increase in MWCNT concentration, even though SiC concentration reduces. This
indicates that the efficacy of MWCNT is more as compared to SiC addition for improvement in wear resistance.

59
 Proceedings of NCAMMM - 2018

Similar fact is also true for improvement in strength. However, irrespective of materials, the wear rate increases
almost linearly with applied load. If one compared the wear rate of different materials, the maximum wear rate is
noted in case of LM13-1.5wt% MWCNT; even greater than that of LM13-10wt% SiC composites. The trend is
similar in

Figure 4: Micro hardness of hybrid composites Figure 5: True stress-strain plot of hybrid composites

all the sliding speed. Similarly, The coefficient of friction as a function of applied load at three different speed of
2m/s, 3m/s and 4m/s are shown in Fig.7(a), (b) and (c) respectively. The coefficient of friction increases with
increase in applied load irrespective of materials. But interestingly it is noted that LM13-1.5 wt% MWCNT
composites exhibited the minimum coefficient of friction. It is also noted that the coefficient of friction is highest in
case of LM13-10 wt% SiC composite. In hybrid composite, the coefficient of friction decreases monotonically with
decrease in MWCNT concentration. This may be due to the fact that MWCNTs act as lubricant. With increase in
MWCNT the lubrication action increases.

Figure 6: Wear rate as function of applied load at (a) 2m/s, (b) 3m/s and (c) 4m/s.

Even though the coefficient of friction decreases in LM13 -1.5 wt% MWCNT composites, it exhibited the poorest
wear resistance. This id due to the fact that when higher amount of MWCNTs are added separately within the
matrix, the agglomeration tendency increases. This led to reduction in strength. The MWCNT mat formation is
reduced. Because of lower strength, asperities may penetrate more leading to higher materials cut due to plaughing
or there is greater tendency of adhesion in localized region. Then the is a possibility of increase in coefficient of
friction. But, because of higher temperature rise, again the slipping action do predominant. Under the presence of
MWCTS the slipping action further increases. When SiC and MWCNTs both are present, the MWCNT entangled
SiC particles and SiC particles act as protuberances on the matrix protecting it from effective asperity contact. Fine

60
Effect of MWCNTs addition on the wear and compressive deformation behavior of LM13-SiC-MWCNTs hybrid composites
and fragmented SiC along with MWCNT get embedded in and mixed into the specimen surface and increase the
subsurface strength particularly. This also causes reduction in wear rate to some extent at higher temperature.
Uniform distribution of MWCNTs also make relatively stable surface matting along with the aligned eutectic silicon
needle. All these fact causes more stable mechanically mixed layer on the subsurface, which causes reduced wear
rate. As the MWCNTs are uniformly dispersed it increases the lubricating nature of the surface.

Figure 7: Coefficient of friction as applied load at (a) 2m/s, (b) 3m/s and (c) 4m/s

In case of only SiC reinforcement, the MWCNTs mats are not there. There is possibility of grater asperity penetration.
Matrix is also not lubricated. Hence there is a possibility of more localized adhesion, delamination of SiC particles as
well greater cutting action. These will increase greater mechanical locking and required higher frictional force for its
sliding as well as wear rate. The MWCNTs when added with SiC may get uniformly distributed and entangled the SiC
particles, making stronger bonding with the matrix. The wear debrises which are formed may also be locked with the
longer and stronger MWCNTs causing restriction in materials removal and increase in wear resistance.

7. Conclusions:
o Increasing the percentage of MWCNTs in LM13-SiC-MWCNTs hybrid composites reduces the wear rate
and coefficient of friction of hybrid composites.
o Addition of MWCNTs in making Al-SiC-MWCNT hybrid composites enhances the compressive strength
of composites.
o Micro hardness of LM13-SiC-MWCNT hybrid composites is also affected with increasing the percentage
of MWCNTs and it increases with increase in percentage of MWCNTs in LM13-SiC-MWCNT hybrid
composites.

References
[1] C.A.I. Merino, J.L. Sillas, J. Meza, J.H. Ramirez, Journal of Alloys and Compounds, 707 (2017) 257-263.
[2] J. Liang, M.C. Saha, M.C. Altan, Procedia Engineering, 56 (2013) 814-820.
[3] S.-m. Zhou, X.-b. Zhang, Z.-p. Ding, C.-y. Min, G.-l. Xu, W.-m. Zhu, Composites Part A: Applied Science and
Manufacturing, 38 (2007) 301-306.
[4] H. Kwon, M. Saarna, S. Yoon, A. Weidenkaff, M. Leparoux, Materials Science and Engineering: A, 590 (2014)
338-345.
[5] T. Huber, H.-P. Degischer, G. Lefranc, T. Schmitt, Composites Science and Technology, 66 (2006) 2206-2217.

61
 Proceedings of NCAMMM - 2018

[6] J.-M. Molina, M. Rhême, J. Carron, L. Weber, Scripta Materialia, 58 (2008) 393-396.
[7] L.-z. Wang, Y. Liu, J.-j. Wu, X. Zhang, International Journal of Minerals, Metallurgy, and Materials, 24 (2017)
584-593.
[8] B. Guo, X. Zhang, X. Cen, X. Wang, M. Song, S. Ni, J. Yi, T. Shen, Y. Du, Materials Characterization, (2017).
[9] A. Alpas, J. Zhang, wear, 155 (1992) 83-104.
[10] Y. Liang, Z. Ma, S. Li, S. Li, J. Bi, Journal of Materials Science Letters, 14 (1995) 114-116.

62
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Cyclic Oxidation Behavior of ZrB 2 -SiC Based Ultra High Temperature Ceramic Composite

Sayandip Sarkar and Manab Mallik*

Department of Metallurgical and Materials Engineering, National Institute of Technology, Durgapur - 713209

Abstract: Present investigation focuses cyclic oxidation behavior of ZrB 2 -25 vol.% SiC (20 vol% SiC p and 5 vol%
SiC w ) composite. Composite was prepared by pressure-less sintering at 2000 °C for 2 hr. The B 4 C and C were
added as sintering additives. The cyclic oxidation test has been conducted at two different temperatures 1200 °C
and 1300°C for 6hrs in air. Mass main was measured after each cycle at every oxidation temperature. The phases
present on the oxide scale were determined by X-ray diffraction (XRD) analysis. The structures of the oxide scales of
the specimens were examined using a field emission scanning electron microscope (FESEM-EDX). Results show
that mass gain occurred due to formation of ZrO 2 and SiO 2 at elevated temperatures. The main oxidation products
were monoclinic ZrO 2 and silica. At 1200 °C and above, the presence of SiC particles markedly improves the
resistance to oxidation of the composite.
Keywords: Zirconium diboride; cyclic oxidation; ceramic composite

1. Introduction
Ultra high temperature ceramics (UHTC) that have very high melting point (>30000C) are suitable for application in
reusable thermal protection system (TPS) of hypersonic vehicles [1]. Transition metal borides, nitrides and carbides
are considered as Ultra high temperature ceramic. Amongst the all ultrahigh temperature ceramics ZrB 2 is very
popular because it has low density in the di-boride family. Beside these it has good thermal and electrical
conductivity, thermal shock resistance properties. Above 10000C preferential oxidation of B 2 O 3 makes the oxide
layer less protective. In the 1000-1400°C range linear oxidation kinetics of ZrB 2 is observed. Oxidation resistance is
one of the key issues for the use of UHTCs. Due to low oxidation resistance, monolithic ZrB 2 and HfB 2 are not
suitable for ultra high temperature applications [1]. The introduction of second phases has shown improvement in
the oxidation resistance and mechanical properties of UHT. Several researchers have reported that the oxidation
resistance of ZrB 2 and HfB 2 improved significantly by addition of SiC, which reduces oxidation rate by forming
silica rich scale [1,2].To improve its oxidation resistance SiC in the range of 10-30 vol% is added as it forms a SiO 2
based glassy layer which provides protection to passive oxidation in the temperature range 1100-16000C or higher
[2-4]. Addition of SiC also improves its flexural strength. Therefore ZrB 2 -SiC ceramics have found a very high
attention in the last decade. Different researchers have been reported on the oxidation behavior of ZrB 2 -SiC
ceramics in isothermal and non-isothermal conditions. But there are very few works on cyclic oxidation behavior of
this composite. Cyclic oxidation results provide us a realistic data on stability of oxide scales under different
oxidation conditions that includes heating to a particular temperature along with the effect of thermal stresses which
is associated with thermal cycles. Therefore, in this study our main aim is to study the cyclic oxidation behavior of
pressure-less sintered ZrB 2 -SiC ultra high temperature ceramic composite.
Cyclic oxidation behavior of ZrB2-SiC based ultra high temperature ceramic composite

2. Experimental Procedure
The ZrB 2 , SiC and B 4 C powders with more than 99% purity have been obtained from Al faesar. Phenolic resin was
added as a binder for the green parts as well as source of carbon. Raw powders and phenolic resin were mixed in a
planetary mono mill using a speed of 250 rpm for 2 h to promote intimate mixing. The ZrB 2 -25 vol.% SiC (20 vol%
SiC p and 5 vol% SiC w ) composite has been prepared by pressure-less sintering at 2000 °C for 2 hr. The B4C and C
were added as a sintering aid. Sintering technique has been followed from procedure described by Mallik et al [5].
The density of the composite was measured using Archimedes principle. The sintered sample was sectioned using
Isomet slow speed diamond saw. The sectioned samples were polished metallographically and X-Ray diffraction
(XRD)analysis delineated the constituent phases in the microstructures. The microstructure of the composite has
been examined using field emission scanning electron microscopes(FESEM). Samples with dimensions of 10 mm X
4 mm X 4 mm were sliced and metallographically polished prior to oxidation studies. The cyclic oxidation test was
done for 6 cycles at 1200°C and 1300°C. One cycle includes measuring the weight of sample before oxidation,
isothermal holding for 1 hour, kept for cooling to room temperature, measuring weight to record mass gain or loss.
The constituent phases of the oxide scales were determined by XRD analysis. The structures of the oxide scales of
the specimens were examined using a FESEM.

3. Results and Discussions

The sintered composite indicates 99% of the theoretical density. The main phases present in the pressure-less
sintered ZrB 2 -SiC composite have been examined by X-ray diffraction (XRD) analysis and pattern is shown in Fig.
1. XRD pattern indicates that constituent phases of the investigated composites are primarily ZrB 2 and SiC.
Micrograph of ZrB 2 -25 vol.% SiC composite is depicted in Fig.2 and the microstructure exhibits uniform
distribution of the phases.

Figure 1: XRD pattern of pressure-less Figure2: Microstructure of ZrB2-25 vol.

sintered ZrB2-25 vol.% SiC composite % SiC composite

The results of cyclic oxidation studies at 1200 °C and 1300 °C for 6 h, as shown in Fig. 3, reveal that a stable oxide
scale is formed after 5 h exposure at 1200 °C, whereas continuous mass gain is observed at 1300 °C. The net mass-
gain of the composite after exposure for 6 h at 1200°C and 1300 °Care2.6 mg/cm2 and 4.9 mg/cm2, respectively
Greater mass gain per unit area is observed for exposure of the specimen at 1300 °C than at 1200°C. The

64
 Proceedings of NCAMMM - 2018

temperature exposure at 1200 °C shows little mass gain up to 3 cycles beyond which the specimen shows little mass
loss and finally mass increases rapidly. Whereas, linear mass gain kinetics is observed for the temperature exposure
at 1300 °C up to 3 cycles beyond which rate of mass gain decreases. The mass gain is attributed to formation of
ZrO 2 and SiO 2 however; the mass loss is caused by evaporation of B 2 O 3 .

The

Figure 4: XRD pattern of oxide scale of

Figure 3: Oxidation behavior of ZrB2- ZrB2-25 vol. % SiC composite after
25 vol. % SiC composite cyclic oxidation at 1300°C

Mechanism of oxidation has been analysed by examination of the oxidation products using XRD and FESEM. A
result of XRD analysis show (Fig. 4) peaks of ZrO2 is the major constituent phase along withZrSiO 4 phase. But
EDX spectra (Fig. 5c) from the oxide scales have shown additional evidence of SiO 2 . Absence of SiO 2 -peaks in the
XRD pattern is due to its amorphous character. Figures 5 a and b show typical FESEM images of the oxide scales
formed onZrB 2 -25 vol. % SiC composite after oxidation at 1200 °C and 1300 °C, respectively. The products of
oxidation comprise of ZrO 2 and SiO 2 rich glassy phase. Oxide scales (Fig. 5a) reveal several cracks and peeled off
oxidation layers on the composite. In addition oxide scale formed at 1300 °C reveals some bubbled morphology.
The reactions involved during oxidation at1200 °C and 1300 °C are:

ZrB 2 (c) + 5/2O 2 (g) = ZrO 2 + B 2 O 3 (l) (1)

SiC(c) + 3/2O 2 (g) = CO(g) + SiO 2 (c) (2)
SiC(c) + O 2 (g) = CO(g) + SiO 2 (g) (3)
SiC(c) + 1/2O 2 (g) = C(c) + SiO 2 (g) (4)

65
Cyclic oxidation behavior of ZrB2-SiC based ultra high temperature ceramic composite

Equation (2) & (3) are related to the passive oxidation of SiC. The outer glassy oxide layer is found to be enriching
with SiO 2 . The subscale of oxide layer contains ZrO 2 with SiO 2. The presence of SiC assists the formation of
borosilicate glass layer. The SiO2 rich layer provides an effective diffusion barrier for oxygen anions, and
consequently protects the composites from oxidation [3, 6, 7]. Hence the gain in mass can be attributed to occur due
to the formation of ZrO 2 and SiO 2 while the loss in mass is considered due to the evaporation of B 2 O 3 and escape
of CO.

(c)
Element Weight% Atomic%

BK 2.16 3.92

OK 60.84 74.54

Figure 5: SEM micrograph of the oxide scale formed onZrB 2 -25 vol. % SiC composite after oxidation at (a) 1200
°C, (b) 1300 °C and (c) typical EDX spectrum of oxide scale

4. Summary
Cyclic oxidation behaviors of ZrB 2 -25 vol.% SiCcomposite have been studiedat 1200 °C and 1300 °Cfor 6hrs.
Monitoring mass change and examining oxide scales draw following conclusions:
(i) Mass gain of ZrB 2 -25 vol.% SiCcomposite increased with increasing temperature and time.
(ii) Mass gainoccurred due to formation of ZrO 2 and SiO 2 , at elevated temperatures.
(iii) The main oxidation products were monoclinic ZrO 2 and silica.
(iv) At1200 °C and above, the presence of SiC particles markedly improves the resistance to oxidation of the
composite.

66
 Proceedings of NCAMMM - 2018

Acknowledgements
The financial supports from sponsors project [Project Number=DST, SERB, SB/EMEQ-251/2013]are gratefully
acknowledged. The authors also express their sincere gratitude to technical personnel at the Central Research
Facility of IIT Kharagpur and COE NIT Durgapur.

References
[1] Levine SR, Opila EJ, Halbig MC, Kiser JD, Singh M, Salem JA.Evaluation of ultra-high temperature ceramics
foraeropropulsion use.Journal of the European Ceramic Society. 2002 Jan 1;22(14-15):2757-67.
[2] Mitra R, Upender S, Mallik M, Chakraborty S, Ray KK. Mechanical, thermal and oxidation behaviour of
zirconium diboride based ultra-high temperature ceramic composites. InKey Engineering Materials 2009 (Vol.
395, pp. 55-68).Trans Tech Publications.
[3] Opeka MM, Talmy IG, Wuchina EJ, Zaykoski JA, Causey SJ. Mechanical, thermal, and oxidation properties of
refractory hafnium and zirconium compounds.Journal of the European Ceramic Society. 1999 Oct 1;19(13-
14):2405-14.
[4] Monteverde F, Bellosi A. The resistance to oxidation of an HfB 2 –SiC composite.Journal of the European
Ceramic Society. 2005 May 1;25(7):1025-31.
[5] Mallik M, Roy S, Ray KK, Mitra R. Effect of SiC content, additives and process parameters on densification
and structure–property relations of pressureless sintered ZrB 2 –SiC composites. Ceramics International. 2013
Apr 1;39(3):2915-32.
[6] Mallik M, Ray KK, Mitra R. Oxidation behavior of hot pressed ZrB 2 –SiC and HfB 2 –SiC composites. Journal
of the European Ceramic Society. 2011 Jan 1;31(1-2):199-215.
[7] Mallik M, Ray KK, Mitra R. Effect of Si 3 N 4 Addition on Oxidation Resistance of ZrB 2 -SiC Composites.
Coatings. 2017 Jun 30;7(7):92.

67
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Effect of CNT-Ni-P Composite Coating on Tribological Behaviour for Brake Pad System

Atul Kumar Harmukh, Santosh Kumar, Sushma Bharti, Subrata Kumar Ghosh
Department of Mechanical Engineering, Indian Institute of Technology (ISM), Dhanbad
Email: subratarec@yahoo.co.in

Abstract: Mechanical and tribological properties of interacting surfaces are major issue for engineering
applications. The aim of the paper is to analyse the surface behaviour in terms of friction and wear of brake pad
system. In the present study, the electroless coating technique is used to improve the surface properties of plain
brake pad. Carbon nanotubes are used as a coating material with base Ni-P to increase the properties like wear and
chemical stability of surface. Energy dispersive x-ray spectroscopy (EDX) test is used to find the constituent of
material as a friction producer, fillers material, binders and reinforcement. The wear and friction parameters are
studied by pin-on-disc type of tribometer. It has been observed that the coated samples show better surface
properties. Friction force and wear rates show lesser in CNT-Ni-P composite coated sample than without coated
sample.
Keywords: carbon nanotube, tribological behaviour, electroless coating, wear, brake pad

1. Introduction
In the brake pad system, the coating technique is most effective to reduce the wear. In the term of surface coating
science, various types of coating process are there, electroless plating is one of them. The main advantage of the
electroless coating gives better strength, uniform coating thickness, good hardness, etc. Z.H.Li et al. [1] concluded
that the CNT-Ni-P composite coating reduces the wear rate as compared to the uncoated and the Ni-coated
substrates of the plain steel surface. The CNT is mainly improve the wear resistance in the CNT-Ni-P composite
coating and Ni and P are mainly used to improve the corrosion resistance of the surface. The decrement in the
coefficient of friction in CNT-Ni-P composite coating with increase of load and wear time.
The Ni-Cu-P/CNT and Ni-Cu-CNT composite coating is used to reduce the wear rate and by this composite coating
the hardness of the surface is increased and due to the tight arrangement of nano CNT the corrosion resistance has
been improved [2]. L.Y.Wang et al. [3] concluded that the certain amount of volume fraction of CNT reduce the
wear and friction. But, further increase the volume of fraction of CNT increase the wear and friction and cracking
and spalling will be started in the worn out surface and coefficient of friction will be increased. Byung-Joo Kim et
al. [4] investigated that the Ni dispersed in the MWCNT coating, the interfacial mechanical property is improved
and rheologoical behavior like suspension viscosity, storage and loss moduli is improved.
The Ni-P-Cr composite coating of double layer were used in the surfaces and it concluded that the double
layer coating provide the much hardness as compared to the electroless coating technique. The nickel and
phosphorous coating in the inner layer and chromium in the outer layer provide the good corrosion resistance as well
as the hardness [6]. Elsa Georjiza et al. [7] were studied on the Ni-B-SiC composite electroless coating method and
the Ni-B improves the corrosion resistance and the SiC provide strength to the surface. As a result, the friction and
wear of the surface are decreased and hardness of the surface is improved. T.Ram Prabhu et al. [8] concluded that
Effect of CNT-Ni-P composite coating on tribological behaviour for brake pad system

the reinforcement of Cu/SiC with Gr hybrid composite coated material improve the corrosive resistance and wear
property of the material. In the composite material the compression and the flexural strength are increased by the
multi layer walled of the material or multi layer reinforcement of the material.
Ke Duan et al. [9] concluded that the Carbon nanotube with the Ni-P composite coating
enhanced the damping quality of material and it also concluded that the damping quality of material depends upon
the coating material. In the pipeline steel material by the use of Ni-P-Ti composite coating, it can reduce the wear
and corrosion resistance of the material and enhance the mechanical properties [10]. By the use of CNT-Ni-P, the
chemical stability and hot hardness of the material was increased and morphology of the surface had been
improved [11]. This investigation is based on the brake pad material to reduce the wear, friction and improve the
mechanical properties and life span through electroless coating of carbon nanotubes.

2. Methodology
The friction and wear test of brake pad material was performed on the pin-on-disc tribometer. The disc is made of
gray cast iron and diameter and thickness were 100 mm and 10 mm respectively. The roughness value of the disc is
0.18 µm. The diameter and length of pin were 6 mm and 20 mm respectively .The composition of the pin material is
investigated by the “Energy dispersive X-Ray spectroscopy” (EDX) test. Figure 1(a) and Figure1(b) shows the
elemental analysis of brake pad samples without coated and with coated respectively.

(a) (b)
Figure 1: Elemental analysis of brake pad pin by EDX (a) elemental analysis of brake pad pin without coating (b)
elemental analysis of brake pad pin with coating

Elemental analysis of brake pad pin of without coated sample and with coated sample has also been shown in Table
1. The electroless coating technique was used to change the tribological behaviour of brake pad material. In this
coating technique, the base solution was prepared by using all the chemicals as given in Table 2. Firstly, all the
chemicals are mixed in deionised water (200ml). The magnetic stirrer was used for 30 min to make the base
solution. In this base solution nickel sulphate, sodium hypophosphite, aluminium chloride and sodium citrate were
used as complexing agent, metal ion source and as a buffering solution respectively.

69
 Proceedings of NCAMMM - 2018

Table 1: Elemental composition of brake pad pin material

Sl. no. brake pad sample without coated brake pad sample with coated
material Wt % material Wt %
1. C 62.46 C 29.71
2. O 28.82 O 41.98
3. Na 0.26 Na 4.12
4. Mg 0.89 Mg 0.27
5. Al 0.89 Al 0.37
6. Si 2.27 Si 0.80
7. S 0.86 P 4.11
8. Cl 0.10 S 0.68
9. K 0.18 K 0.14
10. Ca 0.62 Ca 0.24
11. Fe 1.95 Fe 6.52
12. Cu 0.17 Ni 10.88
13. Ba 0.51 Ba 0.24

Table 2: Bath concentration of base solution(Ni-P)

Sl no. Chemicals Weight (gm)

1. Nickel sulphate 8 gm
2. Sodium hypophosphite 10 gm
3. Sodium citrate 5 gm
4. Aluminium chloride 10 gm

(a) (b) (c)

Figure 2: (a) Coating performed in base solution, (b) CNT mixed solution with base solution, (c) pin-on-disc
tribometer

The brake pad pin is dipped in base solution for 30 min at temperature of ˚C.
80 In this process, Ni -P coating has
been found on the brake pad sample. Base solution is cooled to room temperature and mixed with 2.5 gm of carbon
nano-tube (CNT). Stable solution of carbon nano-tube has prepared with help of ultrasonicator. Again, the brake pad
˚C. The temperature of 80˚C is maintained fo r 30
pin is dipped into the mixed solution of CNT and heated upto 80
min for coating. Elecroless coating arrangement shown in Figure 2 (a) and Figure 2 (b). Tribological properties such
as coefficient of friction and wear of brake pad samples was measured by using Pin-on-disc tribometer as shown in
Figure 2 (c).

3. Results and discussion

Figure 3 shows the field emission scanning electro microscopy (FESEM) images of brake pad sample, which
validates the coating of CNT with Ni-P base material on the surface of brake pad sample.

70
Effect of CNT-Ni-P composite coating on tribological behaviour for brake pad system

(a) (b)
Figure 3: FESEM image of Brake pad pin samples (a) without coating (b) with coating

Wear and Friction investigation has been performed on coated and without coated brake pad pin. Figure 4 shows the
variation of wear with time for brake pad pin with and without coating. It has been observed that the wear rate
increases with time, but it is less for coated material. Figure 5 shows the variation of Friction with respect to time
for brake pad pin with and without coating. It has been observed that the coefficient of friction increases with time,
but it is also less for coated material.
From the experimental investigation, It has been observed that the wear of coated material has lesser value
around 10-20 percent than without coated pin. The reason may be due to the Ni-P composite coating with CNT
make strong complex coordination bond with brake pad material. This composite coating improve the surface
morphology of brake pad and hence, the wear resistance is improved as compare to the uncoated material.

Figure 4: Variation of wear with time at 3kg and 4kg

71
 Proceedings of NCAMMM - 2018

Figure 5: Variation of coefficient of friction with time at 3kg and 4kg

4. Conclusions
In this investigation, it has been revealed that the wear rate and coefficient of friction decreases with time for coated
samples with respect to without coated samples. The results show that coated surface hardness is more than without
coated sample. Therefore, CNT-Ni-P composite coating may be applied to enhance the surface properties of brake
pad.

References
[1] Li ZH, Wang X Q. Preparation and tribological properties of CNT-Ni-P composite coating. Tribology
international. 2006; 39: 953-957.
[2] Yucheng Wu, Rong Ren. Preparation and characterization of Ni-Cu-P/CNT quaternary electroless coatin .
Materials research bulletin. 2008; 43: 3425-3432.
[3] Wang LY, Tu J P. Friction and wear behaviour of electroless Ni-based CNT composite coating. Wear. 2003;
254: 1289-1293.
[4] Kim Byung-Joo,Bae Kyong-Min. Roles of Ni/CNTs hybridization on rheological and mechanical properties
of CNTs/epoxy nanocomposites. Materials Science and Engineering. 2011; 528: 4953–4957.
[5] Alishahi Mostafa, Monirvaghefi Seyed Mahmoud. The effect of carbon nanotubes on the corrosion and
tribological behavior of electroless Ni–P–CNT composite coating.Applied Surface Science.2012; 258: 2439–
2446.
[6] Wang Qin-Ying , Xi Yu-Chen. Study on properties of double-layered Ni–P–Cr composite coating prepared
by the combination of electroless plating and pack cementation. Journal of alloys and compound. 2017; 729:
787-795.
[7] Georjiza Elsa, Gouda Venice. Production and properties of composite electroless Ni-B-SiC coatings. surface
and coating technology. 2017 ; 325: 46-51.
[8] Ramprabhu T, Varma V K. tribological and mechanical behaviour of multilayer Cu/SiC-Gr hybrid
composites for brake friction material application. Wear. 2014; 317: 201-212.

72
Effect of CNT-Ni-P composite coating on tribological behaviour for brake pad system

[9] Duan Ke, Li Li, Yujin Hu. Damping characteristic of Ni-coated carbon nanotube/copper composite. Materials
and Design. 2017; 133: 455-463.
[10] Wang Chuhong, Farhat Coheir. Indentation and bending behavior of electroless Ni-P-Ti composite coatings
on pipeline steel. Surface and coating technology. 2018; 334: 243-252.
[11] Wang Qianzhi, Callisti Mauro. Evolution of structural, mechanical and tribological properties of Ni–
P/MWCNT coatings as a function of annealing temperature. Surface and coating technology. 2016; 302: 195-
201.

73
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Effect of Interphase and Dispersion of CNTs on the Elastic Properties of CNT-Polyethylene

Nanocomposite

Ashish K. Srivastava*, Vimal K. Pathak*, Mithilesh K. Dixit*

* Assistant Professor, Mechanical Engineering Department, Manipal University, Jaipur, India
Email: ashishkumar.srivastava@jaipur.manipal.edu

Abstract: This paper describes the effect of dispersion of carbon nanotubes (CNTs) within the matrix material and
interfacial bonding between the CNT and polyethylene (PE) matrix on the elastic properties of the nanocomposite
material. Van-der Waals (VdW) force based cohesive zone is used to model the interphase zone as the third phase
between CNT and PE matrix. Thereafter, CNTs along with their interphase zone are dispersed into the matrix with
the aid of Boolean based random sequential adsorption (RSA) technique to model an representative volume element
(RVE). Periodic boundary conditions are applied on the RVE and elastic properties of resulting nanocomposite
material are evaluated by numerical homogenization. From the study, it is found that the interphase zone between
CNT and PE, has the positive effect on the elastic properties of PE-nanocomposite. Further, the effect of CNT
alignment on the elastic properties of the nanocomposite is also studied and it is established that the alignment of
CNTs leads to substantial increase in the axial modulus of the nanocomposite. For only 1 % volume fraction of CNT
reinforcement, the axial modulus of nanocomposite enhances by approximately 75.06 %, which is only 19.47 % in
case of randomly oriented CNT-reinforced PE nanocomposite.
Keywords: Carbon nanotube (CNT), Representative volume element (RVE), Random sequential algorithm (RSA),
Cohesive zone model (CSM), Finite element method (FEM), Elastic properties.

1. Introduction
Carbon nanotubes (CNTs), cylindrical fullerene structure of hexagonally arranged carbon atoms possess elastic
modulus and tensile strength of over 1TPa and 150 GPa respectively (1). These cylindrical fullerene structures are
considered as potential candidates for reinforcement to make advanced structural nanomaterials. To transfer the
extraordinary material properties of CNTs to the matrix material and to avoid agglomeration of CNTs, it is
necessary to obtain the good interfacial bonding between CNT and matrix material along with proper dispersion of
the CNTs into the matrix(2). Thus, to analyze this new generation of composite materials, it is important to consider
the effects of the interphase region between CNT and matrix and the dispersion of CNTs in matrix(3).
The high strength and modulus of CNTs are vain if the enforced load on the nanocomposite material is not
transmitted to the CNTs. Thus a strong interphase between nanostructure and matrix is required for effectual load
transfer(4). To predict the behavior of interphase, the molecular simulations (e.g., molecular dynamics and
mechanics, density functional theory etc) and continuum mechanics (e.g., finite element methods, boundary element
methods, meshfree methods etc) have been widely applied by the investigators due to their particular potentialities to
resolve the problems at nano and micro-scales. Regardless of the potentialities of molecular simulations to capture
the effect of force-field between the atoms, and modeling environment more precisely than continuum mechanics, it
Effect of interphase and dispersion of CNTs on the elastic properties of CNT-Polyethylene nanocomposite

is bounded to little length and time scale, which makes it computationally inefficient for the analysis of
nanocomposites having large length scales.
Therefore continuum mechanics-based finite element method (FEM) plays an important role in minimizing
the simulation cost. Many researchers have employed its potentialities to estimate the elastic constants of composite
materials using the approach of representative volume element (RVE)(5). Shokrieh and Rafiee(6) and Joshi and
Upadhyay(7) have used the spring elements for the interfacial region to model the single-wall and multi-wall CNT
reinforced RVE, respectively.
Since the interfacial region has a substantial effect on the effective elastic constants of nanocomposite
material(8), it is necessary to study its elastic behavior on the proper theoretical background. A generalized
expression for the interfacial potential based on van-der-Waals (vdW) energy between nanofiller (i.e., CNT) and
polyethylene matrix was introduced by Jiang et al.(9). Tan et al.(10) utilized the same nonlinear cohesive law, for
CNT/polymer interfacial region to study the macroscopic behavior of nanocomposite, and reported that at little
strain CNTs amend the elastic behavior of nanocomposite, but at big strain, such enhancements disappear owing to
complete debonding of these nanofillers from the matrix material.
The stiffness properties of nanocomposite also depend highly on the type of dispersion (i.e., random/aligned) of
CNTs in matrix(11). Poor CNTs dispersion lead to the formation of agglomerates and make a negative effect on the
mechanical properties of nanocomposite(12). Randomly-dispersed CNT nanocomposite posses isotropic stiffness
properties whereas, aligned-CNT offer transversely isotropic stiffness properties of nanocomposite(13). Joshi and
Upadhyay(14) reported that highest axial stiffness of nanocomposite is obtained due to the alignment of CNTs in the
direction of loading.
The complex geometry of actual nanocomposite possessing dispersed randomly/aligned CNTs can be
realized through the random sequential adsorption algorithm (RSA)(15), which is frequently applied for modeling
random fiber composite. In RSA technique, initially the reinforcements are placed randomly into a volume, and then
new fillers are considered for adsorption only if it does not overlap with the previously adsorbed reinforcement. The
higher volume fraction of reinforcements in the RVE is achieved by applying geometric periodicity and using
various filler sizes and curved fillers. In 2016, Liu et al.(16) proposed a general computationally efficient Boolean
based technique to generate the RVE possessing the non-overlapping or intersecting fibers of arbitrary shapes,
which can determine the orientation and position of new fiber without iterations.
In this paper, vdW energy based CZM as proposed by Jiang et al.(9) is used to predict the elastic modulus of the
interphase region between CNT and PE. The CNTs having interphase zone are dispersed (randomly/aligned) in the
matrix with the help of Boolean based RSA technique to model the RVE. Thereafter, the stiffness properties of
CNT-nanocomposite are predicted with the enforcement of periodic boundary conditions on the RVE.

2. Young's Modulus of Interfacial Region

The vdW interaction between cylindrical fullerene structure, CNT and matrix material, represented by cohesive zone
model (CZM) proposed by Jiang et al.(9) is employed in the present study to evaluate the elastic modulus of the
interfacial region and is given below:
2𝜋𝜋 2 2𝜎𝜎 𝐼𝐼 9 𝜎𝜎 𝐼𝐼 3 3 𝜎𝜎 𝐼𝐼 8 𝜎𝜎 𝐼𝐼 2
𝜙𝜙 = 𝜌𝜌𝑚𝑚 𝜌𝜌𝑐𝑐 𝜎𝜎𝐼𝐼 3 𝜀𝜀𝐼𝐼 �2𝑅𝑅 � − � + 𝜎𝜎𝐼𝐼 � − �� (1)
3 15𝑟𝑟 9 𝑟𝑟 3 2 10𝑟𝑟 8 𝑟𝑟 2

75
 Proceedings of NCAMMM - 2018

Interphase potential (Eq. (1)) is considered as total strain energy of the assumed isotropic interphase material, and
Young's modulus of this interfacial region can be evaluated by using the following expression, as employed by Sears
and Batra(17):
𝜕𝜕 2 𝜙𝜙
𝐸𝐸𝐼𝐼 = (2)
𝜕𝜕 2 𝑆𝑆

where strain (S) at the interfacial distance r and found as:

𝑟𝑟−ℎ 0
𝑆𝑆 = (3)
ℎ0

ℎ0 acts as the equilibrium distance between CNT and matrix material and predicted by the minimization of the
interfacial potential (i.e., 𝜕𝜕𝜕𝜕⁄𝜕𝜕𝜕𝜕 = 0) as:
ℎ0 = 0.8584𝜎𝜎𝐼𝐼 (4)
The elastic modulus of the interfacial region is computed for the equilibrium position (i.e., 𝑟𝑟 = ℎ0 ), and the
corresponding expression, deduced from Eq. (2), is given as:
2 𝜎𝜎 11 𝜎𝜎 5 6 𝜎𝜎 10 𝜎𝜎 4
𝐸𝐸𝐼𝐼 = ℎ0 𝜋𝜋 2 𝜌𝜌𝑚𝑚 𝜌𝜌𝐶𝐶 𝜎𝜎𝐼𝐼 𝜀𝜀𝐼𝐼 �24𝑅𝑅 �� 𝐼𝐼 � − � 𝐼𝐼 � � + 9𝜎𝜎𝐼𝐼 � � 𝐼𝐼 � − � 𝐼𝐼 � �� (5)
3 𝑟𝑟 𝑟𝑟 5 𝑟𝑟 𝑟𝑟

It is important to remark that 𝐸𝐸𝐼𝐼 calculated from Eq. (5) will have a unit of force/length and therefore, required to be
ℎ0
divided by the average of the circumference of the inner and outer surface of CNT [i.e., by 2𝜋𝜋 �𝑅𝑅 + �].
2

3. Generation of Randomly/Aligned Dispersed CNT RVE

In order to model the RVE, RSA algorithm proposed by Liu et al.(16) is modified and RVEs consisting of randomly
dispersed/aligned CNTs are generated by Boolean based operations using FEM based software COMSOL Multi
Physics. Each CNT is described by origin O and two Euler angles θ and ϕ in the xyz co-ordinate system as shown in
Fig. 1. CNTs are randomly oriented in 3-dimension by creating a range of uniform pseudo-random integers between
0o to 1800. To maintain the periodicity of RVE, if any CNT cuts the surface of RVE, the copy of that CNT is
translated to opposite surface and then the part of that CNT lying outside the RVE is chopped to form the periodic
RVE. The obtained RVE of CNT nanocomposite having randomly dispersed and aligned CNTs are shown in Figs. 2
& 3 respectively.

Fig. 1 Coordinate system and Fig. 2 Randomly-dispersed CNT- Fig. 3 Aligned CNT-reinforce RVE
definition of θ and ϕ for CNT RVE.
4. Numerical example
In this work, effective stiffness properties of 1 %, volume fraction CNT reinforced PE nanocomposite are estimated
by the proper utilization of periodic boundary conditions applied on the RVE created through RSA. COMSOL

76
Effect of interphase and dispersion of CNTs on the elastic properties of CNT-Polyethylene nanocomposite

Multiphysics is utilized to perform the FEM analysis on the RVE, and all FEM based calculations, necessary to
predict the homogenized elastic properties of nanocomposite material (i.e., volume-average of the stresses and the
strains) are performed by using the special features of the software. Polyethylene (PE), (𝐸𝐸𝑚𝑚 = 3.4 GPa) and CNT
(50,50), (𝐸𝐸𝐶𝐶𝐶𝐶𝐶𝐶 = 1000 GPa) are taken as the matrix and reinforcement respectively. CNT(50,50) is considered as
equivalent cylinder model. The effect of perfect (i.e., without interphase between CNT and PE) and imperfect (i.e.,
with interphase between CNT and PE) bonding on the elastic constants of nanocomposite are studied.
The dimensions of CNT are given below:
Length, L CNT = 50 nm; Outer radius, R = 3.6 nm; Inner radius, r = 3.2 nm.

5. Verification
Stiffness properties of perfectly bonded CNT-PE nanocomposite obtained from FEM analysis are compared with the
predictions made by Tsai-Pagano model and plotted in Fig. 4. The elastic constants of the nanocomposite are found
to be 3.9082 GPa (taken as the average value of E x , E y, and E z ) and 3.7957 GPa using FEM and Tsai-Pagano
method respectively, thus FEM results are in good agreement with the results obtained from Tsai-Pagano method.

5
Ex
Elastic Properties (GPa)

4 Ey

3 Ez
E Tsai-Pagano
2
Gxy
1 Gxz
Gyz
0
Fig. 4. Stiffness properties of 1% CNT reinforced RVE

6. Results and Discussion

6.1. Elastic modulus and thickness of interphase
The elastic modulus and thickness of interphase between CNT and PE are predicted using Eqs. (4 & 5). The
necessary value of L-J parameters along with volume and area density of PE and CNT respectively, taken from Lu
et al.(18) are given as:
Bond energy, 𝜀𝜀𝐶𝐶−𝐶𝐶𝐻𝐻2 = 0.004656 eV; vdW radius, 𝜎𝜎𝐶𝐶−𝐶𝐶𝐻𝐻2 = 0.3825 nm.
Volume density of PE, 𝜌𝜌𝑚𝑚 = 3.1 ×1028 m-3;
For carbon atoms in CNT, area density of carbon atoms (𝜌𝜌𝐶𝐶 ) in CNT = 3.8177 × 1019 𝑚𝑚−2 .
The elastic modulus and thickness of the interfacial region are estimated as 8.9658 GPa and 0.3283 nm, respectively.
It is to be noted that the obtained Young's modulus of the interphase zone is found to be higher than that of pristine
PE. This observation is in good concurrence with like findings by Zhang et al.(19) about the higher strength of
interphase zone between cylindrical fullerene structure, CNT, and polymer than that of the only polymer matrix.
6.2. Elastic properties of CNT-PE nanocomposite

77
 Proceedings of NCAMMM - 2018

Elastic constants of CNT-PE nanocomposite for randomly dispersed and aligned CNTs are given in Table 1 & 2
respectively. It can be observed that randomly dispersed and aligned CNT nanocomposite show isotropic and
transversely-isotropic behavior, respectively. Resulting values of stiffness properties for the nanocomposite
considering the effect of the interphase zone are found to be higher than that of perfectly bonded CNT-PE
nanocomposite. Which can be explained by the fact that, the assumed hypothetical interphase is found to be stiffer
than PE matrix.
In case of randomly dispersed CNT-PE nanocomposite, perfect bonding assumption leads to approximately
14.95 % enhancement in the elastic modulus of PE. Whereas, this increment is nearly 15.98 % for the condition of
imperfect bonding. On the other hand, CNT alignment in matrix substantially enhances the elastic constant of the
nanocomposite in the direction of alignment. While considering the case of imperfect bonding, approximately 75.06
% enhancement in axial modulus of pristine PE is found but the value of transverse elastic modulus is limited to
only 10.15 %.
Table 1 Stiffness properties of randomly dispersed CNT-PE nanocomposite
Type of Stiffness Properties [GPa]
bonding 𝐸𝐸𝑥𝑥 𝐸𝐸𝑦𝑦 𝐸𝐸𝑧𝑧 𝐺𝐺𝑥𝑥𝑥𝑥 𝐺𝐺𝑥𝑥𝑥𝑥 𝐺𝐺𝑦𝑦𝑦𝑦 𝜈𝜈𝑥𝑥𝑥𝑥 𝜈𝜈𝑥𝑥𝑥𝑥 𝜈𝜈𝑦𝑦𝑦𝑦
Perfect 3.9115 3.7903 4.0229 1.4899 1.5271 1.5273 0.3017 0.2956 0.3073
Imperfect 3.9390 3.8285 4.0620 1.5028 1.5389 1.5429 0.3012 0.2843 0.3068

Table 2 Stiffness properties of the aligned CNT-PE nanocomposite

Type of Stiffness Properties [GPa]
bonding 𝐸𝐸𝑥𝑥 𝐸𝐸𝑦𝑦 𝐸𝐸𝑧𝑧 𝐺𝐺𝑥𝑥𝑥𝑥 𝐺𝐺𝑥𝑥𝑥𝑥 𝐺𝐺𝑦𝑦𝑦𝑦 𝜈𝜈𝑥𝑥𝑥𝑥 𝜈𝜈𝑥𝑥𝑥𝑥 𝜈𝜈𝑦𝑦𝑦𝑦
Perfect 3.7120 3.7101 5.9001 1.3665 1.4877 1.4398 0.3603 0.2956 0.2955
Imperfect 3.7476 3.7426 5.9521 1.3920 1.5080 1.5119 0.3596 0.2957 0.2953

7. Conclusion
VdW interaction based cohesive zone model (CZM) is employed in the present manuscript to predict Young's
constants of interphase. Thereafter, in order to mimic the effect of dispersion of CNTs in the matrix, Boolean based
RSA technique is utilized to model the RVE. The stiffness constants of the nanocomposite material are predicted
from the RVE and based on the study following conclusions are found:
• The obtained Young's modulus of the hypothetical interfacial region is obtained higher than that of pristine
PE.
• Considering the impression of the interphase zone, the values of elastic properties of the nanocomposite
material are found to be higher than that of perfectly bonded CNT-PE nanocomposite.
• Alignment of CNTs in matrix leads to substantial enhancement in the axial modulus of the nanocomposite
material.
References:
[1] Iijima S. Helical Microtubules of graphitic carbon. Nature. 1991;354:56–8.
[2] Fiedler B, Gojny FH, Wichmann MHG, Nolte MCM, Schulte K. Fundamental aspects of nano-reinforced
composites. Compos Sci Technol. 2006;66(16):3115–25.
[3] Shokrieh MM, Rafiee R. A review of the mechanical properties of isolated carbon nanotubes and carbon
nanotube composites. Mech Compos Mater. 2010;46(2):155–72.

78
 Effect of interphase and dispersion of CNTs on the elastic properties of CNT-Polyethylene nanocomposite

[4] Bakshi SR, Agarwal A. An analysis of the factors affecting strengthening in carbon nanotube reinforced
aluminum composites. Carbon. Elsevier Ltd; 2010;49(2):533–44
[5] Joshi UA, Sharma SC, Harsha SP. Analysis of elastic properties of carbon nanotube reinforced
nanocomposites with pinhole defects. Comput Mater Sci.; 2011 Oct;50(11):3245–56.
[6] Shokrieh MM, Rafiee R. On the tensile behavior of an embedded carbon nanotube in polymer matrix with
non-bonded interphase region. Compos Struct; 2010 Feb;92:647–52.
[7] Joshi P, Upadhyay SH. Effect of interphase on elastic behavior of multiwalled carbon nanotube reinforced
composite. Comput Mater Sci; 2014 May;87:267–73.
[8] Odegard GM, Clancy TC, Gates TS. Modeling of the mechanical properties of nanoparticle/polymer
composites. Polymer. 2005;46:553–62.
[9] Jiang LY, Huang Y, Jiang H, Ravichandran G, Gao H, Hwang KC, et al. A cohesive law for carbon
nanotube/polymer interfaces based on the van der Waals force. J Mech Phys Solids. 2006 Nov;54(11):2436–
52.
[10] Tan H, Jiang LY, Huang Y, Liu B, Hwang KC. The effect of van der Waals-based interface cohesive law on
carbon nanotube-reinforced composite materials. Compos Sci Technol. 2007;67:2941–6.
[11] Shi D-L, Feng X-Q, Huang YY, Hwang K-C, Gao H. The Effect of nanotube waviness and agglomeration on
the elastic property of carbon nanotube-reinforced composites. J Eng Mater Technol. 2004;126(3):250–7.
[12] Chanteli A, Tserpes KI. Finite element modeling of carbon nanotube agglomerates in polymers. Compos
Struct; 2015;132:1141–8.
[13] Alian AR, Kundalwal SI, Meguid SA. Interfacial and mechanical properties of epoxy nanocomposites using
different multiscale modeling schemes. Compos Struct; 2015;131:545–55.
[14] Joshi P, Upadhyay SH. Analysis of alignment effect on carbon nanotube layer in nanocomposites. Phys E
Low-dimensional Syst Nanostructures; 2015;66:221–7.
[15] Feder J. Random sequential adsorption. J Theor Biol. 1980;237–54.
[16] Liu H, Zeng D, Li Y, Jiang L. Development of RVE-embedded solid elements model for predicting effective
elastic constants of discontinuous fiber reinforced composites. Mech Mater; 2016;93:109–23.
[17] Sears A, Batra RC. Macroscopic properties of carbon nanotubes from molecular-mechanics simulations. Phys
Rev B. 2004 Jun;69:235406.
[18] Lu WB, Wu J, Song J, Hwang KC, Jiang LY, Huang Y. A cohesive law for interfaces between multi-wall
carbon nanotubes and polymers due to the van der Waals interactions. Comput Methods Appl Mech Eng.
2008;197:3261–7.
[19] Zhang Y, Zhuang X, Muthu J, Mabrouki T, Fontaine M, Gong Y, et al. Load transfer of graphene / carbon
nanotube / polyethylene hybrid nanocomposite by molecular dynamics simulation. Compos Part B;

79
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Tribological Behaviour of Coated Mild Steel with Nano Al2O3-Ni-P Composite Material

Sushma Bharti, Santosh Kumar, Atul Kumar Harmukh, Subrata Kumar Ghosh
Department of Mechanical Engineering, Indian Institute of Technology (ISM), Dhanbad
Email: subratarec@yahoo.co.in

Abstract: The present work is based on surface improvement of mild steel to increase quality and durability of the
product. Surface improvement of the mild steel sample has been done by using electroless coating of nano Al2O3-Ni-
P. Friction and wear characteristics of the mild steel sample was evaluated by using a Pin-on-disc tribometer.
Elemental composition of the mild steel with and without coating was obtained from Energy dispersive X-ray
analysis (EDX) test. Using electroless coating of Al2O3 nanoparticles combined with Ni-P base metal improves the
wear and friction properties and surface hardness by providing a uniform coating over the mild steel sample.
Keywords: Electroless coating; Al2O3-Ni-P; Friction; Wear

1. Introduction
Coatings techniques are used to prolong the life of a component by improving its mechanical properties such as
mechanical strength, increasing wear and corrosion resistance for industrial applications. Various coating methods
are available among which electroless coating widely being used in industries due to its wide ranges of advantages
that includes its high rate of deposition and it also provide uniform coating on intricate/complex geometries [1].Ni-P
is a typical example of electroless plating due to its broad usage in machinery, electronics and automobiles, valves
and aerospace industries due to its uniform coating and excellent properties such as high hardness, good wearand
corrosion resistance, low coefficient of friction, high reflectivity etc [2-3].Incorporation of hard particles such as
SiC, ZrO2, TiO2, graphene, CNT, Carbon nanofibers (CNF’s), diamond, and ZnO further enhanced the tribological
properties of the Ni-P composite coating [4-11]. Promphet et al. [12] investigated on graphene oxide into Ni-P-TiO2
solution and the results obtained exhibits a significant increase in corrosion resistance and electrical conductivity of
the steel.
Karthikeyan et al. [13]concluded that deposition rate and surface roughness are highly influenced by
varying the concentration of reducing agent (sodium hypophosphite). On increasing the concentration of reducing
agent Ni forms amorphous phase thus micro hardness of the coating gets reduced. Heat treatment of the composite
coating was carried out which results in the formation of intermetalic nickel phosphide (Ni3P) which improves the
micro hardness and surface roughness of the composite coating. Popoola et al. [14]investigated the sliding wear
behaviour of Ni-Sn-P composite coating. The results show a considerable increase in wear and corrosion resistance
of the composite coating due to change in microstructure after the addition of Sn to the Ni-P coating on base metal.
Present work deals with the improvement in the tribological properties and morphology of mild steel sample by
using electroless nano Al2O3-Ni-P complex coating because nano alumina particles gave the most promising results.

2. Methodology
Electroless coating is an autocatalytic deposition process which involves reduction of metallic ion salt and oxidation
of chemicals present in the bath solution. The bath solution for electroless coating mainly consist of metallic salt,
Tribological behaviour of coated mild steel with nano Al2O3-Ni-P composite material
reducing agent that donates electrons to other chemical species in a redox reaction, stabilizing agent that prevent
bath from decomposition and provide stability and complexing agent which is used to exert a buffering action and
also prevent precipitation of basic nickel salt. The deposition process mainly based on the redox reaction of bath
compositions. It provides uniform coating on the substrate, without building up at the edges and corners which
increases its application range.

(a) (b)
Figure 1: Experimental setup for electroless coating (a) Base solution for Ni-P coating and (b) Ultrasonicated nano
alumina powder was mixed in the base solution

In the present study, mild steel pin having dia. 6 mm and 20 mm thickness are used as a substrate material, while
disc is made of D2 steel. Experimental setup forelectroless coating is shown in Figure 1. To perform the electroless
coating, initially the coating bath solution was prepared by adding all the chemicals in aqueous solution with
requisite amount. The composition of bath constituent is given in Table 1. These chemicals provide coating of Ni-P
on the mild steel sample. After the coating of Ni-P onbase metal, ultrasonicated nano alumina particles were added
to the bath solution and an electroless Al2O3-Ni-P composite coating was obtained.
Table 1: Bath composition concentration and operating conditions
Coating bath composition
Chemicals Concentration(gL-1)
Nickel sulphate 40
Sodium hypophosphite 20
Tri-sodium citrate 25
Ammonium chloride 50
Nano alumina 6
Operating condition
pH 4.5
Temperature 80-90°C
For homogeneous mixing of chemical ingradients in the bath solution, magnetic stirrer has been used. Heating
source is provided by using heating plate. Temperature of the bath solution is controlled by using thermometer
dipped in bath solution. Thickness of coating is monitored by the total time of plating process i.e. 1 hr.

3. Results and Discussion

The elemental analysis of mild steel substrate without and with coating was obtained by Energy Dispersive X-ray
(EDX) analysis andthe results are given in Figure 1 and Figure 2 respectively, whereasthe composition of elements
of mild steel pin without and with coating has also been shown in Table 1 and Table 2.The results obtained by this
clearly shows that the mild steel substrate without coating mainly contains “Fe” in its elemental form which imparts

81
 Proceedings of NCAMMM - 2018

good tensile strength, toughness, ductility, malleability but it has poor corrosion resistance. Thus its corrosion
resistance will be improved by using electroless coating method using Ni-P base metal. Phosphorous content imparts
good corrosion resistance on the composite coating.

Figure 2. Elemental analysis of mild steel pin without coating by EDX result

Figure 3: Elemental analysis result of mild steel pin with coating

Table 2: Elemental composition of mild steel pin without coating

Sl. No. Element Wt.%

1 C 6.72
2 O 4.23
3 Si 0.23
4 Mn 1.18
5 Fe 87.50
In the present investigation, surface roughness of the mild steel samples with and without coating of Ni-P/nano
alumina were measured by using Mitutoyo Surface Roughness tester. It has been observed that the surface
roughness value decreases with composite coating. The average surface roughness value (Ra) for composite coating
is 1.626±0.1 µm.Micro-hardness of the mild steel sampleswere measured by using a Micro Vicker hardness Tester
(Economet VH-1 MD) under the load of 50 N. Average value has been considered to minimize the error of micro-
hardnessfor both with and without coating,these values are given as 330 HV and 226 HV respectively.Heating of
electroless nickel coating leads to precipitation of phosphides, which acts as a barrier for dislocation movement, thus
the hardness of the coating increased.

82
Tribological behaviour of coated mild steel with nano Al2O3-Ni-P composite material
Table 3: Elemental composition of mild steel pin with coating

Sl. No. Element Wt.%

1 C 15.03
2 O 34.23
3 Na 3.91
4 Mg 0.21
5 Al 2.53
6 Si 0.30
7 P 9.45
8 S 0.62
9 Cl 2.12
10 K 0.31
11 Ca 0.29
12 Fe 1.94
13 Ni 29.07

Tribological properties such as friction and wear of mild steel substrate was measured by usingPin-on-
disctribometer. Results shows that the wear rate and friction force both decreases in case of mild steel pin with
Al2O3-Ni-P composite coating.

Figures 4 and 5 shows the wear and friction force for different types of samples i.e. with coating and
without coating. The conditions for the experiment were fixed at 4 kg and 5 kg load and 300 rpm. The friction and
wear both are less for coated materials. It has also been observed that both wear and friction force increases with
time but after a certain time both the factors decreases for the coated samples. The reason may be due to the
development of grain growth at the high temperature interface.

600

500
wear w/o coating at 5 kg
400
wear (µm)

load
300 wear w/o coating at 4 kg
load
200
wear with coating at 5 kg
100 load
0 wear with coating at 4 kg
1 2 3 4 5 6 7 8 9 10 load

time(min)

Figure 4: Variation of Wear with Time for coated and without coated samples

83
 Proceedings of NCAMMM - 2018

35
30

Friction Force (N)

25 FF w/o coating at 5 kg load
20
FF w/o coating at 4 kg load
15
10
FF with coatng at 5 kg load
5
0
FF with coating at 4 kg
1 2 3 4 5 6 7 8 9 10 load
time(min)

Figure 5: Variation of Friction Force with Timefor coated and without coated samples

4. Conclusions
This investigation is carried out to study the elemental composition, surface roughness, micro-hardness, wear and
friction properties of the electroless Ni-P/nano alumina coated and without coated mild steel samples. Heating of
electroless nickel coating leads to precipitation of phosphides, which acts as a barrier for dislocation movement, thus
the hardness of the coating increased. Electroless Ni-P/nano alumina composite coating results in decrease in wear
and friction force values due to the development of grain growth at the high temperature interface.
References
[1] Wang R., Ye W., Ma C., Wang C., Preparation and characterization of nanodiamond coresmcoated with a
thin Ni-Zn-P alloy film, Materials Characterization, 59 (2008), 108–111.
[2] HoorF. Shafia, Aravinda C.L., Ahmed H.F., Mayanna S.M., Electroless deposition and characterization of
Fe-W-Pt alloys, J. Mater. Sci. Lett. 19 (2000) 1067-1069.
[3] Hamdy Abdel Salam, Shoeib M.A., Hady H., Salam O.F. Abdel, Electroless deposition of ternary Ni-P alloy
coatings containing tungsten or nanoscattered alumina composite on steel, J. Appl. Electrochem. 38 (2008)
385-394.
[4] Chen C.K., Feng H.M., Lin H.C., Hon M.H., The effect of heat treatment on the microstructure of
electroless Ni–P coatings containing SiC particles, Thin Solid Films 416 (2002) 31–37.
[5] Sharma S.B., Agarwala R.C., Agarwala V., Ray S., Dry sliding wear and friction behaviour of Ni–P–ZrO2–
Al2O3 composite electroless coatings on aluminum, Mater. Manuf. Process. 17 (2002) 637–649.
[6] Chen W., Gao W., He Y., A novel electroless plating of Ni–P–TiO2 nano composite coatings, Surf. Coat.
Technol. 204 (2010) 2493–2498.
[7] Hu Q.-h., Wang X.-t., Chen H., Wang Z.-F., Synthesis of Ni/graphene sheets by an electroless Ni plating
method, New Carbon Mate. 27(2012) 35-41.
[8] Li W., Jin H., Hao Y., Chen T., Dai J., Wang Q., The microstructure of nickel layer on single-walled carbon
nanotubes prepared by an electroless coating process, J. Nanomater. 2011(2011) 5.
[9] Tsai T.-K., Chuang C.-C., Chao C.-G., Liu W.-L., Growth and field emission of carbon nanofibers on
electroless Ni-P alloy catalyst, Diam. Relat. Mater. 12(2013)1453-1459.
[10] H. Ashassi-Sorkhabi, Es’hagi M., Corrosion resistance enhancement of electroless Ni-P coating by
incorporation of ultrasonically dispersed diamond nanoparticle Corros. Sci. 77(2013) 185-193.

84
Tribological behaviour of coated mild steel with nano Al2O3-Ni-P composite material
[11] Shibli S.M.A., Jabeera B. Anupama R.I., Incorporation of nano zinc oxide for improvement of electroless Ni-
P plating, Appl. Surf. Sci. 253(2006) 1644-1648.
[12] Promphet Nadtinan, Rattanawaleedirojn, Rodthongkum Nadnudda, Electroless Ni-P-TiO2 Sol-RGO: A smart
coating for enhanced corrosion resistance and conductivity of steel, Surface and coating Technology
325(2017) 604-610.
[13] Karthikeyan S., Ramamoorthy B., Effect of Reducing agent and nano Al2O3 particles on the properties of
electroless Ni-P coating, Applied Surface Science 307(2014) 654-660.
[14] Popoola A.P.I., Loto C.A.,Osifuye C.O., Aigbodion V.S., Popoola O.M., Corrosion and wear properties of
Ni-Sn-P ternary deposits on mild steel via electroless method, Alexandria Engineering Journal 55 (2016)
2901-2908.

85
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Influence of the Delamination Geometry on the Shear Behaviour of Glass/Epoxy

Composites

Subhankar Roy1, Tanmoy Bose1, Kishore Debnath1

1
Department of Mechanical Engineering, National Institute of Technology Meghalaya, Meghalaya, India

Abstract: In recent times, composite materials have become important engineering materials due to its
multifunctional properties like high strength to weight ratio, high stiffness, and low thermal expansion. Polymer
composites provide flexibility to the designer in achieving the property requirement by using different kind of fibers
in different kind of matrices. In the present study, glass fiber/epoxy composite have been manufactured to study their
shear behaviour using Iosipescu shear test fixture. The study includes manufacturing of four types of glass
fiber/epoxy specimens, one without delamination and the other three with different shapes of delamination such as
circular, square and rectangular delamination. The results show that the specimen having circular delamination
possess lowest shear strength, followed by rectangular and square.
Keywords: Glass fiber, Epoxy, Delamination, Iosipescu Shear Test

1. Introduction
A composite material is generally defined as a combination of two or more materials in different proportions to
achieve the required properties that are better than those possessed by the individual materials. Thus a composite
material possesses superior properties like high strength to weight ratio, high tensile strength, and low thermal
expansion. Polymer matrix composites (PMCs) have found application in various fields due to their adequate
strength, light-weight, toughness, and low-cost compared to conventional metals. PMCs can be categorized
according to different reinforcing material such as glass fibre reinforced plastics (GFRP), carbon fibre reinforced
plastics (CFRP), and Kevlar fibre reinforced plastics. Glass fibres possess very good properties like high strength,
stiffness, flexibility, and resistance to chemical reactions as mentioned by Sathishkumar et al. (1). According to
Chavan and Gaikwad(2) and Vinay et al. (3), glass fibre reinforced polymers are extensively used in automotive
industries, construction industries, sports industries, marine industries, home appliances, electrical industries etc.
Another new category of composite materials that is finding extensive applications in the field of aerospace
structures is the fibre metal laminates (FML) namely glass reinforced aluminium (GLARE). GLARE is generally a
stack of aluminium sheets which are bonded with unidirectional glass fibres reinforced epoxy prepregs as discussed
by Guocai and Yang (4). Bhaskar and Srinivas (5) studied the mechanical behaviour of glass fibre composites where
glass fibre is in the form of woven mat and chopped strand mat under different loading conditions. The study
showed that mechanical properties of the composites based on chopped strand mat glass fibre are better than the
woven mat glass fibre. The chopped strand mat glass fibre composite was also studied by Mathapati and Mathapati
(6) by varying the glass fibre content and performing the mechanical tests which illustrates the advantage of using
higher proportion of glass fibre. The experimental results of delamination fracture toughness were studied for
glass/epoxy composites developed by compression resin transfer moulding and conventional resin transfer moulding
process (7). The results indicated that the fracture behaviour is significantly influenced by the fabric structure. The
Influence of the delamination geometry on the shear behaviour of glass/epoxy composites

influence of delamination during drilling of a glass/epoxy composite was studied and a new delamination factor was
proposed by Nagarajan et al. (8). Yang et al. (9) studied the bending, compression, and shear behaviour of stitched
woven glass fibre reinforced epoxy composite fabricated by resin transfer moulding. The shear test was done by
short beam and grooved coupon test approaches for different patterns of stitching. It was concluded that the better
resistance from delamination for increasing density of fibre stitching was in z-direction. The effect on shear
behaviour due to fibre orientation was studied by Almeida Jr. et al. (10). Four different orientations of fibre mats
along with four different types of test methods were considered in the study. The 00 fibre orientation shows better
shear strength as compared to the 900 fibre orientation. The double-notched and v-notched test methods were found
to be more useful for determining the shear strength of the material. Godara et al. (11) investigated the influence of
adding carbon nanotubes in glass/epoxy composites by performing the single fibre push-out test. The results showed
better interfacial shear strength when carbon nanotubes were introduced as an additional reinforcement in
glass/epoxy composites. From the literature, it is evident that the shear loading behaviour for different shapes of
delamination in the context of glass/epoxy composite has not been investigated. In this paper, the shear loading
behaviour of glass/epoxy composite has been carried out using Iosipescu test for circular, square, and rectangular
delamination and then compared with a delamination free specimen.

2. Experimentation
2.1. Selection of materials
E-glass fibre was used as reinforcing material in the form of woven mat. The epoxy resin (Araldite AW106) used for
manufacturing of glass/epoxy composite has creamy viscous colour. The range of curing temperature for chosen
resin was around 200-1500 and that of minimum curing time is 15 hours-5 minutes, respectively.The ratio of
proportion in which epoxy AW106 and hardener HV953 mixed was1:1. The mixture of AW106/HV953 does not
release any volatile constituents and gives good resistance to static and dynamic loads.

2.2. Fabrication of glass/epoxy composites

The following steps have been carried out for fabricating the glass/epoxy composites:
Type A: Composite specimen without delamination - A cast iron mould having flat surface of 200 mm × 200 mm has
been made for manufacturing the composite. The composite was fabricated on the lower open mould and then
covered by the upper mould for applying pressure. A transparent sheet was attached to the flat surface of the mould
to prevent any contact of the resin mixture with the flat surface. E-glass fibre woven mats were cut in size as per
requirement. The Araldite AW106 resin and Araldite HV953 hardener were mixed in 1:1 ratio and stirred until the
mixture becomes uniform. The resin mixture was then applied on the transparent sheet attached to the mould and
spread evenly using a brush. The cut fibre mat was then placed over the resin layer and a uniform pressure was
applied over the entire area using a roller. This helps in proper adhesion of the glass fibre mat and resin without
leaving any air bubbles or voids throughout its thickness. The resin mixture was applied again over the glass fibre
mat and the process was repeated till the required thickness of the composite is achieved. Another transparent sheet
was used to cover the top layer before covering with the upper mould that applies continuous pressure. The mould
was left for curing under room temperature for about 24 hours. After curing, the upper mould was removed and the

87
 Proceedings of NCAMMM - 2018

final glass/epoxy composite specimen was taken out by peeling off the transparent sheets attached to both side of the
specimen. A glass/epoxy composite specimen having 16 layers of glass fibre woven mat and total thickness of 3.6
mm was manufactured by this process.
Type B: Composite specimens with delamination - In addition to the steps mentioned for type A, for creating a
delamination, a Teflon tape cut in the required shape and dimension was introduced at the middle i.e., after the eight
layer of glass fibre mat. The Teflon tape does not allow adhesion between the eighth and ninth layer of glass fibre,
thus creating a delamination in the final specimen. The fabrication of the four types of glass/epoxy composite
specimen was followed by Iosipescu shear testing of the specimens and analysis of the results. The above steps are
represented in the form of flow chart in Fig. 1. The cured glass/epoxy composite was taken out of the mould and is
cut in the form of a double-edged notched specimen as recommended by Odegard and Kumosa (12) for performing
the Iosipescu test. The dimensions of the specimen according to ASTM D 5379-93 standard are: length = 78mm,
width = 20mm, thickness = 3.6 mm, notch angle =900 and notch depth = 4.4mm. All the four types of specimen
ready for undergoing shear test are shown in Fig. 2.

(a) (b) (c) (d)

Fig. 1 Flowchart of the manufacturing process of glass Fig. 2 Glass/epoxy composite specimen (a) without
fibre reinforced composites delamination, and with (b) circular, (c) square, and (d)
rectangular delamination
2.3. Experimental Setup
The experimental setup that was used for carrying out the Iosipescu shear test comprises of an ultimate testing
machine (shown in Fig. 3) along with the Wyoming Iosipescu shear test attachment (shown in Fig. 4) fabricated in-
house. Iosipescu shear test has been developed by Iosipescu (13) in 1967. The test consists of a double-edged
notched specimen subjected to two opposing force couples. This is considered as a type of four point load test for
determining the interlaminar shear strength of the specimen. By considering a notch angle of 900 and notch depth of
22% of the width, a constant shear stress can be obtained. In case of fibre reinforced composite, there is a stress
concentration near the notch tip which is proportional to the fibre volume fraction and the fibre orientation.

88
Influence of the delamination geometry on the shear behaviour of glass/epoxy composites

Fig. 3 UTM setup Fig. 4 Iosipescu shear test fixture

3. Results and Discussions

The properties of glass/epoxy composites depend on its composition, orientation of the fibres, and the number of
fibre layers used. In the present work, 16 layers of woven glass fibre mats were used to bond with the epoxy resin.
The properties of the composites also depend on the bonding strength between the glass fibre and epoxy resin. Thus,
the shear properties of glass/epoxy composites (with and without delamination) were evaluated using Iosipescu
shear test fixture. Table 1 shows the maximum load and the maximum shear stress determined for two sets of
specimen. The fractured specimens after performing the Iosipescu shear test are shown in Fig. 5.
Table 1 Maximum load and maximum shear stress obtained from Iosipescu shear test
Specimen type Maximum load (kN) Maximum shear stress (MPa)
Specimen Specimen Average Specimen Specimen Average
Set 1 Set 2 Value Set 1 Set 2 Value
Without delamination 14.25 15.25 14.75 262.84 281.28 272.06
Circular delamination 13.95 13.00 13.48 257.30 239.78 248.54
Square delamination 13.25 14.50 13.88 244.39 267.45 255.92
Rectangular delamination 12.50 14.50 13.50 230.56 267.45 249.01

Fig. 5 Fractured specimens after performing Iosipescu shear test

89
 Proceedings of NCAMMM - 2018

The shear test on the glass/epoxy composite specimen with double edge notch shows that a maximum load of
14.75kN is obtained for a displacement of around 12 mm for the specimen without delamination (Fig. 6a). The
maximum shear stress for the glass/epoxy specimen without delamination is around 272.06MPa (Fig. 6b). The
specimen breaks drastically after the maximum shear stress is reached. The results obtained for the glass/epoxy
composite with different shapes of delamination namely circular, square, and rectangular are illustrated in Fig. 6.
From the figure, it can be observed that the maximum load (13.48 kN) that can be sustained is much lower in case of
circular delamination followed by rectangular (13.50 kN) and square (13.88 kN) delamination, respectively.
Similarly, the stress versus strain plot shows a same trend where the maximum shear stress for circular delamination
(248.54 MPa) was found to be lower than that of the square (255.92 MPa) and rectangular delamination (249.01
MPa). Thus, it can be concluded that the glass/epoxy composite with circular delamination is more prone to damage
and crack growth as compared to square and rectangular delamination under shear loading. Among the different
shapes of delamination, the sequence of maximum shear strength in descending order was found to be: square,
rectangular, and circular delamination.

Fig. 6 (a) Load versus displacement curve and (b) Stress versus strain curve

4. Conclusions
The glass/epoxy composites were manufactured using the hand layup process. A circular, square, and rectangular
delamination was created in the composite specimen by introducing Teflon tape in between the glass fibre layers.
The Iosipescu shear tests were successfully carried out for all type of specimen viz. with delamination and without
delamination in order to study their behaviour under shear loading. The load versus displacement graph was studied
and the ultimate shear stress was measured for each test specimen. It was observed that the composite specimen
having no delamination has the highest shear strength as compared to the specimens having delamination. Among
the different shapes of delamination, the specimen having circular delamination has the lowest shear strength. It was
also observed that the circular delamination has the most influence on the crack growth and failure of glass/epoxy
composite under shear loading.

90
Influence of the delamination geometry on the shear behaviour of glass/epoxy composites

References
[1] Sathishkumar TP, Satheeshkumar S, Naveen J. Glass fiber-reinforced polymer composites – a review. Journal
of Reinforced Plastics and Composites. 2014;33(13):1258-1275.
[2] Chavan VB, Gaikwad MU. Review on development of glass fiber/epoxy composite material and its
characterizations. International Journal of Science, Engineering and Technology Research. 2016;5(6):2224-
2228.
[3] Vinay HB, Govindaraju HK, Banakar P. Processing and characterization of glass fiber and carbon fiber
reinforced vinyl ester based composites. International Journal of Research in Engineering and Technology.
2015;4(5):401-406.
[4] Guocai W, Yang JM. The mechanical behavior of GLARE laminates for aircraft structures. Journal of the
Minerals, Metals, and Materials Society. 2005;57(1):72-79.
[5] Bhaskar V, Srinivas K. Mechanical characterization of glass fiber (woven roving/chopped strand mat E-glass
fiber) reinforced polyester composites. International Conference on Functional Materials, Characterization,
Solid State Physics, Power, Thermal and Combustion Energy. 2017;1859,020108:1-5.
[6] Mathapati SS, Mathapati SS. Testing and analysis of mechanical properties of E-Glass fiber reinforced epoxy
polymer composites. International Journal of Research and Innovations in Science and Technology.
2015;2(1):46-52
[7] Treber D, Haspel B, Elsner P, Weidenmann KA. Delamination fracture toughness of continuous glass-
fiber/epoxy composites for structural applications. International Journal of Plastics Technology. 2017;21(1):39-
54
[8] Nagarajan VA,Sundaram S,Thyagarajan K,Rajadurai JS, Rajan TPD. Measuring delamination severity of glass
fiber-reinforced epoxy composites during drilling process. Experimental Techniques. 2013;37(2):66-73.
[9] Yang B, Kozey V, Adanur S, Kumar S. Bending, compression, and shear behavior of woven glass fiber-epoxy
composites. Composites Part B: Engineering. 2000;31:715-721.
[10] Almeida Jr. JHS, Angrizani CC, Botelho EC, Amico SC. Effect of fiber orientation on the shear behavior of
glass fiber/epoxy composites. Materials and Design. 2015;65:789-795.
[11] Godara A, Gorbatikh L, Kalinka G, Warrier A, Rochez O, Mezzo L,et al. Interfacial shear strength of a glass
fiber/epoxy bonding in composites modified with carbon nanotubes. Composites Science and Technology.
2010;70:1346-1352.
[12] Odegard G, Kumosa M. Determination of shear strength of unidirectional composite materials with the
Iosipescu and 100off-axis shear tests. Composite Science and Technology. 2000;60:2917-2943.
[13] Iosipescu N. New accurate procedure for single shear testing of metals. Journal ofMaterials. 1967;2(3):537-
566.

91
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Static and Dynamic Mechanical Properties of Glass/Carbon Fiber Reinforced Epoxy

Composite

Nitai Chandra Adaka,b, Suman Chhetria,b, Bipin Kumar Singh a, Saikat Bolar a, Naresh Chandra Murmua,b,
Pranab Samanta*a,b, Tapas Kuila *a,b
a
Surface Engineering and Tribology Division, CSIR-Central Mechanical Engineering Research Institute,
Durgapur; bAcademy of Scientific and Innovative Research (AcSIR), CSIR-CMERI, Campus, Durgapur
Email: ps.iitb@gmail.com and tkuila@gmail.com

Abstract: Fiber reinforced polymer composite materials have proven their great potential in high strength structural
application specially in aerospace and automotive industry due to their high specific strength to weight ratio. In this
regard, carbon fiber (CF) and glass fiber (GF) are most prominent candidate among the different types of synthetic
fibers. But, one of these fiber reinforcement can't fulfill the required properties offered by the designer. Only hybrid
composites may offer the required properties by the selection of fibers and matrixes. In the present study, the CF/GF
reinforced epoxy hybrid composites have been prepared by the vacuum assisted resin transfer molding techniques
and the static and dynamic mechanical properties of the prepared composite have been investigated as per ASTM
standards. The mechanical properties of the hybrid composites were improved as the CF reinforcement content
increased in the matrix material.
Keywords: Fiber; Epoxy; Vacuum infusion; Mechanical properties.

1. Introduction
The demand of fiber reinforced polymeric composite (FRPC) materials in aerospace and transportation industry is
now widespread and gradually increasing day by day. The prime motive of using polymeric composites is to reduce
the weight of the automotive parts as well as increase the fuel efficiency to protect the environment from pollution
[1]. FRPC also offers high specific strength, stiffness, good corrosion resistance, dimensional stability and
conformability compared to monolithic conventional structural materials. Recently, the use of FRPC in aerospace
has increased from 3 to 20% in Airbus and over 50% in Boeing's [2]. In this domain, CF and GF have attracted great
attention for high performance structural applications due to its high specific stiffness among the all synthetic fibers.
Among the different types of polymer, epoxy resin is widely used for production of FRCP due to its low setting
shrinkage, adhesion to fiber, considerable cohesion strength and good thermal properties [3]. Therefore, the
GF/epoxy composite materials are most desirable for structural applications as GF is very cheap compared to CF.
But, the mechanical properties of GF/epoxy composite is low compared to CF/epoxy composite. Therefore a
combination of GF with CF is required to get the assured mechanical properties and produce the developed
composite in an economical way. Manders et al. investigated the hybrid effect of GF/CF/epoxy composite and
reported that the failure strain enhanced of up to 50% as the carbon phase increased [4]. Yerramalli et al. also
studied the hybrid effect of GF/CF/epoxy composite with an overall fiber volume fraction of 30% and noticed
splitting and kinking failures under static and dynamic loading rates [5]. Zhang et al. made GF/CF/epoxy composite
by ‘wet lay-up’ method and reported that it was not a good practice to obtain a high quality hybrid laminate [6].
Static and dynamic mechanical properties of glass/carbon fiber reinforced epoxy composite
Although there has been considerable work devoted on the development of GF/epoxy composites by using CF as
a secondary reinforcement, only a few has focused on the static as well as dynamic mechanical properties of the
GF/CF/epoxy hybrid composite. As epoxy matrix is a viscoelastic material, therefore dynamic mechanical analysis
(DMA) is very crucible with static mechanical properties to predict the overall improvement of the prepared hybrid
composites. Therefore, in this present work, an attempt has been made to investigate the influence of CF on static
and dynamic mechanical properties of GF/epoxy composite.

2. Materials and Experimental Procedure

In this study, commercially available plain weave carbon and glass fabrics with an areal density of 0.2 kg/m2 was
procured from Flips India Engineering (Mumbai) for reinforcement. The low viscosity, liquid modified Bisphenol-A
epoxy resin (LAPOX*C-51) and low viscosity modified cycloaliphatic amine hardener (Lapox AH-428) was bought
from Atul Limited (Gujarat, India) to prepare the matrix system in this work. The mixing ratio of epoxy resin and
hardener was 100:45 by weight and the mixture was cured at room temperature (24 hr).
The composites were prepared by VARTM process. At first, the fabrics were cut into dimension of 150 mm ×
200 mm and placed on the molding tool which was previously coated with a release agent for easy removal of the
prepared composites. The stacking sequence of the lamina was [(0/90) 5 ]s for tensile and DMA test and [(0/90) 7 ]s
for bending test. The lamina thickness was 0.3 mm. A peel ply was placed on the top of the lamina to prevent the
difficult separation of the bagging film from the composites. A nylon mesh was placed on the top of the peel ply to
ensure the continuous resin transfer in the lateral direction. The inlet of the resin supply was fixed with a spiral pipe
around the fabrics. The degassed resin was infused into the fabrics due to the difference in pressure. The vacuum
pressure was 0.2 Pa. The resin was cured with the shape of the perform fabrics. The prepared composite panels were
post cured at 120 OC for 1 h after curing at room temperature for 24 h. Finally, the desired samples were cut from
the panels for mechanical testing. Figure 1 shows the hybrid composite fabrication in VARTM process.

Figure 1. Laminate fabrication and specimens preparation.

3. Characterization

93
 Proceedings of NCAMMM - 2018

The tensile properties of the prepared composite materials were determined according to ASTM D3039 by using a
universal testing machine (UTM) (Tinius Olsen, H50KS). The ultimate tensile strength σ ut = Pmax/A, where Pmax =
Maximum force before failure (N), A = Average cross-sectional area (mm2). The transverse rupture strength or
flexural strength of the composites was determined by using the same UTM according to ASTM D7264. Flexural
strength and flexural modulus was calculated by using the formulae σ f = 3PL/2bd2, where σ f = stress at the outer
surface at mid span (MPa), P = applied force (N), L = support span (mm), b = width of beam (mm), d = thickness of
beam (mm). The dynamic mechanical properties of the composite materials were examined by using a dynamic
mechanical analyzer (DMA 8000 Perkin Elmer) according to ASTM D7028. This test was carried out under flexural
loading mode (duel cantilever) at a frequency of 1 Hz and amplitude of 50 µm. The temperature was ramped from
28 to 200oC with a heating rate of 2oC min-1. At least four specimens were tested for each case. The specimen's
details of the mechanical tests is given in the Table 1.
Table 1: Specimen dimension of different mechanical test

Test Name Specimen's dimension (mm)

Tensile 130 x 12 x 2.5
Flexural 128 x 13 x 4
DMA 45 x 9 x 2.5

4. Results and Discussion

4.1 Static mechanical properties of the prepared composite
The tensile strength of the prepared GF and CF reinforced composites depend upon the strength and modulus of
individual fiber, strength and chemical stability of the epoxy resin and fiber to epoxy resin interaction. The tensile
stress vs. strain (%) curves of the pure GF/epoxy, GF/CF/epoxy and CF/epoxy composites is shown in Figure 2(a).

Figure 2. (a) Tensile and (b) Flexural properties of the different composites

Figure 2(a) emphasise that the tensile strength of the CF/epoxy composites are more than the GF/epoxy composite.
The enhancement of the ultimate tensile strength of the hybrid composites occurred compared to GF/epoxy
composites due to the reinforcement of CF. It was also found that the induced stress in the prepared composites
linearly increased with increase the deformation until failure of the test specimens. It proves the brittle nature of the
prepared composites. In the GF/CF/epoxy hybrid composites, 50% CF was reinforced with the GF. The 50%

94
Static and dynamic mechanical properties of glass/carbon fiber reinforced epoxy composite
reinforcement of CF with the GF increased about 74% ultimate tensile strength of the hybrid composite compared to
GF/epoxy composite. The ductility of the GF/CF/epoxy hybrid composites also enhanced up to 5% than GF/epoxy
composites. The flexural test was performed to predict the bending strength of prepared composites. The typical
flexural stress-strain behaviour of the composites is shown in Figure 2(b). Like tensile properties, the flexural
strength and flexural strain of the GF/CF/epoxy hybrid composites increased about 22% and 7% respectively than
pure GF/epoxy composite. The overall static mechanical properties of the prepared composites are given in Table 2.

Table 2: Static mechanical properties of the prepared composites

Composite Ultimate tensile strength Ultimate flexural strength
GF/epoxy 177± 25 509 ± 30
GF/CF/epoxy 308 ± 20 618 ± 27
CF/epoxy 350 ± 22 732 ± 32

4.2 Dynamic mechanical properties of the prepared composite

DMA determine the dynamic mechanical responses of the prepared composites as a function of temperature,
frequency or time. In DMA, an periodic force mostly sinusoidal force with small amplitude is applied to the test
specimen and the sinusoidal stress and strain curve is recorded as a function of time. The evaluated modulus from
this stress-strain curve is not exactly same as the Young’s modulus. In DMA, the complex modulus (E) is calculated
and the magnitude of E can be expressed as: Complex modulus (E) = E' + i E", where E'= storage modulus and E''=
loss modulus. The E' is real part of E and it is defined as the amount of maximum energy stored in the composite
during one cycle of oscillation. It also emphasise the temperature dependant stiffness and load-bearing capacity of
the composites. The imaginary part of E is termed as E'' and it represents the amount of energy dissipated by
materials in form of heat during sinusoidal deformation. The E'' expresses the viscous response of the composites.
The variation of E' and E" with the increase of temperature of the prepared composites is shown in Figure 3(a) and
3(b), respectively.

Figure 3. Variation of (a) Storage modulus and (b) Loss modulus of the different composites with temperature

The damping property of the prepared composites is determined by a damping factor (tan δ) which is the ratio of E'
and E". It signifies the degree of molecular mobility of the polymer composites. The lower value of tan δ represents

95
 Proceedings of NCAMMM - 2018

the elastic behaviour and higher value of tan δ exhibits a non-elastic behaviour of the composite. The variation of the
tan δ value with temperature of the prepared composites is shown in Figure 4.
The tan δ value of the CF/epoxy composites is lower than the other prepared composites i.e. CF/epoxy is
more elastic than GF/epoxy and GF/CF/epoxy composites. The tan δ is also associated with the glass transition
temperature (T g ) of the composites. The T g is the temperature range where state of prepared composites changes
from a glassy to rubbery. The damping capacity of the polymer composites below T g is low as the polymer chain
segments are in frozen state. The polymer composite also loses its mechanical properties above T g as the cross
linking between polymer molecules breaks gradually with rise of temperature. So, the cross-linking density and
segmental motion of the epoxy resin is totally dependent on T g . The higher T g value of epoxy resin shows the
higher thermal stability of the prepared composites. The T g values of the prepared GF/epoxy, GF/CF/epoxy and
CF/epoxy composites are 45, 47 and 57oC respectively which is determined from the peak of the tan δ values of the
composites.

Figure 4. Variation of tan δ with temperature of the prepared composites

4. Conclusion
The Composites are prepared by VARTM process which decreases the void contents in the prepared composites
than the hand-layup process. The failure occurred in the tensile specimens in the gauge length that proved that the
test worked out evidently. The 74% improvement of ultimate tensile strength of the GF/CF/epoxy hybrid composites
than the GF/epoxy composite prove that the hybrid composites can be used as a replacement of GF reinforced
polymer composites where there is a require of good mechanical properties. The enhancement of the E' of the hybrid
composites also increased the stiffness of the composite than GF/epoxy composite. The FRPC with mixture of GF
and CF will not only increase the strength and stability of the material but also reduce the cost of the material than
costly CF/epoxy composites.

Acknowledgement
Authors are thankful to the Director of CSIR-CMERI. Authors are also thankful to Council of Scientific and
Industrial Research, New Delhi, India for funding MEGA Institutional project (MLP 218112).
Reference
[1] Arao Y, Yumitori S, Suzuki H, Tanaka T, Tanaka K, Katayama T. Mechanical properties of injection-molded
carbon fiber/polypropylene composites hybridized with nanofillers. Composites Part A 2013;55:19–26.

96
Static and dynamic mechanical properties of glass/carbon fiber reinforced epoxy composite
[2] Potluri P, Perezciurezu D, Ramagulam R. Measurement of meso-scale shear deformations for modelling textile
composites. Composites Part A 2006;37(2):303-314.
[3] Ferreira JM, Pires JTB, Costa JD, Zhang ZY, Errajha OA, Richardson M. Fatigue Damage Analysis of
Aluminized Glass Fiber Composites, Materials Science and Engineering A, 2005;407: 1–6.
[4] Manders P W, Bader M G. The Strength of Hybrid Glass/Carbon Fibre Composites: Part 1 Failure Strain
Enhancement and Failure Mode. Journal of Materials Science 1981;16: 2233-2245.
[5] Yerramalli CS, Waas AM. Compressive Behavior of Hybrid Composites. Proceedings of 44th AIAA/
ASME/ASCE/AHS Structures, Structural Dynamics and Materials Conference 2003;1509.
[6] Zhang J, Chaisombat K, He S, Wang CH. Hybrid Composite Laminates Reinforced with Glass/Carbon Woven
Fabrics for Light Weight Load Bearing Structures. Materials and Design. 2012;36: 75-80.

97
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Effect of Nanostructured 8% partially yttria stabilized Zirconia( YSZ) Coating in

Oxidation Behaviour of Inconel Alloy

D Ghosh1*, S Das2, H Roy1

1
Scientist, NDT & Metallurgy Group, CMERI, Durgapur-713209, India
2
Senior Research Fellow, NDT & Metallurgy Group, CMERI, Durgapur-713209, India

Abstract: The nano structured 8% partially yttria stabilized Zirconia (YSZ) powder coating (plasma spray) on
Inconel (IN 718) substrate are made for oxidation studies at high temperature. The oxidation study was carried out
at 1273K for 24 hours in dry air oxidation environment. The reaction kinetics and rate of oxidation are studied and
characterization Of post corroded sample are also studied in SEM, FESEM and XRD. The micron size coating of
similar kind is also investigated in same environment to compare with nanostructured coating. The results clearly
indicate that nano structured coating offers better resistance of oxidation as compared to the micron size coated
specimen. The detail mechanism of corrosion mechanism of both nanostructured and micron size coating is
discussed in details in this paper.
Keywords: Yttria stabilized Zirconia (YSZ), Nano structured Coatings, Oxidation, Inconel

1. Introduction
The 8% YSZ is used as thermal barrier coatings in different industrial applications. It provides thermal insulation
effect to the high temperature zone of gas turbine ,jet engine and also space craft.. This coating can not be directly
applied over the substrate as the coating is not so adherent. The adhesion propery can be improved by applying
McCrAly bond coat on the substrate, before the use of final YSZ coating. This combination of two coatings provides
high temperature oxidation and corrosion resistance along with thermal insulation effect as reported in several
literatures [1–6]. Oxidation at high temperature is a major material degradation mechanism in different industrial
applications. Further oxidation may affect the component integrity and residual life for further applications. The
microstructure of the coating played an important major role for coating integrity, adhesion and other properties. .
So by controlling the microstructure of the coating, it is possible to improve the oxidation resistance for different
industrial applications. The different defects in the coating like porosities, cracking, nature of molten zone formed
during plasma spray coating can affect the high temperature oxidation properties to a significant extent. [7-10]. The
plasma sprayed coating of ceramic oxide employed for oxidation/ resistance has been reported by many researchers
[11-15]. Nano materials are significantly used for improving the different material properties in the past two
decades by improving extraordinary physical and chemical properties.Generally in namostructured coating , the
powder size employed for the coating is restricted to 100 nm. The significant improvement of mechanical and
thermal properties shown by nano structured coatings over the micron size coating is reported in some literatures
[16- 23].
The present study investigates the effect of nano structured YSZ coating over inconel super alloy ( IN
718) on high temperature oxidation behavior under dry air at 1000°C (1273K). The studies are also carried out in
Effect of nanostructured 8% partially yttria stabilized Zirconia (YSZ) coating in oxidation behaviour of Inconel alloy
micron size coating.The reaction kinetic behaviour, oxidation rate and post oxidized sample characterization are
investigated to find out the details mechanism for improving oxidation rate.

2. Materials and Methods

2.1 Substrate material
The material used for these experimental studies was inconel 718 superalloy substrate which are most popular
candidate material for gas turbine and aero engines. Specimens of 20mmX10mmX5mm were cut from a plate for
experimental studies. The specimens are cleaned and grit blasted with alumina powder (grit 60) at a pressure of 5
kg/cm2 in a pressure air blasting machine for improving the adhesion characteristics.
2.2 Preparation of nano YSZ powder
The initial material for preparation of the nano powder is the micron size 8 mol percent yttria stabilized
zirconia(YSZ) powder of 20+/-5µm size, supplied by Powder alloy corporation USA (PAC 2008P). The variation
of size is not much pronounced and considered to be more uniform. The same powder is subjected to planetary ball
milling (Model: Res PM 200, Retsch Ltd, Germany) to obtain the nano size powder. Wet milling is performed for 10
hours to obtain the nanosize less than 100nm. The size of the powder after ball milling operation finally , is then
measured in particle size analyser (Model: Zeta Nano Zs, Malvern, UK). The size variation of nano size powders is
found within 100 nm and is shown in Fig 1.

Statistics Graph (4 measurements)

15
Intensity (%)

10

0
1 10 100 1000 10000
Size (d.nm)

Fig 1: Paticle size analysis histogram of nano milled powder

2.3 Development of coatings
The YSZ powder of micrometer size is directly plasma sprayed to the substrate material by using atmospheric
plasma spray unit. (Model: SG-100, Paraxair USA). The equipment is coupled with 6 axis robotic arm for uniform
coating in all sides of the specimen over the substrate. The NiCrAlY powder is used as bond coat with the substrate
of 150 µm. This coat is given for better coating adhesion .The plasma coating set up for the experimental study is
shown in Fig 2. The nano size powders are first subjected to spray drying to make agglomerated powder for better
flowability during the plasma coating process. The agglomerated nano powder on the substrate shows the nano
features in the coating and often referred as nano structured coating. The same weight of micron size powder is used
for micron size coating for comparison purpose. The plasma spraying parameters for the coating are given in
Table1.

99
 Proceedings of NCAMMM - 2018

Fig 2: Experimental set up for plasma spray Fig 3: Experimental set up for oxidation studies
coating unit with Six axis robotic arm

Table 1: Spray Parameters of atmospheric plasma spray YSZ coating

Micronsize coating Nanostructured coating

Voltage- 42 V 47V
Current- 844A 860A
Arc gas flow rate- 36 litre/ minute 40 litre/ minute
Carrier Gas flow rate ( hydrogen) – 6 litres/ minute 6 litres/ minute
Stand off distance- 125 mm 125mm
Powder flow rate- 40 gms/ minute 34 gms/ minute

2.4 Oxidation test

The corrosion/oxidation test is carried out in a high temperature vertical tubular furnace. A digital weighing balance
is attached at the top of the furnace. Specimens for Oxidation tests are kept in the middle zone (central heating zone
of the furnace before heating the furnace to the desired temperature. Oxidation test is carried out in dry air under
isothermal conditions at 10000 C (1273K) up to 24 hours. The weight change was measured at the end of each time
interval with the help of an electronic balance (Metler Toledo) with a sensitivity of 0.01 mg. After 24 hours
duration, the specimen inside the furnace was subjected to cooling at the rate of +/- 40C. The detail experimental set
up is shown in Fig 3.

3. Results and Discussion

3.1 Oxidation rate and kinetic behavior
The specimens for both the micron size and nano structured coated specimen are taken for kinetic study. The
oxidation rate and kinetic behaviour of micron size and nano structured YSZ coated specimen is presented in Fig4.
The figure shows that the oxidation rate of nano structured YSZ coated specimen is much lower than the micron size
YSZ coated specimen. The figure also suggests the significant increase of oxidation rate in case of nanostructured
YSZ coated specimen. The kinetic behaviour of both the coated specimen follows the parabolic growth rate (Fig 4),
which indicates that the oxidation process is diffusion controlled and governed by the outer cation and inner anion
migration.

100
Effect of nanostructured 8% partially yttria stabilized Zirconia (YSZ) coating in oxidation behaviour of Inconel alloy

Fig 4: Oxidation Rate and Kinetic behaviour of nano structured and micron size YSZ coated specimen

The parabolic rate constant (KP) are calculated by the following equation

 ∆W 
2

  =K p t ... (1)
 A 
Where ΔW/A = weight gain per unit area (mg/cm2)
K P= parabolic rate constant (mg2cm-4s-1) t= time in seconds
The parabolic rate constant of micron size YSZ coated and nano structured YSZ coated specimens are calculated
and found to be 6.2296X10-7 mg2cm-4s-1 and 3.3606X10 -7 mg2cm-4s-1 respectively. The result indicates that there is
significant decrease in parabolic rate constant in case of nano structured YSZ coated specimen.

3.2 Characterizations of Oxidized Samples

3.2.1 Surface morphology - The surface layer of the post oxidized specimens is characterized in Field Emission
Scanning Electron Microscope (Model: SIGMA HD Make: Zeiss Ltd, Germany). The microstructures of micron
size YSZ coated specimen consists of fully molten and some unmelted YSZ particle. The surface feature also
shows porosities and cracks on the top surface of the coating(Fig 5a). The coating is not so adherent( rather fragile
in nature) due to presence of unmelted zone. On the other hand, the coating microstructure of the nano structured
YSZ coating reveals adherent and free from cracks and porosities. It consists mostly with fully molten zone along
with semi molten (Fig 5b and 5c.) At higher magnification the nano structured particles are observed densely
packed in fully molten and semi molten zone( Fig 5c). The semi molten zone at higher magnification shows the
presence of nano YSZ particle at higher magnification (Fig 5d). The top surface of the coating is packed densely
which evidences the good adherence of The coating. This nano structured coating is formed by mixture of full
molten zone and semi molten zone((bimodal structure.) in the plasma spray jet
3.2.2 XRD analysis - Fig. 6 shows XRD pattern of the existing phases on top coat surface after oxidation. The XRD
analysis of micron size coated specimen shows mainly tetragonal zirconia ( strong phase) along with monoclinic
zirconia and tetragonal YO 4 ( weak phase) (Fig 6a). In contrary, the XRD analysis of nanostructured YSZ coating
indicates the presence of tetragonal zirconia( main phase) along with monoclinic zirconia and tetragonal YO 4. the
monoclinic zirconia and tetragonal YO 4 are formed during oxidation of specimens. The only difference is that the
formation of monoclinic zirconia and tetragonal YO 4 is much less in case of nana structured coatings (Fig 6b). The

101
 Proceedings of NCAMMM - 2018

morphology of the nano structured YSZ surface, after oxidation ,show the oxidation products looking like nano
needle shaped. (Fig 5c).

a b
Fully molten
Unmelted zone Zone
Fully molten
Semi molten

Fig 5: Surface morphology of post oxidized(a) micron size YSZ coated specimen(b) nano structured YSZ coated
specimen (c) semi molten zone of nano structured YSZ coating showing typical nano YSZ particles( rod shaped)

The needle shaped oxidation products are possibly tetragonal YO 4 as indicated from XRD analysis (Fig 6b).
Formation of YO 4 crystals causes extraction of yttrium (Y) from YSZ with a consequence of destabilization
followed phase transformation [24]. The YO 4 formation is much less in nano structured YSZ due to packness of
nano structure.

a b

Fig 6: XRD analysis of post oxidized (a) micron size YSZ coated and (b) nano structured YSZ coated specimen

102
Effect of nanostructured 8% partially yttria stabilized Zirconia (YSZ) coating in oxidation behaviour of Inconel alloy
4. Discussion
One of the major reason for the better high temperature corrosion resistance of nano structure coating can be due to
the fact that former forms a stable zirconia, which does not crack as the temperature is cooled down, while the
micron size YSZ coating forms tetragonal zirconia, which transforms to monoclinic phase change and hence further
results the volume change and cracking of the scale. Fig 5a clearly suggests the cracks on the top scale of the post
corroded micron size YSZ coated specimen while Fig 5b and Fig 5c shows the adherent scale without cracking of
the scale. in case of nano structured coatings.during Oxidation, the nanostructured coating results in formation of
tetragonal zirconia as stable phase. This results in resistance to cracking of the scale, which finally restriction in
migration of cation/ anion during oxidation and improves the high temperature oxidation resistance.
In this context, the chemical compositions of YSZ powder is also important. In this experimental studies, 8 mol
percent of Y 2 O 3 is used along with 92% zirconia (ZrO 2 ). In this type YSZ powder, the formation of tetragonal
ZrO 2 phase depends on the particle size of this type of YSZ powder. The nano YSZ powder are extremely fine in
size( within 100 nm) and this size YSZ powder during plasma coating process may help to stabilize the tetragonal
ZrO 2 phase during oxidation, while the micron size powder during coating process results conversion of some
partially stable tetragonal zirconia to monoclinic zirconia phase. The results of the oxidation tests indicates that
nanostructured YSZ coating decreases oxygen diffusion towards NiCrAlY bond coat through YSZ top coat and acts
as a strong barrier for outward cation migration and to some extent inner anion migration. In contrary, the micron
size YSZ coating shows more migration of oxygen towards the substrate and finds more oxidation rate. The
formation of oxidation products( monoclinic zirconia and tetragonal YO 4 ) as a result of oxidation as suggested from
XRD analysis (Fig 6). This clearly indicates that the micron size YSZ coating suffers more oxidation than nano
structured coating.

5. Conclusions
The following conclusions can be made on the basis of results and discussion.
1. The nano structured YSZ coating over inconel substrate is better oxidation resistant than micron size YSZ
coated specimen at dry air at 1273K.
2. Both the YSZ coated specimen follow parabolic rate kinetics, which indicates that the corrosion is governed by
diffusion growth (i.e.cationic and anionic transport) during oxidation process. The parabolic rate constant (K P ) of
the nano structured specimen is much lower than the micron size coated specimen.
3. The surface morphology of the post corroded micron size YSZ coated specimen shows fully molten zone,
unmelted zone along with cracks and porosities in between the two zones. The cracks and porosities behaves
the site for short circuit diffusion path for ionic transport and enhance the corrosion rate. The cross sectional
view of the coating also confirms the presence of thermally grown oxide(TGO) in between coating and substrate,
which provides adequate thermal insulation..
4. The surface morphology of the nano structuted YSZ coated specimen consists of molten and semi molten
zones along with nano particles of YSZ mostly in the semi molten zone. The nano YSZ particles easily
transports in grain boundary and restricts the short circuit diffusion path for cation and anion migration and
improve the oxidation resistance to a significant extent.

103
 Proceedings of NCAMMM - 2018

5. The better oxidation resistance of nano structured YSZ coating can be attributed to the formation of stable
tetragonal zirconia ( ZrO 2 ) which is responsible scale cracking and spallation of the scale during cooling. At the
same time, the micron size YSZ coating forms tetragonal zirconia during oxidation, which transforms to
monoclinic zirconia phase change and hence further results the volume change and cracking of the scale. The
XRD analysis of both the coating justifies the stable and unstable phase after oxidation of both nano structured
YSZ and micron Size YSZ specimen.

References
[1] R Taylor, J.R Brandon , P.Morrel, Microstructure, composition and property relationship of plasma sprayed
thermal barrier coatings, Surface and Coating. Technology, 1992, 50 ,141.
[2] D.J Wortman, B.A Nagaraj , E.C; Duderstadt :Thermal barrier coating for gas turbine use, Material. Science.
Engineering A,1989, 121 ,433.
[3] R.A Miller, Current status of thermal barrier coating-an overview, Surface and Coating Technology. 1987,30
,1.
[4] M Gell , E.H. Jordan, K. Vaidyanathan, K. McCarron, B. Barber, Y. Sohn and V.K. Tolpygo, Bond strength,
bond stress and spallation mechanism of thermal barrier coatings, Surface and Coating Technology,1999, 53,
120–121.
[5] H Liu, . Q Xue, The tribological properties of TZP graphite self lubricating ceramics, Wear, 1998, 198, 143
[6] H Ahn, J Kim Lim D, Tribological behaviour of plasma sprayed zirconia coatings, ` Wear,1997, 203-204,
77
[7] G Fargas , D Casella, L Llanes and M Anglada, Thermal shock resistance of yttria stabilized zirconia with
palmqvist indentation cracks, Journal of Europian. Ceramic. Society. 2003, 23, 107-114
[8] P Bengtsson, T Ericsson T, J Wigren, Thermal shock testing of burner cans coated with a thick thermal
barrier coating, Journal of Thermal Spray Technology.,1998, 7, 340
[9] W.D Kingery, H. K Bowen and D.R Uhlmann Introduction to Ceramics ,John Wiley, New York, 1976, p.
822
[10] R.J., Rratton, A.H Heuer, L.W Hobbs (Eds.), Advances in Ceramics, Science and Technology of Zirconia
,American Ceramic Society, Columbus. OH, 1981, p. 226.
[11] 11. R.J., Rratton, A.H Heuer, L.W Hobbs (Eds.), Advances in Ceramics, Science and Technology of
Zirconia ,American Ceramic Society, Columbus. OH, 1981, p. 226.
[12] D Ghosh and S K Mitra, Effect of Y 2 O 3 Superficial Coating on the High Temperature Corrosion behavior of
2.25 Cr-1 Mo steel in SO 2 +O 2 atmosphere. Journal of Institution Of Engineers (India): Series D, Springer
publications, 2012,9(2), 59-63
[13] D Ghosh, S Mukherjee and S Das, High temperature Oxidation behavior of yttria( Y 2 O 3 ) coated low alloy
steel, Surface Engineering (DOI 10.1179/1743294414Y.0000000271)
[14] S Patil, S C Kuiry , S Seal, Nano crystalline ceria imparts better high temperature protection. Proceeding of.
Royal Society of London A, 2004, 460, 3569-3587

104
Effect of nanostructured 8% partially yttria stabilized Zirconia (YSZ) coating in oxidation behaviour of Inconel alloy
[15] D Ghosh, A K Shukla, S K Mitra ,Effect of ceria coating on corrosion behavior of low alloy steel, Surface
Engineering, 2013,29(8),584-587
[16] D Ghosh, A K Shukla, S K Mitra, Influence of CeO 2 superficial coating on the High Temperature Corrosion
Behavior of 2.25 Cr-1 Mo steel in SO 2 +O 2 atmospheres, Protection of Metals and Physical Chemistry of
Surfaces,2013, 49( 6),749-752,
[17] H Chen, C Ding, P Zhang , P La and S. W Lee, Wear of Plasma sprayed nanostructured zirconia coatings
against stainless steel under distilled water conditions, Surface and Coating Technology, 2003,173, 144-149
[18] H Chen and C.X Ding, Nano structured zirconia coatings prepared by atmospheric plasma spraying, Surface
and Coating Technology.2002, 150, 31
[19] C Berndt and E.J Larernia, Thermal spray processing of nano scale materials-a conference report with
extended abstract, Journal of Thermal Spray Technology. 1998, 7 (3), 411
[20] B.H Kear and G Skandan, Nanostructured Materials.1997, 8 (6), 765
[21] 21. M Cell ,Application opportunities for nanostructured materials and coatings. Materials Science and
Engineering A,1995, 204, 246.
[22] J Karthikeyan, C.C Berndt, J Tikkanen and S Reddy and H,Herman, Plasma spray synthesis of nano material
powder and deposits, Materials Science and Engineering A,1997,238, 275-286
[23] M. Gell, Nanostructured Materials,1995, 9, 997
[24] M Saremi, A.,M., Kevyaniand and M H Sohi, Hot corrosion resistance and Mechanical Behavior of
atmospheric plasma sprayed conventional and nano structured zirconia coatings, 2nd international conference
on ultrafine Grained & Nanostructured Materials (UFGNSM) International Journal of Modern Physics:
Conference Series Vol. 5 (2012) 720–727

105
Sub - theme

Mechanical
Metallurgy
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Effect of Heat Treatment Parameters on the Carbide Spheroidization of 0.48% Carbon

Steels

Nandita Gupta1, S.K.Sen2

1
Foundry Tech Dept, NIFFT, Ranchi, ng_nifft@yahoo.com
2
SAIL, Ranchi, sksen_51@yahoo.com

Abstract: Cast structures unless designed in strict conformity with the natural characteristics of metal solidification,
sometimes contain internal defects or surface imperfections which may seriously affect serviceability. In order to
render them serviceable they must be subjected to a heat treatment which will refine the grain and breakup the
dendritic structure, relieve internal stresses, and develop the desired mechanical properties. Test study on fast
spheroidizing annealing process for successful and early spheroidization on as cast 0.48%C steel samples, forge
reduction method is tried which is easily adaptable in Industries after that subcritical annealing treatment is done. It
is found that with more warm forging enthalpy is more. Hardness within 3hrs of annealing after 20% forge
reduction is reduced indicating of 1.8 µm spheroidized particle. DTA and TGA shows first peak of recovery at
140○C-150○C, second peak transformation from ferrite to austenite at 700○C and last peak at 900○C -920○C.
Microstructure evolution in fully dense warm forged steel shows that carbon diffuses out from cementite lamellae to
ferrite matrix and slowly small cementite particles began to grow by coarsening.
Keywords: Carbon Steel, Thermomechanical Treatment, Spheroidization

1. Introduction
The 0.48%C steel find applications where formability is the critical parameter in addition to strength such as in nuts,
bolts and machinery components. Variety of conventional and non-conventional methods [1-3] had tried for
spheroidization in 0.48% carbon steels by several researchers. Compatibility of the methods of spheroidization
solely depends on individual's choice and commercial factors apart from those of technical reasons. For successful
and early spheroidization, subcritical method of spheroidization is found to be more suitable method. Forge
reduction method on as cast samples is easy adaptable in industries. Even with less sophisticated environment it is
found to be a viable method. In this investigation, the sub-critical process [4-5] is compared with direct cast forge
annealing.

2. Experimental
The present investigation is to study the application of thermomechanical treatment (warm forging) on 0.48% C
steel and also to exploit beneficial effects of grain refinement during repeated cycle of hot deformation by forging.
At first a piece of Armco iron (0.02%C) brick of size 80×80 mm² weighting, 45 kg was placed in induction furnace
with 1600°C tapping and a 15kg melt is cast in open ingot mould covered with anti-piping compound to avoid the
piping during solidification. Reheating and forging at 910°C is done with reduction of sizes from 25mm to 20 mm,
15 mm and 10mm in subsequent reduction cycle of 20%, 40% and 60%, respectively and then the samples are taken
for Dilatometry examination to check the Ac1 and Ar1 temperature. DTA and TGA of as cast and forge samples are
done to evaluate phase transformation peaks and thermodynamic data of the peaks. NDT measurements after
 Effect of Heat Treatment Parameters on the Carbide Spheroidization of 0.48% Carbon Steels
different times of spheroidization are used to measure elastic modulus and damping coefficients. Optical
microscopy, EPMA, SEM, XRD investigations are done to check the progress of the work. The chemical
composition of the sample are shown in Table 1.
Table 1 N2 20% deformation
%Co %C %P %S %Mn % % % % % % % % %
Si Cr Mo Ni Al Cu Fe Nb Ti
0.0049 0.5 <0.0014 0.024 0.374 0.0678 0.0433 <0.0016 0.0274 0.0272 0.6088 98.99 0.0068 0.0052

3. Results and Discussion

It is found that with more than 50% reduction and with holding directly at 710°C for 3 hours (Fig. 4) increase
volume fraction of a spheroidized pearlite and it leads to better combination of strength and toughness showing
proportional relation of these properties. Course austenite grain structure prior to transformation hardly affects the
spheroidization process and does not deteriorate mechanical properties after the strain induced pearlite
spheroidization. The cementite constitution was all orthorhombic systems of Fe3C type (Table 2). This implies that
cementite constitution did not change in the heating process. Here we had examined the process of strain induced
spheroidization of cementite of pearlite in hypoeutectoid steel and also of cementite in hypereutectoid steel to the
technology of direct charging of cast bricks.
To avoid the detrimental effect of the coarse austenite grain structure developed during consequent
annealing and normalizing treatment of the cast bricks. Also it is examined how far the strain induced fine globular
spheroidization of pearlite cementite, produced by hot forging integrated in the direct charging of as cast brick could
substitute the conventional quenching and tempering process of engineering carbon steels, Subcritical annealing at
710±5°C for 3, 6, 9, 12 hours are done. The dilation rate was found 1.31×10-5/°C in steel up to 400°C. An increase
in the dilation rate was observed beyond 400°C. Taking different cooling rates of 10°C/sec and 2°C/sec, phase
transformation temperatures (Ac1 and Ac3) for 0.48%C steel are as follows 671○C to 729°C and 713°C to 729°C
respectively. Around 15°C to 30°C was the difference in critical temperatures from as calculated by lever’s rule.
With deformation it is found that percent of carbide particles are much higher.

Table 2 Nucleating temperature of different phases (Cementite and ε-carbide)

Cementite (Fe3C) Orthorhombic Structure Nucleates at dislocations grain boundaries >200°C
ε-carbide (Fe2.4C) HCP Structure Matrix dislocations <250°C

XRD results (fig1) shows that in 0.48%Carbon steel, three ferrite peaks and one cementite peak was found. The
biggest ferrite peak shows the largest presence of ferrite in the composition[Table3]. the results are same as being
found by earlier researchers [1-6] which justifies the earlier spheroidization via hot forging reduction. As per K.
Aihara [6-8] direct spheroidizing annealing on given below composition was 62.5% deformation in 6 passes at
675°C (0.45%C, 0.23%Si, 0.77%Mn, 0.015%P, 0.018%S) a specimen of 40×40 mm for 1 hour treatment time. Our
chemistry is- (0.48%C, 0.05%P max, 0.6%Mn max, 0.3%Si max, 0.02%Al max, traces of Mo, Ni, Cu, and rest was
iron.) It seems Direct casting and forging annealing within 3 hrs span resulting in 60% spheroidization can be
commercially applied for the production of Spheroidized steel.

107
 Proceedings of NCAMMM - 2018

(a)

(b)
Fig. 1. XRD of steel after 60% deformation and(a) 6 hours and(b)12hrs spheroidization

108
Effect of Heat Treatment Parameters on the Carbide Spheroidization of 0.48% Carbon Steels
Table 3. Phase analysis by X-ray diffraction using peak area of ferrite (α phase) [2026100 (110)] and cementite
phase [2.01100(031)] N2 40%, 12 hours treatment reveal in % Cementite is shown in equation below

Ferrite peak Cementite peak

2θ θ Sinθ D (mm) area F mm² area F mm²
F 18.75 9.375 0.162895 2.182381 3.75×57 9×10.5

Equation 1: {C/(F+C)}*100= (94.5/308.25)*100= 30.66

F= Ferrite (α) 100 in intensity peak, C= Cementite 100 intensity peak
The mass of material under investigation is continuosly followed as a function of temperature or time in DTA/TGA .
curve (Table 4). All the variations resulting from these factors are manifestations of heat and mass transfer the two
important phenomenon.
Table 4. Effect of hot deformation (forging) on enthalpy in different samples as measured by DTA/TGA
Experiments
S No. Sample No. Temp Range Enthalpy Area µv/sec
1 N2 40% HF 720-740°C 9.965 J/Kg 253
2 N2 0% HF 720-740°C 8.969 J/Kg 173

(a)

(b)
Figure 2 (a) and (b): DTA/TGA enthalpy Curve for 0.48%C steel without deformation

109
 Proceedings of NCAMMM - 2018

NDT measurements found that the modulus of elasticity is reduces in 20% deformed samples from as cast steels
whereas in 40% deform sample steel again retain its original modulus of elasticity further deformation increases
further the modulus of elasticity. Poisson's ratio given after simeon poisson (1781-1840) it is the ratio of lateral
strain to longitudinal strain. In this investigation it is found that from lesser deformed samples the more deformed
samples have reduced trend of poisson's ratio(Table 5).

Table-5 Modulus of elasticity and Poissons ratio by NDT Method

Sample Treatment Modulus of Elasticity (KN/mm²) Frequency(Hz) Poisons Ratio
N2 40% Deformation 214.57 8676 0.233

Figure 3a shows the the carbide precipitation and growth which takes place initially primarily at grain or subgrain
boundary sites . With deformation by warm forging there is an increase in thermodynamic potential. As per
M.O.Brien[1,6] it was confirmed that this enhancement in spheroidization is the important achievement under
temperature and strain rate conditions and is is primarily by diffusion and also the presence of the dislocations are
the contributory factor.

Fig. 3 Medium Carbon Steel 0.48%C &40% Deformation (a) 6 hrs annealing and (b) 9 hrs annealing heat treatment

Fig. 4 Medium Carbon steel 0.48% . with 60% Deformation, 3 Hrs Heat Treatment
With 40% Deformations and 9Hrs heating the process of conversion intensifies which further results in cementite
platelet division into parts, transformation of lamellar cementite into granullar form and also coalescence of carbide
particles resulting in free cementite globurization. With 60% deformation and 3Hrs annealing the 0.48%C steel fine
pearlite in ferrite and cementite matrix is created. Iron diffusion plays a key role in spheroidization. No systematic

110
Effect of Heat Treatment Parameters on the Carbide Spheroidization of 0.48% Carbon Steels
variation of ferrite grain resulting from heating is observed. It is found that preexisting retained cementite also acts
as nucleation site of spheroidal cementite. The initial drop in hardness during spheroidization shows the recovery of
grains and more number of sites for heteroge neous nucleation. Further increase in hardness is due to fine
recrystallized grains which provide less nucleation sites. Thereafter it does not decrease until spheroidization is
commenced fully (Fig. 4).

Fig 5. Vickers Hardness of 0.48%C steel with 60% warm forge reduction

4. Conclusion
The hardness when observed by Vickers hardness test in 0.48%C steel, 20% deformation leads to minimum
hardness for 3 hours annealing (Fig. 5). After 40% deformation hardness increases further with increase of
spheroidization time to 12 hours and hardness decreases. With increase in spheroidization time the size of cementite
increases (Table 2) and there is loss of carbon in the matrix thus overall hardness falls after 60% deformation (Fig.
5) rate of spheroidization becomes faster in 0.48%C steel. Fall in hardness to 120 VPN show maximum
spheroidization of the steel. This drop in hardness could also be due to recrystallization of ferrite matrix. Ferrite
grains have become finer (5µm) (Fig.3). Optical microscopy reveal mixed structure consists of (Fig 3,4)ferrite and
bainite after 710°C, 3 hours 20% deformed specimens both the heated annealed steels have very fine ferrite grain.
After 6 hours annealed 60% deformed samples no evidence of the original structure was found now cementite
particles are now quite massive considerably larger than the original thickness of cementite in pearlite phase
spheroidized grain size 1.8µm minimum and 3.2µm maximum. There was not much change in microstructure except
same elongation in grains are found. The XRD results (Fig 1) of further deformation by forging and air cooling

111
 Proceedings of NCAMMM - 2018

(Table 3)show low energy interface between ferrite and cementite in pearlite and that the driving force for
spheroidization is then small. The DTA and TGA measurement (Table 4) is a technique in which the mass of the
material under investigation is continuously followed as a function of temperature or time as it is heated or cooled at
the predetermined rate. The result reveal all the variations is the manifestations of heat and mass transfer
phenomenon. It is found that more the enthalpy faster is the rate of spheroidization.
TGA result shows rate of oxidation is less up to 780°C -800°C. Beyond that rate of oxidation is much
more. The first peak (at 140°C -150°C) fig. 2 is recovery, 2nd peak (700°C onwards) is for spheroidization of
cementite, at 900°C -920°C jump is for the start of transformation from ferrite and austenite to full austenite.
Dissociation of cementite lamellas of pearlite occurs or rather carbon diffuses out from cementite lamellas to ferrite
matrix and slowly small cementite particles again began to grow by coarsening. The mass change is mainly due to
thermo molecular forces which arises from different velocities with which the gas molecules travel from different
temperature regions of the tube. NDT measurements say that modulus of elasticity reduces from as cast steels
initially and with more deformation modulus of elasticity increases. The reading of poisons ratio (Table 5) with
deformation samples show reduced trend of poisons ratio in more deformed samples.

References
[1] J.M.O Brien and W.F Hosford, J. Mater Eng. Performance 1997,Vol 6, pp 69-72
[2] N.Gupta and S.K.Sen "effect of hot forging on spheroidization of medium carbon steel" Steel India,
September 2006, Vol 29, No. 1, pp 26-32
[3] M.Nishimani, M .A ratani and M. Kaka, Kawasaki steel Giho,33 (2001), vol 4 pp151
[4] N.Gupta and S.K.Sen " Spheroidization treatments for steels"Defence Science Journal, October 2006, Vol 56
No. 4, pp665-676
[5] Yoshikazu kawakawa, Y akatoshi Okabe and Yankee Koyama “Development of high carbon history steel tube
with excellent formability” K Kawasaki steel technical report no. 47, December 2002.
[6] J.M.O Brien and W.F Hosford ,Metallurgical and Material Transaction A, April 2002, vol 33A , pp1255
[7] K.Aihara, "Anew Thermomechanical Processing for Spheroidizing Carbide directly in Rolling Line", 33RD
MWSP CONF, PROC, ISS-AIME, 1992, Vol XXIX , 285.
[8] K.E, Thelning ,Steel and its Heat Treatment ,2nd edition ,Butterworth’s London 1986.

112
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Modelling Tertiary Creep of In-Core Pure Aluminium using Hyperbolic Function

K. Vinaya, A. Syeda, M. K. Samala A. Aryab

a
Reactor Safety Division, Bhabha Atomic Research Centre, Mumbai 400085, India,
E-mail: kvinay@barc.gov.in
b
Material Science Division, Bhabha Atomic Research Centre, Mumbai 400085, India

Abstract: Boron free pure aluminium is a popular candidate extensively used as in-core material in nuclear
reactors. High temperature deformation under creep is an important parameter to be understood for the design of
various in-core structure of a nuclear reactors. Through analytical modelling of material characteristic and
benchmarking with experimental data could help reduce number of experiments and time involved. An analytical
model has been developed incorporating LMP, modified Garofalo and Robinson methods. The rupture time is
predicted by modified LMP formulation, the tertiary region is captured through time weighted Robinson formulation
and secondary region is expressed through modified Garofalo equation. Hence, an attempt is made to bridge the
simplistic model from secondary and tertiary region which could be handy designer’s tool. The model is being
extensively implemented to estimate the rupture time, secondary-tertiary transition time to design various in-core
components in nuclear reactors, without invoking cumbersome calculations involving estimation of several material
constants to determine damage coupled creep expressions.
Keywords: creep, pure aluminium, analytical modelling, modified Garofalo, Robinson, time weighted, LMP,
rupture time, secondary, tertiary

1. Introduction
The material under study is reactor grade boron free pure aluminium which is a very good candidate material for
many in-core components of nuclear reactor. The high temperature creep deformation of a material is often
characterized by steady state secondary creep region [1], initiation of tertiary creep region and rupture time.
Hyperbolic sine function used in Garofalo model has great potential to predict the steady state secondary creep
region.[3] There are many damage based expressions to model the tertiary region, however, they are cumbersome
and computationally expensive and may not be friendly for designers [4]. Often, the formulation of damage based
expression is power law based. Hence, an attempt is made to bridge the simplistic model from secondary and
tertiary region which could be handy designer’s tool, which is hyperbolic function based.

2. Experimental Data and Development of Model

Experiments were carried out on reactor grade boron free pure aluminium cylindrical specimens. The raw data
acquired through experiments were processed and plot. The analytical model was development based on the
experimental data and tailored to the material under study. Experimental conditions are tabulated in Table-1 below:
Modelling tertiary creep of in-core pure aluminium using hyperbolic function
Table 1: Details of experimental conditions in creep test
Case-1 Case-2 Case-3 Case-4
o o o
T = 300 C T = 350 C T = 400 C T = 425 oC
σ = 15 MPa σ = 7.5 MPa σ = 3 MPa σ = 4 MPa
T/Tm = 0.45 T/Tm = 0.53 T/Tm = 0.60 T/Tm = 0.65
σ/σy = 0.33 σ/σy = 0.375 σ/σy = 0.6 σ/σy = 0.80
Note: Tm – Melting Temperature (oC); σy – Yield Stress (MPa)

The homologous stresses under consideration have been chosen ranging from 0.3 to 0.8 and homologous
temperatures from 0.45 to 0.65. Ergo, the temperatures and stresses encompass all the quadrants of homologous
temperature to homologous stress matrix. Hence, the model developed would be universal representing all the
ranges of stresses and temperatures. The methodology adopted in modeling the creep behavior predicting secondary,
tertiary and rupture time is by bridging Garofalo, Robinson and LMP models modified and tailored to reactor grade
boron free pure aluminium material. The rupture time is predicted by modified LMP formulation [5], the tertiary
region is captured through time weighted Robinson formulation [6] and secondary region is expressed through
modified Garofalo equation.
dε/dt = A.sinh(σ/G)n.exp(-/RT).sinh(t/tr) (1)

Where, ε – strain, A = 10727, σ = stress in MPa, G = Shear modulus in MPa, Q – Activation energy, R – Ideal
gas constant, T – temperature in K, t – time, tr – estimated rupture time.

P =d sinh (e+f.σ m)+g (2)

Where, P – parameter, d = -1359 K, e = -3.59, f = 1.34 MPa-1, m = 0.359, g = 11359 K, σ = stress in MPa

tr =10P/T-C (3)
Where, tr – rupture time (s), P – parameter (K), T – temperature (K), C – constant

Plots comparing experimental data and estimation by proposed model are made and presented in the figures
below:

20 Creep strain vs Time

at 300 oC and 15 MPa
15
Creep Strain (%)

10 Model

0
0 2000 4000 6000 8000 10000 12000 Time (s)

Figure 1: Comparison between experimental data and model at temperature of 300oC and 15 MPa stress

114
 Proceedings of NCAMMM - 2018

60
Creep strain vs Time
50

Creep Strain (%)

at 350oC and 7.5 MPa
40
30
20 Model
10 Experimental data
0
0 5000 10000 15000
Time (s)
Figure 2: Comparison between experimental data and model at temperature of 350oC and 7.5 MPa stress
30
Creep strain vs Time
25 at 400oC and 3 MPa
Creep Strain (%)

20
15 Model
10 Experimental data
5
0
0 5000 10000 15000
Time (s)

Figure 3: Comparison between experimental data and model at temperature of 400oC and 3 MPa stress

14
Creep strain vs Time
12
Creep Strain (%)

at 425oC and 4 MPa

10
8 Model
6
4 Experimental data
2
0
0 5000 10000
Time (s)

Figure 4: Comparison between experimental data and model at temperature of 425oC and 4 MPa stress

The hyperbolic based model is reasonably accurate and does capture various regions of the creep curve. It is able to
represent secondary, secondary-tertiary transition and tertiary region within experimental uncertainties. The
comparison of transition region from secondary to tertiary region estimated through modified robinson method is
compared with experimental data in the Table 2.

115
Modelling tertiary creep of in-core pure aluminium using hyperbolic function
Table 2: Comparison of transition of secondary to tertiary region

Case-1 Case-2 Case-3 Case-4

o o o
T = 300 C T = 350 C T = 400 C T = 425 oC
σ = 15 MPa σ = 7.5 MPa σ = 3 MPa σ = 4 MPa
Exp = 9000 s Exp = 7970 s Exp = 6500 s Exp = 5700 s
Mod = 9700 s Mod = 8250 s Mod = 6740 s Mod = 5825 s
Error = ~7.8% Error = ~3.5% Error = ~3.7% Error = ~2.2
Note: Exp – Experimental data (s); Mod – Estimated by Model (s)

3. Conclusion
A hyperbolic based simplistic formulation is developed which estimates secondary, secondary-tertiary transition and
tertiary region. The plots of model prediction are extensively tested against experimental data at various stresses and
temperatures. This formulation is time implicit and time weighted based expression. The rupture time is predicted by
modified LMP formulation, the tertiary region is captured through time weighted Robinson formulation and
secondary region is expressed through modified Garofalo equation. Hence, an attempt is made to bridge the
simplistic model from secondary and tertiary region which could be handy designer’s tool. The model is being
extensively implemented to estimate the rupture time, secondary-tertiary transition time to design various in-core
components in nuclear reactors, without invoking cumbersome calculations involving estimation of several material
constants to determine damage coupled creep expressions.

Acknowledgments
Author1 would like to thank Shri K. N. Vyas, Director, BARC for providing an opportunity to study on this subject
matter.

References
[1] Cocks ACF, Ashby MF (1982). “Creep fracture by coupled power-law creep and diffusion under multiaxial
stress.” Metallurgical Science, 16:465–474.
[2] Courtney TH (2000), “Mechanical behavior of materials.”, McGraw-Hill, New York
[3] Garofalo F. (1965), “Fundamentals of Creep and Creep-Rupture in Metals”, McMillan. Series in Materials
Science, McMillan, New York.
[4] Kachanov LM (1999), “Rupture time under creep conditions.” Int. Journal of Fracture, 97:11–18.
[5] Larson, F.R., Miller, E.J. (1952), “Time temperature relationship for rupture and creep stresses”, Trans. ASME,
vol. 74, p. 765- 775
[6] Robinson, E.L. (1938), “Effect of temperature variation on the creep strength of steels”. Trans. ASME, vol. 160,
p. 253-259

116
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Structural Analysis and Measurement of Mechanical Properties of Sputter Deposited

Tungsten Films

Satyajit Sarkar*, Shubhra Bajpai, Ankeet Pati

CSIR-Institute of Minerals and Materials Technology, Bhubaneswar - 751013, Odisha, India
*
Corresponding author: Satyajit Sarkar, E-mail: satyanitdgp@gmail.com

Abstract: Tungsten films were deposited at various working pressures by magnetron sputtering. The effect of
varying deposition pressure on the structural and mechanical properties of the tungsten films have been investigated
here by using GI-XRD, FESEM, nanoindentation and scratch test. The films were found to be highly crystalline
comprising of both α and β-W when deposited at lower pressures whereas α-W becomes predominant at higher
pressures. The films developed at higher pressures were found to possess better mechanical properties like higher
hardness, elastic modulus and scratch resistance as compared to the films deposited at lower working pressures.
Keywords: Tungsten films; Magnetron Sputtering; Mechanical properties

1. Introduction
Tungsten (W) is known to possess some striking features like its high density (ρ = 19.30 g/cm3 [1]), thermal
stability (T m = 3693 K [1]), electrical conductivity (5.49 μΩ cm [1]) and wear resistance [2]. W thin films have
been studied extensively pertaining to the various applications they have, like optical sensors [3], semiconductor
interconnect layer diffusion barriers [4], diffracting layers in x-ray mirrors [5], wear resistant coatings for micro-
electromechanical systems [2] and plasma-facing surface coatings in tokamaks [6]. Tungsten thin films portray two
kinds of crystalline phases, namely body-centered cubic (bcc) phase (α-W) which is thermodynamically stable and
metastable (β-W) A15 phase [1,2]. The properties possessed by these phases are quite different. The two phases, α-
W and β-W, have lattice parameters of 3.16 Å and 5.04 Å, respectively [7]. The electrical conductivity of α-W is
usually more than that of β-W [8]. In comparison to this, it has been found that the metastable β phase possesses
higher hardness of about 24.5 GPa as compared to nearly 21.3 GPa of the α-W phase [9]. Magnetron sputtering
method has been used frequently to deposit tungsten films [1,2] due to its advantages like high deposition rate,
uniform deposition making the process economical and suitable for large scale industrial applications [7,10]. Till
date, there is only one report showing the influence of varying deposition pressures (≥ 20 mTorr) on the mechanical
properties of sputter deposited W films [11]. Further, another report revealed the mechanical properties of W films
deposited at 12 mTorr pressure [12]. W films were sputter deposited in lower pressure range≤ (15 mTorr) and the
corresponding changes in their structural and mechanical properties with change in pressure were investigated here.

2. Experimental
Silicon (100) pieces were used as substrates for deposition of the tungsten films. The substrates were thoroughly
cleaned before deposition using acetone and distilled water in an ultrasonic oscillator sequentially for 5 minutes and
dried using compressed air. Deposition was done in an RF magnetron sputtering system (AJA International). 99.99
% pure tungsten disc (2 inch diameter, 3 mm thick) was used as the sputter target along with argon as the sputtering
gas. Pre-cleaning of the substrates and target were done for 5 minutes in Ar plasma to remove any contamination.
Structural analysis and measurement of mechanical properties of sputter deposited tungsten films
The level of base pressure achieved before deposition was 6 x 10-6 Torr. RF power of 50 W was used along with
target to substrate distance of 80 mm. The films were deposited at working pressures of 5, 7, 10 and 15 mTorr at
room temperature. The Ar flow rate used was 10 sccm. The depositions were carried out for 15 minutes each. The
GI-XRD analysis of the samples was done using X-ray diffractometer (Rigaku) using Cu Kα radiation. Mechanical
properties like hardness and elastic modulus of the films were measured by nanoindentation using UMIS system
(Fischer-Cripps, Australia) consisting of a Berkovich diamond indenter having a tip radius of 150 nm. A load of 30
mN was applied during nanoindentation and a set of 12 tests were done on each sample to understand the average
response of the material. Indents on the same sample were separated apart by 15 µm. The Bruker-CETR UMT
scratch tester comprising of a Rockwell-C spherical indenter of 200 µm diameter was used to obtain the micro-
scratch test results. Data pertaining to various parameters like scratch depth, applied force, lateral force and signals
of acoustic emission were taken during the scratch test. During the tests, a constant load of 5 N was applied for 10 s
initially followed by ramp loading from 5 to 50 N for 63 s at a rate of 1 mm/min which lead to the formation of a
scratch of approximately 1 mm length. After an observation period of 15 days, the film deposited at 5 mTorr was
found completely delaminated from the substrate due to which nanoindentation and scratch tests could not be
performed on it.

3. Results and Discussion

3.1 Structural analysis

15mTorr

10mTorr

7mTorr

5mTorr

PAr

Fig. 1: GI-XRD plots of W films at various working pressures

The XRD plots of the deposited W films has been given in fig. 1. The ICDD database has been referred to for the
reference data of α and β tungsten phase (file no. 00-001-1203) and has been shown as vertical bars in fig. 1. By
observing the XRD peaks in fig. 1, it can be seen that at lower working pressure (≤ 7 mTorr), the films are relatively
crystalline and comprise of both α and β phases of tungsten. However, with increasing working pressure, the
formation of α-W phase became predominant which, in most cases is a desirable phase. The sputtered atoms
undergo excessive collisions and hence get scattered to greater extent randomizing the angle of incidence of the
vapor flux arriving at the substrate. This, in turn, reduces the degree of crystallization. Hence, tungsten atoms
sputtered at relatively very high pressures lacked atomic mobility due to increased collisional scattering as a result of
which they stick to the same site where they impinge. This leads to the entrapment of many lattice defects ultimately
resulting in poor crystal orientation. At lower deposition pressures, the sputtered atoms undergo lesser number of
collisions which improves the mobility of the tungsten atoms on the substrates. This, in turn, favors the formation of

118
 Proceedings of NCAMMM - 2018

multiple crystal structures and the entrapment of lattice defects during deposition decreases. Hence, multiple crystal
orientations, i.e., both α and β tungsten were observed in case of films deposited at lower pressures

Fig. 2: FESEM images of W films deposited at (a) 5 mTorr, (b) 7 mTorr, (c) 10 mTorr and (d) 15 mTorr.

Fig. 2 shows the images of the morphology of the W films deposited at various pressures. It is evident from fig. 2
that the film deposited at 5 mTorr is highly porous and is of relatively low quality. Film deposited at 7 mTorr
possesses comparatively coarse grains with a considerable amount of porosity. In comparison to this, the
morphological features of the films deposited at higher pressures of 10 and 15 mTorr are completely different and
demonstrate higher compactness. The film deposited at 10 mTorr possesses very fine grains as compared to the one
deposited at 15 mTorr having slightly larger grains as is evident from fig. 2. Hence, it is evident from the FESEM
images that the films deposited at higher pressures of 10 and 15 mTorr are much better than those deposited at lower
pressures of 5 and 7 mTorr in terms of the film quality. Furthermore, the differences in the morphological features of
the films deposited at higher pressures when compared to those deposited at lower pressures are justified by the
corresponding changes in the crystal structures of the films as revealed by the GI-XRD results.
The W coatings deposited at lower pressures (i.e., 5 and 7 mTorr) were found to delaminate. The film
deposited at 5 mTorr delaminated completely after 15 days. However, the film deposited at 7 mTorr has shown
slower delamination rate as it was found to be partially delaminated. Coatings were found to be well intact when
deposited at higher pressures. Thin films of refractory metals developed by sputtering tend to possess high residual
stresses [13]. Shen et al. [14] have found that W films are under compressive stresses when deposited in the low
pressure range of 2-12 mTorr. The surface mobility of the adatoms gets enhanced at low pressures primarily due to
the energetic bombardment of sputtered and reflected neutral particles as a result of the atomic peening process [15].
The small target to substrate distance might also be another reason for the energetic bombardment of sputtered
atoms. Waters et al. [16] have reported that W films having high compressive residual stresses are prone to
delamination. Hence, the delamination of the films deposited at lower pressures in this work may be attributed to the
presence of compressive residual stresses within the films.

119
Structural analysis and measurement of mechanical properties of sputter deposited tungsten films
3.2 Mechanical Properties

Fig. 2: P-h characteristic curves of W coating

The Oliver-Pharr analysis was used to calculate the hardness and elastic modulus from the load-displacement curve
[17,18]. Obtained depth versus load (h-P) curves have been shown in fig. 2.

Table 1: Nanoindentation test results of sputtered W films

Sample Hardness Elastic Depth

(H, GPa) Modulus (E, GPa) (h, µm)
7 mT 16.64 208.05 0.1226
10 mT 18.63 224.01 0.1167
15 mT 18.27 229.18 0.1168

The average values of the mechanical properties like elastic modulus and hardness have been given in table 1. The
hardness and elastic modulus of the deposited W films were found to be in a range of 16- 19 GPa and 200-235 GPa
respectively. Theoretical elastic modulus of W is 410 GPa [19]. As reported by Saha et al. [12], the hardness and
elastic modulus of sputtered W film on hard substrate-sapphire were found to be 14-15 GPa and 480 GPa
respectively. In that case, hardness was lower than the theoretical value, but modulus was found to be higher than
the theoretical value. They concluded that higher modulus was due to the partial contribution of the sapphire
substrate (E=440 GPa). Wong et al. [20] found the sputter W hardness to be equivalent to 20.7 GPa. Shih et al. [11]
have found out that sputtered tungsten films have hardness near to 17 GPa when deposited at a working pressure of
20 mTorr. The hardness of the films deposited at further higher pressures˃( 20 mTorr) follows a decreasing trend.
This gives us an indication that films deposited at higher pressures (≥ 20 mTorr) are likely to possess lower hardness
values as compared to those deposited at relatively lower pressures. The low hardness of the film deposited at 7
mTorr as seen from table 1 may be attributed to the presence of porosity and larger grain size as is evident from the
FESEM image in fig. 2. Ozkan et al. [21] have demonstrated that the mechanical properties of W thin films
deteriorate with increase in the amount of porosity induced in the films. Furthermore, the prime reasons for the
higher hardness values of the films deposited at pressures of 10 and 15 mTorr are considered to be the compact and
dense microstructure with relatively smaller grain sizes as seen from fig. 2. The film deposited at 15 mTorr has
slightly lower hardness as compared to the one deposited at 10 mTorr as seen in table 1. This may be attributed to
the relatively larger grain size of the film deposited at 15 mTorr as is evident from fig. 2. The influence of grain size
on the hardness of metals has been given by Hall-Petch equation,

120
 Proceedings of NCAMMM - 2018

H = H 0 + K.d-1/2 (1)

where H represents the hardness, d gives the grain diameter, H 0 and K are material specific constants. Hence, it can
be inferred that hardness of the nanocrystalline material is higher as compared to the bulk coarse grained materials
when the grain size decreases [22].
Furthermore, scratch tests were done to study the scratch resistance of the W films. Images of the scratched coatings
were taken after the tests. The load values at which the coatings delaminated were found out from the acoustic
emission data. Furthermore, the values of the critical load are found out from the images of the scratched coatings.
The coefficient of friction (COF) is given by the following relation:

µ= F x / F z (2)

where µ stands for COF, F x gives the lateral load, and F z is the normal load.

Fig. 3: Coefficient of friction (COF), normal load (F z ) and acoustic emission (AE) data obtained during scratch test
of sputtered W coatings deposited at (a) 7, (b) 10 and (c) 15 mTorr pressure.

The micro-scratch characteristics of the deposited films have been depicted in fig. 3. The damaged area has been
demonstrated by acoustic emission (AE) and represented in the optical micrograph. The origin of the critical loads
(LC 1 and LC 2 ) has been shown by COF and F z . Here x-axis and right y-axis represent the ramping test time and
normal load (F z ) respectively. Observed average values of the parameters are summarized in Table 2. LC 1

121
Structural analysis and measurement of mechanical properties of sputter deposited tungsten films
represents resistance against the initiation of the cracking. LC 2 represents the critical load at which the film
delaminates from the substrate completely ideally referred to as the adhesion strength. The crack propagation
resistance (CPR) of the scratch is analyzed on the basis of relative toughness of the films and is given as follows:

CPR = LC 1 (LC 2 -LC 1 ) (3)

Adhesion strength (LC 2 ) and coefficient of friction (COF) of W coatings were found to be in the range of 27- 34 N
and 0.07-0.11 respectively. It is evident that the films deposited at higher pressures usually portray higher adhesion
strength (LC gets shifted to the right with increasing pressure). It is also analogous to film stability, hardness and
elastic modulus. The load bearing capability (LC 2 -LC 1 ) and calculated CPR value (Table 2) follow a similar kind of
trend. The value of fracture toughness can be indirectly obtained from these data. The average value of the
coefficient of friction was also found to increase with pressure. The prime cause for such behavior might be the
exorbitant accumulation of abraded W in front of the indenter as W deposited at lower pressures is softer.
Comparing all the three micrographs, it can be seen that the 15 mTorr sample shows relatively neat cracking area
over the scratch which is an indication of its steady behavior against cracking.

Table 2: Critical cracking load and coefficient of friction (COF) of W sputter coatings.

Sample Critical Load Critical Load Load bearing Crack Propagation Avg.
(LC 1 ), N (LC 2 ), N capability Resistance COF
(LC 2 -LC 1 ) (CPR)
7 mTorr 17 27 10 170 0.07
10 mTorr 15 30 15 225 0.10
15 mTorr 18.5 34 15.5 286.75 0.11

4. Conclusions
The sputter deposited tungsten films were prone to changes in structure and mechanical properties on varying the
deposition pressure. The films were highly crystalline when deposited at lower pressures (≤7 mTorr), but α-W phase
gradually became predominant at higher pressures. The hardness of the films were found to increase with increase in
deposition pressure upto 10 mTorr followed by a slight decrease for the 15 mTorr sample. The elastic modulus of
the films demonstrated a continuous rising trend with increasing deposition pressure. The film deposited at the
highest pressure of 15 mTorr presented steady cracking characteristics as seen from the neat cracking area when
compared to the films developed at lower pressures. Hence, the films developed at higher pressures of 10 and 15
mTorr demonstrate superior mechanical properties and are more suitable for usage in wear resistant applications.

Acknowledgement
We would like to thank the Director, CSIR-IMMT Bhubaneswar for permitting to publish the results. This work was
supported by the Board of Research in Nuclear Sciences (BRNS) under Department of Atomic Energy (DAE), India
through grant-in-aid project with BRNS sanction no. 39/14/17/2016-BRNS.

References
[1] Martienssen W, Warlimont H. Springer Handbook of Condensed Matter and Materials Data. New York:
Springer, Heidelberg; 2005.

122
 Proceedings of NCAMMM - 2018

[2] Fleming JG, Mani SS, Sniegowski JJ, Blewer RS. Tungsten coating for improved wear resistance and
reliability of microelectromechanical devices, United States Patent No. 62908592001.
[3] Lita AE, Rossenberg D, Nam S, Miller AJ, Balzar D, Kaatz LM, Schwall RE. Tuning of tungsten thin film
superconducting transition temperature for fabrication of photon number resolving detectors. IEEE Trans.
Appl. Supercond. 2005; 15: 3528-3531.
[4] Rossnagel SM, Noyan IC, Cabral C Jr. Phase transformation of thin sputter-deposited tungsten films at room
temperature. J. Vac. Sci. Technol. B. 2002; 20: 2047-2051.
[5] Salditt T, Lott D, Metzger TH, Peisl J, Vignaud G, Hoghoj P, Scharpf O, Hinze P, Lauer R. Interfacial
roughness and related growth mechanisms in sputtered W/Si multilayers. Phys. Rev. B. 1996; 54: 5860-5872.
[6] Rieth M, Dudarev SL, et al. Recent progress in research on tungsten materials for nuclear fusion applications
in Europe. J. Nucl. Mater. 2013; 432: 482-500.
[7] Salamon K, Milat O, Radic N, Dubcek P, Jercinovi M, Bernstorff S. Structure and morphology of magnetron
sputtered W films studied by x-ray methods, J. Phys. D: Appl. Phys. 2013; 46: 095304 (10pp).
[8] O’Keefe MJ, Grant JT and Solomon JS. Magnetron sputter deposition of A-15 and bcc crystal structure
tungsten films. J. Electron. Mater. 1995; 24: 961-967.
[9] Vüllers FTN, Spolenak R. Alpha- vs. beta-W nanocrystalline thin films: A comprehensive study of sputter
parameters and resulting materials properties, Thin Solid Films. 2015; 577: 26–34.
[10] Seshan K. Handbook of Thin Film Deposition Processes and Techniques. Norwich: Noyes; 2002.
[11] Shih KK, Smith DA, Crowe JR. Properties of hard tungsten films prepared by sputtering. J. Vac. Sci. Technol.
A. 1988; 6: 1681-1685.
[12] Saha R, Nix WD. Soft films on hard substrates-nanoindentation of tungsten films on sapphire
substrates. Materials Science and Engineering: A. 2001; 319: 898-901.
[13] Vink TJ, Walrave W, et al. Stress, strain, and microstructure in thin tungsten films deposited by dc magnetron
sputtering. J. Appt. Phys. 1993; 74: 988-995.
[14] Shen YG, Mai YW, et al. Residual stress, microstructure, and structure of tungsten thin films deposited by
magnetron sputtering. J. Appl. Phys. 2000; 87: 177-187.
[15] Thornton JA, Hoffman DW. Stress related effects in thin films. Thin Solid Films. 1989; 171: 5-31.
[16] Waters P, Volinsky AA. Stress and Moisture Effects on Thin Film Buckling Delamination. Exp. Mech. 2007;
47: 163–170.
[17] Bajpai S, Gupta A, Pradhan SK, Mandal T, Balani K. Crack Propagation Resistance of α-Al2O3 Reinforced
Pulsed Laser-Deposited Hydroxyapatite Coating on 316 Stainless Steel. JOM, 2014; 66: 2095-2107.
[18] Das P, Anwar S, Bajpai S, Anwar S. Structural and mechanical evolution of TiAlSiN nanocomposite coating
under influence of Si 3 N 4 power. Surface and Coatings Technology. 2016; 307: 676-682.
[19] Oliver WC, Pharr GM. An improved method for determining hardness and elastic modulus using load and
displacement sensing indentation experiments. J. Mater. Res. 1992; 7: 1564-1583.
[20] Wong KM, Shen YG, Wong PL. Tribological properties of sputtered tungsten and tungsten nitride thin
films. Sci. China Ser. A Math Phys. Astron. 2001; 44, 242-247.
[21] Ozkan T, Demirkan MT, Walsh KA, Karabacak T, Polycarpou AA. Density modulated nanoporous tungsten
thin films and their nanomechanical properties. J. Mater. Res. 2016; 31: 2011-2024.
[22] Sun HL, Song ZX, Guo DG, Ma F, Xu KW. Microstructure and Mechanical Properties of Nanocrystalline
Tungsten Thin Films. J. Mater. Sci. Technol. 2010; 26: 87-92.

123
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17, February, 2018 at
CSIR-CMERI, Durgapur

Study the Effect of Basicity and FeO Contents of Slag on Dephosphorisation of Steel

Sujata Devi, Niladri Sen, Basudev Mishra

RDCIS, SAIL, RDCIS Rourkela Center-769011
Phone : 0661-2511208, E-mail: sujatadevi@sail-rdcis.com

Abstract: Phosphorus removal from steel is a key area of research because of its detrimental effect on mechanical
properties of steel. P in metal needs to be reduced by making a slag rich in P 2 O 5 and by avoiding P 2 O 5 reversal.
CaO fixes the P 2 O 5 by decreasing the activity of P 2 O 5 in the slag and by forming stable compounds like 3CaOP 2 O 5
and 4CaOP 2 O 5 . It is also important to maintain high FeO content of slag as it helps in converting the P present in
the metal to P 2 O 5 . However, beyond an optimum value, FeO replaces CaO which results in decrease in
dephosphorisation. The role of basicity and FeO content of slag on dephosphorization has been evaluated in the
present work. This work was carried out experimentally by using a 100 kg Air Induction Furnace. The result of the
present work shows that with the increase in basicity, dephosphorisation increases up to ‘basicity 3’ after which the
dephosphorisation decreases. At basicity levels higher than 3, a part of CaO remains un dissolved in the slag at
steelmaking temperature. It was also found that with increase in FeO content of slag, percentage of
dephosphorisation increases up to ‘FeO-25%’ after which the percentage dephosphorisation decreases. Though
‘FeO’ contributes to the oxidation of ‘P’ in the metal, it should not be increased above a value of about 25%. This
is because after an FeO level 25%, its increase happens at the expense of CaO and thus slag basicity is reduced.
Thus an optimum balance between FeO and CaO is necessary for proper dephosphorisation. This has been
established in the present work.

1. Introduction
Removal of phosphorus in steelmaking operations has been subject of extensive research since the inception of early
steelmaking technologies. Phosphorus removal is a key area of research because of its detrimental effect on
mechanical properties of steel such as cold shortness, temper embrittlement, poor ductility and strength.
Phosphorous in the steel mainly comes from iron ore, coke and recycled BOF Slag. Due to depletion of good quality
iron ore and increasing demand for low phosphorous steel, effective phosphorous control is an essential requirement.
Therefore numerous researches have been carried out for the removal of phosphorous from the steel since several
decades. Steel containing phosphorous has some adverse as well as good effects. Phosphorous has great influence
towards the solid solution strengthening of ferrite a modest quantity of phosphorous can increase the yield and
ultimate strength of mild steel. It also increases the deep drawability and hardness of steel. Referable to the above
advantages such steels are used for cold forming operation. Phosphorous in steel can also increase corrosion
resistance and metal cutting characteristic. The brittle behavior of steel containing phosphorous, at a temperature
less than recrystallization temperature is called cold shortness. It is a condition of wrought iron, steel, or other metal,
in which the metal, on account of its brittleness, cannot be worked, when cold, without fracture or cracking at the
edges. Phosphorous decreases ductility and hence increase the tendency of the steel to produce crack during cold
working. When Phosphorous, Manganese or Silicon is added to steel, the tensile strength of steel increases and
elongation decreases. A balance between cold formability and hardness needs to be maintained to produce high
Study the effect of basicity and FeO contents of slag on dephosphorisation of steel
strength formable steel. As Phosphorous level increases, ductility of steel decreases severely due to the
embrittlement [1]. The effects of phosphorus (P) on the properties of steels are summarized in Table-1[2]. It can be
seen that P has both positive and negative effects on the steel’s properties.
Table-1: Effects of phosphorus on properties of steels

Sl.No. Property Effect of phosphorus

1 Strength Strong positive (strengthens ferrite)
2 Bake hardenability Positive
3 Ductility Strong negative
4 Galvanneal Can improve resistance to powdering
5 Phosphatability Positive
6 Core loss in motor lamination Strong negative
7 Fracture toughness Strong negative

In metallurgical research publications of several decades, the terminology “capacity” has been used in lieu of the
thermodynamic term the “equilibrium constant” in formulating experimental equilibrium data on slag–gas and
slag–metal reactions, e.g. sulphide capacity, phosphate capacity, nitride capacity. Because of limited existing data
on the activities of reactants in slags, mass concentrations are used in the reaction equilibrium constant, which of
course varies with the slag composition. Similarly, sulphide capacity, phosphate capacity of slag also varies with
the slag composition. In more than 60 years of experimental research into dephosphorization, a range of chemical
reactions have been proposed to describe the transfer of phosphorus between metal and slag. Early researches such
as Turkdogan [3] used a range of molecular representations to describe dephosphorization.

2[P] + 5[O] → (P 2 O 5 ) …..(1)

∆ G° = −740375 + 535.365T J/mol ----(1a)
2[P] + 5[O] + 4(CaO) → (4 CaO. P2 O 5 ) …..(2)
2[P] + 5(FeO) → (P 2 O 5 ) 5[Fe ] …..(3)
where [X] represents species dissolved in the metal phase and (Y) represents species dissolved in the slag phase.
Even at the time they were proposed, researchers expressed doubts about the validity of such chemical expressions.
Consequently, by the 1970’s it was common to assume phosphorus existed in molten slags as an ionic species. This
resulted in different expressions for the phosphorus reaction[4]:
5 3
[P] + [O] + (O2-) → (PO 4 3-) …..(4)
2 2
5 3 5
[P] + (FeO) + (O2-) → (PO 4 3-) + Fe …..(5)
2 2 2
5 3 1 5
[P] + (FeO) + (CaO) → (Ca 3 (PO 4 ) 2 ) + [Fe] …..(6)
2 2 2 2
Despite it being generally believed that dephosphorization occurs through an ionic dissociation into the slag, for
convenience, many researchers still use reaction (1) to perform analysis. Regardless of the equation used, there are
three common understandings implied by the above equations:

125
 Proceedings of NCAMMM - 2018

 A highly basic slag is required to accept phosphorus from the metal into the slag.
 A high oxygen potential must exist to force phosphorus from the metal to the slag.
 Relatively low temperature promotes the thermodynamics of phosphorus removal.
Dephosphorisation is a reversible reaction, at higher temperature backward reaction is favorable which may lead to
rephosphorisation. During dephosphorisation of steel, aim is to reduce “P” in metal, make P 2 O 5 rich slag and also
fix the P 2 O 5 in the slag in combined form so that phosphorous reversible does not take place. CaO is added to the
flux in the form of lime. This lime (CaO), fixes the P 2 O 5 by decreasing the activity of P 2 O 5 by forming a stable
compound like 3CaOP 2 O 5 or 4CaOP 2 O 5 . It is very important to maintain FeO content of slag as as it helps in
converting the P present in the metal to P 2 O 5 . However, beyond an optimum value, FeO replaces CaO in the slag
which results in decrease in dephosphorisation. Activity of CaO should be high which increases the availability of
free dissolved lime at higher basicity slags[5]. The effectiveness of each component of slag on degree of
dephosphorization is needed to be evaluated. In the present work, aim is to evaluate the role or influence of oxides
like CaO, FeO, and SiO 2 of slag on degree of dephosphorisation of steel. Dephosphorisation studies were carried
out in 100 kg induction furnace to find out the effect of individual component of the slag.

2. Experimental
For studies on dephosphorization in steels, experiments were carried out in 100 kg Air induction furnace using 40 kg
low carbon steel. Lining of the induction furnace was done using high MgO ramming mass refractory. Laboratory
grade chemicals like calcium oxide, aluminium oxide and magnesium oxide were used along with iron ore. Amount
of flux added was varied along with slag basicity during several experimental heats. Low carbon steel was used as a
base charge for melting and the initial metal composition is shown in the Table-1. Chemical analysis of slag
components were carried out using XRF and chemical analysis of metal was carried out using OES.
Table-1: Composition of steel used in experiments
C Si Mn P S Al
0.017 0.27 0.89 0.020 0.02 0.03

3. Results and Discussion

In the present work, aim is to evaluate the role or influence of oxides like CaO, FeO, and SiO 2 of slag on degree of
dephosphorisation of steel. This has been carried out by experimentally using Air Induction Furnace.

Experimental Results from Air Induction furnace - Experimental heats numbering 17 comprising slag metal
reactions were carried out using Air Induction Furnace to find out the percentage dephosphorisation of different slag
composition to investigate the role basicity and FeO. During each heat, the samples of slag and liquid metal were
collected and analyzed. The results for metal samples and slag sample are shown in Table-2 and Table-3
respectively. Table-2 shows the chemistry of steel samples that were analyzed after dephosphorisation of steel. It is
observed from this table that the maximum percentage of dephosphorisation achieved is 60%, and minimum
percentage of dephosphorisation achieved is 10%. Similarly Table-3 shows the analysis of slag samples of different
composition which were added during dephosphorisation of steel. It has been observed from this table that the

126
Study the effect of basicity and FeO contents of slag on dephosphorisation of steel
maximum weight percent of P 2 O 5 in the slag achieved is 0.55, and minimum weight percent of P 2 O 5 in the slag
achieved is 0.099.
Table-2: Final metal composition after dephosphorisation and percentage dephosphorisation

C Si Mn P S Al % Dephosphorisation
Heat 1 0.13 0.007 0.35 0.016 0.01 0.113 20
Heat 2 0.12 0.017 0.27 0.02 0.016 0.035 10
Heat 3 0.11 0.0008 0.069 0.008 0.012 0.017 60
Heat 4 0.032 0.0009 0.1 0.009 0.012 0.043 55
Heat 5 0.021 0.02 0.2 0.016 0.02 0.178 20
Heat 6 0.056 0.001 0.183 0.012 0.016 0.192 40
Heat 7 0.018 0.004 0.1 0.009 0.011 0.748 55
Heat 8 0.037 0.003 0.32 0.018 0.013 0.58 10
Heat 9 0.047 0.005 0.2 0.013 0.01 0.43 35
Heat 10 0.024 0.01 0.21 0.016 0.016 1.32 20
Heat 11 0.051 0.001 0.14 0.01 0.012 0.19 50
Heat 12 0.046 0.001 0.1 0.01 0.01 0.24 50
Heat 13 0.054 0.005 0.12 0.011 0.01 0.04 45
Heat 14 0.041 0.001 0.22 0.015 0.011 0.24 25
Heat 15 0.12 0.007 0.23 0.013 0.01 0.058 35
Heat 16 0.04 0.004 0.18 0.01 0.09 0.14 50
Heat 17 0.043 0.024 0.016 0.011 0.14 0.153 45

Table-3: Final slag composition after dephosphorisation

MgO Al 2 O 3 SiO 2 P2O5 CaO Fe T FeO

Heat 1 16.34 5.63 24.21 0.099 26.03 9.05 11.67
Heat 2 14.38 7.67 21.97 0.26 26.35 14.96 19.29
Heat 3 16.55 3.71 20.17 0.55 27.35 21.12 27.24
Heat 4 10.85 2.68 20.17 0.47 28.58 23.09 29.79
Heat 5 21.16 1.79 19.87 0.39 26.02 18.76 24.2
Heat 6 9.45 2.76 20.04 0.48 31.71 16.11 20.78
Heat 7 11.49 4.54 20.62 0.31 15.85 18.55 23.92
Heat 8 15.13 4.31 22.97 0.14 20.97 12.11 15.62
Heat 9 12.33 3.48 24.12 0.33 27.68 11.58 14.93
Heat 10 14.97 3.98 21.46 0.24 24.29 15.53 20.03
Heat 11 10.39 2.5 25.69 0.36 28.51 13.41 17.29
Heat 12 11.22 2.65 25.08 0.3 25.61 14.61 18.84
Heat 13 13.06 3.44 25.42 0.28 24.36 13.28 17.13
Heat 14 16.41 3.28 24.53 0.2 24.83 12.62 16.27
Heat 15 13.3 2.73 27.01 0.2 28.72 9.86 12.71
Heat 16 9.86 3.67 22.53 0.35 30.58 13.52 17.44
Heat 17 20.3 4.13 26.74 0.058 26.12 8.73 11.26

127
 Proceedings of NCAMMM - 2018

Effect of Basicity of Slag on dephosphorisation - Figure-1 shows the variation in the percentage of
dephosphorisation for different slag basicity with FeO-20%, MgO-5%, and Al 2 O 3 -2%. This Figure indicates that
with increase in basicity, percentage dephosphorisation increases up to ‘basicity 3’ after which percentage
dephosphorisation decreases. Basic slags are essential for the achievement of efficient dephosphorisation because
they allow phosphate ions to be readily accepted into the slag. The dissolution rate of CaO increases with increse in
FeO concentration and decreases with basicity of the melt. Any undissolved CaO will not be effective for
dephosphorisation. Higher dissolution of the CaO in slag is imperative for good dephosphorisation. At basicity
higher than 3, a portion of CaO added does not go into the solution. This undissolved CaO increases the melting
point and viscosity of the slag. Higher viscosity of slag hinders effective slag metal reaction. As the melting point of
slag increases, dissolution of CaO is difficult at steelmaking temperature. Thus undissolved CaO does not take part
in the dephosphorisation reaction. Thus the mass transfer of phosphorus from metal phase to slag phase gets
diminished resulting in lesser dephosphorisation at basicity higher than 3.
FeO-20%, Al2O3-2%, MgO-5%
60
%Dephosphorisation

55
50
45
40
2 2.5 2.75 3 3.5
Basicity
Figure-1: Percentage dephosphorisation for different slag basicity with FeO-20%, MgO-5%, Al 2 O 3 -2%

Effect of FeO content of Slag on dephosphorisation - Figure-2 shows the variation in the percentage of
dephosphorisation for different FeO content of a slag having Basicity-3, MgO-5%, and Al 2 O 3 -2%. This Figure
indicates that with increase in FeO content of slag, percentage dephosphorisation increases up to ‘FeO-25%, after
which the percentage dephosphorisation decreases. It is seen that dephosphorisation decreases at 30% FeO. Though
‘FeO’ contributes to the oxidation of ‘P’ from the metal thereby increasing in the deoxidation process, it should not
be increased above about 25% for the above mentioned slag since its percentage increases at the expense of CaO
which reduces slag basicity. In the reaction
2[P] + 5[O] → (P 2 O 5 )
CaO reduces the activity of P 2 O 5 in the slag phase thereby moving the reaction in the forward direction. Thus an
optimum balance between FeO and CaO is necessary for proper dephosphorisation. Similar Figures have been
plotted for different percentages of MgO in the slag. These are shown in Figure-3 and in Figure-4. These have also
been plotted for basicity 2 of slag as shown in Figure-5 and Figure-6. Figure-3 shows the variation in the percentage
of dephosphorisation for different FeO content of slag with Basicity-3, MgO-10%, and Al 2 O 3 -2%. This Figure also
indicates that with increase in FeO content of slag percentage dephosphorisation increases up to ‘FeO-25%’ after
that percentage dephosphorisation decreases. All these Figures show a similar trend. Figure-4 shows the variation in
the percentage of dephosphorisation for different FeO content of slag of Basicity-3, MgO-15%, and Al 2 O 3 -2%. This

128
Study the effect of basicity and FeO contents of slag on dephosphorisation of steel
Figure indicates that with increase in FeO content of slag, percentage dephosphorisation increases up to ‘FeO-20%’
after which percentage dephosphorisation decreases.

Basicity-3, Al2O3-2%, MgO-5%

80

%Dephosphorisation
60
40
20
0
15 20 25 30
%FeO

Figure-2: Percentage dephosphorisation for different FeO content of slag with Basicity-3, MgO-5%, Al 2 O 3 -2%

Basicity-3, Al2O3-2%, MgO-10% Basicity-3, Al2O3-2%, MgO-15%

%Dephosphorisation

60 %Dephosphorisation 40
50
30
40
30 20
20
10
10
0 0
15 20 25 30 15 20 25
%FeO %FeO
Figure-3: Percentage dephosphorisation for different Figure-4: Percentage dephosphorisation for different FeO
FeO content of slag with Basicity-3, MgO-10%, content of slag with Basicity-3, MgO-15%, Al 2 O 3 -2%
Al 2 O 3 -2%

Basicity-2, Al2O3-2%, MgO-5% Basicity-2, Al2O3-2%, MgO-10%

%Dephosphorisation

80 52
%Dephosphorisation

60 50
48
40
46
20 44
0 42
15 17.5 20 25 15 18 20 25
%FeO %FeO
Figure-5: Percentage dephosphorisation for different Figure-6: Percentage dephosphorisation for different
FeO content of slag with Basicity-2, MgO-5%, Al 2 O 3 - FeO content of slag with Basicity-2, MgO-10%, Al 2 O 3 -
2% 2%

It may thus be concluded that from Figures 2, 3, and 4 that maximum dephosphorisation decreases from about 60%
to 40% with increase in MgO content of the slag from 5% to 15%. Morever, with 15% MgO in the slag, maximum
dephosphorisation achieved at 20% FeO, rather then at 25% FeO. These graphs were plotted for slag basicity of 3.

129
 Proceedings of NCAMMM - 2018

Figure-5 shows the variation in the percentage of dephosphorisation for different FeO content of slag of Basicity-2,
MgO-5%, and Al 2 O 3 -2%. This Figure indicates that with increase in FeO content of slag, percentage
dephosphorisation increases up to ‘FeO-25%’. Figure-5.6 shows the variation in the percentage of
dephosphorisation for different FeO content of slag of Basicity-3, MgO-5%, and Al 2 O 3 -2%. This Figure indicates
that with increase in FeO content of slag, percentage dephosphorisation increases up to ‘FeO-25%’ and after that
percentage dephosphorisation decreases.

4. Conclusions
Dephosphorisation study has been carried to find out the influence of oxides like CaO, FeO, and SiO 2 of slag on
degree of dephosphorisation of steel. This has been carried out experimentally using Air Induction Furnace.
• With increase in basicity, dephosphorisation increases up to ‘basicity 3’ after which the dephosphorisation
decreases. This happens because, at basicity values higher than 3, a portion of CaO added does not go into the
solution which increases the melting point and viscosity of the slag. As the melting point of slag increases,
dissolution of CaO is difficult at steelmaking temperature thus undissolved CaO does not take part in the
dephosphorisation reaction
• With increase in FeO content of slag, percentage dephosphorisation increases up to ‘FeO-25%’ after which the
percentage dephosphorisation decreases. Though ‘FeO’ contributes to the oxidation of ‘P’ in the metal, it should
not be increased above a certain value of about 25%. Since its percentage increases at the expense of CaO,
which reduces slag basicity. Thus an optimum balance between FeO and CaO is necessary for proper
dephosphorisation.

References
[1] Swinnerton M. The influence of slag evolution on BOF dephosphorisation.
[2] Basu S. Studies on dephosphorisation during steelmaking(Doctoral dissertation, KTH).
[3] ET T. Slag composition variations causing variations in steel dephosphorisation and desulphurisation in
oxygen steelmaking. ISIJ international. 2000 Sep 15;40(9):827-32
[4] Inoue R, Suito H. Mechanism of dephosphorization with CaO–SiO2–FetO slags containing mesoscopic scale
2CaO· SiO2 particles. ISIJ international. 2006;46(2):188-94.
[5] Chen GJ, He SP. Effect of MgO content in slag on dephosphorisation in converter steelmaking. Ironmaking &
Steelmaking. 2015 Jul 1;42(6):433-8

130
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

The Effect of Al-5Ti-1B Master Alloy Addition on the Microstructure, Hardness and
Mechanical Properties of Hypoeutectic Al-7.6Si Alloy

Sourabh Guptaa,b, Prosanta Biswasa, M.K. Mondala*, Rahul Bhandaric, S. Pramanika

a
Department of Metallurgical and Materials Engineering, National Institute of Technology,
Durgapur-713209, West Bengal, India.
b
Durgapur Steel Plant (Steel Authority of India Limited), Durgapur-713203, West Bengal, India.
c
Luthfaa Polytechnic Institute, Durgapur-713212, West Bengal, India.
E-mail: manas.nitdgp@gmail.com

Abstract: In this investigation, hypoeutectic Al-7.6Si alloy with (2.0 wt.%, 4.0 wt.% and 6.0 wt.% ) and without Al-
5Ti-1B master alloy addition has been developed through gravity casting method. The effects of the Al-5Ti-1B grain
refiner on the microstructural morphology, hardness, mechanical properties and fracture behaviour of the Al-7.6Si
alloy have been studied. The cast unmodified hypoeutectic Al-7.6 Si alloy consisting of needle and rod-like eutectic
Si particles with very sharp corners inside the α-Al phase. The addition of Al-5Ti-1B master alloy in the Al-7.6 Si
alloy, TiB 2 is formed and these TiB 2 are acting as a potential site for nucleation of α-Al grains. Due to this, the
grain refined alloys have globular α-Al grains and a fibrous eutectic Si phase. The addition of 4 wt.% Al-5Ti-1B in
the alloy, the average diameter of α-Al grains decrease to 42.5 µm from 14.19 µm (as cast Al-7.6Si alloy) and
roundness of α-Al grains increases to 0.696 from 0.33 (as cast Al-7.6Si alloy). The bulk hardness, ultimate tensile
strength (UTS) and elongation (%) of the modified alloy are increased. Further, factographs showed that the
cleavage fracture is reduced in the modified alloy and fine dimple formation is increased.
Keywords: Al-7.6Si alloy; Al-5Ti-1B grain refiner; microstructure evolution; hardness; mechanical properties;
Fracture Behaviour

1. Introduction
Nowadays, it is gradually becoming very important for automotive industries to increase energy efficiency to reduce
fuel consumption and pollution. Environmental concerns, government-mandated fuel efficiency standards, raw-
material constraints and global competition are driving the automotive industry to reduce fuel consumption while
maintaining safety, performance, and minimal impact on costs [1]. Switching to light-weight materials, without
sacrificing safety and performance, is being looked at as the most cost-effective way to address these challenges. As
a result, more emphasis is being laid on the increase in production and use of aluminium silicon cast alloys. The Al-
Si alloys are widely used because of their excellent properties such as high fluidity, high specific strength, good
wear resistance, good corrosion resistance, low thermal expansion, high recyclability and low cost of manufacturing
[1,2]. Some of the typical applications include cylinder heads, aircraft stabilizer, and crankcase for small engines,
cellular phone castings and domestic food components [3].
The microstructure of the hypoeutectic Al-Si alloys has coarse-columnar primary α-Al phase and needle-
like or plate-like eutectic silicon phase [4]. The Hypoeutectic Al-Si alloy contains a large volume fraction of primary
α-Al phase [5]. Therefore, the shape, size and distribution of α-Al in the microstructure directly influence the
The effect of Al-5Ti-1B master alloy addition on the microstructure, hardness and mechanical properties of hypoeutectic Al-7.6Si alloy

mechanical properties [6] of the alloy. Thus, the grain refinement of the Al-Si alloy is needed to achieve the fine-
grained structures with superior mechanical properties. Previously, few researchers have investigated the effects of
varying ratios Ti and B on the Al-Si alloy [4,5,7-10]. Recently, the effects of Al-Ti-B-RE master alloy (RE: rare
earth) addition on grain refinement was investigated in the Al-Si alloy. The plate-like or needle-like morphology of
eutectic silicon acts as sites of internal stresses development, which provide an easy path for cracks to propagate
resulting in fracture [4]. Because of it, the strength and ductility of Al-Si alloy are not up to mark [9]. Therefore, the
modification of eutectic silicon is carried out to convert the plate-like or needle-like morphology to the fine fibrous
structure with superior mechanical properties. Sodium (Na)[12], strontium (Sr) [13] and antimony (Sb) [14] are used
for this purpose. However, antimony is generally avoided as it leads to the formation of toxic stibine gas (SbH 3 ) in
the foundry [15]. Modification of eutectic silicon has also been investigated using high-intensity ultrasonic vibration
[16]. Previously, some works have been carried out on grain refinement of hypoeutectic (Si<12wt%) Al-Si alloys. In
an earlier work, Kori et al. [8], investigated the effect of Al-5Ti-1B (up to 0.1wt%Ti) master alloy on Al-7Si alloy. It
was reported that the effect of the Al-5Ti-1B master alloy was poor with respect to boron based master alloy due to
the poisoning effect of Si. The poisoning effect of Si is attributed to its affinity with Ti where it forms titanium
silicide, which coats the nucleating agent in the master alloy rendering them ineffective [17]. However, study the
effect of Al-5Ti-1B addition on microstructure, hardness, mechanical properties and fracture behaviour is not
presently available in existing standard literature. Kori et al. [5] studied the poisoning effect of Al-5Ti-1B master
alloy addition on Al-7Si alloy. It was concluded that the poisoning effect is less at an optimum addition of
0.024wt.% Ti and it was substantiated by measuring the dendritic arm spacing values. Still, there is no study on
mechanical properties. Lee et al. [7] investigated the effect of Al-5Ti-1B master alloy addition on Al-8Si alloy with
titanium additions up to 0.15wt%. It was observed that there is no significant effect of Ti on grain refinement
beyond 0.05wt% Ti in the alloy.
In this research, hypoeutectic Al-7.6 Si alloy has been successfully developed through gravity casting route
with (2 wt.%, 4 wt.% and 6wt.%) and without Al-5Ti-1B master alloy. The main objective of the study is to
investigate the effect of higher concentration Al-5Ti-1B master alloy addition on mechanical properties of the
hypoeutectic Al-7.6Si alloy. Finally, a structure-property correlation is established to understand the various aspects
of properties enhancement.

2. Experiment
The Hypoeutectic Al-7.6Si alloy was developed through melting and gravity casting routes. Initially, small pieces of
commercially pure Al (99.7%) were placed inside a clay graphite SiC crucible and melted using an electrical
resistance furnace at 7600C. After melting of commercially pure Al, small pieces of pure silicon (99.16%) was
inserted into the melt with the help of graphite rod and wait for 45 minutes for complete Si dissolution. After
homogenization, the Al-5Ti-1B master alloy was added to the molten mixture. Then wait for 20 minutes and then
the molten mixture was degassed with hexachloroethane (0.1 wt.%). After that, slag was removed from the top
portion of the melt and it was immediately poured into a cast iron mould, designed as per BS1490 standard [18, 19].
After solidification, samples of 10 mm × 10 mm were machined out from the as-cast billets for optical microscopy
and hardness test. The metallography samples were polished through the standard procedure of aluminium alloy and
etched with Keller’s reagent (1% HF, 1.5% HCl, 2.5% HNO 3 and distilled water). The optical microscopy was

132
 Proceedings of NCAMMM - 2018

carried out by Leica optical microscope. Hardness test was carried out on the Vicker’s hardness testing machine.
PANalytical X’PERT PRO machine using Cu-K α radiation was used for XRD analysis. Further, tensile test
specimens were machined out from the cast billet as per ASTM standard. Tensile tests were carried out on Tinius
Olsen Universal Testing Machine under ambient conditions. Fractured surfaces of the modified and unmodified
alloy were investigated under FESEM (Carl Zeiss, Sigma, UK) to understand the mode of fracture. The volume
percentage (serologically equal area percentage) of different phases and size of the primary α-Al were measured
using the ImageJ image analysis software.

3. Results and Discussion

3.1 Microstructural evaluation
The optical microstructure of the Al-7.6Si alloy with and without Al-5Ti-1B master alloy addition is shown in
Figure 1(a)-(d). The unmodified hypoeutectic Al-7.6Si alloy (Figure 1(a)) exhibits the presence of gray colour
combined dendritic plate-like and needle-like particles with very sharp corners inside the α-Al phase. This gray
colour phase was identified as eutectic Si phase by the EDS spot analysis (Fig. 2). The eutectic Si is evenly
distributed within the primary α-Al phase and the α-Al phase is exhibit like a matrix.

Figure 1: Optical microstructure of (a) Al-7.6Si alloy (b) Al-7.6Si alloy with 2 wt. % Al-5Ti-1B (c) Al-7.6Si alloy
with 4 wt. % Al-5Ti-1B and (d) Al-7.6Si alloy with 6 wt. % Al-5Ti-1B

The addition of Al-5Ti-1B master alloy (2 wt.%, 4 wt.% and 6wt.%) effectively modify and refine both the α-Al
and eutectic Si phase (Fig. 1(b)-1(d)). The alloy with 2 wt.% Al-5Ti-1B master alloy has some globular α-Al grains
with dendritic morphology and a fibrous eutectic Si phase with large spacing (Fig. 1(b)). Further, the Al-7.6Si-4
wt.% Al-5Ti-1B alloy has more amount of globular α-Al grains and a fine fibrous eutectic Si morphology with least
spacing compared to the 2 wt.% Al-5Ti-1B added alloy (Fig. 1(c)) The average size of the α-Al grains are decreased
and roundness (degree of sphericity (DOS)) of the α-Al grain increases, when the concentration of the Al-5Ti-1B

133
The effect of Al-5Ti-1B master alloy addition on the microstructure, hardness and mechanical properties of hypoeutectic Al-7.6Si alloy

master alloy is increased from 2 wt.% to 4 wt.% (Fig. 3). The average α-Al grain size and DOS were measured by
ImageJ image analysis software and calculated according to the equation (1) and (2) [21-22].

Grain Diameter (GD) = 2 Ag π (1)

Roundness (R) = (4πA g Pg2 ) (2)

Where A g and P g are the area and perimeter of the α-Al grains.

Figure 2: Identification of eutectic Si by SEM based EDS spot analysis in the Al-7.6Si alloy.

But, the alloy with 6 wt.% Al-5Ti-1B master alloy has more dendritic α-Al grains with less roundness and the
spacing of fibrous eutectic Si is more as compared to the Al-7.6Si-4 wet.% Al-5Ti-1B alloy (Fig. 1(d)). The average
size of the α-Al grains is also comparatively large and roundness is less (Fig. 3). This may cause of over
modification of the alloy.

Figure 3: The roundness and average diameter of α-Al grains variation of Al-7.6 Si alloy with different additions of
Al-5Ti-1B grain refiner.

Furthermore, the X-ray diiffratoragram of the Al-7.6Si alloy (Fig. 4(a)) and the Al-7.6Si-6 wt.%Al-5Ti-1B alloy
(Fig. 4(b)) shows the presence of α-Al and Si peaks, while the 6 wt.% Al-5Ti-1B master alloy added alloy has the

134
 Proceedings of NCAMMM - 2018

peaks of titanium boride (TiB 2 ) apart from the α-Al and Si peaks. The TiB 2 processes HCP structure with a lattice
parameter a TiB2 = b TiB2 = 3.034 Å c TiB2 = 3.226 Å [23] and the crystal structure of aluminium is FCC with a lattice
parameter a Al = 4.0497 Å [24]. The FCC aluminium matrix has (111) closed packed plane matching with (0001)
closed packed plane of the HCP TiB 2 . The closed packed direction on (0001) plane is [2110] and the interatomic

distance is a TiB2 . Further, interatomic distance in close-packed [110] direction on (111) plane is a Al 2 and misfit

parameter or disregistry is 0.056. Therefore, the TiB 2 will act as potential nucleation sites of α-Al grains during
solidification. The modification and the refinement of both the phases occur due to the formation of the TiB 2 in the
grain refined alloy.

Figure 4: The XRD pattern of (a) Al-7.6Si alloy and (b) Al-7.6%Si-6 wt.% Al-5Ti-1B alloy.

3.2 Hardness
The bulk hardness of the alloys and composites generally depends on the morphology and the volume fraction of
different phases of the alloys [25].

Figure 5: Bulk hardness variation of Al-7.6 Si alloy with different additions of Al-5Ti-1B grain refiner.

Figure 5 shows the bulk hardness of the devolved alloys as a function of Al-5Ti-1B concentration. The bulk
hardness of unmodified alloy is approximately 46 VHN, but the modified alloys have a bulk hardness around
65VHN to 74VHN. This increase in hardness is caused by the microstructural morphology changed, such as
globular α-Al grains and fibrous eutectic Si phase formation in the modified alloy. The Al-7.6Si-4.wt% Al-5Ti-1B

135
The effect of Al-5Ti-1B master alloy addition on the microstructure, hardness and mechanical properties of hypoeutectic Al-7.6Si alloy

alloy has high hardness compare to other modified alloy because it has more globular fine α-Al grains and fibrous
eutectic Si phase with low spacing.
3.3 Mechanical properties
Figure 6 shows the UTS and percentage elongation of the Al-7.6Si alloy as a function of Al5Ti-1B grain refiner
concentration. The un-modified Al-7.6Si alloy has UTS value of 108MPa and % elongation value of 6.60 %. The
grain refinement of Al-7.6Si alloy by the addition of 2.0 wt.% Al-5Ti-1B grain refiner resulted in 33.3% and 47.5%
improvement of UTS and % elongation values. The 4 wt.% Al-5Ti-1B refined alloy has relatively more improved
UTS (58.3%) and elongation (101.5%). But, further increase in the Al-5Ti-1B refiner concentration to 6 wt.% the
UTS and elongation values are slightly decreased compared to previous composition but, much higher than the
unrefined alloy. This improvement of the UTS and ductility attribute to microstructural refinements such as fine
globular α-Al grains and fibrous eutectic Si phase with low spacing (as discussed in section 3.1)

Fig. 6: UTS and % elongation of Al-7.6 Si alloy as a function of Al-5Ti-1B grain refiner concentration

3.3 Fracture Surface

Figure 7 exhibits the fracture surface of the developed alloys. In the present study, various structural defects (micro-
porosity and microcrack) and sharp corners of eutectic Si are the suitable sites for failure initiation (Figure 7(a)-(d)).
Then the crack propagates through the eutectic Si particles by means of cleavage fracture of the eutectic Si particles
(Figure 7(a)-(d)). However, a combined mode of brittle and ductile fracture with dimple formation has been found in
the fracture surface of the developed alloys. The factrograph of the unmodified alloy has mainly cleavage facets of
coarser and dendritic eutectic Si particles (Figure 7(a)). Whereas, the grain refined alloys have a very less number of
cleavage fracture of eutectic Si as the plate-like and needle-like eutectic Si transformed into fibrous morphology.
The decohered particles are observed in the grain refined alloy and the dimple formation is increased (Figure 7(c)-
(d)). The fracture surface factrograph has a great agreement with the microstructural morphology and mechanical
properties of the developed alloys.

136
 Proceedings of NCAMMM - 2018

Figure 7: Secondary electron images of the fractured surface (a) Al-7.6Si (b) Al-7.6Si-2 wt. % Al-5Ti-1B (c) Al-
7.6Si-4 wt. % Al-5Ti-1B and (d) Al-7.6Si-6 wt. % Al-5Ti-1B alloy

4. Conclusion
Effect of the Al-5Ti-1B grain refiner addition on the microstructure, mechanical properties and fracture behavior of
the hypoeutectic Al-7.6Si alloy has been studied and following conclusion are drawn:
• The cast unmodified hypoeutectic Al-7.6 Si alloy consisting of a needle and rod-like eutectic Si particles with
very sharp corners inside the α-Al phase and the α-Al phase is present as like a matrix phase.
• The grain refined alloys have globular α-Al grains and a fibrous eutectic Si phase.
• The bulk hardness, ultimate tensile strength (UTS) and elongation (%) of the modified alloy are increased as
compared to the unmodified alloy.
• Addition of 4 wt.% of Al-5Ti-1B grain refiner to the Al-7.6Si alloy gives the smallest grain size and highest
roundness of α-Al grains compare to the 2 wt.% and 6 wt.% Al-5Ti-1B added alloy. As a result, the Al-7.6Si
alloy with 4 wt.% Al-5Ti-1B grain refiner has the highest strength (171 ± 2), ductility (13.3 ± 0.4) and hardness
(73.8 ± 0.5)
• The cleavage fracture and brittle fracture are reduced in the modified alloy and fine dimple formation is
increased.

Acknowledgement
The authors are like to thank NIT, Durgapur RIG # 2 project for financial support and the Director of National
Institute of Technology Durgapur, India for his continuous encouragement.

137
The effect of Al-5Ti-1B master alloy addition on the microstructure, hardness and mechanical properties of hypoeutectic Al-7.6Si alloy

References
[1] Kaufman JG, Rooy EL. Aluminum alloy castings: properties, processes, and applications. ASM International;
2004.
[2] Davis JR. Aluminium and aluminium alloys, Alloying: understanding the basics. ASM international; 1993:
351-416.
[3] Principles of Purchasing Castings, CD-ROM, American Foundry Society (2002).
[4] Basavakumar KG, Mukunda PG, Chakraborty M. Influence of grain refinement and modification on
microstructure and mechanical properties of Al–7Si and Al–7Si–2.5 Cu cast alloys. Materials
Characterization. 2008 Mar 31;59(3):283-9.
[5] Kori SA, Auradi V, Murty BS, Chakraborty M. Poisoning and fading mechanism of grain refinement in Al-
7Si alloy. InMaterials Forum 2005 (Vol. 29, pp. 387-393).
[6] Olofsson J, Svensson IL, Lava P, Debruyne D. Characterisation and investigation of local variations in
mechanical behaviour in cast aluminium using gradient solidification, Digital Image Correlation and finite
element simulation. Materials & Design. 2014 Apr 30;56:755-62.
[7] Lee YC, Dahle AK, StJohn DH, Hutt JE. The effect of grain refinement and silicon content on grain
formation in hypoeutectic Al-Si alloys. Materials Science and Engineering: A. 1999 Jan 15;259(1):43-52.
[8] Kori SA, Murty BS, Chakraborty M. Development of an efficient grain refiner for Al–7Si alloy. Materials
Science and Engineering: A. 2000 Mar 15;280(1):58-61.
[9] Auradi V, Kori SA. Effect of processing temperature on the microstructure of Al-7Ti master alloy and on
refinement of α-Al dendrites in Al-7Si alloys. Adv. Mater. Lett. 2015;6:252-9.
[10] Rao AP, Das K, Murty BS, Chakraborty M. Microstructural features of as-cast A356 alloy inoculated with Sr,
Sb modifiers and Al–Ti–C grain refiner simultaneously. Materials Letters. 2008 Jan 31;62(2):273-5.
[11] Wang XJ, Cong XU, Muhammad A, Hanada S, Yamagata H, Wang WH. Effects of Al-Ti-B-RE grain refiner
on microstructure and mechanical properties of Al-7.0 Si-0.55 Mg alloy. Transactions of Nonferrous Metals
Society of China. 2014 Jul 1;24(7):2244-50.
[12] Al-Haidary JT, Hanon MM, Abdulsahib YM. Effect of sodium chloride on microstructure and mechanical
properties of Al-Si alloy. Eng. & Tech. Journal. 2015; 33(A): 2187-2197.
[13] Jigajinni SM, Kori SA, Venkateswarlu K. Development of New Combination of Grain Refiner-cum-Modifier
for LM-21 Alloy. Transactions of 61st Indian Foundry Congress. 2013; 1-7.
[14] Chen JX, Lu MJ, Wu SS, Lü SL. Study on Eutectic Microstructure and Modification Mechanism of Al-Si
Alloys. Materials Science Forum 2017 (Vol. 877, pp. 97-103). Trans Tech Publications.
[15] Gruzleski JE, Closset BM. The treatment of liquid aluminium-silicon alloys. Amer Foundry Society; 1990.
[16] Jian X, Meek T, Han Q. Refinement of eutectic silicon phase of aluminium A356 alloy using high-intensity
ultrasonic vibration. Scripta Materialia. 2006 Mar 31;54(5):893-6.
[17] Qiu D, Taylor JA, Zhang MX, Kelly PM. A mechanism for the poisoning effect of silicon on the grain
refinement of Al-Si alloys. Acta Materialia. 2007 Feb 28;55(4):1447-56..
[18] British standard. Specification for Aluminium and aluminium alloy ingots and castings for general
engineering purposes. British standard institution; 1988; BS1490.
[19] Mondal MK, Biswas K, Maity J. Microstructural characterisation of novel 6351 Al–Al4SiC4 in-situ
composite. Materials Science and Technology. 2013 Nov 1;29(11):1394-402.
[20] Mohanty PS, Gruzleski JE. Grain refinement mechanisms of hypoeutectic Al-Si alloys. Acta materialia. 1996
Sep 1;44(9):3749-60.
[21] Taghavi F, Ghassemi A. Study on the effects of the length and angle of inclined plate on the thixotropic
microstructure of A356 aluminium alloy. Materials & Design. 2009 May 31;30(5):1762-7.
[22] Fan Z. Semisolid metal processing. International materials reviews. 2002 Apr 1;47(2):49-85.
[23] Yan HY, Qun W, Chang SM, Ping GU. A first-principle calculation of structural, mechanical and electronic
properties of titanium borides. Transactions of Nonferrous Metals Society of China. 2011 Jul 1;21(7):1627-
33.
[24] Stock SR, Cullity BD. Elements of X-ray Diffraction. Massachusetts, Addison-wesley publishing company
inc. 1978; 2nd edn: 506.
[25] Biswas P, Mondal MK, Roy H, Mandal D. Microstructural evolution and hardness property of in situ Al–
Mg2Si composites using one-step gravity casting method. Canadian Metallurgical Quarterly. 2017 Jul 3;56
(3):340-8.

138
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Influence of Microwave Heat Treatment on Natural Iolite

Shubhashree Swain*, Siddhartha Kumar Pradhan

CSIR-Institute of Minerals and Materials Technology, Bhubaneswar - 751013, Odisha, India
*
Corresponding author, E-mail: shubhashreeswain@gmail.com

Abstract: Microwave (MW) heat treatment of natural iolite gemstone of Odisha, India has been carried out. The
iolite gemstones were heated up to a temperature of 800°C. The gemstones did not show any noticeable
enhancement in color or luster. This makes microwave heat treatment of iolite ineffective in enhancing its aesthetic
values as compared to the previous results published for ruby where color and luster enhancement takes place. The
experimental results have also been verified by Raman spectroscopy analysis and optical absorption spectra of the
iolite samples. The peak intensities in the Raman spectra of the heat treated samples are lower as compared to those
of the untreated samples. Furthermore, there is a decrease in the overall intensity of the optical absorption spectra
of the heat treated iolite samples in the range of 200-800 nm as compared to that of the untreated ones. Hence, the
results demonstrate the ineffectiveness of MW heating in enhancing the aesthetic value of natural iolite gemstones.
Keywords: Natural iolite; Microwave heat treatment; Raman spectroscopy; Optical absorption

1. Introduction
Gemstones are kind of minerals, which when cut and polished properly have very high aesthetic value and can be
used for making ornaments of various kinds. Various kinds of gemstones are extensively used for making jewellery
owing to their luster, color, hardness, etc. Iolite is one of the varieties of gems of the mineral Cordierite. It is a very
well known gemstone which represents the translucent variety of Cordierite. The color of iolite usually varies from
light to deep blue with a purplish tinge. The deep colored stones have more aesthetic value. Iolite is comparatively
cheaper and is usually used as a substitute for more expensive blue gemstones like Sapphire.
Cordierite (Mg,Fe) 2 Al 4 S i5 O 18 is an aluminosilicate primarily popular due to its properties like excellent thermal
shock resistance [1,2], low dielectric constant [2,3], low coefficient of thermal expansion [1,2], high melting point
[1,2], adequate mechanical properties [2,4] and chemical resistance [2,5]. These kind of properties enable the usage
of ceramics comprising of cordierite for various applications like packing materials for electronic packing, refractory
materials, catalytic converter substrates for changing pollutant gases into less harmful ones of automobile engines,
etc [2,6].
Recently, the aesthetic value improvement of Ruby gemstones by microwave (MW) heat treatment was
reported by our group [7]. Hence, the effect of MW heat treatment has also been studied on natural iolite gemstones
which posses non-uniform color saturation in this work. Heating gemstones to enhance the overall aesthetic value is
an ancient process. MW heating has certain advantages which give it an edge over conventional heating. Some of
these benefits incorporate high-speed unvarying heating, energy-saving, good quality microstructures, fast product
growth, and eco-friendly method. During the heating process, the material absorbs electromagnetic energy from the
microwave radiations and converts it to heat [7].
Influence of microwave heat treatment on natural iolite
2. Experimental Method
The unprocessed iolite gemstones were accumulated from Odisha, India. The stones varied in size between 4 mm
and 6 mm. Initially, the raw samples were cleaned with isopropanol to remove any dirt and grease. In the beginning,
to eradicate any dirt and grease, the untreated stones were cleaned with isopropanol. The samples were heated in a
MW furnace (GN Technologies) by keeping in between two SiC pieces which absorbed the MW energy. We
observed an indirect heating at the speed of 15-20 °C/min up to a temperature of 800 °C. Both MW heat treated and
natural gemstones were characterized using UV-Visible spectroscopy (JASCO, V-650, Japan) for optical absorption
and Raman spectroscopy (Seki, STR-500, Japan) to identify the compound phases of impurities.

3. Results and Discussions

The silicate cordierite symbolized by the formula (Mg,Fe) 2 Al 4 Si 5 O 18 possesses a tetrahedral structure. The low-
cordierite prevalent in the space group Cccm can be represented by a crystal chemical formula by [8]:
(M) 2 (T 2 3) 2 (T 2 1) 2 (T 2 6) 2 (T 1 6)(T 1 1) 2 O 18 , (Ch0,Ch¼) (1)
2+ 2+ 2+
where octahedrally coordinated Mg , Fe or Mn ions are symbolized by M, tetrahedral positions are given by T
and cha nnel sites are shown by Ch. Six-membered rings are formed by the T 2 tetrahedra, compared to this the T 1
tetrahedra cross-links these units to formulate a framework structure. The T 2 3, T 2 1 and T 1 6 sites are occupied by Si
and the rest of the two T 1 1 and T 2 6 sites are occupied by Al. Endless channels are formed by the stacked six-
membered rings parallel to the c-axis. Certain excess ions like Na+ and K+, balanced charge deficiencies, or volatiles
like H 2 O and CO 2 can be placed in Ch0 and Ch¼ channel sites [8, 9].

3.1. Microwave heating

The heating behavior of the unprocessed iolite gemstone is portrayed in figure 1.

Figure 1: Temperature profile of iolite during MW heating

The temperature profile fluctuated before rising steeply to reach up to 745°C in 27 minutes. From then on it
increases slowly up to about 800 °C in another 17 minutes. So, it takes less than one hour of time to reach about
800°C. In our previous work related to ruby gemstone [7], the temperature rose to 1500°C in 50 minutes. There, two
distinct temperature rise regimes were demonstrated, where indirect fast heating of the sample by heat absorption
from the SiC susceptors appears to take place up to a temperature of about 800°C. However beyond 800 °C, a kind
of hybrid heating took place due to heat absorbed from the SiC susceptors and the volumetric MW heating of the

140
 Proceedings of NCAMMM - 2018

ruby (alumina) itself. SiC is frequently used as a susceptor material on account of its dielectric behaviour which
makes it an efficient microwave absorber. It is particularly well suited for heating of low-loss ceramics at low
temperatures [10,11]. The temperature profile of iolite gemstone as evident from figure 1 indicates that it is the
indirect heating regime. Since the melting point of iolite is near about 1435°C, it should not be heated to
temperatures near to this. Hence, heating iolite using MW did not help much in the present study.

3.2 Raman spectroscopy analysis

Figure 2: Raman spectra of untreated and heat-treated iolite gemstones

Raman spectra of the untreated and microwave heat-treated iolite gemstones are shown in figure 2. The spectra of
the untreated samples possesses sharp peaks at 255 cm-1, 552 cm-1, 668 cm-1, 970 cm-1, 1010 cm-1 and 1182 cm-1.
Two of the most strong peaks are present at 970 cm-1 and 1010 cm-1 which are primarily due to the stretching
vibration of tetrahedral T 2 1, and T 2 3 sites of SiO 4 tetrahedral [9] mainly resembling the orthorhombic cordierite
[12]. A strong peak was also observed at 668 cm-1, prime reason of which is the two connected M sites arising from
the stretching vibration of T 1 1 Al tetrahedra. Although, the sharp peak at 255 cm-1 is the after effect of bending
vibration of octahedrally coordinated M site. Adding further, the peak at 552 cm-1 is primarily due to the combined
effect of the stretching vibration of M and T 2 6 sites along with the bending vibration of Si bearing T 1 6 and Al
bearing T 2 6 tetrahedra [9]. Similarly, the bending vibration of T 2 1, T 2 3, and T 2 6 sites along with the stretching
vibration of M site is represented by the peak at 576 cm-1. The changes in the symmetry from high-temperature
hexagonal structure to a low-temperature orthorhombic structure is the reason of splitting of the strong band at 565
cm-1 into two peaks, at ~552 cm-1 and ~576 cm-1 [9]. Similar results have also been demonstrated by Majumdar et
al. [13]. But the Raman peak intensities have decreased after the heat treatment, although the peak positions are
same as the untreated spectra.

3.2. Optical absorption studies

Figure 3 depicts the optical absorption pattern of the untreated and microwave heat treated iolite in the UV-visible
range recorded at room temperature. A strong absorption peak has been observed at 270 nm along with a small
hump at 212 nm in the ultraviolet region. The F+ and F centers occurring because of the capture of a single or a
couple of electrons by an oxygen vacancy are primarily responsible for the appearance of these peaks [14].

141
Influence of microwave heat treatment on natural iolite

Figure 3: Optical absorption spectra of untreated and heat-treated iolite gemstones.

The presence of a considerable amount of Fe3+ in the crystal structure is portrayed by the 570 nm peak which in turn
gives rise to blue to violet dichroism in received cordierite samples. The dichroism observed in cordierite is
attributed to the charge transfer between the Fe2+ and Fe3+ ions [15,16]. However, the iolite gemstone demonstrates
an overall decrease in the intensity of the absorption when heat-treated at 800 °C. Hence, such an observation
indicates that MW heat treatment at 800 °C does not enhance the color of iolite as is evident from the experimental
results.

4. Conclusion
MW heating, which is a unique process for rapid thermal excitation has been used to heat iolite gemstone. As
compared to the previous work reported for ruby where visible enhancement in color and luster was achieved by
heat treatment, no such change is obtained in the physical appearance of iolite when heat-treated at 800 °C. The
temperature profile of the iolite gemstones during MW heating indicates the stones get heated primarily in the
indirect heating regime. The Raman spectra portray an overall decrease in the peak intensities of the heat treated
iolite gemstones as compared to the untreated ones. The splitting of the Raman band at 565 cm-1 shows the presence
of a perfect ordering of Al-Si distribution within tetrahedral sites. The optical spectra of the gemstones demonstrate
the presence of considerable amount of Fe3+ in the stones which in turn leads to the charge transfer mechanism
between the Fe2+ and Fe3+ ions giving rise to blue to violet dichroism in cordierite samples collected from Odisha,
India.
Acknowledgments
This work was financially supported by Council of Scientific and Industrial Research (CSIR), New Delhi, India
through Project ESC-206.

References
[1] Yalamac E, Akkurt S. Additive and intensive grinding effects on the synthesis of cordierite. Ceram. Int. 2006;
32: 825-832.
[2] Rundans M, Sperberga I, Lindina L, Steins I. Effect of sintering process and additives on the properties of
cordierite based ceramics. 5th Baltic Conference on Silicate Materials, IOP Conf. Series: Materials Science
and Engineering. 2011; 25: 012009.
[3] Camerucci MA, Urretavizcaya G, Castro MS, Cavalieri AL. Electrical properties and thermal expansion of
cordierite and cordierite-mullite materials. J. Eur. Ceram. Soc. 2001; 21: 2917-2923.

142
 Proceedings of NCAMMM - 2018

[4] Acimovic Z, Pavlovic L, Trumbulovic L, Andric L, Stamatovic M. Synthesis and characterization of the
cordierite ceramics from nonstandard raw materials for application in foundry. Mater. Lett. 2003; 57: 2651
2656.
[5] Gonzalez-Velesco JR, Ferret R, Lopez-Fonseca R, Gutierrez-Ortiz MA. Influence of particle size distribution
of precursor oxides on the synthesis of cordierite by solid state reaction. Powder Technol. 2005; 153: 34-42.
[6] Oliveira FAC, Fernandes JC. Mechanical and thermal behavior of cordierite-zirconia composites. Ceram. Int.
2002; 28: 79-91.
[7] S. Swain, S. K. Pradhan et al. Microwave heat treatment of natural ruby and its characterization. Appl. Phys.
A 2016; 122: 224.
[8] Cohen JP, Ross FK, Gibbs GV. An X-ray and neutron diffraction study of hydrous low cordierite. American
Mineralogist. 1977; 62: 67–78.
[9] Kaindl R, Többens DM, Haefeker U. Quantum-mechanical calculations of the Raman spectra of Mg- and Fe-
cordierite, American Mineralogist. 2011; 96: 1568–1574.
[10] Baeraky TA. Microwave measurements of dielectric properties of silicon carbide at high temperature. Egypt.
J. Sol. 2002; 25: 263-273.
[11] Ding D, Zhou W, Zhang B, Luo F, Zhu D. Complex permittivity and microwave absorbing properties of SiC
fiber woven fabrics. J. Mater. Sci. 2011; 46: 2709-2714.
[12] Haefeker U, Kaindl R, Tropper P. Semi-quantitative determination of the Fe/Mg ratio in synthetic cordierite
using Raman spectroscopy. Am. Mineral. 2012; 97: 1662–1669.
[13] Majumdar AS, Mathew G. Raman-Infrared (IR) Spectroscopy Study of Natural Cordierites from Kalahandi,
Odisha, Journal of geological society of India. 2015; 86: 80-92.
[14] Yang X, Li H, Cheng Y, Tang Q, Su L, Xu J. Growth of highly sensitive thermoluminescent crystal α-
Al 2 O 3 :C by the temperature gradient technique. J. Cryst. Growth. 2008; 310: 3800-3803.

[15] Poon WCK., Putnis A, Salje, E. Structural states of Mg cordierite: IV. Raman spectroscopy and local order
parameter behavior. Jour. Phys. Condens. Mat. 1990; 2: 6361– 6372.

[16] Faye GH, Manning PG, E. H. Nickel. The polarized optical absorption spectra of tourmaline, cordierite,
chloritoid and vivianite: ferrous-ferric electronic interaction as a source of pleochroism, The American
Mineralogist. 1968; 53: 1174-1201.

143
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Experimental Investigation on Tribological Behaviour of Interacting Surfaces with

Microstructral Matrix

Santosh Kumar1, Subrata Kumar Ghosh1, A. Mukhopadhyay2

1
Department of Mechanical Engineering, Indian Institute of Technology (ISM), Dhanbad
2
Department of Mechanical Engineering, Jadavpur University
Email: subratarec@yahoo.co.in

Abstract: Mechanical and tribological properties of interacting surfaces of austenitic stainless steel AISI 304L are
generally changed with different temperature range and cooling medium. The objective of this paper is to analyse
the surface behaviour according to grain growth and microstructural matrix formation. The wear tests of the heat
treated AISI 304L samples (Such as cooled in a different medium, namely in a furnace, air and in water) have been

carried out in a multi tribo tester. X ray diffraction (XRD), Scanning electron microscopy (SEM), Micro hardness,

Fischer Feritscope tests are also performed to find out the effect of microstructure as well as micro hardness on the
wear behavior. The minimum wear has been observed in case of normalized samples compared to other samples
(such as annealed, solution annealed). This may be due to the presence of percentage value of ferritic, martensitic,
carbide (CrC) and other different microstructures matrix and its micro hardness. It has also been observed by
fischer feritscope that the heat treated samples contain δ- ferrite structure (less than 1%).
Keywords: Heat treatment; Wear; Hardness; Microstructure; AISI 304L stainless steel

1. Introduction
Stainless steel is a major component in industrial, commercial, and consumer products. Typical application of AISI
304L stainless steel includes to car headers, various machinery parts, screws, valves, lining of coal hopper, utensils,
and marine applications and like others. Because, AISI 304L stainless steel exhibits excellent corrosion resistance
and formability with different temperature range [1-2]. Peng Wang et al. [3] studied on phase composition and grain
size refinement accomplished by X-ray diffraction. They concluded that the martensitic transformation of ultrafine
grain in AISI 304L vary accordingly pressure and temperature with bond free energy. Many papers were studied on
heat-treatment of AISI type 304 stainless steel at different time-temperature cycle with phase trasformation and
microstructure change[4-5]. Austenitic Stainless steel was sensitized, when exposed to elevated temperature range of
470-750°C causes carbide precipitations at grain boundaries. The few investigation revealed that the sensitized
samples gave the highest hardness value at 6660C, while highest hardness value was obtained on temperature of
10900C for solution annealed 304 stainless steels. This temperature was found to be the optimum to avoid grain
growth on solution annealed 304 stainless steels [6-7]. Wear is associated with interacting surfaces at the interface
between two or more bodies under relative motion control. The study of friction and wear behavior is thus important
in the characterization of annealed AISI 304L stainless steel along with the evaluation of other mechanical
properties [7-9]. Progress of wear finally determines the useful life and the quality of a product. Hence the nature of
friction, wear and their control plays an important role in different engineering operations. Though there are several
research based models and formulations in this regard, but majority of them are not suitable to predict the
Experimental investigation on tribological behaviour of interacting surfaces with microstructural matrix
tribological behavior in a particular case. Thus, iteration of friction and wear data through practical experimentation
in a particular situation is very much important for understanding the interacting surface behaviour [10-12]. The
present work has been carried out to compare the sliding friction and wear behavior of different heat treated AISI-
304L stainless steel against EN-8 steel in dry condition.

2. Methodology

AISI SS-304L stainless steel has been selected as test material and EN-8 stainless steel (AISI 1040) as standard
roller material. AISI SS-304L is having different peaks values in XRD pattern as shown in Figure 1. This peak
values show the crystal size of different type of microstructure in different planes. SS-304L contains single phase
FCC structure of austenitic structure. XRD pattern shows with different peak such as first peak tends to plane [miller
indices plane: 111] at 400-500, second peak tends to plane [miller indices plane: 200] at 500-600, third peak tends to
plane [miller indices plane: 220] at 700-800, fourth and fifth tend to plane [miller indices plane: 220] at 900-1000
respectively. So, pattern demonstrates that uniformity in crystal size, that is, maximum part of SS 304L surface
contain austenitic structure with similar planar lattice.

Figure 1: XRD pattern of ss304L raw sample

Electric muffle furnace was used for heat treatments of the samples. The heat treated AISI 304L samples have been
prepared in a different medium, namely in a furnace, air and water. Initially samples were heated at 660°C and
allowed to dwell for 30 minutes by using the electric muffle furnace for the preparation of Solution annealing
samples. Air cooling has been done in an open environment. Sensitized samples were reheated at temperature of
1040°C with a soaking time of 30 minutes and cooled rapidly in water maintained at room temperature.
Hardness of the heat treated samples has been measured by rockwell hardness tester (Model: MRS-2260) with a
1/16" steel ball indenter under a load of 100 Kgf. The hardness values in HRB scale have been indicated in Table 1.
Scanning electron microscopy (SEM) shows that different types of grain formation of the heat treated AISI 304L
samples (Such as cooled in a different medium, namely in a furnace, air and in water) in Figure 2. AISI 304L
stainless steel undergoes a martensitic transformation under cyclic loading conditions. There are three prominent
matrix or zone in the microstructures namely white, black and mixture of white and black. This matrix or zone are
used for measurement of micro hardness values.

145
 Proceedings of NCAMMM - 2018

Table 1: Hardness of the heat treated samples in HRB scale

Sl. No. Heat Treated Samples
Sensitized Solution Annealed Furnace Cooled Air Cooled
Sample Sample Sample Sample
I 47 57 51 56
II 49 58 53 55
Avg. 49.5 57.5 52.5 54.5

(a): normalized sample (b): annealed sample

(c): sensitize sample (d): solution annealed sample

Figure 2: Scanning electron microscopy of the heat treated AISI 304L samples

Figure 3: XRD graph of heat treated samples as compare to raw material sample
XRD graph shows with different peak for heat treated samples compare to raw material in Figure 3. It has been
observed that plane angle is changed after heat treatment for furnace cooled, air cooled and solution annealed
samples. It may be due to different type of microstructure formation, which is responsible of grain growthment.

146
Experimental investigation on tribological behaviour of interacting surfaces with microstructural matrix
3. Results and Discussion
Micro hardness values have measured by micro hardness tester ‘LM 248 AT’ (LECO, serial no. XM8 116,
Michigan) as shown in Table 2. It has been found that average microhardness of solution annealed samples have
maximum and minimum in sensitize samples. Furnace and air quenched samples have moderate microhardness.

Table 2: Microhardness at different matrix of heat treated samples

Heat Matrix Type Microstructure D 1 (μm) D 2 (μm) Avg. HV

Treatmented (H)
Sample
Sensitized White Austenite 32.48µm 33.16µm 190.54
Sample White Austenite 30.87µm 28.87µm
Black Ferrite 25.50µm 27.29µm 240.23
Black Ferrite 27.88µm 29.02µm
Solution White Martensite 30.48µm 29.16µm 605.20
Annealed White Martensite 31.87µm 32.87µm
Sample Black Carbide(CrC) 25.30µm 29.29µm 756.21
Black Carbide(CrC) 22.68µm 27.05µm
Furnace White Austenite 17.88µm 18.63µm 192.12
Quenched White Martensite 16.16µm 17.93µm 612.12
Sample Black Ferrite 26.39µm 27.43µm 242.70
Black Ferrite 27.89µm 23.14µm
Air White Austenite 28.34µm 27.35 µm 189.30
Quenched Black Ferrite 30.14 µm 32.23 µm 248.60
Sample Black Carbide(CrC) 15.55 µm 166.33 µm 749.40
Mixture Black Mixture of Martensite and 25.44 µm 99.66 µm 599.30
and white carbide

Fischer Feritscope measures the bcc structuraled ferrite percentage in the heat treated samples shown in Table 3. It
has been found that samples have approximately 1% δ-Ferrite and rest is combination of austenite, carbide (CrC)
and Martensite. The heat treated SS-304 stainless steel contain mainly austenite (approximately 95%). Multi Tribo
Tester TR-25 (DUCOM, India) has been utilized for wear test of AISI SS-304L heat treated samples (dimension 20
mm×20 mm×8 mm). The hardness of the roller material (EN-8 stainless steel) is 55 HRC. Figure 4 depicts the
comparative wear curves of four different heated samples of austenitic stainless steel AISI 304. It has been found
that sensitise and solution annealed samples have the maximum and minimum wear respectively. Sensitise sample
contains mainly austenite and ferrite, which have lower microhardness value than carbide(CrC) and martensite and
resulted in more wear. Solution annealed sample have carbide(CrC) and martensite higher microhardness than
others and having less wear.

Table 3: Percentage of bcc phase as obtained from Fischer Feritscope

Heat % Reading Value of bcc phase % Avg.(bcc

Treatmented Phase)
Sample 1 2 3 4 5 6 7 8 9 10
Sensitized 0.58 0.68 0.77 0.85 0.82 0.78 0.72 0.78 0.65 0.85 0.748

147
 Proceedings of NCAMMM - 2018

Sample
Solution 0.35 0.25 0.16 0.47 0.28 0.42 0.27 0.36 0.44 0.49 0.349
Annealed
Sample
Furnace 0.97 1.2 0.85 0.87 0.95 0.97 0.98 1.1 0.93 0.96 0.978
Quenched
Sample
Air 0.74 0.55 0.65 0.62 0.67 0.62 0.68 0.74 0.75 0.66 0.668
Quenched
Sample

250

200
wear in sensitize
150 sample
wear(µm)

wear in annealed
100 sample

50 wear in
normalized
sample
0
0 1000 2000
time(s)
Figure 4: Variation of wear with time for heat treated samples

4. Conclusions
The characterization of microstructure has been revealed by SEM micrography, XRD test and mirohardness test.
Fischer feritscopic test shows the presence of less than 1% δ- ferrite (bcc structure) in all the heat treated samples.
XRD graph demonstrates that plane angular orientation of samples is changed after heat treatment such as solution
annealed, furnace cooled, air cooled and sensitize sample due to grain growthment. It has been revealed that the
solution annealed sample is associated with lowest wear among the other types of heat treated samples. Normalised
and sensitise samples show moderate and maximum wear respectively. The trend of wear rate may be due to the
formation of different microstructure such as austenite, carbide(CrC) and Martensite (that is 99%).

References

148
Experimental investigation on tribological behaviour of interacting surfaces with microstructural matrix
[1] Covert RA, Tuthill AH. Stainless steels: An introduction to their metallurgy and corrosion resistance, Dairy.
Food and Environmental Sanitation. 2000; 20 (7): 506-517.
[2] Parthasarathi NL, Borah U, Albert K Sh, Correlation between coefficient of friction and surface roughness in
dry sliding wear of AISI 316L (N) stainless steel at elevated temperatures. Computer Modelling and New
Technologies. 2013; 17(1): 51-63.
[3] Wang Peng, Zhang Jinzan, Zhang Yang, Wang Haitao, Hua Wentao, Yu Dongli. Preparation of high-
performance ultrafine-grained AISI304L stainless stee lunder high temperature and pressure. Progress in
Natural Science, Materials International. 2016; 26: 404–410.
[4] Meng X, Zhang J, Zhao J, Han W. Influence of annealing treatment on microstructure and properties of cold
sprayed stainless steel coatings. Acta Metall. Sin. (Eng.Lett.). 2011; 24(2): 92-100.
[5] Tanwer AK. Effect of various heat treatment processes on mechanical properties of mild steel and stainless
steel. AIJRSTEM. 2014; 8(1): 57-68.
[6] Tukur SA, Dambatta MS, Ahmed A, Mu’az NM. Effect of Heat Treatment Temperature on Mechanical
Properties of the AISI 304 Stainless Steel. IJIRSET. 2014; 3(2): 9516-9520.
[7] Fontalvo GA, Mitterer C. Comparison of the tribological properties of different cold work steels at
temperatures up to 250°C. Proceedings of 6th. International Tooling Conference, Austria. 2010; 18(02): 223-
236.
[8] Grossmann MA, Bain EC, Principles of Heat Treatment. Metals Park,OH. American Society for Metals. 1964;
5: 290-296.
[9] Chowdhury MA, Nuruzzaman DM, Roy BK, Dey PK, Mostafa MG, Islam MS, Mia R Md. Experimental
investigation on friction and wear of stainless steel 304 sliding against different pin materials.World Applied
Science Journal. 2013; 22 (12): 1702-1710.
[10] Leroux P. Stainless steel analysis using tribology and 3D metrology. NANOVEA. Available
from:http://www.nanovea.com. 2010; 1-9,
[11] Ludema KC. Friction, Wear, Lubrication: A Textbook in Tribology. CRC Press, USA. 1996.
[12] Parthasarathi NL, Borah U, Albert K Sh. Correlation between coefficient of friction and surface roughness in
dry sliding wear of AISI 316L (N) stainless steel at elevated temperatures. Computer Modelling and New
Technologies. 2013; 17(1): 51-63.

149
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Effect of Boron Modified Microstructure on Impression Creep Behaviour of Simulated

Multi-Pass Heat Affected Zone of P91 Steel

Akhil Khajuria*,1, Modassir Akhtar2, Dr. Rajneesh Kumar3, Dr. Jaganathan Swaminanthan4, Dr. Raman
Bedi5, Dr. Dinesh Kumar Shukla6
*1
Department of Mechanical Engineering, NIT Jalandhar, Punjab
2
Department of Metallurgical and Materials Engineering, NIT Warangal, Telangana
3
Project Planning and Engineering Division, CSIR-National Metallurgical laboratory, Jamshedpur, Jharkand
4
Materials Engineering Division, CSIR-National Metallurgical laboratory, Jamshedpur, Jharkhand
5
Department of Mechanical Engineering, NIT Jalandhar, Punjab
6
Department of Mechanical Engineering, NIT Jalandhar, Punjab
E-mail: akhilkhajuria40@gmail.com

Abstract: Existence of microstructural heterogeneity among subzones of weld heat affected zone (HAZ) in P91 steel
after welding results in formation of metallurgical notch at outer HAZ (FGHAZ/ICHAZ) which gives invitation to
type IV cracking as weldments are put into creep exposure. In the present work effect of boron modified
microstructure on creep behaviour of Gleeble™ simulated HAZ produced due to multi-pass weld thermal cycles has
been examined by impression creep technique. Short duration impression creep tests for 20 hours were run at
625 oC and 365 MPa on simulated P91 and P91B multipass HAZs representing CG + FGHAZ and CG + ICHAZ as
well as on as received steel alloys of boron free P91 (P91) and boron added P91 (P91B). It is revealed that
presence of 100ppm boron in P91 steel reduces variation in creep deformation behaviour among P91B base metal
and P91B multi-pass HAZs whereas absence of boron in P91 increases variation among P91 base metal and P91
multi-pass HAZs rendering HAZ as a weakest link in boron free P91 weldments and therefore prone to type IV
cracking.
Keywords: P91 steel, effect of boron, Gleeble™, HAZ simulation, multi-pass welding, impression creep

1. Introduction
The demand to enhance efficiency and reduction in CO 2 emissions has given thrust to increase operating
temperatures and pressures of power generating units. P91, ferritic/martensitic steel, is one such special low carbon
alloy which has been empowering power plant industry from four decades particularly due to its better creep
strength than austenitic steels [1]. Manufacturing components of P91 steel like boiler headers, super heater tubes etc.
in thermal power plants require fusion welding, since they are generally thick sections and it is difficult to join them
by low heat input welding processes such as tungsten inert gas (TIG) welding or electron beam welding (EBW) [2].
Therefore high heat input welding processes like submerged arc welding (SAW) are employed for producing P91
weldments. But, thermal cycles of fusion welding lead to undesirable microstructural changes due to phase
transformations at heat affected zone (HAZ) of weldment degrading mechanical as well as creep properties with
respect to virgin state of P91 steel [3]. Such deprivation in creep strength leads to premature in-service failure of P91
weldments at outer HAZ by Type IV cracking [4]. Type IV cracking is intergranular cracking which occurs due to
Effect of boron modified microstructure on impression creep behaviour of simulated multi-pass heat affected zone of P91 steel

microstructural heterogeneity among sub-zones of HAZ. To overcome this issue, addition of 100ppm boron in virgin
P91 steel has been found useful for enhancing creep life of P91 welded joints, but still final creep rupture takes place
due to Type IV cracking [5]. To maintain smooth functioning of power plants, repairing pre-existing P91 joints suits
as a better option instead of substituting them with new joints due to economic reasons. But, repairing a fusion
welded joint add up thermal cycles to the existing P91- HAZ whose microstructure is again affected due to multi-
layer fusion welding [6]. For multi-phase steels like P91 (P91 contains M 23 C 6 , MX particles), formation of a
subzone in HAZ is governed by phase transformation temperatures Ac1 and Ac3, since secondary phase particles
affect martensite to austenite transformation and subsequently by maximum peak temperature reached in a particular
subzone of HAZ [7]. M 23 C 6 carbides (M = Cr, Fe, Mo) have BCC type of crystal structure and they help to stabilise
martensite laths whereas, MX (M = Nb, V, Ti & X = C, N) particles retard both grain growth and creep deformation
by pinning on grain boundaries. Therefore for single pass weld thermal cycle case, based on decline in peak
temperature from weld metal towards base metal, HAZ is formed into three distinguishable subzones usually named
after grain size and phase transformation temperature i.e. coarse grain HAZ (CGHAZ), fine grain HAZ (FGHAZ)
and inter-critical HAZ (ICHAZ) as shown in Figure 1. However, development of microstructure at HAZ due to
multiple weld thermal cycles is significantly different, since overlaying successive layers of weld bead produce
successive thermal cycles affecting previous existing subzone of HAZ [8]. It can be observed from Figure 1 that
successive layers of subzones of HAZ evolve after triggering reheating of previous CGHAZ during welding multi-
pass thermal cycles. Hence, HAZ subzones formed during multi-pass welding are named according to the degree of
heating. For example, if previous CGHAZ is reheated to a peak temperature of FGHAZ, sub HAZ formed due to
this heating is referred as CG + FGHAZ or super-critically reheated CGHAZ. Similarly, if degree of heating on
CGHAZ corresponds to a peak temperature of ICHAZ, then HAZ subzone formed is referred as CG + ICHAZ or
inter-critically reheated CGHAZ and so on.

Figure 1 Schematic representing evolution of subzones of HAZ after multiple passes of weld thermal cycles

Impression creep is a pioneering development among creep testing methods for characterizing creep behaviour of
creep resistant materials as it uses specimens of small dimensions and less test time than conventional methods of
creep testing [9]. This technique becomes more useful when volume of material to be evaluated for knowing creep
properties is inherently small like heat affected zone (HAZ) in weldments whose size falls in the range of 2 – 3 mm
even for high heat inputs of fusion welding [10]. With the availability of thermo mechanical simulator like

151
 Proceedings of NCAMMM - 2018
Gleeble™, exact replicas of thermally treated microstructures in subzones of HAZ could be produced [11]. Such
simulated samples can be further used for studying creep behaviour by impression creep technique. Creep resistance
of a P91 - HAZ subzone is dependent on its microstructure evolved after receiving a particular weld thermal cycle
which significantly alters fraction of micro-constituents of designed virgin alloy. Numerous studies have been
reported on creep behaviour of subzones of P91-HAZ representing single pass weld thermal cycle whereas creep
behaviour of subzones of boron added P91-HAZ as well as HAZ subzones of boron free and boron added P91
originating from multi-pass weld thermal cycles has not so far been investigated. Taking into account repairing of
P91 weldments by successive weld passes, present work aims to examine creep behaviour of subzones of simulated
multi – pass HAZ of P91 and P91B steels. To maintain the scope of this paper, current investigation only focus on
creep behaviour of Gleeble™ simulated subzones for CG + FGHAZ and CG + ICHAZ of both boron free and boron
added P91 steels susceptible to type IV cracking by impression creep technique.

2. Experimental Procedure
2.1 Steels
As received heats of P91 and P91B were subjected to two furnace heat treatments comprising of normalizing at
1050oC/30 min. and then tempering at 760oC/2hr. Chemical composition of P91 and P91B are shown in Table 1. It
may be noted that P91 contains 22ppm boron. This percentage of boron is ineffective to show any significant effect
on creep properties of 9Cr steels.
Table 1 Chemical composition of P91 and P91B

Element C Cr Mo Si Mn V Nb Ni Al N B Ti Fe
P91 0.11 9.44 0.71 0.21 0.3 0.2 0.05 0.223 - 0.06 0.0022 - Bal.
P91B 0.10 8.26 0.88 0.3 0.33 0.186 0.06 0.01 0.03 0.004 0.01 0.0041 Bal.

2.2 Gleeble™ simulation of multi-pass HAZ

The prerequisites for multi-pass HAZ simulation are heat input (KJ/mm), heating rate – HR (oC/sec), peak
temperature -T p (oC), hold time at T p (sec), inter-pass temperature, and cooling rate (oC/sec) or t 8/5 (sec). Heat input,
heating rate, cooling rate and inter-pass temperature were found from thermal data acquisition system (DAQ)
employed during submerged arc welding (SAW) of 11 mm thick plates of P91. SAW for joining of 11 mm butt joint
of P91 corresponded to a heat input of 2.5 KJ/mm. After analysis of DAQ data, t 8/5 = 30 seconds was selected which
represents a cooling rate of 10 °C/s during exponential cooling from 800 oC to 500 oC with linear heating at heating
rate of 100 °C/s. In case of transformable steels, peak temperatures for subzones of HAZ simulations are kept
according to phase transformation temperatures i.e. Ac1 and Ac3. For determining Ac1 and Ac3; dilatometric
experiments on standard dumbbell shaped specimens as per ASTM-1033 were carried out on P91 and P91B steels
on continuous cooling transformation unit of thermo-mechanical simulator- Gleeble™ 3800. Thereafter, peak
temperatures for multipass HAZ simulations of P91 and P91B were decided to carry out thermal simulations on
Gleeble™. All thermal simulations were conducted on standard square HAZ simulation samples with dimensions 78
mm × 11 mm × 11 mm with copper grips and a span of 30 mm. Two thermocouples of K type were spot welded for

152
Effect of boron modified microstructure on impression creep behaviour of simulated multi-pass heat affected zone of P91 steel

monitoring programmed temperature and obtaining real time simulation temperature data with respect to time. The
exponential cooling from peak temperature was programmed as per equation (1).

T = T P × 𝑒𝑒 −0.47𝑡𝑡/∆𝑡𝑡 …………..(1)

where T = instantaneous temperature (oC), t = time (sec) and ∆𝑡𝑡 = time to cool from 800oC to 500oC (sec)

Summary of dilatometric experiments on Gleeble™ is presented in Table 2. Thermal profile data of simulated HAZ
as obtained from QuickSIM inbuilt software with Gleeble™ was plotted on Originlab plotting tool. Plots of thermal
profiles between peak temperatures versus time are shown in Figure 2.

Table 2 CCT and multi-pass HAZ simulations of P91 and P91B steels
CCT HR (oC/s) T p (oC) Hold at T p (sec) CR (oC/s) Ac1 and Ac3 (oC)
P91 100 1100 20 10 885 and 965
P91B 100 1100 20 10 933 and 1040

Figure 2 Thermal profiles of simulated HAZs of P91 and P91B steels

It can be noted from simulated thermal profiles of P91B in Figure 2(b) that higher peak temperatures have been
taken for simulation of second cycle of FGHAZ and ICHAZ, since transformation temperatures Ac1 and Ac3 of
P91B steel are higher than P91 as revealed by dilatometry experiments. Whereas simulation of first cycle of
CGHAZ was carried out at same peak temperature (1240 oC) for both steels because maximum temperature
measured by DAQ in this subzone of HAZ is exceedingly larger than Ac3 of P91 and P91B.

2.3 Specimen preparation for microstructural characterization and impression creep tests
Figure 3 shows a typical simulated P91 – multi – pass HAZ specimen on Gleeble™ system. Simulated specimens
were cut from thermocouple spot weld location by wire EDM exposing two surfaces for microstructural
investigation and creep tests. After taking out hot region from simulated specimens, a post weld heat treatment
(PWHT) for 760 oC/3hrs was given in a muffle furnace. The exposed surfaces of thermocouple spot weld location
were prepared for optical microscopy. Mechanical polishing of these surfaces by using emery papers was done
(220Grit>320Grit>400Grit>600Grit>800Grit>1000Grit) for removing deep scratches. This was followed by cloth

153
 Proceedings of NCAMMM - 2018
polishing with the use of alumina and water for removing the fine scratches from the sample surface to achieve
mirror like appearance. These prepared surfaces were etched with 28% Nital (HNO 3 + C 2 H 5 OH). Same method
was followed for sample preparation of as received steels. Heyn’s linear intercept method was used to measure prior
austenite grain (PAG) size. For impression creep tests, samples of dimensions 10 mm × 10 mm × 10 mm were used.
Creep tests were performed on Spranktronics impression creep testing machine at 625 oC and a punching stress of
365 MPa corresponding to uniaxial creep stress of 110 MPa in order to satisfy a correlation factor of 3.3 as specified
in literature [12]. An indenter of nimonic alloy with 2 mm diameter was used for this purpose.

Figure 3 Gleeble™ simulated P91 - HAZ sample

3. Results and discussions

3.1 Microstructures of as received steels and simulated HAZs
Both microstructures of P91 and P91B steels as observed in Figure 4 show tempered martensitic structures with only
difference in PAG size. P91 has a smaller PAG size of 18 µm than 28 µm in P91B. Existence of a bigger PAG size
is attributed to the effect of boron which decreases the grain boundary energy and allows grain growth in P91.

Figure 4 Microstructures of as received steels (a) P91 (b) P91B

Also, prior austenite grain boundaries (PAGBs) in P91B appear to avoid cellular grain structure in P91B.
Martensitic laths show healthy presence and are oriented in identical directions inside PAGBs of P91B in contrast to
P91 where owing to small PAGs, relatively weak lath martensitic structure is observed. Figure 5 show
photomicrographs of simulated multi-pass HAZs of P91 and P91B steels. It is noted from Figure 5 (a), that a
microstructure comprising of big grains with neighbouring small grains is obtained in P91-CG + FGHAZ and an

154
Effect of boron modified microstructure on impression creep behaviour of simulated multi-pass heat affected zone of P91 steel

average PAG size of 31 µm is measured. The first thermal cycle during this simulation went to a peak temperature
of 1240 oC which corresponds to CGHAZ resulting in coarsening of grains and subsequently second thermal cycle
of FGHAZ slightly influences the microstructure leading to formation of small grains. FGHAZ thermal cycle
induces heterogeneity and leaves excessive precipitation across the microstructure. Similar observations are noted
for P91B – CG + FGHAZ too other than PAG size which is of the same order of 28 µm as in as received state of
P91B steel. Figure 5(b) show photomicrograph of P91 - CG + ICHAZ which depicts that partial transformation of
martensite is dominant. Since peak temperature during second thermal cycle in this simulation corresponds to
ICHAZ, martensite doesn’t get sufficient energy to fully transform to austenite and therefore final microstructure
after PWHT cycle is only tempered with the presence of mixed grains of partially transformed martensite and
tempered martensite. However, this is not the case in P91B – CG + ICHAZ. A fully martensitic microstructure with
a PAG size of 28 µm is observed. This is attributed to the presence of boron which stabilizes the microstructure even
after weld thermal cycle and PWHT.

Figure 5 Microstructures of simulated P91and P91B - HAZs (a) P91- CG + FGHAZ – After PWHT (b) P91-
CG + ICHAZ –After PWHT (c) P91B - CG + FGHAZ –After PWHT (d) P91B - CG + ICHAZ –After PWHT

3.2 Impression creep behaviour of as received steels and simulated HAZs

155
 Proceedings of NCAMMM - 2018
The impression creep curves between depth of indentation (DOI) and time for 20 hours of creep exposure at 365
MPa and 625 oC presenting comparison of creep behaviour among as received steels and their simulated multi-pass
HAZs are shown in Figure 6.

Figure 6 Impression creep plots at 365 MPa and 625 oC (a) P91 and multi – pass HAZs of P91 (b) P91B and
multi – pass HAZs of P91B

It is observed from Figure 6(a) that base metal of P91 show least resistance to punching stress and reaches to
maximum DOI within the same duration of creep test which is attributed to least PAG size of P91 steel. P91 – CG +
FGHAZ happens to most creep resistant in P91 group which produces a large gap in maximum DOI achieved
among it and base metal of P91 thus separating out variation in creep behaviour. On the other hand, relatively bigger
PAG size of P91B than P91 show small increment in creep resistance as observed in Figure 6(b). Also, maximum
DOI reached in same duration of impression creep falls in a close range for simulated multi – pass HAZs of P91B
and base metal of P91B indicating accomplishment of nearly homogenous creep behaviour which is attributed to
role of boron in microstructural stabilization after HAZ simulation and PWHT.

4. Conclusions
1. No effect on PAG size is observed in simulated multi-pass HAZ of P91B – CG + FGHAZ and P91B – CG +
ICHAZ with respect to virgin P91B steel inferring stabilization of HAZ microstructure after multi –pass welding
and PWHT.
2. Presence of boron in P91 not only enhances creep resistance of this steel, but also produces nearly homogenous
creep behaviour among base metal of P91B and subzones of P91B – HAZ.
3. Existence of mixed grains of partially transformed martensite and tempered martensite weakens the
microstructure and therefore poor creep resistance is shown by P91 – CG + ICHAZ as in contrast to P91B – CG +
ICHAZ where fully tempered martensitic structure is observed which comparatively better creep resistance to same
punching stress of 365 MPa at 625 oC.
4. The impression creep technique is found sensitive to base metals of boron free and boron added P91 steel as well
as different simulated multi-pass HAZ specimens of P91 and P91B. This infers that effect of heat treatments on
creep resistant steels and their weld HAZs could be quickly characterized for creep behaviour using short duration
impression creep tests.

156
Effect of boron modified microstructure on impression creep behaviour of simulated multi-pass heat affected zone of P91 steel

References
[1] Klueh RL. Ferritic/martensitic steels for advanced nuclear reactors. Transactions of the Indian Institute of
Metals. 2009 Apr 1; 62(2):81-7.
[2] Goswami P. P (T) 91 steel—a review of current code and fabrication practices. In Proceedings of the Sixth
International Conference on Advances in Materials Technology for Fossil Power Plants, La Fonda, USA
2010.
[3] Wang Y, Kannan R, Li L. Characterization of as-welded microstructure of heat-affected zone in modified
9Cr–1Mo–V–Nb steel weldment. Materials Characterization. 2016 Aug 31;118:225-34.
[4] Wang Y, Li L. Microstructure evolution of fine-grained heat affected zone in Type IV failure of P91 welds.
Weld. J. 2016 Jan 1;95:27.
[5] Baral J, Swaminathan J, Chakrabarti D, Ghosh RN. Creep behavior of P91B steel in the presence of a weld
joint. Materials Science and Engineering: A. 2015 Apr 17;631:220-9.
[6] Łomozik M, Tasak E. Physical simulation and numerical modelling of X10CrMoVNb 9.10 (P91) steel repair
welding. Materials for Advanced Power Engineering 2006.
[7] Porter DA, Easterling KE, Sherif M. Phase Transformations in Metals and Alloys, (Revised Reprint). CRC
press; 2009 Feb 10.
[8] Wang XL, Tsai YT, Yang JR, Wang ZQ, Li XC, Shang CJ, Misra RD. Effect of interpass temperature on the
microstructure and mechanical properties of multi-pass weld metal in a 550-MPa-grade offshore engineering
steel. Welding in the World. 2017 Nov 1;61(6):1155-68.
[9] Yang F, Li JC. Impression test—A review. Materials Science and Engineering: R: Reports. 2013 Aug
31;74(8):233-53.
[10] Vijayanand NV, Ganesan V, Laha K, Mathew MD. Evaluation of creep deformation behaviour of different
microstuctural zones of 316LN SS weld joint using impression creep testing technique. Materials Science and
Technology. 2014 Aug 1;30(10):1223-8.
[11] Akhtar M. Metallurgical characterisation of simulated heat affected zone in boron modified

P91 steel. 2017 June (Masters dissertation, NIT, Warangal).

[12] Yu HY, Li JC. Computer simulation of impression creep by the finite element method. Journal of Materials
Science. 1977 Nov 1;12(11):2214-22.

157
 THEME

Advanced
Manufacturing
 Casting and Powder
Metallurgy
 Machining
 Forming, Welding and Additive
Manufacturing
Sub - theme

Casting and
Powder Metallurgy
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Counter Gravity Casting- Potentials and Challenges -A Review

Vineet Chak1,a, Himadri Chattopadhyay2,b, Md. Mahfooz Alam1

1
Department of Forge Technology, NIFFT, Ranchi
2
Dept. of Mechanical Engineering, Jadavpur University, Kolkata
a
vineetchak07@gmail.com, bchimadri@gmail.com

Abstract: Materials such as aluminium, magnesium and their alloys are gaining much importance due to their Light
weight, high strength to weight ratio and other unique characteristics. Many researchers have carried out their
studies on the processes of shaping these materials and among all the methods of shaping, casting have emerged as
the most promising. But casting of these materials using conventional methodologies has never been easy and
productive, as in these methods casting defects occurs like gas defects due to melt oxidation and defects due to
shrinkage and pouring. Defects in the final product cannot be compromised as it results into number of problems
such low strength, lack of good surface finish etc. which results into high number of rejections. Casting defects such
as porosity, hot tears, initiation of corrosion etc. which leads to the low mechanical properties of the cast products
and all these defects are initiated or caused during the pouring of metal into the mold. Counter gravity casting
process has emerged as the substitute for casting of such materials, as in this process molten metal is drawn against
gravity into the mold, for a more efficient fill of the mold cavity. The process was patented in 1972 by Hitchiner
Manufacturing and different variants of the process had evolved over the years. Components processed using this
method is quite reliable and can achieve thin walls as well as dimensional accuracy. This process also fulfills the
recent objectives of casting products with most intricate details with thin walls which is really very much difficult to
achieve with conventional casting methods.
Keywords: counter gravity casting (CGC), vacuum, pressure control system and mold filling

1. Introduction
In Counter-gravity casting, vacuum is used to create the difference in pressure between the mold and the surface of
molten metal, due to which molten metal is sucked inside the metal as the pressure on the surface of the molten
metal is higher than the mold cavity and metal flow occurs against the gravity i.e, counter gravity. This approach of
mold filling in this process offers it various advantages over traditional process related to quality, economics, defects
elimination and achieving net-shape cast products. Such advantageous properties of this process has lead to the
growing importance of the technology, especially by power and automobile industries [1] the unique characteristic
of metal flow against gravity for mold filling and also the solidification at desired pressure. Let us consider the
pressure on the surface of molten metal in the crucible as P 2 and the pressure inside the mold cavity as P m, then due
to the difference i.e, (P 2 -Pm) metal flows upward(counter gravity). [2]
There are two methods by which metal can be forced to flow against the gravity as shown in fig. 1(a),
vacuum is utilized in creating the pressure difference between molten metal container and the mold cavity whereas
in fig. 1(b), external pressure is applied on the molten metal and is forced to flow in the mold cavity through the
vertical tube.
Counter gravity casting- potentials and challenges -A review

Fig 1 (a): Counter Gravity mold filling using vacuum Fig 1(b): Counter Gravity filling mold using
externally applied pressure

2. Developments in Counter Gravity Casting

According to the study made by G.D. Chandley [3] the process of filling mold by the use of vacuum was started in
the early 1970s. The metal used was steel and now also used for casting heat resistant super alloys, titanium alloys
are being cast routinely using ceramic crucibles. Further he stated that the mold design enables a greater material
yield and reduction in the energy requirements. The control over the rate of vacuum greatly helps in complete
filling, permitting casting of very high quality in viscous metals, such as metal matrix composites of aluminum. [3-
4]. WanqiJiea et al. [5] in his study, for producing the large sized thin walled cast components of high quality
aluminium, fabricated a casting equipment utilizing the counter gravity technique with resin sand molds. In this the
molten material is drawn upward against the gravity into the mold at desired rate by compressed air, and is allowed
to solidify at larger pressure [4-5].The comparison of casting with counter gravity filling of mold and traditional
casting with gravity pouring is shown in the table below [6-8].

Table 1: Comparison table between counter gravity and traditional casting

Counter gravity casting Traditional casting
Vacuum or external pressure is used for creating the Metal is poured into the mold with the help of gravity.
pressure difference for flow of the molten metal against
the gravity.
Pure molten metal (free from impurities) goes inside the Molten metal with impurities goes inside the cavity.
mold cavity.
Metal efficient process as less material solidifies in The process requires more metal and hence costly.
gating, higher metal yield.
Due to less or no turbulence, gating system can be Gating design is complicated to counter the turbulence
simplified. effect.

3. Research Scope in Counter Gravity Casting

For a comprehensive idea on CGC [4-8] may be referred .There are relatively fewer studies in those area. Figure 2
shows important domains in which further research can be done for counter gravity casting.

159
 Proceedings of NCAMMM - 2018

Fig. 2 Block diagram showing scope of research in CGC

4. Applications
Industries utilizing the benefits of properties of lighter materials such as Aluminium, nickel and magnesium alloys
are the major beneficiaries of counter gravity casted parts, engine parts of Aerospace can be cast, automotive parts
such as wheel, camshaft housing & cover, cylinder head cover. Some other areas of counter gravity cast parts
include sports goods like bicycle frames. The properties of these lighter materials are also enjoyed by Ordnance in
casting parts for missile transporter, ground rocket launcher [9-12].

5. Summary
A comprehensive study concerning the main issues of counter gravity casting technique involving various aspects
like the principle of Counter Gravity Casting, its variations, process parameters, and the process performance
measures in terms of quality of the casting produced. Mechanics of pouring molten metal into the mold has been
illustrated. The technique of filling the mold against gravity has been compared with gravity pouring and the major
advantaged are illustrated. The extent of geometrical abilities of this process is presented on the basis of past studies.
The following interpretations may be made on the basis of the present study.
 Counter gravity casting has been derived from the conventional casting process with difference of drawing the
molten metal into the mold cavity. The metal is drawn into the cavity due to the pressure difference created by
using the vacuum.
 Excellent quality, light weight, and intricate detail components of medium and small size with thin walled
sections from wide variety of alloys can be cast with this process. Counter gravity casting process uses cleaner
melt which avoids post solidification defects. The design of vacuum system plays an important role in creating
the pressure difference to get the molten metal free from impurities inside the mold cavity.
 Vacuum assisted controlled filling of mold enables the process to reduce the occurrence of casting defects
caused due to oxidation and turbulence of molten metal and helps in achieving defect free cast products.

Despite of having several advantages over gravity pouring technique, this process yet required to be explored and
further research in the areas of computational control of mold filling, mathematical modeling and simulation of melt
filling, optimum mold and filling tube design and control of vacuum during the process are required to be done to
utilize the potentials of this process in engineering applications.

160
 Counter gravity casting- potentials and challenges -A review

References
[1] Aremo B, Adeoye MO. Aluminium Countergravity Casting–Potentials and Challenges. In Recent Trends in
Processing and Degradation of Aluminium Alloys 2011. InTech.
[2] Jianmin Z, Yaohe Z, Ping G, WukuI G. A new counter-gravity casting process, mold filling dynamics,
convective heat transfer and influence on mechanical properties: Asian pacific conference for materials and
mechanics at Yokohama, Japan during November 13-16, 2009
[3] Chandley GD. Use of vacuum for counter-gravity casting of metals. Materials Research Innovations. 1999 Jun
1;3(1):14-23.
[4] Sata A, Sutaria M. Scope of Investment Castings Supported by Survey of Foundries in Rajkot Cluster. Indian
Foundry Journal. 2014 Jun;60(6).
[5] Jie WQ, Li XL, Hao QT. Counter-gravity casting equipment and technologies for thin-walled Al-alloy parts in
resin sand molds. InMaterials Science Forum 2009 (Vol. 618, pp. 585-589). Trans Tech Publications.
[6] Shendye S, King B, McQuay P. Mechanical Properties of Counter-Gravity Cast IN718. TMS. 2005 Jun
6;124:133.
[7] Hebsur MG. Processing of IN-718 Lattice Block Castings.
[8] Salonitis K, Zeng B, Mehrabi HA, Jolly M. The challenges for energy efficient casting processes. Procedia
CIRP. 2016 Jan 1;40:24-9.
[9] Fan ZT, Ji S. Low pressure lost foam process for casting magnesium alloys. Materials science and technology.
2005 Jun 1;21(6):727-34.
[10] Campbell J. Stop pouring, start casting. International Journal of Metalcasting. 2012 Jul 1;6(3):7-18.
[11] Chowdhury AK. Advanced Manufacturing Technologies: Proceedings of International Conference on
Advanced Manufacturing Technologies at CMERI, Durgapur During 29-30th November 2007. Allied
Publishers; 2007.
[12] Handbook AS. Volume 15 Casting. Materials Park: ASM International. 2008:416-522.

161
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Tribological Properties of Aluminium Graphite Composite Journal Bearing

S. Ansary, R. Shaikh, M. Sekh, R. Haque, Md.Kamaruzzaman*, S. Haidar*

Department of Mechanical Engineering,AliahUniversity,Kolkata, India.
*Corresponding Author E.Mail: shamimhaidar@yahoo.com, zaman.ece@gmail.com

Abstract: In this present investigation, journal bearings were made by Powder Metallurgy technique under different
compaction pressure, sintering temperatures, sintering time and graphite percentage which ultimately produces
components with different relative densities. The properties like, wear, interfacial friction etc. have been measured
under laboratory conditions. The performances of the Journal bearings were tested under different bearing load and
different sliding velocities. It is observed that the developed aluminium graphite porous composite journal bearings
are more acceptable. It is expected that present work will create an immense importance to eliminate leaded bronze
bearings in one hand and energy saving in other hand by reducing interfacial frictional force.
Keywords: Al-Graphite composites; Powder metallurgy; Journal bearing; Wear; Friction coefficient

Nomenclature
𝑊𝑊 𝑏𝑏 = Balance weight ( Kg.)
𝑊𝑊 𝑑𝑑 = Dead weight ( Kg.)
𝐷𝐷 𝑗𝑗 = Journal diameter ( mm.)
m i = Initial mass of bearing sample before rotation.
m f = Final mass of bearing sample after rotation for t hours
t = Time of rotation.

1. Introduction
Over the last three decades, there has been a considerable interest in the use of aluminum metal matrix composites
(MMCs) which are normally fabricated by using ductile material like Aluminum as the base metal or matrix material
and reinforced by a ceramic second phase particulates e.g. graphite. Several metals and alloys have been used as
matrix materials; however, most Research and Development has been concentrated on aluminum and its alloys. It is
investigated that ferrite-pearlite ductile iron bearing slide with minimum friction against a mild steel journal and also
wear out at the lower rate at high sliding speed among three grade of ductile iron [1]. Lin et al investigated the
10%SiCp/Al-Mg composites by semi-solid mechanical stirring technique. The distribution of SiCp reinforcement in
matrix is improved by the superior wettability between matrix and reinforcement, with increasing Mg content
[2].AMCs with soft reinforcement Gr particles posses better wear characteristics as the thin layer created by Gr
particles prevents metal to metal contact between adjacent sliding surfaces [3].Composites containing the 4wt. % or
6 wt. % graphite particles exhibited lowest wear rate and friction coefficient with variation in sliding speed[4].Hsiao
and Jen (2000) investigated the pure Gr particles and coated Gr particles to form two kinds of 6061 aluminum
alloys[5].The graphitic composites included A356 Al–10% SiC–4% Gr and A356 Al–5% Al 2 O 3 –3% Gr that are
being developed for cylinder liner applications shows that the graphitic composites displayed a transition from mild
to- severe wear for all load and sliding speed combination because of the thicker and more stable tribo-layers on the
contact surfaces of graphitic composites, than that of non-graphitic composites and the A356 Al alloy[6]. Akhlaghi
and Zare-Bidaki (2009) assessed the influence of graphite content on the dry sliding and oil impregnated sliding
Tribological properties of aluminium graphite composite journal bearing

wear characteristics of sintered aluminum 2024 alloy–graphite (Al/Gr) composite materials using a pin-on disc wear
test. The composites with 5–20 wt. % flake graphite particles processed by the technology of in situ powder
metallurgy. And for both dry and oil impregnated sliding experiment it is found that with the increase in Gr content
decrease in coefficient of friction. Al/Gr composites with 5 wt. % graphite exhibited superior wear properties over
the base alloy, whereas at higher graphite addition a complete reversal wear behavior observed. Oil impregnated
Al/Gr composites with 10 wt. % or more graphite particles exhibited higher wear rate than that of the base alloy
[7].Hassan et al (2009) have reported decrease in hardness with increase in % reinforcement of Gr due to increased
porosity and the implication of these observations is that after a certain limit of the % reinforcement of Gr in Al–Gr
composites are not beneficial to add Gr as reinforcement. Hard ceramic particals of SiC may overcome this
difficulty when added as a second reinforcement with high % reinforcement of Gr in Al–Gr composites [8]. Ted et
al (2000) studied that, the hardness and coefficients of thermal expansion of the composites decreases with the
amount of graphite increases. It was found that in case of hybrid composites there is no occurrence of seizure
phenomenon which occurred with a monolithic aluminum. The friction coefficient reduces with the graphite content
increases while the amount of graphite released on the wear surface increases [9]. Jinfeng et al. (2008) investigated
the 40%SiC/5%Gr/Al composites by variation in size of graphite particle addition by squeeze casting technology,
and their friction and wear properties. Results showed that the wear resistance increased by 170 to 340 times as well
as the friction coefficient of composites decreased after the addition of graphite[10]. Basavarajappa et al. (2007)
investigated that the wear resistance of the composites increases with the addition of SiCp and graphite
reinforcements[11]. Sahoo et al. studied wear behaviour of Al-SiCp metal matrix composites and optimization using
taguchi method and grey relational analysis [12].

2. Experimental Method
Pure Aluminum fine powder mixed with graphite powder has been used in this experiment as metal powder.
Maker’s (Loba Chemie Pvt. Ltd) supplied this aluminum and graphite fine powder directly. Average particle size of
Aluminum and graphite powder was 200 mesh and 72 mesh respectively. A cylindrical mixer was used for mixing
the aluminium graphite powder and lubricant. The compacted specimens were sintered at 9000 C for almost three
hours in air tight atmosphere. Sintering was carried out in a muffle type furnace that has capacity of providing
sintering temperature of up to 13000 C. The surface of the specimens was polished with a fine emery paper. The set
up for coefficient of friction measurement of the samples were designed and fabricated. For dry Test Arrangement
The set up consist of test journal, which is connected to a shaft of the motor. The journal was press fitted and keyed
to the motor shaft. The test bearing is fitted to a M.S sleeve. A small spirit level is rigidly connected to the top
surface of the sleeve with a thin layer of araldite. Normally at static condition, the lever connected to the casing
(sleeve) is in balanced condition and the spirit level shows the horizontality of the lever.In dry test the sample was
rotating in the housing in the different speed like 300, 450, 600 RPM with different time duration 20 min, 30 min, 40
min respectively. The friction and wear rate of the different sample were calculated from the different type of speed
of the motor and distance of the balance weight.

163
 Proceedings of NCAMMM - 2018

Fig 1. Friction measurement set-up

3. Equations
The coefficient of friction 𝜇𝜇 is calculated as follows:-
𝜇𝜇 =(𝑊𝑊 𝑏𝑏 𝐿𝐿)/(𝑊𝑊 𝑑𝑑 𝐷𝐷 𝑗𝑗 )
Ware rate was measured using the equation
Wr = (m i -m f )/t
Where, L= Distance from the centre of the lever to the point of balance weight, in mm. W b = Balance weight
(Kg.),W d =Dead weight (Kg.),D j = Journal diameter ( mm.)

4. Results and Discussion

The sample (which were prepared in a shape of a journal bearing) are being tested for different speed of motor and
for different time span. The outcomes of Calculated coefficient of Friction, and wear are given in the table 1 to 4.

Table 1: Coefficient of Friction of composite bearing with 5% graphite contain

RUNNING TIME
RPM

20 MIN. 30 MIN. 40 MIN. AVG. μ

WT.(g)
DEAD

L L L
(mm) μ (mm) μ (mm) Μ
1000 91 0.42 100 0.46 97 0.45 0.46
1500 133 0.45 136 0.46 140 0.47 0.46
300

2000 27* 0.43 30* 0.47 30* 0.47 0.47

1000 97 0.45 103 0.48 106 0.49 0.48
1500 136 0.46 142 0.48 143 0.48 0.48
450

2000 29* 0.45 30* 0.48 32* 0.5 0.48

1000 99 0.46 106 0.49 112 0.52 0.49
1500 142 0.48 142 0.48 142 0.48 0.48
600

2000 30* 0.47 32* 0.5 34* 0.52 0.5

OVERALL AVG. μ 0.483

164
Tribological properties of aluminium graphite composite journal bearing

Table 2: Coefficient of Friction of composite bearing with 7% graphite contain

RUNNING TIME

WT.(g)
DEAD
RPM

20 MIN. 30 MIN. 40 MIN.

AVG. μ
L L L
(mm) μ (mm) μ (mm) μ
1000 56 0.26 58 0.27 60 0.28 0.27
300

1500 83 0.28 83 0.28 86 0.29 0.28

2000 17* 0.28 17* 0.27 18* 0.29 0.27
1000 65 0.3 67 0.31 70 0.32 0.31
450

1500 89 0.3 98 0.33 103 0.35 0.33

2000 21* 0.33 22* 0.34 23* 0.35 0.34
1000 67 0.31 70 0.32 74 0.34 0.32
600

1500 95 0.32 98 0.33 100 0.34 0.33

2000 21* 0.33 23* 0.35 24* 0.37 0.35
OVERALL AVG. μ 0.312

Table 3: Coefficient of Friction of composite bearing with 9% graphite contain

RUNNING TIME
RPM

20 MIN. 30 MIN. 40 MIN.

AVG. μ
WT.(g)
DEAD

L L L
(mm) μ (mm) Μ (mm) μ
1000 65 0.3 70 0.32 74 0.34 0.32
300

2000 22* 0.34 22* 0.34 24* 0.35 0.34

1000 65 0.3 72 0.33 74 0.34 0.33
1500 91 0.31 98 0.33 102 0.35 0.33
450

2000 20* 0.33 23* 0.35 26* 0.36 0.35

1000 78 0.35 78 0.35 81 0.37 0.35
1500 102 0.35 110 0.37 115 0.39 0.37
600

2000 22* 0.34 24* 0.38 26* 0.41 0.38

OVERALL AVG. μ 0.351

165
 Proceedings of NCAMMM - 2018

Table 4: Wear rate, sliding speed and percentage of reinforcement

Sl. Reinforcemen-t (% Speed (RPM) Dead Sliding Distance Wear

No. of Gr) Wt. (mm) Rate (gm/sec)
(gm)
1 5 300 1000 800 0.0008
2 5 300 1500 800 0.0008
3 5 300 2000 800 0.0009
4 5 450 1000 1200 0.0009
5 5 450 1500 1200 0.001
6 5 450 2000 1200 0.0015
7 5 600 1000 2000 0.0013
8 5 600 1500 2000 0.0016
9 5 600 2000 2000 0.002
10 7 300 1000 800 0.0009
11 7 300 1500 800 0.001
12 7 300 2000 800 0.0011
13 7 450 1000 1200 0.0012
14 7 450 1500 1200 0.0014
15 7 450 2000 1200 0.0019
16 7 600 1000 2000 0.0017
17 7 600 1500 2000 0.0021
18 7 600 2000 2000 0.0023
19 9 300 1000 800 0.0023
20 9 300 1500 800 0.0024
21 9 300 2000 800 0.0026
22 9 450 1000 1200 0.0027
23 9 450 1500 1200 0.0029
24 9 450 2000 1200 0.0032
25 9 600 1000 2000 0.0035
26 9 600 1500 2000 0.0037
27 9 600 2000 2000 0.0038
Average Wear Rate 0.0023

0.42
0.41 2 kg
Coefficient of Friction

0.4 load
0.39 1.5 kg
0.38 load
0.37
0.36
0.35
0.34
250 450 650
RPM

Fig. 2: Effect of reinforcement and RPM on coefficient of friction

166
 Tribological properties of aluminium graphite composite journal bearing

It is clear from the table and the graph that the amount of graphite in the aluminium composite journal bearing is
more the coefficient of friction is decreased more as graphite particles have self-lubricating properties. But in the
other hand the hardness of the product is decreased as the graphite percentage is increased. So there have certain
limit or optimum amount of graphite that can be added with the composite journal bearings. It is seen from the data
sheet that after addition of 7% graphite the wear rate is more so it is the optimum position. Coefficient of friction is
also increased with the rotational speed of the motor.

0.004
300 RPM 0.0029
0.0035 0.0027 600

Wear Rate(gm/sec)
450 RPM 0.0025 RPM
Wear Rate(gm/sec)

0.003 0.0023 450

0.0025 0.0021 RPM
0.0019
0.002 0.0017 300
0.0015 RPM
0.0015 0.0013
0.0011
0.001 0.0009
0.0005 0.0007
0.0005
0
800 1000 1200 1400 1600 1800 2000 2200
3 5 7 9
% Reinforcement Sliding Distance(m)

Fig. 2: Effect of reinforcement and sliding distance on wear rate in al-gr composite

The wear of Al–Gr composites is noticeably reduced with the addition of Graphite which facilitates easy machining
compared to Al alloy. But the experiment was not carried out with high Gr reinforcement (> 9%), because it may
show immediate increase of wear as the fracture toughness is decreased with increase in porosity as a result of
increasing Graphite content. It has been seen from outcomes from other researches [9, 11] that, decrease in hardness
is observed with increase in % reinforcement of Gr due to increased porosity, which results in very high were rate.

0.003 0.003 2000 gm

0.0028 600 RPM Load
Wear Rate(gm/sec)
Wear Rate(gm/sec)

0.0026 0.0025
0.0024 450 RPM 1500 gm
0.0022 Load
0.002 0.002
0.0018 300 RPM 1000 gm
0.0016 0.0015 Load
0.0014
0.0012
0.001 0.001

500 1000 1500 2000 2500 250 350 450 550 650 750
Load(gm) Sliding Speed(RPM)

Fig. 3: Effect of load and sliding speed on wear rate in Al-Gr composite

5. Conclusion
Study on tribological properties such as coefficient of friction and wear rate of sintered Al-Gr composite journal
bearing developed through powder metallurgy techniques has been described in this paper. AMCs with soft
reinforcement Gr particles posses better wear characteristics as the thin layer created by Gr particles prevents the
contact between two adjacent sliding metallic surfaces. The wear has been considerably influenced due to the
presence of Gr particulates and the more stable wear is observed. With the increase in Gr. content, formation of
smaller sized wear debris was also observed. It is investigated that Al–Gr sintered composite with 0–7 wt. %Gr

167
 Proceedings of NCAMMM - 2018

shows enhanced wear behaviour. It is thereby presumed, that on machining of Al–Gr compositesdeveloped by this
method will undergosignificant reduction of cutting forces whichwill be attributed to the possible reduction of
friction.

References
[1] Sutradhar G, Chakrabarti A K. A Study on Friction and Wear in Ductile iron Bearings under dry sliding
condition. The 54th International Foundary Congress New Delhi.1987;XXXIII(11):25-30.
[2] Yating Wu, Bin Shen, Lei Liu, Wenbin Hu. The tribological behaviour of electro less Ni–P–Gr–SiC composite.
Wear. 2006;261:201–207.
[3] Rang Chen, Akira lwabuchi, Tomoharu Shimizu, HyungSeop Shin, HidenobuMifune. The sliding wear
resistance behavior of NiAI and SiC particles reinforced aluminium alloy matrix composites. Wear.
1997;213:175-184.
[4] Huijun Yang, Rui-yingLuo, Suyi Han, Midan Li. Effect of the ratio of graphite/pitch coke on the mechanical
and tribological properties of copper–carbon composites. Wear. 2010 may; 268(11):1337–1341.
[5] Hsiao Yeh Chu, Jen Fin Lin. Experimental analysis of the tribological behaviour of electroless nickel-coated
graphite particles in aluminum matrix composites under reciprocating motion. Wear. 2000;239(1):126-142.
[6] Riahi A R, Alpas A T. The role of tribo-layers on the sliding wear behaviour of graphitic aluminum matrix
composites. Wear. 2001 October;250(1-12):1396-1407.
[7] Akhlaghi F, Zare-Bidaki A. Influence of graphite content on the dry sliding and oil impregnated sliding wear
behaviour of Al 2024–graphite composites produced by in situ powder metallurgy method. Wear. 2009
January;266(1-2):37-45.
[8] Hassan A M, Tashtoush G M, Ahmed A K J. Effect of graphite and/or silicon carbide particles addition on the
hardness and surface roughness of Al–4 wt.% Mg alloy. Journal of Composite Material. 2007 February 1;453–
465.
[9] Ted Guo M L, Tsao C-Y A. Tribological behaviour of self-lubricating aluminium/SiC/graphite hybrid
composites synthesized by the semi-solid powder-densification method. Composites Science and Technology.
2000 January 1; 60(1):65-74.
[10] Jin fengLeng, Gaohui Wu, Qingbo Zhan, Zuoyong Dou, Xiaoli Huang. Mechanical properties of SiC/Gr/Al
composites fabricated by squeeze casting technology.ScriptaMaterialia. 2008 May 27;59:619-622.
[11] Basavarajappa S, Chandramohan G, Mahadevan A, Thangavelu M, Subramnin R, Gopalakrishnan P. Influence
of sliding speed on the dry sliding wear behavior and the subsurface deformation on hybrid metal matrix
composite. Wear. 2007 March 15;262(7-8):1007-1012.
[12] Ghosh S, Sahoo P, Sutradhar G. Wear Behaviour of Al-SiCp Metal Matrix Composites and Optimization Using
Taguchi Method and Grey Relational Analysis. Journal of Minerals and Materials Characterization and
Engineering. 2012;10(4236):1085-1094.

168
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Microstructure and Mechanical Properties of Rheocast of ADC12 Aluminium Alloy

Sujeet Kumar Gautama*, Himadri Royb, Aditya Kumar Lohara, Sudip Kumar Samantaa, Goutam
Sutradharc
a
Advanced Manufacturing Centre, CSIR-Central Mechanical Engineering Research Institute, Durgapur, India
b
NDT &Metallurgy Group, CSIR-Central Mechanical Engineering Research Institute, Durgapur-713209, India
c
Department of Mechanical Engineering, Jadavpur University, West Bengal, 700032, India

Abstract: In the present work, the variation in processing parameters (cooling slope angle, slope length) of
rheocasting on its microstructural evolution has been investigated in detail. Furthermore, comparisons between the
conventional cast and rheocast samples have been made at same pouring temperature. The ingot of ADC12
aluminium alloy was melted in resistance heating furnace at 750oC. Subsequently, the liquid metal alloy was poured
in a holding furnace and set at the desired temperature and released on cooling slope plate (made of stainless steel)
at a fixed temperature of 585oC with varying slope angle (30, 45, 60 degree) and slope length (400, 500, 600 mm). It
was found that among all processing parameters, slope angle and slope length, adversely affect the morphology of
the structures (particle size and degree of sphericity). The optimum values of these parameters (slope angle and
slope length) are found to be 45 degrees and 500 mm respectively. These optimum parameters generate the
minimum particle size and the maximum degree of sphericity of the observed microstructure. The results also
indicate that the rheocast samples are having better mechanical properties than corresponding gravity cast samples.
The fracture studies reveal the presence of mixed mode fracture to be predominant fracture mode in case of
rheocast specimens.
Keywords: Rheocasting, Cooling Slope; Tensile properties; ADC12 Aluminium alloy

1. Introduction
Semi solid metal processing (SSM) has potential to be a significant manufacturing technology for automotive
industry, aerospace, electrical and construction industry. This processing technology provides various advantages
such as better mechanical properties, the high degree of dimensional precision, over the conventional method [1-2].
The key element of the SSM is the transformation of microstructure from dendritic to non-dendritic along with the
reduction in the segregation and porosity levels in the castings [3-4]. The rheocasting process has become even more
popular in the recent decade, as it possesses various advantages over thixocasting, as well as less processing cost,
enhancing the casting dimensional accuracy and improvement of die life [5]. Several techniques such as ultrasonic
vibration, electromagnetic stirring, and magneto- hydrodynamic and rapid cooling have been investigated to find out
near equiaxed grain microstructure. Alternatively, the cooling slope process is one of the simplest and best
techniques that use the formation of semisolid slurry for rheocasting and improves the uniformity of microstructure.
Significant works were conducted at semi solid metal processing [6-9 ] on producing SSM via inclined plate (CRP -
Continuous Rheoconversion Process and cooling slope) [10-15].The most common materials used for SSM as
evidenced in the existing literature papers are A356/A357(very low Cu) and A380 (higher Cu 3%) aluminium alloys
as compared to ADC12 at 1.5% level. Regardless of this fact, ADC12 aluminium alloy is a material widely used in
Microstructure and mechanical properties of rheocast of ADC12 aluminium alloy

aluminium die cast industry. However, no complete study about the semi solid metal processing of this alloy has
been available to the best knowledge of the authors. ADC12 aluminium alloy has excellent material properties such
as high castbilty, high fluidity and low shrinkage rate. In contradiction, there are several cast defects in conventional
die casting components e.g. porosity, surface blister and blowholes. Considering that porosities are obtained
especially in die casting because of turbulent flow and defects like surface blister cannot be normally heat treated
[16-17]. To resolve this problem in ADC12 aluminium alloy, the rheocasting process is chosen in this study. The
primary objective of this work is to study the semi solid metal processing of ADC12 aluminium alloy using the
cooling slope technique. In addition, attempts have been made to analyse the effect of processing parameter of
cooling slope technique on the morphology of primary phase aluminium alloy and mechanical properties.

2. Experimental Procedures
ADC12 Al alloy has been used for the present work, liquids and solidus temperatures of this material are 572⁰C and
520⁰C, respectively as validated through DTA result shown in figure 1. Table.1 shows the chemical compositions of
ADC12 Al alloy.
Table 1 Chemical composition of ADC12 aluminium alloy

Element Si Cu fe Mn Mg Zn Ti Cr Ni Pb Sn Al

Wt % 10.57 1.685 0.780 0.314 0.082 0.117 0.021 0.019 0.030 0.65 0.013 balance

Fig. 1 DTA curve of ADC 12 aluminium alloy

In the present work, the cooling slope plate made of stainless steel was used for semi solid slurry generation. ADC12
ingot was fed inside silicon carbide crucible which was located within resistance heating furnace and melted at 750
⁰C. The melt was degassed and poured into holding furnace, which was set at the desired temperature (585 ⁰C). The
holding furnace was located at the top of cooling slope experimental set up as shown in Fig.2. The process was
carried out at the pouring temperatures 585 ⁰C. Cooling slope plate was adjusted with respect to the horizontal plane
at different angles (30o, 45o and 60o) and preheated at the constant temperature of 60 ⁰C by using circulation
underneath. Arrangements to change the length of the cooling slope plate were also made. Three different slope
lengths were considered for the investigation namely, 400 mm, 500mm and 600 mm. Prevention of sticking of
liquid melt on the cooling plate was then taken care by boron nitride coated layer on cooling slope plate. When the
melt was flowing through the cooling slope plate, the temperature was monitored using k- type thermocouples,
which were placed at three different locations and their output was recorded by rapid data acquisition software. At

170
 Proceedings of NCAMMM - 2018

the exit of the cooling slope, the slurry was filled in the preheated copper mould (at 200 ⁰C and then cooled at
ambient temperature). The detailed metallographic examination was carried out from both conventional cast and
those from rheocast samples, using an optical microscope (M/s Axioimager, Carl Zeiss, UK) and Field Emission
Scanning Electron Microscope (FESEM, Model: ∑igma HD, M/s. Carl Zeiss) for comparative study.

Boron nitride
Oil circulating Holding

Cooling slope

K-type

Copper mould

Digital data acquisition

Fig.2 schematic diagram of cooling slope techniques

2.1 Mechanical test

Tensile specimens were fabricated from various rheocast samples to evaluate their tensile properties. The dimension
of the tensile specimen is shown in Fig.3. From each experimental run, three rounded tensile specimens were
machined from solidified billet. Tensile tests were carried out using a 50kN Universal Tensile testing machine (M/s.
Tinius Olsen) at 0.5 mm/sec displacement rate. Obtained results were analysed and effects of the processing
parameters on tensile properties were determined. The broken fracture samples after tensile tests were characterized
using Zeiss Field Emission Scanning Electron Microscope for fractographic analysis.

Fig. 3 Dimensions of tensile samples

3. Results and Discussion

3.1 Microstructure analysis
Figure 4 depicts gravity cast microstructure, which was directly poured into the rectangular copper mould without
using the cooling slope. In the microstructure, the morphology of primary phase of Al alloy having the dendritic
structure with a size of more than 272 μm and average arm spacing 15 μm of the primary phase of aluminium is

171
Microstructure and mechanical properties of rheocast of ADC12 aluminium alloy

observed. In conventional cast, the dendritic morphology of primary phase of aluminium alloy and long needle type
mixture of eutectic Al-Si alloy is revealed.

Elongated and needle-shaped

eutectic mixtures structure

Dendritic structure of primary Al

alloy

In cooling slope process, microstructure consists of the globular primary phase of aluminium alloy throughout the
sample. The formation of globular phase can be thought to arise from the fact that as the melt flows over the cooling
slope plate, it loses its superheat. As a result, the formation of heterogeneous nucleation of the primary phase of
aluminium alloy takes place at contact surfaces between the melt and cooling slope plate. Due to shearing effect,
these particles are detached from the surface of cooling slope plate and are carried along with flowing melt. The
detached particles become spherical in form. Later, it gets converted into semisolid on reaching the exit of slope
plate. The difference in the microstructure of conventional cast vis-à-vis rheocast samples in this research work are
found similar with few earlier reported work on A356 alloys [18-19]. The effect of the change in slope angle and
slope length on various microstructures is studied in this investigation.
3. 1.1. Effect of slope angle - Slope angle governs flow rate and a contact time between the molten metal and slope
surface. Fig 5 (a) shows the change in the particle size and globularity of primary phase of Al alloy due to the
change of slope angle at the constant pouring temperature and slope length. Lower slope angle cannot completely
convert the dendritic morphology into the globular form of the α-Al phase in the microstructures [20]. Below a
certain value of cooling slope angle (i.e., 30 degrees in this case) the molten metal flows slowly and solid shell
forms easily on the cooling slope surface. Increasing the slope angle from 30o to 45o decreases the average size of
the primary α-Al grains and increases the globularity as shown in Fig 5(b) and 5 (c). Increasing the slope angle
results in an increase of the shear stress that helps to break the dendritic microstructure and converts it to nearly
more and more globular and fine grains. When the slope angle is increased further (60o) beyond an optimum value
(45o in this case) the slurry passing over the inclined plate travels at a high speed, decreasing the amount of heat
dissipated from the molten metal that reaches at the end of the inclined plate. As a result, the collected semi-solid
metal contains high liquid fraction and low solid particle in it, which is undesirable [21]. Typical microstructure of
samples poured at 60o slope angle is shown in Fig 5(d).

172
 Proceedings of NCAMMM - 2018

A
B

C D

Fig 5. Microstructure of rheocast samples at constant pouring temperature (585oC) and slope length (500mm): (a)
direct effect of (slope angle); (b) 30; (c) 450; and (d) 600.

3.1.2. Effect of slope length - The microstructures of rheo-cast samples at different cooling slope lengths of 400, 500
and 600 mm at 45-degree slope angle are shown in Fig. 6. It can be clearly seen from Fig 4 and Fig 6 that the
morphological changes are significant with and without using cooling slope. The dendritic morphology of the
gravity cast sample is replaced with equixed grains using cooling slope technique; which facilitates heterogeneous
nucleation due to rapid heat exchange and successive detachment of the formed crystals from the surface due to the
shear has driven the flow of the melt through the slope [22]. The optimum microstructure having equixed fine grain
and the high degree of sphericity is obtained at 500 m length. Further, increase in length increases the particle size
along with an increase in the degree of sphericity. This phenomenon can be attributed to the fact that the increase in
slope length leads to the thickening of the solid layer formed between the melt and the inclined slope. As a
consequence, the rate of heat transfer along with the rate of cooling of the melt flowing on the surface of the inclined
plate decrease, which leads to lowering of the nucleation rate of primary solid phase. Therefore, size of primary
solid phase available for the final microstructure solidified into the mould increases [2].
3.2. Tensile results
Figure7 shows the comparison of tensile results obtained experimentally from the conventional cast and cooling
slope cast at different processing conditions. Morphology of microstructure is having more effect on tensile strength
and elongation[23]. The conventional cast of ADC12 Al alloy has large and elongated particles (dendritic).This
dendritic structure of α-Al effects on mechanical properties because of high strain hardening and micro-porosity.

173
Microstructure and mechanical properties of rheocast of ADC12 aluminium alloy

During the solidification, the formation of microporosity within melt takes place which creates a solid slecton
reducing the fluidity of melt. It also adversely affects the tensile strength and elongation of the alloy. In cooling
slope casting, near globular α-Al phase is obtained which improves its tensile strength and elongation of the alloy.
Thus, in cooling slope casting a likelihood transition of fracture manner from intergranular to transgranular is seen.
Few earlier investigators have compared conventional cast and the rheo cast of Al alloys [24-25].

A B

C D

Fig 6. Microstructure of rheocast samples at constant pouring temperature (5850C) and slope angle (450): (a) direct
effect of processing parameter (slope length); (b) 400 mm; (c) 500 mm; and (d) 600 mm.

Fig7. Tensile strength and % elongation of ADC12 alloy of conventional cast and cooling slope cast at different
processing conditions

3.3. Fracture surface analysis

174
 Proceedings of NCAMMM - 2018

The fracture surface of the conventional cast of tensile specimens is shown in Fig 8 (a) and (b). A crack initiation
point in gravity cast depends on intermetallic structure. It causes the structural defect such as porosity. A crack
initiation structure of these intermetallic phases is plate and long needle type shape. Crack initiation, it is propagated
by cracking of eutectic Al-Si particles and Fe-based intermetallic compound; this is responsible for facets of the
fracture surface. Fig 8(c) & (d), (e) & (f) and (g) & (h) show the factrography of cooling slope cast under different
processing conditions. Generally, mixed mode fracture manner occurs in cooling slope cast alloy due to void
initiation at globular eutectic silicon particles and is responsible for facets in the fracture.

A B C

D E F

G H

Fig 8. FESM fractographs of tensile fracture surfaces at different magnification 1000x & 2000X. A&B conventional
cast. cooling slope cast at different processing condtions C&D angle 30 degree ,length 500mm and pouring
temperature 585°C. E&F angle 45 degree, length 500mm and pouring temperature 585°C. G&H angle 60 degree,
length 500mm and pouring temperature 585°C.

4. Conclusions
The obtained result of this investigation infers the following:
1. Morphology of α-Al phase in conventional cast alloy readily changes from dendritic to the non-dendritic structure
while flowing through the cooling slope plate which facilitates heterogeneous nucleation due to rapid heat exchange

175
Microstructure and mechanical properties of rheocast of ADC12 aluminium alloy

and successive detachment of the formed crystals from the surface due to the shear has driven flow of the melt
through the slope.
2. Since the slope angle and lengths influence the heat extract for the molten metal and thus the cooling rate, which
thus controls the ability to obtain globular alpha aluminium. It is observed that the optimum slope length (500 mm),
which enables smaller particle size along with the maximum degree of sphericity, is obtained at the constant slope
angle of 450. Similarly, the optimum slope angle (smaller particle size and the high degree of sphericity) is observed
at constant slope length of 500 mm.
3. An increase in tensile strength and elongation were obtained in rheocast samples at the optimum value of cooling
slope parameter 450, 5850C, 500 mm than conventional cast.

Acknowledgements
The authors would like to thank Director, CSIR- Central Mechanical Engineering Research Institute (CMERI) for
his kind permission to carry out and publish this work. The authors would like to acknowledge the help rendered by
Central Research Facility, CMERI-Durgapur, for material characterization. The authors express their heartfelt
gratitude to AdMaC group (Anmol khalko, Anup rajak) staffs for their support throughout the work.

References
[1] Gencalp S, Saklakoglu N. Semisolid microstructure evolution during cooling slope casting under vibration of
A380 aluminum alloy. Materials and Manufacturing Processes. 2010 Aug 31;25(9):943-7.
[2] Taghavi F, Ghassemi A. Study on the effects of the length and angle of inclined plate on the thixotropic
microstructure of A356 aluminum alloy. Materials & Design. 2009 May 31;30(5):1762-7.
[3] Spencer DB, Mehrabian R, Flemings MC. Rheological behavior of Sn-15 pct Pb in the crystallization range.
Metallurgical and Materials Transactions B. 1972 Jul 1;3(7):1925-32.
[4] Flemings MC. Behavior of metal alloys in the semisolid state. Metallurgical and Materials Transactions B.
1991 Jun 1;22(3):269-93.
[5] Park C, Kim S, Kwon Y, Lee Y, Lee J. Mechanical and corrosion properties of rheocast and low-pressure cast
A356-T6 alloy. Materials Science and Engineering: A. 2005 Jan 25;391(1):86-94.
[6] Pan QY, Hogan P, Apelian D, Makhlouf MM. the continuous rheoconversion process (CRPTM).
[7] Liu D, Atkinson HV, Jones H. MTDATA thermodynamic prediction of suitability of alloys for thixoforming.
InS2P 8th International Conference 2004 Sep 21.
[8] Apelian D.SSM and Squeeze Casting: Principles & Opportunities. NADCA Transactions 2006
[9] Jorstad J, Apelian D. Pressure assisted processes for high integrity aluminum castings. International Journal of
Metalcasting. 2008 Jan 1;2(1):19-39.10. Xu Q, Apelian D, Makhlouf MM. Research programs.
[10] Birol Y. Cooling slope casting and thixoforming of hypereutectic A390 alloy. journal of materials processing
technology. 2008 Oct 16;207(1):200-3.
[11] Birol Y. A357 thixoforming feedstock produced by cooling slope casting. Journal of Materials Processing
Technology. 2007 May 7;186(1):94-101.
[12] Xu J, Wang TM, Chen ZN, Zhu J, Cao ZQ, Li TJ. Preparation of semisolid A356 alloy by a cooling slope
processing. InMaterials Science Forum 2011 (Vol. 675, pp. 767-770). Trans Tech Publications.

176
 Proceedings of NCAMMM - 2018

[13] Birol Y. Semi-solid processing of the primary aluminium die casting alloy A365. Journal of Alloys and
Compounds. 2009 Apr 3;473(1):133-8.
[14] Haga T, Suzuki S. Casting of aluminum alloy ingots for thixoforming using a cooling slope. Journal of
materials processing technology. 2001 Dec 3;118(1):169-72.
[15] Tian C, Law J, Van Der Touw J, Murray M, Yao JY, Graham D, John DS. Effect of melt cleanliness on the
formation of porosity defects in automotive aluminium high pressure die castings. Journal of materials
processing technology. 2002 Mar 5;122(1):82-93
[16] Zhao HD, Wang F, Li YY, Xia W. Experimental and numerical analysis of gas entrapment defects in plate
ADC12 die castings. Journal of materials processing technology. 2009 May 1;209(9):4537-42.
[17] Das P, Samanta SK, Das R, Dutta P. Optimization of degree of sphericity of primary phase during cooling
slope casting of A356 Al alloy: Taguchi method and regression analysis. Measurement. 2014 Sep 30;55:605-
15.
[18] Vundavilli PR, Mantry S, Mandal A, Chakraborty M. A Taguchi Optimization of Cooling Slope Casting
Process Parameters for Production of Semi-solid A356 Alloy and A356-5TiB2in-situ Composite Feedstock.
Procedia Materials Science. 2014 Jan 1;5:232-41.
[19] Saklakoğlu N, Gencalp S, Kasman Ş, Saklakoğlu İE. Formation of globular microstructure in A380 aluminum
alloy by cooling slope casting. InAdvanced Materials Research 2011 (Vol. 264, pp. 272-277). Trans Tech
Publications. (2011)
[20] Nourouzi S, Baseri H, Kolahdooz A, Ghavamodini SM. Optimization of semi-solid metal processing of A356
aluminum alloy. Journal of Mechanical Science and Technology. 2013 Dec 1;27(12):3869-74.
[21] Salarfar S, Akhlaghi F, Nili-Ahmadabadi M. Influence of pouring conditions in the inclined plate process and
reheating on the microstructure of the semisolid A356 aluminum alloy. In8th Int. Conf. on Semisolid Proc. Of
Alloys and Composites, Cyprus 2004.
[22] Das P, Samanta SK, Ray T, Venkatpathi BR. Mechanical properties and tensile fracture mechanism of
rheocast A356 Al alloy using cooling slope. InAdvanced Materials Research 2012 (Vol. 585, pp. 354-358).
Trans Tech Publications.
[23] Poddar P, Sahoo KL. Microstructure and mechanical properties of conventional cast and rheocast Mg–Sn
based alloys. Materials Science and Engineering: A. 2012 Oct 30;556:891-905.
[24] LÜ SL, WU SS, ZHU ZM, Ping AN, MAO YW. Effect of semi-solid processing on microstructure and
mechanical properties of 5052 aluminum alloy. Transactions of Nonferrous Metals Society of China. 2010 Sep
1;20:s758-62.

177
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Feasibility Study of Forging of Reduced Pure Fe 2 O 3 Briquettes

1
Ritwik Das, 2Manas Kumar Mondal, 3Susanta Pramanik
1
Dept of MME, NIT Durgapur, 2Dept of MME, NIT Durgapur
3
Dept of MME, NIT Durgapur , Email: sus_metnitd@yahoo.com

Abstract: Production of billets and blooms is carried out by continuous casting or mechanical working of ingots
produced from liquid steel. These processes involve consumption of energy and time. The pure ferric oxide
briquettes reduced at 9000C can be forged directly in a press. This briquette after reduction has strength and
thermal energy to sustain mechanical working. This project involves the making of pure iron powder briquettes
along with fly ash as binder and carbon powder. They are then reduced at various temperatures for a definite period
of time. Consequently after reduction they are subjected to forging. This paper illustrates the optimisation of
different temperature and time for necessary mechanical working of briquettes.
Keywords: Mechanical working, reduction of iron ore, forging

1. Introduction
Over the last 40 years, the energy efficient performance and production rate of blast furnace have improved
significantly. The energy consumption of integrated iron making and steelmaking is 150 MJ/T for lump ore and
sinter fines or 650 MJ/T for magnetite pellets or 1050 MJ/T for hematitepellets. Iron making and steelmaking
generates approximately 7% of anthropogenic CO 2 emissions worldwide [1].The integrated iron and steelmaking
process typically consume thermal energy on the order of 16–21 GJ/T for crude steel production [2].The major
amount of thermal energy comes from burning of coke inside the blast furnace. In the integrated iron making and
steel making process, the blast furnace plays major role in reduction of iron ores. The iron ore fines need to be
agglomerated to charge in the blast furnace. The agglomeration techniques of iron ore fines are sintering, pelletizing,
briquetting and nodulising. Extensive research work has been carried out on development of composite briquette
using iron ore and coal as well as on reduction of iron ore and charcoal with help of carbon fines, coke. Much
work has been performed to prepare composite briquette using iron ore fines, coal fines and different binders.
However, limited work has been done on characteristics of reduced composite briquette where charcoal is a
carbon source. Somerville [3] studied about importance of high strength against breakdown of iron ore, carbon
and a binder starch containing briquettes or pellets. The briquettes were prepared with varying carbon-to-iron
mass ratio between 0.18 and 0.26. The green briquettes were reduced at 1300°C and 1350 °C. Holding time was
varied from 3 to 15 min. From the experiment, it was found that reduced briquettes with starch binder were very
strong and 3800 N force required to yield under compression. Briquette strength increases with increasing
holding time and heating temperature. The increasing strength was imputed to the relative distance between iron
oxide particles in the composite briquette mixture. A slag network acts as a binder, developed as iron oxide
particles remain close together at low C/Fe ratios. Subsequently the strength of reduced briquettes is increased.
Long bottom et al. [4-5] observed that reduced coal and iron fines developed a combined slag network between
iron oxide particles, coal and gangue. Consequently the strength also increases due to this network. Slag
Feasibility study of forging of reduced pure Fe2O3 Briquettes

formation will be increased as the reduction temperature is increased or flux is added into the mixture. More slag
formation implies more strength in the reduced briquette. Increasing iron to carbon ratio in the compact
composite mixture facilitate in decreasing the distance between iron oxide particles. This fact imparts strength of
the briquettes or pellets after heating.
This work proposes an alternative method for manufacturing forged products from reduced ferrous
briquettes through mechanical working without melting of iron. An attempt has been made to study the effect of
mechanical working such as forging of the iron briquettes immediately after reduction. The reduction of briquettes
was carried out at 900°C, 1000°C and 1100°C. The holding times for reduction were 5 hours, 6 hours and 7 hours
for each temperature. The briquettes after reduction are taken to a hydraulic press for forging. The briquettes showed
maximum 12% reduction in height. In this paper the experimental details and discussions are expressed at greater
length.

2. Experimental Procedure
2.1 Raw materials
Laboratory gradeFe 2 O 3 fines, fly ash, and carbon fines were used as raw materials to prepare briquettes. Chemical
analysis of Laboratory grade Fe 2 O 3 fines, is shown in Table 1. The raw materials, pure Fe 2 O 3 supplied by Merck
Specialities Private Limited having purity level 98%, graphite fines supplied by LOBA Chemie containing 98%
carbon and fly ash was collected from DVC thermal power plant Durgapur. Chemical analysis of fly ash is shown in
Table 2.Sieve analysis of raw materials is shown in Table 3.Pure Fe 2 O 3 and fly ash were sieved in different
fractions. Finally 75 micron size has been taken for this project work.
Table 1: Chemical analysis of Laboratory grade Fe 2 O 3 fines
Constituents Fe 2 O 3 Fe 3 O 4 SiO 2 Al 2 O 3 Total Insoluble L.O.I
Elements
Amount (%) 98 - - - 1 1

Table 2:Chemical Analysis of fly ash

Constituents SiO 2 CaO Al 2 O 3 MgO MnO Fe 2 O 3 C Zn S Other
elements
Amount (%) 4.92 3.76 1.82 0.64 0.32 42.76 36.2 0.84 1.7 0.61

Table 3: Sieve analysis

Sieve size 1 mm 710 micron 500 micron 355 micron 250 micron 150 micron 75 micron 53 micron
(% weight) (% weight) (% weight) (% weight) (% weight) (% weight) (% weight) (% weight)

Pure Fe 2 O 3 - 2 3 1 2 8 75 9

Fly ash - - - 3 2 5 82 8
Carbon fines - - - 4 3 6 79 8

2.2. Preparation of Briquette

Pure Fe 2 O 3 fines, fly ash powder, and carbon fines were thoroughly mixed in arotating bottle for 6 hours. 40 gm of
the mixture from the stoichiometric mass of pure Fe 2 O 3 and carbon powder was taken. Briquettes were formed by
compacting the 40 gm mass in a die & punch assembly with the load of 4.5 Ton/square cm. Height of the briquette

179
 Proceedings of NCAMMM - 2018

is 4 cm & diameter 2 cm. Briquettes were kept for 24 hours for air drying. After that they were dried in a closed
chamber at a temperature of 2000C for 2 hours.
2.3. Reduction of briquettes
1. Sample enclosed in a steel chamber.
2. Steel chamber introduced inside the furnace after 9000C is attained.
3. N 2 gas was passed through the chamber at a flow rate of 2 lit/min.
4. The temperature of horizontal tube furnace was gradually raised to 9000C and maintained for 5 hours.
5. The sample was taken out & put on the anvil of the press along the axis of manufacture and the reduction in
height was noted.
6. The experiment was also carried out for 6 hours & 7 hours respectively at 9000C.
7. Similar procedure was carried out for10000C and 11000C respectively.

3. Results & discussion

The briquettes before reduction were tested for strength by drop test. The results are tabulated in Table 4.
Table 4: Drop test value for pre-reduced briquettes
Mass of briquette (gm) Load (Ton/square cm) No of drops
30 3 59
30 4.5 63
40 3 67
40 4.5 75

Table 5: Reduction in height during forging of reduced briquettes

Type of Briquette Reduction Temperature (0C) Holding Time (Hours) % of Height Reduction
A 5 -
900 6 -
7 2.5
B 5 -
1000 6 -
7 5
C 5 -
1100 6 7.5
7 12.5

Table 5 shows the result of forging operation. The briquettes having weight of 30 gm showed values in drop test
which is not permissible for mechanical working. Briquettes of category ‘A’ & ‘B’ showedminimum amount of
reduction in height as they are not fully reduced. Category ‘C’ briquettes could respond to mechanical working as
the reduced Fe developed strength with SiO 2 matrix, fly ash as shown in Fig: 1 & Fig: 2.

180
Feasibility study of forging of reduced pure Fe2O3 Briquettes

Figure 1: FESEM image of iron ore graphite Figure 2: FESEM image of iron ore graphite
composite briquette reduced at 10000C composite briquette reduced at 11000C

4. Conclusions
Mechanical working of reduced briquettes showed maximum 12% reduction in height.The increase of temperature
of reduction might be a possible reason for complete conversion of Fe 2 O 3 to Fe. The reduced briquettes got complete
removal of iron oxide with increase in temperature & time of holding. The strength of the briquettes may not be
comparable with mild steel as required for mechanical work.

5. Future scope of work

Manufacture of briquettes from iron ore fines instead of pure Fe 2 O 3 fines. Complete reduction of briquettes to have
more strength with addition of other metals.

References
[1] Kim Y, Worrell E. International comparison of CO 2 emission trends in the iron and steel industry. Energy
policy. 2002 Aug 1;30(10):827-38.
[2] Tanaka K. Assessment of energy efficiency performance measures in industry and their application for policy.
Energy policy. 2008 Aug 1;36(8):2887-902.
[3] Somerville MA. The Strength and Density of Green and Reduced Briquettes Made with Iron Ore and Charcoal.
Journal of Sustainable Metallurgy. 2016 Sep 1;2(3):228-38.
[4] Longbottom RJ, Monaghan BJ, Nightingale SA, Mathieson JG. Strength and bonding in reduced ironsand–coal
compacts. Ironmaking& Steelmaking. 2013 Jul 1;40(5):381-9.
[5] Longbottom RJ, Monaghan BJ, Mathieson JG. Development of a bonding phase within titanomagnetite-coal
compacts. ISIJ international. 2013;53(7):1152-60.

181
Sub - theme

Machining
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Some Experimental Studies on Relative Effects of Employing More Eco-Friendly and

Less Hazardous Vegetable Oil in Drops on Chip Formation and Cutting Forces in High
Speed Machining of Inconel-718

Arijit Dasgupta*, Aayush Dubey, Monojit Deb, Mrityunjoy Mondal and Asit Baran Chattopadhyay
Mechanical Engineering Department, AOT Adisaptagram, Aedconagar 712121, India.
*Corresponding author Tel: + 91 9836436808; E-mail address: arijitdg1994@gmail.com

Abstract: Inconel-718 is widely used for its unique properties, mainly high yield strength, high hot strength,
high heat and corrosion resistance, in aerospace, automobile, chemical and marine engineering. But processing
like machining and grinding of Inconel-718 is very difficult for its high strength, poor thermal conductivity,
work hardenability and chemical reactivity. The problem becomes more acute with increase in metal removal
rate (MRR) specially cutting velocity. In industries, Inconel-718 components are generally machined by suitable
coated carbide inserts at cutting velocity around 60m/min to 80 m/min under profuse cooling by mineral oil.
Machining exotic materials using hard turning, minimum quantity lubrication (MQL) or small quantity liquid
(SQL) is reasonably becoming popular for economy and less health hazards. In the present work, Inconel-718
rods have been turned at cutting velocities up to 120 m/min under different environments including using drops
(SQL) of sunflower oil mainly to investigate its effects, compared to some conventional environments on chip
formation and magnitude of the cutting force components. The experimental results have been quite significant
and have been presented here.
Keywords: Machinability; Inconel-718; coated carbide; Small Quantity Liquid (SQL); cutting force

1. Introduction
Nickel based superalloys are extensively used in aerospace engineering for much higher strength to weight ratio
as compared to denser steels. Inconel-718 is most widely used Nickel based superalloy. Inconel-718 is available
and used in both cast iron and wrought form. Inconel-718 is extensively used for making various components in
aerospace and automobile industries, nuclear and chemical plants, cryogenic appliances, marine engineering,
high pressure compressors, super conducting motors and generators, gas turbine parts like nozzles, blades, shafts
etc. which are subjected to high temperature and stress and so on. Although Inconel-718 is well formable but this
useful superalloy is quite difficult to machine and grind, which is essentially required to provide dimensional
accuracy and surface finish. Large cutting forces, high cutting temperature, poor finish and rapid tool failure are
the major problems and challenges in machining Inconel-718. Increase in cutting velocity and feed aggravate the
problems. Attempts have been made for more effective and efficient machining of Inconel-718 by proper
selection and use of cutting tool inserts as reported by Settineri et al., (2008)[1] and proper cutting fluid
application by Kamata and Obikawa, (2007)[2]. Wang et al., (2003)[3] highlighted that the cutting forces
became more steady with lesser magnitude hybrid turning of Inconel-718 rods by carbide inserts (WG 300
SGPN) where the job surface was locally heated by plasma flame and simultaneously cooling the tool by liquid
nitrogen. Type and method of application of cutting fluid play significant role on machinability of any work
material in terms of magnitude of cutting forces, temperature, quality of machined surface and tool life.
Adhikary et al., (2005)[4] investigated and reported that intensive stresses and temperature at the tool tip is a
Some experimental studies on relative effects of employing more eco-friendly and less hazardous vegetable oil in drops on chip formation
and cutting forces in high speed machining of Inconel-718

major cause of poor machinability and quick damage of the tool nose in machining the high strength and heat
resistive Inconel-718. Conventional methods of cutting fluid application especially soluble oil (oil water
emulsion) does not help much in machining Inconel-718 due to inability of the cutting fluid to reach at the small
chip tool interface. However, use of extreme pressure additive type suitable cutting oil plays some favorable role
especially when employed in the form of jet or mist. Dhar et al., (2002) [5] investigated that, flood cooling
method cause tremendous pollution and health hazards which could be substantially controlled by cryogenic
cooling. Advent of minimum quantity lubrication (MQL) or small quantity of liquid (SQL) method of cutting
fluid application made a breakthrough in respect of economy and control of pollution and health hazards. Behera
et al., (2014) [6] recently reported that proper applications of MQL enables substantial reduction of cutting
forces and surface roughness in machining Inconel-718.
Objective of the present work is to investigate the effects of employing SQL in the form of drops of
sunflower oil in comparison to dry and flood cooling by soluble oil and sunflower oil in respect of chip
formation and cutting forces in turning Inconel-718 by TiAlN coated carbide inserts at wide range of cutting
velocity and feed.

2. Experimental Investigation
2.1 Experimental Procedure and Condition
Inconel-718 rod is straight turned in a powerful rigid centre lathe by only TiAlN coated carbide inserts at
different speed-feed combinations under the different conditions including SQL of sunflower droplets. The
cutting force components (tangential P Z , transverse P Y , and axial P X ) were on line monitored using tool force
dynamometer, data acquisition system and PC. The machining chips were collected after 5 seconds of turning
and its thickness was measured under all the speed-feed environment combinations.

Table 1: Experimental conditions

Machine Tool Heavy Duty Centre Lathe
INCONEL-718
Composition: (in wt %)
Work specimen
Nickel - 52.5%, Chromium - 19.0 %, Iron - 18.5 %,
Niobium - 5.1 %, Molybdenum - 3.0 %, Aluminium - 0.5%,
Titanium - 0.9 %, Carbon - 0.08 %
Material: TiAlN Coated Carbide Type: MCMG
Cutting Tool
Geometry:(NRS) − 6°,−6°,6°,6°,45°,45°,0.08( mm)

Process Parameters
30, 60 , 90 , 120
Cutting Velocity(V C ) (m/min)

Feed(s o ) (mm/rev) 0.06 , 0.08 , 0.1, 0.12

Dry
Flood (soluble oil)
Environment
Small Quantity Lubrication (Sunflower Oil)
Flood (Sunflower Oil)
2.2 Experimental results and discussion
The role of the salient factors on the main or tangential component, P Z of the cutting force in straight turning of
ductile metal is expressed by:
183
 Proceedings of NCAMMM - 2018

PZ = ts oτ s (ζ − tan γ o + 1) (1)
Where, t= depth of cut, s o = feed rate, τ s = dynamic yield shear strength of the work material under the machining
conditions, ζ= chip reduction coefficient, γ o = orthogonal rake of the cutting tool.
From Eq. (1) Bhattacharya A, (1984) [7], P Z seems to be directly proportional to t and s o and depends upon the

value of ζ, (ratio of chip thickness after cut, a 2 and before cut, a 1 ). The value of τs of ductile metals, especially

work hardenable type is affected by cutting strain rate and temperature. The value of ζ plays a significant role on
the cutting forces. The value of ζ is mainly governed by tool rake angle and apparent coefficient of friction, µ a at
the chip tool interface. The value of µ a depends on the chip tool interaction which is favourably controlled by
lubrication cooling and cleaning by proper cutting fluid application. Inconel–718 is characterized by high hot
strength and poor thermal conductivity. Hence P Z etc. are not much affected by cutting temperature but its high
value in machining Inconel–718 is expected to affect tool life and surface integrity. The machinability
characteristic of any material is to some extent indicated by the form, pattern, colour, smoothness and thickening
of the chips. The chips collected after machining are shown in Fig.1 (a), (b), (c) and (d).

(a) dry SQL (b) soluble oil, flood (c) vegetable oil, flood (d) vegetable oil
Fig.1 (a),(b),(c),(d) Machining chips of Inconel-718 developed under different combinations of V C , s o and
environments

The hot strong but ductile Inconel–718 specimen produced more or less continuous ductile chips even
under dry machining Fig.1 (a). The variation in the pattern of chips and distortion with the increase in velocity
and feed may be reasonably attributed to the severe damage of tool tip and built up edge formation in absence of
any cutting fluid. Sizeable reduction in width of the chip occurring frequently might be due to the reduction in
depth of cut, t due to plastic deformation and rapid wear of the tool tip due to high stress and temperature. A few
typically damaged tool tips are shown in Fig.2 (a), (b), (c) and (d). Formation of built up edge (BUE) was also
noted and found to affect the chip configuration. It can be seen from Fig.1 (b) that the application of soluble oil
in the form of flood has substantially improved the chip configuration expectedly due to cooling and reduction
of BUE formation and deformation of the tool tip at least at the early stage of machining. Flood cooling by oil,
even the presently used sunflower oil, has improved the chip configuration further as can be seen from Fig.1 (c).
This is expected because of the lubricating effect of the oil. It is interesting and important to note Fig.1 (d) that
the favourable effect of use of the oil on the chip form almost prevailed even when that oil was employed in

184
Some experimental studies on relative effects of employing more eco-friendly and less hazardous vegetable oil in drops on chip formation
and cutting forces in high speed machining of Inconel-718

SQL as droplets only. So it can be stated that with respect to chip form and chip tool interaction, the effect
(percentage wise) of SQL has not been much less compared to flooding with oil.

(a) dry machining (b) soluble oil, flood (c) vegetable oil, flood (d) vegetable oil, SQL
Fig.2 (a), (b), (c), (d) Typical damaged tool tips while turning the Inconel specimen

The thickness of the chips were measured under all the present parametric combinations and shown in Fig. 3(a),
(b), (c), (d).

so = 0.08 Environments

(a) for s o = 0.06 mm/rev (b) for s o = 0.08 mm/rev

so = 0.10 environments

(c) for s o = 0.10 mm/rev (d) for s o = 0.12 mm/rev

Fig. 3 (a),(b),(c),(d) Chip reduction coefficient(ζ) obtained under the different parametric combinations of V C ,
s o and environments

The random and inconsistent variation in ζ has occurred seemingly due to wide change in the tool geometry and
chip tool interaction for BUE formation and damage of the tool tip under different speed-feed conditions. Fig.3
reveals that dry machining has all along provided much larger value of ζ expectedly. Even flood cooling by
soluble oil has reduced ζ drastically mainly for cooling effect. Flooding of oil has reduced ζ to the maximum
extent quite obviously because of effective cooling and lubrication. Fig.3 indicates that in overall, SQL
application also could reduce ζ substantially particularly at the medium range of feeds. Flooding of the oil
compared to SQL of the oil all along enabled larger reduction in ζ mainly due to additionally bulk cooling. Fig.4
(a),(b),(c),(d) depict the magnitude of the tangential (P Z ), transverse (P Y ) and axial (P X ) components of the
cutting force which was recorded under all the conditions undertaken.

185
 Proceedings of NCAMMM - 2018

(a) work material: INCONEL-718; tool: TiAlN coated carbide; feed s o : 0.06 mm/rev

environments

(b) work material: INCONEL-718; tool: TiAlN coated carbide; feed s o : 0.08 mm/rev

environments

(c) work material: INCONEL-718; tool: TiAlN coated carbide; feed s o : 0.10 mm/rev

(d) work material: INCONEL-718; tool: TiAlN coated carbide; feed s o : 0.12 mm/rev

Fig. 4 (a),(b),(c),(d) Role of the different methods of cutting fluid applications on cutting forces (P Z , P Y , P X )
under different V C – s o combinations.

186
Some experimental studies on relative effects of employing more eco-friendly and less hazardous vegetable oil in drops on chip formation
and cutting forces in high speed machining of Inconel-718

Several observations can be made from the plots in Fig.4 (a),(b),(c),(d).The force P Z has not been significantly
affected by increase of cutting velocity V C expectedly due to high hot strength, ductility, and low thermal
conductivity. But with the increase of feed, P Z increased more or less consistently under all the environments.
Dry machining all along provided larger value of cutting forces. Flooding by oil always reduced the forces to the
maximum extent reasonably, due to more effective lubrication in addition to bulk cooling which helped in
retaining the tool sharpness and favourable chip tool interaction.
It is important to note that machining of the Inconel-718 specimen by the present oil in SQL form also
enabled substantial reduction in P Z at all the speed – feed combinations and outperformed the use of soluble oil
flood. SQL has been very close to flood–oil method in respect of the cutting forces especially in case of P Z . The
irregular rise and fall in P Z under dry cut at different speed–feed combinations are likely due to random and
intensive BUE formation and damage of the tool tip. Fig.4 also visualizes that the thrust components P Y and P X
of the cutting force also got reduced almost equally significantly along with the flood cooling by soluble oil and
even the present oil consistently under the speed–feed combinations undertaken.

3. Conclusion
Based on the experimental observations the following conclusions were drawn:
• Dry machining of Inconel-718 is not recommendable at high feeds and cutting velocities beyond about 60
m/min because it adversely deforms the chip–form and sharply raises the cutting forces almost from
beginning of machining due to rapid damage of the tool-tip due to intensive stress and temperature
concentrated at the small tool-tip.
• Proper selection and method of application of cutting fluid may smoothen chip formation and reduce the
cutting forces to reasonable extent.
• Use of oil, even sunflower oil, in flood condition enabled maximum reduction of cutting forces under the
present investigation, but application of SQL of the same oil performed almost at par with flooding – oil and
better than flood cooling by soluble oil.
• Technological gain of employing SQL of the oil is slightly less compared to that attained by using flood-oil
but the overall socio-economic benefits of SQL method is huge compared to the flood-oil method.

References
[1] Settineri L, Faga MG, Lerga B. Properties and performances of innovative coated tools for turning inconel.
International Journal of Machine Tools and Manufacture. 2008 Jun 1;48(7-8):815-23.
[2] Kamata Y, Obikawa T. High speed MQL finish-turning of Inconel 718 with different coated tools. Journal of
Materials Processing Technology. 2007 Oct 1;192:281-6.
[3] Wang ZY, Rajurkar KP, Fan J, Lei S, Shin YC, Petrescu G. Hybrid machining of Inconel 718. International Journal
of Machine Tools and Manufacture. 2003 Oct 1;43(13):1391-6.
[4] Adhikary S, Halder B, Das S, Chattopadhyay AB. Study of Tool wear and tool life during turning of Inconel 718.
InProc. of the 12th National Conf. On Machines and Mechanisms (NaCoMM 2005) IIT Guwahati (pp. 45 – 52).
[5] Dhar NR, Paul S, Chattopadhyay AB. The influence of cryogenic cooling on tool wear, dimensional accuracy and
surface finish in turning AISI 1040 and E4340C steels. Wear. 2001 Nov 1;249(10-11):932-42.
[6] Behera BC, Chetan SG, Rao PV. Effects on forces and surface roughness during machining Inconel 718 alloy using
minimum quantity lubrication. InProc. of the 5th International & 26th All India Manufacturing Technology, Design
and Research Conference (AIMTDR 2014) 2014 (pp. 12-14).
[7] Bhattacharya A. Metal Cutting: Theory and Practice. Kolkata: Central Book Publishers; 1984.

187
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Topological Surface of H.S.S and Titanium31 using Micro Electro Discharge Machining

Jush Kumar Siddani*, Dr. C. Srinivas**, Dr. N.Nagabhushana Ramesh***

*
Acharya Nagarjuna University-Guntur; jushkumar.siddani@gmail.com
**
Department of Mechanical Engineering, RVR & JC College of Engineering – Guntur, csrinivas@ rvrjcce.ac.in
***
Department of Mechanical Engineering, Anurag Group of Institutions- Hyderabad; drnrameshmech@cvsr.ac.in

Abstract: Micro Electro Discharge Machining Drill is variant EDM processes in a situation where it employs water
as a dielectric fluid, with pressure flushing and rotating nanotube electrode. This paper brings out the relative
features of work pieces i.e., H.S.S and Titanium 31 about their surface roughness and Topological surfaces.
Keywords: Micro Electro Discharge Machining Drill (Micro EDMD), Titanium 31, High Speed Steel (H.S.S),
MRR

1. Introduction
The variant EDM process consists of a low energy pulse, water dielectric and a spark gap, flushing with pressure
through a nano pipe electrode which is assisted by Taylor coquette flow of the dielectric from the rotation of the
electrode. The Electro discharge Machining (EDM) is extensively applied for machining exotic materials and
complex shapes. Drilling of micro holes of 0.3 to 3.0 mm, escape holes in dies and pneumatic valves start holes for
wire cut EDM etc). The setup illustrated in the Fig: 1.

Fig 1. Schematic Representation of Micro EDM Drill

Slender electrode is mounted by a ceramic guide assembly. The nano - tube Tool and workpiece form a pair of
electrodes with electrode negative with uniform gap of few scores of microns. The erosion occurs by high frequency
sparks triggered by a square pulse generator. Water dielectric is employed in place of traditional kerosene pulse
generator. The conventional EDM is different for machining small holes owing to poor erosion rates, taper and
oversize are typical problems [1]. Frequent short circuits from spark gap contamination by erosion debris [2]. Water
Topological surface of H.S.S and Titanium31 using micro electro discharge machining
as dielectric medium is used efficiently which results in smallest spark gaps, fastest solidification of eroded particles
before they coalesce to form larger size debris [3]. The viscosity is superior due to flushing action.
The main advantage of micro EDM is to machine complex shapes of low force. The force applied is small
due tool and workpiece which have no contact during the machining process. The chattel provides advantage to both
the tool and workpiece. Advantages of micro EDM include high aspect ratio, setup cost and better precision with
large design freedom. Micro EDM is a contact less MRR process eliminating mechanical stress rattle on and shaking
problems during machining. Micro EDM is efficient to machine every type of micro holes with high characteristic
ratio.

2. Experiment Plan
The Work pieces are prepared on the Electron Discharge Machine Rapid drill.
Input factors- Work piece materials of High Speed Steel (H.S.S) and Titanium 31 electrode brass tube of 1mm and
3mm diameter tubes with 0.3 mm diameter holes, pulse current 3 Amps and 6 Amps, pulse on times 6 and 10, pulse
off times 4 and 7 (dial positions with increasing order) pulse voltage (100 V) and flush pressure (100 bar) were kept
uniform[4].
Output factors and their estimation mode - Roundness error of 3D-cmm, are observed at the top and bottom of the
hole and the taper. Surface finish roundness indices on Talysurf. Topological surface and SEM. Scanning Electron
Microscope

3. Results and discussion

Surface Characteristics - The results of surface roughness are listed in Table 1.
Table 1. Surface Roughness (R a ) range in Micro EDM Drill (µm)

Materials
Process H.S.S Titanium 31
Current (Low) Current (High) Current (Low) Current (High)

Micro EDM 1.13-2.12 1.77-2.12 0.91-1.02 1.94-2.63

The finding are as expected but illustrate the surface roughness in titanium 31 compared to H.S.S, sizeable variation
in the range of Ra are observed along Titanium 31 surface. The attributed to non uniform attrition occurring to spark
discharge is due to short circuit current surge. The eroded surfaces inTitanium 31 were very smooth with good
appearance compared to H.S.S, The surface finish and geometric accuracy of Titanium 31 surfaces have significant
advantage and lower to H.S.S surfaces.
The high energy of spark discharge produce melt on the spot of it impingement and atomization of liquid
metal by the spark forces and expanding gases. The erosion rates in Titanium 31 are high to facilitate the high
energy pulses and spark alone could not be the reason but also the short circuits between electrode and work peace
with surge of current. Normal spark discharge key amount of molten metal is retained and only a little part is
isolated as atomized droplets. Unstable forces of short circuit provide a large amount expulsion and lower

189
 Proceedings of NCAMMM - 2018
preservation of melt metal. The feature need further explosion of wearing down mechanism of Titanium 31 and
H.S.S. using micro EDM drill.

4. Topological Study
The eroded surface of work materials H.S.S besides Titanium 31 from Micro EDM are shown in the SEM
photographs of Figure 2.

Fig 2. SEM photographs 2 (a, b, c) of H.S.S and 2(d, e, f) Titanium 31

Erosive outcome of spark is substantial similarity of Sparking is a incessant spark discharges; consequently the
articulate sparks were employed. The tool electrode is bust anode within Micro EDM drill. The emblematic
occurrence of passivating film pattern on anode and evaluation of hydrogen on cathode in favour of gaseous bridge
spark gap also ionization of spark channel formation involve the nature of polarity.

190
Topological surface of H.S.S and Titanium31 using micro electro discharge machining
The erosion in each case appears elevated quench effect of circulate liquid prevent vaporization. The sizeable
retained metal which appear contain resolidified spark zone. Reduction is extensive in Micro EDM Drill owing is
soaring quench result in water based electrolyte. The eroded surface of Titinium31and H.S.S show distinctive form
associated along EDM [4] like rupture blisters from dissolve gas, pock marks plus crater after expulsion of melt
metal since spark energy. This removal to be superior on Micro EDM Drill owing to superior since watercourse
based electrolyte also promote oxidation tendency Titanium 31 in (Fig: d, e, f). Erosive result of spark is also seen
from the debris collected. The molten state is clearly seen from the evenly shaped spheroidal particle in H.S.S (Fig.
3a).

(a) (b)
Fig 3. (a) SEM Photographs of debris of H.S.S, (b) SEM Photographsof debris of Titanium 31

However in Titanium 31 the low quench rate water base electrolyte is seen mire pattern with too little time of
spheroidization (Fig. 3b).

5. Conclusions
• Micro EDM Drill has considerably lower erosion rates for drilling micro holes.
• The H.S.S and Titanium 31 leads to erosion rates of high thermal also electrical conductive.
• Surface irregularity follows the likely also normal pattern alike near that of erosion rates.
• Considerable reserved metal which resolidified also exhibit typical kind of spark attrition in the form of gas
pockets.
• The quenching effect of water as working fluid of Micro EDM Drill on Titanium 31 result into unequal debris
stick jointly and appears like sludge form.

References
[1] Crichton, I.M and J.A. McGeough. Theoretical, experimental and computational aspects of electrical discharge
arc machining process. Annals of the CIRP. 1994; 33: 429-433.
[2] Paulo Carlos Kaminski, and Marcelo Neublum Capuano. Micro hole machining by conventional electrical
discharge machine. Int. Jl. Mach. Tools & manufacture. 2003; 43: 143.1149.
[3] Kuneida. M, Lauwers. B, Rajurkar K.P. and Schumacher, B.M. Advancing EDM through fundamental insight
into the process. CIRP Annals-Manufacturing Technology. 2005; 54: 64-87.
[4] Electronica, EDM smart Drill. Manual – Electronic machine tools, Pune, India 2015.

191
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Experimental Investigation of Fiber Laser Cutting of Alumina

Rahul Rakshit1, Umar Arif1, Shakti Kumar, Mukul Anand, Vikas Kumar1*, Alok Kumar Das1
1
Department of Mechanical Engineering, Indian Institute of Technology (Indian School of Mines)
Dhanbad, Jharkhand, *E-mail address: vikas09me50@gmail.com

Abstract: Ceramics are the present need for solving various engineering application problems. The inherent
features of the ceramics have increased its utility in harsh operating environments. At the same time, the brittle
nature and high hardness of the ceramics lead to its difficulty in machining. Laser machining is considered as
one of the best options for accurate and precise cutting of ceramics. In this paper laser cutting of 1mm
thickness, alumina ceramic has been performed using CW 400W fiber laser. Nitrogen has been used as the
assisting gas. The kerf width, dross formation and striation formation of the laser cut sample have also been
investigated.
Keywords: alumina, fiber laser, laser cutting, assist gas, cut geometry

1. Introduction
Ceramics have ahuge demand in industrial fields owing to its excellent mechanical strength, wear resistance,
high corrosion resistance, high electrical resistance and sustainability at high operating temperature. Alumina
ceramics further possess certain unique properties, such as good smooth/ flat surface with minimal porosity,
high thermal shock resistance, high chemical and thermal stability, low warpage and very stable breaking
strength (1). So, alumina ceramics are the leading candidate in the electronic industries as chip resistors, IC
substrates, etc. The conventional methods that are used for machining alumina are mechanical routing, diamond
or abrasive saw cutting because of its high hardness and brittle nature. But uncertain tool wear and tool breakage
leads to fracture of brittle alumina(2). Furthermore, these methods are not suitable for cutting complicated
shapes with sharp edges. Alumina has very low electrical conductivity due to which it becomes difficult to
machine them with wire EDM. So among various non-conventional machining techniques, ultrasonic, abrasive
water jet and laser machining are used for cutting alumina.
Laser beam machining is one of the accepted techniques for ceramic cutting because it produces high
quality and precise cuts at lower costs compared with other techniques.It is an abrasion less technique in which
the material is neither subjected to cutting forces nor vibration. There are different types of laser which include
CO 2 laser, Nd:YAG laser, fiber laser, disc laser and excimer laser. Fiber lasers have numerous distinct
advantages than other types of lasers. Because of high efficiency (20- 40 % wall plug efficiency), compact size,
excellent beam quality, no hard optic resonator and reduced maintenance costs, fiber lasers demand is increasing
for industrial applications(3).Also, the laser mode can be continuous and pulsed (nano,pico,andfemto).
Continuous wave laser generates heat shocks so pulsed mode lasers are generally beneficial for cutting
ceramics.
There are numerous assisting gases such as nitrogen, oxygen, compresses air, helium,and argon that are
used for cutting ceramics. Various research works have shown the influence of these gases on the cut
quality(4,5). Chen et al. (5) performed fiber laser cutting on alumina using N 2 and O 2 gases. It was reported that
N 2 resulted in theblack cutting surface while O 2 did not show any such effect. Yan et al.(6)presented a crack-
free cutting of 6 mm thick alumina using pulsed fiber laser and O 2 as assisting gas. The characteristics of crack
formation were also discussed. Chen et al. (7) gave a detailed analysis of the microstructure and
Experimental investigation of fiber laser cutting of alumina

resolidificationbehavior of the recast layer during CO 2 laser cutting of electronic ceramic. The striation
formation on the cutting edge during laser cutting acts as an obstacle to good quality cut. Yan and Li (4)
performed striation free cutting using nanosecond pulsed Nd:YAG laser. An important analysis was carried out
that vaporization driven cutting and low pulse frequency can control the melt oscillation and striation free cut
can be obtained.

2. Experimental Procedure
In this experiment, 97% pure electronic alumina of 25 mm x 25 mm x 1mm dimension has been used. The laser
machine used for this work is Fibre Laser manufactured by Red Power. It is a continuous wave (CW) single
mode laser which delivers power upto 400 W ata wavelength of 1070 nm. The lens of 80mmfocal length and
0.022mmfocal point diameter is used. The precitic cutting head with thegas nozzle is mounted on the Z- axis.
The workpieces are clamped on a linear motor translation stage that enables the substrates to move along the X-
axis and Y- axis. The maximum translation cutting speed of 5000mm/min can be obtained. In laser cutting, the
assist gas protects the optics from any kind of damage and ejects the molten material out of the kerf. For
conducting this experiment, nitrogen gas is selected as the assisting gas which is blown from the nozzle at 1 mm
standoff distance from theworkpiece top surface.

Fig. 1: Schematic diagram of laser cutting

Table 1: Laser cutting parameters

Parameters Range

Laser power (W) 100- 300

Scanning speed (mm/min) 500-1500
N 2 gas pressure (bar) 4-6
Standoff distance 1mm

A number of trial experiments were performed to set the cutting parameters. It was observed that with
increasing assist gas pressure, the thin alumina was subjected to brittle fracture. With low power density and
high cutting speed, no cut was performed on the alumina. So based on these pilot experiments, the range of
variable parameters was set as shown in table 1. A number of slots were cut on each sample at different
parameters which were used for characterization.

3. Results and Discussion

193
 Proceedings of NCAMMM - 2018

A detailed analysis of the cut geometry was carried out. The variation of kerf width at different operating
parameters was recorded. The striation formation on the cut surface and the debris accumulation at the bottom
surface was studied.

Kerf width - Olympus metallurgical microscope was used for measuring the kerf width. It was measured at four
different locations (Fig. 2a) on a single cut at a particular parameter. The average value was recorded which was
used for studying the variation of kerf width.

(a) Measurement of average kerf width (b) Spattering due to resolidification

Fig. 2: Optical images of cut surfaces

After measuring the average kerf width values at different operating parameters, it was found that kerf width
increased with increasing laser power and assisting gas pressure. When the laser power is increased, it melts the
material to a larger extent thus increasing the kerf width. High gas pressure easily blows away the molten
material and increases the kerf width. But as the scanning speed was increased, the kerf width started
decreasing. This is due to areduction in interaction time of laser energy and theworkpiece.
When the molten material is not completely ejected out of the kerf then it resolidifies on the top surface
(fig. 2b).If the applied laser energy, cutting speed, focus position and assistant gas pressure are not controlled
properly, incomplete melting occurs or traces of molten metal re-solidify. It was observed that by increasing gas
pressure and reducing the scanning speed, the spattering can be controlled to some extent. Sometimes, improper
alignment of coaxial gas flow with the laser beam hinders the gas flow to the bottom-most surface due to which
molten material is ejected out on the top surface and resolidifies.

Burr formation/ debris accumulation - It was observed that debris was getting accumulated on the bottom side
of the cut surface (Fig.3). This formation was taking place because the laser beam energy starts decreasing as it
approaches the bottom of the cut surface. Though complete elimination of burr formation is not possible it can
be controlled to some extent by increasing laser power. To analyze the debris formation on the bottom side of
the cut surface, Field emission scanning electron microscope FESEM (Model: Supra 55, Make: Zeiss, Germany)
equipped with an energy-dispersive X-ray spectroscopy (EDS) was used. The results (fig. 4) showed that the
burr is actually the oxides of the alumina ceramic.

194
Experimental investigation of fiber laser cutting of alumina

Fig. 3: Optical image showing burr formation on the bottom surface of thecut sample

(a) EDS analysis of debris on bottom surface (b) FESEM image of the debris

Fig. 4: EDS analysis of the burr formation on the bottom side of thecut surface

Striation formation - When the molten material is blown out through the kerf, it generates striation marks on the
cutting edge which results in poor cut geometry. It basically happens due to mismatching of cutting velocity and
laser power. Few optical images of the striation formation on the cut geometry is shown in fig 5. It is evident
that when the laser power is increased without changing the scanning speed and gas pressure, a deeper striation
mark (fig. 5b) is formed due to improper removal of molten material.

(a) 100 W power, 1000mm/min scanning speed,and 4 (b) 300 W power, 1000mm/min scanning
bar N 2 gas pressure speed,and 4 bar N 2 gas pressure

Fig. 5: Optical images of striation formation

195
 Proceedings of NCAMMM - 2018

4. Conclusion
After performing the laser cutting operation on alumina ceramic, it can be concluded that
• Kerf width decreases with increasing scanning speed. But increasing laser power and gas pressure increases
the kerf width.
• EDS analysis predicted formation of oxide layer on the bottom side of cut sample
• Heat affected zone is negligible as it was not visible with an optical microscope with 100X magnification.
• A detailed study is required for complete elimination of burr formation.

References
[1] Liang Y, Dutta SP. Application trend in advanced ceramic technologies. 2001;21:61–5.
[2] Tuersley IP, Jawaid A, Pashby IR. Review : Various methods of machining advanced ceramic materials.
1994;42:377–90.
[3] Penn W, Team L. Trends in Laser Material Processing for Cutting ,Welding , and Metal Deposition using
Carbon Dioxide , Direct Diode , and Fiber Lasers. 2005;5706:25–37.
[4] Yan Y, Li L, Sezer K, Whitehead D, Ji L, Bao Y, et al. International Journal of Machine Tools &
Manufacture Nano-second pulsed DPSS Nd : YAG laser striation -free cutting of alumina sheets. Int J
Mach Tools Manuf [Internet]. Elsevier; 2012;53(1):15–26. Available from:
http://dx.doi.org/10.1016/j.ijmachtools.2011.07.006
[5] Chen X, Ji L, Bao Y, Jiang Y. High Quality Fiber Laser Cutting of Electronic Alumina Ceramics.
2011;155:917–22.
[6] Yan Y, Li L, Sezer K, Whitehead D, Ji L, Bao Y, et al. International Journal of Machine Tools &
Manufacture Experimental and theoretical investigation of fibre laser crack-free cutting of thick-section
alumina. Int J Mach Tools Manuf [Internet]. Elsevier; 2011;51(12):859–70. Available from:
http://dx.doi.org/10.1016/j.ijmachtools.2011.08.004
[7] Chen X, Ji L, Bao Y, Jiang Y. Improving cutting quality by analysis of microstructure characteristics and
solidification behaviour of recast layer formation on laser cut ceramic. J Eur Ceram Soc [Internet].
Elsevier Ltd; 2012; 32 (10) :2203–11. Available from:
http://dx.doi.org/10.1016/j.jeurceramsoc.2012.03.020

196
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Influence of different cooling conditions on Machinability during turning of EN-24 steel

Archana Thakur1, Alakesh Manna2, Sushant Samir3

1
PhD research scholar, Punjab engineering college (Deemed to be university) Chandigarh, 160012, India
2
Professor, Punjab engineering college (Deemed to be university) Chandigarh, 160012, India
3
Associate Professor, Punjab engineering college (Deemed to be university) Chandigarh, 160012, India
*Corresponding author E-mail: archana30207@gmail.com

Abstract: Application of cutting fluids in machining process reduces temperature and friction at cutting zone.
However, the use of cutting fluids elevates the total cost of production and health hazards to the operators. A
new method known as minimum quantity lubrication (MQL), in which very less amount of cutting fluid is
required to be sprayed into the cutting zone with compressed air. To enhance the heat carrying capacity and
lubrication of cutting fluid, nanoparticles were added to the cutting fluid. In this research, the nanofluid was
prepared by addition of 0.5wt.% copper oxide (CuO) to the soluble oil. The EN-24 steel was turned with CVD
coated carbide insert (CCMT 09T308) under different conditions i.e. dry, wet, MQL and nanofluid with MQL
cutting environments. The acquired results during turning revealed that the surface roughness height (Ra, µm),
cutting temperature, cutting forces and tool wear were reduced when turning operations were carried out with
CuO based nanofluid with MQL.
Keywords: Turning, MQL, surface roughness height, cutting temperature, cutting force, tool wear

1. Introduction
Cutting fluids play an important role in manufacturing industry. Because during machining very high
temperature was generated due to friction between tool and workpiece interfaces. High temperature at cutting
zone reduces tool life and surface quality. To minimize the problems, cutting fluids were used. Cutting fluids
provides lubricating and cooling effects. With a lots of advantages, cutting fluids has many disadvantages such
as harmful effect to the environment, disposal off cutting fluids increases machining cost, heath hazards to the
operators and wastages of cutting fluids [1-2]. To minimize these effects alternative i.e. minimum quantity
lubrication (MQL) environment was suggested by the different researchers. MQL shows technological and
economical advantages over dry and wet machining [3-7]. The nanometre sized particles can be added to the
base fluid to increase the thermal conductivity of the cutting fluids [8-9]. Manimaram et al. [10] reported that
thermal conductivity increases with addition of copper oxide to base fluid. The effectiveness of MQL process
can be increased if nanoparticles were added to the base cutting fluid [11]. Amrita et al. [12] studied the
performance of nanographite based cutting fluid during machining. They concluded that the addition of nano
particles with conventional fluid enhance the thermal conductivity of cutting fluid. Sharma et al. [13]
investigated the effect of nanofluids during turning of AISI D2 steel. Authors concluded that MQL using
nanofluid reduces temperature and surface roughness. Khandenkar et al. [14] analyzed the effect of Al 2 O 3 based
nanofluid during turning of AISI 4340. They concluded that cutting forces, tool wear and surface roughness
height were reduced when Al 2 O 3 based nanofluid was used over dry and wet turning. In the present set of
research, MQL set-up was developed and used during turning of EN 24 steel. The turning of EN 24 steel was
done under four environments i.e. dry, wet, MQL and CuO based nanofluid with MQL. The acquired results
from different turning environment were compared with each other.
Influence of different cooling conditions on machinability during turning of EN-24 steel
2. Material and Experimental Conditions
EN-24 Steel was used as workpiece material for turning experiments. The CVD coated tungsten carbide (CCMT
090T308) cutting insert was used for turning of EN-24 steel. The experiments were performed on HMT NH 26
Centre lathe. The fabricated MQL set-up is shown in Figure 1. This fabricated set-up was used to supply the air
and cutting oil in the form of mist. MQL set-up consists of nozzle, air control valve, oil container, mixing
camber and pressure gauges. The high velocity air supplied through air compressor, air and cutting oil were
mixed at mixing chamber and supplied to cutting zone through nozzle in the form of mist. In the present study,
soluble oil was used as base cutting fluid. The CuO nanoparticles were used with MQL to increase the
lubricating properties and cooling effect. The CuO nanoparticles were added to the base cutting fluid with the
concentration of 0.5 wt%. The nanofluid was prepared by dispersing 0.5wt% CuO particles in base cutting fluid
by sonication. To measure the turned surface roughness height (R a , µm), the Surfcom 130A surface roughness
measuring instrument, Zeiss, Japan, was used.

Pressure gauge
Fluid container

Control
valve

Mixing chamber

Figure 1 Minimum quantity lubrication set-up

Table 1 Equipments and parameters
Workpiece Material EN-24 Steel (300mm X 80mm)
Tool Holder specification SCLCR 1212F09
Cutting Tool CCMT 09T308
Machine Tool HMT NH 26 Centre lathe
Nano-cutting Fluid Water Soluble oil (20:1) + 0.5 wt%(0.08 vol.%) CuO nanoparticles,
Size of CuO nanoparticles 30-50nm
Surface Roughness Surfcom 13OA
Cutting Forces Tri-axial Turning force dynamometer
Cutting Temperature Infrared thermometer, MT-5
Tool Wear Tool maker microscope
Cutting Speed(Vc) 80 (m/min)
Feed rate(f) 0.16 (mm/rev)
Depth of cut(ap) 0.50 (mm)
MQL Air : 7bar, external nozzle
MQL flow rate 120ml/h
Cutting time(t) 60,120,180,240,300(sec)
Turning Environments (i) Dry turning
(ii) Wet turning
(iii) Turning with MQL
(iv) Turning with MQL + soluble oil + CuO
Cutting forces in all three directions during tuning were measured by tri-axial turning force dynamometer. The
infrared thermometer was used to measure the cutting temperature and tool wear was measured by tool maker

198
 Proceedings of NCAMMM - 2018

microscope. Table 1 represents the detail about experimental condition, cutting tool insert, machine tool and
measuring equipments used during experimental investigation.

3. Results and Discussion

3.1 Surface roughness
The variation of surface roughness height (R a ) with time under different machining environments is shown in
Figure 2. The applications of CuO based nanofluid with MQL and minimum quantity lubrication (MQL) in
turning provide better surface quality compared with dry and wet turning. The CuO based nanofluid with MQL
shows better results as compared with others environment because addition of CuO nanoparticles enhances
thermal conductivity of base fluid, and reduces friction between tool and work interface.

Dry Machining
Wet Machining
MQL
3 Nanofluid with MQL
Surface Roughness(Ra) micron

50 100 150 200 250 300

Machining Time(sec)

Figure 2 Variation of surface roughness with time(s)

The application of nanofluid with MQL provides better cooling and lubrication which decreases temperature at
tool and work interface and tool wear which leads to improve in surface quality. SEM images of turned surfaces
are shown in Figure 3. From Figures 3(a) and (b), it is clear that the surface finish in nanofluid with MQL
turning is better over dry turning. Figure 4 shows the turned workpiece under different machining environment.
From Figures 4(a),(b),(c) and (d), it is clear that the surface quality of the turned surfaces were better in
nanofluid with MQL turning over dry, wet and MQL turning, respectively.

Figure 3 SEM micrograph of surface turned at: (a) Dry environment (b) nanofluid with MQL environment

3.2 Cutting force

The variation of cutting force with time under different machining environments is shown in Figure 5. It is clear
that the main cutting force in CuO nanofluid with MQL is low as compared to dry, wet and MQL machining
environment. In dry machining, cutting force was high as compared to other machining environments, it is due
to absence of cutting fluid in tool-chip interface, tool is wear out quickly because of high friction between tool-
work interface and thermal softening of tool [14].

199
Influence of different cooling conditions on machinability during turning of EN-24 steel

Figure 4 Turned workpiece at different turning environment: (a) Dry turning, (b) Wet turning, (c) MQL turning,
(d) Nanofluid with MQL turning.
In wet machining process, tool wear is less over dry turning because of cutting fluid provides the lubrication and
cooling action during cutting, which reduces the cutting force (through dissipation of heat by convection and
conduction). MQL decreases the cutting force as compared to dry and wet turning. In MQL turning, effective
cooling is there due to aerosolization which brings down the temperature effectively at cutting zone and
subsequently protects against rapid increasing of tool wear thus maintaining the effectiveness of cutting tool and
the cutting forces are increase comparatively less as compared to dry and wet turning. The application of
nanocutting fluid i.e. addition of CuO to the base fluid helps in increase the heat transfer rate by providing
better cooling and lubrication, thereby increase of cutting forces are comparatively very less as compared to the
dry, wet and MOL turning. It clear from Figure 5 that the main cutting force in nanofluid with MQL machining
environment is lower as compared to dry, wet and MQL machining.
Dry Machining
Wet Machining
MQL
Nanofluid with MQL
300
Cutting Force(N)

200

100

50 100 150 200 250 300

Machining Time(Sec)

Figure 5 Variation of cutting force with machining time

3.3 Cutting temperature
Cutting temperature plays significant role in machining process. High temperature in cutting zone reduces tool
life and increases the power consumption. So it is important to control the cutting temperature at cutting zone.
Figure 6 shows the variation of cutting temperature with machining time. It was clearly observed that Nanofluid
with MQL reduces temperature at tool-work interface compared with dry, wet and MQL machining. When CuO
added in base fluid improves its thermal conductivity [10]. The reason for increase in thermal conductivity of
nanofluid is that suspended CuO nanoparticles increase surface area and heat capacity of fluid. The CuO based
nanofluid is used in MQL, it gives better cooling and lubrication effect due to the arosolization of nanofluid by
the pressurized air, which effectively enters the cutting region and reduces the temperature.

200
 Proceedings of NCAMMM - 2018

Dry Machining
Dry Machining Wet machining
Wet Machining 100 MQL
MQL Nanofluid with MQL
110
Nanofluid with MQL
90
Cutting Temperature(Degree C) 100

90 80

Tool Wear(micron)
80
70
70

60 60

50
50
40
40
30

50 100 150 200 250 300 50 100 150 200 250 300
Machining Time(Sec) Machining Time(sec)

Figure 6 Variation of cutting temperature Figure 7 Variation of tool wear with machining time
with machining time

3.4 Tool wear

The variation of tool wear with time under different machining environments was shown in figure 7.Tool wear
was less in nanofluid with MQL turning compared with other machining conditions. In nanofluid with MQL
turning process, cutting fluid penetreate into tool-work interface which reduces the friction between tool and
work hence reduce tool wear. Tool wear reduced by 50.11%, 38.43% and 16.10% compared to dry, wet and
MQL turning, respectively.
3.5 Chip morphology
Figure 8 shows the chip morphology at different experimental conditions. The chips obtain under dry turning
were blue in color presenting high temperature at cutting zone, the chips obtain in wet turning were silver color,
color of chips obtained in MQL turning were golden and in nanofluid with MQL turning the color of chips were
light golden because of better cooling and lubrication effect under MQL environment.

Figure 8:Chips at different turning environment: (a) Dry turning, (b) Wet turning, (c) MQL turning, (d)
Nanofluid with MQL turning.
4. Conclusions
In the research investigation, turning of EN-24 steel was carried out under dry, wet, MQL and nanofluid with
MQL environments. The results acquired from different environments were compared and the following
conclusions were drawn are listed as follows:
a) The machined surface quality was better with nanofluid with MQL turning over dry, wet and MQL turning.
b) Nanofluid with MQL showed the reduction in cutting force when compared with dry, wet and MQL
machining. The cutting forces in nanofluid with MQL turning was reduced to 60.80% over dry turning,
47.15% over wet turning and 25.10% over MQL turning.
c) There was 65.90%, 29.31% and 12.76% reduction in cutting temperature when CuO nanoparticles were
added to base fluid as compared to dry, wet and MQL turning respectively.

201
 Influence of different cooling conditions on machinability during turning of EN-24 steel
d) Tool wear was reduced when turning operations were carried out with nanofluid MQL environment as
compared to dry, wet and MQL turning, it is because of reduction in friction between tool and workpiece.
CuO based nanofluid with MQL in turning was found to be more advantageous over dry, wet and MQL turning.
CuO based nanofluid with MQL in turning produced advantageous results in improvement of surface quality,
reduction of cutting temperature, cutting force and tool wear.

Reference
[1] Amrita M, Srikant RR, Sitaramaraju AV, Prasad MM, Krishna PV. Experimental investigations on
influence of mist cooling using nanofluids on machining parameters in turning AISI 1040 steel.
Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology. 2013
Dec;227(12):1334-46.
[2] Cetin MH, Ozcelik B, Kuram E, Demirbas E. Evaluation of vegetable based cutting fluids with extreme
pressure and cutting parameters in turning of AISI 304L by Taguchi method. Journal of Cleaner
Production. 2011 Dec 31;19(17):2049-56.
[3] Shaji S, Radhakrishnan V. An investigation on solid lubricant moulded grinding wheels. International
Journal of Machine Tools and Manufacture. 2003 Jul 31;43(9):965-72.
[4] Suresh Kumar Reddy N, Venkateswara Rao P. Performance improvement of end milling using graphite as a
solid lubricant. Materials and manufacturing processes. 2005 Jul 1;20(4):673-86.
[5] Dilbag S, Rao PV. Performance improvement of hard turning with solid lubricants. The International
Journal of Advanced Manufacturing Technology. 2008 Aug 1;38(5-6):529-35.
[6] Mukhopadhyay D, Banerjee S, Reddy NS. Investigation to study the applicability of solid lubricant in
turning AISI 1040 steel. Journal of manufacturing science and engineering. 2007 Jun 1;129(3):520-6.
[7] Weinert K, Inasaki I, Sutherland JW, Wakabayashi T. Dry machining and minimum quantity lubrication.
CIRP Annals-Manufacturing Technology. 2004 Jan 1;53(2):511-37.
[8] Xuan Y, Li Q. Heat transfer enhancement of nanofluids. International Journal of heat and fluid flow. 2000
Feb 29;21(1):58-64.
[9] Li CH, Peterson GP. Experimental investigation of temperature and volume fraction variations on the
effective thermal conductivity of nanoparticle suspensions (nanofluids). Journal of Applied Physics. 2006
Apr 15;99(8):084314.
[10] Manimaran R, Palaniradja K, Alagumurthi N, Sendhilnathan S, Hussain J. Preparation and characterization
of copper oxide nanofluid for heat transfer applications. Applied Nanoscience. 2014 Feb 1;4(2):163-7.
[11] Sharma AK, Tiwari AK, Dixit AR. Improved Machining Performance with Nanoparticle Enriched Cutting
Fluids under Minimum Quantity Lubrication (MQL) Technique: A Review. Materials Today: Proceedings.
2015 Jan 1;2(4-5):3545-51. 12. Amrita M, Srikant RR, Sitaramaraju AV. Performance evaluation of
nanographite-based cutting fluid in machining process. Materials and Manufacturing Processes. 2014 May
4;29(5):600-5.
[12] Sharma P, Sidhu BS, Sharma J. Investigation of effects of nanofluids on turning of AISI D2 steel using
minimum quantity lubrication. Journal of cleaner production. 2015 Dec 1;108:72-9.
[13] Khandekar S, Sankar MR, Agnihotri V, Ramkumar J. Nano-cutting fluid for enhancement of metal cutting
performance. Materials and Manufacturing Processes. 2012 Sep 1;27(9):963-7.

202
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Generation of Various Micropatterns by Electrochemical Micromachining

Sandip Kunar1 and B. Bhattacharyya2

Production Engineering Department, Jadavpur University, Kolkata, India
Email: sandip.sandip.kunar@gmail.com1; bb13@rediffmail.com2

Abstract: To fabricate the quality micropatterns at lower cost and short time for industrial applications, a novel
technique i.e. maskless electrochemical micromachining (EMM) is used due to its several advantages such as higher
machining rate, better surface quality, irrespective to material hardness, no thermal effect, etc. The service life and
performance of product depend upon the shape, size, location and area density of micropatterned surfaces. In
maskless EMM method, one masked patterned cathode tool can fabricate more than average twenty five
micropatterned samples with higher machining efficiency. The indigenously developed maskless EMM setup having
EMM cell, tool and workpiece fixture devices, electrical connections and electrolyte circulation systems can
generate better machined samples with better machining accuracy and surface quality. The research article
investigates the effect of major influencing parameters i.e. inter electrode gap and flow rate on machining accuracy,
machining depth and surface finish(Ra) of micro circular patterns produced by this technique. Various types of
micropatterns such as varactor, cascade and square micropatterns were successfully fabricated using NaCl+NaNO3
electrolyte. Experimental investigations recommend lower inter electrode gap and higher flow rate for better
machining accuracy, controlled depth and surface quality.

1. Introduction
Micropatterned surfaces have modified surface morphology in order to get uniform distributed depressions or
asperities with controlled shape and size. Functional micropatterned surfaces concern the significant physical and
chemical phenomena at the micro and nano scale textured surfaces and play a vital role in the advancement of many
engineering fields, such as optics, electronics, information technology, biomimetics, aviation, etc [1]. Modified
active through mask EMM method showed the applicability and machinability of the process for producing micro
dimples with uniform shape and size on the planar and non-planar surfaces [2]. Through-mask EMM was used to
generate micro dimples using pulsed current and the effect of duty cycle and pulse frequency on diameter and depth
of micro-dimple arrays was experimentally investigated. A patterned polydimethylsiloxane (PDMS) mask with
good reusability and flexibility was introduced for fabricating micro-dimple on a cylindrical surface[3]. This
process is expensive and time-consuming for individual masking of workpiece. Sandwich-like electrochemical
micromachining was used to produce uniform and deeper micro dimples using smaller porous metal cathode in
enclosed electrochemical reaction unit. The effect of process parameters such as pore size of metal cathode,
machining voltage and machining time on diameter and machining depth of micro dimples was investigated [4].
Sometimes, the insoluble products and other sludges deteriorate the shape and size of machined samples in enclosed
electrochemical reaction unit. Varactor micropattern was an important microfabricated machined profile generated
by maskless EMM method, which was used in radio frequency communication including voltage controlled
oscillators and filters, microwave applications, etc. The effects of predominant process parameters i.e. duty ratio and
Generation of various micropatterns by electrochemical micromachining
applied voltage on machined surface characteristics i.e. machining accuracy and surface finish of generated varactor
micropattern were investigated[5]. Maskless EMM was used for copper dissolution in acidified and non -acidified
media for fabricating the micropatterns having maximum depth of 4.7μm [6]. The micropatterned depth is less,
which was generated by maskless EMM. Electrochemical etching using laser masking (FELM) was used for laser
masking and anodic dissolution of metal without need of photo-mask for multilayered structure’s generation. A
patterned layer surface was generated by laser marking using a pulsed fiber laser on stainless steel. The patterned
surface was selectively dissolved by electrochemical etching because laser marked area acted as a protective mask
[7]. Sometimes, laser marking comprises inherent limitations such as machining accuracy, surface finish, etc.
A novel concept of maskless EMM has been used to generate different types of micropatterns i.e. circular, varactor,
cascade and square micropatterns with good surface quality and higher productivity. One SU-8 2150 masked
patterned cathode tool can fabricate more than twenty five micropatterned samples. The indigenously developed
maskless EMM setup consisting of EMM cell and other accessories can generate micropatterned samples using
vertical cross flow system with better machining accuracy and surface quality than other flow modes. Maskless
EMM is applied to investigate the influence of machining parameters i.e. inter electrode gap and flow rate on mean
machining accuracy, mean machining depth and mean surface finish during fabrication of different micropatterned
samples. Analysis has been done to acquire the best parameter setting based on micropatterned characteristics of
micro circular pattern.

2. Experimental Planning
Figure 1 shows the maskless EMM experimental setup for generation of different microsurface textured patterns on
stainless steel surfaces (SS-304).

Fig. 1. Developed maskless EMM set-up

Maskless EMM setup was used for generation of different micropatterns on stainless steel surfaces (SS-304). The
experimental setup consists of different accessories i.e. EMM cell, tool and job holding devices, electrical
connections, electrolyte supply systems, etc for carrying out the experimental investigations for generation of
different micropatterned samples. The important feature of electrolyte system is vertical cross flow system, in which

204
 Proceedings of NCAMMM - 2018

flow is parallel between tool and workpiece and flows from downward to upward vertically. This flow maintains the
better surface quality than other modes of electrolytes. The textured patterned tools having diameter of 200μm was
fabricated by UV lithography process on stainless steel wafer bonding with reusable SU-8 2150 mask having
thickness 230µm. To analyze the effect of EMM process parameters on control of the desired performance
characteristics by maskless EMM setup, process parameter combinations were decided properly to obtain good
micropatterned samples. The ranges of IEG and flow rate are 50-200μm and 3.35-6.35 m3/hr respectively. The
constant parameter settings are duty ratio of 30%, pulse frequency of 5kHz, current cut off of 3A, machining voltage
of 10V, electrolyte concentration of NaCl(0.18M/L)+NaNO3(0.26M/L), flow pressure 0.1kg/cm2 and time of 3
minutes. The number of repetition of each experiment is five. The textured characteristics were decided with optical
microscope (Leica DM2500, Germany) and 3D Non-Contact Profilometer.

3. Influence of EMM Parameters on Micropattern Characteristics

Generation of different micropatterns with precise micro-features and particular dimensions by maskless EMM
desires to know proper machining parameter combinations. Hence, to study the micropatterned characteristics,
micropatterns have been fabricated by maskless EMM and effect of process parameters i.e. inter electrode gap and
flow rate on radial overcut, machining depth and surface roughness have been investigated.
To investigate the influence of inter electrode gap on radial overcut, machining depth and surface
roughness of micro circular patterns, other parameters are kept constant i.e. duty ratio of 30%, pulse frequency of
5kHz, machining voltage of 10V, flow velocity of 6.35 m3/hr, electrolyte concentration of
NaCl(0.18M/L)+NaNO3(0.26M/L) and machining time of 3 minutes. As shown in figure 2, the radial overcut
increases with increase in inter electrode gap for micro circular patterns. In lower inter electrode gap, the current
density distribution is more controlled and stray current effect is also less. The generated micro circular patterns are
more uniform in lower inter electrode gap. In higher inter electrode gap, the shape and size of micro circular patterns
are non-uniform in shape and size because the stray current distribution is higher. Very feeble sparking occurs below
50μm gap due to accumulation of electrolysis products and gas bubbles generation. Electrolysis products accumulate
in very lower inter electrode gap and generates short circuit or micro-sparking. The machining depth decreases with
increasing inter electrode gap. In lower inter electrode gap, the depth is controlled because machining localization
increases for uniform current flux distribution. In higher inter electrode gap, the machining depth is uncontrolled and
non-uniform due to lower machining localization. It is also observed that roughness increases with increase in inter
electrode gap. In lower inter electrode gap, roughness is less because controlled machining takes place with reduced
stray current effect. Afterwards, the roughness values increase in higher inter electrode gap due to uncontrolled
etching. For higher machining accuracy, controlled depth and good surface finish, lower inter electrode gap is
recommended. Micro circular patterns have been fabricated by changing electrolyte flow rates with other constant
parameters i.e. inter electrode gap of 50μm, duty ratio of 30%, pulse frequency of 5kHz, machining voltage of 10V,
electrolyte concentration of NaCl(0.18M/L)+NaNO3(0.26M/L) and machining time of 3 minutes as shown in figure
3. Radial overcut decreases with increase in flow rate because the stray current effect is less at higher flow rate due
to quicker removal of sludges from the micromachining zone.

205
Generation of various micropatterns by electrochemical micromachining

Fig. 2. Effect of inter electrode gap on overcut, machining depth and surface roughness

At lower flow rate, the overcut is uncontrolled and non-uinform because lower flow rate performs uncontrolled
machining localization all over the machining zone for improper removal of electrolysis products. The machining
depth increases with increasing flow rate because machining localization increases with higher flow rate for quicker
removal of sludges. In lower flow rate, the depth is lower for micro circular patterns because uncontrolled anodic
dissolution occurs for accumulation of electrolysis products. The surface roughness decreases with increase in flow
rate because because it drags away the metal ions from the grain boundaries as well as from peaks region of the
surface under machining operation. In higher flow rate, the results indicate that increment in flow rate shows
improvement in surface finish. Roughness is higher at higher flow rate because uncontrolled anodic dissolution
takes place. For higher machining accuracy, controlled depth and good surface finish, higher flow rate is
recommended.

Fig. 3. Effect of flow rate on overcut, machining depth and surface roughness

All texturing experiments have been carried out by SU-8 2150 negative photoresist using maskless EMM method.
All masked patterns do not distort during texturing operation and these are used for fabrication of textured samples
more than twenty five times. During experimentation, all masked patterns were observed very carefully that very
little deformation takes place in the masked tools after generation of many samples. Very little deformation has been
ignored during fabrication of micropatterned samples and does not affect the micropatterned samples during

206
 Proceedings of NCAMMM - 2018

fabrication. The masked pattern tool has more utility for generation of micropatterned samples with higher
machining accuracy and surface quality.
Micro circular patterns have been fabricated at a particular parametric machining conditions i.e. inter
electrode gap of 50μm, duty ratio of 30%, pulse frequency of 5kHz, machining voltage of 10V, electrolyte
concentration of NaCl(0.18M/L)+NaNO3(0.26M/L), flow rate of 6.35 m3/hr and machining time of 3 minutes.
Almost all circular patterns maintain the proper and uniform geometrical shape with good surface quality due to
controlled anodic dissolution and uniform current distribution. Very little distortion may not influence the shape of
fabricated textured patterns. Figure 4(a) shows microscopic micro circular pattern with average radial overcut of
23.26μm, average depth of 21.5μm and average surface roughness of 0.0129μm. Figure 4(b) shows 3D view of
circular impression and figure 4(c) shows the depth profile of circular impression having depth of 29.1μm. Figure
4(d) shows the roughness profile of micro circular impression with average surface roughness of 0.0181μm.

Fig. 4. (a) Micro circular pattern (b) 3D view (c) depth profile (d) surface roughness (Ra) profile

4. Fabrication of Various Types of Micropatterns

The fabrication of different micropatterns with better shape and size has a great significant influence in various
advanced applications i.e. biomedical, defence, aerospace, tribological, etc. The research endeavor is continuing for
design and development of new micropattern’s shapes for further improving of application performances.
In these experimental investigations, varactor micropattern was fabricated by UV lithography process on brass wafer
with reusable PMMA (Polymethyl methacrylate) mask having thickness 250µm. The cascade micropatterned tool
having width of 430μm was fabricated by UV lithography process on stainless steel wafer bonding with reusable
SU-8 2150 mask having thickness 230µm. The square pattern having side length of 480μm was fabricated by UV
lithography process on stainless steel wafer bonding with reusable SU-8 2150 mask having thickness 230µm.
Different micropatterned surfaces have been fabricated with parameter combination i.e. inter electrode gap of 50μm,
duty ratio of 30%, pulse frequency of 20kHz, machining voltage of 8V, electrolyte concentration of NaCl (0.18M/L)
+NaNO3 (0.26M/L), flow rate of 6.35 m3/hr and machining time of 30 seconds.
Figure 5(a) shows the microscopic varactor micropattern with average overall overcut of 31.26μm, average depth of
36.1μm and average surface finish (Ra) of 0.0627μm. Figure 5(b) shows the microscopic cascade micropattern with
average width overcut of 32.47μm, average depth of 21.6μm and average surface finish (Ra) of 0.0553μm. Figure
5(c) shows the microscopic square micropattern with average side overcut of 26.29μm, average depth of 8.19μm and
average surface finish (Ra) of 0.060μm. It is also noticed that small values of standard deviation are observed for

207
Generation of various micropatterns by electrochemical micromachining
different micropatterns in overcut, machining depth and surface finish. It is occurred due to uniform current
distribution and good flushing conditions using maskless EMM method.

Fig. 5. Microscopic images of micropatterns fabricated by maskless EMM: (a) varactor micropattern (b) cascade
micropattern (c) square micropattern
5. Conclusions
Micropatterned surfaces generated by maskless EMM method are one of the key micro features in microproducts for
various applications. Machining accuracy and surface quality of textured patterns influence product
functionalization and service life of microproducts. Based on the experimental results, major conclusions can be
drawn as follows:
(i) The micro circular patterns are fabricated with a novel texturing method i.e. maskless EMM method with low
cost and short time. The developed EMM cell is utilized for fabrication of good texturing characteristics with
regular shape and size of textured micropatterns using vertical cross flow electrolyte system. One masked
patterned tool can fabricate more than twenty five machined samples using SU-8 2150 mask and PMMA mask
for different textured patterns.
(ii) Machining with lower inter electrode gap and higher flow rate is recommended for higher machining accuracy,
controlled depth and good surface finish.
(iii) From the analysis of various micrographs for micro circular patterns, it can be concluded that the parametric
setting i.e. inter electrode gap of 50μm, duty ratio of 30%, pulse frequency of 5kHz, machining voltage of 10V,
electrolyte concentration of NaCl (0.18M/L) +NaNO3 (0.26M/L), flow rate of 6.35 m3/hr and machining time of
3 minutes generates the best micro circular pattern.
(iv) Different micropatterns such as varactor, cascade and square micropatterns were successfully fabricated with
higher machining accuracy and good surface quality using maskless EMM method.
Maskless EMM is very helpful for selecting the influencing process parameters for fabricating various micro-
textured patterns of various sizes in different advanced applications e.g. biomedical, defence, aircraft, etc. For better
microtexturing patterns, the design of textured tools and tool movement strategy need to be improved further.

References
[1] Bruzzon AAG, Costa HL, Lonardo PM, Lucca DA. Advances in engineered surfaces for functional
performance. CIRP Annals - Manufacturing Technology. 2008; 57(2):750-769.

208
 Proceedings of NCAMMM - 2018

[2] Ming P, Zhang X, Zhou W, Zhao C, Zhou H, Ge Q. Development of a modified through-mask electrochemical
micromachining for micropatterning nonplanar surface. The International Journal of Advanced Manufacturing
Technology. 2017 Nov; 93(5-8):2613–2623.
[3] Chen X, Qu N, Li H, Xu Z. Pulsed electrochemical micromachining for generating micro-dimple arrays on
a cylindrical surface with a flexible mask. Applied Surface Science. 2015 July 15; 343:141–147.
[4] Zhang X, Qu N, Fang X. Sandwich-like electrochemical micromachining of micro-dimples using a porous
metal cathode. Surface and Coatings Technology. 2017 Feb 15; 311:357–364.
[5] Kunar S, Bhattacharyya B. Micropattern generation using electrochemical micromachining. In: Ahuja BB,
Basu SK, Rajiv B, editors. Proceedings of 6th International & 27th All India Manufacturing Technology, Design
and Research Conference; 2016; Pune. New Delhi: Excel India Publishers; 2016. p. 2092-2096.
[6] Nouraei S, Roy S. Electrochemical process for micropattern transfer without photolithography: A Modeling
Analysis. Journal of The Electrochemical Society. 2008 Feb; 155(2):D97-D103.
[7] Shin HS, Park MS, Chu CN. Electrochemical etching using laser masking for multilayered structures on
stainless steel. CIRP Annals - Manufacturing Technology.2010; 59(1):585–588.

209
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

An Empirical View on Accuracy and Machinability of TiNiCu Shape Memory Alloys

during Wire Electro Discharge Machining

Abhinaba Roy1, Narendranath S.

Department of Mechanical Engineering, National Institute of Technology Karnataka, Surathkal-575025
1
Corresponding author, er.abhi.roy@gmail.com, +91-7022384560

Abstract: TiNiCu shape memory alloys are useful functional materials where depending on the application,
percentage elemental composition of TiNiCu alloy can be varied to provide it with required properties. Wire
electro discharge machining (WEDM) process is most suitable to process this particular group of alloy due to
its thermal nature. Though not determined in the current stage of this investigation, it was found that thermal
conductivity and melting temperature are directly related to the machining response variation exhibited by these
alloys. ANOVA analysis of kerf width has been carried out to figure out the most influential parameter as kerf
width directly influences dimensional accuracy of part produced using WEDM. An empirical variation has been
established in terms of percentage machining accuracy for all the selected alloys using a design of experiment
approach.
Keywords: Wire Electro Discharge Machining, Kerf Width, Material Removal Rate, Surface Roughness

1 Introduction
TiNi based shape memory alloys are well known for their excellent properties over traditional materials.
However, for applications demanding sensitive actuation capabilities, TiNiCu ternary alloy is preferred over
binary TiNi alloy which has a major drawback of large hysteresis. When copper (Cu) is introduced to replace
nickel (Ni) in TiNi alloy system, property of the alloy changes drastically [1]. Addition of copper decreases the
hysteresis response of TiNi shape memory alloy. This results in the decrease of transformation strain, reduction
in pseudoelastic hysteresis and also reduces the sensitivity of the martensitic start temperature to composition.
Among the different compositions of TiNiCu, addition of 5 to 10 at.% Cu is preferred. Although addition of Cu
greater than 10 at.% is most preferred [2], above 10 at.% addition of Cu, the ingot becomes brittle and hence it
exhibits poor machinability which is highly unfavorable for manufacturing sectors. Therefore, in this present
study the copper content is kept constant at 10 at.% for all the three alloys. Conventional machining processes
are not suitable for these alloys as they do not yield required surface finish and dimensional accuracy required in
their applications. Since they are used for their unique mechanical properties, it has to be kept in mind that the
metallurgical properties, dimensional accuracy and the surface finish of the final product has to be as per
requirement. In our previous work, WEDM has been established as the most useful method to perform
machining of TiNi based SMAs [3]. Kerf width, also known as the machining width, determines the accuracy of
the part produced and is indeed a controllable response. Kerf width greatly influences the corner radius during
profile cutting and other geometrical zones that consists of tight angular variation which was the driving force of
the study performed by a group of researchers [4]. Another group of researchers investigated the effect of
WEDM parameters on kerf width and found that pulse on time and wire tension are most significant parameters,
whereas rest of the parameters are irrelevant [5]. Another study emphasized on the wire vibration caused during
machining that effects the kerf width and concluded that low discharge sparks followed by higher wire tension
and small distance between wire guides will restrict wider kerf formation [6]. Since lesser kerf width represents
An empirical view on accuracy and machinability of TiNiCu shape memory alloys during wire electro discharge machining

higher accuracy and most of the other researchers are correlating the process parameters with the kerf width and
other contributing factors, the current investigation deals with variation in kerf width due to change in
percentage elemental composition under similar machining parameters. Ti 50 Ni 40 Cu 10 , Ti 45 Ni 45 Cu 10 and
Ti 40 Ni 50 Cu 10 are the three alloys which are subjected to this study and the variation of kerf width in these alloys
under similar machining condition have been discussed.

2 Material Preparation and Experimentation

In this investigation TiNiCu shape memory alloys were prepared by using Vacuum TIG Arc Melting technique
where obtained melt ingots were melted and remelted six times each to maintain homogeneity. To evaluate the
variation in the response of the selected alloys, four control parameters namely, pulse on time (Ton), pulse off
time (Toff), servo voltage (SV) and wire feed (WF) were considered. Levels of the selected parameters and
output responses are indicated in Table 1 and Table 2 respectively. Taguchi's L16 orthogonal array reduced
number of experimental runs and maximized the apparent effect of machining on the quality of the machined
product. Wire cut EDM (make : Electronica India Ltd.) was used with de-mineralized water as the dielectric
fluid and a 250µm zinc coated brass wire as the wire electrode. Zeiss EVO-18 Scanning Electron Microscope
has been used to measure kerf width. Material removal rate is measured based on the kerf width profile as
shown in Figure 1 and Equation 1. Five measurements were taken along the entire kerf length to obtain a
statistically accurate measurement. Surface roughness of the machined surfaces are noted using Mitutoyo SJ-
301 surface roughness tester.
Table 1 Input parameters and output responses
Levels
Input parameters Unit Symbol
1 2 3 4
Pulse on time µs T on 100 110 120 130
Pulse off time µs T off 20 30 40 50
Servo voltage V SV 15 30 45 60
Wire feed m/min WF 4 6 8 10
Table 2 Output responses
Responses Unit Symbol
Material Removal Rate mm3/min MRR
Kerf Width µm KW
Surface Roughness µm SR

Fig. 1 Kerf width profile

𝜋𝜋 𝑘𝑘 𝑓𝑓 2
[�𝑡𝑡 𝑙𝑙 × 𝑡𝑡 ℎ ×𝑘𝑘 𝑓𝑓 �+ ×� � ×𝑡𝑡 ℎ ]
2 2
𝑀𝑀𝑀𝑀𝑀𝑀 = (𝑚𝑚𝑚𝑚3 /𝑚𝑚𝑚𝑚𝑚𝑚) … … … … Equation 1
𝑇𝑇

Where, t l = travel length (mm), t h = workpiece thickness (mm), k f = kerf width (mm), T = time taken for entire
travel length (minutes)

211
 Proceedings of NCAMMM - 2018

3. Results and Discussion

3.1 ANOVA analysis
In this section, the machining characteristics of the selected alloys based on the input process parameters have been studied using analysis of variance. Table 3 and Table 4
represent F-test and P-test values from ANOVA results obtained from Minitab 17.1.0 for MRR, SR and KW respectively for the selected TiNiCu shape memory alloys. Two
group of researchers have analyzed the experimental data using ANOVA to comprehend the most influencing process parameters that affect the output responses of wire electro
discharge machining [7,8].

Table 3 ANOVA results for MRR and SR

ANOVA results for MRR ANOVA results for SR
Alloy TNC11 Alloy TNC12 Alloy TNC13 Alloy TNC11 Alloy TNC12 Alloy TNC13
Source DF F P %p F P %p F P %p F P %p F P %p F P %p
Ton 3 18.23 0.020 85.31 9.32 0.050 76.32 14.08 0.028 84.04 63.49 0.003 89.86 11.58 0.037 88.88 47.51 0.005 94.32
Toff 3 0.89 0.535 4.19 0.96 0.514 7.85 0.71 0.607 4.24 3.95 0.145 5.58 1.18 0.447 9.08 1.15 0.457 2.27
SV 3 1.27 0.426 5.92 1.02 0.494 8.34 1.20 0.442 7.17 0.56 0.676 0.79 0.04 0.988 0.30 0.85 0.551 1.69
WF 3 0.98 0.506 4.58 0.91 0.528 2.49 0.76 0.585 4.55 2.66 0.222 3.76 0.23 0.874 1.73 0.87 0.545 1.72
Table 4 ANOVA results for KW
Alloy TNC11 Alloy TNC12 Alloy TNC13
Source DF F P %p F P %p F P %p
Ton 3 8.25 0.058 32.07 4.89 0.113 23.75 2.97 0.197 30.08
Toff 3 3.10 0.189 12.07 4.28 0.132 20.79 1.14 0.458 3.84
SV 3 11.14 0.039 43.32 10.20 0.044 49.51 4.30 0.131 43.52
WF 3 3.23 0.181 12.55 1.22 0.436 5.94 1.47 0.379 14.87

Similarly, by the help of ANOVA, the most influencing parameters in the machining of alloy TNC11, TNC12 and TNC13 can be identified as they are our material of interest.
As per the results given in Table 3, pulse on time (T on ) is the most influencing parameter when material removal rate (MRR) is concerned. Percentage contribution (%p) of
pulse on time for alloy TNC11, TNC12 and TNC13 are 85.31 and 5.92, 76.32 respectively. Similarity in behavior for kerf width is also noticed in Table 4. In case of kerf width
(KW), pulse on time and servo voltage are also found as the most influencing parameters as they carry close weightage in P-test results (0.058 and 0.039 for TNC11 for
example). However, servo voltage has more potential to create a wider kerf compared to pulse on time which is evident from low P-test values compared to pulse-on time and
percentage contribution of 43.32%, 49.51% and 43.52% (refer Table 4) for alloy TNC11, TNC12 and TNC13, respectively. However, servo voltage has more potential to create

212
An empirical view on accuracy and machinability of TiNiCu shape memory alloys during wire electro discharge machining

a wider kerf compared to pulse on time which is evident from low P-test values compared to pulse-on time and
percentage contribution of 43.32%, 49.51% and 43.52% (refer Table 4) for alloy TNC11, TNC12 and TNC13,
respectively. Also, for surface roughness pulse on time was found to be most influential parameter. Pulse on
time largely remains as the sole parameter to successfully influence and generate a more even surface showing
very similar results for the alloys TNC11, TNC12 and TNC13 at 89.86%, 88.88% and 94.32% respectively.

3.2 Effect of influential process parameters on kerf width

Fig. 2 Machined at 130 µs, kerf zone of: (a) alloy TNC11; (b) alloy TNC12; (c) alloy TNC13

It can be observed from mean effects plot depicted in Figure 2a that kerf width tends to increase with growing
pulse on duration. Due to higher energy of the discharge spark at high pulse on duration, more material is
removed from the metal matrix leading to wider kerf formation. Secondly, due to lower thermal conductivity of
the alloy TNC11, it was easy for the discharge sparks to melt off more material from its surface compared to
TNC12 and TNC13 and thus kerf width is more in case of TNC11 along the entire range of pulse on duration.
Dielectric fluid jet helps to flush out the molten material during the machining process. It can be noted in Figure
2(c-e) that accumulation of material residue along the kerf boundary is more in case of TNC13 compared to its
counterparts. Since thermal conductivity of TNC13 is more than TNC11 and TNC12, it becomes difficult for the
discharge spark to create necessary thermal energy to dissipate material from the metal matrix. Consecutively,
dielectric fluid jet also struggles to flush out the molten material because it tends to stick faster than it melts in a
given period of time due to faster heat dissipation resulting from higher thermal conductivity of the material.
This in turn leads to lower kerf width of the material having higher thermal conductivity which is TNC13 in this
case. This trend in variation of kerf width is evident from Figure 2(a-b). The variation in behavior can be related
to the findings of a study [9], where the material removal rate is related to the melting temperature and thermal
conductivity of the material. According to their findings, higher material removal rate is due to lower thermal
conductivity of the material which makes it easy to melt off the material from the surface compared to its
counterparts. This can be further justified by the observations noted in a study [10], where it was found that with
increasing Ni content in binary TiNi alloy, the thermal conductivity of the alloys exhibit a periodic rise and drop
which is also applicable for Cu-doped TiNi alloys. Based on their results and observing the behavior of the

213
 Proceedings of NCAMMM - 2018

response of the materials under similar machining condition in the current study, it is clear that thermal
conductivity of TNC11 is least followed by TNC12 and TNC13. Servo voltage considered as one of the major
input parameters in wire electro discharge machining, helps in maintaining an uniform spark gap between the
workpiece and the wire electrode during machining process. As can be seen from mean effects plot of kerf width
depicted in Figure 2b, with increasing servo voltage it gradually decreases. As higher servo voltage tends to
maintain larger spark gap, volume of ionization of dielectric fluid around the machining zone enlarges which
leads to reduced intensity of spark energy that impinges on workpiece surface. Therefore, net melting of
workpiece material reduces and leads to narrower kerf. At low servo voltage, spark gap is minimum which
results in intense spark discharge due to which machined surface suffers heavy degradation which generally
decreases with increasing servo voltage. The obtained results are similar to the findings of a group of researcher
who investigated on kerf width variation using micro-WEDM [6].
3.3 Comparison of machining accuracy based on process parameters
A compact assimilation of kerf width and material removal rate for the selected alloys for Taguchi's L16
orthogonal experiments have been represented in Figure 3.

Fig. 3 Compact bar chart of: (a) kerf width; (b) MRR for L16 experiments
For the ease of understanding, these two histograms are divided into four categories depending on the MRR
value which are category 1, category 2, category 3 and category 4 and tabulated in Table 5. For alloy TNC11,
TNC12 and TNC13, least recorded kerf widths are 297 µm, 300 µm and 297 µm respectively noted at second
experimental run of category 1 for each of the alloys. But the material removal rate for second experimental run
is much lower compared to the entire range of L16 experiments. However, if fifth and seventh experimental runs
of category 2 are compared, it can be seen that seventh experimental run is more suitable for smaller kerf width
even though MRR will be compromised in this particular case and is applicable for all the chosen alloys. In
category 3, eleventh experimental run scores high for machining accuracy whereas it ranks second for MRR
which is a fair rank considering the other three experimental runs. Even though experimental run thirteen and
fourteen exhibit quite smaller kerf and high MRR, due to extreme damage to the material surface these are not
practical options and hence qualify sixteenth experimental run as a more useful option under category 4. The
qualifying set of parameters that succeed to provide a fair machining accuracy, their categories, input process
parameters and percentage accuracy are depicted in Table 5. Along with kerf, operator should also keep in mind
the productivity of the machining process. Therefore, if MRR is too low but kerf is minimum, that should be
reserved for micromachining whereas a little compromise on kerf associated with comparatively better material
removal rate can be used for producing parts that don't require ultra precision and not bound by tight tolerance
limits.
Table 5 Categorical classification of obtained kerf width and suitable experimental runs

214
An empirical view on accuracy and machinability of TiNiCu shape memory alloys during wire electro discharge machining

Suitable
Range of
Category Experimental Run Experimental Process Parameters Percentage Accuracy
MRR(mm3/min)
Run
Ton-100 µs; Toff-30 µs; TNC11-81.2%; TNC12-
1 1 2 3 4 0-1 2
SV-30 V; WF-6 m/min 80.0%; TNC13-81.2%
Ton-110 µs; Toff-40 µs; TNC11-72.4%; TNC12-
2 5 6 7 8 1-4 7
SV-15 V; WF-10 m/min 73.2%; TNC13-75.2%
Ton-120 µs; Toff-40 µs TNC11-72.4%; TNC12-
3 9 10 11 12 4-6 11
SV-30 V; WF-4 m/min 68.4%; TNC13-71.2%
Ton-130 µs; Toff-50 µs TNC11-60.8%; TNC12-
4 13 14 15 16 6-12 16
SV-45 V; WF-4 m/min 63.2%; TNC13-61.2%

4 Conclusions
Experimental investigation of wire electro discharge machining on Ti 50 Ni 40 Cu 10 , Ti 45 Ni 45 Cu 10 and
Ti 40 Ni 50 Cu 10 as cast shape memory alloys has been carried out and following conclusions have been drawn:
• Ti 50 Ni 40 Cu 10 exhibited superior affinity for higher material removal rate and wider kerf formation under
similar machining conditions compared to Ti 45 Ni 45 Cu 10 and Ti 40 Ni 50 Cu 10 which can be attributed to lower
thermal conductivity of Ti 50 Ni 40 Cu 10 .
• As per ANOVA results, T on and SV are proved to be most influential parameters for MRR and KW
respectively. Even though T on has close percentage contribution like SV, based on P-test value SV is a
more dominant factor than T on when KW is concerned. T on was also found to be most influential parameter
to influence SR.
• For a medium-to-average MRR, experimental run 7 and 11 are found to be yielding satisfactory accuracy
for kerf width and suitable for machining with average accuracy applicable for the chosen alloys.

Acknowledgement
The authors would like to express their gratitude to DST-SERB for its funding support under the project bearing
sanction number SB/S3/MMER/0067/2013.

References
[1] A. Ishida, M. Sato, K. Ogawa, Microstructure and shape memory behavior of annealed Ti-36.8 at.% Ni-
11.6 at.% Cu thin film, Mater. Sci. Eng. A. 481–482 (2008) 91–94.
[2] T. Nam, T. Saburi, Y. Kawamura, K. Shimizu, Shape Memory Characteristics Associated with the B2-B19
and B19-B19’ Transformations in a Ti-40Ni-10Cu (at.%) Alloy, Mater. Trans. JIM. 31 (1990) 262–269.
[3] M. Manjaiah, S. Narendranath, S. Basavarajappa, Review on non-conventional machining of shape memory
alloys, Trans. Nonferrous Met. Soc. China (English Ed. 24 (2014) 12–21.
[4] N. Tosun, C. Cogun, G. Tosun, A study on kerf and material removal rate in wire electrical discharge
machining based on Taguchi method, J. Mater. Process. Technol. 152 (2004) 316–322.
[5] A. Shah, N.A. Mufti, D. Rakwal, E. Bamberg, Material Removal Rate , Kerf , and Surface Roughness of
Tungsten Carbide Machined with Wire Electrical Discharge Machining, J. Mater. Eng. Perform. 20 (2011)
71–76.
[6] S. Di, X. Chu, D. Wei, Z. Wang, G. Chi, Y. Liu, Analysis of kerf width in micro-WEDM, Int. J. Mach.
Tools Manuf. 49 (2009) 788–792.
[7] Y.S. Liao, J.T. Huang, H.C. Su, A study on the machining-parameters optimization of wire electrical
discharge machining, J. Mater. Process. Technol. 71 (1997) 487–493.
[8] S. Tilekar, S.S. Das, P.K. Patowari, Process Parameter Optimization of Wire EDM on Aluminum and Mild
Steel by Using Taguchi Method, Procedia Mater. Sci. 5 (2014) 2577–2584.
[9] S.F. Hsieh, S.L. Chen, H.C. Lin, M.H. Lin, S.Y. Chiou, The machining characteristics and shape recovery
ability of Ti-Ni-X (X=Zr, Cr) ternary shape memory alloys using the wire electro-discharge machining, Int.
J. Mach. Tools Manuf. 49 (2009) 509–514.
[10] B. Ramachandran, R.C. Tang, P.C. Chang, Y.K. Kuo, C. Chien, S.K. Wu, Cu-substitution effect on
thermoelectric properties of the TiNi-based shape memory alloys, J. Appl. Phys. 203702 (2013) 203702-1–
7.

215
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Determination of Johnson-Cook Material Model Parameters for Machining Simulations

using Inverse Analysis: A Review

Tarun Kumar Sa, Hemant Gandhia, Chithajalu Kiran Sagarb, and Amrita Priyadrashinic*
a
Student, Mechanical Engineering Department, BITS-Pilani, Hyderabad campus, Telangana-500078, INDIA
b
PhD Scholar, Mechanical Engineering Department, BITS-Pilani, Hyderabad campus, Telangana-500078, cAssitant
Professor, Mechanical Engineering Department, BITS-Pilani, Hyderabad campus, Telangana-500078,
* Corresponding author Tel.:040 66303645.E-mail: amrita@hyderabad.bits-pilani.ac.in

Abstract: Machining plays an important role for transforming the raw materials into finished products. This is the
only process that involves removal of material to get the final shape, size, dimensions and finish. Selection of input
parameters to get optimum outputs during machining is critical and requires exhaustive experimental tests, which
are no doubt expensive. Hence, Finite Element (FE) modeling has come up as one of the best alternative which
helps us to understand the mechanics as well as to select the right kind of cutting parameters with minimum number
of experiments. But FE analysis of the machining process is highly dependent on the constitutive model that is
considered for depicting the material behavior undergoing deformation during machining process. Hence, this
paper aims at providing an overview of different approaches essentially used for identification of parameters of
commonly used constitutive model namely, Johnson Cook Material Model (JCMM). The focus of present paper is to
discuss the capability of inverse analysis for determining JCMM constants so that it can be effectively used for FE
simulations of machining process.
Keywords: Johnson Cook material model, Inverse analysis, Machining, Cutting forces

1. Introduction
Machining is a process that involves removal of material from the surface of workpiece material in order to get final
shape, size, dimensions and finish of finished products. FE modellinghas gained much of the fame in the recent
pastfor modelling and simulating various manufacturing processes including machining. This approach helps in
reducing the number of experimental tests considerably which are nevertheless required for optimisation of the
process; thus, saving much of the material, money, time and effort. However, depicting the material behaviour of
material undergoing plastic deformation accurately during machining process is very crucial for successfully
simulating machining processes. In typical machining processes, the work material undergoes severe deformations
and shearing, with the strains reaching up to 200% and strain rates around 106 /s or more in the shear zone (1,2);
thus making itdifficult to determine the flow stresses of work material under equivalent conditions. These conditions
cannot be replicated by the standardized tensile and compression tests. As a result, flow stress of the work material
determined usingconventional Universal Testing Machine (UTM) become mostly unacceptable to represent the
material behaviour in the above mentioned strains and strain rates. There are sometheoretical models which are
based upon material behaviourat atomic level, but unfortunately, theyare hardly feasible when it comes to material
modelling. Therefore, commonly used material models for describing plastic deformation are semi-empirical
constitutive models. JCMMis one such model which takes into account both stain hardening and strain softening
Determination of Johnson-Cook material model parameters for machining simulations using inverse analysis: a review
phenomena. The present work discusses about the various methods used to predict the JCMM parameters, the
accuracy and the general reliability of the model as well as the corresponding values of the constants.

2. Constitutive Model and Oxley’s Model

The JCMM describes the plastic material behavior at varied range of strain rates and temperatures and thus, taking
into consideration both the strain hardening as well as strain softening. The model is considered to be numerically
robust and is found as an inbuilt model in most of FE software packages (3).The JCMM flow stress is given by(4):
𝜀𝜀̇ 𝑇𝑇−𝑇𝑇𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟
σ = [A + Bε𝑛𝑛 ] [1 + C ln ][1 − ( )𝑚𝑚 ] (1)
ɛ̇ 0 𝑇𝑇𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 −𝑇𝑇𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

where, A is the initial yield strength, B is hardening modulus, n is the strain-hardening exponent, C is strain rate
sensitivity, m is the thermal softening constant, 𝜀𝜀̇ is the reference strain rate, 𝑇𝑇𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚 is melting temperature of
workpiece and 𝑇𝑇𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 is room temperature.
Oxley’s machining model is based on slip-line field theory(5,6). The model determines the outputs, namely, cutting
forces, stresses and average temperature during orthogonal cutting. There are few assumptions based on which the
analysis is done such as plane strain conditions, sharp cutting tools without nose radius and steady state conditions.
While determining the various machining outputs using Oxley’s model, variation of material flow stresses with
strain, strain rate and temperatures are taken into considerations.

3. Methods to Determine JCMM Constants

In the past, various researchers have employed high speed compression tests to estimate the flow stresses of
workpiece materials at relatively higher strains, strain rates and temperatures as compared to that of standard tensile
or compression tests. The maximum strain rates achievable in high compression tests are 450𝑠𝑠 −1 which are much
lower than that expected during machining process. Split Hopkinson Pressure Bar (SHPB) and Taylor’s impact tests
are categorized as high strain rate tests which are generally used to assess the work material flow stress conditions
for simulating the machining simulations. Strain rates as high as 20,000 𝑠𝑠 −1 and temperatures as high as 900 ℃ can
be attained by SHPB tests and strain rates as high as 10,000 𝑠𝑠 −1 can be achieved in Taylor’s impact test(6). Several
researchers have calibrated the material models using the experimental data obtained, especially, from SHPB for
metal cutting simulations(7,8). However, availability of this advanced equipment is very limited because of the
extremely high cost and need of skilled operator.
3.1 Inverse identification of JCMM constants
3.1.1Levenberg–Marquardt Algorithm(9) - Levenberg–Marquardt Algorithm was first published in 1944 by Kenneth
Levenberg and then rediscovered by Donald Marquardt in 1963(9). This algorithm is used to solve non-linear least
square problems. Being a minimization algorithm, it is used to find local minimum of a non-linear function. In
inverse identification problems, it helps to find the error between the outcomes of the FE simulations carried out
using the familiar parameter set and test simulations carried out to recalculate the initial parameter set. Since it is an
iterative process it helps to minimize both the cutting forces and the non-overlapping area of the chip shapes. The
weighted sum of the cutting force difference and the non-overlapping area forms the aggregate error function
𝜖𝜖𝑖𝑖 (x)(9).

𝜖𝜖𝑖𝑖 (x) = |𝑒𝑒𝑖𝑖𝐴𝐴 (𝑥𝑥)| + w.|𝑒𝑒𝑖𝑖𝐹𝐹 (𝑥𝑥)| (2)

217
 Proceedings of NCAMMM - 2018

𝑁𝑁 𝑁𝑁
1 1
𝜒𝜒 2 (𝑥𝑥) = � 𝜖𝜖𝑖𝑖 2 (𝑥𝑥) = �(�𝑒𝑒𝑖𝑖𝐴𝐴 (𝑥𝑥)� + w. |𝑒𝑒𝑖𝑖𝐹𝐹 (𝑥𝑥)|)2 (3)
2 2
𝑖𝑖=1 𝑖𝑖=1
Shrot and Baker(4)developed an adiabatic 2D FE model of the machining process using commercial FE software
ABAQUS. The model was divided into three parts such that the top section, middle section and bottom section
depicting chip surface, sacrificial layer and machined surface, respectively and then an ideal simulation was carried
out in order to recalculate JCMM parameters(10). The simulation is carried out until the error function cannot be
minimized below a predefined correctness or a maximum number of iterations are attained such that(10)
𝜕𝜕𝜖𝜖 𝑖𝑖 𝜖𝜖 𝑖𝑖 �𝑥𝑥+𝛿𝛿 𝑗𝑗 𝑒𝑒 𝑗𝑗 �−𝜖𝜖 𝑖𝑖 (𝑥𝑥)
(𝑥𝑥) = . (4)
𝜕𝜕𝑥𝑥 𝑗𝑗 𝛿𝛿 𝑗𝑗

Two different starting points and three parameters (A, B, n) were used to evaluate the outcome. The first case was
close to the standard set so the algorithm converged in 6 iterations while the second set was far from the standard set
but the algorithm converged in just 4 iterations. The stress-strain curves were within 8% of each other and the
cutting forces also matched very well. After successful identification with three parameters, four parameters were
used with a starting set far from the standard set and the convergence occurred in just eight iterations with cutting
forces error of approximately 10%. The authors suggested that the solution can be improved either by using the
Nelder-Mead algorithm or by using multi stage optimization strategy(10).
3.1.2 Kalman filter (11) - The JCMM parameters have been identified using Kalman filter which is an inverse
analysis method, developed by Rudolf Kalman (11).The approach can be used to estimate the JCMM parameters
based on two experimental cutting parameters. The unidentified constants are represented in a vector form and
assigned some initial values at time t=0. The equation used for iteration is as follows(11):

𝑥𝑥𝑡𝑡 = 𝑥𝑥𝑡𝑡−1 + 𝐾𝐾𝑡𝑡 [𝑑𝑑𝐸𝐸𝐸𝐸𝐸𝐸 − 𝑑𝑑𝑡𝑡 (𝑥𝑥𝑡𝑡−1 )] (5)

such that 𝑥𝑥𝑡𝑡 = (𝐴𝐴𝑡𝑡 , 𝐵𝐵𝑡𝑡 , 𝐶𝐶𝑡𝑡 , 𝑛𝑛𝑡𝑡 , 𝑚𝑚𝑡𝑡 )𝑇𝑇 at any instant , 𝑥𝑥𝑜𝑜 = (𝐴𝐴𝑜𝑜 , 𝐵𝐵𝑜𝑜 , 𝐶𝐶𝑜𝑜 , 𝑛𝑛𝑜𝑜 , 𝑚𝑚𝑜𝑜 )𝑇𝑇 for initial estimates, 𝐾𝐾𝑡𝑡 is the
𝐸𝐸𝐸𝐸𝐸𝐸
𝐹𝐹𝑐𝑐
Kalman gain matrix, 𝑑𝑑𝐸𝐸𝐸𝐸𝐸𝐸 = � 𝐸𝐸𝐸𝐸𝐸𝐸 � is experimentally determined cutting process variables in the vector form such
𝜆𝜆 ℎ
𝐸𝐸𝐸𝐸𝐸𝐸 𝐸𝐸𝐸𝐸𝐸𝐸
that𝐹𝐹𝑐𝑐 is the cutting force and 𝜆𝜆ℎ is the chip compression ratio (CCR).
𝐹𝐹𝑐𝑐𝑡𝑡−1
𝑑𝑑𝑡𝑡 (𝑥𝑥𝑡𝑡−1 )=� 𝜆𝜆 𝑡𝑡−1
�is the cutting parameter calculated from the preceding iteration where 𝐹𝐹𝑐𝑐𝑡𝑡−1 is the primary cutting
ℎ

force and 𝜆𝜆𝑡𝑡−1

ℎ is the chip compression ratio.
The Kalman gain matrix is computed as follows(11):
𝜕𝜕𝒅𝒅𝑡𝑡−1 𝑇𝑇 −1
𝐾𝐾𝑡𝑡 = 𝑃𝑃𝑡𝑡 ( ) 𝑅𝑅𝑡𝑡 (6)
𝜕𝜕𝑥𝑥 𝑡𝑡−1
𝜕𝜕𝒅𝒅𝑡𝑡−1 𝑇𝑇 𝜕𝜕𝒅𝒅𝑡𝑡−1 𝜕𝜕𝒅𝒅𝑡𝑡−1 𝑇𝑇 𝜕𝜕𝒅𝒅𝑡𝑡−1 (7)
𝑃𝑃𝑡𝑡 = 𝑃𝑃𝑡𝑡−1 − 𝑃𝑃𝑡𝑡−1 � � ( 𝑃𝑃𝑡𝑡−1 ( ) + 𝑅𝑅𝑡𝑡 )−1 𝑃𝑃
𝜕𝜕𝑥𝑥𝑡𝑡−1 𝜕𝜕𝑥𝑥𝑡𝑡−1 𝜕𝜕𝑥𝑥𝑡𝑡−1 𝜕𝜕𝑥𝑥𝑡𝑡−1 𝑡𝑡−1
To modify the JCMM parameters, Kalman gain matrix is then multiplied with the difference of the experimentally
measured machining parameters with that of the updated ones. Rows of Kalman gain matrix consist of five JCMM
parameters and the column consists of two cutting process parameters.𝑃𝑃𝑡𝑡 , a simulation covariance matrix, which is
associated with the range of unidentified JCMM parameters at time t is revised at each step.𝑅𝑅𝑡𝑡 , an error covariance
matrix, is fixed at the start of the iteration. Kalman algorithm changes with these two matrices, hence they should be
assigned properly(11).

218
Determination of Johnson-Cook material model parameters for machining simulations using inverse analysis: a review

(∆A) 2 0 0 0 0 
 
 0 (∆B) 2 0 0 0 
𝐹𝐹 2 0
𝑃𝑃0 =  0 0 (∆C ) 2
0 0  𝑅𝑅𝑡𝑡 = � 𝑐𝑐 � (8)
  0 𝜆𝜆2ℎ
 0 0 0 (∆n) 2 0 
 0 (∆m) 2 
 0 0 0
where 𝑃𝑃0 is initial simulation covariance matrix and(∆𝐴𝐴)2 ,(∆𝐵𝐵)2 ...(∆𝑚𝑚)2 are the predicted values of the unidentified
JCMM parameters and 𝑅𝑅𝑡𝑡 is error covariance matrix with diagonal elements as cutting parameters.
The authors(11)carried out the above mentioned algorithm in python by establishing a relation between JCMM
constants and cutting parameters from the available literature (12).An error percentage of 7% was obtained between
the experimental obtained values of cutting forces and chip thickness with that of the predicted values which was
further improved to 2% using the Kalman filter inverse analysis by the author. The method proved to be cost
effective and reliable to find the JCMM parameters. Figure 1(a) and (b) shows the variation in cutting force and
CCR, respectively.

(a) (b)
Figure 1. Graphs showing the difference in (a) cutting force (F c ) and (b) CCR for the JCMM parameters(11)

3.1.3 Determination of JCMM constants using descriptors and proxies(13) - This method uses descriptors, to predict
the material parameters. The descriptors are intermediate quantities that are used to establish a relation between
experimentally found quantities and material parameters. In addition, this technique needs comparatively lesser
number of FE simulations to accurately predict the material parameters(4). Generally, in most of the inverse
identification approaches, material constants are determined from the experiments. However, Baker(2,13)has used
simulations to set the target values. Following are the steps involved:
• Determine the target values for the proxies, namely, cutting forces and shear angles using FE simulations on the
basis of given target parameters.
• Assume initial values of material constants i.e. JCMM constant based on literature available.
• Develop an FE model by taking the assumed values of JCMM constant as inputs for material model and run the
simulations.
• Compute proxies and descriptors and stop iterating if the convergence is achieved.
• Compute new values of descriptor values based on either a linear fit or simple scaling
• Compute corresponding JCMM constants i.e. the updated JCMM constants for the new descriptors.

Three different sets of target parameters were considered by taking four different starting values such that each

219
 Proceedings of NCAMMM - 2018

combination are simulated three times using the developed FE model. The values of mean cutting force as well as
shear angle were computed for each case, which was used for the optimization. The root mean square error between
the predicted values and the target values was the indicator of the quality of the parameter set. Convergence was
assumed when the RMS error was less than 3%. Although most predictions achieved the 3% error, they were unable
to reduce the error value to less than 1-2%.
3.1.4 Extended Oxley’s shear zone theory(14) - This method attempts to formulate an inverse analysis approach to
determine the JCMM parameters using orthogonal cutting tests(6,14). The basic idea is to reduce any kind of errors
which are likely to happen while using analytical models. Firstly, cutting forces, chip geometry (cut chip thickness)
and temperatures were measured experimentally during the cutting tests. The measure values of chip geometry are
used for calculating strain and strain rates with the help of Oxley’s theory based on parallel sided shear zone theory
which are further used to determine the values of cutting forces. The cutting forces obtained using Oxley’s model is
compared with the experimentally measured values of cutting forces. Then, based on the obtained data set, JCMM
parameters are rectified by adjusting the parameters of the model so that the values fit to the obtained flow stress
conditions that are strain, strain rate and temperature dependent. The optimization routine focuses on minimizing the
error of the normalized resultant cutting forces using the least squared method such that root of sum of squares of
each cutting force component yields resultant force. The experimental data for the cutting forces have been taken
from the literature sources (15,16,17). Similarly, Buzyurkin et al.(18) used the JCMM constants A, B and n from the
results obtained from one of the literature(19)and predicted C and m by performing FE simulations based on full
factorial analysis considering all the combinations of input factors. It has been observed that a nonlinear relationship
exists between strain, strain rate and temperature and that of the cutting parameters.
3.1.5 Response Surface Methodology (RSM) based approach to identify JCMM constants (20) - In this approach,
RSM has been used to enhance the inverse methods based on orthogonal cutting tests that are already available in
literature (20,21). High strength aluminium alloyAl6061-T6 is machined using uncoated carbide insert with a back
rake angle (22).Cutting forces and chip thickness were measured experimentally using a three-component Quartz
dynamometer and a digital micrometer, respectively which were considered as input data. While, the outputs
namely, flow stresses, equivalent plastic strain, equivalent strain rate, and the machining temperature were
calculated using Oxley’s shear zone theory(23). The JCMM has been used without any modifications. A nonlinear
regression analysis based on the interior point algorithm was employed to determine the constitutive model
constants. The RSM technique is used in such a way that large number of cutting parameters can be considered
during the optimization routine. Central Composite Design (CCD) is considered while conducting the cutting
experiments. A second-order model has been developed on the basis of RSM and CCD with 95% confidence
level(22):
𝐾𝐾 𝐾𝐾 𝐾𝐾

𝑦𝑦 = 𝛽𝛽0 + � 𝛽𝛽𝑖𝑖 𝑥𝑥𝑖𝑖 + � 𝛽𝛽𝑖𝑖𝑖𝑖 𝑥𝑥𝑖𝑖 2 + � 𝛽𝛽𝑖𝑖𝑖𝑖 𝑥𝑥𝑖𝑖 𝑥𝑥𝑗𝑗 + 𝜖𝜖 (9)

𝑖𝑖=1 𝑖𝑖=1 𝑖𝑖<𝑗𝑗

where, y: response and :ith and jth values of machining process variables, 𝛽𝛽0 , 𝛽𝛽𝑖𝑖 , 𝛽𝛽𝑖𝑖𝑖𝑖 , 𝛽𝛽𝑖𝑖𝑖𝑖 : regression coefficients and 𝜖𝜖:
experimental error of the observations. Following equation states the response in terms of machining variables,
namely, α: rake angle, V: cutting speed and f:feed rate(22):

220
Determination of Johnson-Cook material model parameters for machining simulations using inverse analysis: a review

𝑦𝑦 = 𝛽𝛽0 + 𝛽𝛽1 𝑉𝑉 + 𝛽𝛽2 𝑓𝑓 + 𝛽𝛽3 α + 𝛽𝛽4 𝑉𝑉 2 + 𝛽𝛽5 𝑓𝑓 2 + 𝛽𝛽6 α2 + 𝛽𝛽7 𝑉𝑉𝑉𝑉 + 𝛽𝛽8 𝑉𝑉α + 𝛽𝛽9 𝑓𝑓α (10)
The material constants would be obtained using the solution from the nonlinear regression.
Table 1 summarizes the parameter values of JCMM obtained using various techniques as discussed in the present
work. It can be observed that relative average error percentage for various inverse analysis techniques varies in the
range of 10-16 % which is lower than that of the error percentage found in case of SHPB test.

Table 1. JCMM constants obtained by different inverse analysis method for different work materials

Experimental A B
Method Material n C m Ref. % Error
Conditions (MPa) (MPa)
SHPB HY-100 steel Cutting speed = 1.778 316 1067 0.107 0.027 0.7 Shi et al(7) 19.7
m/s, mm, rake angle =
50, clearance angle =
60.
FEM orthogonal Carbon Steel Cutting speed = 30 152.5 350 0.20 - - Shrot et al 10
simulation m/s, feed= 0.2 mm (6)
Levenberg– AISI 52100 Cutting speed = 33.3 635.9 101.7 0.64 - 2.26 Shrot et al 8
Marquardt m/s, feed= 0.19 mm, (10)
Algorithm using rake angle = 00.
orthogonal turning
model
Kalman filter AISI 4140 Rake angle = 50, 595 580 0.13 0.023 1.03 Agmell et 1.6
using orthogonal Cutting speed = 260 al.(11)
cutting model m/min.
descriptors and Hyperfoam Cutting Speed = 20 300 100 -0.1 0.001 - Baker et 15
proxies using press mm/s al.(13)
forge model
Extended Oxley’s AISI 1045 Rake angle = -50, 553.1 600.8 0.23 0.013 1 Lalwani et 15
shear zone theory Cutting speed = 100 al.(14)
m/min
Response surface Inconel 718 Rake angle =00, feed = 1562 300 0.25 0.016 1.7 Malakizadi 16
method 0.05,0.1, depth of cut = et al.(20)
1 mm

4. Conclusion
The current work demonstrates that identifying the JCMM parameters under strain rates and temperatures, as
expected during machining process, is critical and challenging but not impossible and can be found out with better
greater accuracy if basic understanding of the mechanics of machining processes is known clearly. As far as the
inverse analysis algorithms are concerned, it cannot be concluded that a particular method is better than the other for
identifying the material parameters since the authors have performed the experiments under different cutting
conditions(speed, feed, rake angle, etc.). However, it can be inferred that there is a significant decrease in the error
values in determining the JCMM constants by using the algorithms based on inverse analysis.
References
[1] Baker M. A new method to determine material parameters from machining simulations using inverse identification. Procedia
CIRP. 2015; 31: p. 399 – 404.
[2] Baker M. The influence of plastic properties on chip formation. Computational Materials Science. 2003; 28: p. 556–562.
[3] ShrotA, Bäker M. How To Identify Johnson-Cook Parameters From Machining Simulations. In AIP Conference
Proceedings; 2011.
[4] ShrotA, Bäker M. Determination of Johnson–Cook parameters from machining simulations. Computational Materials
Science. 2012; 52: p. 298–304.
[5] Oxley P, Louis B. The mechanics of machining: an analytical approach to assessing machinability. illustrated ed.: Ellis
Horwood series in mechanical engineering; 1989.

221
 Proceedings of NCAMMM - 2018

[6] ShrotA, Baker M. A study of non-uniqueness during the inverse identification of material parameters. Procedia CIRP 1.
2012;: p. 72 – 77.
[7] CJ Shi, LiuR. The Influence of Material Models on Finite Element Simulation of Machining. Journal of Manufacturing
Science and Engineering. 2004 February; 126(4): p. 849-857.
[8] JVusal, RElchin. Influence of Material Models Used in Finite Element Modeling on Cutting Forces in Machining. In
Materials Science and Engineering; 2016: IOP Publishing.
[9] WMDonald. An Algorithm for Least-squares Estimation of nonlinear parameters. Journal of the Society for Industrial and
Applied Mathematics. 1963 June; 11(2): p. 431-441.
[10] ŠlaisM, DohnalI, ForejtM. Determination of Johnson-Cook Equation Parameters. Acta Metallurgica Slovaca. 2012
september 12; 18: p. 125-132.
[11] AgmellM, AAylin, StåhlJan-Eric. Identification of plasticity constants from orthogonal cutting and inverse analysis.
Mechanics of Materials. 2014; 77: p. 43–51.
[12] AgmellM, AylinA, StåhlJan-Eric. A numerical and experimental investigation of the deformation zones and the
corresponding cutting forces in orthogonal cutting. Advanced Materials Research. 2011;(223): p. 152-161.
[13] BäkerM,ShrotA. Inverse parameter identification with finite element simulations using knowledge-based descriptors.
Computational Materials Science. 2013; 69: p. 128–136.
[14] LalwaniD I, MehtaN K, JainP K. Extension of Oxley’s predictive machining theory for Johnson and Cook flow stress
model. Journal of Materials Processing Technology. 2009; 209: p. 5305–5312.
[15] IvesterR W, KennedyM, Davies M, StevensonR, General Motors, Warren. Assessment Of Machining Models: Progress
Report. Machining Science and Technology. 2000; 4: p. 511–538.
[16] Iqbal S A, MativengaP T, and SheikhM A. Characterization of machining of AISI 1045 steel over a wide range of cutting
speeds. Part 1: investigation of contact phenomena. Proceedings of the Instituition of Mechanical Engineers, Part B: Journal
of Engineering Manufacture. 2007 May 1; 221(5): p. 909-916.
[17] Iqbal S A , MativengaP T, SheikhM A. Characterization of machining of AISI 1045 steel over a wide range of cutting
speeds. Part 2: evaluation of flow stress models and interface friction of flow stress models and interface friction of flow
stress models and interface friction distribution sc. Proceedings of the Institution of Mechanical Engineers, Part B: Journal
of Engineering Manufacture. 2007 Jan; 221(5): p. 917-926.
[18] BuzyurkinaA E., GladkyI L, KrausE I. Determination and verification of Johnson–Cook model parameters athigh-speed
deformation of titanium alloys. Aerospace Science and Technology. 2015; 45: p. 121–127.
[19] AbouridouaneM, KlockeF, LungD, VeselovacD. The mechanics of cutting: In-situ measurement and modelling. In Procedia
CIRP 31; 2015. p. 246 – 251.
[20] MalakizadiA, CedergrenS, SadikI, NyborgL. Inverse identification of flow stress in metal cutting process using Response
Surface Methodology. Simulation Modelling Practice and Theory. 2016; 60: p. 40–53.
[21] ZhangaY, Outeiro J C, MabroukiT. On the selection of Johnson-Cook constitutive model parameters for Ti-6Al-4V using
three types of numerical models of orthogonal cutting. In Procedia CIRP 31; 2015. p. 112 – 117.
[22] DaoudM, JomaaW, Chatelain J F, BouzidA. A machining-based methodology to identify material constitutive law for finite
element simulation. The International Journal of Advanced Manufacturing Technology. 2015 April; 77(9-12): p. 2019–
2033.
[23] Zerilli FJ. Dislocation mechanics-based constitutive equations. Metallurgical and Materials Transactions A. 2004; 35(9): pp
2547–2555.
[24] DabboussiW, NemesJ A. Modeling of ductile fracture using the dynamic punch test. International Journal of Mechanical
Sciences. 2005 August; 47(8): p. 1282-1299.
[25] GuoY B. An integral method to determine the mechanical behavior of materials in metal cutting. Journal of Materials
Processing Technology. 2003 November 10; 142(10): p. 72-81.
[26] BoldyrevI S, ShchurovI A, NikonovA V. Numerical Simulation of the Aluminum 6061-T6 Cutting and the Effect of the
Constitutive Material Model and Failure Criteria on Cutting Forces’ Prediction. Procedia Engineering. 2016; 150: p. 866 –
870.

222
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Effective Study on MRR Based on Different Process Parameters in Micro-Electrical

Discharge Machining

Arjita Das1,2, Sucharita Saha1,2, Sourav Halder1, Kalyan Chatterjee1, Nagahanumaiah1,2

1
Micro Systems Technology Laboratory, CSIR-Central Mechanical Engineering Research Institute, Durgapur, India
2
Academy of Scientific and Innovative Research, New Delhi, India.

Abstract: Micro- Electric Discharge Machining (Micro-EDM) application is rapidly growing in manufacturing of
metal products having micrometric geometric features. However in this process, low energy short-pulsed electric
discharge material removal is slow, enhancing material removal rate (MRR) becomes a significant factor to use it
in batch scale production of micro products. In this paper an attempt has been made to enhance MRR by using
Graphene as an additive in dielectric liquid. In this experimental study Graphene powder of 25µm thickness
particles are suspended in hydrocarbon dielectric liquid; after 15 minutes of sonication it was used as dielectric in
machining of INCONEL 718 samples. The tungsten carbide electrodes of 400µm diameter were used as cathode
and the work samples were immersed inside the dielectric, where craters were machined 10sec in each set of
experiments. In this study, it is noted that the MRR has been increased with average of 33% in volume of material
removal with the use of Graphene additives. Further, the parametric study experiments indicated that pulse-on-
time has significant influence on MRR.
Keywords: Micro EDM, MRR, TWR, surface roughness, dielectric, powder mix, spark gap.

1. Introduction
Micro EDM has emerged to be the most promising manufacturing processes for batch production of micro parts. In
this non-conventional process high frequency pulses of low energy discharge (~150µJ) between the tiny electrodes
in a spark gap of <10µm is used to remove the material to create micro-scale craters on the samples. Combining
CNC controlled movement of tool/work pieces (Micro-EDM milling) is capable of fabricating complex
topographies on any conductive or semi conductive materials irrespective of their hardness and its high aspect ratio.
This machining is mainly used to produce micro-dies, moulds, surgical tools, ultra fashioned jewelleries, and
finishing parts for aerospace and automotive industry.
The demand in Micro-EDM process is to enhance the remove the material rate (MRR) with good surface
finish and less tool wear rate (TWR). However, micro-EDM process is stochastic in nature; there are several process
parameters such as discharge voltage, spark gap, discharge energy, pulse duration, duty cycle including the
breakdown mechanism of dielectric liquid play significant role on its performance. Moreover, the gap phenomenon
and the material removal mechanism in short-pulsed low energy electric discharge are still not understood to a great
extent [1-2] While, MRR is an important response it largely determines the efficiency and cost effectiveness of the
process of EDM, still there is lack of enough scientific evidence due to the uncertain possibilities of discharge
mechanism. Apart from the basic studies, which are focused on understanding the mechanism fundamentally,
researchers have also explored several techniques to enhance MRR. Rotation of tool electrode, reversing the polarity
of the electrodes, optimization of process parameters, and use of additives in dielectric are few important techniques
Effective study on MRR based on different process parameters in micro-electrical discharge machining

to mention [3-4]. Nagahanumaiah et al have studied the plasma characteristic in micro-EDM and reported that the
plasma is cold and dense plasma [5]. Singh and Ghosh et al. [6] states that the removal of material from the
electrode to the presence of an electrical force on the surface of the electrode that would be able to mechanically
remove material and create the craters. Roy et al. reported debris analysis performed in micro-EDM confirming the
influence of non-thermal forces, due to low temperature generated at discharge spot, although material removal
mechanism is not fully understood [7]. Erden et. al [10] proposed that material removal mechanism relating to three
phases of sparking: dielectric breakdown, discharge and erosion. Even reversing the polarity the material removes
with an appreciable amount [9]. It has been reported that the surface material of eroded electrode differ considerably
from initial one as it consists of dielectric pyrolysis products and an alloy between matrix and electrode. The work
piece material may be diffuse into electrode surface and influence its wear resistance [22-23]. Several studies have
reported the influence of performed parametric studies and optimized the process conditions for to have desired
performance. Jahan et al. investigated the performance of micro-EDM with WC electrodes with producing quality
micro-holes by both transistor- and RC-type generators [11]. Son et al. discussed about the pulse duration, on
material removal rate and machining accuracy [12]. Yan et al. described the characteristics of micro-hole of carbide
by using copper tool electrode and investigated the effects various machining parameters [13]. However, in
published literature, influence of dielectric liquid and its breakdown mechanism in micro-EDM has not been studied
in detail. It is reported that in the absence of dielectric the material MRR is reduced in machining of Ti based
composites [ 8]. On the other hand, use of additives in dielectric has been used as one of the possible approach for
enhancing MRR in conventional EDM. [14-17].
However, to the best authors knowledge, there is no systematic study reported the use of grapheme as an
additive in dielectric for micro-EDM, that has been investigated in this paper. This present work reports the use of
two states of di-electric- simple hydro carbon oil and Graphene powder mix in hydrocarbon oil.

2. Materials and Methods

In this experimental study micro-EDM setup built by the authors shown in Figure 1 was used. In this setup, micro
position linear stage (Holmarc make- Model No. TSV 75 Mu01-01) having 1µm travel resolution were used for X
and Y movement of work piece and Z-travel of tool electrode. Tungsten carbide electrode of 400 µm was used to
create the craters on the INCONEL-718 work sample immersed inside the dielectric added with Graphene nano
particles. In this experimental study, 67.8 mg of graphene powder of 25µm thickness particle size were mixed in
60ml of dielectric liquid hydrocarbon oil. This mixture is sonicated for 15 minutes for better mixing prior to use this
suspension in micro-EDM dielectric tank.
The RC-circuit based controller had been developed in-house to regulate the pulsed electric discharge/ A
555 timer is used to generate the high frequency square wave, which is fed to the comparator (LM 358) for amplifying
the signal. Further, the amplified output signal of comparator is used as gate signal in the MOSFET, which is used for
switching purpose. The drain voltage (input) of the MOSFET is the supply voltage. The output of MOSFET is fed to
the capacitor, which is finally connected to the electrodes. The tool electrode is connected to the negative pole of the
capacitor and the work piece is connected to the positive pole. The digital phosphor oscilloscope in DPO7104 has
observed the feedback of the discharges and the resultant breakdown voltage has been calculated as follows.

224
 Proceedings of NCAMMM - 2018

Figure 1: Micro-EDM Experimental setup

Figure 2: Circuit diagram

The experiments were conducted using L-9 orthogonal array. The supply voltage and pulse-on-time were varied at
three levels. In order to get the desired pulse-on-time the capacitor and resistances (R 1 and R 2 ) were varied offline.
However, the capacitor at the discharge circuit was kept constant at 1000µF, thereby by changing the supply voltage
variable discharge energy was obtained. Similarly, by controlling the pulse-on-time keeping constant duty cycle the
pulse frequencies were controlled. All the experiments were conducted for 10 seconds time. The weight of the work
piece has been measured prior to the experiment (wb) and then after 10 seconds of experiments i.e. weight after
machining (wa) were measured . The MRR in volume of material removal per second is calculated as follows.

(wb −wa )
MRR =
ρ∗T

Where, wb= weight of the work piece before experiment; wa= weight of the work piece after experiment; 𝝆𝝆 =
density of the material; T= Time taken for 10 seconds

3. Experimental Results
The experimental table consists of both the process parameters and performance parameters. The process parameters
voltage, capacitance, and Time on are optimised by the Taguchi method and the performance parameter measured is
material removal rate. The experimental MRR value is calculated in plain Hydrocarbon oil once and secondly

225
Effective study on MRR based on different process parameters in micro-electrical discharge machining

graphene mixed hydrocarbon oil. This MRR is further related to discharge voltage i.e. the observable parameter. The
experimental table is given below.
Table 1: Experimental table

Ex Volt Pulse – on – time controlled by Material Removal Rate (mm3/s) Discharge Voltage(V)
p. age varying Capacitance, R 1 and R 2 (as measured in
No. (V) Osiloscope)
Capaci R1 Ω R2 Ω T on With Without Enhance With Without
tor (µ-sec) Graphe additive ment in Graphe additive
(µF) (Calcul additive % additive
ated)
1 20 .0047 3.9K 14.7K 60 0.00273 0.00225 21.33 14 6
2 20 .047 590 2.2K 90 0.00252 0.00196 28.57 10 7
3 20 .47 68 301 120 0.00383 0.00312 22.75 18 5
4 25 .0047 5.7K 22K 90 0.00262 0.00183 43.16 11 7
5 25 .047 680 3.01K 120 0.00215 0.00159 33.22 7 7
6 25 .47 39 147 60 0.00309 0.00237 30.38 13 8
7 30 0047 6.8K 30.1K 120 0.00268 0.00201 33.33 11 8
8 30 .047 390 1.47 60 0.00293 0.00217 35.02 14 9
9 30 .47 57 220 90 0.00243 0.00166 46.38 8 11

Figure 3: Discharge curve of Micro-EDM

In this experiment, a fine Graphene powder of 25µm thickness is suspended in the dielectric fluid. The added
powder improves the breakdown characteristics of the dielectric fluid. The insulating strength of the dielectric
fluid decreases and as a result, the spark gap distance increases, enlarged spark gap distance makes the flushing
of debris uniform. This results in much stable process thereby improving material removal rate and surface
finish. Figure (4) given below show the principle of powder mixed EDM, The various additives mixed in the
working fluid were Al, Cr, Cu and SiC. It was found that the concentration, size, density, electrical resistivity
and thermal conductivity of powders significantly affect the machining performance. Addition of appropriate
amount of powders to the dielectric fluid resulted in increased MRR with decreased TWR and required surface
finish.

226
 Proceedings of NCAMMM - 2018

Figure 4: powder mix with di-electric

4. Conclusion
Reviewing from the experiments, it has been concluded that the material removal rate is more with the additive mix
graphene powder in di-electric medium as compare to simple di-electric liquid. It is evident that there is great
influence on MRR of different dielectrics, whether it is pure or additive-mixed dielectrics in micro-EDM. The
resultant performance measures such as material removal rate, tool wear rate, overcut, and microhole accuracy
during micromachining on any conducting materials. The conclusions that can be drawn from the experimental
results obtained during this research study of microhole machining on Inconel 718 by employing two state
dielectrics.
(i) The MRR is comparatively more in the powder mix di-electric than the simple di-electric by an average of
32.9%.
(ii) It has been noted that the discharge voltage is directly proportional to material removal rate by relation.

Although, from prior research studies, it was concluded that MRR was high with powder mix, but the present work
with graphene powder as additive significantly presented better result related to discharge voltage.

References
[1]. Mahendran S, Devarajan R, Nagarajan T, Majdi A. A review of micro-EDM. InProceedings of the international multi
conference of engineers and computer scientists 2010 Mar 17 (Vol. 2).
[2]. Pham DT, Dimov SS, Bigot S, Ivanov A, Popov K. Micro-EDM—recent developments and research issues. Journal of
Materials Processing Technology. 2004 Jun 10;149(1):50-7.
[3]. Sundaram MM, Pavalarajan GB, Rajurkar KP. A study on process parameters of ultrasonic assisted micro EDM based on
Taguchi method. Journal of Materials Engineering and Performance. 2008 Apr 1;17(2):210-5.
[4]. A.B Chounde, M. M pawar. Study on pulsed DC power supply Parameters for Micro-EDM. International Journal of
Advanced Research in Electrical,Electronics and Instrumentation Engineering. 2014.
[5]. Nagahanumaiah, Ramkumar J, Glumac N, Kapoor SG, DeVor RE. Characterization of plasma in micro-EDM discharge
using optical spectroscopy. Journal of Manufacturing Processes. 2009 Jul 31;11(2):82-7.
[6]. Singh A, Ghosh A. A thermo-electric model of material removal during electric discharge machining. International
Journal of Machine Tools and Manufacture. 1999 Apr 30;39(4):669-82.

227
Effective study on MRR based on different process parameters in micro-electrical discharge machining

[7]. Roy S, Nagahanumaiah, Roy H, Ghosh D. Characterization of Debris Produced in µ-EDM Process. Multi- Material micro
manufacture. 2009 Sep.
[8]. Paul G, Roy S, Sarkar S, Hanumaiah N, Mitra S. Investigations on influence of process variables on crater dimensions in
micro-EDM of γ-titanium aluminide alloy in dry and oil dielectric media. The International Journal of Advanced
Manufacturing Technology. 2013 Mar 1:1-9.
[9]. Gangadhar A, Shunmugam MS, Philip PK. Pulse train studies in EDM with controlled pulse relaxation. International
Journal of Machine Tools and Manufacture. 1992 Oct 1;32(5):651-7.
[10]. Erden A. Effect of materials on the mechanism of electric discharge machining(E. D. M.). ASME, Transactions, Journal
of Engineering Materials and Technology. 1983 Apr 1;105:132-8.
[11]. Jahan MP, Wong YS, Rahman M. A study on the quality micro-hole machining of tungsten carbide by micro-EDM
process using transistor and RC-type pulse generator. Journal of materials processing technology. 2009 Feb
19;209(4):1706-16.
[12]. Son S, Lim H, Kumar AS, Rahman M. Influences of pulsed power condition on the machining properties in micro EDM.
Journal of Materials Processing Technology. 2007 Jul 23;190(1):73-6.
[13]. Yan BH, Huang FY, Chow HM, Tsai JY. Micro-hole machining of carbide by electric discharge machining. Journal of
Materials Processing Technology. 1999 Mar 15;87(1):139-45.
[14]. Kansal HK, Singh S, Kumar P. Technology and research developments in powder mixed electric discharge machining
(PMEDM). Journal of Materials Processing Technology. 2007 Apr 12;184(1):32-41.
[15]. Kibria G, Sarkar BR, Pradhan BB, Bhattacharyya B. Comparative study of different dielectrics for micro-EDM
performance during microhole machining of Ti-6Al-4V alloy. The International Journal of Advanced Manufacturing
Technology. 2010 May 1;48(5):557-70.
[16]. Talla G, Gangopadhayay S, Biswas CK. State of the art in powder-mixed electric discharge machining: A review.
Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture. 2017
Dec;231(14):2511-26.
[17]. Yeo SH, Tan PC, Kurnia W. Effects of powder additives suspended in dielectric on crater characteristics for micro
electrical discharge machining. Journal of Micromechanics and Microengineering. 2007 Oct 8;17(11):N91.
[18]. Behrens A, Ginzel J. Neuro-fuzzy process control system for sinking EDM. Journal of Manufacturing processes. 2003 Jan
1;5(1):33-9
[19]. Pawade MM, Banwait SS. A brief review of die sinking electrical discharging machining process towards automation.
American Journal of Mechanical Engineering. 2013 Jan 23;1(2):43-9.
[20]. Liaoyuan Z, Tingting L, Fanjiang M, Zhenming J, Ruiqi J. Research on the Key Technology of Micro EDM Motion
Control System.
[21]. Yeo SH, Aligiri E, Tan PC, Zarepour H. A new pulse discriminating system for micro-EDM. Materials and
Manufacturing Processes. 2009 Dec 21;24(12):1297-305.
[22]. Soni JS, Chakraverti G. Cracking behavior of high carbon high chromium tool steel duringEDM-a study. InProc. Conf.
14th AIMTDR, IIT, Bombay 1990 (p. 263).
[23]. Soni JS, Chakraverti G. Production of surface finish and accuracies with rotating electrode in electro-discharge
machining. InProc. Conf. 14th AIMTDR, IIT Bombay 1990 (Vol. 297).
[24]. Kunieda M, Lauwers B, Rajurkar KP, Schumacher BM. Advancing EDM through fundamental insight into the process.
CIRP Annals-Manufacturing Technology. 2005 Jan 1;54(2):64-87.

228
Sub - theme

Forming, Welding
and Additive
Manufacturing
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Simulation of Deep Drawing Deformation Behaviour under Simple Loading Path: A

Study

A. K. Ranaa,*, A. K. Singha, A. Maitraa, B. Mukherjeea, A.Bhuiyaa, S.Dattaa

a
Mechanical Engineering Department, Academy of Technology, Aedconagar, Hooghly- 712121, W.B, India
*
Email of corresponding author: amit.rana@aot.edu.in

Abstract: Deep drawing is a significant metal forming process which is used in sheet metal forming
operations. By this process complex shapes can be manufactured with significant accuracy. In this study, the
prediction of deformation behaviour of aluminium during an axi-symmetric cup drawing operation is studied
numerically. For this purpose, 3D Finite Element (FE) simulation of a simple cup drawing process has been
conducted using ABAQUS software. The simulation is done by considering the isotropic hardening and von
Mises stress criteria. As the deep drawing process commences, the variation of generated von Misses stresses
and plastic equivalent strain are observed at different stages of deformation and the change in strain energy
generation of aluminium sheet with time has been analysed.

Key words: deep drawing process, finite element simulation, strain energy

1. Introduction
Now a days, there is a great concern about weight reduction for further reduction in fuel consumption of
automobile and hence less emission of air pollutants. The characteristic properties of aluminium, high strength
stiffness to weight ratio, good formability and good corrosion resistance potential make it the ideal candidate to
replace heavier materials (steel or copper) in the car to respond to the weight reduction demand within the
automobile industry. Aluminium alloy sheets are being widely employed in making components for automobile
and ship building due to their excellent properties like high specific strength, corrosion resistance
and weldability. Although cast aluminium alloys are being employed for a considerable number of
components, the use of forming products of aluminium alloy sheets is still limited because the formability
of aluminium alloy sheet is still poor due to lack of understanding of flow behaviour during deformation.
Many studies have been carried out on numerical simulations of deep drawing process. For example, Meinders
et al. [1] investigated the behaviour of tailored blanks during deep drawing using the finite element code.
Authors simulated the deep drawing of two products using Tailored Blanks and correlated the experimental and
simulation results. Takuda et al. [2] investigated the deformation behaviour and the temperature change in deep
drawing of an aluminium alloy sheet and successfully predicted the forming limit and necking site by the
simulation. Yoshihara et al. [3] investigated deep-drawing process of a circular cup using magnesium alloy
material and found that wall thickness of the drawn cup depends on the initial BHF value. Padmanabhan et al.
[4] evaluated the orientation of blank sheets rolling direction during deep drawing process and the effect of
anisotropy in the tailor-welded blank and noticed that the punch force for deep drawing increases with
anisotropy in the blank sheets. Jawad and Mohamed’s [5] studied the effects of varying punch nose radius used
in the deep drawing process by carrying out the numerical simulation process in commercially finite element
program code. Compared the experimental work with numerical results, it was concluded that large frictional
force is applied to the metal by the edge of the punch, but not by its flat section. Also the work required to form
a part with large nose radius is more than that required to form a part with small punch nose radius. Hama et al.
Simulation of deep drawing deformation behaviour under simple loading path: a study

[6] studied the tool modelling accuracy on a square-cup deep-drawing operation using finite element simulation
and quadratic parametric surfaces proposed by Nagata patch. It was presented that the total number of tool
elements can be reduced to about 10% of the polyhedral model. Rodrigues et al. [7] determined a multi-step
analysis for admissible blank-holder forces in deep drawing process for stamping friction stir welded tailored
blanks and the formability behaviour of similar and dissimilar combinations of AA 5182-H111 and AA 6016-T4
aluminum alloys were previewed. Using the theoretical stress based criteria; analytical FLDs were plotted and
compared the formability limits with the principal strains in the cup walls obtained by the numerical simulation,
thereby concluded that the procedure can be used to determine the maximum BHF for different tailor welded
blanks. Arab and Javadimanesh [8] examined the deep drawing process of axi-symmetric cylindrical cup for the
prediction of critical thickness. From the above studies, it is observed that the essentiality of the proper selection
of process parameters to the deep drawing process is very crucial. The process parameters selection is also
depends on the base material. It is essential to study the process numerically before experimentation. The
objective of the present FE study is to analyse the deformation during deep drawing of Aluminium sheet with
specific dimensions of die, holder, punch, punch velocity and aluminium sheet.

2. Finite Element Simulation

The simplest deep drawing operation for the production of cylindrical cup from a flat circular blank of sheet
metal is conducted by Abaqus FE simulation. The parameters that affect the deep drawing process such as sheet
thickness, mass density, and coefficient of frictions and punch velocity are considered. The dimensions of the
deep drawing forming tools are described in Fig. 1.Only one quarter of the global circular shape is modeled, due
to geometrical symmetry. The dimensions of the die (Fig.1 (a)) are: radius of the support plate = 50 mm, radius
of the hollow cylinder = 22 mm, fillet radius = 6.5 mm, height of the die = 30 mm. The punch (Fig.1 (b))
dimensions are: radius of the punch = 20 mm, height of the punch = 25 mm and fillet radius of the punch = 5
mm. Punch velocity is 200 mm/s. Release velocity of punch remains same and release velocity of holder is 50%
of punch velocity. The downward progression of the punch has been termed as load; while the release of both
the punch and the holder has been termed as unload. The dimensions of the holder (Fig.1 (c)) are as follows:
radius of the support plate = 50 mm, radius of the hollow cylinder = 22 mm, height of the holder = 10 mm, fillet
radius = 6.5 mm. The dimensions of the sheet (Fig.1 (d)) are: radius of the sheet = 40 mm, thickness of the sheet
= 1.1 mm. The mass density is 2.7E-006 kg/mm3. The elastic properties, (i.e. Young’s modulus = 70000 MPa
and Poisson’s ratio = 0.33) and plastic properties are assigned. The specifications of the plastic properties of
material are yield stress (σ 0 = 35 MPa), strain hardening exponent (n=0.275) and hardening coefficient
(k=112MPa).The movement for die and holder are completely restricted. The punch and sheet are allowed to
move along Y axis. The punch, holder (pressure plate) and the die (mould) are assumed as discrete rigid
materials with shell feature in the model. The plastic flow a property of the material is calculated using the
described procedure. It is used as an input for the FE simulation analysis. The flow stress σ at any plastic strain
εp is the summation of yield stress σ 0 and plastic part of the stress value. The quasi-static plastic stress–strain
behaviour can be represented by equation (1).

σ = σ 0 + k (ε p ) n (1)

230
 Proceedings of NCAMMM - 2018

a b

c d

Fig.1. Detailed part dimensions (a) die, (b) punch, (c) holder and (d) sheet

3. Results and Discussion

In this numerical simulation the von Messes criteria for failure and isotropic hardening have been considered.
Zhang et al.[9] established a consistent relationship between the stress and plastic strain components for the
deep drawing application. Isotropic hardening can be considered in a plasticity model if the progress of the yield
surface is in such a way that at any state of hardening, with no translation, it relative to a uniform expansion of
the initial yield surface. Eight node linear brick, reduced integration, hourglass control elements were chosen for
meshing the sheet to capture the sheet deformation behaviour.In the simulation the sheet has been considered as
a deformable structure while the punch which interacts with the sheet is not.

Fig. 2. Section of deep drawing set for simulations

231
Simulation of deep drawing deformation behaviour under simple loading path: a study

a b

Fig. 3. Distribution of equivalent von Mises stress (a) and equivalent plastic strain (b) for different locations of
the sheet at step time (load) = 3.2000E-02

Deep drawing process has performed, using the analysis results that had obtained by the finite elements
simulation modelled. The assembly of die, punch, holder and sheet with mesh is illustrated in Fig. 2. After the
commencement of deep drawing process at step time (load) 3.2000E-02 (Fig.3.), various deformation zones
were observed. Figure 3(a) shows that during deep drawing the sheet periphery will be more stressed compared
to the centre region. Jawad and Mohamed [5] had also observed the same deformation nature. Authors had
concluded that this occurred as that large frictional force is applied to the sheet by the edge of the punch but not
by its flat section. The neck region develops maximum amount of plastic equivalent stress and strain as shown
in fig.3

a b

Fig. 4. Distribution of equivalent von Mises stress (a) and plastic strain equivalent (b) at step time (load):
4.0000E-02

At step time (load) 4.0000E-02, maximum region of the sheet shows high stress (Fig.4 (a)). Figure 4 (b) shows
that maximum equivalent plastic strain has been developed at neck region. If the deep drawing process is further
continued, there are chances where the material might fail. As the deep drawing process further progresses, the
sheet starts to show wrinkles which are one of the primary defects in deep drawing operations. Figure 5
illustrates the phenomenon of wrinkling. If any lateral pressure of is not applied, the compressive hoop stress
can cause the flange to fold or wrinkle. The analysis was done for time period of 0.04 (load) and 0.02 (unload).
If the understated parameters can be changed properly, then the chances of wrinkling can be reduced. The
parameters are (1) friction between the blank, blank holder, punch and die cavity, (2) sheet shape and thickness
and (3) punch speed. If the time period is increased to 0.06 (load) and 0.04 (unload) certain changes were
observed. The progression limit (Fig 6) of the punch increases giving more drawability. As the drawability
increases the effect of wrinkling also increases which leads to tearing. Therefore, FE simulation is revealed the
failure mode of wrinkle. In this study, determination of strain localization is very much helpful to understand the
early stage of failure/fracture. It is found that the local failure mode is strongly related to the stress state in the

232
 Proceedings of NCAMMM - 2018

sheet material.

a b

Fig. 5. Distribution of equivalent stress (a) and plastic strain equivalent (b) at step time (unload) = 2.0000E-02

a b

Fig. 6. At the tearing due to wrinkling: von Mises stress (a) and equivalent plastic strain (b) at time period 0.06
(load) and 0.04 (unload)

Fig. 7. Results of strain energy Vs time at 0.04 Fig. 8. Results of strain energy Vs time for 0.06
(load) and 0.02 (unload) (load) and 0.04 (unload)

From Fig. 7 it was obtained that wrinkling was initiated at 0.056 sec time when the strain energy recorded was
29×10-3J. Maximum wrinkling was observed at a step time of 2.0000E-02 (unload). Figure 8 also shows the
variation strain energy with time and it agrees to the result as obtained in Fig. 7 except there is a change in load
and unloads time. From the Fig. 8 it was found that wrinkling was initiated at 0.074 sec time when the strain
energy recorded was 32×10-3 J. The initial increase in strain energy shows that the material which is initially in
contact during this time period has crystallographic restriction of slip, thus the material in contact starts
absorbing kinetic energy of the punch. After an increase in strain energy there is a point where it attains a
maximum value. This indicates the maximum energy absorption capability of the sheet metal before any
crystallographic restriction of slip. Then there is a certain drop in strain energy which indicates the time period
during which the material no longer absorbs energy. The drop in this energy can be attributed to the
crystallographic restriction of slip taking place internally. The increase in strain energy again after the certain
drop shows that new unstrained molecules take part in absorption of rest part of the energy. The last portion of

233
Simulation of deep drawing deformation behaviour under simple loading path: a study

the curve shows a constant behaviour that indicates the limitation of the unstrained molecules to absorb more
energy.

4. Conclusion
Finite element model is adopted to predict the complex deformation behavior, strain localization. The von Mises
plastic stress and equivalent plastic strain increases as the sheet is continuously deep drawn. The maximum
stress is developed at the bending region (neck). At first, there is an increase in strain energy absorption. When
wrinkling starts there will be decrease in strain energy absorption. One should consider the maximum amount of
strain energy absorption before occurrence of wrinkling and crystallographic restriction of slip while using
aluminium in different applications. According to the need of the application, friction between the blank, blank
holder, punch and die cavity, sheet shape and thickness, punch speed can be manipulated to get more strain
energy absorption capability of aluminium. Thus this finite element simulation study provides a cost and time
effective method of determining the forming conditions, blank holder force and forming limitations during the
deep drawing of aluminium sheet without conducting experiment sets each time.

Reference
[1] Meinders T, Van den Berg A, Huetink J. Deep drawing simulations of tailored blanks and experimental
verification. Journal of Materials Processing Technology. 2000 Jun 1; 103(1):65-73.
[2] Takuda H, Mori K, Masuda I, Abe Y, Matsuo M. Finite element simulation of warm deep drawing of
aluminium alloy sheet when accounting for heat conduction. Journal of Materials Processing Technology.
2002 Jan 15; 120(1):412-8.

[3] Yoshihara S, Manabe KI, Nishimura H. Effect of blank holder force control in deep-drawing process of
magnesium alloy sheet. Journal of Materials Processing Technology. 2005 Dec 30;170(3):579-85.

[4] Padmanabhan R, Baptista AJ, Oliveira MC, Menezes LF. Effect of anisotropy on the deep-drawing of mild
steel and dual-phase steel tailor-welded blanks. Journal of Materials Processing Technology. 2007 Apr 12;
184(1):288-93.

[5] Jawad WK, Mohamed JH. Studying the effect of punch nose radius on deep drawing operation. Eng.
&Tech. 2008; 26(1):55-73.

[6] Hama T, Takamura M, Makinouchi A, Teodosiu C, Takuda H. Effect of Tool-Modeling Accuracy on

Square-Cup Deep-Drawing Simulation. Materials transactions. 2008 Feb 1;49(2):278-83.

[7] Rodrigues DM, Leitao C, Menezes LF. A multi-step analysis for determining admissible blank-holder
forces in deep-drawing operations. Materials & Design. 2010 Mar 31;31(3):1475-81.

[8] Arab N, Javadimanesh A. Theoretical and experimental analysis of deep drawing cylindrical cup. Journal
of Minerals and Materials characterization and engineering. 2013 Nov 26;1(06):336-42.

[9] Zhang Y, Zhang Q, Qin X, Sun Y. A Consistent Relationship between the Stress and Plastic Strain
Components and Its Application in Deep Drawing Process. Mathematical Problems in Engineering. 2017
Jan 11; 2017;1-6.

234
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Thermal and Mechanical Response in FSSW of Sandwich Sheets at Different Dwell

Periods

Pritam Kumar Rana1, R. Ganesh Narayanan1*, Satish V Kailas2

1
Department of Mechanical Engineering, Indian Institute of Technology Guwahati – 781039, India
r.pritam@iitg.ernet.in, ganu@iitg.ernet.in (* corresponding author)
2
Department of Mechanical Engineering, Indian Institute of Science Bengaluru – 560 012, India
satvk@iisc.ac.in

Abstract: Lightweight sandwich sheets have shown extensive use in automobile industries for making body
parts. There are limited joining techniques available for sandwich sheets due to variation in its chemical and
physical properties. Friction Stir Spot Welding (FSSW) is proposed to join AA5052-H32/HDPE/AA5052-H32
sandwich composite in the present work. The thermal and load response during FSSW is investigated. Effect of
dwell period on the mechanical performance on the joint produced is also evaluated by conducting lap shear
test. The sandwich sheet is prepared by placing High Density Polyethylene (HDPE) sheet in between two
AA5052-H32 sheets. Two different levels of dwell period are chosen for the preliminary study. The joint
behaviour in lap shear test is correlated with the temperature profile and load response of the composite during
joining. The joint behaviour of sandwich sheet is also compared with that of the bimetallic sheet in order to
understand the effect of polymer on the strength of the joint. Bimetallic joints are prepared by FSSW of two
overlapping sheets of AA5052-H32. The axial load signifies the energy input during the process. The
temperature developed depends upon the energy input. The temperature field around the joint affect the stirring
of material and decides the quality of the joint. The results obtained suggest that the dwell period does not affect
the mechanical performance in lap shear loading. However, it is possible to observe the effect in some dwell
period between or beyond this range. Bimetallic sheet performance is found to perform better than sandwich
sheet.
Keywords: Friction stir spot welding, sandwich sheet, temperature profile, load response, dwell time

1 Introduction
In recent years, the application of polymer core sandwich sheet in automobile market is increasing for fulfilling
the high demand of light weight vehicle. It is mainly used in bonnet, top floor panels and outer panels(1). The
advantage of using sandwich sheet is, it can combine the property of many materials such as low density, high
bending resistance, energy absorption, ability to make smoother aerodynamic surface, better formability, reduce
noise and vibration (2–4). A sandwich can be defined as a composite which constitutes of two faces, or skins,
separated by and linked to a core which is less stiff and less dense (5). The joining of sandwich sheet is very
difficult due to large difference in the physical and chemical properties of core and skin material. Any type of
fusion welding is not suitable for sandwich sheet because the solubility of metal in polymer is very poor and
also the polymer gets heavily degraded before metal melt or plasticise (6). Few attempts to join sandwich sheet
by fusion welding suggest that it is very difficult to avoid polymer degradation (7,8).Joining of sandwich sheet
is also attempted by self-pierce riveting (SPR), where several problems were encountered such as difficulty in
constraining polymer inside the composite and cavity formation around the rivet (9).
Thermal and mechanical response in FSSW of sandwich sheets at different dwell periods

In the present work the joining of sandwich sheet is attempted by friction stir spot welding (FSSW), which
is a solid state welding process. It is believed that some problems like polymer degradation, cavity formation,
weight addition etc. can be avoided using this process for welding. Though there are many FSSW parameters in
the present analysis, the effect of dwell time in the joint behaviour is attempted for understanding. In FSSW, the
dwell time is the duration of stirring phase during which the tool rotates in the work-piece without axial
movement (10). As the dwell time changes, the actual heating period changes. With increasing dwell time, the
accumulated heat inside the joint give rise in temperature. The high temperature coupled with intense stirring
changes the microstructure by dynamic recrystalization which eventually changes the mechanical performance
of the joint. FSSW is a process variant of friction stir welding (FSW) process, where a specially designed
rotating tool is plunged in to overlapping sheets at a single point and is retracted from the joint after the process
completion (11).

2 Experimental Procedure
In this section, the technique used for welding, specimen preparation, load response measurement,
temperature evaluation and mechanical testing of the joint are discussed.

2.1 Sample preparation

In the present work, 2 mm thick AA5052-H32 alloy sheet and 1 mm thick HDPE sheets are used to prepare the
specimen. The sandwich sheet is fabricated by placing HDPE sheet in between two AA5052-H32 sheets, while
the bimetallic sheet is prepared by placing two AA5052-H32 sheets in overlapping condition. No adhesive is
used at any interface in order to avoid the effect of adhesive on the joint behaviour. The properties of AA5052-
H32 sheet is evaluated in the rolling direction by standard tensile testing. The yield strength, ultimate tensile
strength, total elongation, strength co-efficient, and strain hardening exponent of AA5052-H32 are respectively
found to be 155±1 MPa, 215±1 MPa, 9±1 %, 0.16, and 356±3 MPa. The tensile test of HDPE sheet is also
conducted and found that the yielding has started from beginning and the sheet is not failing but continuously
elongating after attaining peak stress. The ultimate tensile strength and uniform elongation of HDPE sheet is
found to be obtained 29±0.5 MPa and 17.4±2.4 MPa, respectively. The metallic sheets are sheared in leg
shearing machine and polymer sheets in hand shearing machine into desired dimension. Two types of samples
are prepared both for sandwich and bimetallic sheets- specimen for load and temperature measurement, and
specimen for lap shear test. The specimen for load and temperature measurement is prepared by cutting the
metallic and polymer sheets of size 180 mm × 60 mm. The dimensions of the specimen for lap shear test are
schematically illustrated in Fig. 1.

Fig. 1 Specimen for lap shear test

236
 Proceedings of NCAMMM - 2018

2.2 Joint preparation

A three axis vertical type FSW machine (Make: ETA Technologies) to produce all FSSW joints. All the
surfaces of metallic sheets are perfectly cleaned by acetone to remove dirt and grease. The polymer sheet is
cleaned by soap and dried completely. All the required pieces of specimens are placed on the bed of the machine
in overlapping condition. The specimen is perfectly clamped so that there is no movement in any direction. Two
different parameter set are chosen for the joining as listed in Table 1. Only the dwell time is varied from 5 sec.
to 20 sec. This range of dwell time is chosen based on the literatures available for FSSW of aluminium and
FSSW of polymer, since no work is done to see the effect of dwell time on FSSW of sandwich sheet before.
Zhang et al. (12)have varied dwell time from 5 sec. to 15 sec. in FSSW of 5052-H112 aluminium alloy sheets.
Bilici and Yukler(13)have used dwell time in the range of 8-90 s, where after 30 sec., no change in joint strength
is observed. Based on the literature, some intermediate range of dwell time is selected for FSSW of sandwich
sheet. Other parameters like rotational speed, plunge depth, and plunge speed are kept constant to avoid
compounding effect. To keep consistency in the results, same parameter set is used for bimetallic sheet also. A
straight cylindrical pin with flat shoulder tool is used to fabricate FSSW joints. The shoulder diameter, pin
diameter and pin length are 10 mm, 4 mm and 3mm respectively.

Table 1 FSSW process parameters used in the present work

Parameter Rotational speed Tool plunge depth Tool plunge speed Dwell time
set (rpm) (mm) (mm/min.) (Sec.)
1 1600 3.6 8 05
2 1600 3.6 8 20
2.3 Temperature evaluation
To investigate the behaviour of polymer inside the sandwich sheet, it is more important to know the temperature
inside the joint rather than temperature on the surface. The FSSW is done at the edge of the overlapping sheets.
It is done in such a way that half diameter of the tool remains always outside of the specimen. The temperature
is measured using six numbers of K-type thermocouples. The thermocouples are attached to the edge of the
specimens at a distance 4 mm, 6 mm, and 8 mm away from the tool axis on the upper sheet and 0 mm, 2 mm, 4
mm away from the tool axis on the lower sheet. Though the temperature generated inside the joint in actual
FSSW is larger than the FSSW at the edge, it is believed that a qualitative analysis can be achieved using this
procedure on the dwell time effect.

2.4 Load response measurement

The load response is measured by the load cell attached in the machine. Six different indexes are being
measured continuously with elapsed time during the process. These are rotational speed, tool axial movement,
torque, axial load, transverse load and tool movement in transverse direction. The variation in all the indexes are
plotted with respect to time to see the effect of process parameters and the composite system.

2.5 Mechanical performance

The mechanical performance of the joint produced is evaluated by conducting lap shear test on a hydraulically
controlled universal testing machine (Model: UTE 20, Make: FIE). The crosshead speed is kept at 1 mm/min
and the test is carried out at room temperature. The load-extension behaviour of sandwich and bimetallic FSSW
are compared at two dwell times.

237
Thermal and mechanical response in FSSW of sandwich sheets at different dwell periods

3 Results and Discussion

In this section, variation in the load and torque during FSSW of sandwich and bimetallic sheets is studied. The
axial load variation with time during dwell period is observed. The effect of dwell period on temperature
distribution across the weld is also investigated. Finally, the joint performance during lap shear test at two dwell
periods is compared and correlated.

3.1 Temperature distribution

The temperature profile around the joint at six different locations for sandwich and bimetallic sheets welded at
two different dwell period is shown in Fig. 2. The temperature on the upper sheet is always higher than the
lower sheet irrespective of the dwell time and the material combination. This is due to larger contact area
between tool shoulder and upper sheet in comparison to lesser contact area between tool pin and lower sheet.
There are two temperature peaks obtained during heating. The first peak occurs few seconds after tool pin
touches the upper sheet, while the second peak occurs after shoulder touches the upper surface. A small decrease
in temperature after first peak is due to reduction in friction between tool and material at plasticised state. The
temperature near the keyhole (4 mm from tool axis on the upper sheet)is largest and decreases in outward
direction from the tool axis.
Higher temperature is recorded at higher dwell period both in sandwich and bimetallic sheet. This is due to
larger heating time. It is observed from the temperature profile that rate of heating is larger than the rate of
cooling. So, the heat generation will always be greater than the heat dissipation. Hence, larger heating period
results in larger heat generation which gives rise in the temperature.
The peak temperature in sandwich sheet is lesser than the bimetallic sheet at both the dwell periods. This is
due to smaller contact area between tool and metallic portion in sandwich sheet. Further a part of generated heat
might be consumed during melting of polymer.

Fig. 2 Temperature distribution around the weld at six different locations for sandwich sheet welded at dwell
period of (a) 5 Sec, (b) 20 sec. and bimetallic sheet at dwell period of (c) 05 sec. and (d) 20 sec

238
 Proceedings of NCAMMM - 2018

3.2 Load response

A typical variation in the axial load, transverse load, and torque during FSSW of bimetallic sheet at 5 sec. dwell
time is shown in Fig. 3a. Local fluctuation in data can be ignored. The transverse load and the torque value are
not changing much once plunging starts. Major variation occurs in the axial load. Since, the present work is
focussed on the effect of dwell time, the axial load response at changing dwell period is considered for further
study. The axial load response during dwell period for sandwich and bimetallic sheet at two different dwell
periods is shown in Fig. 3b. For initial 4 – 5 sec., the load drops from its maximum and remains almost constant
for rest of the period in all cases. This initial load drop occurs due to decrease in flow stress at elevated
temperature. So, there is no effect of dwell period on the axial load variation neither on sandwich sheet nor on
bimetallic sheet. However, there is a difference between the load response of sandwich sheet and bimetallic
sheet. Initial load requirement of sandwich sheet is lesser but load drop during dwell period is also lesser than
the bimetallic sheet. This results in large load requirement in sandwich sheet especially at larger dwell period.

Fig. 3 Load, torque and tool axial displacement variation (a) of bimetallic FSSW for 5 sec. dwell time, (b) axial
load evolution during dwell period

3.3 Mechanical performance

In FSSW, the joint strength is expressed in terms of failure load because the exact load bearing area is unknown.
Finding joint efficiency is not practical in FSSW. The load – extension behaviour of sandwich and bimetallic
sheet in lap shear test at two dwell period is shown in Fig. 4. The maximum load and extension at the two dwell
periods are almost the same for sandwich and bimetallic system. However, this does not confirm that the FSSW
joints are always unaffected with dwell time. It is possible to observe the effect in some intermediate dwell
period between 5 sec. and 20 sec. or beyond this range. The lap shear strength and the joint ductility of sandwich
sheet is lesser than the bimetallic sheet at any dwell time. It is believed that the lesser peak temperature resulted
in reduced stirring in sandwich sheet.

Fig. 4 Load – extension behaviour of sandwich and bimetallic FSSW at two different dwell time

239
 Thermal and mechanical response in FSSW of sandwich sheets at different dwell periods

4. Conclusions
The sandwich and bimetallic sheets are successfully joined by FSSW process considering two dwell times 5 sec.
and 20 sec. The temperature profile, load requirement during FSSW, and lap shear test results are compared for
analyses. Following conclusions are made from the results:
1. Higher temperature is developed at higher dwell period for both sandwich and bimetallic sheet systems. This
is likely due to larger heating rate than heat dissipation rate. Further, lesser peak temperature developed for
sandwich sheet than bimetallic sheet due to smaller frictional contact area and heat flux contribution for
polymer fusion.
2. During FSSW, initial load drop from peak value is observed during dwell period due to reduction in flow
stress at high temperature. With increasing dwell time, the average load required in axial direction is
independent of dwell period. The load drop in sandwich sheet is lesser than the bimetallic sheet. The larger
peak temperature for bimetallic sheet would result in larger drop in axial load.
3. The lap shear tensile load of the joint is not affected by the dwell time in both the material combinations. In
comparison, the bimetallic sheet shows larger lap shear load than sandwich sheet. It is believed that higher
temperature developed in bimetallic sheet resulted in more intense stirring yielding better joint performance.

Acknowledgement: Authors from IIT Guwahati thank CIF, IIT Guwahati for permitting them to use the UTM
for mechanical tests

References
[1] Burchitz I, Boesenkool R, van der Zwaag S, Tassoul M. Highlights of designing with Hylite - A new
material concept. Mater Des. 2005;26(4):271–9.
[2] Librescu L, Hause T. Recent developments in the modeling and behavior of advanced sandwich
constructions: A survey. Compos Struct. 2000;48(1):1–17.
[3] Kim KJ, Kim D, Choi SH, Chung K, Shin KS, Barlat F, et al. Formability of
AA5182/polypropylene/AA5182 sandwich sheets. J Mater Process Technol. 2003;139(1–3 SPEC):1–7.
[4] Rao MD. Recent applications of viscoelastic damping for noise control in automobiles and commercial
airplanes. J Sound Vib. 2003;262(3):457–74.
[5] Gibson LJ, Ashby M. Cellular Solids, structure and properties. Cambridge university press. 1997.
[6] Filho STA, Santos JF dos. Joining of Polymers and Polymer–Metal Hybrid Structures: Recent
Developments and Trends. Polym Eng Sci. 2009;49:1461–1476.
[7] Salonitis K, Drougas D, Chryssolouris G. Finite element modeling of penetration laser welding of
Sandwich materials. Phys Procedia. 2010;5(PART 2):327–35.
[8] Gower HL, Pieters RRGM, Richardson IM. Pulsed laser welding of metal-polymer sandwich materials
using pulse shaping. J Laser Appl. 2006;18(1):35–41.
[9] Pickin CG, Young K, Tuersley I. Joining of lightweight sandwich sheets to aluminium using self-pierce
riveting. Mater Des. 2007;28(8):2361–5.
[10] Bilici MK, Yükler Aİ, Kurtulmuş M. The optimization of welding parameters for friction stir spot
welding of high density polyethylene sheets. Mater Des. 2011;32(7):4074–9.
[11] Gerlich A, Avramovic-Cingara G, North TH. Stir zone microstructure and strain rate during Al 7075-T6
friction stir spot welding. Metall Mater Trans A. 2006;37(9):2773–86.
[12] Zhang Z, Yang X, Zhang J, Zhou G, Xu X, Zou B. Effect of welding parameters on microstructure and
mechanical properties of friction stir spot welded 5052 aluminum alloy. Mater Des. 2011;32(8–9):4461–
70.
[13] Bilici MK, Yukler AI. Effects of welding parameters on friction stir spot welding of high density
polyethylene sheets. Mater Des. 2012;33(1):545–50

240
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

A Study on the Influence of Shielding Gas on TiN Decomposition in Laser Surface

Alloying

Muvvala Gopinath, G Sai Krishna, and Ashish Kumar Nath*

Department of Mechanical Engineering, Indian Institute of Technology, Kharagpur, India
* Corresponding Author: aknath@mech.iitkgp.ernet.in

Abstract: Laser surface alloying (LSA) of AISI 1020 steel with TiN ceramic particles was carried out in Argon
and Nitrogen as shielding environment at different scan speeds. The influence of type of shielding gas
environment on decomposition of TiN particles and crack formation in alloyed layer was investigated. Also, the
influence of laser scan speed on distribution of ceramic particles was studied. Further, the molten pool thermal
history during LSA was monitored using an IR pyrometer and the possibility of prediction of TiN decomposition
from the profile of thermo-cycle recorded is discussed. SEM, XRD and wear tests were carried out to study the
influence of laser scan speed and type of shielding gas on distribution and decomposition of ceramic particles
and its influence on wear characteristics of the laser alloyed surface.

1. Introduction
In materials engineering applications, mild steels are widely used because of their good weldability, ductility,
high strength and relatively low cost [1, 2]. However, their low hardness and poor wear characteristics restrict
their applications. Therefore, there has been some interest in enhancing the surface hardness and wear
characteristics of these materials to meet the unrelenting demand [3, 4]. Among various surface engineering
techniques, Laser surface alloying (LSA) process with dispersed ceramic particles became more popular
because of the inherent properties of laser which includes precise focusability with high power densities
resulting in rapid heating and cooling rates (103-8 K/s) by self quenching mechanism with minimum HAZ and
distortion of substrate. Among various ceramic reinforcements like TiC, TiN, SiC etc, TiN is one of the
commonly used material as a reinforcement phase owing to its high degree of hardness, good wear
characteristics and excellent chemical stability [5, 6]. However, it suffers a severe limitation due to its tendency
to decompose when exposed to high temperature [7]. One of the other major problems associated with LSA of
ceramic particles is their distribution in the alloyed layer which affects the hardness and wear characteristics of
the coating. Kim et al. [7] reported the possibility of suppression of TiN decomposition using Nitrogen as
shielding environment. However, this requires a further elaborate study as the influence of laser process
parameters was not investigated in this study. Therefore, the present study aims at developing an understanding
of influence of process parameters and shielding environment on decomposition and distribution of TiN
particles during LSA. Further, the study also involves monitoring of molten pool thermal history during LSA
and attempts to identify the decomposition of TiN from the recorded thermo-cycle profile.

2. Experimental Setup
Laser surface alloying of AISI 1020 mild steel substrate with TiN was carried out using a 2 kW Yb-Fiber laser
(IPG photonics, Model no.YLR 2000) operating at 1.07 μm wavelength. Fig. 1 and Table 1 shows the schematic
of experimental setup and process parameters used in the current study. A 200 μm thin layer of TiN powder in
the form of slurry (mixed with 2% PVA) [8] was coated on the substrate surface using a coating machine
A study on the influence of shielding gas on TiN decomposition in laser surface alloying

(Model no: K101 control coater, RK Print Coat Instruments Ltd., UK) and baked in a furnace at 150 °C for
about 1 h to allow the moisture to evaporate and promote the coating to bond onto the substrate followed by
irradiating using the laser beam in Argon and Nitrogen environment. The molten pool thermal history during the
LSA process was monitored using an IR pyrometer mounted with a 1064 ± 25 nm notch filter to isolate the
reflected laser radiation from the pyrometer detector whose details were discussed elsewhere [9]. SEM (Zeiss,
EVO 18 Research) and XRD (PANalytical - Empyrean, Cr source) analysis was carried out to investigate the
effect of process parameters on evolution of microstructure and phases respectively. Wear test was carried out
on the surface developed with 30% overlapped tracks using a pin-on-disk setup (Make: DUCOM, TR-201-M3)
with a 6 mm WC ball as counter body with load, track diameter, RPM and duration set at 29.4 N, 5 mm, 300
and ~ 61 min with 300 m as resulting sliding distance, respectively.

Laser beam

Focusing lens
Shielding gas
Carrier gas (Ar/N 2 ) (Ar/N2 )

Temp (°C)
Shrouding gas (Ar/N 2 ) Monitoring molten pool
thermal history

IR Pyrometer Time (ms)

Melt pool
Pre-placed
powder Alloyed layer
Substrate Data acquisition and
analysis system

Fig.1. Schematic of experimental setup

Table. 1 Laser surface alloying process parameters

Laser Power (W) Powder layer thickness (µm) Spot diameter (mm) Scan speed (mm/min)
1400 200 3 200 - 1400

3. Results and Discussion

3.1 Molten pool thermal history
Fig. 2(b) represents a typical thermal history of the molten pool recorded using an IR pyrometer during LSA at a
laser power, scan speed and spot diameter of 1400 W, 1000 mm/min and 3 mm respectively in Nitrogen
environment. The molten pool thermal history essentially consists of two stages: heating cycle (0A) and cooling
cycle (AD). As the laser beam irradiates, the substrate temperature rapidly rises (0A) above the melting point
and as it advances, the molten pool behind it starts cooling (AD) which involves various stages. The Zone AC
represents the molten pool lifetime. It may be observed that, the slope of the cooling cycle between AC
exhibited changes due to initiation of nucleation and solidification process. The zone BC, termed as
solidification shelf corresponds to the base material, AISI 1020 mild steel. Further, it may be noted that, within
the zone AB, there was a small change in the slope of the thermo-cycle as indicated by arrow, which may be due
to solidification of Ti resulting from decomposition of fine dust like TiN particles present in TiN powder used as
depicted in Fig. 2(a). The zone CD corresponds to solid phase cooling of the alloyed surface. Based on the
various stages of thermo-cycle, heating and cooling rates were calculated in the all the cases. It was observed
that the heating rate in case if Nitrogen as shield gas is slightly higher due to exothermic reaction between Ti
and Nitrogen. However, cooling rate, molten pool lifetime and solidification shelf times were observed to be the
same.

242
 Proceedings of NCAMMM - 2018

Fig.2. (a) Morphology of TiN particles (b) Typical thermo-cycle depicting molten pool thermal history
(1400 W, 1000 mm/min, Nitrogen environment)
3.2 Influence of laser scan speed on distribution of TiN particles
Fig. 3 depicts the distribution of TiN particles in the alloyed surface with different laser scan speeds in Argon as
shielding environment respectively. It can be observed that with decrease of laser scan speed from 1000
mm/min to 600 mm/min, the TiN particles exhibited more uniform distribution with reduced agglomeration.
The distribution activity of the TiN particles is mainly determined by the intensity of Marangoni flow as well as
the viscosity of the liquid metal which are strongly temperature dependent [10-13]. Therefore, with the decrease
of scan speed, heat input increases rising the molten pool temperature reducing the viscosity and enhancing the
intensity of Marangoni flow causing the agglomerates of TiN particles break apart, uniformly distributing them
in molten pool as shown in Fig. 3(b). Also, the increase in molten pool volume with decrease of scan speed or
increase of heat input further favours TiN particles uniform distribution.

(a) (b)

Fig.3. Cross-section of LSA samples in argon environment processed at 1400 W laser power, (a) 1000 mm/min
and (b) 600 mm/min

3.3 Influence of shielding environment on TiN particle decomposition and wear properties
Fig. 4(a) and (b) shows the thermo-cycle and TiN particle condition corresponding to samples processed at 1400
W laser power and 1200 mm/min scan speed in Argon and Nitrogen environment respectively. It may be
observed that, the thermo-cycle corresponding to LSA in Argon environment (Fig. 4(a)) exhibited multiple
solidification shelves. In general multiple solidification shelves occur in case of materials with binary system
having different liquidus temperatures. It may be observed that, in case of Argon environment, the TiN started
decomposing resulting in porous particles as highlighted with free Ti precipitate around these pores. This free Ti
in Fe matrix, during its course of solidification is expected to cause a solidification shelf.

243
A study on the influence of shielding gas on TiN decomposition in laser surface alloying

Fig. 4 Influence of TiN decomposition on molten pool thermo-cycle profile (a) Argon environment and (b)
Nitrogen environment

Therefore, the solidification shelf S1 corresponds to that of substrate material whereas one near to 1700 °C
corresponds to the solidification of resultant Ti from TiN decomposition. However, under the same processing
conditions, with the use of Nitrogen as shielding environment, the presence of solidification shelf S2 was not
observed, as there was no decomposition TiN which is evident from the microstructure of TiN particles in Fig.
4(b). Therefore, the monitoring of molten pool thermal history may provide the information of TiN particle
condition inside the molten pool without carrying out any destructive and time consuming analysis like cross-
sectioning, polishing, etching and analyzing under SEM or optical Microscope. This reduces the lead time in
optimizing the process parameters.
3.4 Effect of laser scan speed and shielding environment on wear properties
Fig. 5 depicts the influence of laser scan speed as well as shielding gas environment on the wear resistance of
the laser alloyed surface developed with 30% overlap between the tracks. It can be observed that with decrease
of laser scan speed, wear rate was found to decrease which may be attributed to the uniform distribution of TiN
particles in the alloyed layer. However, at 400 mm/min, the wear rate was found to increase which is because of
decomposition of TiN particles as well as the increased dilution. Inspite of heavy dilution, coatings developed in
Nitrogen environment showed a far better performance as the TiN particles were found to be intact in this case
whereas in Argon environment major portion of TiN was decomposed making the coating brittle, resulting in
brittle fracture during the wear test shown in Fig. 5(b) which was minimum in case of Nitrogen environment
(Fig. 5(c)).

(a) Nitrogen Argon (b) (c)

4
Wear rate (10-5 gm/s)

0
200 600 1000 1400
Scan Speed (mm/min)

Fig. 5. (a) Influence of laser scan speed and shielding environment on wear resistance of alloyed surface and
wear track morphology (b) Argon and (c) Nitrogen environment (1400 W, 400 mm/min)

244
 Proceedings of NCAMMM - 2018

4. Conclusion
Laser surface alloying of AISI 1020 mild steel substrate with TiN was carried out and molten pool thermal
history during the alloying process was recorded using an IR pyrometer. The following conclusions are drawn
from the study:
1. Moderate scan speeds favor uniform distribution of TiN particles in laser surface alloying.
2. TiN particles were found to decompose during LSA in Argon environment whereas use of Nitrogen as
shielding gas found to suppress this.
3. Recorded molten pool thermo-cycle is capable of depicting the TiN particles condition in the alloyed layer.
4. Moderate scan speed and Nitrogen environment found to result in improved wear resistance.

Acknowledgements
Authors gratefully acknowledge the financial support from the Department of Science and Technology,
Government of India, under the FIST Program-2007 (SR/FIST/ETII-031/2007), and Ministry of Human
Resource Development and Department of Heavy Industries, Government of India, under the IMPRINT
Program-2017 for Project-6917.

References
[1] Sarkar S, Gopinath M, Chakraborty SS, Syed B, Nath AK. Analysis of temperature and surface hardening
of low carbon thin steel sheets using Yb-fiber laser. Surf. Coat. Tech. 2016;302:344–58.
[2] Majumdar JD. Development of wear resistant composite surface on mild steel by laser surface alloying with
silicon and reactive melting. Materials Letters. 2008;62:4257–59.
[3] Chatterjee S, Shariff SM, Padmanabham G, Majumdar JD, Roy Choudhury A. Study on the effect of laser
post-treatment on the properties of nanostructured Al2O3–TiB2–TiN based coatings developed by combined
SHS and laser surface alloying. Surf. Coat. Tech. 2010;205:131–38.
[4] Masanta M, Ganesh P, Kaul R, Nath AK, Roy Choudhury A. Development of a hard nano-structured multi-
component ceramic coating by laser cladding. Mater. Sci. Eng: A. 2009;508:134–40.
[5] Pu Y, Guo B, Zhou J, Zhang S, Zhou H, Chen J. Microstructure and tribological properties of in situ
synthesized TiC, TiN, and SiC reinforced Ti3Al intermetallic matrix composite coatings on pure Ti by laser
cladding. Appl. Surf. Sci. 2008;255:2697–03.
[6] Weng F, Yu H, Chen C, Liu J, Zhao L, Dai J. Microstructure and property of composite coatings on
titanium alloy deposited by laser cladding with Co42+TiN mixed powders. J. Alloy. Compd. 2016;686:74-
81.
[7] Kim TH, Seong BG. Titanium nitride laser-beam surface-alloying treatment on carbon tool steel. J. Mater.
Sci. 1990;25:3583-91.
[8] Gopinath M, Karmakar DP, Nath AK. Online monitoring of thermo-cycles and its correlation with
microstructure in laser cladding of nickel based super alloy. Opt. Laser. Eng. 2017;88:139–52.
[9] Gopinath M, Karmakar DP, Nath AK. Online assessment of TiC decomposition in laser cladding of metal
matrix composite coating. Mater. Design. 2017;121:310–20.
[10] Agarwala M, Bourell D, Beaman J, Marcus H, Barlow J. Direct Selective Laser Sintering of Metals. Rapid
Prototyping J., 1995;1(1):26–36.

245
A study on the influence of shielding gas on TiN decomposition in laser surface alloying

[11] Gu DD, Hagedorn YC, Meiners W, Meng GB, Batista RJS, Wissenbach K, Poprawe R. Densification
Behavior, Microstructure Evolution, and Wear performance of Selective Laser Melting Processed
Commercially Pure Titanium. Acta Mater. 2012;60(9):3849–60.
[12] Gu D, Wang H, Dai D. Laser Additive Manufacturing of Novel Aluminum Based Nanocomposite Parts:
Tailored Forming of Multiple Materials. J. Mater. Sci. Eng. 2016;138(2):021004-1-11.
[13] Yuan P, Gu D, Dai D. Particulate migration behavior and its mechanism during selective laser melting of
TiC reinforced Al matrix nanocomposites. Mater. Design. 2015;82:46–55.

246
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Influence of Heat Input on Shear Strength and Macro-Hardness of 316 Austenitic

Stainless Steel Cladding onto E250 Low Alloy Steel by GMAW Process

Manas Kumar Saha, Lakshmi Narayan Dhara, Santanu Das

Department of Mechanical Engineering, Kalyani Govt. Engineering College,
Kalyani-741235, West Bengal

Abstract: 316 Austenitic stainless steel is one well accepted cladding material having good corrosion resistance
properties. Cladding, a surfacing process not only enhances corrosion, erosion resistance properties but also
improves mechanical properties. Heat input plays a pivotal role for producing good cladding with respect to
weld bead geometry, cooling rate, etc. In the present experiment, three welding torch travel speed were taken
keeping welding current and voltage constant, so that three different heat input were produced using 100% CO 2
as shielding gas. Single layer 50% overlapped cladding was done. Shear strength test and macro hardness test
were performed over test samples prepared from clad part. Results showed that both macro hardness and shear
strength of the clad layer were more than that of base metal and values of both increase with increase in heat
input.
Keywords: Cladding, heat input, shear strength, hardness

1. Introduction
Cladding is one of the surfacing techniques, in which few millimetre deposition of a filler material is produced
onto the substrate layer-by-layer with the objective to improve the corrosion or wear resistance properties of the
substrate. It is also used for improving hardness and is termed as hard facing. Cladding enhances the strength of
the surface of parent metal up to a certain limit without change the microstructure of the base material. Cladding
makes a new surface layer with different compositions which, in general, is harder than the base material.
Comparing with other surfacing techniques, cladding has some distinct advantages, such as enhancement of
hardness, corrosion and/or erosion resistance of job surfaces, providing good bonding and favourable
microstructure of clad part [1-4].
Clad parts are widely used in oil refineries and power generation industries, offshore oil production,
chemical, fertilizer, nuclear and steam power plants, food processing, and petrochemical industries [3-4]. Clad
part are produced by various techniques such as rolling, explosion welding, different fusion welding, resistance
welding, etc. Gas metal arc welding is a semi-automatic, effective, economic and user-friendly technique that is
used extensively to producing cladding. The quality of cladding can be further improved by proper selection of
process parameters. Better selection of process parameters gives optimum bead geometry that ultimately
improves quality of cladding in terms of corrosion/erosion resistance properties and mechanical strength [2-6].
If one uses 100% CO 2 as shielding gas in GMAW process, it is known as metal active gas (MAG) welding.
Though MAG welding produces not very smooth weld bead and some weld defects like spatters, but also it is
much economical. Selection of optimum process parameters of MAG welding can produce favourable
microstructure, satisfactory corrosion resistance properties and very good mechanical strength. Heat input plays
a vital role to produce good quality cladding.
Unlike welding, cladding width and reinforcement should be grater but penetration should be optimum
so that dilution will be of minimum value without sacrificing the shear strength of the weld joint. Less value of
Influence of heat input on shear strength and macro-hardness of 316 austenitic stainless steel cladding onto E250 low alloy
steel by GMAW process
depth of penetration lowers dilution but increases the risk of low bond strength which ultimately enhances the
chance of lowering the shearing strength of cladded layer. Little approaches are been observed to check the
shear strength of the cladded layer so far.
In the present work, 316 austenitic stainless steel weld bead were produced on low alloy steel (E250)
by GMAW process applying 100% CO 2 as shielding gas.Process parameters such as welding current and torch
travel speed were selected in three levels, keeping welding voltage constant so that nine different heat input had
been produced and nine weld bead samples were produced. Analyzing the weld bead geometry components, 3
best samples with reference to bead geometry were selected and the corresponding process parameters were
identified. In the second part of the experiment cladding was performed using best three heat input and
corresponding process parameters. Macro hardness testing and shear testing of clad layer were performed to
check the influence of heat input on mechanical strength of cladded layer.

2. Experimental Procedure
The chemical composition of base metal and filler metals, i.e. E250 and 316 austenitic stainless steel are shown
in Table 1, and Table 2 respectively. 316 austenitic stainless steel filler electrode diameter is 1.2 mm. ESAB
K400 model of MIG/MAG machine (Model: Auto K400) using AC source, semi-automatic nozzle carriage rail
guided vehicle (RGV) uses DC motor. Fig. 1 shows the RGV.
Table 1: Chemical composition of Base material (E250)
%C %Si %Mn %P %S %Cr %Al
0.1985 0.1402 0.4976 0.0609 0.0308 <0.0011 .0003
%Ni %Co %Cu %Nb %Ti %V %Sn
0.0253 0.0059 0.0053 0.0096 0.0024 0.0024 0.0137
%Pb %As %Zn %B %Fe %Zr
0.0126 0.0662 <0.0009 0.0012 <98.8810 0.0025
Table 2: Chemical composition of 316 Austenitic Stainless Steel
%C %Si %Mn %P %S %Cr %Mo
0.0758 0.1824 1.1017 0.0289 0.0076 15.0464 2.0906
%Ni %Al %Co %Cu %Nb %Ti %V
9.9370 0.0105 0.0741 0.3417 0.0023 0.0316 0.0475
%W %Sn %Ce %B %Fe
0.0261 0.0103 0.0103 0.0010 <70.9524

Fig. 1: Welding gun mounted on a motor driven vehicle Fig. 2: Shear testing device fitted on to UTM
3. Experimental Procedure
3.1. Bead on plate experiments
316 austenitic stainless steel weld beads were produced on E250 base metal by GMAW process using 100%
CO 2 as shielding gas. Welding Current and torch travel speed were taken in three levels such as 140 A, 170A,
200A and 360mm/min, 390mm/min, 420 mm/min respectively so that nine number of samples were formed.

248
 Proceedings of NCAMMM - 2018

Welding voltage was kept constant during the experiment.The whole experiments were replicated twice.
Equation (1) is used to find heat input, Q.
𝑉𝑉×𝐼𝐼×60
Q= ×η (1)
𝑆𝑆×1000

where Q = Heat Input (kJ/mm), V = Voltage (V), I = Current (A), S = Travel Speed (mm/min) and η =
Efficiency (In this experiment, it is taken as 0.8).
The heat input and other process parameters are shown in Table 3. The samples were cut in transverse
direction by power saw, ground and etched with 2% nital solution and checked under tool maker’s microscope.
Weld bead geometry components like reinforcement height, bead width and depth of penetration were measured
with the help of scale inside the microscope.
Table 3: Heat input and other process parameters of Bead-on-plate welding
Sample Voltage, V Current, I Travel speed, S Heat input
No. (V) (A) (mm/min) (kJ/mm)
1 27 140 360 0.504
2 27 140 390 0.465
3 27 140 420 0.434
4 27 170 360 0.612
5 27 170 390 0.564
6 27 170 420 0.524
7 27 200 360 0.720
8 27 200 390 0.664
9 27 200 420 0.504

3.2. Cladding experiments

Single layer 50% overlapped 316 austenitic stainless steel cladding were performed on E250 low alloy steel by
GMAW process using 100% CO 2 as shielding gas. Three sets of heat inputs along with their subcomponents
process parameters such as welding current and torch travel speed were selected on the basis of better results of
weld bead geometry formation in previous experiments.
3.3. Macro hardness test
Macro hardness tests for the samples made were performed by Rockwell hardness testing machine (Make: Fine
Testing Machine, India, Model: TSM/FTM, Sl. No. 97/068). Hardness of test specimen is measured using C
scale by applying 150 KN load and a Diamond indenter with 136˚ apex angle. Rockwell testing machine is used
for hardness testing.
3.4. Shear test of cladded part
Shear strength of clad sample is measured by a shear testing device as shown in Fig, 2 which is attached in a
universal testing machine. Shear strength is calculated by following formulae.
Shear strength = shear load/ Area (2)

4. Results and Discussion

4.1. Results of bead-on-plate experiments
Visual inspection results reveal that most of welds of higher heat input conditions possess good quality. It is
visually observed that Width of weld increases with welding heat input as expected. In all cases continuity in
weld deposition is found. Spatters are varied from high range to low range with different heat input conditions.
High spatter at lower current indicates the possibility of globular mode of metal transfer. Table 4 shows the
results obtained from visual inspection.

249
Influence of heat input on shear strength and macro-hardness of 316 austenitic stainless steel cladding onto E250 low alloy
steel by GMAW process
Table 4: Results from visual inspection of bead-on-Plate Experiments
Sl. Sam Voltage Current Travel Heat Blow hole Continuity in Spatter
No. ple (V) (A) Speed input deposition
No. (mm/min) (kJ/mm) 1st 2nd 1st 2nd 1st 2nd
Repli Repli Replica Replica Replic Repli
cation cation tion tion ation cation
1 S3 27 140 420 0.434 Nil Nil Cont. Cont. high Med.
2 S2 27 140 390 0.465 Nil Nil Cont. Cont. High Med.
3 S1 27 140 360 0.504 Nil Nil Cont. Cont. Med. Med.
4 S6 27 170 420 0.524 Nil Nil Cont. Cont. Low Med.
5 S5 27 170 390 0.564 Nil Nil Cont. Cont. High High
6 S4 27 170 360 0.612 Nil Nil Cont. Cont. Med. Med.
7 S9 27 200 420 0.617 Nil Nil Cont. Cont. Low Med.
8 S8 27 200 390 0.664 Nil Nil Cont. Cont. Med. Med.
9 S7 27 200 360 0.720 Nil Nil Cont. Cont. Low Med.

4.2. Weld bead geometry of bead-on-plate experiments

Table 5 shows the results of various weld bead geometry of bead-on-plate experiment using 100% CO 2 as
shielding gas. The results of bead-on-plate experiments clearly indicate that at higher input condition weld bead
width, reinforcement and depth of penetration increases with increasing heat input. It is seen that within welding
speed of 360-420 mm/min and 200A of welding current, good reinforcement and weld bead width at optimum
penetration is observed in both sets of experiments. Three heat input conditions (0.617 kJ/mm, 0.664 kJ/mm,
and 772 kJ/mm) are selected for cladding based on visual inspection, weld bead geometry and their effects on
heat input from bead on plate experiments.
Table 5: Values components of bead on plate experiments with 100% CO 2 gas shield
Sample Voltage Current Travel Heat input Weld bead Height of Depth of
No. (V) (A) Speed (kJ/mm) width Reinforcement Penetration
(mm/min) (mm) (mm) (mm)
S3 27 140 420 0.434 5.000 1.950 1.650
S2 27 140 390 0.465 5.500 2.250 1.900
S1 27 140 360 0.504 5.680 2.450 2.100
S6 27 170 420 0.524 5.755 2.615 1.040
S5 27 170 390 0.564 5.900 2.745 1.400
S4 27 170 360 0.612 5.855 2.585 1.425
S9 27 200 420 0.617 6.005 2.500 1.500
S8 27 200 390 0.664 6.060 2.800 1.790
S7 27 200 360 0.720 6.500 3.135 2.490

4.3. Macro hardness test

The three zones, on which the values are measured, are the base metal, the weld bead zone and the heat effected
zone. Average hardness value of base plate E250 is 41 HRC. Table 6 shows hardness values of three zones.
From the result it is observed that hardness value is decreasing from clad zone (48-61) to base metal (41.44).
Hardness of the heat affected zone is 43-52. Fig.3 shows graphical representation of hardness at different zone.
4.4. Shear test
Table 7 shows shear stresses of clad samples. Shear strength of clad samples are greater than base metal as
standard ultimate shear strength of 0.2 % carbon steel at hot rolled condition is given by 3150 Kg/cm2.

250
 Proceedings of NCAMMM - 2018

Table 6: Hardness values of bead on plate experiments

Sample Heat Hardness value on weld bead zone (HRC) Hardness value on the interface zone (HRC)
No. input 1st value 2nd 3rd value Average 1st value 2nd 3rd value Average
kJ/mm value value value value
S3 0.434 50 52 49 50 43 44 41 43
S2 0.465 47 51 45 48 42 45 43 43
S1 0.504 53 50 55 53 45 44 44 44
S6 0.524 55 51 48 51 44 46` 42 44
S5 0.564 51 54 49 51 49 47 45 47
S4 0.612 58 55 58 57 53 50 52 52
S9 0.617 57 52 59 56 51 49 48 49
S8 0.664 61 55 57 58 49 51 47 49
S7 0.720 64 57 61 61 53 50 46 49

64
62
60 Hardness value at
Hardness, HRC

58
56 weld metal zone
54
52 Hardness value at
50
48 interface zone
46
44 Hardness of Base
42
40 metal E250
0 1 2 3 4 5 6 7 8 9
Test samples
Fig. 3: variation of hardness at different zones of cladded part.
Table 7: Shear strength values of weld clad done with a shielding gas of 100% CO 2
Sample Heat Input Shear area Shear force Shear stress Shear stress
No. (kJ/mm) (mm2) (kN) (kN/mm2) (kg/cm2)
1 0.617 56.04 20.08 0.358 3580
2 0.664 116.62 44.19 0.378 3780
3 0.720 75.82 42.40 0.559 5590

6000
strength, kg/cm²

(100% CO2)

4000 (100% CO2

Shear

2000
Poly. ((100%
CO2)
0
0.617 0.664 0.72
Heat input, kJ/mm
Fig. 4: Bar chart of shear strength with heat input of clad samples
Table 7 shows bar chart of shear strength with heat input of clad samples. From the chart it is obtained that shear
strength increases with increase of heat input. The failure or crack occurs generally at three Places. These areas
are clad layer, Strain localization in base metal, i.e. heat affected zone and base metal. From the observations, it
is seen that in all the specimens, shear occurs at heat affected zone or at the base metal. In no case, shear occurs
at weld bead. Fig. 4 shows graphical representation of shear strength against heat input, which shows shear
strength increases with increase in heat input within experimental domain.
316 Austenite stainless steel filler electrode contains alloying elements like Cr, Ni, Mo, Si, Mn, which are
responsible to increase hardness and shear strength. The austenitic phase present in 316- SS converts into

251
Influence of heat input on shear strength and macro-hardness of 316 austenitic stainless steel cladding onto E250 low alloy
steel by GMAW process
retained austenite or martensite on cooling which has more mechanical strength than base metal. Moreover high
carbon equivalent of 316-SS signifies more possibility of formation of martensite in cladding causing more
hardness.

5. Conclusion
From the results obtained from various e tests performed it can be concluded as follows:
 It is observed from visual inspection that welds with higher input are found to be of good quality than lower
heat input
 After cladding, hardness of austenitic stainless steel clad portion is found to be significantly higher than low
alloy steel base plate, as it has the presence of hardening elements like chromium, nickel, molybdenum, etc.
 Results of shear test clearly indicate that shear strength increases with increasing heat input. Shear strength
of all the clad samples is a lot higher than base metal. Comparing both sets of shear stress test, maximum
shear stress is obtained in the present investigation at 0.772 kJ/mm heat input with 200A welding current,
27V weld voltage and 360 mm/min weld torch travel speed.
 Finally, it can be concluded that a higher heat input of 0.772 kJ/mm may be recommended for cladding to
have maximum shear strength at the weld joint within the experimental domain.

References
[1] Saha M K and Das S. A review on different cladding techniques to resist corrosion. J. of the Association of
Engineers, India. 2016; 86: 51-63.
[2] Datta S, Nandi G, Bandyopadhyay A, Pal P. Grey-based Taguchi method for optimization of bead
geometry in submerged arc bead-on-plate welding. Int. J. of Adv. Mfg. Tech. 2009; 45: 276-286.
[3] Mondal A, Saha M K, Hazra R, Das S. Influence of heat input on weld bead geometry using duplex
stainless steel wire electrode on low alloy steel specimens. J. of Cogent Engg. 2016; 3: 1-14.
[4] Kannan T, Murugan N. Effect of flux cored arc welding process parameters on duplex stainless steel clad
quality. J. of Material Processing Tech. 2005; 176: 230–239.
[5] Chakrabarti B, Das S, Das H, Pal TK. Effect of process parameters on clad quality of duplex stainless steel
using GMAW process. Trans. of the Indian Inst. of Metals. 2013; 66(3): 221-230.
[6] Verma A K, Biswas B C, Roy P, De S, Saren S, Das S. On the effectiveness of duplex stainless steel
cladding deposited by gas metal arc welding. e-Proc. IIW IC 2014, Seoul, Korea. 2014.
[7] Buchanan VE, Shipway PH, McCartney DG. Microstructure and abrasive wear behaviour of shielded arc
welding hardfacings used in the sugarcane industry. Wear. 2007; 263: 99-110.
[8] Gualco AH, Svoboda G, Surian, ES, De Vedia LA. Effect of welding procedure on wear behavior of
modified martensitic tool steel hardfacing deposit. Mat. & Des. 2010; 31(9): 4165-4173.
[9] Saha M K, Mondal A, Hazra R, Das S. On the variation of hardness of duplex stainless steel clad layer
deposited by flux cored arc welding. Reason- A Technical Journal. 2016; XV: 1-6.

252
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Effect of Laser Cladding Parameters on Clad-Track Uniformity

Debapriya Patra Karmakar, Gopinath Muvvala, Shams Perwez, Ashish Kumar Nath
Department of Mechanical Engineering, Indian Institute of Technology, Kharagpur
E-mail: debapriya.mse10@gmail.com

Abstract: Uniformity of the Stellite-6 clad-tracks deposited by pre-placed powder laser cladding at different
parameters has been investigated. Stellite-6 clad-tracks were deposited on AISI H13 tool steel at different laser
power and scan speed. The cross-sectional profiles of clad-tracks were measured by a non-contact type surface
profilometer and the irregularities of a track has been quantified by the difference of track widths and heights
along a track over average width and height of that particular track. These two parameters were termed as
width and height irregularity and found to be influenced by the laser power and scan speed. Slow scan speed
and large energy density lead to higher height irregularity possible due slow solidification rate. Width
irregularity found to decrease with scan speed and appeared to be influenced by the distribution of laser power
in the laser spot at different laser parameters.

1. Introduction
Laser cladding has become very popular in fields of different layer deposition on working surface of
components to improve the mechanical and chemical properties like hardness, wear, and corrosion resistance
etc. Hence different modes of laser cladding are being used for depositing layers of different materials on work-
surfaces, repair works of worn out or damaged surfaces as well as fabrication of three-dimensional objects. In
these types of process deposition or material addition is performed layer by layer. Further, each layer consists of
multiple overlapped tracks as building blocks. Hence, the layer dimension is controlled by the track profile.
Many researchers investigated the profile and geometry of laser claddings of different materials with respect to
the process parameters for different category and modes of laser cladding. Liu et al. (1) studied single track pre-
placed powder laser cladding and performed geometric modeling. Farnia et al. (2) investigated the effects of
pulse duration and overlapping factor on melting ratio for pre-placed laser cladding using a pulsed Nd:YAG
laser. Barekat et al. (3) and D’Oliveira et al. (4) studied different aspects of co-axial laser cladding and
investigated the correlation between the processing parameters and geometrical characteristics of laser clad-
tracks. Literature on the effect of process parameters on mechanical properties and microstructure of laser
claddings of various materials are also available (5, 6). But there is not much information about the uniformity
or irregularity in a single clad-track or clad layer deposited by laser cladding. Process parameters leading to non-
uniform track are expected to be discarded as they are not suitable uniform deposition. But when there is a
requirement of a particular mechanical property or microstructure obtained by a particular set of process
parameters which causes intended microstructure and properties but don’t give desired track uniformity, then
there is a need of information about uniformity/irregularity and influence of processing parameters on it so that a
judicious compromise on particular criteria can be made to select an acceptable parameters.
In the current work, the effect of laser power and scan speed on uniformity or irregularity of laser deposited
cald-tracks has been investigated.
Effect of laser cladding parameters on clad-track uniformity

2. Experimental Details
A 2kW Yb-fiber laser (IPG, YLR 2000) integrated with a 5-axis CNC machine was used to conduct the laser
cladding experiment. The simplest mode of laser cladding i.e. pre-placed powder laser cladding had been
performed to deposit the clad-tracks of Stellite-6, which is a well-known wear and corrosion resistance cobalt-
chromium alloy, on AISI H13 tool steel substrate material. Laser scans at different power and scan speed were
conducted on 1mm thick pre-placed Stellite-6 powder layer applied with polyvinyl alcohol as a binder. The
process parameters are listed in Table 1.
Optical top views and the microstructure of some of the tracks are shown in Fig.1. The microstructure
of selected parameters are similar to wear and erosion resistance coatings investigated by other researchers (7,
8). Profiles of clad-tracks were measured by a non-contact type laser based displacement sensor (Micro-Epsilon
ILD-2300) mounted on a CNC controlled high precision motorised translation stage (Thorlabs MTS50Z8) at
different positions along the tracks. The track profiling was done perpendicular to tracks. Minimum and
maximum cross-section(c/s) profiles were identified after several measurements. Different cross-sectional
profiles perpendicular to the tracks were plotted. Profile comparisons of two such tracks with the corresponding
track-images and position of the profile sections are shown in figure 2.

Table 1 Process parameters for laser cladding

Pre-placed powder layer thickness Laser spot diameter Laser Power Scan Speed
1mm 3mm 1200W 200 mm/min
1400W 400 mm/min
1600W 600 mm/min

From these profiles, the difference between the maximum and minimum track widths (Δw) and heights (Δh)
were measured. Average width (w av ) and height (h av ) were also calculated. Uniformity or irregularity of each
Δw Δh
clad-tracks was quantified in terms of two parameters defined as (width irregularity) and (height
w av h av

irregularity), expressed in percentage. High irregularity value represents less uniformity of clad-track. These two
parameters were plotted against the laser power and scan speed along with the average width and height to
understand their behavior aganist these two parameters.
View
Top

2mm 2mm 2mm

Microstructure
of polished and
etched cross
sections

1200W 200mm/min 1200W 400mm/min 1200W 600mm/min

Figure 1 Top view and cross-sectional microstructure of clad-tracks formed at 1400W power

254
 Proceedings of NCAMMM - 2018

large large
small c/s c/s c/s small c/s

Figure 2 Track profiles at different position of two tracks formed at 200 and 600 mm/min scan speed
and 1200W laser power.

3. Results and Discussion

Δw Δh
Variations of width ( ) and height ( ) irregularity, (Δw) and (Δh) with respect to the process parameters
w av h av
are shown in figures 3 and 4.

Large height irregularity percent is observed for tracks scanned at a slower speed (200mm/min) for all
the power levels. For all the experimental laser power levels slower scan speed caused high energy density
(120-160 J/mm2). The solidification front propagation of the melt-pool at slow speed and high energy density is
also expected to be slow, giving enough time to flow of molten metal. But the molten metal flow become
intermittent due to inertia of the melt-pool which causes formation of uneven clad-track with appearance
similar to multiple liquid metal droplets joined together and solidified afterward. These clad-tracks could be
compared to weld-beads formed by globular metal transfer in case of arc welding. However at higher scan speed
and lesser energy density melt-pool solidification is expected to be faster, which causes immediate solidification
without any uneven flow of molten metal from the melt-pool of laser cladding. Hence, at higher speed clad-
tracks were found to be more uniform or less irregular. At the time of process parameter selection for clad-layer
to be formed by multiple overlapped track, uniformity or irregularity is also has to be considered along with
thickness, aspect ratio, microstructure, cooling rate etc. If parameters are selected based on desired
microstructure, cooling rate etc, with some compromise at uniformity level, then post-processing allowance and
number of layer deposition need has to be increased to counteract the height irregularity so that desired final
shape and size can be obtained.

255
Effect of laser cladding parameters on clad-track uniformity

Δh (mm)
0.6
0.5
0.4
0.3
0.2
0.1
0

hav (mm)
1.5
1.25
1
0.75
0.5
0.25
0
Δh/hav (%)
60
50
40
30
20
10
0
200mm/min

400mm/min

600mm/min

200mm/min

400mm/min

600mm/min

200mm/min

400mm/min

600mm/min
1200W

1200W

1200W

1400W

1400W

1400W

1600W

1600W

1600W
Figure 3 Height difference (Δh), average height (h av ), and percentage irregularity of height (Δh/h av ) variations
at different parameters

The percentage irregularity of track-width found to be relatively less significant (<20%) compared to height
irregularity. Clad-width is expected to be controlled by the laser spot diameter and the side walls of the clad-
tracks are also remain surrounded by un-melted pre-placed powder particles to some extent at the time of
melting and solidification. Hence the irregularity could be less affected by laser parameters. But overall width
irregularity was found to decrease with scan speed in terms of the width difference (Δw) along the tracks. At
higher scan speed effective spot diameter gets reduced, which could be responsible for relatively uniform
distribution of laser energy in effective laser spot diameter causing less width difference of different sections
along a track. In case of slow speed, the uneven distribution of power in the spot causing high intensity near
core region and less intensity near the boundary region could be the reason of lateral irregularity or width
irregularity of formed clad-tracks. However percentage irregularity gets magnified when width difference (Δw)
is divided by the average width (W av ) at higher speed levels as higher speed causes narrow tracks. Thus, width
irregularity information is also needs at the time of selection of track-to-track lateral overlap percentage besides
other factor like aspect ratio, desired layer-thickness etc. if clad-tracks with relatively larger levels of width
irregularity has been selected based on desired microstructure formation or due to limited flexibility of cladding
process parameters.

256
 Proceedings of NCAMMM - 2018

Δw (mm)
0.5
0.4
0.3
0.2
0.1
0

wav(mm)
4
3
2
1
0

Δw/wav (%)
15

10

0
200mm/min

400mm/min

600mm/min

200mm/min

400mm/min

600mm/min

200mm/min

400mm/min

600mm/min
1200W

1200W

1200W

1400W

1400W

1400W

1600W

1600W

1600W
Figure 4 Width difference (Δw), average width (w av ), and percentage irregularity of width (Δw/w av ) variations
at different parameters

Further, the width and height irregularity is expected to be related to each other and complementary to some
extent as melt-pool created by laser radiation can flow and solidify in both height and width direction. Thus,
direct correlation or mapping of these two parameters with laser parameters may not give pure increasing or
decreasing trend but may be useful in process parameter selection and adjustment of parameters if needed.

4. Conclusions
Based on the experimental results the following conclusions can be made.
1. Height irregularity at 200 mm/min is found to be significantly large and decreased significantly at higher
scan speeds.
2. Slow scan speed and large energy density lead to higher height irregularity caused by slow solidification
rate.
3. Width irregularity is found to be less dependent (<15%) to laser parameters compared to height
irregularity (max. value >50%).
4. Width irregularity found to decrease with scan speed and appeared to be influenced by the distribution of
power in the laser spot at different laser parameters.

257
Effect of laser cladding parameters on clad-track uniformity

Thus, height irregularity needed to be accounted for resulting layer waviness and final thickness with tolerance
limit, whereas the width irregularity information could be useful for the selection of lateral overlap in case of
multi-track claddings.

Acknowledgements
Authors gratefully acknowledge the financial support from the Department of Science and Technology,
Government of India, under the FIST Program-2007 (SR/FIST/ETII-031/2007), and Ministry of Human
Resource Development and Department of Heavy Industries, Government of India, under the IMPRINT
Program-2017 for Project-6917.

References
[1] Liu H, Qin X, Huang S, Hu Z, Ni M. Geometry modeling of single track cladding deposited by high power
diode laser with rectangular beam spot. Optics and Lasers in Engineering. 2018; 100:38–46.
[2] Farnia A, Ghaini FM, Sabbaghzadeh J. Effects of pulse duration and overlapping factor on melting ratio in
preplaced pulsed Nd:YAG laser cladding. Optics and Lasers in Engineering. 2013; 51: 69–76.
[3] Barekat M, Razavi RS, Ghasemi A. Nd:YAG laser cladding of Co–Cr–Mo alloy on γ-TiAl substrate.
Optics & Laser Technology. 2016; 80: 145–152.
[4] D’Oliveira ASCM, da Silva PSCP, Vilar RMC. Microstructural features of consecutive layers of Stellite 6
deposited by laser cladding. Surface and Coatings Technology. 2002; 153: 203–209.
[5] Weng F, Yu H, Chen C, Dai J. Microstructures and wear properties of laser cladding Co-based composite
coatings on Ti–6Al–4V. Materials and Design. 2015; 80: 174–181.
[6] Telasang G, DuttaMajumdar J, Padmanabham G, Tak M, Manna I. Effect of laser parameters on
microstructure and hardness of laser clad and tempered AISI H13 tool steel. Surface & Coatings
Technology. 2014; 258: 1108–1118.
[7] Singh R, Kumar D, Mishra SK, Tiwari SK. Laser cladding of Stellite 6 on stainless steel to enhance solid
particle erosion and cavitation resistance. Surface & Coatings Technology. 2014; 251: 87–97.
[8] Xu G , Kutsuna M, Liu Z, Yamada K. Comparison between diode laser and TIG cladding of Co-based
alloys on the SUS403 stainless steel. Surface & Coatings Technology. 2006; 201: 1138–1144.

258
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Comparison of Laser Marking on Aluminum, Stainless Steel and Copper Sheet using
Nd:YVO 4 Laser

A. Roy1, N. Kumar2, Santanu Das3 and A. Bandyopadhyay4

1
Power Engineering Department, Jadavpur University, Salt Lake City, Kolkata, India
2,4
Department of Mechanical Engineering, Jadavpur University, Kolkata, India
3
Mechanical Engineering Department, Kalyani Govt. Engineering College, Kalyani, India
E-mail: 1angshuman_roy01@yahoo.co.in, 2nikhilju2013@gmail.com, 3sdas.me@gmail.com,
4
asishbanerjee@yahoo.com

Abstract: Laser marking on metals, plastics, ceramics, semiconductors and natural materials is used in different
industries for the purpose of product identification like name, number, barcode, logo, decorative arts, etc. In the
present work, laser marking has been conducted on three different materials like aluminum, stainless steel and
copper sheet with 1.5mm thickness using Nd:YVO 4 laser having a wavelength of 1064nm. The input parameters
have been selected as laser power, scanning speed and pulse frequency. Responses like marking width and
marking depth have been measured, and influence of process parameters on these responses has been explored.
It is found that both marking width and marking depth increase with the rise in laser power, while hike in
scanning speed and frequency cause lowering of these responses.
Keywords: Aluminum, Stainless steel, Copper, Marking depth, Marking width, Laser marking, Nd:YVO 4 laser

1. Introduction
Mark depth (MD) and mark width (MW) are strongly affected by the laser source and also by the process
parameters, such as power, pulse frequency and scanning speed. Laser based marking is performed by the
principle of melting and vaporization of unwanted material from the parent material. Main advantages of laser
marking are that it is non-contact, permanent, highly repeatable, eco-friendly and is a highly flexible process
compared to conventional marking processes like ink stamp marking, electro chemical marking, punching, etc.
and it can be automated easily [1].
Qi et al. [1] investigated to find out pulse frequency to have the most significant effect on mark width and
mark depth for better quality marking. Valette et al. [2] worked on two different steels, martensitic Z30C13 and
austenitic 316L successfully. Chen et al. [3] in an interesting work tried marking on an eggshell surface
employing a 100 W CO 2 laser engraving system. It had beam diameter of 6 mm and rate of pulse repetition of 1
kHz. Marx et al. [4] reported laser marking on plants and fruits. For this experiment, various laser wavelengths
used were 193 nm to 10,600 nm to mark a simple 9 mm2 sized 2-D barcode on the surface of apple and
rhododendron cuttings. Depending on the laser wavelength, the process steps of laser marking, the risk of fungal
infection and the marking pattern alteration due to product storage were investigated. The marking process
variables, such as laser energy, laser wavelength, pattern size, pattern design, etc. considerably influenced
representation accuracy of the patterns used, which was important for later image processing. Amara et al. [5]
investigated fibre laser colour marking on steels using pulsed fiber laser with a wavelength of 1064 nm. The
amount of the absorbed energy varied with laser beam power, marking velocity and laser pulse frequency.
Femtosecond laser marking of stainless steel was performed by Astarita et al. [6]. They studied laser marking on
titanium cold sprayed coatings on aluminium. The marking tests were carried out under different experimental
Comparison of laser marking on aluminum, stainless steel and copper sheet using ND:YVO4 laser

conditions varying the main process parameters, such as laser pulse power and laser scanning speed. Condition
for obtaining maximum penetration depth and width of the marks were acquired, and occurrences of internal
damage induced by the process were studied.
Recently, authors of this paper carried out a series of investigations [7-9] on laser marking/engraving on
different materials. Pulsed mode of Nd:YVO 4 laser system was used to mark on AISI 304 Stainless Steel [7]
and aluminum [8] thin sheets. They employed Response Surface Methodology of the Design of Experiment for
undertaking the experimental work to evaluate optimum process parameters corresponding to desired mark
depth and width. They also explored [9] Surface Roughness Obtained during Laser Engraving on stainless steel
specimens.
The aim of this work is to find out the influence of laser power, pulse frequency and scanning speed on the
quality characteristics of marked samples. Mark depth and mark width are considered to be the characteristic
marking geometrical features.

2. Experimental Procedure
In this experimental work, aluminum, copper and stainless steel specimens of 20 mm x 30mm x 1.5 mm thick
sheets have been chosen. Specimens are cleaned by acetone to make it oil free and also ultrasonic cleaning is
conducted for removing unwanted dust particles. All the experiments have been carried out using Nd:YVO 4
(Neodymium-doped Yttrium Orthovanadate) laser system (Make: Electrox, Model: EMS 100). Spot diameter of
laser beam is 50 µm. All the samples are marked at the same combination of process parameters (by varying one
parameter and keeping the other two parameters constant). Thereafter, samples are mounted in ND-RD (ND:
Normal direction; RD: Rolling direction) section. The mounted samples are then cloth polished to see the
marking grooves. Pictorial views of marked samples (Al, Cu and SS) are shown in Fig. 1(a-c) respectively.
Depth and width of marking are measured using an optical microscope (Make: Leica; Model: DSLM). Table 1
shows detailed process control parameters used. Fig. 2 shows view of few marked samples of Al, Cu and SS.
Table 1. Process control parameters and values
Parameters, Units Notations Values
Power, % of 12 W P 60 70 80 90
Frequency, kHz f 4 8 12 16
Speed, mm/s s 1 5 10 15

(a) (b) (c)

Fig. 1. Photographic views of marked samples of (a) Al, (b) Cu and (c) Stainless Steel (SS) materials

260
 Proceedings of NCAMMM - 2018

(a) (b) (c)

(d) (e) (f)

Fig. 2. Pictorial views of few marked samples of
[(a) Al at s= 5 mm/s f= 10 kHz and P= 90% of 12 W, (b) Al at s= 15 mm/s f= 10 kHz and P= 90% of 12 W,
(c) Cu at s= 12 mm/s f= 5 kHz and P= 90% of 12 W, (d) Cu at s= 8 mm/s f= 10 kHz and P= 90% of 12 W,
(e) SS at s= 5 mm/s f= 10 kHz and P= 90% of 12 W, (f) SS at Cu at s= 5 mm/s f= 8 kHz and P= 90% of 12 W.]

3. Results and Discussion

Table 2 through Table 4 show the measured values of responses, i.e. marking width (MW) and marking depth
(MD), at varied levels of power (P), scanning speed (s) and frequency (f) respectively. Marking width and depth
are listed for all three metallic materials which are aluminum, copper and stainless steel thin sheets. With these
data, Fig.3(1-f) have been constructed to follow the nature of variations of each of the two response variables
against the three process parameters.
Table 2. Measured responses at different values of power
P f s Al Cu SS
(W) (kHz) (mm/s) MW (µm) MD (µm) MW (µm) MD (µm) MW(µm) MD (µm)
60 11.0 48.5 12.2 17.2 14.1 8.15
70 11.2 56.2 14.5 20.4 17.6 39.75
10 8
80 13.5 56.3 15.5 21.6 17.8 42.95
90 15.1 70.6 18.2 26.7 40 46.25

Table 3. Measured responses at different values of speed

s f P Al Cu SS
(mm/s) (kHz) (W) MW (µm) MD (µm) MW (µm) MD (µm) MW (µm) MD (µm)
1 19.3 82.4 22.2 32.65 18.4 16.7
5 18 64.2 20.5 18.65 18.2 16.6
10 90
10 17.4 61.9 18.4 18.4 15.6 16.3
15 15.3 48.4 12.6 17.0 11.4 15.5

261
Comparison of laser marking on aluminum, stainless steel and copper sheet using ND:YVO4 laser

Table 4. Measured responses at different values of frequency

f s P Al Cu SS
(kHz) (mm/s) (W) MW(µm) MD (µm) MW(µm) MD (µm) MW(µm) MD (µm)
4 15.7 83.3 20.7 42.95 28.4 75.1
8 5 90 15.3 68.2 19.3 37.55 27.9 61.2
12 14.5 64.4 13.9 31.6 26.7 58.4
16 13.7 60.2 13.3 22.15 21.3 17.7

Fig. 3 shows variation of process parameters with the responses considered in this experimental
investigation. The variation in marking depth and marking width for different materials depends on their thermal
conductivity and density.

(a) (b) (c)

(d) (e) (f)

Fig. 3. Plots of (a) f vs MW, (b) f vs MD, (c) s vs MW, (d) s vs MD, (e) P vs MW and (f) P vs MD

From Fig. 3(a, b), it is seen that with increase in pulse frequency, marking width and marking depth decrease on
the whole. This may be due to the fact that with increase in pulse frequency, peak power decreases resulting in
decrease in mark width and mark depth. Fig. 3(c, d) show the plots of marking depth and marking width against
scanning speed. A clear trend of decreasing mark depth and marking width with increasing scanning speed can
be observed. Low heat input with high speed is the possible reason behind. It is found from Fig. 3(e, f) that with
increase in laser power, both marking width and marking depth increase. With increase in laser power, heat
input to the work-piece increases, and consequently, large volume of material will be melted and/or vaporised.
Hence, large marking width and marking depth that constitute to the volume of weld material are expected to be
obtained with an increase in heat input.

262
 Proceedings of NCAMMM - 2018

4. Conclusions
From the experimental investigation, the following inferences may be drawn.
 Marking width and marking depth increase with increasing laser power that is a common phenomenon
happening with the increase in heat input.
 With increase in frequency and scanning speed, both marking width and marking depth decrease and this
nature is a usual one.

Acknowledgement
The authors are grateful to the School of Laser Science and Engineering, Jadavpur University, Kolkata and
CSIR-CGCRI, Kolkata for providing the facility for this research work.

Reference
[1] Qi J, Wang KL and Zhu YM, A study on the laser marking process of stainless steel. Journal of Materials
Processing Technology, 2003, 139(1-3): 273-276.
[2] Valette S, Steyer P, Richard L, Forest B, Donnet C and Audouard E, Influence of femtosecond laser
marking on the corrosion resistance of stainless steel. Applied Surface Science, 2006, 252(13): 4696-4701.
[3] Chen MF, Hsiao WT and Huang WL, Laser coding on the eggshell using pulsed-laser marking system.
Journal of Materials Processing Technology, 2009, 209(2): 737-744.
[4] Marx C and Rath T, Investigations on laser marking of plants and fruits. Biosystems Engineering, 2013,
116(4): 436-446.
[5] Amara EH and Noukaz A, Experimental investigations on fiber laser color marking of steels. Applied
Surface Science, 2015, 351: 1-12.
[6] Astarita A, Leone, Minutolo MC and Velotti SC, Study of the laser marking process of cold sprayed
titanium coatings on aluminum substrates. Optics &Laser Technology, 2016, 83:168–176.
[7] Roy A, Kumar N, Bandyopadhyay A and Das S, Empirical modeling and optimization of laser marking of
aluminum sheet using response surface methodology. Proceedings of the ICEM-2016, Ranchi, Jharkhand,
2016.
[8] Roy A, Kumar N, Bandyopadhyay A and Das S, Optimization of pulsed Nd:YVO 4 laser marking of AISI
304 stainless steel using response surface methodology. Proceedings of the 7th International Conference of
Materials Processing and Characterization, Hyderabad, India, 2017: Materials Today: Proceedings
(Accepted for publication).
[9] Roy A, Kumar N, Bandyopadhyay A and Das S. Exploring Surface Roughness Obtained during Laser
Engraving. Proceedings of the National Conference on Advanced Functional Materials Processing &
Manufacturing (NCAFMPM-2017), Durgapur, 2017.

263
 THEME

Precision Engineering
and Metrology
 Monitoring and Control in
Manufacturing
 Statistical Quality Control and
Optimization
 Signal Processing and
Machine Learning
Sub - theme

Monitoring and
Control in
Manufacturing
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Prediction of Cutting Forces in High Speed Ball-End Milling Considering Inertial Forces

Mithilesh K Dikshit*1, Vimal Pathak1, K.J. Uke2, A.B. Puri3, Atanu Maity4

1
Department of Mechanical Engineering, Manipal University Jaipur, Jaipur, India-303007
2
Condition Monitoring and Structural Analysis, CSIR-CMERI, Durgapur - 713209
3
Department of Mechanical Engineering, NIT Durgapur, Durgapur – 713209
4
Engineering Design Group, CSIR-CMERI, Durgapur – 713209
*Corresponding author: dixit.mithilesh@gmail.com

Abstract: In the present research inertial force compensation method has been employed to overcome the dynamic
effect of the workpiece together with dynamometer. Cutting forces (tangential, radial and axial forces) in high speed
ball end milling have been compensated by considering inertia forces of workpiece-dynamometer’s equivalent mass.
The accelerations of the workpiece in machining coordinate system have been measured with accelerometers.
Cutting force coefficients have been determined using average forces based on mechanistic method. The proposed
method enables accurate determination of cutting force coefficients for realistic cutting operations in high speed
ball end milling.
Keywords: Ball end milling; cutting forces; inertia force; force coefficients

1. Introduction
Ball end milling is one of the most versatile end milling processes to produce intricate shape in the industries such as
automotive, aerospace, and die and mould etc. In general, high speed machining (HSM) may be defined on the basis
of cutting speed. In HSM cutting speed is significantly higher than those used in conventional machining. Schulz
and Moriwaki [1] had done first comprehensive study of high speed milling for the manufacturing of die and mould
and concluded that better productivity can be obtain using high speed milling. Dewes and Aspinwall [2] had
observed that higher productivity could be achieved by using HSM. The authors had observed that the
manufacturing cost was reduced significantly while using HSM compared to traditional machining processes and
nonconventional machining processes (like EDM, ECM etc.). In addition, certain specific part configurations such
as thin webs which cannot easily obtainable by conventional machining can be effectively produced by HSM.
In end milling, extensively used force models are: lump mechanism force model and mechanistic force
model. In former force model, a single cutting force coefficient is adopted which is the combined form of shearing
effect on rake face and ploughing effect at the flank edge while in the later force model, the cutting forces are
distinguished into shearing forces and edge forces. Edge forces arise due to the ploughing action at flank edge which
is function of cutting edge length (dS) in contact with the workpiece. Shearing action is responsible for metal failure
which occurs on the rake face and depends on chip cross sectional area. In both existing cutting force models, the
cutting forces are directly related to uncut chip cross sectional area, with cutting force coefficients. Thus, prediction
of cutting force coefficients is essential to estimate cutting forces in ball end milling. Precise determination of
cutting force coefficients leads to accurate prediction of cutting forces and consequently in process simulation.
Within this scope, a large number of researches have been done for the determination of cutting force coefficients.
In general, two basic approaches are used: (1) Mechanics of cutting and (2) Mechanistic approach. A good number
of researches have been done on mechanistic approach based on average cutting forces in slot milling for the
Prediction of cutting forces in high speed ball-end milling considering inertial forces

identification of cutting force coefficients. Using average cutting forces, Zhu et al. [3] characterized the specific
cutting energies in ball end milling. Wan et al. [4] proposed a two-fold solution for the calibration of instantaneous
cutting force coefficients for general end mills. Lazoglu and Liang [5] presented an analytical expression for
dynamic cutting forces for multi-flute ball end milling cutters in frequency domain. Dhupia and Girsang [6]
employed least squares approximation method to estimate cutting force coefficients in ball end milling. Cutting
forces in ball end milling of sculptured surfaces was estimated by Sun et al. [7]. A chip thickness model was
proposed considering tool path and feed rate scheduling Tools of varying geometries were used. Gao et al. [8] had
applied cubic polynomial curve fitting method to determine cutting force coefficients in bull-nose end milling.
Cutting force determination was based on mechanistic approach employing average force per revolution of tool.
Wang et al. [9] had studied the effect of cutting speed in flat end milling process and reported insignificant effect of
cutting speed on the magnitude of cutting forces. However, the experiments were conducted on low speed range (70-
95 m/min).From the previous researches, it has been observed that most of the researches, cutting force coefficients
have been evaluated at low cutting speeds (15-65 m/min). Since, in mechanistic approach cutting forces are
primarily depending on uncut chip cross sectional area. At higher rotational speed chip formation mechanisms may
be affected and may be responsible for possible change in cutting force coefficients [10]. In context to this, Dikshit
et al. [10] determined specific cutting force coefficients in ball end milling considering wide range of rotational
speed and reported that shear cutting force coefficients decreases on increase in rotational speed.
In the present research, cutting force coefficients in ball end milling have been determined based on
average cutting force using linear mechanistic model. An inertial force compensation method has been employed to
overcome the dynamic effect of the workpiece together with dynamometer. The inertia force of workpiece together
with dynamometer’s equivalent mass has been taken into account and cutting force coefficients have been
determined using the compensated forces in tangential, radial and axial directions. The accelerations of the
workpiece in machining coordinate system have been measured with accelerometers. Cutting force coefficients have
been determined using the polynomial curve fitting method, Dikshit et al. [11]. The proposed method allows fact
and accurate determination of cutting force coefficients in the desired range of axial depth of cuts for realistic
cutting operations in industrial machining process.

2 Mechanistic Cutting Force Model

Mechanistic force model is well established and has been adopted by various researchers for different geometrical
shape of milling end cutters.The mechanistic cutting force model include booth shearing and ploughing forces. The
ball end milling cutter consists of cylindrical and spherical parts. On the spherical part, the helix angle of the cutting
edge is not constant and varies with axial depth of cut. To simplify these complexities, a uniform discretization of
the cutting edge on ball end mill has been done along the axis of the cutter into a finite number of disks as shown in
Fig. 1 (a).

265
 Proceedings of NCAMMM - 2018

(a) (b)

Fig. 1 (a) Front and (b) top view of ball end milling cutter

The cutting edges lie on the spherical portion of the cutter. Let us consider an elemental ith cutting edge ‘P’ on jth
disc. The position of elemental cutting edge P can be located by elevation (z) from tool tip, radial distance (Rz)
which can be measured from tool axis, axial immersion angle (κ) of the cutting edge, and radial immersion angle (β)
shown in the Fig. 1 (b). The radial position of the ith element ‘P’ on the jth cutting edge as function of elevation ‘z’
may be given by:

β ( z )= θ + ( j - 1) φ - δ ( z ) (1)
ij p i

where ‘θ’ is the spindle rotation angle, φp is the pitch angle (φp = 2π/N), j is tooth number (0, 1 for a two flute
cutter), δi(z) is radial lag angle varying between 0 and δ0, and varies with the local helix angle ψi for ith element. In
most of the literatures, chip thickness formulation proposed by Martellotti for the straight end mills has been
adopted and the same was modified for ball end milling process. In the present research, instantaneous uncut chip
thickness model for ball end milling process has been adopted based on trochoidal tool path geometry of cutting
edge as reported by Dikshit et al.[11].In the present study mechanistic cutting force model has been adopted in
which two fundamental factors namely shearing effect and edge effect have been considered [11]. As stated above,
due to the complicated geometry of ball-end milling cutter, the cutting forces vary in different discs. For each axially
discretized disc, the cutting forces may be regarded as an independent element. The total force acting on the cutter
may be evaluated through the numerical integration of the forces acting on each discretized disc along the axis of the
tool. To avoid the radial run-out cutting force in one revolution of the cutter was recorded. Average cutting forces
(in X, Y, and Z- directions)can be obtained by integrating differential forces along the axis of ball end milling cutter
in the engagement region over one revolution of the cutter and can be given by [10]:

dz fdz dz fdz dz fdz

F ( x) =
− [ K Q + kte P ] + [ K rs B + Kts C ] ; F ( y ) =[ K re P − kteQ ] + [ − K rs C + Kts B ] ; F ( z ) =− [ K ae A] + [ K as P ]
φ p re 4φ p φp 4φ p φp φp

(2)

Where,

266
 Prediction of cutting forces in high speed ball-end milling considering inertial forces

β β β β
P cos ( β )  ex =
=
β st
; Q sin ( β )  ex =
β st
;A [ β ]ββexst =
;B cos ( 2 β )  ex =
β st
; C  2 β − sin ( 2 β )  ex
β st
(3)

From Eq. (2) it can be seen that the shear forces and ploughing forces are segregated and these are expressed as
linear function of feed. Similarly, experimental forces may also be segregated into edge forces and shear forces.
Thus, the average force components can be given in general as:

Fx = Fxe + f × Fxs ; Fy = Fye + f × Fys ; Fz = Fze + f × Fzs (4)

Hence, on comparing theoretical cutting forces (Eq. (2)) and experimental cutting forces (Eq. (4)), the cutting force
coefficients can be given by:
4φP
Fxs − CKts
4φP BFys + CFxs ap φP Fzs
=Kts = ; K rs = ; K as
a p B2 + C 2 B ap P
(5)
φP
  PKte −
Fye
φP  PFye − QFxe  ap φP Fze
= = =
( )
Kte ; K re ; K ae
a p  P 2 + Q 2  P ap A
 
Where, Kts, Krs, Kas, Kte, Kre, Kae are the cutting force coefficients where ‘s’ represents shear and ‘e’ represents edge
coefficients respectively. Average cutting forces ( FX ,Y , Z ) have been measured from experiments as discussed in

section 4.

3. Inertia Mass Force Compensation

The force signals measured at higher rotation speeds might acquire signals which are characterized by frequencies
components that could approach the dynamometer’s resonance frequency or even higher. Thus, it could be result in
significant distortion in the measured force signals. Many compensation methods have been reported previously viz.
frequency response matrix also known as Transmissibility, Kalman filter estimation and accelerometrical
compensation method. Castro et al. [12] proposed a correction method based on transmissibility method and error
removed from the dynamometer dynamic readings using a mathematical approach. Giradin et al. [13] latter use this
method including crosstalk contributions.

In accelerometrical compensation method, the forces (inertia forces) generated due of the workpiece-
dynamometer’s equivalent mass must be compensated to minimize the dynamic effect of the dynamometer. In this
process, inertial components of the cutting forces should be removed from the measured forces. The accelerations of
the workpiece together with dynamometer are measured in a global reference system using accelerometers. In this
context, very few researches are available in previous literatures regarding the force compensation using
accelerometrical compensation or inertial compensation. This process was firstly reported by Lapoujolade et al. [14].
The accelerations of the top plate of the dynamometer along with workpiece was measured and finally the author
had removed the inertial components of the forces measured with the dynamometer. Omitting the inertia forces, the
modified cutting forces were given by:
=
Fmod Fmeasure − Finertia (6)

267
 Proceedings of NCAMMM - 2018

Where Fmod, Fmeasure, and Finertia are modified cutting forces, measured cutting forces through dynamometer and
inertia force along with equivalent mass of dynamometer and workpiece, respectively.

4. Experimental Setup and Results

High speed ball end milling tests are performed on aluminum alloy of grade Al2014-T6 at different spindle speeds
and feeds. Feed rates were chosen for finishing to semi-roughing operations and are varied linearly. All the tests
have been conducted on a 3-axis vertical CNC milling machining centre, a Mikron, VCP 710.Two fluted solid
carbide ball end mill cutter having 10 mm diameter and 30o helix angle with TiAlN mono layer coating
(CoroMillPlura series, Sandvik) has been used for conducting the experiments. A 3-component dynamometer
(Make: Kistler, model 9257B) has been used for force data acquisition. A multichannel charge amplifier of type
5070A has been employed for getting the output in terms forces.Dynamometer together with workpiece was rigidly
clamped on the vertical milling machine’s bed. In addition, two accelerometers Type VIB 6.142R/DEX have been
used to measure the acceleration of workpiece in high speed ball end milling process. The accelerometers have been
placed at right angle to each other on the workpiece. One is in the direction of main feed direction and other in the
direction of cross-feed direction. VibXpert 6 has been used for data acquisition and acceleration recording in both
time and frequency domain. Experimental setup along with all the components are shown in Fig. 2.

a b e

Dynamometer Charge amplifier

c d

Workpiece VibXprt Accelerometer

Fig. 2. Experimental set up: (a) Dynamometer, (b) charge amplifier, (c) workpiece and (d) VibXprt (e)
accelerometer sensor

Several tests are conducted using different rotational speeds and axial depth of cuts. For each tests local rotational
speed has been employed. Five different feeds have been selected within the range as specified by the tool
manufacturers. Machining conditions used for force and acceleration measurements are shown listed in Table 1.

Table 1. Selected parameters for ball end milling experiments

Cutter Ball-end mill: Φ 10 mm
SandvikCoroMillPlura: R216.42-10030-Al10G 1620
Workpiece Al2014-T6
Cutting speed (rpm) 4586, 5026, 6880, 10052, 14210
Feed per tooth (mm/tooth) 0.02, 0.07, 0.12, 0.17, 0.22
Depth of cut (mm) 0.2, 0.6, 1, 1.4, 1.8
Radial depth of cut (mm) 0.3, 0.5
The cutting force coefficients are determined using the average cutting forces (Eq. 5). The average cutting forces
have been measured in high speed ball end milling experiments using the parameters indicated in Table 1. The
maximum acceleration of workpiece together with dynamometer in X and Y-directions are recorded as 0.36 m/s2

268
Prediction of cutting forces in high speed ball-end milling considering inertial forces

and 0.31 m/s2 respectively. Z-direction acceleration is found to be very small which can be neglected. Mass of the
workpiece together with the dynamometer was taken prior to experiment and was found to be 11.7 kg. After every
cutting experiment mass was recorded.Cutting forces obtained from the ball end milling tests are modified by
compensating inertial forces (which is product of mass and acceleration) using Eq. (6).The cutting force coefficients
at 10052 rpm are obtained as:

Table 2. Determined values of cutting force coefficients

Kts Krs Kas Kte Kre Kae
1221.2 -275 -561 9.06 69.24 31.93
Cutting Forces (N)

88

38

-12
0.1 0.11 0.12 0.13 0.14 0.15
Time (S)

Fx-comp Fy-comp Fz-exp

Fig. 3 Experimental and compensated cutting forces at axial depth of cut 0.6 mm, feed per tooth 0.17 mm,
radial depth of cut 0.7 mm and cutting speed 10052 rpm
Experimental cutting forces without inertia force compensation and with compensation are shown in Fig. 3. Results
have been validated through prediction of the cutting forces using the determined cutting force coefficients obtained
from compensated forces as shown in Table 2. The predicted cutting forces are shown in Fig.4. From Fig. 4, it can
be seen that the predicted cutting forces are much closer to the cutting forces compensated with inertia forces. The
maximum deviation of the predicted forces from the experiential forces with compensation have been estimated as
6.3 %, 4.93% and 11.1% in Fx, Fy and Fz respectively.

5. Conclusions
In high speed ball end milling process, due to high rotational speed, high harmonic may generate at entry and exit of
the cutter into/from the workpiece, which is in the neighbourhood of the dynamometer’s natural frequency and
causes dynamic noise in cutting force measurement. An inertial force compensation method has been employed to
overcome the dynamic effect of the workpiece with dynamometer.

269
 Proceedings of NCAMMM - 2018

Fig. 4 Predicted cutting forces obtained after compensated cutting force coefficients at depth of cut 0.6 mm,
feed per tooth 0.17 mm, radial depth of cut 0.7 mm and cutting speed 10052 rpm.
The inertia force of workpiece together with dynamometer’s equivalent mass has been taken into account and
cutting force coefficients have been determined using the compensated forces. Cutting force coefficients in ball end
milling have been determined based on average cutting force using linear mechanistic model. The following
conclusions have been drawn from the present research:

1) In high speed ball end milling, tooth passing frequency is very close to the natural frequency of the
dynamometer. Therefore, accuracy of measurements can be improved by compensating inertia forces.
2) Dynamic effect of dynamometer along with workpiece has been compensated with inertia forces. Cutting
force coefficients have been obtained using the compensated forces using average force method.
3) Predicted cutting forces show good agreement with the experimental cutting forces compensated with
inertia forces, as compared to non-compensated experimental forces.

References
[1] Schulz H, Moriwaki T. High-speed machining. CIRP Annals-Manufacturing Technology. 1992 Jan
1;41(2):637-43.
[2] Dewes RC, Aspinwall DK. A review of ultra high speed milling of hardened steels. Journal of materials
processing technology. 1997 Sep 1;69(1-3):1-7.
[3] Zhu R, Kapoor SG, DeVor RE. Mechanistic modeling of the ball end milling process for multi-axis
machining of free-form surfaces. Journal of Manufacturing Science and Engineering. 2001 Aug
1;123(3):369-79.
[4] Wan M, Zhang WH, Qin GH, Tan G. Efficient calibration of instantaneous cutting force coefficients and
runout parameters for general end mills. International Journal of Machine Tools and Manufacture. 2007
Sep 1;47(11):1767-76.
[5] LazogluI, Liang SY. Analytical modeling of force system in ball-end milling. Machining. Sci. & Tech.
1997; 1(2): 219-234.
[6] Dhupia J,Girsang I.Correlation-based estimationof cutting force coefficients for ball-end milling, Mach.
Sci. & Tech. 2012;16(2):287-303.
[7] Sun Y, Ren F, Guo D, Jia Z. Estimation and experimental validation of cutting forces in ball-end milling of
sculptured surfaces. Int. J. Mach. Tools Manuf.2009;49:1238–1244.
[8] Gao G, Baohai W, Dinghua Z, Ming L.Mechanistic identification of cutting force coefficientsin bull-nose
milling process. Chinese Journal of Aeronautics. 2013;26(3):823–830.
[9] Wang F, Zhao J, Li A. Experimental study on cutting forces and surface integrity in high-speed side milling
of Ti-6Al-4V titanium alloy. Machining Sci. & Tech. 2014;18:448-463.

270
Prediction of cutting forces in high speed ball-end milling considering inertial forces

[10] Dikshit MK, Puri AB, Maity A. Analysis of rotational speed variations on cutting force coefficients in high
speed ball end milling. J Braz. Soc. Mech. Sci. Eng. 2017;39(9):3529-3539.
[11] Dikshit MK, Puri AB, Maity A, Banarjee AJ.Determining cutting force coefficients from instantaneous
cutting forces in ball end milling. Int. J. Machining and Machinability of Mat. 2016;18(5-6):552-571.
[12] Castro LR, Vieville P, Lipinsik P. Correction of dynamic effects on force measurements made with
piezoelectric dynamometers.Int. J. Mach. Tools Manuf. 2006;46:1707-1715.
[13] Girardin F, Remond D, Rigal J. High frequency correction of dynamometer for cutting force observation in
milling. J. Manuf. Sci.Engg. 2010;132(3):312002-8.
[14] Lapoujolade F, Coffignal G, Pimont J. Cutting force evaluation during high speed milling. In: proceeding
of 2nd international conference IDMME. May 27-29, 1998.

271
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Electromechanical Characterization of Dielectric Elastomer Actuator based Pump for

Optimum Volume Flow Rate

Amit Kumar, Dilshad Ahmad, Karali Patra*

Indian Institute of Technology Patna, Bihta, Patna 801103, India
*
Corresponding Author: kpatra@iitp.ac.in

Abstract: Recently dielectric elastomer is potentially used for actuator, sensor and energy harvesting applications
due to its large deformation when stimulated electrically. However, developing a dielectric elastomer actuator
based pump to get optimum volume flow rate is still a challenging task. The present work is an experimental attempt
to determine the optimum volume flow rate of dielectric elastomer actuator based pump with operating voltage,
initial dome deflection and active area percentage. Experimental results show that an operating voltage of 8 kV,
active percentage area of 45% and initial dome deflection of 30 mm can obtain optimum volume flow rate for the
dielectric elastomer actuator based pump. The highest volume flow rate obtained is 320ml/min.

1. Introduction
Fluid pumps have large numbers of application in the field of automobile, aerospace, biomedical, drug delivery,
electronic device cooling, biological science etc. The wide applications motivate the researchers to better understand
the fundamental science of these pumps for high performance design. Conventional motor pumps are often large,
heavy and noisy to be fit within the system. Attempting to solve the problems in biomedical, aerospace and other
engineering as well as biological zone, materials are very important and definitely the field of material science
developed new advanced material for research. Potentially one of the best and alternatives actuator materials
available to achieve higher actuation has been dielectric elastomer. When subjected to a high electric field dielectric
elastomer membrane sandwiched between compliance electrodes undergoes large deformations with a fast response
speed [1–3]. Researchers have been developing various suitable methods to achieve higher and higher planer
actuation of a dielectric elastomer actuators [2,3]. Strains over 100% have also been obtained in many ways in past
decades e.g., by providing pre-stretch to elastomer before electrode coating(4), by swelling an elastomer with a
solvent(4),and by spraying charge on an electrode-free elastomer [5]. Further harnessing snap-through instability in
dielectric elastomer made possible to achieve a giant areal strain of 1692% [6].
Dielectric elastomer due to its flexibility, high specific energy density, high actuation strain [7], high
toughness [8] and noise free operation makes it an exciting material for automobile, aerospace and electronic device
cooling etc. high flow rate actuator based pump. Also, dielectric elastomer membrane has gained importance to be
applied as pump for fluidics and biomedical applications [9]. Despite of its potential use very few work has been
done to obtain optimum volume flow rate in dielectric elastomer actuator based pump. Cabuz et al. developed
dielectric elastomer actuator based pumps earlier that have flow rate in few micro litres per minute [10] Feng et al.
developed piezoelectric actuator based micro pump and it is demonstrated to have a flow rate of 30ml/min [11].
Since no one has investigated the effect of voltage, initial dome shape and active area percentage on the
volume flow rate, the present work is to lead the concept for dielectric elastomer actuator based pump with the
 Electromechanical characterization of dielectric elastomer actuator based pump for optimum volume flow rate
method of higher planer actuation to achieve high flow rate. The work is focused towards the effect of parameters
(e.g., Peak to peak voltage, initial dome deflection and different ratio of active area) on the volume flow rate.
However the initial parameters have been approximated from literatures [1,9] which leads to maximum value that
we can achieve without any instability in thedielectric elastomer diaphragm used in pump.

2. Working Principles
The basic principle of operation of dielectric elastomer actuator is the columbic electrostatic force. When an electric
field is applied across the compliance electrodes of a sandwiched dielectric elastomer actuator, charge flows through
an external conducting wire from one electrode to another and accumulates on the electrodes. The charges of the
opposite signs on the two electrodes generate electrostatic force between the electrodes and cause the membrane to
deform. This electrostatic force between two electrodes compresses the elastomer that reduces the thickness of the
elastomer and expands the active area of the material due to material incompressibility. The electrostatic pressure
which is also called Maxwell stress is given as follows [12]:
2
V 
σ M ε=
= rε o E
2
ε r ε o   (1)
t 

Where, ε r is the relative permittivity or dielectric constant of the material and ε o is the permittivity of the free space

and E is the applied electric field, defined as the applied voltage V divided by the thickness of the elastomer t .

The dielectric elastomer diaphragm under static pressure in pump can be modelled as circular plate with clamped

edge of which the flexural rigidity D and maximum out of plane deflection Wmax is given by equation (2) and
equation (3) respectively (13).

E × t3
D=
12 (1 − υ 2 )
(2)

q × r4
Wmax = (3)
64 D

Where, E is the modulus of elasticity, υ is the Poison’s ratio and r is the radius of circular dielectric elastomer
diaphragm and q is the load intensity uniformly distributed over the surface.

On application of electric field, thickness of DEA diaphragm t decreases resulting in decrease in its
flexural rigidity D which intern increases maximum out of plane deflection of diaphragm. With application of
varying electric field, a pulsating deflection of DEA diaphragm is obtained.

3. Experimental Setup and Sample Preparation

Figure 1 shows the experimental setup developed in the laboratory to perform all the experiments. The dielectric
elastomer samples (VHB 4910) are pre-stretched to 100% and carbon grease of diameter 25 to 40 mm is applied on
both sides of these prestretched specimens. The specimens were then fitted on the cylindrical chamber equipped

273
 Proceedings of NCAMMM - 2018

with pressure sensor and inlet pressure control valve. A laser displacement sensor was fitted to determine the
dynamic actuation of elastomer in real time mode. All the data was attained through LabVIEW data acquisition
system connected with a computer.

Figure 1: Schematic diagram of experimental set up

Axis of
Laser Displ. Computer symmetryy
NI-
Sensor Volume
9219
Analog flow-rate
Current Image

(a) Camera (b)

Figure 2: (a) LabVIEW and data acquisition interface (b) deflected dome shape diaphragm of dielectric elastomer

Table 1: Components and its specifications used in the experiment

Component Specifications
VHB Tape (VHB4910) Make: 3M, Thickness : 1mm
Density : 960 kg/m3
Laser Displacement Sensor Make : RIFTEK, Model : RF603-60/100
Range : 60mm-160mm, Output : 4-20 mAmp
Webcam Make : LOGITECH, Model : C920 HD
Resolution : 1920x1080 pixels
Frame rate : 15 fps
NI-9219 (AI UNIVERSAL) Make : NATIONAL INSTRUMENTS
Input Voltage Range : ±60V
Input Current Range: ±25mA
High Voltage Converter Make : XP Power
Model : Q101, Maximum o/p curr:50μA

Sample preparation is the important part of testing dielectric elastomer. Proper stretching with good coating of
electrode is necessary for high performance of dielectric elastomer actuator. It is observed that for increasing the
actuation strain of DEA, pre-straining of elastomer is required. Pre-straining can be done through uniaxial, bi-axial
and pure shear ways. Pre-straining causes the reduction in thickness due to which there is increase in actuation strain
as the Maxwell stress increases for the corresponding thickness value (14). During electromechanical actuation at
certain electric field, there is exponential thinning of DEA occurs until catastrophic failure, called electromechanical

274
Electromechanical characterization of dielectric elastomer actuator based pump for optimum volume flow rate
instability (EMI). For the experiments the sample are bi-axially pre-stretched up to 100%, 200% and 300% of its
initial size to see the variation in the mid-point deflection of the dome shape of diaphragm. The detailed description
of all the components and its specifications are given in Table 1.

4. Results and Discussion

For characterization of dome shaped dielectric elastomer diaphragm for pump application, mid-point deflection of
different dome shaped samples was acquired from laser displacement sensor and stored into the PC. Thus, pulsating
deflection with higher amplitude which gives rise to the possibility of a pumping action with enhanced flow rate is
obtained. It is also observed that there is lag between input sine wave and output deflection due to viscoelastic loss
in DE diaphragm as reported in earlier work (9).A MATLAB code is used to calculate the area and its geometric
centroid which acted as the input for the Pappus’s theorem which provides a relation to calculate the volume from 2-
D image for any axisymmetric area (16). The deflection observed is also axisymmetric as shown in Fig. 2(b) and
also reported earlier(15). Hence continuous frames of the video are processed to get the data of volume in different
frame and in turn at different point of time as shown in figure 2 (a). From the pulsating displacement of the
diaphragm we can plot the volume flow rate which is the difference of maximum and minimum volume covered
under the diaphragm.

(a (b

Figure 3(a) Volume flow rate (ml/min) vs. Peak to Peak voltage (kV). (b) Volume flow rate (ml/min) at diffent
initial dome deflection

Different experiments were repeated at 100% pre-stretched sample at 1 Hz frequency and at different peak to
peak voltage. Fig. 3(a) shows the plot of the result. It is observed that the flow rate increases with increase in peak-
to-peak voltage as the Maxwell stress increases with increasing voltage resulting in increase in deflection. Also
increasing the voltage across DE elastomer diaphragm beyond 8 kV results in electromechanical instability (EMI) of
the sample at early stage of actuation. Another set of experiments were performed with 100% pre-stretched sample
with varying initial dome deflection keeping the peak to peak voltage at 8 kV and frequency 1 Hz. With the repeated
experiment of dome shaped diaphragm at different initial dome deflection, it is observed that the volume flow rate
increases with increase in initial dome deflection as plotted in Fig. 3(b). Further increase in initial dome deflection
results in clotting of carbon particles on the surface of pre-stretched DE elastomer. Maxwell stress act upon the area
which acts as a capacitor i.e., the area coated with compliant carbon grease also termed as active area. Hence another
set of experiments performed to observe the effect of change in active area percentage upon the volume flow rate.

275
 Proceedings of NCAMMM - 2018

350
4 KV 6 KV 8 KV
Volume flow rate (ml/min)
300

250

200

150

100

50

0
0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70%
Active area percentage

Figure 4: Volume flow rate vs Percentage Active area

Figure 4 shows an increasing trend is up to 45% of active area as there is enough area to compensate the decrease in
thickness due to Maxwell stress but going beyond that result in wrinkling on the peripheral surface causing sudden
drop in volume flow rate. The expected reason for the wrinkling is the clamped edge of the circular diaphragm
which does not allow any further expansion of active area on application of Maxwell stress due to applied voltage.

5. Conclusion

The present work successfully investigated the effects of electromechanical parameters like voltage, active
percentage area and initial dome deflection for getting higher volume flow rate in dielectric elastomer actuator based
pump. It is experimentally verified that an applied voltage of 8kV, active percentage area of 45% and initial dome
deflection of 30mm are the desired parameters to obtain optimum volume flow rate for dielectric elastomer actuator
based pump. The highest volume flow rate obtained is about 320ml/min which is quite high compared to earlier
literature. This work will be useful for designing a dielectric elastomer actuator based pump to get higher volume
flow rate for applications in automobiles, aerospace and electronic device cooling.

References
[1] Carpi F, Rossi DD, Kornbluh R, Pelrine R, Larsen PS. Dielectric Elastomers as Electromechanical Transducers.
Vol. 53, Journal of Chemical Information and Modeling. 2013. 1689-1699 p.
[2] Sahu RK, Sudarshan K, Patra K, Bhaumik S. Evaluation of area strain response of dielectric elastomer actuator
using image processing technique. 2014; 9056:90562C.
[3] Patra K, Sahu RK. A visco-hyperelastic approach to modelling rate-dependent large deformation of a dielectric
acrylic elastomer. Int J Mech Mater Des. 2015;11(1):79–90.
[4] Pelrine RE, Kornbluh RD, Joseph JP. Electrostriction of polymer dielectrics with compliant electrodes as a
means of actuation. Sensors Actuators A Phys [Internet]. 1998;64(1):77–85. Available from:
http://www.sciencedirect.com/science/article/pii/S0924424797016579
[5] Keplinger C, Kaltenbrunner M, Arnold N, Bauer S. Rontgen’s electrode-free elastomer actuators without

276
Electromechanical characterization of dielectric elastomer actuator based pump for optimum volume flow rate
electromechanical pull-in instability. Proc Natl Acad Sci [Internet]. 2010;107(10):4505–10. Available from:
http://www.pnas.org/cgi/doi/10.1073/pnas.0913461107
[6] Keplinger C, Li T, Baumgartner R, Suo Z, Bauer S. Harnessing snap-through instability in soft dielectrics to
achieve giant voltage-triggered deformation. Soft Matter [Internet]. 2012;8(2):285–8. Available from:
http://xlink.rsc.org/?DOI=C1SM06736B
[7] Sahu RK, Saini A, Ahmad D, Patra K, Szpunar J. Estimation and validation of maxwell stress of planar
dielectric elastomer actuators. J Mech Sci Technol. 2016;30(1):429–36.
[8] Ahmad D, Patra K. Fracture Behavior of Dielectric Elastomer under Pure Shear Loading. IOP Conf Ser Mater
Sci Eng [Internet]. 2017; 229:12035. Available from: http://stacks.iop.org/1757-
899X/229/i=1/a=012035?key=crossref.6669a6281f1b926b1f035ad3db16ba45
[9] Saini A, Ahmad D, Patra K. Electromechanical performance analysis of inflated dielectric elastomer membrane
for micro pump applications. Process SPIE. 2016;9798:979813.
[10] Cabuz C, Herb WR, Cabuz EI. The dual diaphragm pump. Tech Dig MEMS 2001 14th IEEE Int Conf Micro
Electro Mech Syst (Cat No 01CH37090) [Internet]. 2001; 519–22. Available from:
http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=906593
[11] Feng GH, Kim ES. Piezoelectrically actuated dome-shaped diaphragm micropump. J Microelectromechanical
Syst. 2005;
[12] Pelrine R. High-Speed Electrically Actuated Elastomers with Strain Greater Than 100%. Science (80- ).
2000;287(5454):836–9.
[13] Timoshenko S, Woinowsky-Krieger S. Thoery of Plates and Shells. 1959. p. 591.
[14] Kofod G. The static actuation of dielectric elastomer actuators: how does pre-stretch improve actuation? J Phys
D Appl Phys [Internet]. 2008; 41(21):215405. Available from: http://stacks.iop.org/0022-
3727/41/i=21/a=215405
[15] Li Z, Wang Y, Foo CC, Godaba H, Zhu J, Yap CH. The mechanism for large-volume fluid pumping via
reversible snap-through of dielectric elastomer. J Appl Phys. 2017;122(8).
[16] Weisstein, Eric W. "Pappus's Centroid Theorem." From MathWorld - A Wolfram Web
Resource. http://mathworld.wolfram.com/PappussCentroidTheorem.html

277
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

R&D Activities for Enhancing New Product Quality: A Combined Approach of Analytic
Hierarchy Process and Structural Equation Modeling Approach

Sudeshna Roy*1, Nipu Modak1 and Pranab K Dan2

*1, 1
Mechanical Engineering Department, Jadavpur University, Kolkata, India
2
Rajendra Mishhra School of Engineering Entrepreneurship, IIT Kharagpur, India

Abstract: Make in India initiates of making India a global manufacturing hub by encouraging the large as well as
small and medium enterprises (SMEs) to manufacture their own products within the country. The development of
manufacturing sectors enlightens the necessity of new product development (NPD) for achieving competitive
advantages in global perspectives. Research and development (R&D) activities has become an inevitable
phenomenon to introduce new products for industrial sustainability. This study recognizes the fundamental factors
such as technological advancement and intellectual capital closely associated with R&D activities and their impact
on NPD success. Like success factors, cost, quality and development time of new products are identified as success
measures to quantify the NPD success of the firm. Analytic hierarchy process (AHP) has been applied to prioritize
the success factors as well as success measures depicting the R&D as the most important factor succeeded by
technological developments and intellectual capital. Moreover, product quality is analyzed as the most vital
measure among the three measures. This empirical study assimilates the primary data from 76 Indian
manufacturing companies to develop the interrelationship model inferring the effect of R&D activities and its
associated factors on the quality of the new product using structural equation modeling (SEM) approach. The result
expresses the positive effect of R&D activities on product quality which is again be influenced by the technological
developments and intellectual capital.

Keywords: R&D activities, technological advancement, intellectual capital, AHP, SEM

1. Introduction
New product development (NPD) offers industrial sustainability of the firm by introducing innovative products as
per market demand. It ensures better firm’s performance for sustaining in the volatile and competitive market
environment (Buyukozkan and Arsenyan, 2012). There are various crucial factors termed as critical success factors
(CSFs) influence the NPD success of the firm which is again quantified by the success measures. Research and
development (R&D) has been previously been discussed as one of the vital CSFs of the NPD for introducing new
products. Continuous R&D practice helps to modify the existing products similarly innovates new products
(Nicholas et al., 2015; Ernst 2010). Learning is considered as a secondary constituent of R&D activities which
directly encourages R&D through sharing knowledge and creative ideas by overcoming cultural barriers (Medeiros
et al., 2014;Roy and The´rin, 2008). The technological advancement helps to provide concrete platform to R&D by
offering required technical resources for NPD(Mendes and Ganaga, 2013). Similarly, intellectual capital is the
intangible asset mainly classified as human capital, relational capital and structural capital related to R&D activities
of the firm (Chen et al., 2006) which indulges the NPD success through assuring high-end competitive advantages
(Zhou and Fink, 2003). Like success factors there are success measures like cost, quality and development time of
R&D activities for enhancing new product quality: a combined approach of analytic hierarchy process and structural equation
modeling approach
new products (Kazerouni et al., 2014; Huang, Soutar and Brown, 2004). This study is a novel approach to develop a
structural framework considering R&D activities and associated factors along with the most vital success measure
prioritized by AHP based on the primary data collected from Indian manufacturing industries.

2. Methodology
The Analytic Hierarchy Process (AHP)is one of the multi-criteria-decision-making method developed by Thomas L.
Saaty (1980). It is a structured technique of pairwise comparisons based on the experts’ opinion for prioritization
(Saaty 2008, Saaty 1987). Structural equation modelling (SEM) is a multivariate statistical approach use to develop
relationship among variables by hypothesis testing (Hoyle, 1995). This is the combination of confirmatory factor
analysis and multiple regressions to test the interdependencies of measured variables on latent constructs (Hair et al.,
2010). By using AHP, this study calculates the weights of the R&D based success factors and identify the most
essential success measures on the basis of the success factors considered. Identifying product quality as the most
vital success measure, SEM approach is adopted to develop a structural model using IBM SPSS AMOS 21.0
software depicting the interrelationship of R&D and its associated factors and their combined effect on product
quality. The primary data from 76 experts of Indian manufacturing companies have been collected for analysis
purpose. This work involves formulation of mentioned hypotheses for developing interrelationships among
aforesaid constructs using SEM based on the primary data set obtained from survey. These hypotheses are:
H1a: R&D has a positive impact on product quality (PQ).
H1b: Learning acts as a secondary construct of R&D and influences R&D activities.
H2a: Technological development (Tech) motivates product quality (PQ).
H2b: Technological development (Tech) escalates R&D activities as well.
H3a: Intellectual capital (IC) positively encourages product quality (PQ).
H3b: Intellectual capital (IC) has a positive impact on R&D activities also.

Fig. 1: Path model of constructs

3. Results
3.1 Weight calculation using analytic hierarchy process
This study identifies the best alternative among the three success factors such as cost, quality and development time
by calculating priority vectors using AHP based on Saaty’s scale as shown in Table 1. This technique also prioritizes
the success factors by calculating their weights and ranks them based on the survey data collected from

279
 Proceedings of NCAMMM - 2018

manufacturing experts shown in Table 2. The list of priority vectors of cost, quality and development time
considering technological development, R&D and intellectual capital are listed in Table 3, 4 and 5 respectively.
Priority vectors for calculating weights of success factors are enlisted in Table 6. Finally, the decision matrix
considering the success factors as criteria and measures as alternatives is listed in Table 7.
Table 1. Saaty’s 9 Point Scale
Scale points Definition of Scale points Explanation
1 Equally Important Two attributes are equally important
3 Somewhat more important One attribute is slightly more important than other
5 Much more important One is moderately more important than other
7 Very much Important One is strongly preferred over other
9 Strongly Important One attribute is extremely more important over the other
2,4,6,8 Intermediate Values

Table 2. Number of respondents from various manufacturing sectors

Firm Level Characteristics Frequency Firm Level Characteristics Frequency
Fabrication 13 Burners and heaters 6
Electrical equipment 10 Material handling equipment 6
Industrial valves 10 Cell and battery 3
Textile machineries 9 R&D sectors 2
Fire fighting equipment 9 Air Ventilators 2
Hydraulics & pneumatic 6
Total number of respondent 76

Table 3.Priority Vectors of cost, quality and development time considering technological developments
C1: Technological Cost Quality Time Priority
Developments Vectors
Cost 1 0.26 1.41 0.19
Quality 3.87 1 3.87 0.654
Time 0.71 0.29 1 0.157
λ max = 3.071, CI = 0.035, CR = 0.061
Table 4.Priority Vectors of cost, quality and development time considering R&D activities
C2: R&D Cost Quality Time Priority
Vectors
Cost 1 0.33 3.16 0.261
Quality 3 1 5 0.636
Time 0.32 0.2 1 0.103
λ max = 3.049, CI = 0.026, CR = 0.042
Table 5.Priority Vectors of cost, quality and development time considering technological developments
C3: Intellectual Capital Cost Quality Time Priority
Vectors
Cost 1 0.33 1.41 0.229
Quality 3 1 2.45 0.575
Time 0.71 0.41 1 0.196
λ max = 3.034, CI = 0.017, CR = 0.029
Finally, comparing the importance of the three decision criteria which are the success factors for NPD, the
judgmental matrix is formed.

280
R&D activities for enhancing new product quality: a combined approach of analytic hierarchy process and structural equation
modeling approach
Table 6. Judgmental matrix comparing the success factors of NPD
Success Factors Tech R&D IC Priority Vectors
Tech 1 0.41 3 0.295
R&D 2.45 1 3.87 0.583
IC 0.33 0.26 1 0.122
λ max = 3.049, CI = 0.025, CR = 0.043
Now, the decision matrix is developed by considering the previous priority vectors of success measures such as
cost, quality and time based measures along and the final priorities of the measures is been calculated.

Table 5. List of manifest variables of latent constructs

Success Factors Tech R&D IC Final Priority
(0.3) (0.58) (0.12)
Cost 0.19 0.261 0.229 0.227
Quality 0.654 0.636 0.575 0.624
Time 0.157 0.103 0.196 0.149

According the calculation, the final priorities of success measures, i.e. cost, quality and time based measures of
NPD are listed as (0.227, 0.624, 0.149) in Table 5. It depicts the quality of the product is the most important measure
of the NPD success followed by the cost and the time based measures. Again the priority vector of success factors of
NPD, i.e. technological developments, R&D activities and intellectual capital is enlisted as (0.295, 0.583, 0.122) in
Table 4. This portrays the weight of R&D activities are much crucial for NPD succeeded by technological
developments and intellectual capital.
3.2 Analysis of measurement validity
A thorough data survey is performed based on the 7 point Likert scale to identify the importance as well as
implementation rate of these R&D activities along with its associated factors.In this questionnaire, 1 stands for
strongly disagree and 7 for strongly agree for denoting importance of the factors whereas 1 for very low and 7 for
very high for implementation as well as measuring product quality. The opportunity to express the own indicators
they have recognized during measuring the constructs and the obstacles they have faced during implementing the
aforesaid success factors offers additional novelty to this study. The reliability and consistency of the survey data is
tested by composite reliability (CR), average variance extracted (AVE) and Cronbach’s Alpha reliability testing
using IBM SPSS 21.0 software. The CR values greater than 0.5 is considered as highly reliable where as its values
in between 0.3 to 0.5 is moderate. For AVE, values greater than 0.5 indicate the reliability. In case of α, values
should be either greater or equals to 0.8 is treated as reliable one and can be further used for analysis purpose (Ong
et al., 2004). The structural model depicting the interrelationship of R&D activities along with its associated factors
and their impact on product quality is developed by SEM approach using IBM SPSS AMOS 21.0.
3.3 Measurement model results
Estimation of model fit has been performed by confirmatory factor analysis. SEM estimates the path values for
analyzing the fitness of developed structural model. Besides, it also performs various model fitness tests like chi-
square including degree of freedom, goodness-of-fit (GFI), adjusted-goodness-of-fit (AGFI) and root mean square of
error approximation (RMSEA). The results of fitness tests depict that the values are within desired range asχ2 =
409.696, degrees of freedom = 224, χ2 /degrees of freedom = 1.829, RMSEA = 0.049, GFI = 0.91, AGFI = 0.892
(Hair et al., 2010). Standardized regression weights (SRWs) of manifests are range from 0.51 to 0.96 and values of

281
 Proceedings of NCAMMM - 2018

reliability and consistency indices such as CR, AVE and αare 0.65 to 0.78, 0.49 to 0.57 and 0.778 to 0.836
respectively which is also reliable. These values are enlisted in Table 7.

Table 7. Constructs and indicators including reliability indices and path estimates
Constructs and manifests including reliability indices SRW
1. Research and Development Activities (R&D) [CR=0.75; AVE= 0.53; α=0.802] -
X1: Qualification and experience of the R&D team members 0.83
X2: Investment in R&D infrastructure and methods for innovations 0.65
X3: Vision and direction of R&D management 0.68
X4: R&D culture 0.87
X5: Number of patents 0.79
X6: Learning (L) [CR=0.69; AVE= 0.49; α=0.778] -
X61: Debriefing of all NPD experiences of NPD team members 0.96
X62: Existence of NPD manuals to assist managerial decision-making 0.87
X63: Active participation in in-house training program 0.81
X64: On-the-job training 0.84
X65: Collective review to report the progress of NPD activities 0.93
X66:Maintenance of database containing factual information 0.69

Constructs and manifests including reliability indices SRW

2. Technological Advancement (Tech) [CR=0.72; AVE= 0.51; α=0.791] -
X7: Investment for up-gradation of technological infrastructure 0.67
X8: Adoption of newly launched and state-of-the-art technologies 0.73
X9: Implementation of flexible manufacturing systems 0.51
X10: Usage of lean manufacturing 0.54
X11: Application of cellular manufacturing 0.65
X12: Adoption of design for manufacturability and assembly (DFMA) 0.86
3. Intellectual Capital (IC) [CR=0.65; AVE= 0.46; α=0.753] -
X13: Enhancement of human capital 0.52
X14: Enrichment of process capital 0.57
X15: Improvement of relational capital 0.61
X16: Betterment of innovation capital 0.59
3. Product Quality (PQ) [CR=0.78; AVE= 0.46; α=0.753] -
X17:Meeting quality guidelines 0.93
X18:Achieved product performance goal 0.74
X19:Achievement of design goal 0.82

3.4 Structural model results

Structural model development has been developed to portray the interrelationship model. This also shows the
appropriate model-to-data fit as χ2 = 462.56, degrees of freedom = 236, χ2 /degrees of freedom = 1.96, RMSEA =
0.043, GFI = 0.886, AGFI = 0.854 (Hair et al., 2010). Structural model represents the hypothesized relationship
between R&D activities along with its associated factors and product quality which in turn escalates NPD success.
Figure 2 shows the structural model developed by IBM SPSS AMOS 21.0. Values of path estimates among the
constructs are enlisted in Table 8.

282
R&D activities for enhancing new product quality: a combined approach of analytic hierarchy process and structural equation
modeling approach

Fig. 2: Structural model after execution

Table 8. Statistics of Path Estimates

Path Description Hypotheses Estimates
R&D PQ H1a 0.76
R&D L H1b 0.56
Tech PQ H2a 0.65
Tech R&D H2b 0.78
IC PQ H3a 0.59
IC R&D H3b 0.63

4. Conclusions
This study highlights the inevitable role of R&D activities for industrial sustainability in Indian manufacturing
industries through a detailed data survey. The AHP priorities the product quality as the most important measure of
NPD success among cost, quality and development time. The structural model developed by SEM depicts the impact
of R&D, learning, technological advancement and intellectual capital on product quality to escalate the NPD
success. The model also represents that learning is a secondary construct of R&D which helps to develop high
quality R&D activities which in turn motivates the product quality for achieving NPD success. Again, technological
advancement and learning both encourages R&D activities as well as better quality of the newly developed product
for NPD success. Moreover, the AHP analyses the weights of the factors and identifies technological advancement
as the critical one succeeded by the R&D and intellectual capital.

283
 Proceedings of NCAMMM - 2018

References
[1] Büyüközkan G, Arsenyan J. Collaborative product development: a literature overview. Production Planning
& Control. 2012 Jan 1;23(1):47-66.
[2] Chen YS, James Lin MJ, Chang CH. The influence of intellectual capital on new product development
performance–the manufacturing companies of Taiwan as an example. Total Quality Management and
Business Excellence. 2006 Dec 1;17(10):1323-39.
[3] Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate Data Analysis: A Global
Perspective, 7th edition. Pearson Education.
[4] Hoyle, R. H. (1995). The Structural Equation Modelling Approach: Basic Concepts and Fundamental Issues,
In structural equation modelling: Concepts, issues, and applications, (pp. 1-15.), R.H. Hoyle (editor),
Thousand Oaks, CA: Sage Publications, Inc.
[5] Lau AK. Critical success factors in managing modular production design: Six company case studies in Hong
Kong, China, and Singapore. Journal of Engineering and Technology Management. 2011 Sep 30;28(3):168-
83.
[6] Medeiros JF, Ribeiro JL, Cortimiglia MN. Success factors for environmentally sustainable product
innovation: a systematic literature review. Journal of Cleaner Production. 2013;1:1-1.
[7] de Sousa Mendes GH, Miller Devós Ganga G. Predicting success in product development: The application of
principal component analysis to categorical data and binomial logistic regression. Journal of technology
management & innovation. 2013 Nov;8(3):83-97.
[8] Nicholas J, Ledwith A, Aloini D, Martini A, Nosella A. Searching for radical new product ideas: Exploratory
and confirmatory factor analysis for construct validation. International Journal of Technology Management.
2015 Jan 1;68(1-2):70-98.
[9] Roy MJ, Thérin F. Knowledge acquisition and environmental commitment in SMEs. Corporate Social
Responsibility and Environmental Management. 2008 Sep 1;15(5):249-59.
[10] Saaty TL. Decision making with the analytic hierarchy process. International journal of services sciences.
2008 Jan 1;1(1):83-98.
[11] Saaty RW. The analytic hierarchy process—what it is and how it is used. Mathematical modelling. 1987 Jan
1;9(3-5):161-76.
[12] Ong CS, Lai JY, Wang YS. Factors affecting engineers’ acceptance of asynchronous e-learning systems in
high-tech companies. Information & management. 2004 Jul 31;41(6):795-804.
[13] Zhou AZ, Fink D. The intellectual capital web: a systematic linking of intellectual capital and knowledge
management. Journal of intellectual capital. 2003 Mar 1;4(1):34-48.

284
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

An Artificial Neural Network Approach for Predicting Flank Wear, Cutting Force, and
Surface Roughness for Turning Operation using Ceramic Tool Insert

Ananda Rabi Dhar1, Bipin Kumar Singh2, Nilrudra Mandal2, Shibendu Shekhar Roy1
1
Mechanical Engineering Department, National Institute of Technology, Durgapur – 713209, West Bengal
2
Centre for Advanced Materials Processing, CSIR-Central Mechanical Engineering Research Institute, Durgapur,
India

Abstract: The role of economical machining in today’s competitive manufacturing sector is well understood.
Machining operation such as turning on lathe is till now not fully analytically understood because of the
involvement of multiple physics in the process and the dynamic nature of the problem. Statistical regression
analysis is used to address the problem with prediction of machinability parameters like flank wear, cutting force,
and surface roughness using ceramic tool inserts. Three separate regression equations for these parameters are
obtained in terms of cutting conditions viz. cutting speed, feed rate, and depth of cut from a previous literature
experimenting with typical ceramic tool insert using Response Surface Methodology (RSM). These equations are
used to produce the large number of sample data within range of operations. The generated sample data is then
used to train a Back Propagation Neural Network (BPNN) built to model the input-output relationship of the same
turning operation. Cutting conditions such as cutting speed, feed rate and depth of cut are selected as input to the
neural network model andflank wear of the inserts, cutting force, and surface roughness corresponding to these
conditions are designed as the output of the network. The trained network is tested with separate test samples
generated by the same regression equations. The mean absolute percentage error (MAPE) during both training and
testing are recorded after altering network topology viz. number of neurons in hidden layers, learning rate,
momentum constant, steepness coefficient in activation function. It is observed that the predictive model performs
reasonably well with the given case.
Keywords: Ceramic Tool Insert, Turning, Regression, Back Propagation Neural Network (BPNN)

1. Introduction
Manufacturing is the most important cause of economic growth. Keeping in view the strong correlation among
machine tool, machining and manufacturing, advances in tool development in parallel with determination of optimal
machining conditions is considered of immense importance now [1]. There have been enormous efforts to optimize
the machining conditions to successfully invest in the development of any new cutting tool especially made of hard
materials like ceramics, carbides, and nitrides. Even traditional machining process such as turning operation often
lies at the crossroads of science and arts. The modern automated CNC machines till now are mostly operated by
skilled human operators for setting the optimal cutting conditions. The most critical parameters for cutting are the
cutting velocity, feed rate and depth of cut for a given tool and work material conjugate. These are to be set so
precisely that a desired surface finish is obtained with maximum throughput and least cost of tooling.
An artificial neural network approach for predicting flank wear, cutting force, and surface roughness for turning operation using
ceramic tool insert

Ceramic tool inserts are gaining attention with every new addition and improvement of material. The tool performs
machining of alloy and high speed steels with improved accuracy without losing material characteristics at an
elevated temperature over a sufficient span of time.
In our study, an Artificial Neural Network (ANN) based approach is used to develop a model to predict the
machinability parameters for the ceramic tool insert from the given cutting conditions. The model uses a Back
Propagation Neural Network (BPNN), which internally uses a supervised learning scheme for updating the
connection weights. The regression equations obtained from the previous work of the author [2] are used to generate
sample data within operation range for training the BPNN. A separate sample test set of data is used to validate the
model. The mean absolute percentatge error (MAPE) during the training and testing is observed after altering the
network topology parameters. The optimal configuration is selected for the predictive model.

2. Literature Review
Many distinguished researchers have invested dedicated researches to understand the mechanism of machining fully
by using all sorts of methods – analytical, statistical, and soft computing based approaches.
Several analytical models for the analysis of the turning process had been developed by various researchers [3-4].
The exact solutions of the differential equations obtained thereby correlating the input parameters and physical
phenomena of turning process are too difficult to obtain. The involvement of multiple physics such as
thermodynamics, fluid mechanics, and structural dynamics makes the analysis much harder. Consequently, the
simulation models developed so far by using different software packages suffer huge deviations from the
experimental outcomes.
Nowadays, statistical regression analysis (from the experimental data according to some DOE) is widely
used [2, 5-7] to predict the machinability behaviour for different types of advanced tool developments. Mandal et al.
[5-7] worked on development of statistical regression models using RSM for predicting flank wear, cutting forces,
and surface roughness for tuning operation with newly developed ceramic tool insets.
Currently, neural network based expert systems are being largely applied in industry, for applications such
as machine condition monitoring, robotics, manufacturing processes and design [8-13]. There is an increasing trend
in the manufacturing domain to apply ANN and other soft computing techniques in conjunction with it for both
approximation and optimization. Muthukrishnan and Davim [8] applied ANOVA and ANN technique to predict
surface roughness of Al–SiC composite while turning with PCD insert. The researchers also showed that ANN
methodology was faster giving higher accuracy compared to ANOVA.Pananikumar [9] uses Taguchi’s L 16 , 4-level
orthogonal array with grey relational grade analysis to optimize drilling parameters for glass fibre-reinforced
polymer (GFRP) composite materials. Sharma [10] estimated cutting forces and surface roughness for hard turning
using neural networks. Mandal et al [11] applied RSM and ANN for predicting flank wear of ZTA tool insert and
compared the performances of the two techniques.

3. Proposed Methodology
In the previous work of author [2], experiments as per the Design of Experiments (DOE) principle were conducted
on turning centre with varying cutting parameters viz. cutting velocity (V), feed rate (f) and depth of cut (d) using a

286
 Proceedings of NCAMMM - 2018

Cr 2 O 3 doped Zirconia Toughened Alumina (ZTA) tool insert and AISI 4340 steel work piece. After a specific time
of cutting operation, three machinability parameters viz. flank wear (V b ), cutting force (F z ), and surface roughness
(Ra) were measured. After collecting data from the required number of trial runs, a statistical regression technique
named Response Surface Methodology (RSM) was applied and a regression model was developed to predict the
probable machinability outcomes from the cutting parameters.
In our current study, the predictive model developed essentially has two modules – a data generation from the
regression equations and a back propagation neural network to predict the outcome from given input data.
3.1Experimentation and data generation
Three ‘uncoded’ regression equations are obtained from previous literature of the author [2]. The range of operations
is determined from the same literature as depicted in Table 1. The regression equations are presented in Eq. (1), (2),
and (3). Interested readers can refer to [2] for the detailed experimental setup and regression analysis.

Table 1. Input and output parameters and operation range

Parameters Type Units Notations Max Value Min Value

Cutting Speed Input m/min V 420 140
Feed Rate Input mm/rev f 0.24 0.12
Depth of Cut Input mm d 1.5 0.5
Flank Wear Output mm W NA NA
Cutting Force Output Newton F NA NA
Surface Roughness Output μm R NA NA

W = 0.0253 + 0.000336 *V + 0.2500* f + 0.05200* d(1)

F = 367.5 - 0.8595 *V + 945.0 *f+ 64.6 * d+ 0.001099 * V*V - 183.3 * f *d(2)

R = 0.531 + 0.02212* V + 0.8840* d - 0.000042 * V*V(3)

Random real values are generated within the range for all input parameters and output values are produced using Eq.
(1), (2), and (3), thereby producing a sample data. 1000 such data sample are prepared for training the model and
100 such samples are generated for testing the same.
3.2 Back propagation neural network (BPNN)
The neural network can be categorized into unsupervised and supervised types [12]. The supervised type is selected
to build the current model, which uses aBack Propagation Neural Network written in C# (MS VS 2012). The
network updates itself by learning from the error or deviation from the target output. The network has input layer
containing three neurons for capturing inputs for cutting speed, feed rate, and depth of cut. It has output layer
containing three neurons for generating outcomes in the forms of flank wear, cutting force, and surface roughness.
Apart from the input and output layers, it has one or more intermediate layers known as hidden layers each
containing specific number of neurons. In our work, only single hidden layer is considered. The number of neurons
in the hidden layer is set by trial and error based on network performances. Each neuron is connected to all the

287
An artificial neural network approach for predicting flank wear, cutting force, and surface roughness for turning operation using
ceramic tool insert

neurons in previous layer and the next layer as well. Fig. 1 illustrates the typical network topology with different
layers and neurons. Each connection has a strength known as connection weight. The regular back propagation
algorithm which is also known as gradient descent is applied, where the deviation E of outputs with respect to the
target values are obtained in each iteration for each sample and the same is fed back to the network for updating the
connection weights. The presence of actual target outputs during training makes it a type of supervised learning.

Input Layer Hidden Layer Output Layer

Weights Weights

Cutting Speed Flank Wear

Feed Rate Cutting Force

Depth of Cut Surface Roughness

Fig. 1 Artificial Neural Network topology with different layers along with neurons with

The outputs of each neuron j, 𝑂𝑂𝑗𝑗 𝑖𝑖n the network (hidden layers and output layer) in each iteration are obtained by
using the Eq. (4) which sums up all the contributions of the previous later neurons to produce 𝑌𝑌𝑗𝑗 and then by
activation function or squashing function, which produces an output between -1 and 1 by a bipolar sigmoid function
in Eq. (5). Here, 𝑆𝑆 is the steepness coefficient of the sigmoid function, M is the number of neurons in previous layer.

𝑀𝑀

𝑌𝑌𝑗𝑗 = �(𝑋𝑋𝑖𝑖𝑖𝑖 ∗ 𝑊𝑊𝑖𝑖𝑖𝑖 ) (4)

𝑖𝑖=1

2
𝑂𝑂𝑗𝑗 = −1 (5)
1 + 𝑒𝑒 (−𝑆𝑆∗𝑌𝑌𝑗𝑗 )

The deviations or errors for all outputs are computed in instance-incremental mode using Eq. (6). The weights are
updated using the Eq. (7). Here, 𝜂𝜂 is learning rate, α is momentum constant, ∆𝑊𝑊𝑖𝑖𝑖𝑖 (𝑛𝑛)and ∆𝑊𝑊𝑖𝑖𝑖𝑖 (𝑛𝑛 − 1)denotes
weight update in nth and (n-1)th iterations respectively; N is the total number of output neurons.

288
 Proceedings of NCAMMM - 2018

𝑁𝑁
1
𝐸𝐸 = �(𝑇𝑇𝑗𝑗 − 𝑂𝑂𝑗𝑗 )2 (6)
2
𝑗𝑗 =1

𝜕𝜕𝜕𝜕(𝑛𝑛)
∆𝑊𝑊𝑖𝑖𝑖𝑖 (𝑛𝑛) = (1 − 𝛼𝛼) ∗ 𝜂𝜂 ∗ + 𝛼𝛼 ∗ ∆𝑊𝑊𝑖𝑖𝑖𝑖 (𝑛𝑛 − 1)(7)
𝜕𝜕𝑊𝑊𝑖𝑖𝑖𝑖

The input values and the output target values are normalized using Eq. (8) and (9) before feeding into network,
where Max and Min denote respective maximum and minimum values in actual inputs and targets in the training
samples.𝑉𝑉𝑥𝑥 and𝑉𝑉𝑇𝑇 are the actual values of the input and target parameters respectively.

2 ∗ (𝑉𝑉𝑥𝑥 − 𝑀𝑀𝑀𝑀𝑀𝑀)
𝑋𝑋 = − 1.0 (8)
(𝑀𝑀𝑀𝑀𝑀𝑀 − 𝑀𝑀𝑀𝑀𝑀𝑀)

1.7 ∗ (𝑉𝑉𝑇𝑇 − 𝑀𝑀𝑀𝑀𝑀𝑀)

𝑇𝑇 = − 1.0 (9)
(𝑀𝑀𝑀𝑀𝑀𝑀 − 𝑀𝑀𝑀𝑀𝑀𝑀)

4. Results and Discussions

The parameters defining the network topology like number of neurons in hidden layers, learning rate, momentum
constant, and value of steepness coefficient in activation function were determined correctly after running several
trials of training. In each atomic configuration, one parameter value is altered by keeping the others fixed. MAPE for
1 |𝑇𝑇𝑖𝑖 − 𝑃𝑃 𝑖𝑖 |
each output parameter is compared. 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 = ∑𝑃𝑃𝑖𝑖=1( ∗ 100), where P is total number of samples; T i and P i
𝑃𝑃 𝑇𝑇𝑖𝑖

are target and prediction, respectively. The overall MAPE in prediction (for 3 output parameters)with the optimally
tuned settings was found to be 0.209 for training and 0.2183for testing data. To test the robustness of the model, 4
such sample test runs with different test data of 100 sample size generated from the regression equations were
conducted and tested with the optimal configuration of the network. The results are shown in Table 2,whichare quite
satisfactory in terms of the performance matrics [13].

Table 2. Results showing MAPE of 4 sample test runs

MAPE in
MAPE in MAPE in MAPE in Overall Prediction
Sample Test Run
Flank Wear Cutting Force Surface Roughness (for 3 output
parameters)
1 0.3892 0.1379 0.1279 0.2183
2 0.3703 0.1622 0.1166 0.2164
3 0.3400 0.1518 0.1108 0.2008
4 0.2080 0.1570 0.2456 0.2035

5. Conclusions
The result obtained in this study clearly suggests that the model can perform reasonable well in predicting the
machinability parameters like flank wear, cutting force, and surface finish, given the cutting condition viz. cutting
speed, feed rate, and depth of cut. It should be mentioned that the process of manually changing settings of network
parameters by trial and error needs to be replaced by coupling the BPNN with other soft computing tools like

289
 An artificial neural network approach for predicting flank wear, cutting force, and surface roughness for turning operation using
ceramic tool insert

Genetic Algorithm (GA), Fuzzy Logic Control (FLC), etc. The use of batch mode BPNN or radial basis function
with fuzzy classifier needs to be also investigated in a view to improve the model.

References

[1] R. Venkata Rao. Advanced modeling and optimization of manufacturing processes. Springer series in advanced
manufacturing; 2011.
[2] Singh, B. K., Mondal, B., and Mandal, N. Machinability evaluation and desirability function optimization of
turning parameters for Cr 2 O 3 doped zirconia toughened alumina (Cr-ZTA) cutting insert in high speed
machining of steel. Ceram. Int.;2015,vol. 42, pp. 3338–3350.
[3] Chinchanikar S., Choudhury S. K.. Investigations on Machinability Aspect of Hardened AISI 4340 Steel
Different Levels of Hardness Using Coated Carbide Tools, Int. Journal of Refractory Metals and Hard
Materials; 2013,vol.38, pp. 124-133.
[4] Salvatore F., Saad S., Hamdi H. Modelling and Simulation of Tool Wear During the Cutting Process, 14th
CIRP Conference on Modelling of Machining Operations (CIRP CMMO) Procedia CIRP 8; 2013,pp. 305-310.
[5] Mandal, N., Doloi, B., Mondal, B, Das, R. Optimization of flank wear using Zirconia Toughened Alumina
(ZTA) cutting tool: Taguchi method and Regression analysis, Measurement; 2011, vol. 44,pp. 2149–2155.
[6] Mandal, N., Doloi, B., Mondal, B.Force Prediction Model of Zirconia ToughenedAlumina (ZTA) Inserts in
Hard Turning of AISI 4340 Steel Using Response Surface Methodology. International Journal of Precision
Engineering and Manufacturing; 2012, vol. 13, pp. 1589-1599.
[7] Mandal, N., Doloi, B., Mondal, B. Predictive modeling of surface roughness in high speed machining of AISI
4340 steelusing yttria stabilized zirconia toughened alumina turning insert. Int. Journal of Refractory Metals and
Hard Materials;2013,vol. 38,pp. 40–46.
[8] Muthukrishnan, M., Davim, J. P. Optimization of machining parameters of Al/SiC-MMC with ANOVA and
ANN analysis, J. Mater. Process. Technol.;2009, vol. 209,pp. 225–232.
[9] Palanikumar, K. Experimental investigation and optimization in drilling of GFRP composites, Measurement;
2011, vol. 44, no. 10, pp. 2138– 2148.
[10] Sharma V. S., Dhiman S., Schgal R., Sharma S. K. Estimation of cutting forces and surface roughness for hard
turning using neural networks. Journal of Intelligent Manufacturing; 2008, vol. 19,pp. 473–483.
[11] Mandal, N., Doloi, B., Mondal, B. Application of back propagation neural network model for predicting flank
wear of yttria based zirconia toughened alumina (ZTA) ceramic inserts. Transactions of the Indian Institute of
Metals;2015, vol. 68, no. 5, pp. 783–789.
[12] Pratihar, D.K. Soft Computing,Narosa Publishing House, India; 2008.
[13] Datta, S., Deepanshu, Pratihar, D. K. Modelling of input-output relationships of metal inert gas welding process
using soft computing-based approaches. International Journal of Computational Intelligence Studies; 2017, vol.
6, no. 1, pp.1-28.

290
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Optimization of Machining Parameters during Hard Turning of AISI D3 Steel using

Fuzzy-TOPSIS Approach

Debabrata Ratha, M. Priyadarshinib, K. Palc*, S.Pandad

a
Research Scholar, Veer Surendra Sai University of Technology, Sambalpur-768018, India
b
M.Tech. Scholar, Veer Surendra Sai University of Technology, Sambalpur-768018, India
c,d
Associate Professor, Veer Surendra Sai University of Technology, Sambalpur-768018, India
*Corresponding Author

Abstract: Present day industries have to meet various challenges on economical productivity along with efficient
quality of the finished machined parts during turning processes. During this work, an optimal parametric setting has
been developed for minimum surface roughness and maximum material removal rate at less cutting force for hard
turning of AISI D3 hardened steel using Fuzzy-TOPSIS method. The combined parametric effects obtained from
cutting speed along with feed rate and depth of cut are systematically investigated. Mixed (black) ceramic matrix
Al 2 O 3 and TiCN tool has been employed to accomplish the experiments. Taguchi’s standards L 25 Orthogonal Array
have been used for experimentation. By selecting suitable cutting parameters, the turning operation performance
characteristic like cutting force, surface roughness and material removal rate can be improved together.
Keywords: AISI D3 steel, cutting force, Taguchi, Fuzzy-TOPSIS

1. Introduction
Today metal cutting is one of the most significant processes in most economically developed industries. In order to
meet the challenge, hard turning is considered to be an effective metal machining or cutting process for steel of
hardness value above 45 HRC, because of minimum setup time, increased productivity along with process
flexibility, less power consumption, improved surface integrity, lower product costs and insignificant environmental
disputes exclusion of cutting fluid. Multi-objective optimization has been performed to establish a trade-offs
between various input parameters to achieve desired value of the responses.
Aouici et al. (2012) optimized by using RSM the effect of input process parameters at various levels on
both surface quality along with cutting force during hard turning of AISI H1 using CBN tools where cutting force
was affected mainly by both depth of cut and work piece hardness. Asilturk and Akkus (2011) discussed on
optimization of machining parameters by minimizing the surface roughness values during dry turning of hardened
AISI4140 (51HRC) using Al 2 O 3 + TiC with carbide coated cutting tools and concluded that, surface roughness is
mainly affected with feed. Gaitonde et al. (2009) studied experimentally the effects of DOC (i.e., depth of cut)
along with machining operation time on cutting force generated, power consumed, surface quality and turning tool
wear while turning AISI D2 steel with CC 650, CC650WG and GC 6050 WH ceramic inserts. The study revealed
that CC650WG insert’s performance was better for tool wear and machined surface roughness whereas CC650
insert was used to reduce the machining force along with specific cutting force and power. Lalwani et al. (2009)
investigated experimentally the overall effect of process parameters upon cutting force generated along with surface
quality during turning of MDN250 steel using coated ceramic cutting tool inserts. Aslan et al. (2007) optimized the
Optimization of machining parameters during hard turning of AISI D3 steel using Fuzzy-TOPSIS approach

machining parameters for flank wear of the tool inserts along with surface quality using Taguchi orthogonal array
while turning AISI 4140 hardened steel which offers 63 HRC with the use of (Al 2 O 3 + TiCN) mixed ceramic
cutting tool inserts. Hessainia et al. (2013) studied by using Taguchi Design Of Experiment method and RSM, the
effect of machining factors and cutting phenomenon on surface quality during turning of hardened 42CrMo4 steel
and concluded that feed rate mainly affects the surface roughness. Davim and Figueira (2007) systematically
investigated on the machinability aspects of D2 steel by ceramic turning tools by using statistical techniques
approach. Li B (2012) gone through an extensive review on tool wear estimation by using both theoretical analysis
and numerical simulation technologies. He described that FEM is a powerful tool for prediction of the cutting
process variables and the hidden Markov models were also introduced to estimate tool wear in turning process.
Saini S et al. (2012) gone through brief survey for analysis of the effects of cutting parameters on residual stresses,
surface roughness and tool wear during hard turning and described that, compressive residual stresses are generated
for the fresh tool inserts whereas tensile residual stresses are generated with the propagation of tool wear. The
surface roughness is reduced due to increase in cutting speed but increases with the corresponding increase in depth
of cut and feed. The tool wear is mainly influenced by the cutting speed as compared to feed and depth of cut.
Dogra M et al. (2010) surveyed on the recent research progress by using CBN tool for tool wear rate, surface
roughness and machined chip formation during hard turning operation. Different modelling techniques like FEM,
soft computing and various mathematical approaches are also discussed and concluded that for better performance,
saw – tooth chip formation is desirable in hard turning. Sahoo A K et al. (2013) investigated experimentally the
flank wear, tool life along with the cost analysis and also developed a mathematical model. In this, the work piece
material was coated carbide inserts (TiN/TiCN/Al 2 O 3 /ZrCN).

2. Experimental details
The work material is taken as AISI D3 steel which is heat treatable and offers hardness within the range 58-64 HRC
and exhibits excellent stability during heat treatment. The tool used for the experiment is of WIDIA make and has
composition of mixed (black) ceramic matrix Al 2 O 3 and TiCN. The tool characteristics are showed in Table 1.
Table 1. Tool characteristics
ISO number Grade Dimension
Do L 10 S Re Di
CNGA120408T02020 CW2015 12.70 12.90 4.76 0.8 5.16

2.1 Experimental design with Taguchi method

The range selection for all the process parameters is done on the basis some trial experiments that are done
maintaining same level for all the factors and varying the levels of each factor for some experiment of which the
surface quality was found to be acceptable. Three process parameters with five levels of machining parameters
designed in the experiments and are shown in table 2. The experiments have been conducted with straight turning
operation and Taguchi experimental design with L 25 (54) orthogonal array with three columns and twenty-five rows.
The experimental results of material removal rate (MRR) and Surface roughness (SR) with 25 runs are depicted in
Table 3. To check the repeatability of the machining process used for the experiments; three additional experimental

292
 Proceedings of NCAMMM - 2018

tests have been carried out for randomly chosen experiments of run no.7, 11 and 19. An absolute % error in terms of
MRR and SR value were found to be within 1-3%which may be acceptable.

Table 2. Levels of machining parameters

Controlled parameters Level 1 Level 2 Level 3 Level 4 Level 5
Speed 88 148 192 250 325
Feed rate 0.04 0.05 0.06 0.07 0.08
Depth of cut 0.1 0.3 0.5 0.7 0.9

Table 3. Process performance parameter values of for each experimental setting

Exp. Speed Feed Depth MRR SR Force
No. of cut (mm3/min) (µm) (N)
1 88 0.04 0.1 70.08 2.12 277.405
2 88 0.05 0.3 262.35 1.52 71.58
3 88 0.06 0.5 517.8 1.06 102.407
4 88 0.07 0.7 831.53 2.16 8.539
5 88 0.08 0.9 1200.24 3.52 22.92
6 148 0.04 0.3 372.08 1.48 115.23
7 148 0.05 0.5 726 1.08 18.2
8 148 0.06 0.7 1199.1 1.54 7.99
9 148 0.07 0.9 1976.52 4.57 201.75
10 148 0.08 0.1 236.24 2.54 116.71
11 192 0.04 0.5 753.6 1.06 23.574
12 192 0.05 0.7 1296.4 0.54 15.84
13 192 0.06 0.9 1964.92 3.08 90.26
14 192 0.07 0.1 268.17 1.51 167.1704
15 192 0.08 0.3 1624.6 2.05 101.56
16 250 0.04 0.7 1569.16 1.04 97.99
17 250 0.05 0.9 2131.65 4.52 22.04
18 250 0.06 0.1 116,28 0.53 188.687
19 250 0.07 0.3 1040.76 3.12 30
20 250 0.08 0.5 1962 1.08 105.501
21 325 0.04 0.9 2207.16 3.10 45.88
22 325 0.05 0.1 324.25 0.52 82.29
23 325 0.06 0.3 1159.74 0.54 59.342
24 325 0.07 0.5 2231.60 1.07 24.555
25 325 0.08 0.7 3510.64 0.56 91.11

2.2 Optimization by using fuzzy-TOPSIS approach

In this study, maximizing the MRR along with minimizing both the surface roughness and cutting force values are
considered to be the objectives.
During the optimization, (9 × 3) performance matrix thus prepared, given as, A= [a mn ]

Where, m =1, 2, 3. . . .9 (experimental run in number)

293
Optimization of machining parameters during hard turning of AISI D3 steel using Fuzzy-TOPSIS approach

n = 1, 2, 3, 4 (performance parameters in number).

Table 4. Linguistic variables for important weight

Importance Fuzzy weights

L 01 (Lowest) 0;0;0.1

L 02 (Lower) 0;0.1;0.3

L 03 (Low) 0.1;0.3;0.5

M (Medium) 0.3;0.5;0.7

H 03 (High) 0.5;0.7;0.9

H 02 (Higher) 0.7;0.9;1

H 01 (Highest) 0.9;1;1

Table 5. Machining responses and their importance

Machining responses Decision makers
DM 1 DM 2 DM 3 DM 4
MRR H 01 H 03 H 01 H 02
SR L 01 L 02 L 03 L 01
FORCE M L 02 L 01 L 03

Normalized performances matrix was formulated by using equation (1).

a mn
X mn = (1)
�∑9𝑚𝑚 =1 a mn 2

Where, a mn represents actual value with mth attribute and nth experiment run, whereas X mn represents corresponding
normalize value (Table 6).
The aggregate fuzzy weights of the performance parameters are shown in Table 7. Three weighted performance
matrix developed are shown in Table 8.
With reference to the above performance matrix, both H+ (set of positive ideal value) along with H- (set of negative
ideal value) are represented as
H+= [max. (h mn ) n ε M] or [min (h mn ) n ε M’], m= 1,2,3,4….9 (2)
+ + +
={h 1 , h 2 ,………, h 9 }
H- = [min. (h mn ) n ε M] or [max (h mn ) n ε M’], m= 1,2,3,4….9 (3)
={h 1 -, h 2 -,………, h 9 -}
Where, M={1; 2; 3; 4} and M’= {1; 2; 3; 4}
M and M’ both are involved with performance parameters of higher value the better one and lower value the better
one respectively.
Distance of each experimental result was calculated by using the following equations:

294
 Proceedings of NCAMMM - 2018

d m + = ∑4𝑛𝑛=1 𝑑𝑑(h mn , h n + } m=1,2,3,4……..9 (4)

-
dm = ∑4𝑛𝑛=1 𝑑𝑑(h mn , -
h n } m=1,2,3,4……..9 (5)
n
Proximity of specific experimental run with ideal sol was calculated by using the closeness coefficient and is
expressed as,
dm −
C+= (6)
dm− + dm+

Table 6. Normalize performance matrix

Exp. Run MRR SR FORCE

1 0.675 0.411 143.545
2 9.460 0.211 9.557
3 36.850 0.103 19.562
4 95.032 0.426 0.136
5 197.992 1.132 0.980
6 19.028 0.200 24.768
7 72.441 0.107 0.618
8 197.616 0.217 0.119
9 536.926 1.908 75.925
10 7.670 0.590 25.408
11 78.054 0.103 1.037
12 230.988 0.027 0.468
13 530.642 0.867 15.197
14 9.884 0.208 52.129
15 362.748 0.384 19.240
16 338.412 0.099 17.911
17 624.516 1.867 0.906
18 1.858 0.026 66.411
19 148.872 0.890 1.679
20 529.066 0.107 20.762
21 669.545 0.878 3.927
22 14.450 0.025 12.631
23 184.856 0.027 6.569
24 684.455 0.105 1.125
25 1693.887 0.029 15.484

Table7. Aggregated fuzzy weights

Machining responses Aggregated fuzzy weight

MRR 0.75 0.9 0.975
SR 0.025 0.1 0.25
FORCE 0.1 0.225 0.4

295
Optimization of machining parameters during hard turning of AISI D3 steel using Fuzzy-TOPSIS approach

Table 8. Weighted Performance Matrix

Expt. Run MRR SR FORCE

1 1.772 0.154 104.070
2 24.832 0.079 6.929
3 96.731 0.039 14.183
4 249.458 0.160 0.099
5 519.730 0.425 0.710
6 49.948 0.075 17.957
7 190.158 0.040 0.448
8 518.743 0.081 0.086
9 1409.430 0.716 55.046
10 20.135 0.221 18.421
11 204.891 0.039 0.752
12 606.344 0.010 0.339
13 1392.935 0.325 11.018
14 25.945 0.078 37.793
15 952.213 0.144 13.949
16 888.333 0.037 12.986
17 1639.355 0.700 0.657
18 4.878 0.010 48.148
19 390.789 0.334 1.217
20 1388.798 0.040 15.053
21 1757.555 0.329 2.847
22 37.932 0.009 9.158
23 485.247 0.010 4.762
24 1796.693 0.039 0.815
25 4446.454 0.011 11.226

3 Result and discussions

In this present study, optimization of the experimental data is carried out for analyzing the various impacts of input
machining parameters with corresponding output responses obtained. The average closeness coefficient is calculated
and is shown in Table 10 and concluded that better performance is represented with the higher value of closeness co-
efficient and represented as the optimal combination level of machining process parameters, i.e.; A5B5C4.

296
 Proceedings of NCAMMM - 2018

Table 9. Closeness coefficient at each parametric combination

Expt. No. A B C C+
1 1 1 1 0.0001
2 1 2 2 0.0266
3 1 3 3 0.0408
4 1 4 4 0.0774
5 1 5 5 0.1366
6 2 1 2 0.0297
7 2 2 3 0.0643
8 2 3 4 0.1366
9 2 4 5 0.3202
10 2 5 1 0.0230
11 3 1 3 0.0675
12 3 2 4 0.1558
13 3 3 5 0.3263
14 3 4 1 0.0200
15 3 5 2 0.2289
16 4 1 4 0.2150
17 4 2 5 0.3827
18 4 3 1 0.0131
19 4 4 2 0.1082
20 4 5 3 0.3246
21 5 1 5 0.4083
22 5 2 1 0.0290
23 5 3 2 0.1283
24 5 4 3 0.4174
25 5 5 4 0.1976

Table10. Responses for closeness co-efficient

Process Average of closeness coefficients

Parameters Level 1 Level 2 Level 3 Level 4 Level 5
Speed 0.056304192 0.114758317 0.159712213 0.208733921 0.39608727

Feed rate 0.144121964 0.131680568 0.129026196 0.188645284 0.342121901

Depth of cut 0.017042715 0.104306279 0.182921686 0.3164993 0.314825932

4 Conclusions
The hard turning on AISI D3 steel has been carried out successfully so as to obtain a finish surface using mixed
ceramic matrix Al 2 O 3 and TiCN tool inserts without using coolant. The machined parameters selected are controlled
optimally to improve the surface quality with suitable cutting force and concluded that:
1. The optimization of objective response can be done efficiently and in simplified way by using Taguchi.

2. From the analysis it can be shown that both depth of cut and cutting speed are the important input parameters
which affect the cutting force along with surface roughness and MRR.

3. By selecting suitable cutting parameters, the turning operation performance characteristic like cutting force along
with surface quality and MRR can be improved together.

297
Optimization of machining parameters during hard turning of AISI D3 steel using Fuzzy-TOPSIS approach

Acknowledgement
The authors thank to V.S.S.U.T., Burla, Orissa, India to avail their facilities and kind support for conducting the
research work.

References
[1] Aouici H, Yallese MA, Chaoui K, Mabrouki T, Rigal JF. Analysis of surface roughness and cutting force
components in hard turning with CBN tool: Prediction model and cutting conditions optimization.
Measurement. 2012;(45):344–353.
[2] Asiltürk I, Akkus H. Determining the effect of cutting parameters on surface roughness in hard turning using
the Taguchi method. Measurement. 2011;(44):1697–1704.
[3] Gaitonde VN, Karnik SR, Figueira L, Davim JP. Machinability investigations in hard turning of AISI D2 cold
work tool steel with conventional and wiper ceramic inserts. Int. Journal of Refractory Metals & Hard
Materials. 2009;(27):754–763.
[4] Lalwani DI, Mehta NK, Jain PK. Experimental investigations of cutting parameters influence on cutting forces
and surface roughness in finish hard turning of MDN250 steel. Journal of materials processing technology. 2 0
0 8;(206):68-73.
[5] Aslan E, Camuscu N, Bingoren B. Design optimization of cutting parameters when turning hardened AISI 4140
steel (63 HRC) with Al 2 O 3 + TiCN mixed ceramic tool. Materials and Design. 2007;(28):1618–1622.
[6] Hessainia Z, Belbah A, Yallese MA, Mabrouki T, Rigal JF. On the prediction of surface roughness in the hard
turning based on cutting parameters and tool vibrations. Measurement. 2013;(46):1671–1681.
[7] Davim JP, Figueira L. Machinability evaluation in hard turning of cold work tool steel (D2) with ceramic tools
using statistical techniques. Materials and Design. 2007;(28):1186–1191.
[8] Li Bin. A review of tool wear estimation using theoretical analysis and numerical simulation technologies. Int.
Journal of Refractory Metals and Hard Materials. 2012;(35):143–151.
[9] Saini Sanjeev, Ahuja Inderpreet Singh, Sharma Vishal S. Residual Stresses, Surface Roughness, and Tool Wear
in Hard Turning: A Comprehensive Review. Materials and Manufacturing Processes. 2012;(27):583–598.
[10] Dogra Manu, Sharma Vishal S, Sachdeva Anish, Suri Narinder Mohan, Dureja Jasminder Singh. Tool
Wear,Chip Formation and Work piece Surface Issues in CBN Hard Turning: A Review. International Journal of
Precision Engineering and Manufacturing. 2010;11(2):341-358.
[11] Sahoo AK, Sahoo B. Performance studies of multilayer hard surface coatings (TiN/TiCN/Al 2 O 3 /TiN) of
indexable carbide inserts in hard machining: Part-II (RSM, Grey relational and techno economical approach).
Measurement. 2013;(46):2868–2884.

298
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Online Experimental Characterization of Micro-EDM Dressing on Ti6Al7Nb Biomedical

Material

M. S. Shaha, Probir Sahab,*

a
Research Scholar, Indian Institute of Technology, Patna-801103, India
b
Associate Professor, Indian Institute of Technology, Patna-801103, India
*Corresponding Author

Abstract: Micro-EDM dressing is one of the recent methods used to fabricate high aspect ratio arrayed micro-
structures, which have a great demand in various micro-engineering fields ranging from electrical to biomedical. In
this paper, evaluation of micro-EDM dressing performances based on online monitoring at three different pulse
energy settings estimated. Novel biomedical material Ti6Al7Nb was used as workpiece material. Online monitoring
responses such as contributing and non-contributing pulses, pulse frequency (fp), pulse type density (dp) during
machining were analysed. The online experimental responses showed that machining gap very much effected under
different pulse discharge energy settings. Also, increase in ideal time was observed along the machining length.
Keywords: Micro-EDM dressing, Contributing and Non-Contributing Pulses, Pulse frequency, Pulse type density

1. Introduction
Demand for high aspect ratio arrayed micro-rod (s) is required in various micro-engineering fields ranging from
electrical to biomedical. Current arrayed Ti6Al4V biomaterials are used as an implant material in biological
interfaces, which comprises a brain machine interface, neural interface, retina implant, cochlear implants, and micro-
needle for syringes [1]. Although, Ti-6Al-4V has already proved to be an excellent biological interface material,
however, the material is slowly being replaced by a new titanium alloy namely, Ti-6Al-7Nb because the later results
to comparable elastic modulus that of bone, less allergic and non toxic [2]. Machining micro-features on such
material is an extremely challenging task as it is hard and difficult-to-machined. An emerging micro-EDM (micro-
electrical discharge machining) dressing process has been shown to hold promise over competing technologies such
as LIGA, chemical etching and wire-EDM for safety and economical manufacturing. Micro-EDM dressing and its
variant processes, is basically a contact-less thermo-electric process, which has the ability to machine any
conductive materials regardless of their chemical and mechanical properties. Metal removal is achieved by
generation of repeated sparks struck between workpiece and electrode separated by a small IEG (inter-electrode gap)
submerged in dielectric fluid. Each spark generates a plasma channel by dielectric fluid breakage and act as a cutting
tool. Availability of new advanced CNC systems and RC pulse generators help in stabilizing the machining process
that can able to obtain large aspect ratio micro-rod. In addition, such advanced machined features give us the
flexibility to use a wide range of process parameter settings which have considerable effect on the performance
measures.
Liao et al. [3] developed pulse discriminating system and further this system used to design a better servo
feed controller in a micro-EDM related process. Jahan et al. [4] observed percentage of short circuit pulse decreases
with increase in both gap voltage and capacitance in micro-EDM drilling to certain value. Aligiri et al. [5] used a
Online experimental characterization of micro-EDM dressing on Ti6Al7Nb biomedical material

real time developed pulse system to understand the material removal volume during the micro-EDM drilling
process. Mastud et al. [6] studied the time-evolution of the process mechanics for micro-EDM drilling and micro-
EDM dressing, as a function of voltage and capacitance. They concluded micro-EDM dressing is more stable than
drilling due to the presence of higher percentage of normal discharge time. Hsue et al. [7] proposed a fundamental
observation of gas bubbles and debris in deep-cavity in EDM process to achieve higher machining efficiency.
Experimental observations showed that higher voltage split the main bubble and result in higher debris expulsion
efficiency. Jain et al. [8] fabricated tapered micro pillars of 300-360 μm height and taper angle of 2.7
˚ - 3.6˚ on
Ti6Al4V by micro-EDM dressing and noticed decrease in taper angle with increase in voltage.
Unlike Ti6Al4V material, not all ranges of process parameters are suitable for machining newly adopted Ti6Al7Nb
material. To the best of author knowledge no work has been reported for machining of micro-EDM dressing of
Ti6Al7Nb material. Through study of waveforms (voltage and current) in the discharge gap of micro-EDM dressing
can help the industries to improve their productivity. Thus, it has been seen that no work has been carried out on
processing Ti6Al7Nb biomedical material in micro-EDM dressing process.

2. Experimental Details
All the experiments were performed on CNC Mikrotools high precision three-axis micro-machining centre. The
machine has a positional resolution of 0.1 μm and an accuracy of ±1 μm. A photograph of experimental setup is
shown in Fig. 1. Three sets of discharge energy level were chosen to perform micro-EDM dressing experiments.
Number of experiments in a particular pulse energy were performed for different machining lengths i.e. 250, 500,
750 and 1000 μm. Ti6Al7Nb rods of 500 μm in diameter were used as the workpiece. Whereas, brass plate of
thickness 500 μm and 300 μm of predrilled hole was used as the cathode. EDM oil is used as dielectric in the
experiments. The machining conditions used in this study are listed in Table I. Although, the process is capable of
producing arrayed micro-rods simultaneously but in this work, fabrication of single rod was considered to be
sufficient for study. Also, all the samples were cleared in an ultrasonic bath before they undergone characterization
studies in order to remove any loose particles left on the rods and tools after fabrication.

Fig. 1 Experimental setup

300
 Proceedings of NCAMMM - 2018

Online monitoring responses such as pulse frequency, pulse type density, percentage of contributing and non-
contributing pulses of the fabricated structures were selected as the response parameters. An NI digitizer card
(model: 5122) with 100 MS sampling rate was used to capture the signals from current and voltage probe. The
current probe model was Fluke i30 and voltage probe model was Yokogawa 701938. Signals from current and
voltage probes were imported via the digitizer card to the LabView software. The pulses were discriminated together
with voltage and current product data to acquire corresponding average energy based on an algorithm written in
LabView [9].

Table 1. Machining conditions

Workpiece material Ti6Al7Nb rod of Ø 500 μm
Tool electrode Pre-drilled hole of 300 μm on brass plate of 500 μm thickness
Feed rate (μm/s) 10
Pulse discharge energy E1 (110V, 0.1nF) 006.05 X 10-7 J
E2 (100V, 1.0nF) 050.00 X 10-7 J
E3 (090V, 10nF) 405.00 X 10-7 J
Machining length 25 μm, 500 μm, 750 μm and 1000 μm

1. Result and discussions

During the experiments discharge energy was varied to understand the changes in the output online responses. This
section presents a detailed experimental characterization study of the micro-EDM dressing process for fabrication of
different segment lengths of the micro-rods. Online study of effects of discharge energy for fabricating different
segments of micro-rods includes:

3.1 Pulses occurrence

Machining stability during micro-EDM dressing process can be determined by calculating percentage of various
pulses according to the pulse discrimination system of RC-type pulse generator suggested by Nirala et al. [9]. They
classified pulses into normal, effective and arcing. Figure 2 shows one snap shot of the voltage and current
signatures taken in between the machining process. Figure 3a shows that percentage of normal pulses increases with
pulse discharge energy. This is due to increase of spark explosive force at higher pulse discharge energy setting that
could accelerate the dielectric circulation and debris removal during the process. Thus, maintaining the healthy IEG
(inter-electrode gap) at this condition. Figure 3a also depicts that for any discharge energy setting, with the
increment of machining length constant percentage of normal pulses were observed. Unlike in micro-EDM drilling,
additional flushing i.e. both in radial and axial direction exists in micro-EDM dressing and that is sufficient enough
to evacuate the debris, thus maintaining the same IEG quality at all the segments during the machining. Figure 3b
shows that for any segment of length, percentage of arcing pulses always high for lower pulse energy setting. This is
due to decrease of spark explosive force at lower pulse discharge energy setting, which results to more accumulation
of localized debris in the machining gap. These localized debris agglomeration lowers the IEG and keeps the
dielectric to remains in the ionization state causing arcing. Figure 3b also depicts that for any discharge energy
setting, along the length of micro-rods, decrease in the percentage of arcing pulses was observed. However, at

301
Online experimental characterization of micro-EDM dressing on Ti6Al7Nb biomedical material

highest pulse energy setting percentage of arcing reaches to lowest for longest micro-rod. Since the tool wear is high
at higher pulse energy setting that provides an easy access for dielectric flushing.

Fig. 2 Typical voltage and current pulse events in micro-EDM dressing

3.2 Pulse frequency (fp)

Pulse frequency is defined as the rate at which number of contributing pulses occurs over a period of time. Equation
1 represents expression for the number of contributing pulses:
1
number of contributing pulses = nn + neff (1)
2

Where,

nn = total number of normal pulses, and

neff = total number of effective pulses

Figure 4a shows the variation of pulse frequency with the machining length at three different pulse energy settings.
At any segment of machining for lowest energy setting, fp was found to be small. This is due to greater accumulation
of debris which increases the ideal time. However, for any particular length of machining, fp for highest energy pulse
setting was found to be still smaller than pulse energy at intermediate setting. This is because large debris produces
in highest pulse energy setting choke the narrow IEG (90V compare to 110V) and again machining gets ideal. But
less compare to E1 due to presence of high explosive force favours the flushing. Thus, in short, variations along the
segment of micro-rod for any discharge pulse energy are due to either occurrence of change in contributing pulses
(Np) or change in cycle time or even both. In other word, lowest energy pulse setting shows more stability as
because minimum fluctuation in fp is found (refer Fig. 4b).

302
 Proceedings of NCAMMM - 2018

(a) (b)
Fig. 3 Variation of percentage of: (a) normal pulses with machining length; (b) arcing pulses with machining length

(a) (b)

Fig. 4 Variation of pulse frequency with machining depth a for all energies settings
b for first energy setting

3.3 Pulse type density (dp)

Pulse type density is defined as the percentage of contributing pulses present at that instant per unit time. It is
estimated using the following Eq. 2:

1
𝑛𝑛 𝑛𝑛 + 2 𝑛𝑛 𝑒𝑒𝑒𝑒𝑒𝑒
% contributing pulse = x 100 (2)
𝑛𝑛 𝑛𝑛 + 𝑛𝑛 𝑒𝑒𝑒𝑒𝑒𝑒

Figure 5a depicts that percentage contribution of effective pulses in material removal compare to normal pulses are
very less (only 3-8 %). The reason for small occurrence of effective pulses is associated with the material properties
of Ti6Al7Nb. Since, melting point of this material (around 1860 K) is very high and it only gets electrically eroded
under normal pulse and negligible material gets removed under effective pulses. Moreover, above equation contains
only half of the effective pulses and thus can be neglected. In another word dp depends upon cycle time (ON time
and OFF time) together with ideal time. Figure 5b shows dp is more for higher pulse discharge energy for any length
of micro-rod fabrication. This is due to pulses (E3) at higher capacitance have longer charging and discharging time
and provide enough time to evacuate the debris effectively. Thus, reduces the ideal machining time and improves dp.

303
Online experimental characterization of micro-EDM dressing on Ti6Al7Nb biomedical material

Figure 5b also depicts for any discharge energy setting, for fabricating longer micro-rods dp found to be decreasing.
This is again dp is not only effected by ON time and OFF time of the capacitor but do also influenced by ideal time.
During machining longer micro-rod at low pulse energy setting (E1) ideal time increases due to the increase in
concentration or reattachment or resolidification of debris to the tool plate surface caused by low explosive spark in
static IEG. Thus, large retraction to servo controller (i.e. increase in ideal time) is needed to stabilize the IEG
condition. In this way, dp gets decrease at higher machining length.

(a) (b)
Fig. 5 (a) Variation of percentage of effective pulses with machining length (b) Variation of pulse type
density with machining length

4. Conclusions
The aim of the current research was to achieve micro-rods on Ti6Al7Nb workpiece material using micro-EDM
dressing process. A thorough study of experimental characterization of this processe online has been carried out.
Following are the main conclusions:
• Percentage of normal pulses increases with pulse discharge energy, however for any discharge energy
setting, even for fabricating longer micro-rods almost constant percentage of normal pulses were observed.
• For any segment of machining length percentage of arcing pulses always high for lower pulse energy
setting and for any discharge energy setting for fabricating longer micro-rods, decrease in the percentage of
arcing pulses were observed.
• fp found to be small and more stable for lowest pulse energy setting.
• Dp was found to be more for higher pulse discharge energy for any length of fabricating micro-rod.
• dp found to be decreasing along machining length for any discharge pulse energy setting.

References
[1] T. Fofonoff et al., “A highly flexible manufacturing technique for microelectrode array fabrication,” Proc.
Second Jt. 24th Annu. Conf. Annu. Fall Meet. Biomed. Eng. Soc. [Engineering Med. Biol., vol. 3, pp. 4–5,
2002.

[2] H. Hermawan, D. Ramdan, and J. R. P. Djuansjah, “Metals for Biomedical Applications,” 2009.

304
 Proceedings of NCAMMM - 2018

[3] Y. S. Liao, T. Y. Chang, and T. J. Chuang, “An on-line monitoring system for a micro electrical discharge
machining (micro-EDM) process,” J. Micromechanics Microengineering, vol. 18, no. 3, p. 35009, 2008.

[4] M. P. Jahan, T. Saleh, M. Rahman, and Y. S. Wong, “Development, Modeling, and Experimental Investigation
of Low Frequency Workpiece Vibration-Assisted Micro-EDM of Tungsten Carbide,” J. Manuf. Sci. Eng., vol.
132, no. 5, p. 54503, 2010.

[5] E. Aligiri, S. H. Yeo, and P. C. Tan, “A new tool wear compensation method based on real-time estimation of
material removal volume in micro-EDM,” J. Mater. Process. Technol., vol. 210, no. 15, pp. 2292–2303, 2010.

[6] S. Mastud, R. K. Singh, J. Samuel, and S. S. Joshi, “Comparative analysis of the process mechanics in micro
electrical discharge machining and reverse micro EDM,” Proc. ASME 2011 Int. Manuf. Sci. Eng. Conf. June
13-17, 2011, Corvallis, Oregon, USA, pp. 1–10, 2011.

[7] A. W. J. Hsue and T. J. Hao, “Comparison on gas bubble and pulse trains of deep-cavity electrical discharge
machining with/without rotary ultrasonic assistance,” Int. J. Adv. Manuf. Technol., no. 415, pp. 1–12, 2016.

[8] V. K. Jain, V. Suthar, and A. V Kulkarni, “Fabrication of tapered micro pillars on titanium using electric
discharge micromachining,” ASME 2014 Int. Manuf. Sci. Eng. Conf. MSEC 2014 Collocated with JSME 2014
Int. Conf. Mater. Process. 42nd North Am. Manuf. Res. Conf., vol. 2, no. 2, pp. 97–113, 2014.

[9] C. K. Nirala and P. Saha, “Evaluation of μEDM-drilling and μEDM-dressing performances based on online
monitoring of discharge gap conditions,” Int. J. Adv. Manuf. Technol., vol. 85, no. 9–12, pp. 1995–2012, 2016.

305
Sub - theme

Statistical Quality
Control and
Optimization
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Parametric Analysis of Surface Roughness of Electroless Ni-Co-P Coating using Response

Surface Method

Subhasis Sarkar1, Jhumpa De2, Rajat Subhra Sen1, Buddhadeb Oraon1, Gautam Majumdar1
1
Department of Mechanical Engineering, Jadavpur University, Kolkata, India
2
Department of Mechanical Engineering, Academy of Technology, Hooghly, India

Abstract: Electroless Nickel-Cobalt-Phosphorous coating has been deposited on Copper substrate following central
composite design of experiment. The surface roughness values have been measured after one hour of deposition.
The surface roughness of the coated sample has been considered as a response. Initially, a regression model has
been developed considering full factorial design of experiment with six additional central points to establish a
relation between response and process parameters. The p-values have also been evaluated from first order
regression equation to find the significant process parameters. The response surface methodology has been
considered to develop a mathematical model to study the response surfaces using central composite design of
experiment. The surface and contour plots predict the variation in surface roughness with variation in different
process parameters.
Keywords: Electroless Ni-Co-P coating, p-value, CCD, RSM

Nomenclature
th
Xi Coded value of i independent variable (dimensionless)
Zi Actual value of ith process parameters (gm lit-1)

R̂ a Estimated roughness (µm)

1. Introduction
The electroless Nickel based plating is a chemical coating process. Unlike electrolytic coating, the electron is
supplied by reducing agent instead of external current supply, to deposit the coating. The coating bath comprises of
Nickel salt, reducing agent, complexant, buffers, activators, surfactants etc. To make ternary or quaternary alloy
coatings, one or more metal salts may be added to the plating bath. These types of ternary or quaternary alloy
coatings have been prepared to impart new properties in the coating to meet the industrial needs. The coating is
deposited uniformly on to the substrate irrespective of the geometry of the surfaces. Surface roughness is the
deviation or undulation of a surface from a mean level for a particular sample length. The surface roughness largely
depends on the amount of porosity present on the coated surface and on the deposition rate [1]. The amount of
porosity and rate of deposition leading to variation in roughness depend on the plating condition. Some mechanical
properties like hardness, corrosion resistance and wear resistance also depend on the roughness value of the coating.
Depending upon the alloy composition surface roughness varies [2]. Addition of surfactant in Ni-P bath reduces the
roughness of the coated surface [3]. With the increase in coating thickness of electroless Nickel coating the surface
roughness decreases [1],[4]. The non-metallic inclusions have no effect on amount of porosity[1]. The surface
roughness of electroless Nickel based plating increases with the increase in deposition time [5]. The nucleation of
Parametric analysis of surface roughness of electroless Ni-Co-P coating using response surface method

electroless Nickel based plating is done by connecting with active metal, immersing in strong reducer or immersing
in Palladium Chloride solution. Lower nucleation density increases the surface roughness of the electroless Ni-P
coating [6]. The roughness of the coated surface does not vary below 65 (nm) coating thickness in Nickel coated
optical fibre using electroless technique. The surface roughness increases with increase in coating thickness after
crossing the threshold thickness value of 65 (nm) [7]. Cobalt is added in the Ni-P matrix to enhance electromagnetic
shielding property, saturation magnetization, remanence and coercivity of the electroless deposits [8]. Literatiure
shows that the concentration of Cobalt Sulphate has significant effect in the weight percentage of Nickel and Cobalt
in electroless Ni-Co-P coating [9].
In the present study electroless Nickel-Cobalt-Phosphorous (Ni-Co-P) coating has been deposited on Copper
substrate by varying the concentration of Cobalt salt, reducing agent and the bath temperature following design of
experiment. The surface roughness values have been measured and analyzed statistically to find the significant
parameters. A mathematical model has been developed to predict the surface roughness of electroless Nickel-
Cobalt-Phosphorous coating on Copper substrate.

2. Experimental Details
The electroless Nickel-Cobalt-Phosphorous coating has been deposited on Copper substrate of (20 × 15 × 0.1) mm3.
At first the substrates have been cleaned in distilled water followed by acid pickling and cleaning in distilled water.
The substrates have been activated by immersing in Palladium Chloride solution at 55oC before immersing them in
the coating bath. The coating has been deposited from a chemical bath following the conditions given in Table 1.
After the deposition for one hour, the coated samples have been taken out and rinsed in distilled water. After that the
samples have been dried out and stored for the further study.

Table 1: Coating deposition condition

Factors Amount
Nickel Sulphate Hexahydrate 25 (gm lit-1)
Cobalt Sulphate Heptahydrate 5. 272-18.728 (gm lit-1)
Sodium Hypophosphite 15.272-26.728 (gm lit-1)
Trisodium Citrate Dihydrate 15 (gm lit-1)
Ammonium Sulphate 10 (gm lit-1)
Temperature 75.3-88.7 (oC)
pH of solution 5
Deposition Time 1 (hour)
Activation temperature 55 (oC)
Bath Volume 250 (cm3)

The surface roughness of the coated samples has been measured by Taylor Hobson Precision Instrument Surtronic
3+. After calibrating the Taly Surf with the standard specimen, the electroless Ni-Co-P coated samples have been
tested under the same. 20 runs have been considered to synthesise the coating following the Central Composite

307
 Proceedings of NCAMMM - 2018

Design of experiments (CCD). Each runs contains 3 specimens of the size stated before. After evaluating the
roughness values for each run, the instrument has been checked with the standard specimen. The EDS spectra and
SEM micrographs of the electroless Ni-Co-P coated samples have been obtained in Oxford X max 50, SOF.

Table 2: Actual and coded values of different process parameters

Actual Values Coded Values
o
CoSO 4 ·7H 2 O NaH 2 PO 2 ·H 2 O Temperature ( C) CoSO 4 ·7H 2 O NaH 2 PO 2 ·H 2 O Temperature
-1 -1
(gm lit ) (gm lit )
Z1 Z2 Z3 X1 X2 X3
5.272 15.27 75.272 −α −α −α
8 18 78 -1 -1 -1
12 22 82 0 0 0
16 26 86 1 1 1
18.728 26.728 88.78 +α +α +α

3. Data Collection and Statistical Analysis

3.1 First order regression analysis
A first order regression equation has been developed to establish the relation between the response and the process
parameters. The surface roughness have been considered as the response in the present study. To develop the first
order regression equation, a full factorial design of experiment along with six central points has been considered
[10].
Table 3: Observed data considering full factorial design with six central points in a randomized run order.

Coded values of the process parameters

Standard Order Run Order X1 X2 X3 Ra
4 1 1 1 -1 0.991
6 2 1 -1 1 0.891
15 4 0 0 0 0.751
18 5 0 0 0 0.779
19 6 0 0 0 0.734
20 8 0 0 0 0.765
16 9 0 0 0 0.739
3 10 -1 1 -1 0.573
2 12 1 -1 -1 0.969
17 14 0 0 0 0.705
8 15 1 1 1 0.576
1 17 -1 -1 -1 0.874
7 19 -1 1 1 0.705
5 20 -1 -1 1 0.721

308
Parametric analysis of surface roughness of electroless Ni-Co-P coating using response surface method

The coded values of process parameters are given in Table 2. Table 3 contains the observed data to form the
regression equation. The p-values have been calculated from equation (1) to identify the process parameter and
interaction of the process parameters which can affect surface roughness significantly.
The fitted regression equation is given in equation (1) With R2 value= 95.39%

R̂ = 0.7695 + 0.0693 X − 0.0762 X − 0.0643 X + 0.0030 X X − 0.0065 X X − 0.0590 X X − 0.0778 X X X

a 1 2 3 1 2 2 3 3 1 1 2 3

Table 4: The p – values of the respective coefficients for the response.

Ra p1 p2 p3 p12 p23 p13 p 123
Surface roughness 0.003 0.002 0.004 0.837 0.659 0.006 0.001
It is evident from Table 4, that all the main effects have significantly affected the response. Also, the interactions
of Cobalt Sulphate-temperature and Cobalt Sulphate-Sodium Hypophosphite-temperature have significant effect on
the response.
4. Results and Discussions
4.1 Response Surface methodology
Response surface methodology (RSM) has been adopted to analyze the problem both statistically and
mathematically.
Table 5: Observed data considering CCD.
Coded values of the process parameters
Run no. Ra
X1 X2 X3
1 1 1 -1 0.991
2 1 -1 1 0.891
3 0 0 1.682 0.658
4 0 0 0 0.751
5 0 0 0 0.779
6 0 0 0 0.734
7 1.682 0 0 0.882
8 0 0 0 0.765
9 0 0 0 0.739
10 -1 1 -1 0.573
11 0 -1.682 0 0.992
12 1 -1 -1 0.969
13 -1.682 0 0 0.494
14 0 0 0 0.705
15 1 1 1 0.576
16 0 0 -1.682 0.953
17 -1 -1 -1 0.874
18 0 1.682 0 0.777
19 -1 1 1 0.705
20 -1 -1 1 0.721

309
 Proceedings of NCAMMM - 2018

Using a rotatable Central Composite Design (CCD) of experiment the response surfaces have been developed. The
CCD contains 8 factorial points, 6 central points and 6 axial points. The observed data following CCD is given in
Table 5.
The predicted response surface equation considering surface roughness as the response is given in equation (2).
2 2 2
R̂a = 0.7458 + 0.0883 X 1 − 0.0711 X 2 − 0.0740 X 3 − 0.0220 X 1 + 0.0474 X 2 + 0.0195 X 3 + 0.0030 X 1 X 2
− 0.0065 X 2 X 3 − 0.0590 X 3 X 1
(2)
2
With R value= 95.39% and p-value=0.005
Since the p-value of second order response surface equation has been found well below 0.05, therefore, the
predicted response surface provides excellent fitting to the observed data.

1.4
1.3

estimated roughness
1.2 1.2
estimated roughness

1.1
1 1
0.9
0.8
0.8
0.7
0.6

2
0.4
2 1 2
0 1
1 2 0
1 -1 -1
0
0 X3 -2 -2
-1 X2
-1
X2 -2 -2
X1

Fig.2: Response Surface and contour plot for

Fig.1: Response Surface and contour plot for
Rˆ a = f ( X 2 , X 3 ) , Hold value: X 1 = 0 .
Rˆ a = f ( X 1 , X 2 ) , Hold value: X 3 = 0

1.4

1.2
estimated roughness

0.8

0.6

0.4
2
1 2
0 1
0
-1 -1
X1 -2 -2
X3

Fig. 3: Response Surface and contour plot for Rˆ a = f ( X 3 , X 1 ) , Hold value: X3 = 0.

4.2. Interpretation of Response Surface Plot

Fig. 1-3 shows the response surface and contour plots for Rˆ a = f ( X 1 , X 2 ) , Rˆ a = f ( X 2 , X 3 ) and

Rˆ a = f ( X 3 , X 1 ) respectively. From Fig. 1-3, it can be observed that the concentrations of Cobalt Sulphate,
reducing agent and temperature are the significant process parameters.

310
Parametric analysis of surface roughness of electroless Ni-Co-P coating using response surface method

Fig. 4: EDAX spectra of electroless Ni-Co-P coating, deposited at central point condition.
After depositing the electroless Ni-Co-P coating at central point condition, the SEM micrograph and EDS spectra
have been studied. Fig. 4 shows the EDS spectra of the electroless Ni-Co-P coating deposited at the central point
condition. The coating contains 66.31 weight % Nickel, 15.98 weight % Cobalt 17.71 weight % Phosphorous.

Fig. 5: The SEM micrograph of the Ni-Co-P coated samples at central point condition.

From Fig. 5, it is evident that granular inter connected grains structure has been observed in the as-deposited
coating at central point condition. Fig. 5 also indicates that proper nucleation has taken place in central point
condition.
5. Conclusions
From Table 4, it can be concluded that the concentrations of Cobalt Sulphate, Sodium Hypophosphite,
Temperature and the interactions of Cobalt salt, temperature and the interaction of the three factors have significant
effect on variation of surface roughness. The response surface model having p-value less than 0.05 converges
excellently with the observed data. From Fig 1-3, it is evident that the concentration of Cobalt salt, Sodium
hypophosphite and Temperature have significant effect on predicting surface roughness.

References
[1] Tomlinson WJ, Mayor JP. Formation, microstructure, surface roughness, and porosity of electroless nickel
coatings. Surface engineering. 1988 Jan 1;4(3):235-8.
[2] Liu Z, Gao W. The effect of substrate on the electroless nickel plating of Mg and Mg alloys. Surface and
coatings technology. 2006 Mar 15;200(11):3553-60.

311
 Proceedings of NCAMMM - 2018

[3] Sudagar J, Lian JS, Jiang Q, Jiang ZH, Li GY, Elansezhian R. The performance of surfactant on the surface
characteristics of electroless nickel coating on magnesium alloy. Progress in Organic coatings. 2012 Aug
31;74(4):788-93.
[4] Ernst P, Wadsworth IP, Marshall GW. Porosity of electroless nickel coatings investigated using different
porosity tests and their application. Transactions of the IMF. 1997 Jan 1;75(5):194-8.
[5] Watanabe H, Honma H. Fabrication of nickel microbump on aluminum using electroless nickel plating. Journal
of the Electrochemical Society. 1997 Feb 1;144(2):471-6.
[6] Homma T, Tanabe M, Itakura K, Osaka T. Tapping Mode Atomic Force Microscopy Analysis of the Growth
Process of Electroless Nickel‐Phosphorus Films on Nonconducting Surfaces. Journal of the Electrochemical
Society. 1997 Dec 1;144(12):4123-7.
[7] Shiue ST, Yang CH, Chu RS, Yang TJ. Effect of the coating thickness and roughness on the mechanical
strength and thermally induced stress voids in nickel-coated optical fibers prepared by electroless plating
method. Thin Solid Films. 2005 Aug 1;485(1):169-75.
[8] Aal AA, Shaaban A, Hamid ZA. Nanocrystalline soft ferromagnetic Ni–Co–P thin film on Al alloy by low
temperature electroless deposition. Applied Surface Science. 2008 Jan 30;254(7):1966-71.
[9] De J, Banerjee T, Sen RS, Oraon B, Majumdar G. Multi-objective optimization of electroless ternary Nickel–
Cobalt–Phosphorous coating using non-dominant sorting genetic algorithm-II. Engineering Science and
Technology, an International Journal. 2016 Sep 30;19(3):1526-33.
[10] Oraon B, Majumdar G, Ghosh B. Parametric optimization and prediction of electroless Ni–B deposition.
Materials & design. 2007 Dec 31;28 (7):2138-47.

312
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Warpage and Shrinkage Minimization in an Injection Molded Component using Improved

Particle Swarm Optimization Algorithm

Vimal Kumar Pathak1, Mithilesh Dixit2, Ashish Shrivastava

1, 2, 3
Department of Mechanical Engineering, Manipal University Jaipur, Rajasthan

Abstract: This paper presents an application of newly developed modified particle swarm optimization in
determining optimal injection molding conditions for minimum warpage and shrinkage. Polybutylene terephthalate
(PBT) and polyethylene terephthalate (PET) were injected in thin wall relay component under different processing
parameters: packing pressure, melt temperature, and packing time. Further, Taguchi’s L 9 orthogonal array is
employed for conducting simulation analysis to consider the interaction effects of the above parameters. A
predictive mathematical model for shrinkage and warpage in terms of the above process parameters is developed
using non-linear regression model. The developed model is further optimized and results are compared with
classical PSO algorithm. The confirmation simulation trials confirm the effectiveness and robustness of MPSO
algorithm in determining minimized shrinkage and warpage values.
Keywords: MPSO algorithm, optimization, injection molding, simulation

1. Introduction
The plastic injection molding (PIM) process is extensively used for producing intricate shaped plastic parts with
distinctive geometric features and it also has short production cycles. Basically, PIM consists of processes like
filling, packing, cooling and ejection [1]. Injection molding helps in producing products of computer,
communication, and consumer electronic (3C) such as movable computers and mobile phones which are generally
thin, light, short and small [2-3]. However, with the increasing demand of more complex products having less wall
thickness, the PIM process is prone to face more challenging tasks [4]. Consequently, the quality of the parts
produced using PIM process is highly affected by the appropriate selection of the various process parameters and the
mold design [5-6]. In contrast, the inappropriate process parameter values can lead to produce part defects, result in
long lead times and high cost [7].
Past studies suggest that shrinkage and warpage are very much associated with machining parameters.
Moreover, the effect of gate geometry and packing parameters on the final part was studied [7] and found that a
thinner gate gains a more equable shrinkage in the process having same packing pressure. Galantucci et al. [8]
applied double skin model for investigating the warpage defect and concluded that melt temperature was among the
most influencing factor for minimizing warpage. Huang and Tai [9] in their study applied computer simulation and
experiment for analyzing the factor affecting warpage in a thin shell injection molded components. Tang et al. [10]
applied Taguchi method for minimizing the warpage in the design of improved plastic injection mold. Similarly,
Taguchi and ANOVA are used in a study for obtaining optimal shrinkage injection molding conditions [11].
Similarly, Park and Dang [12] in their study suggests that runner and cooling channel geometry can
improve the final quality of products. Choi [13] found that residual stress is an important factor that affect the
shrinkage and warpage defect in injection molded components. Liao et al. [14] in their study found that the packing
Warpage and shrinkage minimization in an injection molded component using improved particle swarm optimization algorithm

pressure was the most influencing factors. Kikuchi and Koyoma [15] has applied finite element method for studying
the relation among warpage, part thickness and anisotropy. Moreover, several studies found have used response
surface methodology (RSM) individually or integrating with genetic algorithm (GA) for determining the interaction
and relationship among factors and process parameters [16-17]. Similar studies have been found that uses neural
model [18], modified complex method [19], grey-fuzzy logic for thin shell feature [20] for optimization of warpage
in different thermoplastic parts. The current study takes into consideration a newly developed modified particle
swarm optimization (MPSO) algorithm for optimal injection molding process parameter determination.
In this study, a systematic methodology is presented using regression analysis and determining the
optimum process parameters using a newly developed MPSO algorithm [21]. Particle swarm optimization (PSO)
algorithm was found to be widely used as a powerful tool for solving optimization problems due to the easiness in
its concept having fewer settings to be adjusted in comparison to other nature inspired optimization algorithms [22].
However, basic PSO algorithm has few shortcomings, like weak local search that led to entrapment in local minima
which highly changes the convergence performance. For improving the exploitation behavior of basic PSO, a new
and modified variant of particle swarm optimization i.e. MPSO is suggested. An efficient procedure based on greedy
selection has been used for determining best particle position amid the newly generated and current candidate
solution centered on the fitness value.

2. Simulation Details
An electronic relay part, with an overall dimension of 72 mm × 34 mm × 48.5 mm was designed in CREO-3.0
software and was used as a 3D mesh type model. Moreover, the model consists of 29, 812 elements. The analysis
were performed using three thickness values of 0.8, 0.9 and 1.0 mm. Figure 1 shows the mesh file of relay
component with the cooling channels. The parameters considered for shrinkage and warpage analyses are melt
temperature (A), packing pressure (B) and packing time (C). The values of these parameters are provided in Table 1.
For performing all simulation, the injection time was taken as 3s. A L 9 (33) Taguchi array was selected as the

experimental design for each of the three factors.

Figure 1. Electronic relay mesh file with cooling channel
Table 1.Injection molding parameters and their levels
Factors Description Coded Unit PBT Levels PET Levels
Symbol 1 2 3 1 2 3
A Melt Temperature, t m X1 °C 254 266 278 252 264 276
B Packing Pressure, p p X2 MPa 25 30 35 22 28 34
C Packing Time, p t X3 S 10 15 20 12 18 24

314
 Proceedings of NCAMMM - 2018

3. Results and Discussion

The Taguchi method was used to predict the effect of injection molding process parameters on the shrinkage and
warpage. The measured values of shrinkage and warpage and the signal-to-noise ratios are measured and reported in
Table 2. For the current study, smaller the better characteristic was selected when calculating the S/N ratio, which is
based on equation 1, as shown in Table 2. It was found that the best process parameter value can be determined by
selecting the level with the best values at each factor.Based on shrinkage results in Table 2, itwas found that PET
material had higher S/N ratio (-6.02) value than the PBT material, because it exhibits least shrinkage as compared to
PBT. Similarly, the warpage result shows that PET also has least warpage in comparison to PBT warpage values.

Table 2. Shrinkage and warpage values of PBT and PET material

Exp. PBT PET PBT PET
Shrinkage % S/N ratio Shrinkage % S/N ratio Warpage (mm) S/N ratio Warpage (mm) S/N ratio
1 2.1982 -7.1293 2.0409 -6.2237 0.2432 12.2841 0.1682 15.3867
2 2.1392 -2.3451 1.972 -6.0301 0.2164 13.2956 0.1608 15.8743
3 2.0932 -6.8343 1.9145 -5.5843 0.2813 11.0166 0.1619 15.815
4 2.1523 -7.0749 1.8832 -5.4029 0.1341 17.4521 0.2521 11.9685
5 2.0123 -6.1452 1.9923 -6.0940 0.1931 14.2844 0.2912 10.7161
6 1.9450 -5.7231 1.9639 -5.927 0.1857 14.6237 0.1972 14.1018
7 1.9197 -5.5938 2.021 -6.1512 0.2076 13.6554 0.2102 13.5473
8 1.9921 -6.0927 1.9715 -6.0211 0.1596 15.9397 0.1821 14.6143
9 1.9772 -6.0312 1.9218 -5.7439 0.1801 14.8897 0.1717 15.305

For further analyzing the obtained results and determining the significance of each parameter, regression analysis
and ANOVA test was performed. Based on the generalized regression model, the following analytical model was
obtained for shrinkage of PBT and PET in the form of coded unit:
Shrinkage (S1) = 0.326440 + 0.201507𝑋𝑋1 – 0.034872𝑋𝑋2 +0.065912𝑋𝑋3 – 0.003173𝑋𝑋12 – 0.029771𝑋𝑋22 – 0.004712𝑋𝑋32 +
0.033242𝑋𝑋1 𝑋𝑋2 + 0.023560𝑋𝑋1 𝑋𝑋3 + 0.048322𝑋𝑋2 𝑋𝑋3
Shrinkage (S2) = 0.462270 + 0.305915𝑋𝑋1 – 0.048202𝑋𝑋2 + 0.079210𝑋𝑋3 – 0.002320𝑋𝑋12 – 0.041725𝑋𝑋22 – 0.003712𝑋𝑋32 +
0.045902𝑋𝑋1 𝑋𝑋2 + 0.043450𝑋𝑋1 𝑋𝑋3 + 0.054721𝑋𝑋2 𝑋𝑋3
In addition, the model significance is further validated using ANOVA analysis. The ANOVA results for PBT
material are shown in Table 3. The model was built for 95% confidence level. The correlation coefficient (𝑟𝑟 2 ) of the
developed analytical models for PBT and PET was found to be 0.973 and 0.962 (nearer to the ideal value of 1). The
adequate value of regression coefficient indicate that the model is significant and further analysis and predictions
can be performed. All the linear, square and interaction terms have significant effect on the response output. It was
clearly seen from Table 3 that the F-value is significantly higher for both the ANOVA table, indicating that the
model is significant. There is only 0.05% chance that such a high model F-value may have occurred due to noise.
Furthermore, the performance of the developed model of warpage for both material was tested using ten randomly
selected experiments other than the used in Table 4. Similarly, the following analytical model for warpage was
developed using non-linear regression model for both the material i.e. PBT and PET.
Warpage (W 1 ) = 1.23966 + 0.772017𝑋𝑋1 + 0.005873𝑋𝑋2 - 0.347592𝑋𝑋3 - 0.027631𝑋𝑋12 – 0.000987𝑋𝑋22 – 0.012283𝑋𝑋32 +
0.000533𝑋𝑋1 𝑋𝑋2 + 0.010937𝑋𝑋1 𝑋𝑋3 + 0.000272𝑋𝑋2 𝑋𝑋3
Warpage (W 2 ) = 1.78218 + 0.817812𝑋𝑋1 + 0.007291𝑋𝑋2 + 0.192205𝑋𝑋3 - 0.051140𝑋𝑋12 – 0.001723𝑋𝑋22 – 0.027831𝑋𝑋32 +
0.000612𝑋𝑋1 𝑋𝑋2 - 0.033147𝑋𝑋1 𝑋𝑋3 + 0.000721𝑋𝑋2 𝑋𝑋3

315
Warpage and shrinkage minimization in an injection molded component using improved particle swarm optimization algorithm

The developed regression model significance is further tested using ANOVA analysis. The ANOVA results for PBT
and PET material are shown in Table 4. The model was built for 95% confidence level. The correlation coefficient
(𝑟𝑟 2 ) of the developed analytical models for PBT and PET was found to be 0.941 and 0.924 (nearer to the ideal value
of 1). The adequate value of regression coefficient indicates that the model is significant and further analysis and
predictions can be performed. It was clearly seen from Table 4 that the F-value is significantly higher for both the
ANOVA table, indicating that the model is significant.

Table 3. ANOVA result for shrinkage model

Source DF Sum of squares (SS) Mean square (MS) F-Value P-value
Regression 9 1.52827 0.16980 253.43 0.000 Significant
Linear 3 1.37281 0.45760 682.98 0.000
Square 3 0.01430 0.00477 7.12 0.001
Interaction 3 0.13092 0.04031 60.16 0.000
Residual error 41 0.02744 0.00067
Lack of fit 5 0.02205 0.00441 6.58 0.000
Pure error 36 0.00438 0.00012
Total 50 1.5572

Table 4. ANOVA Analysis for warpage model

Source DF Sum of squares (SS) Mean square (MS) F-Value P-value
Regression 9 202.317 22.48 40.43 0.000 Significant
Linear 3 112.34 37.45 67.36 0.000
Square 3 43.22 14.41 25.92 0.000
Interaction 3 29.23 9.74 17.52 0.000
Residual error 41 22.78 0.556
Lack of fit 5 16.26 3.252 5.56 0.000
Pure error 36 6.52 0.181
Total 50 225.097

4. Modified particle swarm optimization algorithm

The classicalparticle swarm optimization is a nature-inspired algorithm proposed byKennedy and Eberhart in
mid-90s. The PSO is based on mimicking orsimulating the social behavior of birds in a flock [25-26]. Each particle
in the classical PSO is initialized randomly in n-dimensional search area. Each particle performance in search space
is measured using a fitness function that depends solely on the optimization problem. The personal best position of
each particle can signify as the best position that particle𝑖𝑖 has been at so far. New position and velocity for 𝑖𝑖 𝑡𝑡ℎ
particle is updated at each iteration and can be expressed as:
𝑣𝑣𝑖𝑖 (𝑡𝑡 + 1) = 𝑣𝑣𝑖𝑖 (𝑡𝑡) + 𝑐𝑐1 ∗ 𝑟𝑟1 �𝑝𝑝𝑖𝑖 (𝑡𝑡) − 𝑥𝑥𝑖𝑖 (𝑡𝑡)� + 𝑐𝑐2 ∗ 𝑟𝑟2 (𝑔𝑔𝑖𝑖 (𝑡𝑡) − 𝑥𝑥𝑖𝑖 (𝑡𝑡)) (1)
𝑥𝑥𝑖𝑖 (𝑡𝑡 + 1) = 𝑥𝑥𝑖𝑖 (𝑡𝑡) + 𝑣𝑣𝑖𝑖 (𝑡𝑡 + 1) (2)
where 𝑟𝑟1 and 𝑟𝑟2 are the two uniformly distributed independent random numbers in the interval [0,1]. The two
accelerating coefficients𝑐𝑐1 and 𝑐𝑐2 values are normally 2 each for nearly all applications, 𝑝𝑝(𝑡𝑡)and 𝑔𝑔(𝑡𝑡) are the best
individual position of a particle and global best position parameter of entire swarms respectively.
In this study, a new variant of basic PSO is proposed for optimizing the dimensional characteristics in an injection
molded components. The two imperative factors generally considered during design of any optimization algorithm
are exploration and exploitation. In classical PSO, new solution is usually substituted by the old one without

316
 Proceedings of NCAMMM - 2018

comparing which solution gives better results. This behavior displays the lack in exploitation ability of classical PSO
and shows only exploration tendency, making it difficult to find the best possible solutions. Due to the lack in
exploitation behavior, classical PSO has few disadvantages. The first problem is the weak local search ability in
finding the best solution and the second may led to entrapment of particle in local minimum solutions. The proposed
modified version of PSO algorithm produces different swarm position as well as new fitness solution which is
primarily based on the new proposed search equations (3) and (4). 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 is the particle best position, 𝑔𝑔𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 is the
particle global best position. 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 is the random number between 0 and 1 that controls the rate at which the
population evolves. The best solution of the current population are one of the useful sources that can be used for
improving the convergence performance.
Also, Eq. (3) drives the new candidate solution which is also about the best solution of the previous
iteration. Therefore, the proposed updation and search equation as defined by Eq. (3) and (4) will help in increasing
the exploitation capability of the basic PSO. Since the selection strategy in the optimization algorithm is usually

(3)
𝑣𝑣𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 + 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟(𝑔𝑔𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 − 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 )

𝑥𝑥𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 + 𝑣𝑣𝑛𝑛𝑛𝑛𝑛𝑛 (4)

considered as exploitation, as the fitness solution of the individual is used to determine whether or not an individual
must be exploited. Therefore, in the proposed MPSO algorithm, particle swarm employs greedy selection among
two parallel fitness functions to compare and consequently update the best candidate solution, helping in improving
the exploitation behavior of the algorithm. The steps of the proposed modified PSO algorithm is depicted in Figure
2.

5. Optimization problem formulation

In the present study, a mathematical model of the shrinkage and warpage is minimized to attain optimum values of
packing time, melt temperature and packing pressure. The optimal parameters values are required to have improved
injection molding process. To improve the final accuracy of injection molded components, the shrinkage and
warpage in S 1 , S 2 and W 1 and W 2 needs to be minimized. Now the present problem is formulated as optimization
problem:
Minimize f (t m , p p , p t ) = S 1
f (t m , p p , p t ) = S 1 , f (t m , p p , p t ) = W 1 , f (t m , p p , p t ) = W 1
For solving the present optimization problem, a new computer programis developed in MATLAB software for
minimizing the objective functions in which modified PSO is implemented as the solver. The MPSO program
employed different settings of PSO parameters for predicting the values of the injection molded parameters and
obtain minimized values of shrinkage and warpage for final components. For proving the effectiveness of the
proposed MPSO algorithm, its results are compared with those obtained from the standard PSO algorithm. The
parameters for both the algorithms are set as: c1, c2 = 2.05, number of population size = 10.

317
Warpage and shrinkage minimization in an injection molded component using improved particle swarm optimization algorithm

Start

Initialize population randomly

Set parameters 𝑚𝑚, 𝑛𝑛, 𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 , 𝑤𝑤𝑚𝑚𝑚𝑚𝑚𝑚 , 𝑐𝑐1 , 𝑐𝑐2 , 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚, 𝑣𝑣

Update position and velocity for each particle

𝑣𝑣(𝑖𝑖, 𝑗𝑗) = 𝑤𝑤 ∗ 𝑣𝑣(𝑖𝑖, 𝑗𝑗) + 𝑐𝑐1 ∗ 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟(1) ∗ �𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝(𝑖𝑖, 𝑗𝑗) − 𝑥𝑥(𝑖𝑖, 𝑗𝑗)� + 𝑐𝑐2 ∗ 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟(1) ∗ (𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔(𝑖𝑖, 𝑗𝑗) − 𝑥𝑥(𝑖𝑖, 𝑗𝑗);
𝑥𝑥(𝑖𝑖, 𝑗𝑗) = 𝑥𝑥(𝑖𝑖, 𝑗𝑗) + 𝑣𝑣(𝑖𝑖, 𝑗𝑗)

Applying boundary constraint

B
𝑖𝑖𝑖𝑖 𝑥𝑥(𝑖𝑖, 𝑗𝑗) < 𝐿𝐿𝐿𝐿(𝑗𝑗); 𝑥𝑥(𝑖𝑖, 𝑗𝑗) = 𝐿𝐿𝐿𝐿(𝑗𝑗);
𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑖𝑖𝑖𝑖 𝑥𝑥(𝑖𝑖, 𝑗𝑗) > 𝑈𝑈𝑈𝑈(𝑗𝑗); 𝑥𝑥(𝑖𝑖, 𝑗𝑗) = 𝑈𝑈𝑈𝑈(𝑗𝑗)

Evaluate fitness after 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 and 𝑔𝑔𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 update

Position update
Update 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 and fitness 𝑓𝑓0
𝑣𝑣𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 + 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟(𝑔𝑔𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 − 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 )
𝑥𝑥𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑝𝑝𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 + 𝑣𝑣𝑛𝑛𝑛𝑛𝑛𝑛

Update 𝑔𝑔𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 and fitness 𝑓𝑓0

Evaluate new objective function value
based on above update 𝑓𝑓1

Greedy selection b/w

fitness 𝑓𝑓0 𝑓𝑓1

Yes No

Update fitness as 𝑓𝑓2 and Update fitness as 𝑓𝑓2 and

corresponding position 𝑥𝑥2 corresponding position 𝑥𝑥2
𝑓𝑓0 = 𝑓𝑓2 𝑎𝑎𝑎𝑎𝑎𝑎 𝑥𝑥0 = 𝑥𝑥2

Is the termination
B
criterion met?

Stop

Figure 2. MPSO algorithm flowchart

6. Results and Discussion

The results predicted for minimized shrinkage and warpage using PSO and proposed MPSO for optimized values of
melt temperature, packing pressure and packing time are shown in Table 9. It can be seen from Table 9 that the

318
 Proceedings of NCAMMM - 2018

predicted values of injection molded relay component by MPSO algorithm shows significant improvement over
PSO results as well as the simulation results by 36.47 % for PBT shrinkage and 19.31 % for PBT warpage
respectively. Similar results are found for PET material having shrinkage reduction of 30.73% and warpage
reduction of 28.79%. The convergence graph of MPSO algorithm in comparison to standard PSO for PBT material
is shown in Fig. 10. It is observed from Fig. 10 that MPSO algorithm requires only 20 iterations to converge to the
optimum solution as compared to the basic PSO which needs about 60 iterations for optimum solution. The low
values of shrinkage and warpage confirms that the proposed MPSO algorithm provides improved results. This will
enhance the final accuracy and quality of the injection molded component and hence the result of injection molding
process will be improved.
From the obtained results, it is clear that shrinkage and warpage is more when packing pressure and packing time is
less. The predicted MPSO results has slightly higher value of packing pressure as compared to initial values. In
general, to balance the filling pattern, an adequate packing pressure is required that also allow sufficient packing
time. Also the sufficient packing time allow uniform cooling across both the thickness and part and also reduces the
residual stress and hence the shrinkage and warpage are reduced.

Table 5. Optimum parameters prediction using PSO and MPSO

Melt temp (T m ) Packing press (P p ) Packing time (P t ) S1 (%) S2(%) W1 (mm) W2 (mm)
PBTi PETi PBTi PETi PBTi PETi PBTi PETi PBTi PETi
278 252 25 34 10 12 1.9197 1.9145
Initial value
266 252 25 28 15 18 0.1341 0.1608
PSO 279.32 254.02 27.12 29.65 10.87 12.07 1.4203 1.5182
269.72 253.91 26.79 29.01 14.67 17.89 0.1221 0.1365
MPSO 281.09 254.15 28.54 35.17 10.34 12.53 1.2196 1.3298
268.26 253.12 27.91 28.34 15.91 18.22 0.1082 0.1145
Objective function Value

101

10° (a) (b)

Figure 3.Convergence plot of (a) Shrinkage (b) Warpage of PBT material

In order to test the adequacy of the developed mathematical model and justify the use of newly developed MPSO
algorithm, four simulation runs were performed for shrinkage (S1, S2) and warpage (W1, W2). The shrinkage and
warpage results for PBT material is shown in Figure 4 for the chosen parameter values. The data from the
confirmation trials and their comparison with the predicted values of shrinkage and warpage using MPSO algorithm
is shown in Table 6. From Table 6, it is clearly seen that the predicted values of shrinkage and warpage are more
accurate for the predicted model in comparison to the default simulation parameters.

319
Warpage and shrinkage minimization in an injection molded component using improved particle swarm optimization algorithm

Table 6. Confirmation simulation trials

S. Parameters Shrinkage (%) Warpage (mm)

No. A B C Simulation Predicted Simulation Predicted
1 278 34 8 2.2289 2.2003 0.2922 0.2812
2 280 36 10 1.8722 1.7621 0.3681 0.3603
3 282 38 12 2.8021 2.745 0.2690 0.2231
4 281.09 35.17 10.34 2.321 2.1098 0.2901 0.2521

Figure 4. Shrinkage and warpage analysis using optimized parameter

7. Conclusion

This paper presents an integrated methodology for developing mathematical models and predicting the values of
shrinkage and warpage by correlating them with the process parameters of plastic injection molding for making the
electronic relay component using PBT and PET material. The parameters considered for the prediction of shrinkage
and warpage are Melt temperatureT m , Packing pressure P p and Packing timeP t . To find the optimum value of process
parameters, the analytical model using regression analysis was developed. To further improve the optimum values a
recently developed modified particle swarm optimization algorithm was used. The results obtained of ANOVA
analysis after conducting confirmation experiments shows that the analytical model of the shrinkage and warpage
are fairly comparable and is well in agreement with the simulation values. The predicted response of injection
molding processalso shows an improvement, using a newly developed MPSO algorithm.

References
[1] Noble J, Walczak K, Dornfeld D. Rapid tooling injection molded prototypes: a case study in artificial
photosynthesis technology. Procedia CIRP. 2014 Jan 1; 14:251-6.
[2] Chiang KT, Chang FP. Analysis of shrinkage and warpage in an injection-molded part with a thin shell feature
using the response surface methodology. The International Journal of Advanced Manufacturing Technology.
2007 Dec 1; 35(5-6):468-79.
[3] Chen CC, Su PL, Lin YC. Analysis and modeling of effective parameters for dimension shrinkage variation of
injection molded part with thin shell feature using response surface methodology. The International Journal of
Advanced Manufacturing Technology. 2009 Dec 1; 45(11-12):1087.
[4] Ozcelik B, Erzurumlu T. Determination of effecting dimensional parameters on warpage of thin shell plastic
parts using integrated response surface method and genetic algorithm. International communications in heat
and mass transfer. 2005 Aug 1; 32(8):1085-94.

320
 Proceedings of NCAMMM - 2018

[5] Liao SJ, Hsieh WH, Wang JT, Su YC. Shrinkage and warpage prediction of injection‐molded thin‐wall parts
using artificial neural networks. Polymer Engineering & Science. 2004 Nov 1; 44(11):2029-40.
[6] Chiang KT. The optimal process conditions of an injection-molded thermoplastic part with a thin shell feature
using grey-fuzzy logic: A case study on machining the PC/ABS cell phone shell. Materials & design. 2007 Jan
1; 28(6):1851-60. \
[7] Hassan H, Regnier N, Pujos C, Arquis E, Defaye G. Modeling the effect of cooling system on the shrinkage
and temperature of the polymer by injection molding. Applied Thermal Engineering. 2010 Sep 1; 30(13):1547-
57.
[8] Leo V, Cuvelliez CH. The effect of the packing parameters, gate geometry, and mold elasticity on the final
dimensions of a molded part. Polymer Engineering & Science. 1996 Aug 1; 36(15):1961-71.
[9] Galantucci LM, Spina R. Evaluation of filling conditions of injection moulding by integrating numerical
simulations and experimental tests. Journal of materials processing technology. 2003 Oct 20; 141(2):266-75.
[10] Huang MC, Tai CC. The effective factors in the warpage problem of an injection-molded part with a thin shell
feature. Journal of materials processing technology. 2001 Mar 1; 110(1):1-9.
[11] Tang SH, Tan YJ, Sapuan SM, Sulaiman S, Ismail N, Samin R. The use of Taguchi method in the design of
plastic injection mould for reducing warpage. Journal of Materials Processing Technology. 2007 Feb 2; 182(1-
3):418-26.
[12] Park K, Ahn JH. Design of experiment considering two-way interactions and its application to injection
molding processes with numerical analysis. Journal of Materials Processing Technology. 2004 Feb 28;
146(2):221-7.
[13] Park HS, Dang XP. Design and simulation-based optimization of cooling channels for plastic injection mold.
InNew Technologies-Trends, Innovations and Research 2012. InTech.
[14] Choi DS, Im YT. Prediction of shrinkage and warpage in consideration of residual stress in integrated
simulation of injection molding. Composite Structures. 1999 Dec 1; 47(1-4):655-65.
[15] Liao SJ, Hsieh WH, Wang JT, Su YC. Shrinkage and warpage prediction of injection‐molded thin‐wall parts
using artificial neural networks. Polymer Engineering & Science. 2004 Nov 1; 44(11):2029-40.
[16] Kikuchi H, Koyama K. Warpage, anisotropy, and part thickness. Polymer Engineering & Science. 1996 May
1; 36(10):1326-35.
[17] Kurtaran H, Erzurumlu T. Efficient warpage optimization of thin shell plastic parts using response surface
methodology and genetic algorithm. The International Journal of Advanced Manufacturing Technology. 2006
Jan 1; 27(5-6):468-72.
[18] Chiang KT. Modeling and optimization of designing parameters for a parallel-plain fin heat sink with confined
impinging jet using the response surface methodology. Applied Thermal Engineering. 2007 Oct 1; 27(14-
15):2473-82.
[19] Lin JC. Optimum cooling system design of a free-form injection mold using an abductive network. Journal of
Materials Processing Technology. 2002 Jan 15; 120(1-3):226-36.
[20] Li CG, Li CL. Plastic injection mould cooling system design by the configuration space method. Computer-
Aided Design. 2008 Mar 1; 40(3):334-49.
[21] Chiang KT, Chang FP. Application of grey-fuzzy logic on the optimal process design of an injection-molded
part with a thin shell feature. International Communications in Heat and Mass Transfer. 2006 Jan 1; 33(1):94-
101.
[22] Pathak VK, Singh AK, Singh R, Chaudhary H. A modified algorithm of Particle Swarm Optimization for form
error evaluation. tm-Technisches Messen. 2017 Apr 20; 84(4):272-92.

321
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Parametric Optimization of Aluminum Metal Matrix Composite (AMC) with

Reinforcement of Coconut Shell Ash

G.Srinivasarao1, Siva Sankara Raju2 , K. Vikash Kumar3

1
Department of Mechanical Engineering, RVR & JC. College of Engineering, Guntur, Andhra Pradesh, 522019,
E-mail:gsraorvr@gmail.com
2
Department of Mechanical Engineering, Gandhi Institute of Engineering and Technology, Gunupur, Odisha -
765022, India.
3
PG student, Department of Mechanical Engineering, Gandhi Institute of Engineering and Technology, Gunupur,
Odisha - 765022, India.

Abstract: This study has been described on optimization of tribological characteristics of Al-coconut shell ash
(CSA) reinforced composite prepared with stir casting route. In this study, three operating variable (i.e., load,
sliding speed, and % of CSA) and three response (i.e. wear (µm), wear rate (mm3/m), and coefficient of friction) are
considered. The experiments are designed and performed at all combination of input parameters. The influence of
each parameter on the response is established using response tables and response graphs. Based upon experimental
data, the model equations for each response were developed with multiple linear regression. The models give the
factor effects of individual parameters. Interaction effects provide additional information to understand the detail
behavior of parameters. It revealed that load is the most influencing effect on wear behavioral responses. An
optimum parameter combination was obtained by using genetic algorithm.
Keywords: Composites, Coconut shell ash, Design of experiments, Genetic algorithm, Regression.

1. Introduction
Aluminium matrix composites (AMCs) are gained extensive concentration owing to better specific strength, and
exceptional tribological and machinability properties in excess of the unitary alloys(1–4). However, AMCs have
better-quality properties but unable to expose due to intricacy in fabrication and characterization. In general, Stir
casting technique is a cheapest and widely used in mass production for automobile industries. Wear is a linear
dimensional loss in the material occurred due to abrasion. Extensive studies on the tribological performance of AMC
are considering with the effect of variables such as particulate size, shape, volume, of reinforcements, and
load/pressure, velocity, sliding distance and other extensive operating conditions. Similarly, the tribological
behaviour of composite has also been enhanced due to distribution of reinforcement in matrix which influences the
affect of interfacial bonding between matrix and reinforcement (1,5–8). Sahin (3) optimized the wear behaviour of
Al-SiC-MMCs and reported that the effect of abrasive particles size and volume is played a key role in abrasion
mechanism. Kok and Ozdin (1) studied the tribological performance of Al-2024- Al 2 O 3 composite. Present work is
aimed to determine the optimal tribological performance of Al-1100-CSA composites reinforced with 5, 10 and 15
volume %. Moreover, response table and response graphs are used to investigate the significant parameters, effect
on tribological performance of CSA composite.
Parametric optimization of aluminum metal matrix composite (AMC) with reinforcement of coconut shell ash

2. Materials and methods

2.1.Matrix material
1100 Al (99.0%Cu) is heavy pure aluminium based 1xxx series alloy. The suggested properties of this alloy over
the other materials are low in density, better electrical conductivity and work hardened material. This alloy is mainly
preferred in few common application i.e., fins, spun hollowware, rivets, chrome works, gift wares and stamping
plates etc. The composition of material is shown in Table 1.

Table 1 Elemental composition of Al-1100

Fe Si Cu Zn Mn Residual Bal.
0 - 0.95% 0 - 0.95% 0.05 - 0.2% 0 - 0.1% 0 - 0.05% 0 - 0.15% Al

2.2 Preparation of coconut shell ash

Coconut shell is obtained from local market. Its surface skin is cleaned and smoothened. Then, the material is
crushed and grounded in jaw crusher and ball mill respectively. The grinded powder is mesh to size of 100, and then
is burnt at 1450 °C for 3hrs. Burnt ash is again grounded up to 4 hr. The obtained particles are sieved to a fine mesh
of 240 size approximately equivalent to 60μm(9). Table 2, shows the compounds present in CSA.

Table 2 Chemical Composition of CSA

Constituents Fixed C SiO 2 Fe 2 O 3 Al 2 O 3 MgO CaO K2O
% of Wt. 4–6 40-45 8-9 23-25 2–3 3-3.5 0.7 - 0.95

Figure 1. Raw coconut shell and coconut shell ash

2.3 Preparation of Al- CSA-Composites

Al-1100 alloy ingot is cut into small pieces. This pieces are preheated in muffle furnace at a temperature of 572K
(300 ºC) for 1 hr. The preheated material is devolved to a bottom pouring furnace whose temperature is 1472K (800
ºC). Entire melting process has been discussed elsewhere (4,9). The cast materials such as 0% CSA (base alloy),
5% CSA, 10% CSA and 15% CSA composites are prepared with variation of 5% volume.

323
 Proceedings of NCAMMM - 2018

2.4 Experimental procedure

In this work, the experiments are carried out based on full factorial design which consists of 27 experimental runs
for three factors at three levels. The three operating variables such as Load (L), % of coconut shell ash (R) and
sliding distance (D) at three different levels as shown in Table 3. Experiments were carried out in random manner in
order to reduce the errors. Table 4 shows, the tribological performances such as wear (W), wear rate (WR
((mm3/m)*10-3)), and coefficient of friction (C.F) for all combination of experiments.

Table 3. Parameters and their levels

Factor symbol Factor (units) Low(-1) Medium( 0) High(+1)

L Load (N) 10 30 50
R % of CSAp, (% of vol.) 5 10 15
D Sliding distance,(m) 1000 2000 3000

Table 4. Experimental layout with results

Run L R D W WR CF Run L R D W WR CF
1 10 5 1000 191 5.556 0.265 15 30 10 3000 460 2.436 0.226
2 10 5 2000 193 2.815 0.201 16 30 15 1000 268 5.808 0.355
3 10 5 3000 385 2.173 0.188 17 30 15 2000 312 3.115 0.292
4 10 10 1000 185 5.000 0.224 18 30 15 3000 426 2.321 0.28
5 10 10 2000 180 2.423 0.185 19 50 5 1000 480 7.111 0.663
6 10 10 3000 210 2.159 0.197 20 50 5 2000 521 4.167 0.501
7 10 15 1000 165 3.462 0.249 21 50 5 3000 685 4.012 0.391
8 10 15 2000 175 2.327 0.234 22 50 10 1000 415 6.923 0.582
9 10 15 3000 268 1.936 0.271 23 50 10 2000 505 3.962 0.445
10 30 5 1000 380 6.333 0.412 24 50 10 3000 525 2.846 0.359
11 30 5 2000 391 3.222 0.299 25 50 15 1000 398 6.615 0.565
12 30 5 3000 478 2.704 0.238 26 50 15 2000 468 3.654 0.453
13 30 10 1000 310 6.500 0.351 27 50 15 3000 478 2.641 0.392
14 30 10 2000 285 3.115 0.263

3. Data analysis and discussion of results

Design of experiments is aiming at obtaining the relationship between operating variables and tribological
performances. This study mainly concerning with determination of effect of variables on individual tribological
response and combined response affect on tribological behaviour of CSA composites.

3.1 Analysis of the process parameters and their interactions- From the experimental data, level means for each
response are computed and shown in Table 5. The influence of each variable parameter (i.e., L, R and D) on the
tribological performances (i.e., W, WR and CF) is carried out using level mean analysis.

324
Parametric optimization of aluminum metal matrix composite (AMC) with reinforcement of coconut shell ash

A level mean is the average value of response at specific level. The response graphs are representing the
effect of process parameters with individual tribological performance. The slope of the line indicates the effect
variation at each level with corresponding response. The interactions of input parameters are also having significant
effect on responses. The influence of process parameters (i.e., L, R and D) on the tribological performance (Wear
(W), wear rate (WR) and coefficient of friction (CF)) are observed, which is shown in Fig.2.

Table 5. Level mean values for response

level Wear(µm) Wear rate(mm3/m)*10-3 Coefficient of friction
L R D L R D L R D
-1 216.9 411.6 310.2 3.09 4.23 5.92 0.224 0.351 0.407
0 367.8 341.7 336.7 3.95 3.93 3.20 0.302 0.315 0.319
+1 497.2 328.7 435.0 4.66 3.54 2.58 0.483 0.343 0.282
Difference 280.3 82.9 124.8 1.57 0.69 3.34 0.259 0.039 0.125

Figure 2. Response graphs

From the response table and response graphs, it is identified that, L is the dominating factor on the responses such as
wear and wear rate, whereas D is dominating parameter on CF. The tribological performances (i.e. wear, WR and
CF) increase with increasing load. While the load increases the composite losses its ability to carry the load due to
fracture of reinforced particles and formation of debris, which causes the loss of material vis-a-vis ploughing,
delamination, and plastic deformation. At a higher load, the frictional force increases with greater dissipation of
energy leading to a higher temperature at contact surface which results in higher value of CF. Wear increasing with
increase of sliding distance and decreases with addition of reinforcement (CSA) due to hard ceramic compounds
present in CSA which leads to greater bonding of the matrix. The addition of CSA particles having minor effect on
CF and better value is observed at 10% of CSA, whereas more the value of the sliding distance, lesser the value of
CF. The optimal conditions for wear, WR and CF are L 1 R 3 D 1 , L 1 R 3 D 3 and L 1 R 2 D 3 respectively.

325
 Proceedings of NCAMMM - 2018

3.2 Modeling
Second order models for wear, WR and CF are developed using full factorial design experimental data. The input
data to SPSS software is provided in actual form of factors. Model equations for each response are given below.
Wear= 253.358+7.01L-31.044 R+ 1.138 R2 + 1.601*10-5 D2 (R2 =0.937)

Wear rate= 9.147+0.067L-0.069R- 0.005D+1.052*10-6 D2 – 1.147 * 10-5 LD (R2 =0.965)

CF=0.441+0.006L-0.031R +0.001 R2+ 2.567* 10-8 D2 – 2.442*10-6 LD+ 4.95* 10-6 RD (R2 =1.00)

The R2 value of 0.942 specifies that 94.2% of the variation in the wear has been given by the model. Similarly, R2
value of WR is 0.942 and CF is 1.00. From regression equation, the interaction between WR with LD, interaction of
CF with LD and RD is identified.

Figure 3. 3-D surface plots

From the above equations, it can be observed that the L, R, D and their square terms and interaction terms also
coming into the model. The square terms of parameters indicates the quadratic nature of response. The three
individual responses are not having the same optimum input values. To satisfy all the responses, it requires multi-
response optimization and has been performed by using genetic algorithm.
The optimum values of L, R, and D are 10, 13.61 and 1795 respectively and corresponding values of responses
wear, WR and C.F are 165.28, 3.32 and 0.42. The interaction affects can also be observed from the 3-D surface plots

326
Parametric optimization of aluminum metal matrix composite (AMC) with reinforcement of coconut shell ash

which are shown in Figure 3. The confirmation test has been performed and it is observed that % of error with in
limit, hence the model is adequate. So, the model equations are corroborating the process parameters and
tribological performance, with in working range.

4. Conclusions

1. Aluminum 1100 alloy CSA composites (i.e. Al-5%CSA, Al-10%CSA, Al-15% CSA) are prepared indigenously
with a laboratory scale by stir-casting process.
2. Full factorial experiments for three process parameters at three levels were conducted to examine the tribological
performance (i.e., wear, WR and CF) on CSA composites.
3. The wear resistance of composite increases with increasing CSA volume and decreased with increasing load.
4. Load (L) is the dominating factor on the responses such as wear and wear rate (WR), whereas Sliding distance
(D) is dominating parameter on coefficient of friction (CF).
5. Optimum parameters were predicted through genetic algorithm technique (load =10 N, percentage of coconut
shell ash= 13.61 and sliding speed = 1795 m) to achieve the best responses.

References
[1] Kök M, Özdin K. Wear resistance of aluminium alloy and its composites reinforced by Al2O3 particles. J Mater
Process Technol [Internet]. 2007 Mar 1 [cited 2016 Feb 13];183(2–3):301–9. Available from:
https://www.researchgate.net/publication/248252569_Wear_resistance_of_aluminium_alloy_and_its_composite
s_reinforced_by_Al_2O_3_particles
[2] Rohatgi PK, Schultz B. Lightweight Metal Matrix Nanocomposites - Stretching the Boundaries of Metals.
Mater Matters. 2007;2(4):1–6.
[3] Sahin Y. Abrasive wear behaviour of SiC / 2014 aluminium composite. Tribiology Int. 2010;43:939–43.
[4] Siva Sankara Raju R, Panigrahi MK, Ganguly RI, Srinivasa Rao G. Investigation of Tribological Behavior of a
Novel Hybrid Composite Prepared with Al-Coconut Shell Ash Mixed with Graphite. Metall Mater Trans A
[Internet]. Springer US; 2017;48(8):3892–903. Available from:
https://link.springer.com/article/10.1007/s11661-017-4139-1
[5] Hatch JE. Aluminum Properties and Physical Metallurgy. In: ASM, editor. ASM International. Ohio, US; 1984.
p. 1–24.
[6] Kato K. Wear in relation to friction — a review. Wear. 2000;241:151–7.
[7] Rodopoulos CA. Metal Matrix Composites. In: Advanced Materials by Design. 2004. p. 99–117.
[8] Rohatgi PK. Metal-matrix Composites. Def Sci J. 1993;43(4):323–49.
[9] Raju RSS, Rao GS. Assessment of Tribological performance of Coconut Shell Ash Particle Reinforced Al-Si-Fe
Composites using Grey-Fuzzy Approach. Tribol Ind. 2017;39(3):364–77.

327
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Parametric Investigation of E-Jet Micro manufacturing Process: Taguchi Robust

Design Approach

Raju Das1, Amit Ball1, Atul Priya1, Shibendu Shekhar Roy1, Naresh Chandra Murmu2
1
Department of Mechanical Engineering, NIT-Durgapur, 713209, India
2
Principal Scientist, CSIR-CMERI, Durgapur, 713209, India
E-mail: ssroy99@yahoo.com, Tel.: +913432754702; Fax: +913432543447

Abstract: E-jet printing is a high-resolution micro manufacturing process. It finds its application ranging from
printed electronics to biotechnology, microelectromechanical (MEMS) systems, optoelectronics, etc. Suitable
selection of operating environment could result in better printing performance. In this paper, Taguchi robust
design approach has been employed to investigate different operational strategies of E-jet micro manufacturing
process. A favorable operating environment of E-jet has been identified with the help of signal to noise (S/N)
ratio analysis. Applied pressure, nozzle substrate gap, and applied voltage have been chosen as operational
control variables. Droplet deposition rate has been selected as printing performance, which is to be optimized
through the current case study. For experimentation purpose, Taguchi Orthogonal array (OA) has been used.
According to the experimental layout different combination of operational control variables have been chosen
for analysis. Optimal factor level combination of control variables has been proposed.
Keywords: Microfabrication, E-jet, Ink-jet printing, Taguchi robust design, signal to noise (S/N)

1 Introduction
Application of lithographic techniques in the field of printed electronics is well known. However, it suffers
some limitations such as complex pre and post-processing operation, contamination; material constraints and
unable to handle flexible, stretchable substrate. Flexible, stretchable substrate is essential for fabricating human
wearable electronics also cutting-edge foldable display devices. To meet the current technological needs, search
for new complementary technologies is a very active area of research. Recently inkjet printing technology finds
its foothold in the printed electronics industry. Traditionally conventional inkjet printing technologies such as
thermal inkjet (TIJ) and piezo-based inkjet (PIJ) has been used for microfabrication. However,the above-
mentioned inkjet methods suffers from resolution aspect. High-resolution feature generation is problematic with
those technologies. As the ejected material dimension is the order of two as compared to nozzle dimension, for
high-resolution printing micron or even nano size nozzles have to be employed, which is difficult to
manufacture. Nozzle clogging with high viscous ink material also poses a problem. Aforementioned limitations
can be successfully mitigated by a new inkjet printing technology known as Electrohydrodynamic (EHD) inkjet
printing or E-jet printing. It is a microfabrication technology where high-resolution deposition of functional
material is achieved with the help of externally applied electric field. Due to its ability process wide variety of
functional material, ability to incorporate flexible stretchable substrate (a surface on which material is deposited)
and mask-less vacuum free additive in nature fabrication process, it finds its application ranging from printed
electronics to sensor fabrications, biotechnology, photonic and plasmonic device fabrication as a high resolution
patterning medium. It is a simple microdispensing system where ink material gets pulled instead of pushed with
virtually no material wastage. In E-jet printing a voltage difference is applied between the conducting nozzle
Parametric investigation of E-Jet micro manufacturing process: Taguchi robust design approach

and the substrate. When the applied electric field overcomes the surface tension of the ink material, the pendant
ink meniscus at the tip of the nozzle deforms and form a cone shape whose diameter is smaller than the nozzle
diameter, which is known as Taylor cone. The deposition of the material occurs at the tip of the cone not at the
tip of the nozzle thus producing a smaller feature size than the ejector opening. The entire operation of E-jet
printing can be summarised by the formation of Taylor cone followed cone relaxation as a result of deposition.
Both continuous conductive lines, as well as discrete droplets, can be realised with the help of E-jet printing
depending on the application. Various ejection modes can be observed while printing namely- micro dripping
mode, pulsating jet mode, stable cone jet mode, spraying. For drop on demand kind of operation pulsating jet
mode of ejection is useful, whereas continuous feature creation stable cone-jet mode is required. Both
continuous and pulsed voltage source can be used during operation. However, pulsed mode gives more
flexibility to the user while operating. The main components of E-jet printing system are- ink reservoir and
pressure actuation system, conducting nozzle, substrate, a moving stage where the substrate is mounted, an
external voltage source and a vision system usually a high-speed camera to observe stable printing regime.
The entire process operation is composed of generation of Taylor cone due to the application of electric
field and relaxation of the same due to material ejection. The whole system is comprised of several sub-systems,
which are shown in the following schematic diagram. It consists of ink chamber, pressure application system,
external voltage supply system and a computerized moving stage on which substrate is mounted.

Computer
Pressure supply

Ink chamber

Nozzle
Voltage supply

Substrate

XY stage
Fig. 1 Schematic diagram of the EHD printing system

The deposition of liquid droplets with the help of electrical forces was discovered long ago [1]. Since then,
electrically driven deposition technique of functional material found its application in diverse field. As the
electrohydrodynamic induced inkjet microfabrication method has several advantages over other conventional
microfabrication methodologies, it has been successfully implemented in the areas of printed electronics,
biotechnology, optical and plasmonic applications [2-4]. A research group from university of Illinois [5],
successfully printed several functional material in the range from sub-micron to nano level with the help of E-jet
printing. They demonstrated the feasibility of EHD printing in the area of printed electronics. For the
improvement of the printing process, different aspect of the printing process has been investigated over the
years. The available literature could be subdivided into following groups namely- research work devoted to
understand the fundamental principles of the technology and demonstrate its potential, i.e., achieving high-
resolution feature size [6-8], research work deals with control and throughput issues of the problem [9-13],
studies on ink properties [14-15] and specific application-oriented findings[16-17].

329
 Proceedings of NCAMMM - 2018

After going through the literature,it has found that no work has been attempted to predict the optimal
operating condition for droplet deposition frequency of E-jet printing process with the application of Taguchi
philosophy. In this study, an attempt has been made to propose a parameter setting which would result in
improved performance of the process. The following section of the article has arranged in the following order:
first a brief overview of the methodology followed by data collection, then its analysis and discussion of
gathered information and lastly some concluding remarks.

2. Methodology
Taguchi philosophy has been widely applied [18-19] for optimization, quality improvement studies of a given
process. It is a very simple technique involves less computational work and easy to implement for off-line
quality control purpose. The selection criteria for optimal parameter setting in this methodologyis signal to noise
ratio (S/N) ratio analysis. It the ratio between a response of the process and its deviation from the ideal operating
condition. Here the response is termed as signal and contribution of uncontrollable factors which leads to the
deviation is termed as noise. According to the methodology, parameter setting which yields maximum signal to
noise (S/N) ratio is most likely to be the optimal one. As the signal to noise (S/N) ratio increases the noise effect
during the printing operation decreases which makes the process more robust. There are three types of S/N ratio
available namely- smaller-the-better, larger-the-better, nominal-is-the-best; application of each depends on
given problem statement. In this study, larger-the-better criteria has been used for data analysis [18].
Larger-the -better-

1 (1)
S/N= −10 log
n
∑ y2
Where n is the number of observations at a particular factor level combination and y is the value of the response
variable. In the present study, n=1; y is the value of average droplet deposition frequency.

3.0 Data analysis & Discussion

In this present study, the experimental data has been taken from previous parametric studies conducted by Graph
[20]. The experimental layout chosen for conducting the experiment was L 8 OA (orthogonal array). It is a two
level design of experiment technique to explore the parametric region. It is fractional factorial design involves
eight no. of experiments. In the present work applied voltage, ejection pressure and the stand-off height between
the nozzle and the substrate taken as operational parameters which are to be tuned and average droplet
frequency or droplet deposition rate taken as printing performance parameter. The level values of process
parameters are given in Table 1. Table 2 shows OA experimental layout.

Table 1. Factors and levels

Code Factor Level
1 2
A Pres. (psi) 0.25 1.0
B Gap (µm) 30 50
C Volt. (V) Low High

330
Parametric investigation of E-Jet micro manufacturing process: Taguchi robust design approach

Table 2. L8 Orthogonal Array

Experiment no Control Factors
A B C
1 1 1 1
2 1 1 2
3 1 2 1
4 1 2 2
5 2 1 1
6 2 1 2
7 2 2 1
8 2 2 2

The observation data set is given in Table 3, which was used for calculation of signal to noise(S/N) ratio values.
Eq 1. has been used to calculate signal to noise (S/N) ratio values. Since the larger deposition frequency is of
desired response of the E-jet process, larger is better criteria is chosen for analysis.

Table 3. Control variables and response variable [20]

Experiment no. Pres. (psi) Gap (µm) Volt. (V) Ave. Drop Freq. (Hz)
1 1 30 Low 50.88
2 1 30 High 383.64
3 1 50 Low 74.08
4 1 50 High 326.69
5 0.25 30 Low 83.38
6 0.25 30 High 285.14
7 0.25 50 Low 39.77
8 0.25 50 High 253.13

Table 4. Signal to Noise Ratio values

Experiment no. S/N ratio
1 34.13
2 51.67
3 37.39
4 50.28
5 38.42
6 49.10
7 31.99
8 48.06

Calculated signal to noise (S/N) ratio values for each observation is tabulated in Table 4. The estimatedsignal to
noise (S/N) ratio values were used to determine the effect of each parameter on the response as well as to predict
the optimum values of process parameters. The results of Taguchi analysis were given in Table 5.

Table 5. Results of Taguchi Analysis

Factor S/N ratio ∆=Abs(Max.-Min.) Rank
Level 1 Level 2
Pres. (psi) 43.37 41.90 1.48 2
Gap (µm) 43.33 41.93 1.40 3
Volt. (V) 35.48 49.78 14.30 1

331
 Proceedings of NCAMMM - 2018

According to the Taguchi robust design,higher signal to noise (S/N) ratio value leads to better performance of
the process irrespective of the objective. As noise factors are inthe denominator, higher the S/N ratio value
means the contribution of the noise factors has minimal effect. FromTable 5. It can be seen that maximum signal
to noise (S/N) ratio values for pressure, gap or standoff height, and voltage are at low, low, and high level
respectively. The predicted optimal level of operation setting in coded form is A1 B1 C2. Delta ∆ values [21] in
Table5. represents the relative importance of process parameters on the response. Maximum the delta ∆ value,
more is the importance. From the rank column, it can be seen that applied voltage has the maximum influence
regarding the on droplet deposition frequency.

4. Conclusion
In the present work, an attempt has been made to improve the Electrohydrodynamic based ink-jet printing
operation by selecting the process parameter levels at the desirable environment. A single objective problem of
EHD printing has been analyzed with the help of Taguchi robust design approach. Experimentations have been
conducted according to the L 8 orthogonal (OA) experimental plan. Optimal level combination of process
parameters was proposed based on S/N ratio analysis. The predicted optimal factor combination is A1 B1 C2.
Influential factors affecting the response have also been identified. Applied voltage has found to be the most
influential control factor over the investigated region. The above methodology can be used for parametric
optimization of other processes and off-line quality control.

Acknowledgement:
The authors gratefully acknowledge all the support from Department of Mechanical Engineering, NIT Durgapur
and Surface Engineering and Tribology Section, CSIR-CMERI Durgapur.

References
[1] Zeleny J. Instability of Electrified Liquid Surfaces. Physical Review. 1917 July 1; 10(1). doi:
http://dx.doi.org/10.1103/PhysRev.10.1
[2] Jayasinghe S. et al. Electrohydrodynamic jet processing: an advanced electric-field-driven jetting
phenomenon for processing living cells. Small. 2005 Dec 6; 2(2):216-219. doi: 10.1002/smll.200500291
[3] Lee D. et al. Electrohydrodynamic printing of silver nanoparticles by using focused nanocolloid jet.
Applied Physics Letters. 2007 Feb ; 90:081905. doi: http://dx.doi.org/10.1063/1.2645078
[4] Park J. et al. Nanoscale patterns of oligonucleotides formed by electrohydrodynamic jet printing with
applications in biosensing and nanomaterials assembly. Nano Letters. 2008 Octo 29; 8(12): 4210-4216. doi:
10.1021/nl801832v
[5] Park J.U. et al. High-resolution electrohydrodynamic jet printing. Nature Materials. 2007 Aug 5;6:782-
789.doi:10.1038/nmat1974
[6] Chen C.H. et al. Electrohydrodynamic “drop-and-place” particle deployment. Applied Physics Letters. 2006
Mar; 88:154104.doi: 10.1063/1.2191733
[7] Chen C.H. et al. Scaling laws for pulsed electrohydrodynamic drop formation. Applied Physics
Letters.2006 Aug; 89:124103. doi: http://dx.doi.org/10.1063/1.2356891
[8] Choi H.K. et al. Scaling laws for jet pulsations associated with high-resolution electrohydrodynamic
printing.Applied Physics Letters.2008 Mar; 92:123109. doi: 10.1063/1.2903700

332
 Parametric investigation of E-Jet micro manufacturing process: Taguchi robust design approach

[9] Barton K, Mishra S, Alleyne A. et al. Control of high-resolution electrohydrodynamic jet printing. Applied
Physics Letters. 2008 Mar; 92:123109. doi: 10.1016/j.conengprac.2011.05.009
[10] Li J.L. On the meniscus deformation when the pulsed voltage is applied. Journal of Electrostatics. 2006 Jan;
64(1):44-52. doi:https://doi.org/10.1016/j.elstat.2005.04.005
[11] Xu L. et al. Electrohydrodynamic Deposition of Polymeric Droplets under Low-Frequency Pulsation.
Langmuir. 2011 April 20; 27(10):6541-6548. doi: 10.1021/la201107j
[12] Mishra S. et al. High-speed and drop-on-demand printing with a pulsed electrohydrodynamic jet. Journal of
Micromechanics and Microengineering. 2010 Aug 20; 20:095026. doi:10.1088/0960-1317/20/9/095026
[13] Kim Y.J. et al. On-demand electrohydrodynamic jetting with meniscus control by a piezoelectric actuator
for ultra-fine patterns. Journal of Micromechanics and Microengineering. 2009 Sep 14; 19:107001.
doi:10.1088/0960-1317/19/10/107001
[14] Yu M. Design optimization of ink in electrohydrodynamic jet printing: Effect of viscoelasticity on the
formation of Taylor cone jet. Materials and Design. 2016 Jan; 89(5):109-115. doi:
http://dx.doi.org/10.1016/j.matdes.2015.09.141
[15] Yu J.H. et al.: Effect of viscosity of silver nanoparticle suspension on conductive line patterned by
electrohydrodynamic jet printing. Applied Physics A. 2007 Oct; 89(1):157-159. doi: 10.1007/s00339-007-
4210-7
[16] Lim S. Electrohydrodynamic printing of poly (3, 4-ethylenedioxythiophene): poly (4-styrenesulfonate)
electrodes with ratio-optimized surfactant. RSC Advances. 2015 Dec 18; 6:2004-2010. doi:
10.1039/c5ra19462h
[17] N C et al. EHD Printing of PEDOT: PSS Inks for Fabricating Pressure and Strain Sensor Arrays on Flexible
Substrates. SPIE Proceedings. 2015 June 2; 9494. doi: 10.1117/12.2177415
[18] Bagci E et al.A study of Taguchi optimization method for identifying optimum surface roughness in CNC
face milling of cobalt-based alloy (stellite 6). The International Journal of Advance Manufacturing
Technology. 2006 July; 29:940. doi: 10.1007/s00170-005-2616-y
[19] Yousefieh M et al. Taguchi Optimization on the Initial Thickness andPre-aging of Nano-/Ultrafine-Grained
Al-0.2 wt.%Sc Alloy Produced by ARB. Journal of Materials Engineering and Performances. 2016 Oct;
25(10):4239-4248. doi: 10.1007/s11665-016-2273-3
[20] Philip G. A 2nd generation electrohydrodynamic jet (e-jet) printing system, parametric studies of e-jet
nozzles and integrated electrode e-jet deposition. University of Illinois at Urbana-Champaign. M.S. Thesis.
2011 May doi: http://hdl.handle.net/2142/24288
[21] Manivel D et al.Optimization of surface roughness and tool wear in hard turning of austempered ductile
iron (grade 3) using Taguchi method. Measurement. 2016 Nov; 93:108-116.
doi:http://dx.doi.org/10.1016/j.measurement.2016.06.055

333
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018 at
CSIR-CMERI, Durgapur

Data-based Modeling and Continuous Adjustment of CNC Turning Process: A Case Study

O M Vinod1 and P B Dhanish2

1
Faculty of Mechanical Engineering, LBS College of Engineering Kasaragod and PhD scholar, NIT Calicut.
Email:omvinodom@gmail.com
2
Faculty of Mechanical Engineering, NIT Calicut. Email: dhanish@nitc.ac.in

Abstract: Achieving dimensions on target is an important requirement in machining operations. This necessitates
adjustment of the process and we need rules to make this adjustment. This paper reports the implementation of an
adjustment rule using a forecasting model based approach. CNC turning was carried out without any adjustment
and the best fitting model forecasting the dimension with respect to time was determined. A number of auto
regressive integrated moving average (ARIMA) models were attempted and ARIMA (0,1,2) was found to be the best
suited one. An extended version of the ordinary exponential weighted moving average (EWMA) equation was
developed to carry out process adjustment using this model. Adjustment was carried out continuously using this
model and the results show good performance based on mean squared deviation (MSD) from target.

1. Introduction
Turning is one of the commonly used discrete part manufacturing operations. Ensuring dimensional accuracy in
turning operations is a challenge while machining hard and difficult to machine materials mainly due to the presence
of tool wear. As the machining proceeds, the cutting tool wears out, causing the effective depth of cut to decrease,
resulting in gradual increase of the diameter of the work-piece being machined. The general approach of machining
industries is to make periodic adjustments of cutting tool after specified intervals in order to obtain desired
dimensions. Usually such attempts are aimed at compensating the last observed deviation from target. But these
periodic adjustments do not make full use of the information available from the process.
Methods of process adjustment techniques in machining industries have been discussed by many authors.
Grubbs [1] proposed an optimum procedure for setting machines or adjusting processes based on a harmonic rule.
The issues of adjusting a machine with faulty setup have also been analyzed and resolved by Trietsch [2], Colosimo
et al. [3] and Pan [4]. These techniques are not suitable in situations where tool wear is present. Cai et al. [5]
presented an economic design of control chart for trended process. But they gave only secondary importance for
dimensional variations. A procedure to detect drift time in the case of a linearly drifted production process was
suggested by Fahmy and Elsayed [6]. Based on the estimate of the drift time, they developed a new adjustment
procedure. Their method was found to be not effective for small trend rated processes. Jensen and Vardeman [7]
computed an optimal adjustment policy for a real machining process with deterministic process drift and random
adjustment error. Their policy was obtained by considering the cost of adjustment in addition to the error. Sullo and
Vandeven [8] developed an optimal adjustment strategy for a process with run to run variation in a machining
scenario. Serious consideration of sampling and adjustment costs relative to the cost of inadequate quality was the
major motive for their strategy. Liang et al. [9] presented an integrated approach to cutting tool parameter selection
Data-based modeling and continuous adjustment of CNC turning process: A case study
and tool adjustment decisions for multi pass turning operation. Even though loss of quality was considered, they
gave more importance to machining cost, tool cost, tool replacement cost and adjustment cost.
Most existing work on adjustment of machining processes gives importance to other cost aspects rather than off
target cost. No efforts were found to model and adjust machining process based on sequentially collected data. There
is chance that statistical modeling of such process provides reasonably good results. While a number of theoretical
approaches have been proposed for adjustment, real studies on actual machines do not appear to have been reported.
In this paper, we carry out a turning operation on a CNC machine and measure the diameter of part produced.
Different statistical models to predict the process are compared. Thereafter, we have used the best model for
adjusting process in real time and the improvement in performance of the process is evaluated in terms of mean
squared deviation (MSD) from target.

2. Methodology
The role of adjustment cost and monitoring cost are neglected in this study. We assume that the off target cost is
highly significant. We assume that it is a responsive system in which the change in input is reflected at the output in
the next time period itself. Full effect of change in input is assumed to be obtaining at the output in the next time
instant. The problem under consideration here is to minimize deviation of measured diameter from target. The
process under control is modeled as 𝑦𝑦𝑡𝑡 = 𝛼𝛼 + 𝛽𝛽𝑥𝑥𝑡𝑡−1 + 𝛾𝛾𝑡𝑡 . Here 𝑦𝑦𝑡𝑡 is the response variable, 𝛼𝛼 is the offset of the
process from target, 𝛽𝛽 is the process gain (i.e, change in output for a unit change in input) and 𝛾𝛾𝑡𝑡 is the process
disturbance at time t. 𝑥𝑥𝑡𝑡−1 shows the level of control variable at the end of (𝑡𝑡 − 1)𝑡𝑡ℎ period. For a process on target,
offset is zero. For machining scenarios, value of process gain can be taken as unity. In such a situation, 𝑥𝑥𝑡𝑡−1 = (𝑇𝑇 −
𝑦𝑦�𝑡𝑡 ). Here 𝑇𝑇 is the target value of quality characteristic variable and 𝑦𝑦�𝑡𝑡 is its forecast value. Without loss of
generality, we can set the target as zero, implying that deviation from target is the concerned variable. In that case,
𝑥𝑥𝑡𝑡−1 = −𝑦𝑦�𝑡𝑡 , where 𝑦𝑦�𝑡𝑡 is the forecast value of deviation from target. For the purpose of modeling turning process, a
series of 100 numbers of diameter data collected from the unadjusted process is utilized. Modeling is done using
auto regressive integrated moving average (ARIMA) technique suggested by Box and Jenkins [10]. Different low
order candidate ARIMA models were fitted to the data set with the help of an R platform (R studio version-
0.98.1102). Suitability of various models in fitting the process is verified by checking the randomness and normality
of residuals. Here, relative goodness of fits of various ARIMA models is measured using Akaike Information
Criteria (AIC) [11] values.

3. Experimental details
The CNC lathe used for experimental study was a Jobber LM model machine. Mild steel was used as the work piece
and we used carbide cutting tool for machining. 200 mm long work piece was prepared through rough turning in an
ordinary center lathe. Initial diameter of the work piece was kept as 16 mm and target diameter as 15 mm.
Machining was carried out by supporting the work piece at both ends, in the presence of sulfur based cutting fluid.
Spindle speed was maintained at 500 rpm and feed 0.1mm/rev. A single pass turning operation was carried out. The
length of the cutting tool travel was 10 mm. For measurement of diameter, we used a digital micrometer of
resolution 0.001 mm. We carried out a repeatability study of the micrometer and verified that it is capable.
4. Modeling of CNC turning operation

335
 Proceedings of NCAMMM - 2018

Figure 1 shows various plots obtained from the unadjusted process. The mean level of deviation from target
(Fig.1(a) ) indicates an upward drift. For further analysis, the trend in the data series is removed by differencing the
series. The first differenced series (Fig.1(b) ) appears to be varying about a fixed mean, indicating stationarity. This
result is further verified using stationarity tests.

Fig. 1 Plots of unadjusted process. (a) deviation from target; (b) first difference of deviation; (c) ACF and (d) PACF

Box-Jenkins methodology suggests correlograms (ACF and PACF plots) for identifying candidate ARIMA
models. Horizontal dotted lines in these plots indicate 95% confidence interval. ACF plot (Fig.1(c) ) shows a
gradual decay. No significance is shown at any lag in the case of PACF plot (Fig.1 (d) ). Decaying nature of ACF is
an indication of presence of moving average (MA) components in the series. By observing these plots, we do not
anticipate the presence of any autoregressive (AR) terms in the series. Still during the preliminary model selection,
models with lower AIC values containing AR terms are also considered. Some probable low order ARIMA models
and their corresponding AIC values are shown in Table 1. Suitability of fitted models is verified through checking
randomness and normality of residuals. Here ARIMA (0,1,2) model (also called as IMA (1,2) model) is found to be
having least value of AIC, indicating that it is the best model. The estimated values of moving average parameters
𝜃𝜃1 and 𝜃𝜃2 for this model are 0.4132 and 0.2381 respectively.

Table 1 AIC values of candidate ARIMA models

Model ARIMA ARIMA ARIMA ARIMA ARIMA ARIMA ARIMA

(0,1,0) (1,1,0) (2,1,0) (3,1,0) (0,1,1) (0,1,2) (0,2,2)
AIC value -815.56 -819.34 -819.50 -820.41 -822.44 -825.89 -806.98

5. Forecasting equation for IMA (1,2) process

The forecasting equation corresponding to IMA (1,2) model is obtained as an extension of ordinary EWMA
equation as below.

The general equation of IMA (1,2) process is 𝑦𝑦𝑡𝑡 = 𝑦𝑦𝑡𝑡−1 + 𝑎𝑎𝑡𝑡 − 𝜃𝜃1 𝑎𝑎𝑡𝑡−1 − 𝜃𝜃2 𝑎𝑎𝑡𝑡−2 ..........................(1)

336
Data-based modeling and continuous adjustment of CNC turning process: A case study
Here 𝑦𝑦𝑡𝑡 is the quality characteristic variable and 𝑎𝑎𝑡𝑡 is a white noise with mean zero and fixed variance. Let 𝑦𝑦�𝑡𝑡 =
forecast of 𝑦𝑦𝑡𝑡 obtained at the end of (𝑡𝑡 − 1)𝑡𝑡ℎ period. In that case we can write 𝑦𝑦�𝑡𝑡+1 = 𝑦𝑦𝑡𝑡 + 𝑎𝑎𝑡𝑡+1 − 𝜃𝜃1 𝑎𝑎𝑡𝑡 − 𝜃𝜃2 𝑎𝑎𝑡𝑡−1 .
Here the best guess for 𝑎𝑎𝑡𝑡+1 is zero. But 𝑎𝑎𝑡𝑡 and 𝑎𝑎𝑡𝑡−1 can be taken as the forecast errors (𝑦𝑦𝑡𝑡 − 𝑦𝑦�𝑡𝑡 ) and (𝑦𝑦𝑡𝑡−1 −
𝑦𝑦�𝑡𝑡−1 ) respectively. Substituting these values in this equation, we get

𝑦𝑦�𝑡𝑡+1 = 𝑦𝑦𝑡𝑡 − 𝜃𝜃1 (𝑦𝑦𝑡𝑡 − 𝑦𝑦�𝑡𝑡 ) − 𝜃𝜃2 (𝑦𝑦𝑡𝑡 −1 − 𝑦𝑦�𝑡𝑡−1 ) or

𝑦𝑦�𝑡𝑡+1 = (1 − 𝜃𝜃1 )𝑦𝑦𝑡𝑡 + 𝜃𝜃1 𝑦𝑦�𝑡𝑡 − 𝜃𝜃2 (𝑦𝑦𝑡𝑡−1 − 𝑦𝑦�𝑡𝑡−1 )............................(2)

If 𝑦𝑦�𝑡𝑡 stands for the EWMA calculated at the end of period t, and 𝑦𝑦�𝑡𝑡 is used as the forecast for period (𝑡𝑡 + 1), we
can write 𝑦𝑦�𝑡𝑡 = (1 − 𝜃𝜃1 )𝑦𝑦𝑡𝑡 + 𝜃𝜃1 𝑦𝑦�𝑡𝑡−1 − 𝜃𝜃2 (𝑦𝑦𝑡𝑡 −1 − 𝑦𝑦�𝑡𝑡−2 ), which is an extension of the ordinary EWMA equation, that
provides minimum mean square error (MMSE) value for IMA (1,1) process [12]. For practical purposes, initial
values of 𝑦𝑦�𝑡𝑡−1 and 𝑦𝑦�𝑡𝑡−2 can be taken as the target value. Each new observation can lead to the new updated value of
EWMA.

6. Implementation of adjustments
For implementing the adjustment, we manipulated radial position of the cutting tool with the help of a wear offset
adjustment provision available in the CNC lathe. Forecasting equation for IMA (1,2) process is utilized for adjusting
position of cutting tool. Magnitude of cutting tool adjustment is taken as half of the predicted deviation of diameter.

Fig. 2. Results of adjustment. (a) Deviation from target; (b) Forecast error; (c) Incremental adjustment and (d)
Cumulative adjustment

Adjustment is made in opposite to the direction of deviation. Fig. 2 provides results of adjustment. Proper
monitoring of deviation from target (Fig. 2 (a) ) or forecast errors (Fig. 2 (b) ) can reveal the presence of assignable
causes in the process. Such information can be utilized for combining statistical and engineering process control
approaches (SPC and EPC). Here incremental adjustment (Fig. 2 (c) ) indicates the marginal tool advancement,
whereas cumulative adjustment (Fig. 2 (d) ) shows the level of cutting tool position with respect to the work piece.
Process adjustment using IMA (1, 2) model is found to be resulting in a reduction of up to 54% in MSD value.

337
 Proceedings of NCAMMM - 2018

Details of typical adjustment process are shown in Table 2. Equations used to calculate various results of
adjustments are shown below.

Deviation = (Observed diameter – Target diameter)

Forecast (EWMA) = 𝑦𝑦�𝑡𝑡 = (1 − 𝜃𝜃1 )𝑦𝑦𝑡𝑡 + 𝜃𝜃1 𝑦𝑦�𝑡𝑡−1 − 𝜃𝜃2 (𝑦𝑦𝑡𝑡−1 − 𝑦𝑦�𝑡𝑡−2 )
Adjustment = - 0.5 𝑦𝑦�𝑡𝑡
Cumulative Adjustment = (Previous Cumulative Adjustment+ Adjustment)
Forecast Error = ( 𝑦𝑦𝑡𝑡 − 𝑦𝑦�𝑡𝑡−1 )
Deviation in case of no adjustment = (Current deviation – 2* Previous Adjustment)

Table 2 Typical example for adjustment using IMA (1,2) model (T= 15 mm)
Time Diameter after Deviation Forecast of Adjustment Cumulative Forecast Deviation in
index adjustment from target deviation /mm Adjustment error /mm case of no
/mm /mm /mm /mm adjustment /mm
1 14.951 -0.049 -- -- -- -- --
2 14.931 -0.069 -- -- -- -- --
3 14.941 -0.059 -0.02295 0.011477 0.011477 -0.059 --
4 14.965 -0.035 -0.01359 0.006797 0.018274 -0.01205 -0.05795
5 14.967 -0.033 -0.0164 0.008199 0.026474 -0.01941 -0.04659
6 14.97 -0.03 -0.01928 0.009642 0.036115 -0.0136 -0.0464
7 14.978 -0.022 -0.01692 0.008462 0.044577 -0.00272 -0.04128
8 14.995 -0.005 -0.00738 0.003688 0.048265 0.011925 -0.02192
9 15.007 0.007 0.002268 -0.00113 0.047131 0.014376 -0.00038
10 14.996 -0.004 -0.00198 0.000988 0.048119 -0.00627 -0.00173

6. Conclusions
Different ARIMA models have been used to model a CNC turning process. IMA (1,2) model was found to be the
best among these models. Continuous adjustment was then implemented using this model, aiming to get a minimum
MSD value for the turned job. An extended forecasting equation of IMA (1,1) model was developed for
implementing adjustment based on IMA (1,2) model. This version of EWMA equation is simple and easy to apply,
provided the process parameters are known. Briefly, this work shows how to implement continuous adjustment
techniques based on the development of a data-based model for turning process. Some possible future works are
implementation of adjustments using other low order ARIMA models and exploring the feasibility of regression in
modeling and adjustment of turning process.

References

338
 Data-based modeling and continuous adjustment of CNC turning process: A case study
[1] Grubbs, Frank E. An Optimum Procedure for Setting Machines or Adjusting Processes. J qual Technol
[Internet]. 1983 Oct [cited 2017 Feb 21]; 15(4):186-189. Available from: http://asq.org/qic/display-
item/?item=5449
[2] Trietsch D. Process setup adjustment with quadratic loss. IIE Transactions [Internet]. 2000 Apr [cited 2017 Jun
19]; 32(4): 299-307. Available from: http://doi.org/10.1023/A:1007639319116
[3] Colosimo BM, Pan R, Del Castillo E. Setup adjustment for discrete-part manufacturing processes with
asymmetric cost functions. Int J Prod Res [Internet]. 2005 Mar [cited 2017 Feb 21]; 43(18): 3837-3854.
Available from: http://doi.org/10.1080/00207540500147158
[4] Pan R, Del Castillo E. Scheduling methods for the statistical setup adjustment problem. Int J Prod Res
[Internet]. 2003 [cited 2017 Jun 19]; 41(7): 1467-1481. Available from:
https://doi.org/10.1080/0020754031000065142
[5] Cai DQ, Xie M, Goh TN, Tang XY. Economic design of control chart for trended processes. International
Journal of Production Economics [Internet]. 2002 Sep [cited 2017 Feb 21]; 79(2): 85-92. Available from:
http://doi.org/10.1016/S0925-5273(01)00113-X
[6] Fahmy HM, Elsayed EA. Drift time detection and adjustment procedures for processes subject to linear trend.
Int J Prod Res [Internet]. 2006 [cited 2017 Mar 23]; 44(16): 3257-3278. Available from:
http://doi.org/10.1080/00207540500410242
[7] Jensen KL, Vardeman SB. Optimal adjustment in the presence of process drift and random adjustment error.
Technometrics [Internet]. 1993 Nov [cited 2017 Mar 23]; 35(4): 376-389. Available from:
http://www.jstor.org/stable/1270271
[8] Sullo P, Vandeven M. Optimal adjustment strategies for a process with run-to-run variation and 0-1 quality loss.
IIE Trans [Internet]. 1999 Dec [cited 2017 Feb 19]; 31(12): 1135-1145. Available from:
http://doi.org/10.1023/A:1007644129082
[9] Liang M, Mgwatu M, Zuo M. Integration of cutting parameter selection and tool adjustment decisions for multi
pass turning, Int J Adv Manuf Technol [Internet]. 2001 Jun [cited 2017 Feb 21]; 17(12): 861-869. Available
from: https://doi.org/10.1007/s001700170097
[10] Box G, Jenkins G. Some recent advances in forecasting and control. Applied Statistics [Internet]. 1968 [cited
2017 Feb 21]; 17(2): 91-109. Available from: http://www.jstor.org/stable/2985674
[11] Akaike H. Fitting autoregressions for predictions. Ann Inst Stat Math [Internet]. 1969 Dec [cited 2017 Mar 23];
21(1): 243-247. Available from: link.springer.com/content/pdf/10.1007/BF02532251.pdf
[12] Muth JF. (1960). Optimal properties of exponentially weighted forecasts of time series with permanent and
transitory components, Journal of American statistical association [Internet]. 1960 Jun [cited 2017 Jun 21]; 55:
299-306. Available from: https://www.jstor.org/stable/2281742

339
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

An Integrated Entropy-Combinative Distance-Based Assessment (CODAS) Method for

Aerospace Material Selection

1
Anirban Roy, 2Prasenjit Chatterjee, 3Shankar Chakraborty, 4Suprakash Mondal
1, 4
Department of Mechanical Engineering, Mallabhum Institute of Technology,
Bankura - 722122, West Bengal, INDIA. E-mail id: spmondal@gmail.com
2
Department of Mechanical Engineering, MCKV Institute of Engineering,
Howrah - 711204, West Bengal, INDIA.
3
Department of Production Engineering, Jadavpur University, Kolkata - 700032, West Bengal, INDIA.

Abstract: In recent days, for a specific part or product material selection is one of the difficult task for a decision
maker (DM) due to number of materials available in the market with their conflicting properties. New exotic
materials with impressive properties confused decision maker to select the most favorable material. In this research
study aerospace material selection is taken into consideration. Generally, highly reliable light weight materials like
aluminum, titanium and magnesium alloys are used as aerospace material. The main objective of this paper is to
search out the most appropriate aerospace body material among the large number of alternative material available
globally. An integrated entropy-COmbinative Distance-based ASsessment (CODAS) method is applied to search out
most feasible material for aerospace body material. A real time case study is developed with the help of materials
manual and different e-sources and subsequently solved the problem by the adopted integrated framework. It is
found that the Ti-10V-2FC-3Al alternative material emerged out as most potential material.
Keywords: CODAS Method, Entropy Method, Aerospace Material Selection

1. Introduction
Material selection with minimum cost and improved performance is a difficult task. In the design and improvement
of any structural elements, selection of material is one of the most emulating issues and it is also vital for the success
and to fulfill the demands of minimize cost and enhanced performance. Being the key role of engineering design
process, proper selection of materials can momentously minimize the manufacturing cost and enhance organization
competitiveness, customer gratification, and profitability. An improper material selection can adversely affect
productivity, profitability and brand value of the manufacturing organization [1]. Usually, expert personnel are
choosing a material by implementing the trial and error methods with involvement of huge cost or build on
gathering of past data leading to west of time. With increase in the availability of number of materials, exclusion
processes consume more time and become complex [2]. It is well known that material plays a critical role in
engineering design. Selection of material is one of the toughest problem in the design and prosperity of outcome,
and it is also troublesome for the advancement and competitiveness of the organizations. In engineering design
material selection needs clarity of the functional requirement for each and every part and various important criteria
also desires to be discriminated. Productivity, profitability and brand value of an organization can be affected by an
inaccurate selection of material, and it increases the producer responsibility. Corrosion and wear are the significant
An integrated entropy-combinative distance-based assessment (CODAS) method for aerospace material selection

factors that the surface of the engineering parts like aircraft body used in aerospace industry must envisage. By
taking a proper decision the need for protection and enhancement of the mechanical properties can be satisfied to
some extent. Aluminum alloys have its outstanding property [3] of mild weight, corrosion resistance for a structural
application in an aircraft parts. One of the best aluminum alloy is preferred for its great strength, tremendous low
temperature properties (does not reveal ductile to brittle transition and has impressive toughness even at cryogenic
temperatures close to absolute zero), weld-ability and capable to take smooth surface finish. No or little specific
machine ability data are available this aluminum alloys through conventional means such as turning, drilling milling,
and grinding. Titanium alloys are also used as a aircraft body material. Ti alloys have very high tensile strength and
toughness (even at extreme temperature).They are light in weight, have remarkable corrosion resistance and the
ability to withstand extreme temperatures. When a material is selected from a set of materials, a perfect
understanding of functional requirements for every component is necessary and various significant criteria need to
be considered. According to the nature of the operations, types of choices, servitude, among the attributes, types of
unreliability and expectations of DMs, quantity and quality of the accessible data and judgments the selection of
multi-criteria decision-making (MCDM) frame work or method should be done carefully.
It is well understood that the aerospace material selection is one of most sensitive issue. So a potential
MCDM framework can handle the material selection problem. Here, to determine criteria weight, entropy method is
extremely dependable for the information measurement and give high level accuracy. On the other hand, in CODAS
method, Euclidean and Taxicab distances from the negative-ideal position indicates the general performance of an
alternative material. The CODAS utilize the Euclidean distance as the preliminary measure of assessment. The
Taxicab distance is utilized for evaluation when Euclidean distances of two alternatives are very near to each other.
The threshold parameter is set according to the closeness of Euclidean distances.

2. Literature Review
A variety of material selection framework has been developed to support design engineer to select the appropriate
material for a specific engineering application and to improve the efficiency in product design and improvement
process. Debnath et al. [4] adopted Technique of Order Preference by Similarity to the Ideal Solution (TOPSIS) in
fuzzy environment to select the best aluminum alloy for aerospace and automotive industries. Yang et al. [5]
proposed Fuzzy TOPSIS method to improving the materials selection process in the field of automotive material
selection area. Nasab and Anvai,[6] applied Complex Proportional Assessment (COPRAS) and TOPSIS methods to
find out best material among the available alternative. Data envelope analysis (DEA) was also used as an MCDM
tool in material selection problem. It was established that DEA can be applied to handle problem by considering a
classical remark, but MCDM cannot be generally replaced by DEA. Maity & Chakraborty [7] proposed a framework
to select the best tool steel material based on preference ranking organization method for enrichment evaluation
(PROMETHEE II) method. AISI M2 and AISI T1 were the two most excellent choices and AISI 4140 was the less
favored tool steel material. Jahan and Edwards [8] was examined the limitations of normalization techniques and
proposed ways of improving their use in the engineering design decision-making process. Emphasis was placed on
materials selection application in aerospace and biomedical engineering. Patel and Prajapati [9] applied multi-
objective optimization on the basis of ratio analysis (MOORA) method to select the best blanking die material for

341
 Proceedings of NCAMMM - 2018

industrial application. Chatterjee and Chakraborty [10] focused on gear material selection problem with the help of
complex proportional assessment (COPRAS) and additive ratio assessment (ARAS)-based methods. Obtained result
was compared with the past researchers.Sharma et al. [11] applied fuzzy AHP and fuzzy TOPSIS method to select
the best material for an axle in motorcycle. Furthermore ranking comparison was performed between fuzzy AHP
and fuzzy TOPSIS.
Although number of research work has been proposed in the past on materials selection employing
different mathematical approaches (especially MCDM methods), any prior study has not demonstrated the
application of CODAS method, for resolving aerospace material selection problems.

3. Integrated Entropy-CODAS METHOD

Entropy is a term that measures the ambiguity related with indiscrimination incident of the anticipated information
content of a definite point and this indefinite is symbolized by a separate probability distribution. This Entropy
Method enumerates the weights of the criteria from the given decision matrix and it is free from the decision-makers
preference [12].
In CODAS method, the preferences of alternatives are estimated by adopting two measures. The primary
measure deals with Euclidean distance from the negative-ideal of alternatives by utilizing 1st -norm indifference
space for attributes. On the other hand secondary measure deals with Taxicab distance and employing 1st –norm
difference space. Hence it is found that, alternative having maximum distances from the negative ideal solution are
much admissible. If the Euclidean distance ties for two alternatives then Taxicab distance is utilized as secondary
measure [13]. The steps involved in the hybrid entropy-CODAS method are:
Step 1. Construction of the decision-making matrix (Y), as shown below:
 y11 y12 ... y1m 
y ... y2 m 
[ ]
Y = yij n× m
=  21
 
y22
   
 
 yn1 yn 2 ... ynm 

(1)
(
where yij yij ≥ 0 ) indicates the performance values of 𝑖𝑖th alternative of 𝑗𝑗th criterion (𝑖𝑖∈{1,2,3,…,𝑛𝑛} and
𝑗𝑗∈{1,2,3,…,𝑚𝑚}).
Step 2. Calculation of normalized decision matrix ( n ). Here linear normalization is used as follows:
ij

 yij

 max yij
nij = 
i

 min yij
 i
 yij
(2)
Step 3. Calculation of criteria weight by entropy method using the following sub stapes[12]
(a) The output entropy cj of the jth factor,
c j = K ∑ ( L ij ln L ij ) , ( 1 ≤ j ≤ m )
n
i =1

(3)
(b) Calculation of variation coefficient of jth factor hj
hj = 1 − c j , ( 1 ≤ j ≤ m )
(4)

342
An integrated entropy-combinative distance-based assessment (CODAS) method for aerospace material selection

(c) Calculation of weight of the entropy Ej

hj ,(1 ≤ j ≤ m)
Ej =
∑
n
i=1
hj
(5)
Step 4. Calculation of weighted normalized matrix as follows:
rij = w j ⋅ nij
(6)
where w j (0∠w j ∠1) indicates the relative importance of 𝑗𝑗th criterion, and ∑
m
w j = 1.
j =1

Step 5. Determination of the negative-ideal solution:

[ ]
ns = ns j 1× m
(7)
ns j = min rij
i

(8)
Step 6. Calculation of the Euclidean (Ei ) and Taxicab (Pi ) distances from the negative-ideal solution of the
alternatives:

∑ (r − ns j )
m
Ei =
2
j =1 ij

(9)
Pi = ∑ j =1 rij − ns j
m

(10)
Step 7. Construction of the relative assessment matrix:
Ra = [hik ]n×n
(11)
hik = (Ei − Ek ) + (Ψ (Ei − Ek ) × (Pi − Pk ))
(12)
where 𝑘𝑘∈{1,2,…,𝑛𝑛} and 𝜓𝜓 represent the threshold values to admit the impartiality of the Euclidean distances of two
alternatives, and is present below:
Ψ ( y ) = 1 if y ≥ τ
(13)
Ψ ( y ) = 0 if y ∠τ

The threshold parameter (𝜏𝜏) value ranging from 0.01 and 0.05 can be adopted by the decision maker. Here 𝜏𝜏=0.02 is
considered for the calculations
Step 8. Calculation of assessment score of every alternatives:
H i = ∑k =1 hik ,
n

(14)
Step 9. The highest Η𝑖𝑖 score alternative is the most excellent choice among the alternatives.

4. Illustrative Example
Selection of material for aircraft body parts is very difficult due to wide range of materials with their competent
attributes available globally. DM often face problem for conflicting mature of the criteria. In this developed case
study aerospace body material selection problem is considered. This aerospace material selection decision making
problem is developed using different manufactures catalog and different materials manual [14-17].

343
 Proceedings of NCAMMM - 2018

Table 1 Aerospace body materials selection decision matrix

Materials EM(Gpa) FT(Mpa/m) SS(Mpa) TS(Mpa) TC(W/mk) D(g/cc) C(Per kg)

Al-2014-T3 (A1) 71 25 290 480 120 3 300
Al-AA2017-T4 (A2) 72.4 26 262 427 134 2.79 300
Ti6Al4V Grade 5 (A3) 114 43 760 1170 6.7 4.43 1600
Ti-10V-2FC-3Al (A4) 107 44 669 1260 7.8 4.65 3000
Al-6061-T6 (A5) 68.9 96.5 207 310 167 2.7 250
Al-7075-T651 (A6) 71.7 27 330 573 130 2.81 501
Al-7070-T7452 (A7) 70 32 280 524 157 2.83 750
Al-5052 (A8) 70.3 120 138 228 140 2.68 250
Al-6063-T6 (A9) 68.9 68.9 150 240 200 2.7 220

Table 2 Normalized decision matrix

Materials EM FT SS TS TC D C
A1 0.6228 0.2083 0.3816 0.3810 0.6000 0.8933 0.7333
A2 0.6351 0.2167 0.3447 0.3389 0.6700 0.9606 0.7333
A3 1.0000 0.3583 1.0000 0.9286 0.0335 0.6050 0.1375
A4 0.9386 0.3667 0.8803 1.0000 0.0390 0.5763 0.0733
A5 0.6044 0.8042 0.2724 0.2460 0.8350 0.9926 0.8800
A6 0.6289 0.2250 0.4342 0.4548 0.6500 0.9537 0.4391
A7 0.6140 0.2667 0.3684 0.4159 0.7850 0.9470 0.2933
A8 0.6167 1.0000 0.1816 0.1810 0.7000 1.0000 0.8800
A9 0.6044 0.5742 0.1974 0.1905 1.0000 0.9926 1.0000
The developed material selection problem considered nine alternative and their seven conflicting selection criteria.
The alternatives are the alloys of aluminum and titanium such as Al-2014-T3, Al-AA2017-T4, Ti-6Al-4V, Ti- 10V-
2FC-3l, Al-6061-T6, Al-7075-T651, Al-7050-T7452, Al-5052, Al-6063-T6. The adopted alloys are good
mechanical properties like good Elastic Modulus , better Fracture Toughness, fair Shear Strength, enough Tensile
Strength , high thermal conductivity, low Density and low Cost per kg. Considered criteria’s has a great influence on
the aircraft body parts material selection problem. Among the considered attributes Elastic Modulus (EM), Fracture
Toughness (FT), Shear Strength (SS), Tensile Strength (TS) and thermal conductivity (TC) are beneficial in nature
rest, density (D) and cost (C) are non beneficial in nature whose lower values are desirable.

Table 1 represent aerospace material selection decision making problem which consists of nine candidate
alternatives and seven selection criteria and Table 2 depicts normalized values of the problem which are calculated
using Eq. (2). Weights of criteria are calculated using Eqs. (3-5) of entropy method and results are depicted in Table
4. The negative ideal solution is estimated by utilizing the weighted normalized values. From the obtained values the
Euclidean and Taxicab distances from negative-ideal solution of alternative materials are also calculated. Table 3
present the relative assessment matrix (𝑅𝑅𝑎𝑎) and assessment scores (Η𝑖𝑖) which are computed by utilizing Eqs. (11) to
(14). The calculations are executed by considering 𝜏𝜏=0.02.

344
An integrated entropy-combinative distance-based assessment (CODAS) method for aerospace material selection

Table 3 Relative assessment matrix and assessment score of the alternatives

A1 A2 A3 A4 A5 A6 A7 A8 A9 Hi
A1 0.0000 -0.0032 -0.2945 -0.3048 -0.1023 -0.0150 -0.1746 -0.0770 -0.0575 -1.0289
A2 0.0032 0.0000 -0.2894 -0.2997 -0.0189 -0.0118 -0.1696 -0.0719 -0.0524 -0.9104
A3 0.2945 0.2894 0.0000 -0.0031 0.1922 0.1835 0.0195 0.2175 0.2370 1.4304
A4 0.3048 0.2997 0.0031 0.0000 0.2026 0.1938 0.1302 0.2279 0.2474 1.6094
A5 0.1023 0.0189 -0.1922 -0.2026 0.0000 0.0071 -0.0083 0.0253 0.0448 -0.2048
A6 0.0150 0.0118 -0.1835 -0.1938 -0.0071 0.0000 -0.0154 0.0341 0.0535 -0.2853
A7 0.1746 0.1696 -0.0195 -0.1302 0.0083 0.0154 0.0000 0.0977 0.1172 0.4331
A8 0.0770 0.0170 -0.2175 -0.2279 -0.0019 0.0052 -0.0102 0.0000 0.0195 -0.3387
A9 0.0575 0.0524 -0.2370 -0.2474 0.0066 0.0137 -0.0017 -0.0195 0.0000 -0.3753

Table 4 Criteria weights

Criteria EM FT SS TS TC D C
Weight 0.0218 0.1747 0.1746 0.1838 0.2220 0.0185 0.2046

Table 5 Ranking of the alternatives

Materials A1 A2 A3 A4 A5 A6 A7 A8 A9
Rank 9 8 2 1 4 5 3 6 7

The ranking of alternative materials are achieved according to the assessment scores which is displayed in
Table 5. It is found that 4 (Ti-10V-2FC-3Al) is the best and 𝐴𝐴1 (Al-2014-T3) is the worst aerospace body material
according to the estimation of the entropy based CODAS method.

5. Discussion
The criteria weights are the most important parameter for the DM for accessing the most excellent alternative. The
first phase of the framework deals with entropy method to determine the criteria weight corresponds to each
objective. Based on the criteria weight, CODAS method is applied to ranking preorders of the considered materials
are obtained. Generally in CODAS method, Euclidean distance and Taxicab distance are used to evaluate the
alternative materials. The Euclidean and Taxicab distance are estimated with the help of negative-ideal solution.
Hence, the larger distance achieved by the alternative get more preference. On the other hand, in the CODAS
method, the primary measure is Euclidean distance and secondary measure is Taxicab distance. One developed case
study is utilized to demonstrate the CODAS method.

6. Conclusions
The integrated entropy-CODAS method is presented for selection of aerospace body material. Here obtained results
says that 𝐴𝐴4 (Ti-10V-2FC-3Al) is the most feasible choices. The appropriate material selection place a major roll for
minimizing the corrosion and failures of the body parts in aerospace industry. This novel integrated MCDM based
method has the ability for other material selection problems in the aerospace industry. Hence, the above mentioned
method can apply for other parts of the aerospace industry like cockpit, fuselage, winglet, slats, spoiler, elevator etc.
This potential MCDM based framework may useful for the researcher on the aerospace material selection problems.

345
 Proceedings of NCAMMM - 2018

References

[1] Chatterjee P., Athawale V. J., & Chakraborty S.,Materials selection using complex proportional assessment
and evaluation of mixed data methods. Materials & Design, 2011; 32(2); 851-860.
[2] Anojkumar L, Ilangkumaran M, Sasirekha V. Comparative analysis of MCDM methods for pipe material
selection in sugar industry. Expert Systems with Applications. 2014; 41(6):2964-80.
[3] Dursun T, Soutis C. Recent developments in advanced aircraft aluminium alloys. Materials & Design. 2014;
56:862-71.
[4] Bhattacharjee P, Debnath A, Chakraborty S, Mandal UK. Selection of optimal aluminum alloy using TOPSIS
method under fuzzy environment. Journal of Intelligent & Fuzzy Systems. 2017; 32(1):871-6.
[5] Yang SS, Nasr N, Ong SK, Nee AY. Designing automotive products for remanufacturing from material
selection perspective. Journal of Cleaner Production. 2017; 153(1):570-579
[6] Mousavi-Nasab SH, Sotoudeh-Anvari A. A comprehensive MCDM-based approach using TOPSIS,
COPRAS and DEA as an auxiliary tool for material selection problems. Materials & Design. 2017;
121:237-53.
[7] Maity SR, Chakraborty S. Tool steel material selection using PROMETHEE II method. The International
Journal of Advanced Manufacturing Technology. 2015;78(9-12):1537-47.
[8] Jahan A, Edwards KL. A state-of-the-art survey on the influence of normalization techniques in ranking:
Improving the materials selection process in engineering design. Materials & Design (1980-2015). 2015;
65:335-42.
[9] Patel S.S., & Prajapati J. M., Multi-criteria Decision Making Approach: selection of Blanking Die Material.
International Journal of Engineering, 2017; 30(5):800-806.
[10] Chatterjee P, Chakraborty S. Gear material selection using complex proportional assessment and additive
ratio assessment-based approaches: a comparative study. International Journal of Materials Science and
Engineering. 2013;1(2):104-111.
[11] Sharma A, Sharma A, Sachdeva A. Selection of the Best Material for an Axle in Motorcycle using fuzzy
AHP and Fuzzy TOPSIS Methods. MIT International Journal of Mechanical Engineering, 2014; 4(1):29-36.
[12] Dashore K, Pawar SS, Sohani N, Verma DS. Product evaluation using entropy and multi criteria decision
making methods. International Journal of Engineering Trends and Technology (IJETT). 2013;4:2183-7.
[13] Keshavarz Ghorabaee M, Zavadskas EK, Turskis Z, Antucheviciene J. A New Combinative Distance-based
Assessment (Codas) Method For Multi-Criteria Decision-Making. Economic Computation & Economic
Cybernetics Studies & Research. 2016; 50(3).
[14] www.azom.com, date of accession 10-12-2017.
[15] www.matweb.com, date of accession 12-12-2017.
[16] www.makeitfrom.com, date of accession 15-12-2017.
[17] www.kyocera-sgstool.eu, date of accession 13-12-2017.

346
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Regression Model Formulation for Prediction of Clad Layer Characteristics in an In-

House Built Coaxial Nozzle based DMD System

Anirban Changdar, Piyush Pant*, A.K.Lohar

Advanced Manufacturing Centre, CSIR-Central Mechanical Engineering Research Institute,
Durgapur, India, *piyupant@gmail.com

Abstract: Direct metal deposition is a blown powder based additive manufacturing technology, in which powder
particles are fused using a high energy laser source. In the present work a regression model using response
surface method has been developed considering laser power, scan speed and powder flow rate as the process
parameters. Clad layer height and capture efficiency were considered as the response. The study also
investigates deposition for a positive laser defocus condition which demonstrates a better surface quality and
maximizes the material deposition efficiency. The developed regression model provides useful information to
control responses and ensure building a desired clad layer height as per prototyping requirements.
Keywords: RSM, layer height, direct metal deposition, capture efficiency

1. Introduction
In direct metal deposition process the thermal energy of the laser beam is used to melt the substrate and build
the solid metal part by injecting metal powder into a laser generated molten zone. Synchronized movement of
the laser source and powder injection mechanism completes the layers in order to create a part. The most
influential process parameters to be taken into consideration for optimal height per layer are laser power,
scanning speed and powder mass flow rate. A number of studies have investigated and developed relationships
between the process variables and the geometrical, quality characteristics of the clad on different materials using
statistical and analytical models.
Choi and Chang [1] employed statistical techniques to develop the empirical relations for deposited
layer thickness, micro hardness, and porosity for the deposition of H13 steel using a CO2 laser. Layer height
influences the microstructure and mechanical properties, as well as the geometric accuracy, of LBAM-fabricated
parts[2]. Lee [3] used the Taguchi technique to achieve a maximum powder catchment efficiency during the
LBPD of a Co-alloy powder on a steel substrate using a low power pulsed Nd:YAG laser. A maximum powder
catchment efficiency of 12.3% was reported for 0.62 g/s of the powder feed rate, and 6.7 mm/s of the scanning
speed. Balu et.al [4] performed an multi-objective optimization using central composite based response surface
method to develop an empirical relationship between the input variables and quality characteristics. In their
study, the influence of laser power, scanning speed, powder flow rate, and percentage overlap on uniformity
index, dilution, and micro-hardness is determined. Onwubolu et al. [5] utilized response surface methodology
(RSM) to optimize a single-track Diamalloy deposit based on the clad angle: a geometrical feature obtained
from the cross section of the deposited material. A laser power of 2 kW, scanning speed of 5mm/s, and powder
mass flow rate of 0.067 g/s, had been reported as an optimal value to achieve a geometrically perfect clad.
Mahmood et.al [6] investigates the role of processing parameters: laser power, scanning speed, powder flow rate
and gas flow rate, on the material utilization efficiency in laser metal deposited Ti6Al4V. A two-level full
factorial design of experiment was used in their investigation. Process of direct metal deposition involves
Regression model formulation for prediction of clad layer characteristics in an in-house built coaxial nozzle based DMD system

material wastage. To minimize material wastage in DMD process, and to enhance the economy of the process
qualitative analysis were performed by researchers [7].
The aim of the present study is to develop a regression model based on response surface method for an
in-house developed direct metal deposition system. In this work laser is defocused in order to maximize the
powder catchment by increasing the laser spot diameter at the surface of substrate. The developed model
successfully predicts the clad layer height and capture efficiency within the defined range of process parameters.
Later, a multi layer wall deposition is done at optimum value of process parameters.

2. Experimental Setup
The experiments were performed on a Multi Material Deposition (MMD) system, designed and developed in
CSIR-CMERI, Durgapur shown in Fig. 1. The high-power diode laser has 1.2kW maximum power output with
wavelength of 980 nm ± 10 nm. The Z axis of the CNC stage delivers the fibre-guided laser beam and metal
powders coaxially blown through a nozzle to the laser beam spot. 316 L stainless steel powder of 45 µm particle
size was used in the experiment. The powder was preheated to remove moisture for achieving smooth flow. The
nozzle standoff distance is maintained at 10 mm from the mild steel substrate.

Fig. 1: Experimental setup of MMD system Fig. 2 Schematic of layer height measurement method

Measurement procedure - The clad layer height is measured using image processing of a layer cross-section as
shown in Fig. 2. For the measurement of capture efficiency the following relation has been used [6]. The amount
of powder sent to the deposition zone is measured using the powder mass flow rate relation.

𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑓𝑓𝑓𝑓𝑓𝑓 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑋𝑋 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 ℎ 𝑜𝑜𝑜𝑜 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

Mass of powder sent to the deposition region (Wp) = (g)
𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

Mass of substrate before deposition = Ws (g)

Mass of the substrate after deposition = Ws' (g)
Mass of the powder deposited on the substrate (Wf) = Ws'-Ws (g)
Catchment efficiency (η %) = [(Wf)/Wp]*100

3. Design of Experiments
Experiments were conducted according to central composite design which is best suited for Response Surface
Methodology (RSM) application. This method enables lesser number of experiments with better accuracy [8].
20 experimental runs were done by considering three factors and their corresponding two levels. Table 1 enlist
the considered parameters and their corresponding levels in which the experimentations were done.

348
 Proceedings of NCAMMM - 2018

Table 1: Factors and their levels

Level 1 Level 2
Symbol Control Factors (-α) Nominal (+α)
P Power (W) 300 450 600
Scan Speed
V 4 7 10
(mm/sec)
Powder flow rate
F 1.05 2.35 3.5
(gm/min)

Fig. 3 Single layer clad deposition

4. Results and Discussion
Central composite design of experiments was developed to perform the experimentation and for RSM analysis.
A single-track laser deposition was done based on the defined levels of input parameters .Twenty such runs
were made at various combination of parameters. The measured responses against the defined input variables
are represented in Table 2. Two iterations for each experiment run has been performed to minimize the
experimental error. Statistical regression model against each output variable were established by following
Response surface method utilising Minitab Software version 17 [9].

Table 2 Results of RSM based experiments

Run No. Power Scan speed Powder flow Height achieved Capture efficiency
(W) (mm/s) rate (g/min) (mm) (%)
1 300 4 3.5 0.459 13.99
2 450 7 3.5 0.321 30.17
3 300 10 1.05 0.090 21.84
4 600 4 3.5 0.550 39.68
5 300 7 2.35 0.133 11.74
6 450 7 2.35 0.243 17.94
7 300 4 1.05 0.161 10.94
8 450 7 2.35 0.245 17.89
9 600 10 1.05 0.092 12.84
10 450 7 1.05 0.155 21.68
11 450 10 2.35 0.199 10.95
12 300 10 3.5 0.149 10.68
13 600 7 2.35 0.271 17.55
14 450 4 2.35 0.364 17.10
15 450 7 2.35 0.240 17.80
16 600 4 1.05 0.249 18.08
17 450 7 2.35 0.243 20.57
18 450 7 2.35 0.250 18.54
19 600 10 3.5 0.299 23.46
20 450 7 2.35 0.243 18.36

4.1 ANOVA Analysis and development of response surface (RS) regression model
Analysis of Variance (ANOVA) has to be done in order to develop a Response Surface Model as well as for
sorting out significant input parameters and interactions. In this case significant terms are those which have p-
value less than 0.05 because the confidence level for the present model is 95% [10]. The results of ANOVA

349
Regression model formulation for prediction of clad layer characteristics in an in-house built coaxial nozzle based DMD system

against the defined responses are shown in Tables 3 and 4.The R2 value is used to analyze the fit between the
model predicted values and the experimentally obtained values. Higher R2 values indicates better correlation
between them. In this case R2 values are high enough which indicate a good fit between the predicted and
measured data. Suitable response models were established as per the ANOVA analysis are shown below in the
equations (1) and (2).

Table. 3 Analysis of Variance for Capture efficiency

CF DF Adj. SS Adj. MS F P
Model 9 893.521 99.280 52.79 0.000
Linear 3 579.050 193.017 102.63 0.000
P 1 255.880 255.880 136.05 0.000
V 1 64.033 64.033 34.05 0.000
F 1 259.137 259.137 137.78 0.000
Square 3 189.817 63.272 33.64 0.000
2
P 1 31.235 31.235 16.61 0.002
2
V 1 43.804 43.804 23.29 0.001
2
F 1 183.246 183.246 97.43 0.000
2-way interaction 3 386.657 128.886 68.53 0.000
PxV 1 105.448 105.448 56.07 0.000
PxF 1 200.945 200.945 106.84 0.000
VxF 1 80.263 80.263 42.68 0.000
Err 10 18.808 1.881
Lack-of-Fit 5 13.336 2.667 2.44 0.175
Pure Error 5 5.472 1.094
Total 19 912.328
S = 1.37141 R-Sq. = 97.94% R-Sq.(adj) = 96.08% R-Sq. (pred)=85.93%

η = -32.88 + 0.1589 P + 11.09 V - 27.19 F - 0.000150 P2 - 0.4435 V2 + 5.263 F2- 0.00807 P*V+ 0.02671 P*F -
0.844 V*F --- (1)

Table 4. Analysis of Variance for Layer Height

CF DF Adj. SS Adj. MS F P
Model 7 0.242461 0.034637 69.59 0.000
Linear 3 0.231821 0.077274 155.25 0.000
P 1 0.024353 0.024353 48.93 0.000
V 1 0.101655 0.101655 204.23 0.000
F 1 0.105813 0.105813 212.58 0.000
Square 2 0.007983 0.003992 8.02 0.006
P2 1 0.002793 0.002793 5.61 0.035
2
V 1 0.007982 0.007982 16.04 0.002
2-way interaction 2 0.016765 0.008382 16.84 0.000
PxF 1 0.003017 0.003017 6.06 0.030

350
 Proceedings of NCAMMM - 2018

VxF 1 0.013748 0.013748 27.62 0.000

Err 12 0.005973 0.000498
Lack-of-Fit 7 0.005970 0.000853 1279.22 0.000
Pure Error 5 0.000003 0.000001
Total 19 0.248434
S = 0.0223103 R-Sq. = 97.60% R-Sq.(adj) = 96.19% R-Sq. (pred)= 85.41%

H = 0.067 + 0.001259 P -0.0844 V + 0.1130 F - 0.000001 P2 + 0.00555 V2 + 0.000104 P*F -0.01105 V*F --(2)

The positive and negative signs before every term in the equations describe the kind of effect the it has on the
responses. The equations should be scrutinized to check the accuracy of the developed model [11].

4.2 Validation of the model

Figure 3a and 3b shows normal probability plots of the developed models. The plots shows that the residuals
tends follow the linear prediction of the model hence it is quite acceptable.

(a) (b)
Figure 3. (a) Normal probability plot for capture efficiency, (b) Normal probability plot for height

Distribution of the residuals distribution along the straight diagonal line verifies that errors of the model are
distributed normally and the developed statistical relationship is done properly..

4.3 The effect of process parameters on the responses

To understand the interaction of the process parameters with the responses contour plots were generated using
statistical tool Minitab as shown in figure 4a and 4b .It is clear from the plots that both the responses maximize
with increase in power and powder flow rate and varies inversely with the scan speed. Although the disparity in
clad height is more noticeable by the change in scanning speed than laser power and powder flow rate, whereas
the disparity of capture efficiency is more pronounced by changing powder flow rate and power.

4.4 Optimization of the input variables and confirmation

For pursuing optimization of multiple responses with response surface method, the desirability function is used.
In Minitab the responses are optimized using response optimizer. The optimization is done by keeping in mind
the limitations of the system at the same time maintaining a good desirability function. Desirability function is
used for determining quality of the output responses yearned for. The value of the function is bound between 0
and 1. Thus the decided optimized factors along with the desirability factor are given in Table 5.

351
Regression model formulation for prediction of clad layer characteristics in an in-house built coaxial nozzle based DMD system

(a) (b)

Figure 4. Contour plots of a). capture efficiency b). layer height

Table 5. Optimization results with desirability

Desirability P (W) V (mm/s) F(g/min)
factor
0.5041 600 4 2.35

Finally, confirmatory experiments were done to verify the optimised results of the model. Iterative experiments
were conducted at the optimal values. The average values of the characteristics were obtained using
confirmatory experiments and compared with the predicted values. The results are given in Table 6. Based on
the optimum values a ten layer wall is built which is shown in figure 5. The achieved wall height for the ten
layer wall deposit was measured to be 4.010 mm.
Table 6. Confirmatory tests results
Response Clad height Capture efficiency
Predicted optimal value 0.41mm 21.36
Confidence intervals 0.38mm ≤ 0.41≤0.43mm 18.77≤ 21.36≤ 23.94
Result from confirmation 0.39mm 19.20
experiment.

Figure 5. Clad wall built at the optimum value of process parameters

352
 Proceedings of NCAMMM - 2018

5. Conclusion
In the present work a regression model using response surface method has been developed considering laser
power, scan speed and powder flow rate as the process parameters. Clad layer height and capture efficiency
were considered as the responses. The study were performed for a positive laser defocus condition which
demonstrate a better surface quality and maximizes the material deposition efficiency. The developed regression
model provides useful information to control responses and ensure building a desired clad layer height as per
prototyping requirements. The following can be concluded from the work:
• The developed model was found to be in good agreement with the experimental data and is within 5 % error.
• Study predicts that both the responses varies directly with the powder and powder flow rate, and inversely
with the scan speed.
• The optimum values were obtained at 600 W power, 4 mm/s scanning speed and 2.35 g/min powder mass
flow rate level.
• The measured responses obtained at the optimal level of the input factors were well within the confidence
interval of the developed model.
• Based on an optimum values a clad wall is built.

The model developed in the present study can be used to minimize the powder wastage while development of
functional component. Future research work will carry the influence of these processing parameters considering
the thermal aspects during deposition for a laser defocused condition.

References
[1] Choi J, Chang Y. Characteristics of laser aided direct metal/material deposition process for tool steel.
International Journal of Machine Tools and Manufacture. 2005 Apr 30;45(4):597-607.
[2] Shamsaei N, Yadollahi A, Bian L, Thompson SM. An overview of Direct Laser Deposition for additive
manufacturing; Part II: Mechanical behavior, process parameter optimization and control. Additive
Manufacturing. 2015 Oct 31;8:12-35.
[3] Lee HK. Effects of the cladding parameters on the deposition efficiency in pulsed Nd: YAG laser
cladding. Journal of materials processing technology. 2008 Jun 20;202(1):321-7.
[4] Balu P, Leggett P, Hamid S, Kovacevic R. Multi-response optimization of laser-based powder deposition
of multi-track single layer hastelloy C-276. Materials and Manufacturing Processes. 2013 Feb
1;28(2):173-82.
[5] Onwubolu GC, Davim JP, Oliveira C, Cardoso A. Prediction of clad angle in laser cladding by powder
using response surface methodology and scatter search. Optics & Laser Technology. 2007 Sep
30;39(6):1130-4.
[6] Mahamood RM, Akinlabi ET. Processing parameters optimization for material deposition efficiency in
laser metal deposited titanium alloy. Lasers in Manufacturing and Materials Processing. 2016 Mar
1;3(1):9-21.
[7] Akinlabi ET, Mahamood RM, Shukla M, Pityana S. Effect of scanning speed on material efficiency of
laser metal deposited Ti6Al4V. oral presentation at the World Academy of Science, Engineering and
Technology (WASET 2012). 2012 Nov 1:28-.
[8] Montgomery DC. Design and analysis of experiments. John Wiley & Sons; 2017.
[9] Minitab 17 Statistical Software (2010). [Computer software]. State College, PA: Minitab, Inc.

353
Regression model formulation for prediction of clad layer characteristics in an in-house built coaxial nozzle based DMD system

[10] Acherjee B, Kuar AS, Mitra S, Misra D. A sequentially integrated multi-criteria optimization approach
applied to laser transmission weld quality enhancement—a case study. The International Journal of
Advanced Manufacturing Technology. 2013 Mar 1;65(5-8):641-50.
[11] Farahmand P, Kovacevic R. Parametric study and multi-criteria optimization in laser cladding by a high
power direct diode laser. Lasers in Manufacturing and Materials Processing. 2014 Dec 1;1(1-4):1-20.

354
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Optimization of Micro Hardness and Fracture Toughness of Zirconia Toughened

Alumina (ZTA) under Different Compacting Pressures and Sintering Temperatures
using Response Surface Methodology (RSM)

Subhrojyoti Mazumder1, Kunal Ghosh1, Himadri Roy2, Nilrudra Mandal1*

1
Centre for Advanced Materials Processing, 2NDT & Metallurgy Group
CSIR-Central Mechanical Engineering Research Institute, MG Avenue, Durgapur-713209
*
Corresponding Author: n_mandal@cmeri.res.in

Abstract: Compaction pressure and sintering temperature are two most important factors considered in powder
metallurgy process which affect the performance of any developed product. Visualizing the effect of these two
parameters and optimizing of these twoon mechanical properties i.e. micro hardness and fracture toughness of
zirconia toughened alumina (ZTA) ceramics is extremely important towards developing ideal ceramics which
can be used for structural application like cutting inserts, sealing materials etc. In this work, ZTA ceramics have
been compacted at different pressure (6, 7 and 8 ton/cm2) followed by sintering at the varioustemperatures1550,
1600 and 1650oC for evaluation of micro hardness and fracture toughness. The micro structural
characterization has been carried out by means of Field Emission Scanning Electron Microscopy (FESEM).
The Energy Dispersive Spectroscopy (EDS) ensured the weight percentage of alumina (~87%) and zirconia
(~13%) into the composite matrix. The tetragonal phase of the developed ceramics has been identified by X-ray
diffraction method (XRD). From the analysis, it can be predicted that the grain growth and tetragonality factor
are the two prime responsible phenomena for variation of micro hardness and fracture toughness of the
developed ceramics. The optimized compaction pressure and sintering temperature towards maximum micro
hardness and fracture toughness have been found out as 6ton/cm2& 1596.89OC respectively with 90.60%
desirability level using Response Surface Methodology (RSM).
Keywords: Micro hardness, fracture toughness, compaction pressure, sintering temperature, RSM.

1. Introduction
In current manufacturing era, extensive research is going on towards advancement ofceramic materials which
can possess excellent material properties.In this direction, researchers are constantly trying to optimize the
processing parameters towards development of good product. In that aspect of research, enhancement of the
mechanical characteristics of materials is of great concern in order to make them suitable for many industrial
applications. For example, ceramiccomponents show the exceptional performance while they are being used as
cutting tool material for machining operationsbecause of their promising mechanical properties viz., high
hardness, enough fracture toughness, flexural strength etc [1-4]. The material process parameters are
dynamically correlated to influence the desired properties of the materials. It has already proven that the
sintering temperature dominants the grain size, phase composition which eventually influences the mechanical
properties of the materials [5]. Rodrigo A. Barbieri et al. have reported that the characteristics strength of
alumina spring can be increased by 32% with increase of sintering temperature from 1550OC to 1650OC due to
overall reduction in porosity [6]. Moreover, the inter-particle voids can be filled by incorporating the liquid
phase during sintering resulting improvement in densification [7]. Beside the sintering temperature, compaction
pressure has the direct impact on porosity [8]. Zamri Yusoff et al. have described a linear correlation between
Optimization of micro hardness and facture toughness of zirconia toughened alumina (ZTA) under different compacting
pressures and sintering temperatures using Response Surface Methodology (RMS)

compaction pressure and microhardness by using the ANOVA with statistical model in the aluminium
composite. Compaction pressure has been reported as the function of apparent porosity since the porosity can be
minimized significantly with the increase of compaction pressure from 150 to 300 MPa [9]. In current scenario
of research, it is more important to optimize the process parameters to achieve the highest possible level of
productivity. Many researchers have casted off the response surface methodology (RSM) as a potential
statistical tool to optimize their input variables so as to get the desired response as outcome [10-13]. So,
keeping all these points in mind, an effort has been made to evaluate the combined effect of a range of process
parameters viz., compaction pressures and sintering temperatures into the commercially available zirconia
toughened alumina (ZTA) powder. In this present study, the micro hardness and fracture toughness have been
calculated and a statistical model have been developed using response surface methodology (RSM) by
considering these result values as response factors. The optimization of the process parameters have
beendonewith the help of desirability approach.

2. Experimental details
2.1 Preparation of Materials
Commercially available ZTA powder (Zirox Technologies, India) is processed as initial powder. The powder is
at firstmilled in planetary ball mill (FRITSCH, Pulverisette 5) using alumina balls (10 mm diameter) with
acetone as medium for 24 hours in order to get fine particle size. The ball milled slurry is dried in oven at 100oC
for 8 hours. The dried cake is then grounded in mortar pestle to get the final powdered sample. The average
particle size distribution have been calculated using dynamic light scattering (DLS) method with the help of
Malvern Zetasizer instrument and is presented in the Figure 1 which indicated an average particle size of 0.7µm.

Figure 1. Particle size distribution obtained from dynamic light scattering method

The end powder is compacted into circular pallet of diameter 10 mm and thickness 3.5 mm by uniaxial press
(CARVER, Indiana USA) at three pressure levels 6, 7 and 8 ton/cm2. The compacted palletsare sintered in high
temperature furnace at the temperatures 1550, 1600 and 1650⁰C with dwell time 1 hour. The density of sintered
sample is measured with the help of Archimedes principle. The sintered samples are first polished in silicon
carbide plate followed by polished in Bain Polisher using diamond paste (0.5-1micron). The polished samples
areultrasonically cleaned in acetone using bath ultrasonicator. The average roughness values have beenmeasured
as 0.1µm using surface finish analyser (Surtronic Duo, Taylor and Hobson) for all the finished samples.

2.2 Mechanical characterization

356
 Proceedings of NCAMMM - 2018

The mechanical properties viz., micro hardness (HV0.5) and fracture toughness (K IC ) of the samples have been
carried out in micro hardness tester (INNOVATEST, Falcon500). Dwell time for both hardness and fracture
toughness have been set as 10 sec corresponding to the loads 4.9N and 24.5N respectively. At least five
indentations are taken at different places on the each pallet to make sure the repeatability of the results for each
hardness and fracture toughness data. An average value has been reported.Fracture toughness is calculated from
the Evans and Charles formula [13],

K IC = 0.16 (1)

where, K IC = Fracture toughness (MPa.m0.5), c = Average length of the cracks measured in the tip of the Vickers
mark (micron), a = Half average length of the diagonal of the Vickers mark (micron),H = Vickers hardness
(MPa). The details of the process parameters are given in the Table 1.

Table 1. Details of the process parameters, bulk densities, volumetric shrinkage and mechanical
behaviour of all samples

Sl Sample Process parameters BulkDen Volumetric Hardness Fracture

No. Designation Compaction Sintering sity shrinkage (GPa) Toughness
pressure Temperature (g/cm3) (%) (MPa.m0.5)
(ton/cm2) (oC)
1. Z 6 -1550 6 1550 3.233 39.65 18.38±0.29 5.93±0.39
2. Z 7 -1550 7 1550 3.242 40.50 19.03±0.32 5.39±0.80
3. Z 8 -1550 8 1550 3.239 40.03 19.40±0.09 4.77±0.60
4. Z 6 -1600 6 1600 3.210 38.38 19.36±0.25 6.49±0.55
5. Z 7 -1600 7 1600 3.254 38.56 19.42±0.29 5.99±0.62
6. Z 8 -1600 8 1600 3.245 38.75 19.45±0.40 5.83±0.31
7. Z 6 -1650 6 1650 3.248 39.23 16.93±0.38 6.41±0.26
8. Z 7 -1650 7 1650 3.248 39.65 17.86±0.43 5.79±0.69
9. Z 8 -1650 8 1650 3.244 39.65 18.35±0.53 5.91±0.41

3. Statistical modelling
3.1 Response surface methodology
RSM is a mathematical as well a statistical tool which is used to develop the model holding several variables
and to analyse the response of interest. It is a well-known tool to optimize the responses of complex
mathematical models as well [11]. In this study, compaction pressure and sintering temperature have been
chosen as process parameters and micro hardness and fracture toughness have been selected as the response
factors. Central composite design (CCD) is used forexperiments. The performance test indicated 9 trials and the
subsequent results of the trials are depicted in the Table 1.
In this study, the response (Y) is the function of micro hardness (H) and fracture toughness (F) and that can be
expressed as Y = ƒ(H,F). (2)

The second order regression equation which is used to satisfy the responses for ‘n’ number of factors can be
given as

Y = A0 + (3)

357
Optimization of micro hardness and facture toughness of zirconia toughened alumina (ZTA) under different compacting
pressures and sintering temperatures using Response Surface Methodology (RMS)

where A 0 is the free term in the equation, coefficients of A 1 , A 2 ,……….A n are the linear terms, A 11 ,
A 22 ,……..A nn are quadratic terms and A 12 , A 13 ,……..A n-1.n are the interaction terms.
Therefor for two factors, the equation can be expressed as
Y = A 0 + A 1 H + A 2 F + A 12 HF + A 11 H2 + A 22 F2. (4)
The coefficients for the equation (4) have been calculated by means of multiple regression method and
coefficient values have been found out using the Deign Expert software (Version 8.0.1).
3.2 Modelling of micro hardness
After calculating the coefficient values in Equation 4, the mathematical model of the micro hardness has been
determined. Significance test of the model and all the terms have been carried out using analysis of variance
(ANOVA) method. The backward elimination procedure has been applied for removing the insignificant model
terms. The resulting ANOVA of the micro hardness for modified quadratic model is shown in Table 2. The
model F value 18.59929 shows that it is significant. From this table, it can be concluded that the main effects of
compaction pressure (A), sintering temperature (B), sintering temperature2 (B2) are the noteworthy model terms.
The R-squared value (0.91776) is very high, almost close to 1, which is desirable.
The final equation in terms of coded factor can be expressed as
Micro hardness (H coded ) = 19.40952 + 0.4219224 × A - 0.6123658 × B -1.086006867 × B2. (5)
And the final equation in terms of actual factor can be expressed as
Micro hardness (H actual ) = - 1076.019262 + 0.4219224 × Pressure + 1.377841473× Sintering Temperature -
0.000434403× Sintering Temperature2. (6)
Table 2. Final ANOVA for micro hardness
Source Sum of squares df Mean square F value P-value
Prob>F
Model 5.676884 3 1.892294712 18.59929 0.003833 significant
A-Pressure 1.068111 1 1.06811107 10.49842 0.022946 significant
B-Sinter Temp 2.249951 1 2.249951238 22.11468 0.005325 significant
B2 2.358822 1 2.358821829 23.18476 0.004819 significant
Residual 0.508701 5 0.10174018
Cor Total 6.185585 8
Std. Dev. 0.318967 R-Squared 0.91776
Mean 18.68552 Adj R-Squared 0.868416
C.V. % 1.70703 Pred R-Squared 0.679779
PRESS 1.980756 Adeq Precision 11.95522

3.3 Modelling of fracture toughness

Similarly, the significance test of the model and all the terms for the modelling of fracture toughness have been
carried out using analysis of variance (ANOVA) method. Again the backward elimination procedure has been
applied for removing the insignificant model terms. The resulting ANOVA of the fracture toughness for
modified quadratic model is shown in Table 3. The model F value 13.89588 shows that it is significant. From
this table it can be decided that the main effect of compaction pressure (A), sintering temperature (B), sintering
temperature2 (B2) is the notable model term. The R-squared value (0.892905) is very high, almost close to 1,
which is again desirable. Therefore, the final equation in terms of coded factor can be depicted as
Fracture toughness (F coded ) = 6.104458538 - 0.387234984× A + 0.33645151×B-0.405036165 × B2
(7)
358
 Proceedings of NCAMMM - 2018

While the final equation in terms of actual factor can be written as

Fracture toughness (F actual ) = - 416.708378 - 0.387234984× Pressure +0.525175322× Sintering Temperature-
0.000162014× Sintering Temperature2. (8)
Table 3. Final ANOVA for fracture toughness
Source Sum of squares df Mean square F value P-value
Prob>F
Model 1.907011901 3 0.635670634 13.89588 0.007347 significant
A-Pressure 0.899705599 1 0.899705599 19.66773 0.006797 significant
B-Sinter Temp 0.679197712 1 0.679197712 14.84738 0.011962 significant
B2 0.32810859 1 0.32810859 7.172513 0.043917 significant
Residual 0.228726377 5 0.045745275
Cor Total 2.135738278 8
Std. Dev. 0.213881452 R-Squared 0.892905
Mean 5.834434428 Adj R-Squared 0.828648
C.V. % 3.665847212 Pred R-Squared 0.631724
PRESS 0.786540918 Adeq Precision 10.63176
3.4 Confirmation run
To validate the developed model, all the 9 confirmation run experiments are performed as described in Table 4.
The predicted values for the response parameters have beencalculated from the equations (6) and (8). The
percentage errors for the micro hardness and the fracture toughness are varied from -2.07 to 1.96% and -4.16 to
4.60% respectively.
Table 4. Confirmation run experiments
Run Factors Responses
no. A: Pressure B: Sintering Micro Hardness (GPa) Fracture Toughness (MPa.m0.5)
(ton/cm2) Temperature Predicted Experimental Error% Predicted Experimental Error%
(OC) value value value value
1 6 1600 18.51 18.38 -0.72 5.75 5.93 3.11
2 8 1650 18.94 19.03 0.50 5.36 5.39 0.48
3 8 1550 19.36 19.4 0.22 4.98 4.77 -4.16
4 6 1550 18.99 19.36 1.96 6.49 6.49 -0.04
5 8 1600 19.41 19.42 0.06 6.11 5.99 -1.89
6 7 1650 19.83 19.45 -1.92 5.72 5.83 1.95
7 7 1600 17.29 16.93 -2.07 6.42 6.41 -0.22
8 7 1550 17.71 17.86 0.84 6.04 5.79 -4.09
9 6 1650 18.13 18.35 1.20 5.65 5.91 4.60

4. Results and discussion

4.1 Density and volumetric shrinkage analysis at different stages of process parameters
Table 1 comprises the variations in the densities, volumetric shrinkage of the samples with the variation of
process parameters viz., compaction pressure and sintering temperature.
The density is increased while the compaction pressure and sintering temperature both are increased
simultaneously.It has been observed from the Figure 2 that the lowest %volumetric shrinkage (~38.38%)
occursin the case of Z 6 -1600. At 1600OC, the %volumetric shrinkages follow a decent trend while compacted at
different pressure conditions. With the increment of compaction pressure, the voids can significantly be
minimized which obviously influences the volumetric shrinkage into the matrices. The FESEM micrograph
359
Optimization of micro hardness and facture toughness of zirconia toughened alumina (ZTA) under different compacting
pressures and sintering temperatures using Response Surface Methodology (RMS)

(Figure 4) shows a significant size difference between the alumina and zirconia particles. So there may be a
tendency of the zirconia particles to betrappedinto the inter-particle voids produced by the comparatively bigger
alumina particle with the increment of compaction pressure as well. Moreover there may be significant chances
of getting volumetric shrinkage after sintering at the higher temperature as well as the higher pressure since
there may be more chances of grain growth due to the higher degree of closeness of particles rather than the
closeness of the particles compacted comparatively lower pressure. Therefore, the grain size can be broadened
with higher level of pressures (Figure 2) resulting in greater volumetric shrinkage.

Figure 2. Grain growth phenomena with different levels of compaction

40.6
40.4 6 ton/cm2
40.2 7 ton/cm2
8 ton/cm2
Volumetric shrinkage (%)

40.0
39.8
39.6
39.4
39.2
39.0
38.8
38.6
38.4
38.2
1550 1600 1650
Sintering Temperature (OC)

Figure 3.Volumetric shrinkage for all samples at different compacting pressure and sintering temperature

4.2 Effect of variables on micro hardnessand fracture toughness

From the Table 1, it is cleared that the maximum hardness (~19.45GPa) has been observed in the case of Z 8 -
1600 and the maximum fracture toughness is found in the case of Z 6 -1600 (~6.49MPa.m0.5). Z 6 -1600 shows an
increment in hardness while sintered at 1600OC rather than 1550OC but beyond 1600OC again it shows lower
hardness. From the XRD (X’Pert PRO, PANalytical B.V., PW3040/60, Netherland) analysis it can be predicted
that in all the compaction pressure levels the fracture toughness has been increased while sintered from 1550OC
to 1600OC due to the increase of intensity of tetragonal zirconia phases (Figure 7) which signifies the probability
of retention of more metastable stages which ultimately increases the volumetric expansion at comparatively
higher temperature and as a result the crack propagation can be hindered. This may be the reason for obtaining
the higher toughness value when sintered at 1600OC. Again beyond 1600OC the tetragonal zirconia phasesare
lessened resulting decrease in the toughness value. FESEM micrograph confirms the grain size phenomena in
Figure 4. At 1600OC sintering temperature the m-ZrO 2 (monoclinic) phase is shifted to t-ZrO 2 (tetragonal) phase.
The post sintered bulk density has been increased with the increase of compaction pressure as found in Figure 3
and the grain growth is increased with small amount as depicted in FESEM micrographs in Figure 5. The
hardness value has increased almost in all sintering temperatures while the compaction pressure is increased
360
 Proceedings of NCAMMM - 2018

from 6 to 7ton/cm2 due to the lesser porosity and good densification of the samples as found out from the
micrographs in Figure 5. Though the hardness value is decreased drastically when compacted in 8 ton/cm2
having the corresponding sintering temperature as 1650°C (Figure 6). This may be the reason of degree of
closeness of the compacted particle as described earlier. Due to the higher compaction there may be a good
chance to spread the grain boundary more conveniently which ultimately resultsin bigger grain size and as a
consequence more chances of porosity as well as poor surface strength which ultimately reduces the micro
hardness significantly.

Figure 4. Microstructural changes at the following compaction pressures and sintering temperatures
corresponding to two different magnifications 15kX and 50kX:(a) (7ton/cm2, 1550oC) 15K X, (b) (7ton/cm2,
1550OC) 50K X, (c) (7ton/cm2, 1600OC) ) 15K X, (d) (7ton/cm2, 1600OC) 50K X, (e) (7ton/cm2, 1650OC) 15K X, &(f)
(7ton/cm2, 1650OC) 50K X .

361
Optimization of micro hardness and facture toughness of zirconia toughened alumina (ZTA) under different compacting
pressures and sintering temperatures using Response Surface Methodology (RMS)

Figure 5. Microstructural changes at the following compaction pressures and sintering temperatures
corresponding to two different magnifications 15K X and 50K X: (a) (6ton/cm2, 1600oC) 15K X, (b) 6ton/cm2,
1600OC) 50K X, (c) (7ton/cm2, 1600OC) 15K X , (d) (7ton/cm2, 1600OC) 50K X, ; (e) (8ton/cm2, 1600OC) 15K X & (f)
(8ton/cm2, 1600OC) 50K X .

8.0
20.0
2 7.5
6 ton/cm2
6ton/cm 7 ton/cm2
19.5 7ton/cm2 7.0 8 ton/cm2
Fracture toughness (MPa.m0.5)

8ton/cm2
19.0 6.5
Hardness (GPa)

6.0
18.5
5.5
18.0
5.0
17.5 4.5

17.0 4.0

3.5 (b)
16.5 (a)
3.0
1550 1600 1650 1550 1600 1650
Sintering temperature (OC) Sintering temperature (OC)

Figure 6. Mechanical behaviour of all samples at different compacting pressures and sintering temperatures: (a)
Hardness, (b) Fracture toughness.

362
 Proceedings of NCAMMM - 2018

1550oC
Al2O3
Al2O3 1600OC Al2O3

Al2O3 1650OC
t-ZrO Al2O3 Al2O3
2
t-ZrO t-ZrO
m-ZrO m-ZrO m-ZrO 2 Al O m-ZrO
2
2 2 2 2 3 m-ZrO
2 2
Al2O3
Al2O3 Al2O3

t-ZrO Al2O3
Al2O3 2 Al2O3
t-ZrO t-ZrO
m-ZrO 2 Al O 2
2 m-ZrO m-ZrO 2 3 m-ZrO m-ZrO
2 2 2 2

Al2O3 Al2O3
Al2O3

Al2O3
t-ZrO Al2O3 Al2O3
2
t-ZrO t-ZrO
m-ZrO 2 Al O m-ZrO 2
2 m-ZrO m-ZrO 2 3 2 m-ZrO
2 2 2

20 30 40 50 60
2Theta (Degree)

Figure 7. XRD pattern of the ZTA ceramics compacted at 7 ton/cm2 and sintered at different temperatures 1550,
1600 and 1650OC.

Figure 8. EDS images for the sample Z 8 -1650 confirming the presence of Al 2 O 3 and ZrO 2 into the final
samples

EDS study ensures the presence of Al 2 O 3 and ZrO 2 and the weight percentages of both into the final sintered
matrixareapproximately 87 and 17wt% respectively. The smaller brighter particle shows the higher intensity of
zirconia content while the larger darker particle shows the maximum intensity of alumina content as ascribed in
Figure 8.
4.3 Optimization of parameters
In this work, desirability function optimization has been used for maximizing the micro hardness and fracture
toughness of the developed ceramic. For this process, compaction pressure and sintering temperature have been
kept in range. The optimization result as per desirability level is reported in Table 5. In this optimization
process, 6 ton/cm2 and 1596.89 OC have been calculated as optimized compaction pressure and sintering
temperature corresponding to the highest value (0.906) of the desirability function in order to achieve maximize
value of micro hardness (~19.02GPa) and fracture toughness (~6.47MPa.m0.5) simultaneously.

363
Optimization of micro hardness and facture toughness of zirconia toughened alumina (ZTA) under different compacting
pressures and sintering temperatures using Response Surface Methodology (RMS)

Table 5 Optimization result

Solution Process parameters Response parameters Desirability Remarks
number Pressure Sintering Hardness Fracture
(ton/cm2) Temperature (GPa) Toughness
(OC) (MPa.m0.5)
1 6 1596.89 19.02 6.47 0.906 Selected
2 6 1597.35 19.02 6.47 0.905
3 6 1596.24 19.03 6.46 0.905

5. Conclusions
The following conclusions can be made from the present study:
• Compaction pressure and sintering pressure both affected the micro hardness and fracture toughness of final
samples significantly.
• Compaction pressure enhanced the micro hardness and fracture toughness while increasingfrom 6 to
8ton/cm2 but up to the 1600OC sintering temperature. But beyond this temperature both the mechanical
properties reduces drastically.
• Sintering temperature plays the most of the influencing parameter in improving the mechanical properties as
well.
• ANOVA statistical model is best suited seamlessly to optimize the process parameter while the goal is to
achieve the highest values of micro hardness and fracture toughness together.
• Using this RSM technique, optimization of complex multiple variables is possible easily.

Acknowledgement
Authors would like to acknowledge Nanomission, Department of Science and Technology (DST), Govt. of India
for the financial supportas a project (SR/NM/NT-1063/2015) to execute this research. The authors are also
grateful to all the laboratory co-workers of CAMP Group, CSIR-CMERI, Durgapur, India for their kind support.
Special thanks to Dr. Bijay Kumar Show, Department of Metallurgical and Materials Engineering, NIT,
Durgapur, India for carrying out the XRD analysis. Lastly authors are thankful to Prof. (Dr.) Harish Hirani,
Director, CSIR-CMERI, Durgapur for the continuous encouragement to write this paper.

References
[1] Jianxin D, Tongkun C, Xuefeng Y, Jianhua Y, Self-lubrication of sintered ceramic tools with CaF2
additions in dry cutting, International Journal of Machine Tools & Manufacture 46 (2006) 957–963.
[2] Mondal B, Mandal N, Development of Ce-PSZ-/Y-PSZ-Toughened Alumina Inserts for High-Speed
Machining Steel, Int. J. Appl. Ceram. Technol., 11 [2] 228–239 (2014).
[3] Muthuraja A, Senthilvelan S, Development of tungsten carbide based self lubricant cutting tool material:
Preliminary investigation, Int. Journal of Refractory Metals and Hard Materials 48 (2015) 89–96.
[4] Mondal B, Chattopadhyay A.B, Virkar A, Paul A, Development and performance of zirconia-toughened
alumina ceramic tools, Wear, 156 (1992) 365-383.
[5] Dou J, Zhang C, Chen C, Zhang X, Effects of sintering temperature on the properties of
alumina/hydroxyapatite composites, J Sol-Gel SciTechnol (2017) 84:23–27.

364
 Proceedings of NCAMMM - 2018

[6] Barbieri R.A, Perottoni C.A, Zorzi J.E, Influence of Sintering Temperature on the Mechanical Properties
of Alumina Springs, Int. J. Appl. Ceram. Technol., 9 [3] 599–605 (2012).
[7] Aksel C, Mechanical Properties of Alumina-Mullite-Zircon Refractories, Key Engineering Materials Vols.
264-268 (2004) pp. 1791-1794.
[8] Barbosa L.P, Lima E.P.R, Rodrigues D, Filho F.A, Effect of compacting pressure on liquid phase sintering
of ASTM 2124 alloy, Materials Science Forum Vols 660-661 (2010) pp 623-628.
[9] Yusoff Z, Jamaludin S.B, The Influence of Particle Sizes and Compaction Pressure on Surface Hardness of
Aluminum Composite Fabricated Via Powder Metallurgy, Australian Journal of Basic and Applied
Sciences, 5(11): 133-140, 2011.
[10] Elsen S.R, Ramesh T, Optimization to develop multiple response hardness and compressive strength of
zirconia reinforced alumina by using RSM and GRA, Int. Journal of Refractory Metals and Hard Materials
52 (2015) 159–164.
[11] Noordin M.Y, Venkatesh V.C, Sharif S, Elting S, Abdullah A, Application of response surface
methodology in describing the performance of coated carbide tools when turning AISI 1045 steel, Journal
of Materials Processing Technology 145 (2004) 46–58.
[12] Choudhury I.A, El-Baradie M.A, Surface roughness prediction in the turning of high-strength steel by
factorial design of experiments, J. Mater. Process. Technol. 67 (1997) 55–61.
[13] Evans A. G, Charles E. A, Fracture Toughness Determination by Indentation, J.Am. Ceram. Soc., Vol. 59,
(1976), pp. 371-372.

365
Sub - theme

Signal Processing
and Machine
Learning
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

An Expert System based Approach for Selection of Wear Resistant Materials for Steel
Plant Applications

K.K. Singh, C. Mandal, Santosh Kumar

Research & Development Centre for Iron & Steel, Steel Authority of India Limited, Ranchi-834002, Jharkhand,
India

Abstract: A framework for rule based expert system has been designed and developed by capturing the
knowledge available in the area of wear prevention and control in the form of knowledge rules. These rules
have been associated with certain criterion for their applicability in specific situations that the user indicates
during consultation with the expert system. The applicable knowledge rules upon execution, suggests a list of
potential materials with explanation and cautions regarding the application. The specific application examined
here is steel plant components. The expert system framework encompasses various wear situations like solid
particle wear (abrasion and erosion), slurry wear and sliding and rolling wear and the conditions influencing
the dominance of these factors have been encapsulated in the suggested selections. The range of material choice
has been restricted to metallic materials, mainly, with a little coverage of non-metallic, ceramics, hard facing
and coating materials. The expert system qualitatively assesses the nature of practical wear related problems
that may arise in component service. The practicing engineers can use the solutions provided by the system as a
general guideline for selection of materials for tribological applications. A more comprehensive expert system
requires a much better incorporation of wear design principles. The wear resistance of various materials needs
to be understood in unambiguous terms that can be translated into knowledge rules. This requires a long and
sustained discussion and communication among the tribologists, material scientists, artificial intelligence
experts and computer programmer.
Keywords: Wear Resistant Materials, Expert System, Material Selection, Wear

1. Introduction
Wear can be defined as the progressive loss of materials from the surfaces of contacting bodies in relative
motion. The conditions that affect the nature and extent of wear are load, speed, temperature, dynamic loading
(i.e. variation in loads during the operation), presence or absence of external particles in the system and their
natures, external environment, lubrication condition, some kind of damage in the system and assembly. The
wear mechanisms that will manifest in a given system will depend on the prevailing conditions and their
proportions. The material characteristics that are considered for providing solutions to wear problems are
material composition, bulk and surface treatments, bulk mechanical properties, surface characteristics, structural
features. The macroscopic and microscopic examinations as well as simulation based studies will help material
engineers in determining the source of wear, type of wear and accordingly a solution can be devised by
tweaking with the material composition or treatment. Sometimes minor or major alterations in the tribological
systems are also required [1].
The steel production in an integrated iron and steel works involves handling, storage and transportation
of very large volumes of raw materials; such as iron ore and coal; and intermediate products; such as sinter and
coke; in the primary areas of steel production; such as blast furnace, coke oven and sinter plant. At this stage of
hot metal production, because of the aggressive nature of the raw materials and intermediate products mentioned
An expert system based approach for selection of wear resistant materials for steel plant applications

here, the components involved are subjected to severe wear classified as abrasion and erosion as a result of the
interactions between the working surfaces of the components involved and the raw materials and intermediate
products. The classic examples of components at the primary stage of steel production are protective liners of
material handling equipment, crushing hammers, hot and cold screens, impeller casings and impeller blades and
many more. During the course of production of liquid steel as well as at the pre-finishing and finishing stage, the
components involved are subjected to heavy mechanical and thermal stresses, causing severe wear classified as
adhesion, mechanical and thermal fatigue, high temperature oxidation as a result of the interactions between the
component surfaces itself as well as the component surfaces and intermediate products; such as blooms and
billets, slabs, sheets and coils. The classic examples of components at the secondary stage of steel production
are several kinds of working and auxiliary rolls, shears and knives, mill housings and chokes and many more
[2].
The judgement on suitability of the material is dependent on the nature and extent of wear occurring in
a given set of operating conditions. It therefore becomes important for the material engineers to possess an
adequate knowledge of the tribological system, wear mechanisms and the characteristics of engineering
materials in order to suggest the most appropriate materials for tribological applications [3].
The use of artificial intelligence techniques like expert systems is being made to undertake the role of
human expert for applications in the areas where a higher level of abstraction is required to process the available
information to make conclusive recommendations such as in case of selection of wear resistant materials. Since
wear is not an intrinsic material property but a system dependent characteristic, no absolute value can be
assigned to a material for its wear resistance. The wear resistance of a material is always reported as a relative
index in comparison to a reference material. The designer or the engineer can use this information to judge the
comparative wear resistance of a material under examination in relation to some other material. It is worth
mentioning here that these relative wear indices of materials are always for a given set of operating conditions,
either actual or simulated in the laboratory. If the operating conditions differ in case of the situation under
consideration, such information on the relative wear index for a different set of operating conditions will not be
of much use. Besides, the data and information on relative wear index of materials reported by different
agencies may differ due to difference in sets of operating conditions. It is therefore not only confusing for the
designers and the engineers to make use of such informations but there is also an issue of reproducibility and
scalability. It is due to these reasons; that the handbooks and several other kinds of databases are not of much
use for the purpose of selecting wear resistant materials. The human expert who can consider all the relevant
factors and who can possess and process the available information at a higher level of abstraction seems to be
the only solution when suggestions on wear resistant materials are sought [4].
The expert system, which is primarily a computer program that maps the human thinking and decision
making process will succeed if the available pieces of knowledge and information have been translated into
unambiguous terms so as to enable the users to provide inputs in an easy manner. The solutions provided by
these expert systems to the users need to be sufficiently explanatory in order to assist the user to make use of it
easily. Although the expert systems that have been developed for the purpose of selection of wear resistant
materials have genuinely attempted to address all these issues, the wide uses of these programs are not known.
This testifies to the intricacies and complications involved in addressing issues like wear resistance of a material
and suitability of a material for a particular application [4].

367
 Proceedings of NCAMMM - 2018

2. Design Methodology
(1) Identification of key entities or objects or variables along with their value sets which encompasses the
description of tribological system, nature and extent of wear and set of engineering materials for such
applications [5].
(2) Identification of the set of questions for user consultation and obtaining inputs in regard to the tribological
system description. Based on the entities identified for system description or nature and extent of wear or
wear resistant material, the questions have been framed with options for answers from the value sets of those
entities [6].
(3) Identification of interrelations amongst the entities or objects or variables which makes the knowledge rules
for the expert system [7].
(4) The value sets for the entities or objects or variables constitutes the boundary conditions of the applicability
of knowledge rules [8].
(5) The execution of applicable knowledge rules based on the user supplied information to suggest the materials
along with material description [9].
The schematic representation of the said expert system is presented in Figure 1.

Figure 1: Schematic Representation of Expert System for Material Selection

3. Development Process
The development of the expert system application includes
(1) Development of hyperlinked HTML (Hyper Text Markup Language) pages to incorporate the design
principles (Stage-I). As has been discussed earlier, the expert system shells, the commercial off-the-shelf
expert system development tools, are typically used to develop expert system programs. However, other
programming tools; such as structured, procedural and object oriented programming languages; can also
be used to develop such programs [10]. In the recent times, the use of modern web based tools has also
been made to develop expert system applications. In the present work web pages have been generated
using HTML with hyper linking in a manner so as to support the program logic [11].
(2) Use of commercial an off-the-shelf expert system shell to implement the design principles (Stage-II). The
same design principles can be incorporated using commercial off-the-shelf expert system development
tools, i.e. expert system shells [12]. This will only change the look and feel of the application, but will
not change the basic functionality of the expert system program [13].
4. Program Description

368
An expert system based approach for selection of wear resistant materials for steel plant applications

The entities in the knowledge base are all the terms defined by the designer and the developer for all the pieces
of knowledge and information and the attributes and values that these entities can take are also assigned by the
designer and the developer. The rules, known as knowledge rules, are the interrelation of entities amongst each
other based on their attributes and values. E.G., if contact exists between two or more bodies and these bodies
are in relative motion with each other, then it is a case of wear. Similarly, if particles are present and the
carrying media is air or gas, then it is a case of solid particle erosion or abrasion. Further, if the particles are
sliding over a body, then it is a case of 2-body abrasion, if particles are being crushed between the contacting
bodies, then it is a case of 3-body abrasion, if particles are striking the component’s surface at an angle (low or
high), then it is a case of either low or high angle abrasion. Next, if it is a case of 2-body or low angle erosion
and the particle present is iron ore or sinter or coke, or the hardness of the particles present is in the range of
450-600 Vicker’s hardness number, then High Cr Iron or Alumina Ceramic are the most suitable materials. The
expert system for selection of wear resistant materials for steel plant applications maps all those situations, that
relate to wear problems in a steel plant into the form of entities or objects or variables (entities, objects and
variables are the terms which are used interchangeably with the same meaning) with certain value sets or
attributes (value sets and attributes are the terms which are used interchangeably with the same meaning)
assigned to them. The user is asked questions that have been framed based on these entities or objects or
variables with the options/choices for answering based on the assigned value sets or variables. The expert
system, based on the answers for various questions that it receives from the user, processes the information by
making one or more of the knowledge rules applicable for the given situation and suggests the materials. The
materials to be suggested by the expert system too are mapped as entities or objects or variables with value sets
or attributes as applicable (“Yes”) or not applicable (“No”) in a given situation.
The complete set of entity/object/variable and their value sets along with their interrelation in the form of IF
(portion) and THEN (portion) are presented in Table 1.

5. Typical Program Outcome

A typical example of material selection is for choosing protective liners for transfer chutes used in transporting
and handling of iron ore, coke and sinter in a Raw Materials Handling Plant (RMHP) and Coke and Sinter
weighing hoppers of Blast Furnace (BF) of a steel plant. This example is to make the reader understand how the
expert system will help the user in deciding the appropriate materials for such applications. The liners in these
applications are subjected to abrasion by iron ore, coke and sinter under low stress conditions. The user actions
are laid out in the following sequence:
1. The user will start interacting with the expert system program by answering “Yes” to questions relating to
existence of contact and relative motion and accordingly the program will judge the applicability of the
entity relating to existence of wear.
2. Next, the user will be presented with questions relating to presence or absence of external particles for which
the answer will be “Yes” in this particular case. The program will judge the applicability of the entity
relating to a case of particle wear. The user, upon answering as “Yes”, will immediately be prompted to the
question listed relating to carrying media of external particles for which the answer will be “Air/Gas”. The
program will judge the applicability of entity relating to abrasion or erosion wear. The program then will
take the user to question relating to the nature of interaction between the particles and component’s body for
which the applicable answer is “Sliding”. The program will judge the applicability of entity relating to the

369
 Proceedings of NCAMMM - 2018

case of 2-Body Abrasion wear. The program will now try to judge the degree of severity of wear based on
the particle description and so the user will be presented with further related questions.
Table 1: Network of Entities/Objects/Variables (Value Sets) – Rules – Recommendations
IF (portion) THEN (portion)
Contact Exists (Y/N)
Wear Situation Exists (Y/N)
Relative Motion Exists (Y/N)
Wear Situation Exists (Y/N) Particle Wear (Y/N)
External Particles Present (Y/N) Slurry Wear (Y/N)
External Particles Carrying Media (Air/Liquid) Adhesive Wear (Y/N)
Particle Wear (Y/N) Abrasion Wear (Y/N)
External Particles Motion (Sliding/Trapped/Impacting/Striking) Erosion Wear (Y/N)
Abrasion Wear (Y/N) 2BodyAbrasion (Y/N)
External Particles Motion (Sliding/Trapped/Impacting) 3BodyAbrasion (Y/N)
Gouging (Y/N)
Erosion Wear (Y/N) Low Angle Erosion (Y/N)
Impingement Angle (Low/High) High Angle Erosion (Y/N)
Adhesive Wear (Y/N) Rolling Wear (Y/N)
Component Motion (Rolling/Sliding) Sliding Wear (Y/N)
Impact Condition
(Nil or Low/Moderate/High)
Stress Condition
(Nil or Low/Moderate/High)
Particles Velocity (Low/High)
Particles Concentration (Low/High)
External Particles Shape
(Regular/Irregular)
Wear Resistant Materials
External Particle Size
(Carbon & Alloy Steel)
(Fine/Coarse)
(Tool Steel)
External Particles Description
(Bearing Steel)
(Name or Hardness)
(Stainless Steel)
Slurry Description
(Mn-Steel)
(Name or Miller No.)
(Cast Iron)
&
(Heat Resistant Steel)
2BodyAbrasion (Y/N)
(Hard facing Alloys & Composite Plates)
3BodyAbrasion (Y/N)
(Surface Coatings & Treatments)
Gouging (Y/N)
(Ceramics)
Low Angle Erosion (Y/N)
(Rubber & Synthetic Rubber)
High Angle Erosion (Y/N)
(Polyurethane)
Slurry Wear (Y/N)
(UHMWP)
Contact Stresses (Low/Moderate/High)
(Non-Ferrous Alloys)
Contact Type (Point/Line/Area)
Lubrication Condition
(Poor/Good)
Operating Temperature
(Low/High)
Contact Body Speed (Low/Moderate/High)
Mild Wear (Y/N)
Severe Wear (Y/N)

370
An expert system based approach for selection of wear resistant materials for steel plant applications

3. From those information inputs the materials based on High Chromium Iron Alloy and Alumina Ceramic
Tiles will be suggested. The details on these materials will be fetched from the file system. It is worth
mentioning here that while answering for questions, the user can estimate damage levels from the Miller
Number in case of slurry type and Dry Particle Abrasivity/Erosiveness in case of particle type. The program
provides a help in choosing the likely value based on knowledge about the particles present in the system
(14).
Let us consider another example for use of the expert system for material selection for circular trimming shear
and scrap chopper blades, which are used in combination in pickling lines of cold rolling mills of steel plant,
respectively for trimming the edges of hot rolled coils for a smoother and accurately sized feedstock for cold
rolling and chopping it to make handling of the scrap, so generated, easy. It is a typical case of rolling and
sliding wear with moderated impact conditions. The requirements on the part of materials for such applications
are adequate wear resistance with moderately good breakage (chip off) resistance so that the edges do not get
blunt frequently and at the same time it does not chip off or breaks due to impact.
The expert system for selection of material in such a case will ask questions relating to existence of
contact and relative motion, presence or absence of external particles, type of motion and contact between the
bodies, levels of impact/stress and speed the component is encountered with, material of conforming body,
temperature and lubrication condition. The typical answer for these questions would be “Yes” and “Yes”, “No”,
“Sliding/Rolling” and “Point/Line”, “Low” and “Slow”, “carbon & Alloy Steel/Tool & Bearing Steel”,
“Ambient” and “Inadequate/Adequate” respectively. The suggested materials for this particular application will
be X50CrMoV51/AISI-D2 grades of tool steel materials.

6. Conclusions
The expert system program under discussion has been developed in order to assist practicing engineers and
designers to devise solutions in regard to materials selection for combating wear in steel industry without the
help of a human expert. The expert system will facilitate the users in answering a few questions about the
specific application areas that they will be examining. These questions have been framed keeping in view
broader reference to wear situations that usually prevail in steel plants. Upon receiving the responses of the
users against the questions, the expert system keeps evaluating the responses sequentially to intermittently and
terminally determine the nature and extent of wear that might be prevailing in the given situation. These
determinations are based on intelligently processing the user’s responses with the help of knowledge rules that
have been defined based on the established wear theories, published literature and the rule of thumb. At the end,
when all the responses are processed, the expert system will suggest a set of materials that will be most
appropriate for use in the given situation. The typical cases discussed in the text for selection of materials for
shearing knives and wear protective liners elaborate on how the expert system works in real situations.

Acknowledgements
The authors wish to express their deep sense of gratitude to the management of Research & Development Centre
for Iron & Steel, Steel Authority of India Limited, Ranchi-834002, Jharkhand, India for their constant support
and encouragement while pursuing the work and for the permission granted to publish the outcome.

371
 Proceedings of NCAMMM - 2018

References
[1] Eyre TS. Source book on wear control technology. ASM, Metals Park. 1978:1-0.
[2] Landsdown AR, Price AL. Materials to resist wear. A guide to their selection and use. The Pergamon
Materials Engineering Practice Series, Oxford: Pergamon Press, 1986. 1986.
[3] Phelps A. Material selection for wear resistance; University of Dayton Research Institute; Dayton; Ohio @
Ch. 41; Handbook of Material Selection; Myer Kutz (Editor); Willey (2002); 1520p; January 2002
[4] Franklin SE, Dijkman JA. The implementation of tribological principles in an expert-system
(“PRECEPT”) for the selection of metallic materials, surface treatments and coatings in engineering
design. Wear. 1995 Feb 1;181:1-0.
[5] Woydt M. Modern methods to retrieve innovative material solutions for Tribosystems (c). Tribology &
Lubrication Technology. 2000 May 1;56(5):26.
[6] İpek M, Selvi İH, Findik F, Torkul O, Cedimoğlu IH. An expert system based material selection approach
to manufacturing. Materials & Design. 2013 May 31;47:331-40.
[7] Blundell JK, Greenway RB. Expert system for material selection. InNumerical Techniques for Engineering
Analysis and Design 1987 (pp. 633-647). Springer, Dordrecht.
[8] Zarandi MH, Mansour S, Hosseinijou SA, Avazbeigi M. A material selection methodology and expert
system for sustainable product design. The International Journal of Advanced Manufacturing Technology.
2011 Dec 1;57(9):885-903.
[9] Lau P, Chan CW. Expert System For Material Selection of Process Equipment. InTechnical
Meeting/Petroleum Conference of The South Saskatchewan Section 1995 Jan 1. Petroleum Society of
Canada.
[10] Basan R, Franulović M, Križan B. Web-based material data knowledge base and expert system. In15th
International Research/Expert Conference" Trends in the development of machinery and associated
technology"-TMT 2011 2011 Jan 1.
[11] Nofal M., Fouad KM. Developing web based semantic expert system; IJCSI International Journal of
Computer Science Issues; Vol. 11, Issue 1; No. 1; January 2011
[12] Kim W, Song YU, Hong JS. Web enabled expert systems using hyperlink-based inference. Expert Systems
with Applications. 2005 Jan 31;28(1):79-91.
[13] P. Mayer, D. LeFrancois; SURFWEAR - Expert system for engineering of wear resistance surfaces and for
wear problem diagnosis; Ontario Hydro (Research Division); 800 Kipling Avenue; Toronto; Ontario;
Canada M8Z 5S4; pp330-306 & 307-311 @ Material Performance Maintenance : Proceedings of
International Symposium on Material Performance Maintenance; Ottawa; Ontario; Canada; 18-21 August
1991; (eds.) R.W. Revle, V.S. Sastri, M. Elbonjdaini, E. Ghali, D.L. Piron, P.R. Roberge, P. Mayer
[14] Slurry Erosion : Uses, Applications and Testing Methods : A Symposium; Miller/Schmidt (editors);
ASTM, STP, 946

372
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Identification of Diseased Cells using Image Processing

a
Puja Mitra, bSamik Dutta, aAbhiram Hens, aNagahanumaiah
a
Micro Systems Technology Laboratory, bPrecision Engineering & Metrology Laboratory, CSIR-Central
Mechanical Engineering Research Institute, Durgapur, India

Abstract: Biomedical research is updating with the improved accuracy of detection and decision making due to
the advancement of computer vision. Therefore, in the present study, an attempt has been made to classify
various diseases by analyzing the microscopic images of the diseased cells from the shape features of those
cells. For quantifying the shapes, different geometric shape factors viz. circularity, roundness and SGF are
evaluated from the edge detected images for each type of cells. Based on the quantified shape factors, those
diseased cells are identified. Circularity is found the most promising feature for this automatic identification of
the diseased cells. Also the identification of nucleus is successfully performed in this work as the evaluation of
the shape of nucleus is often required for cell diagnostic purpose.
Key words: Image processing, shape factors, geometric features, thresholding, edge detection, cell classification

1. Introduction
Currently, image processing techniques are being widely used in the area of biomedical research. The image
processing approach is a preferable method because it is cost effective, automated, remote and fast way to
identify diseases provided the availability of some suitable images. It was found that from the microscopic
image of an ensemble of cells, circulating tumour cells (CTCs) can be isolated by applying proper image
analysis method [1]. Similarly, diseased cells related to anaemia, leukaemia, malaria can also be detected by
proper image analysis. A number of earlier works have been reported which attempted to analyze the images of
cells for their detection where image segmentation or thresholding took the lead role to segment the target cell
from its background [2]. Whenever a healthy cell is infected by a disease, there is a change in its geometric
features and shape. In order to determine the changes, various shape features such as area, perimeter, diameter
and circularity were evaluated [2]. Although image processing is a promising approach towards biomedical
engineering, the following aspects need to be addressed: (a) feature extraction to extract of relevant features for
accurate identification of cells, (b) implementation of robust image enhancement method, and (c) filtering of
noise.
In the present work, an attempt has been made to identify circulating tumor cells from the normal blood
cells from a blood sample. Also, various shape factors are extracted, in this study, from the processed images to
differentiate between CTCs and normal WBCs; leukaemia affected cell and normal WBCs; anaemia affected
and normal RBCs; malaria affected and normal RBCs. This work mainly focuses on the image segmentation and
feature extraction process. Section II provides detailed description of the applied methods. Section III describes
results and discussion; and Section IV outlines the conclusion.

2. Methodology
In this work, image pre-processing is performed to enhance the microscopic cell images followed by the
thresholding technique to segment the required cell pixels from the background pixels. Next, edge detection is
utilized to detect the boundary pixels of the cell, from which the shape analysis can be performed to extract
Identification of diseased cells using image processing

shape features for identification or classification of various types diseased cells. All the analyses are performed
in MATLAB R2013a® environment.
2.1 Isolation of Circulating Tumor Cells(CTC)
Automatic identification of Circulating Tumor Cells (CTC) in blood samples is very important for the
development of automatic machines for haematological purpose. Therefore, in this work, isolation of CTC cells
from the other cells is performed by using image analysis technique. The steps of image analyses techniques
adopted for the isolation of CTCs are depicted in Fig. 1.

Gray level Pre-processed CTC isolated alongwith its

Color Image
Image Image shape factor determined

Figure 1. Flowchart describing isolation of CTCs from other healthy blood cells

The microscopic images of blood samples are generally of low contrast images and therefore they are needed to
be enhanced. To enhance the contrast, Contrast Limited Adaptive Histogram Equalization (CLAHE) is applied
here. In CLAHE, an adaptive histogram equalization method is applied locally where this user-defined local
regions can be produced by dividing an image into several sub-images. The over amplification of artefacts can
be avoided by utilizing this localized adaptive histogram equalization method, as the contrast enhancement can
be limit by the user [3]. Thresholding is applied on the enhanced image for segmenting and isolating the CTCs
from the normal cells. In image thresholding, the pixels having gray level values above the threshold are
assigned to 1 pixel value and considered as foreground pixels and other pixels are assigned to 0 pixel values or
background pixels. Therefore, choosing an optimum threshold value is important to obtain accurate results and
therefore in this technique, Otsu's optimal thresholding technique is used.

2.2 Differentiate between diseased cells and normal cells

For differentiating of diseased cells from normal cells, geometric or shape features are extracted from the edge
detected image of the original gray scale image. This edge detection is performed here by using Canny edge
detection technique [3]. Edge detection has a vital role in image segmentation. Here, objects of interest are
detected according to the sharp changes in intensity of the image. Noise is a very important factor in edge
detection. Canny edge detection is carried out. Since, even a little noise can effect edge detection; all necessary
edges are detected by removing noise without affecting any changes in the features of the images with the help
of canny edge detector.

The steps adopted here are depicted in Fig. 2.

Gray level Edge detected Geometric Feature

Color Image
Image Image Extraction

Figure 2. Flowchart showing characterization of blood cells for diseased cells detection

Feature extraction plays a major role in the research area of pattern recognition and image processing. Here in
this paper, we have concentrated on Geometrical features. The following dimensionless shape parameters have
extracted for differentiating diseased and normal cells.
Circularity: is proportional to the ratio between area and the square of its perimeter. The scale varies according
to the shape. A cell is considered to be round when the circularity is equal to 1.

374
 Proceedings of NCAMMM - 2018

𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷 𝟐𝟐
𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪 =
𝟒𝟒 × 𝝅𝝅 × 𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨

Roundness: It is also similar to the circularity measure.

𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷𝑷 𝟐𝟐
𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪 =
𝑨𝑨𝑨𝑨𝑨𝑨𝑨𝑨
Shape geometric factor (SGF): is the ratio between larger diameter of the cell and smaller diameter of the cell.

𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫
𝑺𝑺𝑺𝑺𝑺𝑺 =
𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺𝑺 𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫𝑫

3. Results and Discussion

Isolation of circulating tumor cells (CTCs) from healthy blood cells is shown in Fig. 3. Figs 3(a) and 3(e) show
the original images of blood cells containing CTCs. The original images are converted to the gray scale images
from further processing. The gray scale images are then analyzed by CLAHE (contrast limited adaptive
histogram equalization) algorithm to enhance the contrast of the images and the results are shown in Figs. 3(c)
and 3(g). Next, thresholding based on Otsu's method was applied to the images for isolating the large sized
CTCs. Isolated CTCs are shown in Fig. 3(d) and 3(h).

Figure 3. Detection and isolation of CTC: Images in the first column [(a) and (e)] shows the original image of
blood cells containing CTC from two different sources [4, 5]. Images in second column [(b) and (f)] show the
corresponding gray scale image and images in the 3rd column [(c) and (g)] show the corresponding CLAHE
images and finally the last column [(d) and (h)] shows the images of isolated CTC.

Next in order to analyze the shape of the CTCs, three different images are taken from three different
sources which are shown in Fig. 4(a), 4(d) and 4(g). All these images are converted to gray scale (shown in Fig.
4(b), 4(e) and 4(h)) and finally edges of the cells are detected by using Canny edge detection method which are
shown in the images of the 3rd column of Fig. 4. From the final images, various geometric shape factors (as
described in section 2.2) were found out. It was found that circularity values of different CTCs lie between 1 and
2 whereas roundness values lie between 0.5 and 1.0. Another factor SGF shows a large range of values which is
not suitable for classifying the cells. It is important to note that circularity and roundness of normal WBC is near
6.58 and 0.15 respectively. It indicates that circularity of CTC is much lower than that of normal WBC whereas
roundness of CTC is higher than that of normal WBC.

375
Identification of diseased cells using image processing

Figure 4. Shape analysis of CTCs of different morphologies. Images in the first column [(a), (d) and (g)] show
the original image of CTCs taken from different sources [6,7]. Images in second column [(b), (e) and (h)] show
the corresponding gray scale image and finally images in the 3rd column [(c), (f) and (i)] show the Canny edge
detected images.

Next we concentrated on the analysis of leukaemia affected WBCs. For comparison with a normal
WBC, an image of normal WBC is chosen and shown in Fig 5(a). It shows that circularity value of leukaemia
affected cells is slightly greater than 1.0 whereas that of the normal WBC is much greater (6.58). Again
roundness values of leukaemia affected cells are slightly less than 1.0 which is significantly greater than that of
normal WBC (0.15). However, SGF values of the leukaemia affected cells shows a wide range without giving
any indicative measure for comparing with the normal WBCs. The analysis shows that the geometric factors like
circularity and roundness can be used to differentiate the leukaemia affected cells from the normal ones.

Figure 5. Shape analysis of normal white blood cell and leukaemia cells: Images in the first column [(a), (d)
and (g)] show the original images where (a) is the image of normal WBC and (d) and (g) are that of affected
WBCs taken from different sources [1, 8, 9]. Images in second column [(b), (e) and (h)] show the corresponding
gray scale image and images in the 3rd column [(c), (f) and (i)] show the corresponding Canny edge detected
images.

Shape factors are also evaluated for normal RBCs and anaemia affected RBCs. It was found that
circularity values of normal RBCs are higher than that of affected RBC although roundness value is
indistinguishable for this case. It is also found that SGF value of normal RBC is lower than that of affected
RBC.

376
 Proceedings of NCAMMM - 2018

Figure 6. Shape analysis of normal red blood cell and anaemia affected cell. (a) Original image of normal red
blood cell captured from electron microscope [10] and corresponding (b) Gray scale image and (c) Canny edge
detected image; (d) Original image of sickle-cell anaemia captured from electron microscope [10] (e) Gray scale
image of sickle-cell anaemia (f) Canny edge detected image.

We attempted to analyze the images of malaria affected RBCs and normal RBCs in the same above mentioned
method and the results are shown in Fig. 7. It shows that circularity is reduced in case of malaria infected RBC.
However, roundness and SGF which show little increased value for infected RBC are not found to be suitable
parameter for identification of affected cells in this case.

Figure 7. Shape analysis of normal red blood cell and malaria infected cell. (a) Original image of a normal red
blood cell (top view) taken from healthy patients in the Department of Infectious Disease, Jagiellonion
University Hospital, Krakow (11) (b) Gray scale image (c) Canny edge detected image of normal RBC (d)
Original image of a malaria infected cell (top view) of Plasmodium vivax schizont (12)(e) Gray scale image (f)
Canny edge detected image of malaria cell.

4. Conclusion
The present study mainly concentrates on detection and isolation of different diseased cells by identifying their
shape using image processing techniques. In order to quantify the shapes, three shape factors (circularity,
roundness and SGF) are evaluated for all the cells. It was attempted to see if the cells show any particular
change in the values of the shape factors when affected by particular disease. Our results show the indications of
certain change in these values for diseased cells. It can be concluded from this study that the circularity feature
can be utilized to differentiate CTC from normal WBCs; leukaemia affected and normal WBCs; anaemia
affected cells from normal RBCs and malaria affected cells from normal RBCs. Therefore, this study can be
utilized further with lots of future experiments to prove the reproducibility of the results for identification of
diseased cells in real-time environment.

Acknowledgement

377
Identification of diseased cells using image processing

Authors are acknowledged to the Director, CSIR-CMERI, Durgapur for the continuous motivation towards this
research and also to the Micro Systems Technology Group for enormous support to perform this study.

References

[1] Nazlibilek S, Karacor D, Ercan T, Sazli MH, Kalender O, Ege Y. Automatic segmentation, counting, size
determination and classification of white blood cells. Measurement. 2014;55:58-65.
[2] Taherisadr M, Nasirzonouzi M, Baradaran B, Mehdizade A. New approch to red blood cell classification
using morphological image processing. Shiraz E-Medical Journal. 2013;14(1):44-53.
[3] Gonzalez RC, Woods RE. Digital image processing prentice hall. Upper Saddle River, NJ. 2002.
[4] ltd. CB. Images of Circulating Tumor Cells.
[5] Fang J. Ultrafast camera captures rogue cancer cells in blood. ZD Net. 2012.
[6] Stott SL, Hsu C-H, Tsukrov DI, Yu M, Miyamoto DT, Waltman BA, et al. Isolation of circulating tumor
cells using a microvortex-generating herringbone-chip. Proceedings of the National Academy of Sciences.
2010;107(43):18392-7.
[7] Yu M, Stott S, Toner M, Maheswaran S, Haber DA. Circulating tumor cells: approaches to isolation and
characterization. The Journal of cell biology. 2011;192(3):373-82.
[8] Agaian S, Madhukar M, Chronopoulos AT. Automated screening system for acute myelogenous leukemia
detection in blood microscopic images. IEEE Systems journal. 2014;8(3):995-1004.
[9] Mohapatra S, Patra D, editors. Automated cell nucleus segmentation and acute leukemia detection in blood
microscopic images. Systems in Medicine and Biology (ICSMB), 2010 International Conference on; 2010:
IEEE.
[10] Sickle-cell Anemia. Human Diseases And Conditions.
[11] Kozicki M, Czepiel J, Biesiada G, Nowak P, Garlicki A, Wesełucha-Birczyńska A. The ring-stage of
Plasmodium falciparum observed in RBCs of hospitalized malaria patients. Analyst. 2015;140(23):8007-16.
[12] Malaria. Diagnostic Findings.

378
 National Conference on Advanced Materials, Manufacturing and Metrology (NCAMMM - 2018) on 16-17 February, 2018
at CSIR-CMERI, Durgapur

Brain MR Image Analysis using Discrete Wavelet Transform with GLCM Feature
Analysis

1,*
Srinivasan Aruchamy, 2Partha Bhattacharjee, 3Goutam Sanyal
1. Robotics and Automation Group, CSIR-CMERI, Durgapur, . Information Technology Processing Group,
CSIR-CMERI, Durgapur, 2. Department of CSE, NIT, Durgapur
*Corresponding Author: srinivaspsg@gmail.com

Abstract: Computer Aided Diagnosis (CAD) research is playing vital role in any disease identification with
the improved accuracy of detection and decision making due to the development of computer vision. Therefore,
in the present study, an attempt has been made to classify some of the neurodegenerative diseases by analyzing
the brain MR images of the diseased and normal. This work mainly focuses on identification of Alzheimer's,
Mild Cognitive Impairment (MCI) and normal. The proposed algorithm consists of two stages. First stage
comprises of image pre-processing, Skull stripping and Region of Interest (ROI) extraction. In the second
stage the pre-processed image is converted to wavelet domain and Gray level co-occurrence matrix feature
were calculated and used for classification. Also the identification of Alzheimer's is successfully performed in
this work.
Keywords: Image processing, wavelet transform, Alzheimer's, GLCM, edge detection, image classification

1. Introduction
Computer Aided Diagnosis (CAD) [1] systems with image processing is playing vital role in disease
identification and classification. Neurodegenerative diseases are disease which affects the brain nerve cells and
leads to the memory loss. According to reports about 35 million persons are suffering with Alzheimer s disease
and the number is predicted to raise in future [2]. Alzheimer's disease (AD) is one of the primary disease
affects the elderly persons [3]. This disease is a progressive and irreversible in nature but proper early
diagnosis and treatment will control the rate of growth. Manual identification is prone to error and it is
completely depended on the ability of the individuals. For identification of the progression of AD, different
methodologies has been used and found in the literature. It includes MRI, positron emission tomography (PET)
[4][5], white matter and gray matter volumetric area calculation and cerebrospinal fluid(CSF). Most of the
method utilizes machine learning techniques to classify the disease either Normal control (NC), Alzheimer's
disease (AD), Mild Cognitive Impairment (MCI). Krashenyi I et al. [6] proposed a fuzzy logic based classifier
for AD diagnosis. In their method fuzzy inference system uses 24 features from region of interest and the
proposed system uses first three statistical moments for classification of AD and Normal control subjects. Qing
Li et al. [7] developed multi kernel supervised within-class-similarity discriminative dictionary learning
algorithm. Dolph et al. [8] proposed multiclass deep learning method for classifying Alzheimer's disease (AD).
They developed two different learning models by mainly including features from sub-cortical area. They
classified AD with the help of texture as feature by using fractal Brownian motion co-occurrence matrix. They
used ADNI dataset [9] to evaluate the performance and achieved the classification accuracy of 51.4% and
56.8% for different classifiers. This work mainly focuses on the image segmentation and feature extraction
process. Section 2 provides detailed description of the proposed method. Section 3 describes results and
Brain MR image analysis using discrete wavelet transform with GLCM feature analysis

discussion and section 4 outlines the conclusion.

2. Methodology
In this work, image pre-processing is performed in several steps. It includes skull stripping, region of interest
(RoI) extraction and contrast limited adaptive histogram equalization (CLAHE) [10]. Skull stripping [11] is the
process of removing extra-meningeal tissues from the brain MR image. Region of Interest (RoI) is selected by
horizontal and vertical scanning of each input image. The RoI selection algorithms scan the image horizontally
and vertically and it stops when it encounters first white pixel and last white pixel in the horizontal scan line as
well as the vertical scan and marks the indexes which will be used to extract the RoI. This RoI window will be
calculated dynamically based on the view of MR image (side, front, top).

Figure 1. Flowchart describing proposed methodology to detect Alzheimer's disease

2.1 Discrete Wavelet Transform

Discrete Wavelet transform decomposes the input into set of orthogonal wavelets. For a given image I, which is
of the size M×N and having intensity f (x, y) is transformed to wavelet domain through the mathematical
expression.

𝑀𝑀−1 𝑁𝑁−1
1
𝑊𝑊∅ ( 𝑗𝑗0, 𝑚𝑚, 𝑛𝑛) = � � 𝑓𝑓(𝑥𝑥, 𝑦𝑦)∅𝑗𝑗 0,𝑚𝑚 ,𝑛𝑛 (𝑥𝑥, 𝑦𝑦)
√𝑀𝑀𝑀𝑀 𝑥𝑥=0 𝑦𝑦=0

𝑀𝑀−1 𝑁𝑁−1
1
𝑊𝑊𝜓𝜓𝑖𝑖 ( 𝑗𝑗, 𝑚𝑚, 𝑛𝑛) = � � 𝑓𝑓(𝑥𝑥, 𝑦𝑦)∅𝑗𝑗𝑖𝑖 ,𝑚𝑚 ,𝑛𝑛 (𝑥𝑥, 𝑦𝑦) , 𝑖𝑖 = {𝐻𝐻, 𝑉𝑉, 𝐷𝐷}
√𝑀𝑀𝑀𝑀 𝑥𝑥=0 𝑦𝑦=0

where H, V, D are horizontal, vertical and diagonal directions, respectively. The Fig. 2 depicts the 2-D
wavelet transform
2.2 Second Order Statistical Parameters of the image
The given image is converted into a Gray Level Co-occurrence Matrix (GLCM). In this method particular
380
 Proceedings of NCAMMM - 2018

pixel pair is considered though out the image and how many times repeatedly it occur in the matrix is
estimated. This value is stored in the GLCM at the position equal to the value of the pixel pair. The benefit of
GLCM is that many parameters of the image can be extracted from it.

Figure 2. 2-D wavelet decomposition

Figure 3. Method of calculating GLCM horizontally

GLCM is formed out of pixel pairs along horizontal direction. Similarly, GLCM can be found for pairs in
vertical and diagonal direction as well. In this project, we have taken GLCM in four directions with respect to
x axis i.e. along 0, 45, 90 and 135 degrees [12]. Finally the mean of GLCM is calculated and used for the
feature to perform classification.

3. Results
The experiments were carried out with ground truth images available at public database Open Access Series of
Imaging Studies (OASIS) [13]. The database contains 416 subjects out of which 160 subjects were male and
256 subjects were female. The subjects where marked with Clinical Dementia Rating (CDR) [14] and Mini-
Mental State Examination (MMSE) [15] based on single visit or more than one visits separated by at least one
year. 206 subjects were with CDR value 0 which indicates non-demented and 43 (CDR value 1 and 2) were
marked with moderate to high Alzheimer's disease. Two classes have been predefined. The healthy and
slightly defective have been put in class 1 and the significantly defective have been put in class 2. Nearest
neighbor classifier has been implemented using four distance approaches namely Euclidean [16], Correlation
[17], Cosine [18] and City block [19]. To enhance the quality of the image Discrete Wavelet Transform
(DWT) has been applied to the image. To compare the effect of Discrete Wavelet Transform we have
separately observed the results for First level decomposition and Second level decomposition.
To estimate the effectiveness of the proposed method, receiver operator characteristics analysis is

381
Brain MR image analysis using discrete wavelet transform with GLCM feature analysis

applied on the images available in the database. This assessment considers four values, namely True positive
(TP) are the subjects whom we predicted Alzheimer's affected and they do have the disease. True negative
(TN) are predicated as they don't have disease. False positives (FP) are predicted as Alzheimer's affected but
they don't actually have the disease. False negatives (FN) were predicted as no, but the actually do have the
Alzheimer's. Table 1 shows the confusion matrix of first level decomposition.

Table 1: DWT First level decomposition confusion matrix

Euclidean distance Correlation Cosine Cityblock
N=249 Predicted Predicted Predicted Predicted Predicted Predicted Predicted Predicted
No Yes No Yes No Yes No Yes
Actual 130 76 138 68 143 63 137 69
No
Actual 17 26 13 30 11 32 15 28
Yes

Table 2: DWT Second level decomposition confusion matrix

Euclidean distance Correlation Cosine Cityblock
N=249 Predicted Predicted Predicted Predicted Predicted Predicted Predicted Predicted
No Yes No Yes No Yes No Yes
Actual 167 39 152 54 148 58 143 63
No
Actual 7 36 13 30 10 33 15 28
Yes

Table 3 shows list of performance parameters described in Tables 1 and 2.

Table 3: Performance of the proposed method

Euclidean Correlation Cosine Cityblock

distance
Sl. No. Parameter Equation
1st 2nd 1st 2nd 1st 2nd 1st 2nd
level level level level level level level level
1 Accuracy (TP+TN)/Total 62.7 85.5 68.5 75.3 71.9 74.8 67 70
2 Misclassification Rate (FP +FN)/ Total 32.1 9.3 26.3 19.5 22.9 20 27.7 24.8
3 True Positive Rate TP/ Actual Yes 0.6 0.83 0.69 0.69 0.74 0.76 0.65 0.65
4 False Positive Rate FP / Actual No 0.36 0.18 0.33 0.26 0.30 0.28 0.33 0.30
5 Specificity TN / Actaul No 0.63 0.81 0.67 0.73 0.69 0.71 0.66 0.69
6 Precision TP / Predicted Yes 0.25 0.48 0.30 0.35 0.33 0.36 0.28 0.28

4. Conclusions
The method for computer aided Alzheimer's disease is reported for processing the MR images of the brain. In
this we have done MR image pre-processing and skull stripping as a pre-processing step. Discrete wavelet
transform is used to decompose the image into two levels. In each level of decomposed image, GLCM
parameters were calculated and mean value is estimated. The KNN classifier is used for classification with
different distance parameters (Euclidean distance, Correlation, Cosine distance and city block distance).
Euclidean distance with second level decomposition gives better results compared to other methods.

References
[1] Fang C, Li C, Cabrerizo M, Barreto A, Andrian J, Loewenstein D, Duara R, Adjouadi M. A Novel
Gaussian Discriminant Analysis-based Computer Aided Diagnosis System for Screening Different Stages of
Alzheimer. In2017 IEEE 17th International Conference on Bioinformatics and Bioengineering (BIBE) 2017
382
 Proceedings of NCAMMM - 2018

Oct 1 (pp. 279-284). IEEE.

[2] Tung, James, et al. "Everyday Patient-Care Technologies for Alzheimer's Disease." IEEE Pervasive
Computing 12.4 (2013): 80-83.
[3] Zhou Q, Goryawala M, Cabrerizo M, Wang J, Barker W, Loewenstein DA, Duara R, Adjouadi M. An
optimal decisional space for the classification of Alzheimer's disease and mild cognitive impairment. IEEE
Transactions on Biomedical Engineering. 2014 Aug;61(8):2245-53.
[4] Cuingnet R, Gerardin E, Tessieras J, Auzias G, Lehéricy S, Habert MO, Chupin M, Benali H, Colliot O,
Alzheimer's Disease Neuroimaging Initiative. Automatic classification of patients with Alzheimer's disease
from structural MRI: a comparison of ten methods using the ADNI database. neuroimage. 2011 May
15;56(2):766-81.
[5] Martinez-Murcia FJ, Górriz JM, Ramírez J, Ortiz A. A structural parametrization of the brain using
hidden Markov models-based paths in Alzheimer’s disease. International journal of neural systems. 2016
Nov;26(07):1650024.
[6] Krashenyi I, Popov A, Ramirez J, Gorriz JM. Fuzzy classification of Alzheimer's disease using statistical
moments. InElectronics and Nanotechnology (ELNANO), 2015 IEEE 35th International Conference on 2015
Apr 21 (pp. 409-412). IEEE.
[7] Wu X, Li Q, Xu L, Chen K, Yao L. Multi-feature kernel discriminant dictionary learning for face
recognition. Pattern Recognition. 2017 Jun 30;66:404-11.
[8] Dolph CV, Alam M, Shboul Z, Samad MD, Iftekharuddin KM. Deep learning of texture and structural
features for multiclass Alzheimer's disease classification. InNeural Networks (IJCNN), 2017 International
Joint Conference on 2017 May 14 (pp. 2259-2266). IEEE.
[9] http://adni.loni.usc.edu/
[10] Zuiderveld K. Contrast limited adaptive histogram equalization. InGraphics gems IV 1994 Aug 1 (pp.
474-485). Academic Press Professional, Inc..
[11] Aruchamy, Srinivasan, Ravi Kant Kumar, Partha Bhattacharjee, and Goutam Sanyal. "Automated skull
stripping in brain MR images." In Computing for Sustainable Global Development (INDIACom), 2016 3rd
International Conference on, pp. 2043-2047. IEEE, 2016.
[12] Tadashi Araki, Nobutaka Ikeda, Devarshi Shukla, Pankaj K. Jain,Narendra D. Londhe, Vimal K.
Shrivastava, Sumit K. Banchhor,Luca Saba, Andrew Nicolaides, ShoaibShafique, John R. Laird and Jasjit S.
Suri. PCA-based polling strategy in machine learning framework for coronary artery disease risk assessment
in intravascular ultrasound: A link between carotid and coronary grayscale plaque morphology. Elsevier
Journal 16 February 2016.
[13] http://www.oasis-brains.org/
[14] http://knightadrc.wustl.edu/cdr/cdr.htm
[15] https://en.wikipedia.org/wiki/Mini%E2%80%93Mental_State_Examination
[16] Danielsson PE. Euclidean distance mapping. Computer Graphics and image processing. 1980 Nov
1;14(3):227-48.
[17] Sotiriou C, Neo SY, McShane LM, Korn EL, Long PM, Jazaeri A, Martiat P, Fox SB, Harris AL, Liu ET.
Breast cancer classification and prognosis based on gene expression profiles from a population-based study.
Proceedings of the National Academy of Sciences. 2003 Sep 2;100(18):10393-8.
[18] Liao H, Xu Z. Approaches to manage hesitant fuzzy linguistic information based on the cosine distance

383
Brain MR image analysis using discrete wavelet transform with GLCM feature analysis

and similarity measures for HFLTSs and their application in qualitative decision making. Expert Systems with
Applications. 2015 Jul 15; 42(12):5328-36.
[19] de Souza RM, De Carvalho FD. Clustering of interval data based on city–block distances. Pattern
Recognition Letters. 2004 Feb 29;25(3):353-65.

384