MATERIALS AND MANUFACTURING PROCESSES

Vol. 19. No. 3, pp. 475-486, 2004

Effect of Mechanical Alloying

and Sintering on Ni-Ti Powders

S. K. Sadrnezhaad* and A. R. Selahi

Department of Materials Science and Engineering,

Sharif University of Technology, Tehran, Iran

ABSTRACT

Commercially pure nickel-titanium powders were mechanically milled in a ver-

tical attritor mill under protective atmosphere for various times from 10 to 24 hr.
Products were then compacted and sintered at different temperatures for different
times. Amorphization and interatomic phase formation were determined
by X-ray diffractometry, scanning electron microscopy and differential scanning
calorimetry. Porosity, virtual density, transition temperatures and the amount of
Ni3Ti first increased and then decreased with the milling time. Presence of oxygen
in the milling atmosphere showed partial crystallization of NiTi intermetallic
compound accompanied with titanium oxide formation.

Key Words: Mechanical alloying; Ni-Ti; Powder; Amorphization;, Intermetallic

compounds; Sintering; Nitinol; Milling.

*Correspondence: S. K. Sadrnezhaad, Ph. D., Department of Materials Science and

Engineering, Sharif University of Technology, Tehran, 11365, Iran; Fax: +9821-6005717;
E-mail: sadrnezh@sharif.edu or sarnezh@yahoo.com.

475

DOI: 10.1081/AMP-120038656 1042-6914 (Print); 1532-2475 (Online)

Copyright © 2004 by Marcel Dekker, Inc. www. dekker. com
476 Sadrnezhaad and Selahi

1. INTRODUCTION
Mechanical alloying was first developed by Benjamin and Volin in 1974.[1] It
served as a means of combining the advantages of gamma prime precipitation
hardening for intermediate strength and oxide dispersion strengthening in a nickel-
base high temperature super alloy. This was also used by Batezzati et al.[2] in1988 to
prepare mechanically alloyed Ni-Ti powders. Mechanical alloying is now known as
an alternative way for production of nanocrystalline intermetallic materials.[3,4]
It is well known that, besides the three stable intermetallic phases (NiTi2 , NiTi,
and Ni3Ti), several other metastable phases can form during solidification of Ni-Ti
equiatomic alloy.[5] Ageing supersaturated Ni-rich alloy can cause the segregation of
Ni4Ti3 and Ni3Ti2 phases before the formation of Ni3Ti,[6-9] whereas NiTi—whose
high-temperature crystal structure is B2 (CsCl type) with a lattice parameter of about
0.3 nm—reveals the shape memory effect generally referred to as SME. Annealing
followed by cold working and aging can result in formation of a rhombohedral
phase before martensitic transformation.[10,11]
The SME is a thermoelastic process that can be induced by cooling of the B2
phase to a temperature below Ms or by application of an externally applied
stress.[12,13] During cooling, the so-called parent phase undergoes a diffusionless
transformation to a martensitic product with a triclinic distortion of the B19 type [14]
lattice or a monoclinic type[15,16] crystal structure. In Ni-rich alloys, the Ni4Ti3
precipitate is the closest phase that can induce memory effect.[8,11] This can be
achieved by annealing of the alloy at 500ºC or less.[17-19]
Although considerable achievements have been obtained on Ni-Ti Shape
memory alloy (SMA) during past years, there are still controversies on how to
precisely control the SMA characteristics. [20] An important issue is the choice of
alternative production methods as they affect the chemical homogeneity, the
porosity, and the contaminants added to the melt. It is, therefore, agreed upon that,
although the conventional methods like electric arc melting, vacuum induction
melting or combustion synthesis [21] are suitable for production of Ni-Ti alloys,
powder metallurgy must also be considered as a beneficial way of production of the
near-net-shape, fine-grain material that can be applicable to the aerospace,
automotive, or biomedical industry. [22]
The formation of B2 phase in NiTi is very sensitive to the chemical composi-
tion of the alloy. Melting methods have shortcomings due to (1) gas absorption,
(2) elemental vaporization, (3) segregate formation, and (4) crucible contaminate
absorption. [23]
New powder-metallurgical techniques such as self-propagation high-
temperature synthesis, liquid-phase sintering, explosive shock wave compression,
and mechanical alloying have more recently been reported for production of NiTi
from pure elemental powders of nickel and titanium.[23-27] Each method has its own
advantages, disadvantages and need to further development.
Mechanical alloying is a complex process comprised of fragmentation,
deformation, cold welding, and short range diffusion occurring within a layer of
[28]
powder particles trapped between surfaces of two colliding balls. These events
occur in a high-energy ball mill, which is usually one of four common devices. One is
a vertical ball mill, such as attritor mill, in which the balls and the metal powders
Effect of Mechanical Alloying and Sintering on Ni-Ti Powders 477

are charged into a stationary vertical tank and are agitated by impellers radiating
[28,29]
from a central rotating shaft. A Second configuration is the vibratory mill
exemplified by the SPEX shaker mill. The third is a conventional horizontal ball
[29]
mill, and the last is the planetary ball mill.
In this study, a modified attritor mill and a planetary ball mill were used for
mechanical alloying of the Ni-Ti powders. The products were then compacted and
sintered to make disklike tablets. The properties of the tablets were determined and
compared with those produced via other alternative routes.

2. EXPERIMENTAL PROCEDURE

Commercially pure powders of Ni (99.9%) with an average particle size of about

5 µm and Ti (99.9%) with an average particle size of about 30 µm were mixed
equiatomically and milled in a vertical attritor mill working with chromium steel
balls of either 7.5 or 4.8 mm diameters. To increase the efficiency of the process, the
[29,30]
dead zone usually formed at the bottom of the mill was eliminated by addition
[31]
of an extra impeller to the end of the shaft. The milling process continued for up
to 24 hr at two rotational speeds of 350 and 710 rpm under argon atmosphere.
Ten grams of mixed powders were first milled in separate stages, and X-ray
samples were taken at the end of each stage. During milling, the materials were kept
cold by holding the external surface of the container isothermal so that the
nonisothermal processes could be avoided. Twenty-two grams of a similar powder
were then poured into the vial and milled for another 10 hr separated into five 2-hr
intervals. An X-ray map was taken after 20 hr of the final product. In a separate
experiment, 20 g of mixed powders were milled at the same conditions but with a
higher milling rate. The samples were characterized by cobalt radiation
( λ = 1.7903 Å ) of a Philips diffractometer.
The mechanically alloyed powders were compacted at 500 MPa using a single
action press in a floating die and then annealed at 1000 ° C for 4 hr. Their virtual
density was measured by means of floatation method, and an image analyzer
determined their porosity. The cross-section of the samples was metallographically
analyzed and their matrix composition was determined by Electron Probe Micro–
Analyzer (EPMA). The transition temperatures of the samples were measured by a
differential scanning calorimetry (DSC) analyzer, which heated the samples up to
200°C and cooled them down to the room temperature with a heating/cooling rate of
5°C per minute.

3. Results and Discussion

Figure 1 illustrates the micrographs of the pure powders used in these studies.
Figure 2 shows the dot map patterns of a typical particle from within a sample that
was milled for 20 hr. The distribution of nickel and titanium in the particle is a
measure of the progress of the alloying process. A rough estimation implies that the
particle consists of around 35 to 40 % nickel and 60 to 65 % titanium, indicating that
the diffusion process has partially occurred due to the milling operation.
478 Sadrnezhaad and Selahi

(a) (b)
Figure 1. Micrographs of pure powders used to make samples: (a) Ni and (b) Ti. (View this
art in color at www.dekker.com.)

Both grain boundary and volume diffusion can be responsible for the alloying
process. It has been proven earlier that the formation of high diffusivity paths, such
as dislocations during mechanical alloying, increases the rate of inter-diffusion of the
alloying elements.[32] Observations show that the amounts of the diffused materials
will increase by further milling of the sample. An increase in the milling speed, as well
as the ball to powder ratio, up to 10:1 will also increase the diffusion rate.
X-ray diffraction (XRD) patterns indicate a typical broad peak of the
amorphous phase that appears after 10 hr of milling at 350 rpm (Fig. 3). The
reason for interchange of the principal titanium peaks after 8 hr of milling, which is
observable in Fig. 3 is not clear to us. A probable cause might be the formation of
preferred orientations due to cold working and fracturing of the titanium grains
accompanying the accumulation of the dislocations during milling.
Figure 4 clearly indicates the formation of B19' (martensite) phase from B2
(austenite) in the sintered samples. Table 1 summarizes the transition temperatures
of the samples measured after sintering. It is shown that an increase of around 40°C
in the transition temperatures occurs due to up to 16 hr milling, whereas about 20°C
decrease occurs due to an additional 2 hr milling. These changes are consistent with
the ones obtained from comparison of the peaks depicted in Fig. 4.
To understand the reason, one should notice the effect of grain size on stability
of the martensite phase. According to the previous investigations, as the grain size
decreases, martensite will form easier and would thus be in a more stable
condition.[33] Refinement of the grain size causes, therefore, increasing of the
transition temperatures. Further milling, however, results in storage of an excess
energy in the powders. This causes the decreasing of the virtual temperatures existing
for recovery, recrystallization, and grain growth, resulting in reduction of the
transformation temperatures illustrated in Table 1. Every mechanism that has
a stronger effect causes a greater change in the transformation temperatures of the
samples.
Figure 5 shows the variation of the density of the sintered samples with the
milling time. As is seen, the density increases with the milling time. This is due to the
negative molar volume change for formation of TiNi from its pure constituents:
Effect of Mechanical Alloying and Sintering on Ni-Ti Powders 479

Figure 2. X-ray maps of a typical powder particle milled for 20 hr: (a) a powder particle,
(b) distribution of Ti inside the particle, and (c) distribution of Ni inside the particle.
480 Sadrnezhaad and Selahi

a b

Figure 3. XRD patterns for Ni-Ti mixture after milling in argon for (a) 2, 4, 6, 8 and 10 hr
and (b) 12, 14, 16, 18, and 20 hr.

Figure 4. XRD patterns of sintered Ni-Ti powders milled in argon gas for (a) 12, (b) 14,
(c) 16, (d) 18, and (e) 20 hr.

Table 1. Effect of milling time on the transition temperatures of the samples made by
sintering of the mechanically alloyed powders.
________________________________________________________________________________
Milling Time (hr) As (ºC) Af (ºC) Ms (ºC) Mf (ºC)

12 106.5 129 119 100

14 142.5 164 148 126
16 152.0 164 150 142
18 130.0 146 135 121
Effect of Mechanical Alloying and Sintering on Ni-Ti Powders 481

Figure 5. The change of density with the milling time.

Figure 6. The change of porosity of the sintered sample with the milling time.

∆VNiTi = − 0.71 cm3/mol

Figure 5 indicates that after milling for 18 hr, the density of the specimen starts
to decrease. The reason for this decreasing effect is not obvious to us, but its
occurrence is assessed with further experiments. Chemical reactions accompanied by
exchange of different phases of highly reactive unstable nature during prolonged
mechanical alloying may be responsible for this effect.
Variation of the porosity of the sintered sample with the milling time, shown in
Fig. 6, indicates an unusual trend between 10 and 12 hr. It first increases and then
decreases with the milling time. To understand the reason, one can pay attention to
the major sources of porosity in the sample:
• Initial porosity of the specimen before sintering.[34]
• Porosities produced due to the faster diffusion of nickel in titanium.[35]
• Shrinkage pores that form because of the solidification of the liquid phase formed
during sintering.[36]
• Expansion, which occurs because of the internal combustion of Ni with Ti to
form NiTi during sintering.[37]
Mechanical alloying, of coarse, reduced the effects of the second and the forth
source by application of the milling energy to the specimens. The presence of oxygen,
482 Sadrnezhaad and Selahi

even at the smallest level, in the milling atmosphere showed, however, that partial
crystallization of the NiTi intermetallic compound was accompanied by titanium
oxide formation. Gas absorption, partial oxidation, and clustering of the powders
were accompanied with great amounts of heat evolution. This heat was constantly
removed to keep the initial isothermal milling conditions.
Figure 7 shows the microstructure of the sintered specimens containing two
kinds of precipitates: (1) one with a round-shaped edge, and (2) another with a
sharp-edged shape. The chemical analyses of both kinds and the matrix were

a b

c d

(e)

Figure 7. Microstructure of samples made of powders milled for (a) 10 hr, (b) 12 hr, (c) 14 hr,
(d) 16 hr, and (e) 18 hr. (View this art in color at www.dekker.com.)
Effect of Mechanical Alloying and Sintering on Ni-Ti Powders 483

Table 2. Chemical composition of precipitates and the matrix in a sintered specimen.

_____________________________________________________________________________
Shape Ni atom% Ti atom% Compound

Sharp Edge 75.60 27.20 Ni3Ti

Round 57.76 42.57 Ni3 Ti2
Matrix 49.36 50.69 NiTi
_____________________________________________________________________________

Figure 8. Variation of the amount of unstable precipitate Ni3Ti with the milling time.

35 40 45 50 55 60
2θ

Figure 9. XRD pattern of Ni-Ti mixture milled for 24 hr under atmosphere and then
annealed for 1 hr at 920 ºC under purified argon.

determined by EPMA. The results indicated the formation of two compounds; Ni3Ti
and Ni3Ti2 (Table 2); both being considered as common precipitates in the NiTi
alloying process. It was seen that the milling time changed the amount of the
intermetallic phases (Fig. 8), but the amount of Ni3Ti2 was so small that it could not
be measured with image analyzer.
Figure 8 shows that the amount of the intermetallic compound Ni3Ti increases
with the milling time to a maximum value of about 4% that occurs at t = 12 hr. This
quantity then decreases to values lower than 0.5%, with the milling time exceeding
16 hr. Difference in the diffusion coefficients of the nickel and titanium atoms, as
well as their particle size, seems to be responsible for these changes. Further,
interatomic diffusion of the constituent elements allows the mixture to approach the
484 Sadrnezhaad and Selahi

stoichiometric NiTi chemical composition and results in reduction of the amount of

the metastable Ni3Ti phase. Sintering will also affect the change of the composition
of Ni3Ti towards NiTi. Small amount of meta-stable compound Ni3Ti2 can also form
in the samples milled especially for lower periods (10-12 hr). The presence of the
Ni3Ti and Ni3Ti2 compounds has adverse effects on high-temperature workability of
the specimens,[38] and their elimination would improve the hot rolling characteristics
of the mechanically alloyed specimens.
Experimental studies showed that the milling under atmosphere resulted in the
formation of TiO, Ti 2 O3 , and TiO2 phases; whereas TiO and Ti2 O3 were unstable
and diminished with the milling time. Enthalpies of the oxidation reactions
facilitated the formation of the intermetallic NiTi crystalline compound. Figure 9
illustrates the relative amounts of the phases produced after 24 hr milling under
atmospheric conditions followed by 1 hr annealing at 920°C under purified argon.

4. CONCLUSIONS

1. The amounts of amorphization and diffusion are functions of the milling

time, milling speed, and the charge ratio.
2. The porosity percentage will decrease with increasing of the milling time.
3. The amounts of the intermetallic compounds generally decrease with
increasing of the milling time.
4. Two opposite parameter will affect the transition temperatures: (1) the
refinement of the powder particles, and (2) the stored energy in powders that
cause faster recovery, recrystallization, and grain growth.

5. ACKNOWLEDGMENTS

The authors want to acknowledge the financial support of Sharif University of

Technology. They also want to thank Mr. Daneshmaslak and Mr. Ahmadiyan for
their assistance in milling and XRD tests.

6. REFERENCES

1. Benjamin, J.S.; Volin, T.E. The mechanical alloying. Metallurgical Transactions

1974, 5, 1929−1934.
2. Battezati, L.; Cocco, G.; Schiffini, L.; Enzo, S. Thermal properties of
mechanically alloyed Ni50Ti50 powders. Journal of Materials Science and
Engineering 1988, 97, 121−124.
3. Stoloff, N.S.; Liu C. T.; Deevi, S.C. Emerging applications of intermetallics.
Intermetallics 2000, 8, 1313−1320.
4. Westbrook, J.H.; Fleischer, R. L. Intermetallic Compound: Structural
Applications; John Wiley and Sons: New York, 2000; 17−47.
Effect of Mechanical Alloying and Sintering on Ni-Ti Powders 485

5. Massalski, T. Binary Alloy Phase Diagrams; Murray, J.L., Bennett, L.H.,

Baker, H., Eds.; ASM, Metals Park: OH, 1986; 2847.
6. Ishida, A.; Sato, M.; Takei, A.; Nomura, K.; Miyazaki, S. Effect of aging on
Shape memory behavior of Ti-51.3 at pct Ni thin films. Metallurgical and
Materials Transactions A 1996, 27A, 3753−3759.
7. Lopez, H.F.; Salinas-Rodriguez, A.; Rodriguez-Galicia, J.L. Microstructural
aspects of precipitation in a Ti-rich Ni−Ti alloy. Scripta Materialia 1996, 34 (4),
659−664.
8. Comstock, R.J.; Buchheit, T.E.; Somerday, M.; Wert, J. A. Modeling the
transactions stress of constrained shape memory alloy single crystals. Acta
Materialia 1996, 44 (9), 3505−3514.
9. Miyazaki, S.; Kurooka, S.; Ishida, A.; Nishida, M. Effect of heat-treatment on
deformation behavior associated with R-phase and martensitic transformations
in Ti-Ni thin films. Transactions of the Materials Research Society of Japan 1994,
18B, 1041−1044.
10. Miyazaki, S.; Otsuka, K.; Wayman, K. Morphogical changes associated with
the R-phase and martensitic transformations in Ti-Ni single crystals. ISIJ
International 1989, 29 (5), 423−429.
11. Nishida, M.; Honma, T. All-round shape memory effect In Ni-rich TiNi alloys
generated by constrained aging. Scripta. Metallurgica 1984, 18, 1295−1298.
12. Mellor, B.G.; Guilemany, J.M. Stabilized stress induced martensite―its
application in two way shape memory training processes. Scripta Metallurgica
et Materialia 1990, 24, 241−245.
13. Okamoto, Y.; Hamanaka, H.; Miura, F.; Tamura, H.; Horikawa, H. Reversible
changes in yield stress and transformation temperature of a NiTi alloy by
alternate heat treatments. Scripta Metallurgica 1988, 22, 517−520.
14. Wayman, C.M.; Duering, T.W. An Introduction to Martensite and Shape
Memory, Engineering Aspects of Shape Memory Alloys, Shape Memory Alloy
Technology Conference, Michigan State University, August 15−17, 1988;
Butterworth-Heinemann: London, 1990; 3−19.
15. Tang, Weijia. Thermodynamic Study of the low-temperature phase B19 and the
martensitic transformation in near-equiatomic Ti-Ni shape memory alloys.
Metallurgical Transactions A 1997, 23A, 537−544.
16. Otsuka, K.; Ren, X. Martensitic Transformations in nonferrous shape memory
alloys. Materials Science and Engineering 1999, 30A, 89−105.
17. Todoroki, T.; Tamura. H. Effect of heat treatment after cold working on the
phase transformation in TiNi alloy. Transactions of Japan Institute of Metals
1987, 28 (2), 83−94.
18. Honma, T.; Nishida, M.; Yamauchi, K. Ti−Ni Alloy Articles Having a
Property of Reversible Shape Memory and a Method of Making the Same. US
Patent 4,707,196, November 1987.
19. Zarandi, F.M.H.; Sadrnezhaad, K. Thermomechanical study of combustion
synthesized Ti-Ni shape memory alloy. Materials and Manufacturing Processes
1997, 12 (6), 1093−1105.
20. Sadrnezhaad, K.; Mashhadi, F.; Sharghi, R. Heat treatment of Ni-Ti alloy for
improvement of shape memory effect. Materials and Manufacturing Processes
1997, 12 (1), 107−115.
486 Sadrnezhaad and Selahi

21. Igharo, M.; Wood, J.V. Compaction and sintering phenomena in Titanium-
Nickel shape memory alloys. Powder Metallurgy 1985, 28 (3), 131−139.
22. Takasaki, A. Mechanical alloying of the Ni-Ti system. Physic Stat. Sol. 1998,
169, 183−191.
23. Rozner, A.G.; Heintzelman, E.F.; Buehler, W.J.; Gilfrich, J.V. Effect of
addition of oxygen, nitrogen and hydrogen on microstructure and hardness of
cast TiNi intermetallic compound. Transactions of the ASM 1965, 58, 415−418.
24. Green, S.M.; Grant, D.M.; Kelly, N.R. Powder metallurgical processing of
Ni−Ti shape memory alloy. Powder Metallurgy 1997, 40 (1), 43−47.
25. Igharo, M.; Wood, J.V. Compaction and sintering phenomena in Titanium-
Nickel shape memory alloys. Powder Metallurgy 1985, 28 (3), 131−139.
26. Li, Bing-Yun; Rong, Li-Jian; Li, Yi-Yi; Gjunter, V. E. A recent development in
producing porous Ni–Ti shape memory alloys. Intermetallics 2000, 8, 881−884.
27. Chank, Y.; Berger, S.; Weiss Brook-Levinson, B. Z. Solid state amorphization
by mechanical alloying–an atomistic model. Acta Materialia 1994, 42 (11),
3679−3685.
28. Rydin, R.W.; Maurice, D.; Courtney, T.H. Milling Dynamics: Part I. Attritor
dynamics: Results of a cinematographic study. Metallurgical Transactions A
1993, 24A, 175−185.
29. Cook, T.M.; Courtney, T.H. The effect of ball size distribution on attritor
efficiency, Metallurgical and Materials Transactions A 1995, 26A, 2389−2397.
30. Wang, K.Y.; Shen, T.D.; Wang, J.T.; Quan, M.X. Characteristics of the
mechanically alloyed Ni 60 Ti 40 amorphous powders during mechanical
milling in different atmospheres. Journal of Materials Science 1993, 28,
6474−6478.
31. Maurice, D.; Courtney, T.H. Milling dynamics: part II, Dynamics of a SPEX
mill and a one dimensional milling. Metallurgical and Materials Transactions A
1996, 27A, 1973−1979.
32. Pabi, S.K.; Das, D.; Mahapatra, T.K.; Manna, I. Mathematical Modeling of
the mechanical alloying kinetics. Acta Materialia 1998, 46 (10), 3501−3510.
33. Cohen, M. Martensitic nucleation―revisited. Materials Transactions Journal,
1992, 33 (3), 178−181.
34. Gordon, D. Powder Metallurgy, the Process and Its Products; McGraw Hill
Publisher, 1990.
35. Bastin, G.F.; Rieck, G.D. Diffusion in Ni-Ti systems: Part I. Metallurgical
Transactions A 1974, 5A, 1817−1826.
36. Germani, R.M. Phase diagram in liquid phase sintering treatments. JOM
1986, 26−29.
37. Mirabolghasemi, S.H.; Selahi, A.; Haeri, E.; Sadrnezhaad, S.K. Thirteenth
Iranian Foundrymen Society Conference, Tehran, Iran, 2001, 237−244; [In
Persian].