Accepted Manuscript

Title: Polishing of fused deposition modeling products by hot

air jet: Evaluation of surface roughness

Authors: Mohamed Adel, Osama Abdelaal, Abdelrasoul Gad,

Abu Bakr Nasr, AboelMakaram Khalil

PII: S0924-0136(17)30291-1
DOI: http://dx.doi.org/doi:10.1016/j.jmatprotec.2017.07.019
Reference: PROTEC 15316

To appear in: Journal of Materials Processing Technology

Received date: 31-1-2017

Revised date: 16-5-2017
Accepted date: 13-7-2017

Please cite this article as: Adel, Mohamed, Abdelaal, Osama, Gad, Abdelrasoul, Nasr,
Abu Bakr, Khalil, AboelMakaram, Polishing of fused deposition modeling products by
hot air jet: Evaluation of surface roughness.Journal of Materials Processing Technology
http://dx.doi.org/10.1016/j.jmatprotec.2017.07.019

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Polishing of fused deposition modeling products by
hot air jet: Evaluation of surface roughness

Mohamed Adela,*, Osama Abdelaala, Abdelrasoul Gada, Abu Bakr Nasra,

AboelMakaram Khalila

a
Department of Mechanical Engineering, Assiut University, Assiut, 71515,
Egypt

*
The corresponding author:
E-mail address: m.a.abdelhakeem@aun.edu.eg, Tel: +201007379579, Fax:
+20 88 2335572

Graphical abstract:

1
Abstract:
In this study, a novel low-cost polishing process is applied on the surface of Fused
Deposition Modeling (FDM) products. The developed polishing technique impinges
a jet of hot air exit from a nozzle to FDM surfaces. The hot air locally melts the
staircase on the surface and leaves it smoother by the effect of sintering
phenomenon. Accordingly, the process introduces three main parameters: air jet
temperature; air jet velocity; and nozzle translational velocity over the part surface.
An experimental test rig was constructed to evaluate the polishing process and its
parameters using surfaces with average pre-processed Ra values of 7.5∓0.5 μm. The
process shows significant and reproducible improvements in surface roughness
inherent with a glossy surface; whereas, an average reduction ratio up to 88% was
reached which corresponds to Ra of 0.85 μm. It was found that there is an allowable
range of nozzle translational velocity for every combination between jet velocity
and jet temperature; otherwise, lower nozzle velocity than allowable causes
overheating and surface deterioration. Furthermore, the study presents in-depth
investigation to these deterioration phenomena appeared on the surface. As a result,
this investigation demonstrated the possible defects in FDM part surfaces and also
evaluated different process parameters. Moreover, it was observed that surface
defects are reduced in the polished surfaces. For a concise conclusion, it was found
that the condition of low jet velocity and high jet temperature gives the best
polishing result over the allowable nozzle velocities.

Keywords: fused deposition modeling - polishing - hot air- Surface Roughness -

Surface defects – sintering

1. Introduction:

Currently, Additive manufacturing (AM) technologies become not only a rapid

prototyping technique but also a way of batch production of complex parts. The
technology reduces the lengthy steps of traditional manufacturing and presents a

2
progress toward rapid manufacturing. One of the most common techniques in AM is
fused deposition modeling (FDM) owing to its simplicity and low manufacturing
cost. The technology deals with wide range of polymer-based materials and also
non-polymer materials like ceramics (Bellini et al., 2005) or metals (Masood and
Song, 2004). FDM parts are manufactured by extruding a semi-molten filament
material through an automatically controlled nozzle in the machine head; the nozzle
fuses and extrudes the material while the machine head is moving to deposit the
layers of the part. Figure1 presents the manufacturing technique and its main
components. This layer-based manufacturing yields stair-like surface in all AM
techniques. Hence, AM techniques and specially FDM (Pal and Ravi, 2007) are
characterized by higher surface roughness comparing with traditional
manufacturing techniques.

Studies divided surface roughness treatment of FDM parts into two categories: pre-
manufacturing methods and post-processing methods. In pre- manufacturing
methods: they determine the optimum design parameters of the manufacturing
process like layer thickness, part orientation, raster angle/width, etc. Practically, The
influence of pre-manufacturing methods depends much on the part shape and may
result to rough surface which requires post-processing.

Various studies attempt to solve the roughness problem by applying post-processing

methods. Many of these studies employed traditional polishing techniques which
generally depend on removing material from the surface. Other studies suggested
adding material to the surface so, they may be considered as addition techniques.

Abrasive polishing techniques as sand paper are more common between different
polishing techniques. The technique is simple, low cost, and can be used with other
techniques. Ahn and Lee, 2007 grinded Stereolithography (SL) part surfaces with
sand paper after coating them with paraffin wax. Recently, Boschetto and Bottini,
2015 employed barrel finishing to FDM parts, but the technique may need hours to
smoothen FDM parts in a rotating barrel contains abrasive balls. Fluidic jet
polishing is another polishing technique using fluids combined with abrasive
3
particles that hit the surface to erode its staircase. Leong et al., 1998 tested abrasive
jet polishing at jeweler parts manufactured by SL. The process directs a jet of fluid
containing abrasive media to the polished surface to remove its staircase. Also,
Williams and Melton, 1998 tested abrasive flow finishing using a similar concept to
abrasive jet while the fluid flows in a passage that one of its surfaces is the part to
be polished. Generally, Fluid jet polishing is quite simple but needs preparation and
proper tuning to process parameters including processing time, pressure, and grit
size otherwise deterioration may occur to the surface. Galantucci et al., 2009 tried
chemical polishing where an ABS part surfaces were treated with acetone as a
chemical solvent. The optimum parameters were an immersing time of 300 sec at
solvent concentration of 10% acetone and 90% water but the process cannot control
enough the material removal rate at different part surfaces.

Studies presented in addition techniques include coating methods like infiltration,

electroplating, and also painting. Electroplating and painting are usually employed
respectively to change surface characteristics or improve appearance. Ahn et al.,
2004 used infiltration, grinding and heating to provide translucent FDM parts.
Zinniel, 2011 developed a vapor smoothening method to ABS parts using the vapor
of acetone. The technique deposits the vapor of acetone which is a strong etching
solvent to ABS material. This technique improves significantly the Ra values of
ABS parts but it generally needs preparations and use of proper etching solvent if
available to different FDM materials.

Unfortunately, material removal or material addition techniques introduce

dimension change in the surface and mostly require abrasion tools, forces, materials
or/and time. The present study developed a new polishing process that is neither
abrasive nor additive and it uses a low-cost finishing technique.

2. The developed polishing Process:

The general concept of the developed process is to locally melts the surface of FDM

4
parts using a jet of hot air. Surface melting will fuse the staircase (the stair tips) in
the surface of FDM parts and these activates the sintering phenomenon (Hamidi et
al., 2016); this phenomenon promotes coalescence to the staircase found on the
surface and leaving it smoother. Figure 2 presents schematic drawings to the

proposed polishing process. To melt the surface, the heated jet temperature is
sustained at higher temperature than the melting temperature of part material. For
continuous processing operation, the air nozzle must pump the heated air over part
surface with appropriate translational velocity; this translational velocity should be
lower enough to heat the surface till melting.

The phenomenon of roughness improvement is that: when hot air melts the staircase
in the surface; it activates a peculiar phenomenon characterized with sintering. This
phenomenon is spread in several application including AM techniques and primary
used in selective laser sintering and rotational molding process. Figure 3 presents
schematics to the sintering phenomenon. When two particles are brought into
contact under ambient temperature near melting; the two particles coalesces and

removes the void in-between during a step known as densification. The Heating up
(of sintered particles) reduces the viscosity of material and trigger surface tension
force to re-shapes the two particles against the resistance of viscous flow forces.
Figure 4 shows surfaces polished experimentally with hot air.

The developed process is closely similar to laser polishing process where the laser
light is the heating source instead of hot air. Laser polishing technique is
characterized by high reduction ratios of surface roughness and often used with hard
materials ( like diamond or titanium alloys) where abrasive tools are no longer
appropriate (Bordatchev, 2014). Also, laser polishing is rarely used with polymer
materials probably owing to its relative cost.

5
The advantage of using hot air polishing instead of laser (if laser is applicable to
FDM parts) is that:
(1) The forces promote sintering/smoothing of the surface are not only surface
tension force but also there are flow forces: stagnation pressure/force and shear flow
forces. (Jones et al., 2015; Löhner and Drummer, 2015) and others state the
importance of ambient pressure in sintering rate and products quality (porosity, and
mechanical strength). (2) the laser system may introduce high initial cost compared
with the price of FDM parts itself while the developed process is more simple and
less cost. (3) For possible manual application, finishing by hot air is safer than laser.
(4) The developed technique also uses low jet velocities and hence, consumes a
moderate amount of energy – for a quite efficient heater- to pump and heat it.

3. Experimental setup:

Figure 5a shows a schematic diagram of the used test rig. The test rig pumps a
filtered air with a controlled flow rate and temperature on a moving FDM specimen
-and the air nozzle is fixed- with a controlled translational velocity. This was
accomplished through a fresh air exits from the compressor tank and then passes
through an air filter and later through a pressure regulator. After that, the filtered air
passes through a flow controller and then to a heater and finally exits from the air
nozzle to the surface to be processed. Figure 5b shows the experimental set up of
the test rig. The shown 3-D printer provided the specimen with a controlled
translational velocity –up or down direction- relative to the fixed air nozzle. the
frictionless x-y suspension - placed in the top of the printer - hangs the heater, other
attached components, and provides flexibility to air nozzle to move to any point in
the x-y plane. Also, a small manufactured vice was used for fixing and aligning the
test specimen relative to the air jet.

The used flow meter (Alicat, USA) is an industrial type mass/volume flow
meter/controller. It is a venturi type with a differential digital pressure sensor and
has a temperature sensor to compensate the change in viscosity and density of the
treated air. Its function is to read/control the flow rate of the used air.
6
The used air heater's body is made of a high heat-resistance transparent glass for
easy troubleshooting and it used a nickel-chrome coil which was connected to an
industrial on/off controller using a J type thermocouple. The thermocouple was
installed near to the air nozzle opening (about 1cm) to indicate the air exit
temperature. the parts: exit hose of the heater, thermocouple connection, and air
nozzle were thermally insulated with ceramic fiber so that the obtained temperatures
of air exit are with good accuracy.

Test conditions and test specimen:

This section will handle the possible conditions that may affect the process results.
As indicated in figure 5a, the motion direction of the polishing process is the same
as the deposition direction (perpendicular to the staircase). Also, the hot air jet is at
right angle to the processed surface. The air nozzle diameter (d) is 2.5 mm and
opening is at an apparent distance (L) of 5-0.5 mm from the processed surface.
Hence, the dimensionless apparent distant (L/d) between nozzle opening and the
processed surface is equal to two; this distance is considered small so that the
thermal losses of the heated jet to the ambient fresh air is considerably small
(Goldstein, 1990). All the data collected from the polished surfaces correspond to a
single pass test. This means that the test surface was exposed to the heated jet for
one time only (down to up).

Figure 6 shows the test specimen; it is made of Poly-Lactic Acid (PLA) using
"Ultimaker Original" machine with 0.4 mm machine nozzle opening. The site of
Ultimachine cites the common PLA type of PLA4043D-Natureworks. The melting
temperature within extrusion is 210℃ and can be processed at 190-230 ℃; (Greco
et al., 2014a, 2014b) discuss the sintering of PLA.

The chosen specimen shape is simple to exclude further complications related to

the shape or jet incident angle. Its shape ensures six side faces that can be
processed. Also, the specimen surfaces/faces have sufficient wall thickness of larger
than 2.5 mm to withstand polishing. The surface roughness of the pre-processed

7
faces is from 7 to 8 μm –measured by R a - while faces with values other than that
are excluded. Actually, the roughness value of 7.5∓0.5μm is attained by using a
layer thickness around 0.1 mm and surface angle of 90° (vertical surface); this R a
value is possibly attained also by other combinations of layer thickness and surface
angle but still the resulting Ra value is the proper presenting parameter to all these
combinations. This chosen R a value tests the actual importance of this polishing
process. The value of 7.5∓0.5μm is considered the limiting value to several printer
machines -by manufacturing only- and hence, lower R a values than 7.5∓0.5μm will
be achievable only with the aid of the developed polishing process.
The surface roughness data was measured by a digital portable roughness tester of
type: STR-6210S. The measurements were carried out around faces mid-plane
which is the place of processing with hot air and then the average values were
obtained for pre-processed and post-processed surfaces.

4. Experimental Results:

Now, the process variables are reduced to three input parameters to be tested on the
resulting roughness improvements. These three process parameters are air jet
temperature, air jet velocity, and air nozzle translational velocity over the part
surface. Other possible process parameters mentioned in section 4 were held
constants. The following items will present the effect of each process parameter on
the polishing process. The investigation of the obtained trends with in-depth
discussion is presented in section 6.3.

5.1 The polishing process behavior at different process parameters:

Figure 7 presents the achieved R a reduction ratio with the nozzle translational
velocity at different jet velocities and temperatures. Each point in the figure
presents a polished face at certain process parameters. At vertical axis, there is the
achieved R a reduction ratio% calculated by:
R a reduction%= [100*(Raprep-rocessing-Rapost-processing)/ Rapre-processing]

8
To describe the process behavior with present of three process parameters, Figure
7a shows the variation between R a reduction and nozzle velocity at different jet
velocities while jet temperature was held constant at 235 ℃; also figure 7b shows
this variation at different jet temperature while jet velocities were held constant at
20 m/sec. The constant parameters (jet temperature of 235℃, jet velocity of 20
m/sec.) used in figure 7 are chosen to be about in the middle of the tested range of
jet temperature (215-265 ℃) and jet velocities (14-28.5 m/sec.). These ranges were
chosen based on primary tests performed by trials – after consider the material
melting temperature- and it indicated that lower ranges result to larger processing
time which cause large deformation or change to the specimen shape.

All surface roughness data (by R a ) were measured by RC filtering (RC filtering
is a traditional 2- stage filter with phase difference) except points with negative
reductions in figure 7 as will be explained. Filtering gives more stable and precise
results indicating roughness only and excluding other unwanted surface features
like inclination and waviness.

Figure 7 presents several polishing conditions and will be useful to recognize the
possible process parameter for each required reduction ratios. Various polishing
conditions/parameters can results to the same reduction ratio while using certain
condition will be preferred (as will be demonstrated). The sintering phenomenon (in
figure 3) interprets the noticed increase in R a reduction ratio with decreasing air
nozzle velocity where longer sintering time – or lower velocities- results to better
sintered/polished surface.

A main result in figure 7 is that the reduction in R a value increases with decreasing
the nozzle velocity till the ultimate value and then, further decrease in nozzle
velocity results to local waviness and/or possibly scattered holes in the surface
similar to pitting. This condition of local waviness and/or pitting is represented in
figure 7 by negative reduction percentage. This deterioration mode also results by
increasing jet temperature or jet velocity from this ultimate condition. Figure 8

9
presents these two described conditions (the ultimate reduction condition and the
failed ones).
The traditional measuring (by RC filtering) of this failed surface –especially the
failed by waviness mode- may give roughness values similar to the maximum
reduction point which indicates that filtering excludes the introduced waviness and
makes the two cases are similar. To distinguish between these two conditions, other
readings were taken - in addition to RC readings- to the tested surfaces using DP
(Digital Primary profile or no filtering) before and after processing to detect the
start of surface deterioration. Fortunately, DP detected the introduced waviness
which represented in figure 7 by negative reduction percentages – calculate by DP
readings- while it also continued to detect a positive R a reduction in the remaining
points which represented/calculated in the figure by RC readings.

On the positive side, the overall variation of R a reduction with nozzle translational
velocity in figure 7 is similar to the one in laser polishing process. it is known that
there are two energy level in laser polishing; the first one is Surface Shallow
Melting (SSM) (Lamikiz et al., 2007) which represents an appropriate range of
energy-rate characterized by adjacent R a improvement. The second one is Surface
Over Melting (SOM) which is characterized by higher energy rate with increased
roughness values of the processed surface than the initial one (Ramos et al., 2002).

5.2 The effect of jet velocity on polishing process performance:

Another studied parameter beside nozzle velocity (nozzle velocity is discussed in

section 5.1) is air jet velocity. The exit velocities of the heated jet were obtained
by direct measuring of volume flow rate of air and then, applying the continuity
equation - considering the change of temperature/density of air. These flow
rates are 2.5, 3, 3.5, 4, and 5 lit./min. while the pre-heated air is at about 28℃.

The effect of jet velocity on process performance will be demonstrated by studying

the following two sections. The figures of these sections were generated by reusing

10
the data in figure 7a. These sections will only present the obtained jet velocity
behavior while section 6.2 investigates and evaluates these behaviors.

5.2.1 The effect of jet velocity on the maximum achieved Ra:

The results shown in Figure 7a present the R a reduction data achieved at different
jet velocities and constant jet temperature of 235℃. The points of maximum Ra
reduction achieved are the points of interest to be observed. These points are unique
and representative for each specific jet velocity curve. Hence, they will be used to
represent their own jet velocity curves. Figure 9 and table1 presents the
variation between the maximum Ra reduction points in figure 7a and its
corresponding jet velocities. Table 1 contains the numerical R a values -
instead of percentages - of the data presented in figure 9. The error bar in the
figure represents the scatter in the measured R a reduction ratio while the line curve
was conducted between the averages of the collected data.

The figure indicates that the possible R a reduction achieved decreases with
increasing jet velocity and hence, the increase in jet velocity limits the
maximum reduction achieved (investigation to this fact at end of section 6.2). Also,
the achieved reduction becomes constant at jet velocity of about 14 m/sec.; this

indicates that this jet velocity or a pit lower is preferred as it achieved the highest
reduction between different jet velocities. Moreover, polishing with low jet
velocities can reduce pumping and heating energy to the used air and can simplify
the polishing system. This interesting property will be useful to several aspects.

5.2.2 The effect of jet velocity on polishing machinability:

This section discusses the nozzle velocity range (or the machining range) which
produce the highest Ra reductions for every jet velocity curve; this highest R a
reductions range from reductions more than 68% (R a of less than 2.4 μm at table1).
This nozzle velocity range is collected from figure 7a by observing each jet
velocity curve. These collected ranges of nozzle velocity are considered
11
representative to each specific jet velocity. Hence, they will be used to represent
their own jet velocity curves (in similar manner to section 5.2.1). Figure 10 shows
these nozzle velocity ranges represented by the bar at each corresponding jet
velocity. Also, the points generating the line curve in the figure are the
recommended nozzle velocity to produce the maximum R a reduction value for a
good wall quality of the processed surface (section 6.2 will discuss the role of wall
quality).

The figure illustrate an increase in nozzle velocity with the increase of jet velocity.
This variation is logic where the amount of heat rate delivered by air increases
with increasing jet velocity and as there is no overheating/damage surfaces
inbetween the data (in the figure); hence, the nozzle velocity should increase. An
important observation in the figure is that lower jet velocity permits wider range of
nozzle velocity or wider machining range. Actually, Figure 9 supports this result as
lower jet velocities permits higher reduction ratios or wider machining range.

5.3 The effect of jet temperature on polishing process performance:

The final studied parameter is air jet temperature. The way generating and
illustrating the variations between parameters will be presented in two sections in a
similar manner to section 5.2. The figures of these sections are generated by reusing
the data in figure 7b. These sections will only present the obtained jet temperature
behaviors while section 6.3 investigates and evaluates these behaviors.

5.3.1 The effect of jet temperature on the maximum achieved Ra:

The results shown in Figure 7b present the R a reduction data achieved at different
jet temperatures and constant jet velocity of 20 m/sec. The points of maximum R a
reduction achieved are the points of interest to be observed. These points are unique
and representative for each specific jet temperature curve. Hence, they will be used
to represent their own jet temperature curves. Figure 11 and table 2 present the

12
variation between the maximum R a reduction achieved and its corresponding jet
temperature.

The figure indicates that the maximum R a reduction achieved increases with
increasing jet temperature. This confirms the fact that, higher temperatures provides
more sintering and hence smoother surface. Also, it is clear that jet temperature at
235℃ is the critical point above which, the increase in jet temperature causes a
considerable increase in R a reduction than those below 235℃ hence; polishing with
a jet temperature higher than 235℃ is preferred.

5.3.2 The effect of jet temperature on polishing machinability:

The final relation presented in figure 12 was between jet temperatures and the
corresponding possible nozzle velocity range (or the machining range) to obtain
a R a reduction more than 68 %. This relation was established by collecting
nozzle velocity range at every jet temperature (In a similar manner to section 5.2.2).
The bar in figure 12 represents the nozzle velocity ranges while the points
generating the line curve are the recommended nozzles velocity to produce the
maximum R a reduction value for a good wall quality of the processed surface.

The data in Figure 12 indicates that jet temperature has more effect on
machinability than jet velocity (figure 10) as it provides wider nozzle velocity
ranges. This observation is noticed by comparing the curves in figure 7a (jet
velocities of 23 m/sec and 28 m/sec, at 235 ℃) with those in figure 7b (jet
temperatures of 250 ℃ and 265 ℃, at 20 m/sec). The polishing condition at high
jet temperature and moderate jet velocity lasts more effectively at high nozzle
velocities than that at moderate jet temperature and high jet velocity. This
observation relies on the concept of Newton law of heating (qtranfered= h.A.∆T);
which states that the temperature difference (between the jet and the polished
surface) is the driving force to cause heat transfer and hence attain melting; for
example, much high jet velocity with cold jet temperature will indeed has no effect

13
on the polished surface while much low jet velocity with high jet temperature can
affect on the surface.

Also, it was observed that jet temperature at 235℃ was critical point above which
the nozzle velocity became constant which was also noticed in figure 11. This may
refer to a thermal property in PLA-biopolymer material that can affect the
performance of sintering process. A possible cause of this critical point is the
behavior of material through heating it till melting. (Gibson and Shi, 1997)
illustrated a similar behavior for different types of polymer materials within
sintering. Semi-crystalline and crystalline polymers ultimately become viscous with
a corresponding large increase in density as soon as they reach a certain/critical
temperature point. This behavior is also found in amorphous polymers but with
lighter power. Consequently, it appears that this experimental critical temperature
and the reported one seem to be related. Whether this is the case or not, there is a
single nozzle velocity can be used with a range of jet temperatures and still
achieving the best reduction.

6 Discussion:

This section will investigate the different deterioration modes appeared in the
surfaces as shown in figure 8b. Once these modes are discovered and avoided; we
will respectively recognize the effect of different parameters on polishing process,
and will achieve higher and safer reductions.

6.1 The nature of FDM technique and its possible defects:

The data presented in figure 7 indicates that the decrease in nozzle velocity reduces
surface R a values till certain ultimate limit above which notable amount of
waviness and/or pitting in the surface arises. A careful examination to these pitting
holes indicated that they have dark appearance without near bottom or looks like a
through hole in a wall - see figure 13.

14
This description guides directly to the nature of FDM technology itself. As
mentioned, FDM parts are mainly manufactured by extruding small wire from
material filament. This wire is deposited behind each other to form a layer and then
layers are stacked above each other to form the part. Hence, the wall of part surface
is actually composed of several vertical layers (several walls) stacked behind each
other rather than a solid dense one. Within manufacturing, the deposition is not
completely perfect and defects may occur in surface layers in form of discontinuity
or incomplete adhesion between layers.

Figure 14 shows a schematic cross section in FDM part for a perfect and a defected
one. The figure is promoted with microscope's pictures to illustrate the defect types.
The defected part illustrates two main defect types: discontinuity on the extruded
wire or incomplete adhesion between layer boundaries. Discontinuity on the
extruded wire may occur when the extruding head translates from one layer to
other, a possible clogging in machine nozzle, or sudden change in extruding
velocity. The incomplete adhesion may occur by insufficient overlap, low extrusion
temperature, or inaccurate machine head.

6.2 Troubleshooting of the deterioration modes in polished surfaces:

The first investigated deterioration mode is the pitting form. This mode was found
to be a direct result of surface defects in form of microscopic holes (in figure 13) in
the surface. These defects/holes enlarge within processing to form this pitting
surface. Defects enlargement is primarily done by high air jet velocities. Figure 15
presents a schematic to the interaction between air jet and a defect in the wall.
When a jet with high velocity hits the bottom surface of the defect/hole, its flow
momentum turns 90° and produces side forces at the defect walls. These forces
impose the defect to enlarge.

This effect appears in figures 13a, 8b. The figures show holes placed in the middle
of the polished area which is exactly the place of maximum jet velocity. Also,
15
figure 8b shows the highest pitted surface between the failed ones which actually
processed at the highest jet velocity. The effect of jet velocity explains the reason of
fast deterioration to this face as it achieved only R a of 2.3μm (see tables 1, 2). In
contrast, it can be predicted that low jet velocity can prevent or lag the occurrence
of deterioration. Actually, this condition is also verified by noting the effect of jet
velocity on the achievable R a reduction in figure 9. The figure indicates that lower
jet velocity permits higher reduction, in other words, it delays the occurrence of
surface deterioration.
The demonstrated interaction of jet velocity interprets the observation obtained
from figure 10 about machinability. As jet velocity decrease, surface deterioration is
delayed and hence; permits larger R a reductions and wider nozzle velocity range (or
wider machining range).

This nature of jet flow demonstrates the difference (or the possible confusion)
between the static pressure applied by (Spence and Crawford, 1996) which
improves sintering rate and this dynamic pressure type (stagnation pressure) which
result from jet flowing nature. In contrast with static pressure by (Spence and
Crawford, 1996), high jet velocities introduce defect enlargement and limit sintering
performance.

Figure 15 also presents a schematic to the second deterioration mode, which is the
waviness. When the heated jet impinges unsupported surface; the surface may
suffer permanent deflection. This defection is introduced by the stagnation pressure
force that results from jet flow at a partially fused surface. This demonstrates also
another side effect of high jet velocity as it also promotes the occurrence of
waviness mode.

6.3 Evaluation of polishing process parameters:

The discussion in section 6.2 evaluates the effect of jet velocity on polished
surfaces. On the other side, this discussion also demonstrates important
characteristics of the polished surfaces which enhance the polishing process; it

16
indicates that the quality of the outer surface layer affects the maximum roughness
reduction. Moreover, only two or three adjoining layers to the first one will be
important and no matter what is the entire wall thickness; their role is to support the
outer first layer against possible deflection/waviness within machining hence, the
interlayer adhesion quality (in-between them) represents their importance.
Accordingly, thicker surface layer – especially the outer one- can enhance the
characteristics of polished surfaces against possible deteriorations and so permits
larger reduction. Therefore, manufacturing with a large nozzle opening of the
extruder (in the printer machine) can improve surface characteristics and hence,
improves polishing process performance. Also, sufficient overlap between surface
layers can introduce similar improvements.

On the other hand, high jet temperature permits higher reduction ratio (see figure
11) by decreasing the viscosity of staircase in the surface and causing better
sintering. This fact was verified by (Hamidi et al., 2016), and confirmed by the
condition of maximum Ra reduction (reduction of 88%) obtained at the highest jet
temperature as shown in Figure 16.

Another result is that, the conditions with lower reduction ratio than 88% are not
sufficiently sintered rather than, they failed early by the effect of jet velocity.
Furthermore, increasing jet temperature provides faster and/or local processing than
increasing jet velocity; this condition can compensate the weakness (weak adhesion
between surface layers) in the processed wall against possible waviness providing a
safer method.

6.4 The healing effect of the polishing process:

Another advantage of the polishing process is easily recognized by microscopical

examination of the polished surfaces. This will demonstrate that the defects
observed in pre-processed surfaces (in figure 14) are reduced in both amount and
apparent diameters as shown in Figure 17. This improvement is owing to
coalescence and bubble removal properties introduced by sintering phenomenon

17
upon melting (see Figure 3). Proper sintering (with lower jet velocities) activates
healing forces to coalesce discontinuities in surface and boundary layers while
larger jet velocities tend to enlarge surface defects.

The healing effect may have further improvement to the mechanical properties like
tensile or bending strength of the processed parts. The micro-pores or
discontinuities found in the surface layers represent the weakest points in part
integrity which initiate factures occurring under loads. (Rodríguez et al., 2003)
studied the effect of similar voids that may result. (Gurrala and Regalla, 2014)
studied the effect of coalescence between FDM part layers on the resulted part
strength and concluded that there is just sufficient neck growth but, not enough to
fully coalesce. Based on his study, it can be predicted that surface sintering with
sufficient coalescence between surface layers can improve the mechanical
properties (like tensile or bending strength) of the polished parts.

7 Conclusion:

In this study, a new post-processing technique was developed by using hot air to
locally melt the staircase surface (its microscopic corners) of FDM parts. The
results indicate that the technique causes a significant reduction in Ra values up to
88% from a single pass test. Also, the technique can reduce the defects in the
processed surface leaving the surface shiny.

The process is valid within a certain translational velocity of the air nozzle where
lower velocities can cause overheating and so surface waviness with the possibility
of surface pitting while higher velocities cause lower Ra reduction values or
possibly no improvements.

Also, the deterioration modes characterized with surface waviness and/or pitting is
dependent on wall quality which represented by its interlayer adhesion quality and
the amount of surface defects. A qualified wall permits higher Ra values with the
least appearance of waviness in case of deterioration. This qualified wall may be

18
attained by manufacturing of FDM parts with a large nozzle opening of the extruder
(in the printer machine) or with sufficient overlap between surface layers.

Also, it was observed that high jet velocity limits the maximum achievable Ra
reduction as it causes fast deterioration to the surface in form of surface pitting.
lower one permits safer processing and higher Ra reduction.

In contrast to jet velocity, higher jet temperatures permit higher reduction ratio and
reduce the defect on the final surface. This also permits faster and local melting to
the surface and hence, introduces less appearance of waviness.

The best-suggested combination found between parameters is the highest jet

temperature with low to moderate jet velocity values. It will be preferred to work
with jet temperature higher than 235℃ and jet velocity lower than 20 m/sec.

8 Acknowledgement:

This work was supported by "Science and Technology Development Fund;

(STDF)" office which belongs to "The Egyptian Ministry of State for Scientific
Research" under Grant type: RSTDG and project IP: 12612. Thanks to STDF patent
office for their efforts in patent submission. The authors would like to acknowledge
the "Industrial Technology Transfer Unity; (ITTU)" office in Assuit University for
their technical support.

19
References:

Ahn, D., Lee, S., 2007. Improving the surface roughness of SL parts using a coating
and grinding process. Int. J. Precis. Eng. Manuf. 8.

Ahn, S., Lee, C.S., Jeong, W., 2004. Development of translucent FDM parts by
post-processing. Rapid Prototyp. J. 10, 218–224.
doi:10.1108/13552540410551333

Bellini, A., Shor, L., Guceri, S.I., 2005. New developments in fused deposition
modeling of ceramics. Rapid Prototyp. J. 11, 214–220.
doi:10.1108/13552540510612901

Bordatchev, E., 2014. Performance of laser polishing in finishing of metallic

surfaces. Int. J. Adv. Manuf. Technol. 73, 35–52. doi:10.1007/s00170-014-
5761-3

Boschetto, A., Bottini, L., 2015. Surface improvement of fused deposition modeling
parts by barrel finishing. Rapid Prototyp. J. 21, 686–696. doi:10.1108/RPJ-10-
2013-0105

Galantucci, L.M., Lavecchia, F., Percoco, G., 2009. Experimental study aiming to
enhance the surface finish of fused deposition modeled parts. CIRP Ann. -
Manuf. Technol. 58, 189–192. doi:10.1016/j.cirp.2009.03.071

Gibson, I., Shi, D.P., 1997. Material properties and fabrication parameters in
selective laser sintering process. Rapid Prototyp. J. 3, 129–136. doi:Doi
10.1108/13552549710191836

Goldstein, R., 1990. Effect of entrainment on the heat transfer to a heated circular
air jet impinging on a flat surface. J. Heat Transfer 112.
doi:10.1115/1.2910430

Greco, A., Maffezzoli, A., Forleo, S., 2014a. Sintering of PLLA powders for
rotational molding. Thermochim. Acta 582, 59–67.
doi:10.1016/j.tca.2014.03.005

20
Greco, A., Maffezzoli, A., Forleo, S., 2014b. Rotational molding of bio-polymers.
AIP Conf. Proc. 1593, 333–337. doi:10.1063/1.4873794

Gurrala, P.K., Regalla, S.P., 2014. Part strength evolution with bonding between
filaments in fused deposition modelling. Virtual Phys. Prototyp. 9, 141–149.
doi:10.1080/17452759.2014.913400

Hamidi, A., Farzaneh, S., Nony, F., Ortega, Z., Khelladi, S., Monzon, M., 2016.
Modelling of sintering during rotational moulding of the thermoplastic
polymers. Int. J. Mater. Form. 519–530. doi:10.1007/s12289-015-1239-6

Jones, J.B., Wimpenny, D.I., Gibbons, G.J., 2015. Additive manufacturing under
pressure. Rapid Prototyp. J. 21, 89–97. doi:10.1108/RPJ-02-2013-0016

Lamikiz, a., Sánchez, J. a., López de Lacalle, L.N., Arana, J.L., 2007. Laser
polishing of parts built up by selective laser sintering. Int. J. Mach. Tools
Manuf. 47, 2040–2050. doi:10.1016/j.ijmachtools.2007.01.013

Leong, K.F., Chua, C.K., Chua, G.S., Tan, C.H., 1998. Abrasive jet deburring of
jewellery models built by stereolithography apparatus ( SLA ). J. Mater.
Process. Technol. 83, 36–47. doi:http://dx.doi.org/10.1016/S0924-
0136(98)00041-7

Löhner, M., Drummer, D., 2015. The effect of pressure variations on the formation
of gas inclusions in the rotational molding process. J. Polym. Eng. 35, 9–12.
doi:10.1515/polyeng-2014-0145

Masood, S.., Song, W.., 2004. Development of new metal/polymer materials for
rapid tooling using Fused deposition modelling. Mater. Des. 25, 587–594.
doi:10.1016/j.matdes.2004.02.009

Pal, D., Ravi, B., 2007. Rapid tooling route selection and evaluation for sand and
investment casting. Virtual Phys. Prototyp. 2, 197–207.
doi:10.1080/17452750701747088

Ramos, J., Bourell, D., Beaman, J., 2002. Surface over-melt during laser polishing

21
 of indirect-SLS metal parts, in: MRS Proceedings. Cambridge University
Press, New York, USA. doi:10.1557/PROC-758-LL1.9

Rodríguez, J.F., Thomas, J.P., Renaud, J.E., 2003. Mechanical behavior of

acrylonitrile butadiene styrene fused deposition materials modeling. Rapid
Prototyp. J. 9, 219–230. doi:10.1108/13552540310489604

Spence, A.G., Crawford, R.J., 1996. Removal of Pinholes and Bubbles from
Rotationally Moulded Products. Proc. Inst. Mech. Eng. Part B J. Eng. Manuf.
210, 521–533. doi:10.1243/PIME_PROC_1996_210_151_02

Williams, R.E., Melton, V.L., 1998. Abrasive flow finishing of stereolithography

prototypes. Rapid Prototyp. J. 4, 56–67. doi:10.1108/13552549810207279

Zinniel, R.L., 2011. Vapor smoothing surface finishing system. Google patents,
US8075300 B2.

22
List of figures:

Note: the text below the figure is the primary figure caption while the text before it contains
extra notes about the figure.

Fig. 1. FDM technique and its main components.

Fig. 2. Description of the proposed polishing process.

23
Fig. 3. (a) Description of the sintering phenomenon occurs between two particles.
(b) The sintering phenomenon occurs between staircase layers in a FDM part
surface.

Fig. 4. Experimentally polished surfaces; (a) Line polishing test – Ra reduction of

51%; initial Ra of 5.1μm; (b) magnification of boundary region - Ra reduction of
77%; initial Ra of 8 μm.

24
Fig. 5. Test rig of the developed finishing technique: (a) Schematic diagram; (b)
picture of the experimental setup.

Fig. 6. The test specimen used.

25
Fig. 7. Variation between nozzle velocity and the resulting Ra reduction; (a) at
constant jet temperature; (b) at constant jet velocity

26
Fig. 8. Surfaces processed at 235℃, Vjet of 28.5 m/sec.; (a) ultimate reduction -
Vnozzl of 0.95 mm/sec.; (b) negative reduction - Vnozzle of 0.85 mm/sec.

Fig. 9. Variation between jet velocity and maximum achieved reduction.

27
Fig. 10. Variation between jet velocity and air nozzle velocities for R a reductions
more than 68%.

Fig. 11. Variation between jet temperature and maximum achieved reduction.

28
 Fig. 12. Variation between jet temperature and air nozzle velocities for R a
reductions more than 68%.

Fig. 13. Description of Pitting holes, (a) defected face, (b) further magnification to
"a" – the picture rotated180°, (c) further magnification to the hole in "b".

29
Fig. 14. Possible part/wall quality - Sectional elevation to the part, (a) defected wall
(b) perfect wall.

30
Fig. 15. Schematic of deterioration modes in case of overheating to the polished
surface.

Fig. 16. Surfaces processed at T of 265℃, Vjet of 20 m/sec., (a) ultimate velocity
condition, (b) negative reduction condition.

31
Fig. 17. Pictures demonstrate the healing effect, (a) unprocessed areas (b) processed
to Ra reduction 70%, (c) processed to Ra reduction 75%.

32
 Table 1 Lowest Ra value at each jet velocity curve found in figure 7a.

Air jet velocity (m/sec.) 14 17 20 23 28.5

Average Ra values (μm)- stander 1.2- 1.4- 1.55- 1.8- 2.3-

deviation 0.2 0.2 0.3 0.5 0.9

Table 2 Lowest Ra value at each air temperature curve found in figure 7a.

Air temperature (℃) 215 225 235 250 265

Average Ra values (μm)- 1.7- 1.57- 1.55- 1.1- 0.85-
stander deviation 0.26 0.1 0.3 0.1 0.1

33