FGM fabrication

In this chapter the description of the procedure to manufacture the functionally graded (FG)
materials is presented. In this research, FGM specimens were manufactured from the
mixture of ceramics and polymers. Alumina (Al 2O3) is used for the ceramic constituent.
Epoxy-resin is used for the polymer constituent. Based on the extensive literature review in
Chapter 2, it was found that, among three effective techniques used for manufacturing
FGMs; namely, thermal spraying, powder metallurgy and infiltration techniques, the
ceramic-polymer FGMs can only be manufactured appropriately by using the infiltration
technique. Therefore, a multi-step sequential infiltration technique is used to fabricate FG
beam specimens made from alumina and epoxy-resin in this research. In this chapter, the
information in relation to FGM fabrication will be provided in detail as follows:
5.1 Specimen fabrication using multi-step sequential infiltration technique
5.2 Microstructure analysis
5.3 Alternative technique of specimen fabrication

5.1 Specimen Fabrication using multi-step sequential infiltration

technique
Alumina-epoxy composite beams were manufactured using a multi-step sequential
infiltration technique. The earliest development of the technique was proposed by Cichocki
et al. (Cichocki et al., 1998) to produce graded composite materials.
The key principle of the infiltration technique is to make an alumina piece containing a
graded network of porosity which will be infiltrated by epoxy in the final step, in order to
obtain alumina-epoxy composite specimens. By using this technique, the composite
specimens can be represented in the form of interpenetrating-network (IPN) structured
composites which contain two phases of materials that are alumina and epoxy phases. In
several previous investigations (Clark, 1992, Prielipp et al., 1995, Lange et al., 1990), it
was affirmed that the infiltration technique is an effective methodology that can be used for
producing graded composite materials where their material compositions are continuous.
Composite materials having the interpenetrating network structures are different from
traditional fibre-matrix and particle-matrix composite materials in terms of compositional
distribution and pattern which then lead to different behaviour when they are subjected to
mechanical loads. Therefore, this chapter will give overviews of the production process for
making graded composite beams.
The unique advantage of the infiltration technique is that it is suitable for making graded
composites with a wide range of material constituents. As mentioned in Chapter 2 of the
literature review, for example, Neubrand et al. (Neubrand et al., 2002) successfully
employed the infiltration technique to produce ceramic-metal graded composites which
were used for investigating residual stress analysis. To obtain graded composites which
have a large disparity in elastic properties, ceramic-polymer could be used as the material
constituents. Thus, in the investigations of Tilbrook et al. (Tilbrook et al., 2006), the
alumina-epoxy graded composite system in which material compositions were varied along
an axial direction, was produced using the infiltration technique for the purpose of
investigating crack propagation. In this research, the material compositions between
alumina and epoxy of the FG beam specimens will be graded across the thickness direction
of the specimens. The outline of the fabricating process is illustrated in Fig. 5.1.

Fig. 5.1 Schematic of fabricating process of graded composite specimen

5.1.1 Foam preparation
According to the schematic of the production process, it began with the foam preparation
process. This is to construct a porous network structure inside the specimens. The network
structure was achieved by using open-celled polyurethane foam (PU) as an imprint for such
a structure. The polyurethane foam used in this preparation was commercial foam whose
average cell size was approximately 0.8 mm. The PU foam was cut to the required sizes
(length 100 mm and breadth 12 mm), then the foam pieces were sliced to three different
thicknesses (h), 2 mm, 4 mm and 6 mm. There were many ways to slice the foam pieces in
order to obtain a series of very thin foams. For instance, in the foam preparation process of
(Tilbrook, 2005), the PU foam was immersed with water and frozen to retain shape before
it was cut to size with a bandsaw. This method of cutting the foam did not yield a perfect
result because the cut foam pieces were wet and required thawing and drying later. Usually,
the drying process of the wet foam pieces took around 12 hrs to dry in an oven at 45 0C.
Therefore, in this fabrication, the PU foam was cut and sliced by using a hot-wire cutter
that can be considered as a better way, without requiring any process of freezing, thawing
and drying.
The foam pieces with different thicknesses were placed in a uniaxial hot press and
compressed to the same constant thickness of 1 mm, producing a series of foam pieces of

different density. In order to retain the compressed shape, a heat treatment cycle had to be
applied during this compressing of foam. A schematic of the foam compression process is
shown in Fig. 5.2. As seen in Fig. 5.2, a pressure of 300 kPa was applied to compress the
foam pieces. The temperature was ramped up to 180 0C with the increasing rate of 1
0C/min, and was held for 1 hour before letting the temperature cool down naturally to room
temperature. The foam was then removed from the hot press. Unevenness at the edges due
to thermal stress was found which was removed before being utilised in the subsequent
stages.

The uniaxial hot press that was used in this foam preparation process is shown in Fig. 5.3.
It is seen that the brass spacer containing the foam pieces having different thicknesses was
positioned between the top and bottom platforms. Both platforms were set to heat up to 180
0C using a temperature control box. The pressure of the hot press was controlled at 300 kPa
by an hydraulic system. The spacer under the mentioned pressure and temperature was held
for 1 hour in order to ensure that the foam pieces inside had retained the new shapes. The
spacer was then allowed to cool down naturally before removal of the compressed foam
pieces.

5.1.2 Slip casting and drying processes

To manufacture graded composite specimens, a slip casting process was used to create
alumina phases in specimens. A useful and appropriate optimum slip formula for the
infiltration technique was determined by Neubrand et al. (Neubrand et al., 1999).
Therefore, the alumina slip used to manufacture the specimens in this research was
produced using the same formula as proposed in (Neubrand et al., 1999, Tilbrook, 2005).
Within the slip formula, several additives were incorporated in order to improve the quality
of the slip casting process with a colloidal suspension of very fine alumina particles. For
example, a wetting agent was used to aid foam infiltration, an antifoaming agent to prevent
air entrapment, and binders and plasticisers to create green strength and prevent drying
cracks (Rutgers, 2005). The slip contents were presented as follows:
Alumina Slip Formula
45 mL of alumina slip with a 25vol% solids loading can be obtained with the
following formula:
32.525 g Distilled water
0.145 g Ammonia solution (25% concentrate)
0.464 g Dolapix CE64 (Zschimmer and Schwarz, Lahnstein, Germany)
0.276 g Glydol N 109 NEU (Zschimmer and Schwarz, Lahnstein, Germany)
0.094 g Contraspum KWE (Zschimmer and Schwarz, Lahnstein, Germany)
0.927 g Duramax B-1052 (Rohm & Haas, Lauterborg, France)
46.350 g Alumina Powder (Taimicron TM-DAR, Taimei Chemical Co Ltd,
Japan)
The 99.99 % purity alumina powder (Al2O3), powder size of approximately 1m, was
chosen to make the alumina slip because small powder particles can penetrate easily into
the pores of the compressed foam. All of the chemical ingredients used to make the alumina
slip in this research are shown in Fig. 5.4.

A very good mixture of all ingredients (the liquid components) with the alumina powder
according to the formula can be achieved by the aid of a magnetic stirrer (IEC magnetic
stirrer) and an ultrasonic bath (UNISONICS). All of the chemical ingredients and alumina
powder were weighed sequentially corresponding to the formula and mixed together in a
beaker. A teflon-coated magnetic bar was added to the beaker, which was then placed on
the magnetic stirrer. The magnetic stirrer was operated for 10 minutes before the beaker
was moved to the ultrasonic bath for 10 minutes. However, one cycle of mixing, was found
to be insufficient to obtain a good mixture. Therefore, the mixing process with the magnetic
stirrer and the ultrasonic bath was repeated again. The mixing process was carried out at
least 3 times in order to produce the strongest mixture of all ingredients. It was found that
the mixing process led to a reduction of large agglomerates and produced good dispersion.
The alumina slip was prepared one day before casting for ageing purposes. Before the
alumina slip was used for casting in the following day, it was re-mixed or re-dispersed by
the magnetic stirrer and the ultrasonic bath again.
In the slip casting process, the casting mould, which consists of pieces of Perspex, was used
to create the graded composite specimens. The Perspex was made from plaster of Paris. The
pieces of the Perspex were fitted together to the dimension of the compressed foam (length
100 mm and breadth 12 mm). They were placed on a Plaster of Paris block in a vacuum
chamber, as shown in Fig. 5.5. The benefit of the Plaster of Paris block is that it can absorb
moisture from the alumina slip and let the specimen dry appropriately. To minimise the
problem of adhesion between the surfaces of the cast specimen to the mould walls, the
Perspex pieces were lined with Teflon tape. The filter paper (ADVANTEC) was placed
between the compressed foam and the plaster of Paris block. This is done in order to avoid
the likelihood of contamination of the alumina slip by particles or soluble ions from the
plaster. To clearly understand the slip casting set up, the cross-section of apparatus is
illustrated in Fig. 5.5.

After preparing the casting mould, the series of compressed foam pieces were stacked into
the mould. The highest density foam piece was placed in the mould first, followed by the
lighter ones. Due to the density of the foam pieces being much lighter than that of the
alumina slip, the foam pieces would normally float in the slip causing the problem of
discontinuity between adjacent layers. To solve this problem, there were several attempts to
glue the foam pieces together in order to ensure phase connectivity across the stepinterfaces.
Using glue to stick the foam pieces together was one of many attempts to create
the connectivity; however, the glue tended to cause warping of the foam, producing
delamination. Another attempt was to use a heat treatment with slight pressure to construct
the foam piece connection. However, these foam joining techniques presented above were
not effective in producing a suitable bond. Thus a better way to join the foam pieces was to

place the foam pieces so they were restrained slightly by the mould walls or the Perspex
pieces without using glue and then apply a little pressure when placing them into the
mould. Consequently, good connectivity between the adjacent layers can be obtained.
Slip casting was performed in the vacuum chamber (LABEC) by placing the mould
assembly into the chamber. At the beginning stage of the slip casting process, the alumina
slip was filled into the foam pieces at a slow filling rate and using a small amount of the
slip; otherwise, the foam pieces would have floated on the surface. To enhance infiltration
of the slip casting process after the slip was poured over the foam pieces the pressure of the
chamber was lowered to around 100 mbar that led to air bubbles leaving the foam. When
the lowered pressure was released to atmospheric pressure, the slip was then forced into the
porosity of the foam pieces. To minimise the amount of air bubbles, the cycle of varying
pressure could be repeated several times.
After leaving the casting specimen within the mould for 1 or 2 days, the water was
observed to have been absorbed by the Plaster of Paris block. As a result, the slip volume
decreased. At this stage, the foam pieces were joined and glued together by more colloid
alumina slip. To complete the slip casting process, an additional slip was required with the
cycle of varying pressure again. The mould containing the specimen was kept in a closed
plastic container to prevent rapid moisture evaporation from the top surface, which might
result in drying cracks. After being kept in the plastic container for around 7 days, the
specimen had hardened sufficiently to retain the shape and was then removed from the
casting mould. The set of the casting mould and some sufficiently dry specimens are shown
in Fig. 5.6.

The dry specimen was relocated onto a piece of uncompressed polyurethane foam in order
to aid air movement to all surface areas in another closed container. The specimens were
kept in the container for around 7 days. Subsequently, the lid of the container was gradually
opened to promote slow moisture removal from the specimen within the period of 3-5 days.
The final drying process was to move the specimen into an environmental chamber at 50 0C
for 12 hours.
5.1.3 Foam burning out and sintering
After the casting specimen dried completely, the graded porous network structures inside
the specimen were then manufactured by burning out the compressed foam pieces. The
specimen whose size was 100 mm x 12 mm x 4 mm (length x breadth x thickness) was
fired with a low temperature regime, to first pyrolyse the foam pieces and debinder the

ceramic phase. The temperature profile used for the whole process of foam pyrolysis is
illustrated in Fig. 5.7 with the details of temperature increment. For example, starting with
the room temperature, the following temperature was used: 25 0C to 235 0C at 300C/hour
held for 1 hour (foam burnout), 235 0C to 400 0C at 300C/hour held for 1 hour (binderburnout),
400 0C to 800 0C at 400C/hour held for 1 hour (ash burnout). The specimen was
allowed to cool down naturally to room temperature before it was heated up again for
sintering the ceramic phase at 1400 0C, held for 1 hour. The temperature profile of sintering
process is shown in Fig. 5.8. The furnace, (Ceramic engineering furnace manufacturer,
Sydney, Australia) as shown in Fig. 5.9, was used for both burning out foam and sintering
ceramic processes. It was found that sintering reduces the dimensions of the specimen in
every direction, and leaves an alumina specimen with layers of graduated porosities.

5.1.4 Polymer infiltration

The alumina specimen having layers of graduated porosities was ready for polymer
infiltration, using a vacuum chamber, in order to obtain an alumina-epoxy composite
specimen. The alumina specimen was placed into a silicone mould which was suitable for
this process because the specimen can be removed easily when polymer is cured. Epoxy
resin (Epofix, Struers, Denmark) was chosen to fill the graduated porosities inside the
alumina specimen under varying pressures and was then cured at room temperature. By
varying the pressure, air bubbles were forced to leave the specimen and epoxy resin could
infiltrate the graduated porosities more easily. The cycles of varying pressure within a
range of 100 to 1000 mbar were required until ensuring the amount of air bubbles was
minimised. The next step was to leave the epoxy resin cured at room temperature for
around 24 hours. After curing, excess epoxy was removed by polishing with coarse
sandpaper. Finally, the surfaces of specimen were ground and polished with diamond paste.

5.2 Microstructure Analysis

The graded composite beam specimens made from alumina-epoxy which were produced
using the above procedure were selected randomly to section across the thickness direction
to examine the quality of the graded area. Then the sectioned specimen was mounted by
epoxy resin before grinding and polishing with sandpapers and diamond paste, respectively.
For polishing with the diamond paste, the specimen was first polished using an automatic
polisher (Leco VP-50). The used diamond paste has a diamond particle size of 3 m which
was reduced to 1 m at the final step.
Images of the polished surfaces, obtained by optical microscopy with a Nikon 200
microscope and digital camera, were used to characterise the pure alumina and epoxy
phases over the graded region. One such cross-section image was shown in Fig. 5.10. A
measurement of areas, using the areal technique, in the cross-section yielded the
percentages of material compositions in each graded region. The top and bottom surfaces
were referred to as the pure alumina and epoxy resin layers, respectively.

Fig. 5.11 illustrates the whole material composition at the graded region of the specimen.
This includes information about percentages of alumina and epoxy resin in each layer
across the graded region.

As shown in Fig. 5.11: the cross-section image, a few pores were found due to imperfect
epoxy infiltration. It was also seen that the variation of material compositions throughout
the cross-section presented in the form of layer or step changes rather than a continuous
change over the section. By using the areal technique, one can obtain the percentages of
alumina and epoxy resin in each layer in the graded region. It is clearly seen that the top
layer was for a pure alumina and followed with (70% alumina-30% epoxy), (40% alumina60% epoxy) and (20% alumina-80% epoxy) layers, whereas, the bottom layer was for a
pure epoxy resin layer. According to the material compositions or material volume fraction
throughout the cross-section, these were plotted and illustrated in Fig. 5.12. The area underthe
step graph as shown in Fig. 5.12 matches the percentage of alumina (Vc); the proportion
of epoxy resin is indicated by the complementary area of the graph.

The process of slip casting was further investigated. A microstructure analysis displayed
the result of an imperfectly graded composite specimen when the alumina slip was filled
into the compressed foam pieces with a high filling rate and too much slip. The
microstructure of the imperfect specimen that was made by a high rate of alumina slip
infiltration is shown in Fig. 5.13. It can be seen that there are a number of buffer layers of
alumina across the specimen (seen in white colour layers). This is due to the compressed
foams density being much lighter than that of the slip. Therefore, the foam pieces easily
float on the slip between layers so that the buffer layers are constructed. The buffer layers
cause the problem of porosities occurring inside the specimen, especially at the middle zone
of the specimen. It is because the buffer layers have blocked or sealed a way for the epoxy
resin to come into the lower zone. This can be rectified by filling the alumina slip into the
foam pieces with a slow filling rate. The foam pieces were allowed to absorb all of the slip
in the previous infiltration before more slip was added to complete the slip casting process.

5.3 Alternative technique of specimen fabrication

The principles and ideas of powder metallurgy and multi-step sequential infiltration
techniques as presented in Chapter 2 are combined to propose an alternative technique to
produce layered graded specimens. The main advantage of the combined technique is that it
does not require the processes of slip casting and drying; therefore, time taken to
manufacture specimens is greatly reduced. The combined technique utilises the fact that
small particles of mothballs can be used to mix with the ceramic powder in each layer with
different compositions in order to produce graded porosities. The mothballs are burnt out at
low temperature during the sintering process; this leaves the ceramic specimen having
graded porosities. After sintering, the ceramic specimen is solidified and ready for the
process of polymer infiltration, consequently, the graded composite specimen made from
ceramic and polymer can be obtained.
This proposed idea of a combined technique can be proven by manufacturing specimens
from the new process. Mothballs are selected to mix with ceramic powder as they can be
burnt out completely at around 200 oC. The production process of the specimens using the
combined technique began with grinding mothballs to small particle sizes whose diameter
are around 0.5-1.0 mm. The 99.99% purity alumina powder (Taimicron TM-DAR, Taimei
Chemical Co Ltd, Japan) was mixed with the ground mothballs in different bottles which
were desired to have different material compositions. In this fabrication, the five layer
composite specimens were required with the percentages of materials in each layer being
(100 vol.% alumina-0 vol.% mothball); (20 vol.% alumaina-80 vol.% mothball); (40 vol.%
alumina-60 vol.% mothball) and (70 vol.% alumina-30 vol.% mothball). However, the last
layer contains pure epoxy which can be created in the final step of polymer infiltration.
Hence, four bottles with different percentages of alumina powder and ground mothballs
were obtained. The alumina powder and ground mothballs with CMC bonder powder were
then blended for every bottle by using a ball milling machine (Star Machinery Pty Ltd) as
shown in Fig. 5.14. The ball milling machine was set to run for 4-5 hours at 200 rpm in
order to obtain a good mixture of the material compositions.

After the materials in each bottle were mixed, they were brought to fill and form in a steel
die. Fig. 5.15 (a) illustrates the steel die. Fig. 5.15 (b) displays the press machine (Enerpac
13 series), which was operated at 30 kN to compress materials in the steel die. For layered
graded specimens, it was noted that the mixed powders were stacked sequentially layer by
layer with a layer-wise compositional distribution. The first layer (at the bottom of the die)
was given for pure alumina layer and followed with (20 vol.% alumaina-80 vol.%
mothball); (40 vol.% alumina-60 vol.% mothball) and (70 vol.% alumina-30 vol.%
mothball), respectively.

The green compacted specimen was removed from the die slowly with a small application
of force through the compressing bar of the die in order to avoid cracks. The green
specimen was then sintered using the high temperature furnace which is shown in Fig. 5.9.
The increasing temperature profile of this sintering was used on the same principle as
presented in Fig. 5.8. During the sintering, the mothballs were burnt out. After sintering,
one can obtain the alumina specimen having graded porosities. To complete the fabrication
process, the graded porosities inside the alumina specimen were infiltrated by epoxy-resin(Epofix,
Struers, Denmark) with the aid of varying pressure in the vacuum chamber as
shown in Fig. 5.16.

The layered graded specimen made from alumina and epoxy using the combined technique
was finalised in a similar fashion to the infiltration technique by grinding and polishing the
surfaces. The specimen was then sectioned to examine the quality of material gradients
across the graded direction using optical microscopy. It was found that the specimen
obtained via the combined technique had much more porosity than that produced by the
multi-step sequential infiltration technique. To show the combined technique leading to
porosities inside the specimen, Fig. 5.17 presents diagrams of layered graded specimen
fabrication via this technique. The diagrams can be used to explain the mechanisms of
graded construction in the specimen. The process of layer preparation is shown in Fig. 5.17
(a). The mixtures of ground mothballs and alumina powder are used to create layers
according to the desired layer-wise compositional distribution. The sintering process leads
to mothballs being burnt out and leaves the alumina specimen with layers of graded
porosities. The subsequent step is to infiltrate epoxy into the graded porosities of thealumina
specimen as shown in Fig. 5.17 (b). It can be seen that the epoxy profile 1 cannot
pass into the following layer, but the epoxy profile 2 can penetrate through the porosity
connection. However, very few of the porosity connections are created inside the specimen
using the combined technique. Therefore, there is much porosity left without epoxy
infiltration in the specimen as shown in Fig. 5.17 (b).

Due to few porosity connections, the final specimen produced from alumina and epoxy
using the combined technique may have a lot of porosities as seen in Fig. 5.18. The
specimen in this figure contains 3 phases: epoxy, alumina and porosity phases. Only the topand
bottom layers can be classified as the perfect layers which are the layers of pure epoxy
and pure alumina respectively.

According to the diagrams of layered graded specimen fabrication, it can be concluded that
the combined technique which is developed from the combination of ideas of powder
metallurgy and infiltration techniques is not suitable for manufacturing the layered graded

specimen. This is because the combined technique cannot produce the interpenetratingnetwork
(IPN) structures which are very important to minimise the amount of porosity.
Hence, to produce the layered graded specimen made of alumina and epoxy, the multi-step
sequential infiltration technique as presented in Section 5.1 is strongly recommended to be
used for producing such specimen. In Fig. 5.11, it is clearly seen that the interpenetratingnetwork
(IPN) structures are created in the specimen produced by the multi-step sequential
infiltration technique and the number of porosities is minimal. Thus, the layered graded
specimens fabricated by the multi-step sequential infiltration technique are used for further
investigation of vibration testing in this research.