Materials Science and Technology

ISSN: 0267-0836 (Print) 1743-2847 (Online) Journal homepage: http://www.tandfonline.com/loi/ymst20

Organic fibre modified ceramic shell moulding for

investment casting

S. Jones, C. Yuan & K. Lewis

To cite this article: S. Jones, C. Yuan & K. Lewis (2002) Organic fibre modified ceramic shell
moulding for investment casting, Materials Science and Technology, 18:12, 1565-1573, DOI:
10.1179/026708302225007772

To link to this article: http://dx.doi.org/10.1179/026708302225007772

Published online: 19 Jul 2013.

Submit your article to this journal

Article views: 10

View related articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ymst20

Download by: [McMaster University] Date: 06 April 2016, At: 13:24

 Organic Ž bre modiŽ ed ceramic shell moulding
for investment casting
S. Jones, C. Yuan, and K. Lewis

Investment casting research is being carried out by the University of Birmingham sponsored by the EPSRC and a
consortium of industrial companies. The programme is aimed at developing a fundamental understanding of the
process, with a view to routinely producing sound, net shape castings. The casting of liquid metals to produce solid
objects is a manufacturing process, which has been practised for over ® ve thousand years, with investment casting
being one of the oldest known metal shaping methods. The technique itself has tremendous advantages in the
production of quality components and key bene® ts of accuracy, versatility and integrity. As a result the process is
one of the most economic methods of forming a wide range of metal components. Environmental and economic
pressures have, however, resulted in a need for the industry to improve current casting quality, reduce manufacturing
costs and explore new markets for the process. Optimisation of the mechanical and physical properties of the
ceramic shell will be fundamental to achieving these aims. This paper reports on the work carried out to determine
the applicability of the use of ® bres as reinforcement media for investment shells and to explore methods by which
the use of organic ® bres can improve ceramic shell performance. MST/5425

The authors are in the IRC in Materials Processing, The University of Birmingham, Edgbaston, Birmingham B15 2TT, UK
([email protected]). Manuscript received 11 February 2002; accepted 19 March 2002.
# 2002 IoM Communications Ltd. Published by Maney for the Institute of Materials, Minerals and Mining.
Downloaded by [McMaster University] at 13:24 06 April 2016

the shell mould is achieved by removal of moisture.5

Introduction Colloidal binders produce ceramic moulds with very low
green strengths which are prone to cracking during wax
The investment casting process involves the production of removal and handling.6 This is overcome by the use of
engineering castings using an expendable pattern.1 The liquid polymer additions, either latex based for alkaline
principles can be traced back to 5000BC 2 when Early Man systems or PVA based for acidic binders. Polymers increase
employed the method to produce rudimentary tools. This
was followed by centuries of use for jewellery and artistic
components3 before the advent of the Second World War
saw the process developed for aerospace and subsequently
engineering components. 4 The term `investment’ casting
derives from the characteristic use of mobile ceramic slurry,
or `investments’ , to form a mould with an extremely smooth
surface.5 These are replicated from precise patterns and
transmitted in turn to the casting.
The basic steps in the production of an investment cast
component using a ceramic shell mould are shown in
Fig. 1. Primary and secondary ceramic slurries are prepared
from ® ne mesh refractory ® llers and a liquid binder system.
Pattern waxes are produced and an assembly is prepared
according to the required casting speci® cations (Fig. 2). The
assembly is then dipped into the primary slurry, sprinkled
with coarse refractory stucco and dried. The purpose of
the stucco is to minimise drying stresses in the coatings by
presenting a number of stress concentration centres, which
distribute, and hence, lower the magnitude of the local
drying stress. The second main purpose of the stucco is to
present a rough surface, thus facilitating a mechanical bond
between the primary coating and the back up or secondary
investment. When the primary coat has set, (air dried until
the binder gels) the assembly is systematically dipped into
the secondary slurry and stuccoed until the required thick-
ness of shell is built up. The particle size of the stucco is
increased as more coats are added to maintain maximum
mould permeability and to provide bulk to the mould. Each
coating is thoroughly hardened between dippings.
Although a small number of UK foundries are continu-
ing to use alcohol based binders, the majority of larger
companies are switching production to the use of water
based, colloidal silica binders. These binders usually con-
tain between 20 and 30 wt-% amorphous SiO 2 particles,
stabilised by Na + counter ions. Gellation of the colloid to
produce a binding phase for the refractory particles within 1 Basic principles of investment casting process6

DOI 10.1179/026708302225007772 Materials Science and Technology December 2002 Vol. 18 1565
 1566 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting

ratio of 50 : 1, the loading rate chosen was 20 g per litre

of binder liquids. Shorter ® bres can be loaded at higher
rates.

CERAMIC SHELL SPECIFICATIONS

The ceramic shell within this study was designed to be
representative of a standard shell used for aluminium
alloy casting. The slurry speci® cations are contained in
Table 1, shell build in Table 2, and sample details in
Table 3.

ELECTRON MICROSCOPY
Scanning electron microscopy (SEM) was used to view the
a b structure of ceramic mould sections and ® bre morphology.
The instruments used were a Jeol 840A SEM and an
2 Example of standard injected wax pattern and assem- ISI ± 100A SEM. Samples were mounted and gold coated
bly for investment casting before analysis.

the ceramic green strength and have a tendency to reduce ROOM TEMPERATURE FLAT BAR STRENGTH
the level of `wet-back’ of moisture into previously dried MEASUREMENT
coats when new layers are added.7 Strength measurements were carried out in accordance
Thus, a ceramic mould consists of individual layers of with BS 1902. Injected wax bars were used as formers for
Downloaded by [McMaster University] at 13:24 06 April 2016

® ne refractory material and granular refractory material the ceramic shells according to speci® cations given in the
held together by a binder that has been set to a rigid gel. `organic ® bre speci® cation’ and `ceramic shell speci® cations’
Investment casting is a time consuming, labour intensive subsections above. After dipping the shells were steam
process producing complex, high value added components boilerclave dewaxed at 8 bar pressure for 4 min, followed
for a variety of specialised industries in a range of alloys. by a controlled depressurisation cycle at 1 bar min ­ 1 . Test-
Currently the moulds take between 24 and 72 h to produce pieces, of approximately 20 680 mm, were cut using a
due to the need to use controlled moisture removal for grinding wheel and either tested green or ® red at 1000 °C for
water based colloidal gellation on every coat. Drying and 60 min prior to testing. For hot strength measurements,
strength development is the most signi® cant rate limiting ® red samples were tested at 800 °C, at a load application
factor in reduction of lead times and production costs for rate of 5 kg min ­ 1 , in three point bend mode with an outer
the industry. Therefore, scienti® c development within the span of 80 mm.
industry is fuelled by the need towards reduction in pro-
duction times, improvement of shell performance for the
reduction of casting scrap rates and the reduction of metal ± ROOM TEMPERATURE EDGE STRENGTH
mould reaction to improve as cast component surfaces. MEASUREMENT
One recently patented method for reducing shell produc- The edge, or wedge test,9 was developed at The University
tion times and enhancement of ceramic mould properties of Birmingham in conjunction with Rolls Royce plc. The
in the autoclave during wax removal is the use of organic test is speci® cally designed as a quality control procedure to
® bre additions into ceramic slurries,8 thus negating the need determine the actual strength of the ceramic shell at the
for expensive liquid polymer additions. The ® bres, known most vulnerable edges where strength and shell build are
as WexPerm*, are organic and water insoluble. They are vital to prevent cracking. Edge testing was carried out in
therefore dispersed in the binder, as opposed to being accordance with procedures set out in Ref. 9 (see Fig. 3).
dissolved as in the case of water soluble polymers. The ® bres The wedge angle used was 29°. The load was applied at a
are available in a range of sizes and aspect ratios but are rate of 0.25 mm min ­ 1 .
typically less than 3 mm in length. Fibre length and dia-
meter determines both the slurry behaviour and the ceramic PERMEABILITY MEASUREMENT
coating characteristics. Unlike the polymer additions, ® bres
do not create foam and the presence of bubbles in the slurry Impervious mullite rods of dia. 10 mm, to one end of which
is very low, normally zero. a 40 mm diameter table tennis ball has been attached, were
The ® bres have a range of unique properties, leading to used as formers for the ceramic shells according to speci® -
enhanced shell build and the possibility of using fewer coats cations given in the subsection above on `organic ® bre
for equivalent ceramic thickness and mechanical perfor- speci® cation’ (Fig. 4). After drying, the plastic balls were
mance. The following experimentation was designed to burnt out at 700 °C for 5 min at a heating rate of no more
determine the properties of ® bre modi® ed shell systems and than 3 K min ­ 1 . Shell samples were then ® red at 1000 °C for
to compare them to equivalent polymer modi® ed shells, 60 min at a heating rate of 10 K min ­ 1 . Testing was carried
thus allowing the advantages and disadvantages of the use out in accordance with BS 1902:Section 10. 2:1994, at a test
of such a system to be speci® ed for the investment industry. temperature of 800 °C and with a slight adjustment of the
It is important that any potential new material be capable of equation for dynamic viscosity of air.
performing as well as, if not better than current systems.

Results and discussion

Experimental
ORGANIC FIBRE MORPHOLOGY
ORGANIC FIBRE SPECIFICATION Figure 5 shows the structure and morphology of the organic
The ® bres used in this study are 1 mm length nylon with an ® bre additions. The additions are nylon, with a ® bre dia-
average diameter of 19 ± 21 mm in diameter. At this aspect meter of approx 20 mm and a length of 1 mm, giving an
aspect ratio of 50 6. Figure 6 shows a higher magni® cation
*Registered and patented by Wex Chemicals Ltd. micrograph detailing the surface of the ® bres. Of note is

Materials Science and Technology December 2002 Vol. 18

 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting 1567

3 Experimental setup for edge strength testing

Downloaded by [McMaster University] at 13:24 06 April 2016

5 Secondary electron SEM image of nylon ® bre

additions

SHELL BUILD COMPARISONS

Comparisons of the ceramic shell thickness achieved for
polymer modi® ed and ® bre modi® ed shell systems can be
seen in Table 4 and Fig. 7. The use of ® bres increases the
shell thickness by 0.71 mm (15%) over a primary + 4 coat
shell build, which equates to the ability to reduce the
number of secondary coats required to produce an equi-
valent shell thickness. This has signi® cant production time
and material cost implications when producing investment
shells.

Table 3 Sample speci® cations for aluminium shell

4 Schematic section through permeability testpiece Sample Speci® cation Shell build
coated with ceramic shell
1 Polymer Primary 1
Secondary 1
the relatively smooth nature of the nylon surface, which will 2 Fibre Primary 1
have implications for the relative ease of ® bre pull-out Secondary 2
during the application of mechanical force.

Table 1 Slurry speci® cations for aluminium shell preparation*

Binder Silica Polymer Fibre addition, (g L ­ 1 Refractory
Slurry Content, wt-% addition, wt-% of binder liquids) Filler loading

Primary 26 6 NA{ ­ 200 mesh zircon 77% total

­ 200 mesh fused 75% zircon
silica 25% silica
Secondary 1 22 8 NA{ ­ 200 mesh alumino-silicate 57%
Secondary 2 22 N/A 20 ­ 200 mesh alumino-silicate 57%

* All percentages are wt-%.

{NA not available.

Table 2 Shell build speci® cations for aluminium shell coating

Coating Stucco Dip time, s Drain time, s Air speed, ms ­ 1 Dry time, h

Primary 50/80 Molochite 30 60 0. 4 24

Secondary 1 30/80 Molochite 30 60 3 1.5
Secondary 2 30/80 Molochite 30 60 3 1.5
Secondary 3 30/80 Molochite 30 60 3 1.5
Secondary 4 30/80 Molochite 30 60 3 1.5
Seal 30 60 3 24

Materials Science and Technology December 2002 Vol. 18

 1568 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting

6 Higher magni® cation secondary electron SEM images of nylon ® bre additions

FLAT BAR ROOM TEMP FRACTURE

STRENGTH
A comparison of the strengths achieved at room tempera-
Downloaded by [McMaster University] at 13:24 06 April 2016

ture in the three point bend mode for the shell samples is
given in Table 5. The high green strength of Shell 1 is a
direct result of the polymer content, which is re¯ ected by
the reduction in strength as the sample is ® red at 1000 °C
and the polymer burns out. Surprisingly, ® bre reinforced
ceramic has a lower green strength than the polymer
modi® ed system (62% of the polymer strength) and an
equivalent ® red strength suggesting that the ® bre is not
in¯ uencing green strength in the manner of a standard
composite material.
In practice, however, moulds produced with ® bre addi-
tions are less susceptible to autoclave cracking8 suggesting a
higher green strength. Further detailed comparison of load
bearing capacity can be achieved using the adjusted fracture
load (AFL). 1 0 The AFL is de® ned as the load necessary
to break a normalised 10 mm wide testpiece across a 70 mm
span, thus allowing the actual thickness of the shell to 7 Image highlighting the shell thickness increase pro-
become a factor. Results are given in Table 6. It can be seen duced with organic ® bre modi® ed slurries
that the load bearing capacity of the polymer system is still
slightly higher when the moulds are un® red. When ® red, the
® bre modi® ed system has a higher load bearing capacity This suggests, as does the AFL that the ® bre modi® cation
even though both polymer and organic ® bre will have burnt should have adequate mechanical properties in the green
off at these temperatures. The AFL suggests that the ® bre state to withstand the mechanical demands of wax removal.
system will produce a ceramic shell with very little difference Figure 10 shows the stress ± strain results for the ® red
in three point bend strength from the polymer modi® ed polymer modi® ed Shell 1 specimens, tested at room tem-
system making it a viable alternative. perature. The average stress to failure is approximately
4 MPa and the average strain at failure is approximately
2 .25 610 ­ 3 . Figure 11 shows the equivalent test results for
FLAT BAR ROOM TEMPERATURE
STRESS ± STRAIN Table 5 Strength results for shell samples
Figure 8 shows the stress ± strain results for the un® red
Shell Modulus of St.
polymer modi® ed Shell 1 specimens tested at room sample Status rupture, MPa dev. r
temperature. The average stress to failure is approximately
8 MPa and the average strain at failure is approximately Shell 1 Green 7.8 0. 7
2 610 ­ 3 . Figure 9 shows the equivalent test results for the Shell 1 Fired* 4.8 0. 3
® bre modi® ed system (Shell 2), giving an average stress to Shell 2 Green 4.7 0. 4
Shell 2 Fired* 4.7 0. 5
failure of approximately 5 MPa and an average strain to
failure of 1 .5 610 ­ 3 . Though the maximum green strength
* Fired at 1000°C for 60 min.
of the ® bre system in ¯ at bar testing is lower, the actual
strain accommodated by the system is only slightly less.
Table 6 Adjusted fracture load strength for shell samples

Table 4 Shell thickness comparisons Shell Adjusted fracture Standard

sample Status load, N deviation, N
Shell Sample Thickness, St. dev.,
sample Speci® cation Status population mm mm Shell 1 Green 16. 6 1. 9
Shell 1 Fired* 10 1. 2
1 Polymer Green 10 4.6 0.26 Shell 2 Green 13. 8 1. 8
2 Fibre Green 10 5.31 0.39 Shell 2 Fired* 12. 6 1. 5

Materials Science and Technology December 2002 Vol. 18

 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting 1569

11 Stress ± strain diagrams at room temperature for ® red

® bre modi® ed Shell 2
8 Stress ± strain diagrams at room temperature for
un® red polymer modi® ed Shell 1 the ® bre modi® ed system (Shell 2), giving an average stress
to failure of approximately 4 .5 MPa and an average strain
to failure of 1 .5 610 ­ 3 . Though the maximum strength of
Downloaded by [McMaster University] at 13:24 06 April 2016

both shell systems decreases due to loss of organic material,

the strain to failure remains very similar to that experienced
by the un® red ceramic. Again this suggests, as does the
AFL that the ® bre modi® cation will produce a shell with
adequate mechanical properties to withstand metal casting.

ROOM TEMPERATURE EDGE STRENGTH

The apparent reduction in green strength exhibited by
organic ® bre modi® ed ceramics may be as a result of the test
system chosen. It is necessary to compare the shell systems
using a test more speci® cally designed to indicate mould
strength during the dewax procedure. Most autoclave shell
cracking and damage occurs on the edges of components,
such as trailing edges and sharp corners, where reduced shell
build and high stress results in ceramic failure.1 1 The edge
test is designed to ascertain the load bearing capacity of the
shell in the most vulnerable thin sections and edges of the
moulds.
Table 7 gives the comparison of the edge strength data.
9 Stress ± strain diagrams at room temperature for
un® red ® bre modi® ed Shell 2
Again, the ® bre-modi® ed shell appears to exhibit a lower
green strength than the polymer system. In terms of edge
strength, however, the ® bre system is exhibiting a value
which is 84% of the polymer shell which is closer than that
obtained with ¯ at bar tests. The ® red strength of the ® bre
shell appears lower (73% of polymer system), which would
be expected with a shell system containing increased poro-
sity on vulnerable edges.
The edge test results also begin to give clues as to the
mechanism by which the ® bre shells prevent autoclave
cracking. The shell build on the edges is much higher, ® bres
giving an increase of 50% of that of the polymer edge thick-
ness. This is much higher than the 15% thickness increase
exhibited on ¯ at specimens. The slurry and stucco retention
due to the presence of ® bres is easily noted during the dipping
process. Fibre modi® ed slurries have a much reduced drain
time, possibly due to ® bre alignment or electrostatic forces

Table 7 Comparison of edge strength results for shell

systems
Edge strength, Edge thickness,
Shell Status Stucco MPa (mm)

1 Green Fluid bed 1.58+ 0.22 2.49+ 0.23

1 Fired* Fluid bed 1.44+ 0.26 2.43+ 0.26
2 Green Fluid bed 1.32+ 0.18 3.51+ 0.35
2 Fired* Fluid bed 1.15+ 0.19 3.63+ 0.51
10 Stress ± strain diagrams at room temperature for ® red
polymer modi® ed Shell 1 * Fired at 1000°C for 60 min.

Materials Science and Technology December 2002 Vol. 18

 1570 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting

Table 8 Comparison of AFL edge strength results for using linear regression. The results are tabulated in Table 9.
shell system In the un® red state, the ® bre shell has a lower modulus
Shell Adjusted Standard
and therefore will be softer and more likely to deform. This
sample Status fracture load, N deviation, N reduced stiffness may be preventing elastic cracking during
dewax and allowing the ® bre shell to deform slightly with-
Shell 1 Green 1. 7 0.3 out failure. Unlike the polymer shell, the ® bre shell does
Shell 1 Fired* 1. 4 0.2 not reduce its modulus during the ® ring cycle. Again,
Shell 2 Green 2. 8 0.6
Shell 2 Fired* 2. 6 0.6
this may be advantageous in the casting cycle where high
metal weights require a strong shell, which should undergo
* Fired at 1000°C for 60 min. minimum deformation while the metal is still liquid. The
higher modulus may also result in a stiffer shell, which is
easier to remove after casting.
generated by friction in the slurry. Reduced drain times
lead to increased slurry coat thickness, which in turn allows
the pick up and retention of extra stucco grains.
The normalised AFL results for the edge strength are FLAT BAR HIGH TEMPERATURE (800°C)
given in Table 8. These results show clearly that the load STRENGTH
bearing capacity of the ® bre-modi® ed shells is much higher The strength at 800 °C for the polymer modi® ed shell is
on the vulnerable edges where the majority of autoclave tabulated in Table 10. The average strength at 800 °C is
cracking takes place. In the green state the shells have a 65% 2 .8 MPa, which is 58% of the ® red strength at room tem-
higher load bearing capacity. In the ® red state, the strength perature. As before, the AFL normalises the data, giving
increase is equivalent to 86% higher than the polymer a load bearing capacity for this shell system of 5 .4 N at
system. Thus, due to the increased shell build on the edges 800 °C. The strength at 800 °C for the ® bre modi® ed shell is
of components, the ® bre systems have a much higher load tabulated in Table 11. The average at 800 °C is 2 .0 MPa,
Downloaded by [McMaster University] at 13:24 06 April 2016

bearing capacity in both the green and ® red states. The which is 43% of the ® red strength at room temperature. As
increase is suf® cient for the shells to be produced with fewer before, the AFL normalises the data, giving a load bearing
coats and still maintain an equivalent, if not superior, capacity for this shell system of 6 .2 N at 800 °C.
strength and performance in both the green and ® red states. The hot strengths for both shell systems are approxi-
mately equivalent when measured using standard ¯ at bar test
methods, suggesting no difference in mechanical properties
APPARENT YOUNGS MODULUS E during the casting cycle. The AFL for the ® bre modi® ed
As the stress ± strain results exhibit elastic behaviour for the shell is slightly higher, giving a slightly higher load bearing
shell systems, taking only the straight line elastic portion of capacity for the ® red shell. Therefore, it will be possible to
the plots into account, it is possible to calculate an apparent reduce the number of coats applied to ® bre-modi® ed shells
Young’s modulus (measure of the stiffness of a material) and still maintain an equivalent hot strength.

Table 9 Comparison of Young’s modulus results for shell systems

Average youngs Standard
Sample Description Status modulus, GPa deviation r , GPa

1 Polymer modi® ed Green 4.359 0.439

1 Polymer modi® ed Fired 2.610 0.387
2 Fibre modi® ed Green 3.668 0.491
2 Fibre modi® ed Fired 3.365 0.676

Table 10 Flat bar strength at 800°C for the polymer modi® ed Shell 1
Sample Status, °C Thickness, m Width, m Max. load, N MOR, MPa psi AFL, N

1 800 0.00438 0.02149 11.6 3.0 428. 49 5. 7

2 800 0.00446 0.01951 9.8 2.7 384. 56 5. 3
3 800 0.00458 0.02026 12 3.0 430. 01 6. 2
5 800 0.00437 0.01837 10 3.0 434. 11 5. 7
6 800 0.00437 0.02037 9.6 2.6 375. 83 4. 9
7 800 0.00432 0.02014 9.8 2.7 397. 07 5. 1
8 800 0.00449 0.02087 10 2.5 361. 96 5. 0
Average 0.00442 10.4 2.8 401. 7 5. 4
1.0 0.2 29. 3 0. 5

Table 11 Flat bar strength at 800°C for the ® bre modi® ed Shell 2
Sample Status, °C Thickness, m Width, m Max. load, N MOR, MPa psi AFL, f~0.1

1 800 0.00534 0.02089 10.5 1.9 268.43 5.3

2 800 0.00601 0.02127 12.9 1.8 255.71 6.4
3 800 0.00594 0.01846 10.6 1.7 247.84 6.0
4 800 0.00552 0.01798 13.2 2.5 366.92 7.7
5 800 0.00572 0.0199 13.4 2.2 313.42 7.1
6 800 0.00601 0.02065 12.5 1.8 255.22 6.4
7 800 0.00506 0.02053 9.7 1.9 281.03 5.0
8 800 0.00561 0.01942 11 1.9 274.09 5.9
9 800 0.00556 0.01991 14.4 2.5 356.30 7.6
Average 0.00565 11.7 2.0 282.8 6.2
1.4 0.3 39.7 0.9

Materials Science and Technology December 2002 Vol. 18

 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting 1571

13 Secondary electron SEM image of green fracture sur-

12 Comparison of ® red permeability at 800°C for shell face of ® bre modi® ed ceramic Shell 2 showing intact
samples ® bres protruding from the surface

FLAT BAR HIGH TEMPERATURE (800°C) casting cycle. The slightly higher standard deviation does,
STRESS ± STRAIN however, suggest that the use of ® bres produces a slightly
more inconsistent shell structure. These results suggest that
Figure 12 shows the stress ± strain results for the polymer the ® bre system will produce a ceramic shell with greatly
modi® ed Shell 1 specimens (® red) tested at 800 °C. The increased ® red permeability from the polymer modi® ed
average stress to failure is approximately 2 .8 MPa and the system making it a enhanced alternative.
average strain at failure is approximately 0. 18%. Figure 13
Downloaded by [McMaster University] at 13:24 06 April 2016

shows the equivalent test results for the ® bre modi® ed

system, giving an average stress to failure of approximately MORPHOLOGY OF GREEN AND FIRED
2 .0 MPa and an average strain to failure of 0 .28%. The FRACTURE SURFACES
strengths of the two systems are approximately equivalent Un® red and ® red fracture surfaces of ® bre modi® ed shell
at casting temperatures. However, the strain accommodated samples were analysed using SEM to determine ® bre
by the ® bre modi® ed shell is 0.1% higher at casting tem- alignment and the effect of organic ® bre additions during
peratures suggesting that the shell would be less likely to fail fracture. Figure 13 shows the green fracture surface with
during the casting process due to shell wall movement. ® bres, which appear to have been pulled from the ceramic
network. This is as expected with a composite system of this
FIRED PERMEABILITY nature, where the energy required to remove ® bres in this
manner increases the strength required to break the sample
During casting, liquid metal at high temperature is poured
and ultimately the strength.
into the ceramic mould and allowed to cool. The mould
However, Fig. 13 also illustrates that the ® bres appear
itself is full of air and this is displaced by incoming
to be intact, with the end chopped as per the virgin ® bre
liquid metal. It is important that this air leaves the cavity
(Fig. 5). This suggests that there was very little resistance
quickly and completely to prevent trapping and subsequent
and little force required to pull the ® bres from the matrix,
porosity, bubbles and more seriously mis-run or `non-® ll’ of
which explains the relatively low green strength recorded for
areas of the components where trapped air has produced a
these samples. The low resistance is undoubtedly a result of
positive back pressure and prevented ¯ ow of liquid metal.
There are several mechanisms by which air is removed: the very smooth nature of the nylon ® bre surface, illustrated
direct displacement through the pouring cup; absorption earlier in Fig. 6 and again in Fig. 14 where the ® bre cavity
into the alloy; and through the permeation of air through after pull out has very smooth walls. Figure 15 shows the
the open porosity network of the ceramic shell. Permea- ® red fracture surface of the ® bre modi® ed shell sample. The
bility of the shell is the most important factor, with most cavities left behind by the volatilisation of the nylon during
investment shells having a minimum of 30% open porosity the ® ring process can be clearly seen. The connected nature
in the structure.6 Mould ® ll is improved by increasing the of this porosity enhances the porosity of the ® red shell,
open porosity and hence the permeability of the ceramic providing a great deal of connected pathways through the
shell. A comparison of the permeability of the shell samples ceramic and greatly increasing the permeability of the cera-
at 800 °C after previously being ® red at 1000 °C to simulate mic during casting.
casting conditions can be seen in Table 12 and represented
graphically in Fig. 12. The addition of 1 mm nylon ® bres
dramatically increases the ® red permeability by a factor of 3 Summary and conclusions
at aluminium alloy casting temperatures.
The permeability is unaffected by time at temperature It is becoming imperative that the investment casting indus-
and remains consistent throughout the test. This is of great try improves current casting quality, reduces manufacturing
bene® t to the casting process, allowing increased transport costs and explores new markets in order to remain competi-
of air displaced from the mould cavity and providing tive. Optimisation of the mechanical and physical properties
a consistent data input for computer simulations of the of the ceramic shell will be fundamental to achieving these

Table 12 Permeability results at 800°C for shell samples (® red at 1000°C for 60 min prior to testing)
Shell Number Temperature, Test duration, Permeability, Standard deviation,
sample of samples °C min m2 610 ­ 1 3 m2 610 ­ 1 3

Shell 1 3 800 5 1.612 0.183

10 1.601 0.152
15 1.625 0.126
Shell 2 3 800 5 5.793 0.239
10 5.458 0.332
15 5.435 0.334

Materials Science and Technology December 2002 Vol. 18

 1572 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting

15 Micrographs of ® red fracture surface (SEM) showing

cavities left behind by volatilised nylon ® bres

the ® bre system is exhibiting a value which is 84% of the

polymer shell which is closer than that obtained with ¯ at
bar tests. The ® red strength of the ® bre shell appears lower,
which would be expected with a shell system containing
increased porosity on vulnerable edges.
The edge test results also begin to give clues as to
the mechanism by which the ® bre modi® ed shells prevent
autoclave cracking. The shell build on the edges is much
Downloaded by [McMaster University] at 13:24 06 April 2016

14 Secondary electron SEM image of green fracture sur- higher, ® bres giving an increase of 50% of that of the
face of ® bre modi® ed ceramic Shell 2 showing ® bre polymer edge thickness. This is much higher than the 15%
pull-out from opposing surface thickness increase exhibited on ¯ at specimens. The slurry
and stucco retention due to the presence of ® bres is easily
aims. Production of the mould is time consuming, currently noted during the dipping process. Fibre modi® ed slurries
taking between 24 and 72 h depending upon the compo- have a much reduced drain time, possibly due to ® bre align-
nent, due to the need to use controlled moisture removal. ment or electrostatic forces generated by friction in the
Drying and strength development are the most signi® cant slurry. Reduced drain times lead to increased slurry coat
rate limiting factors in the reduction of lead times and thickness, which in turn allows the pick up and retention of
production costs for the industry. extra stucco grains.
The recently patented method of incorporating organic The normalised AFL results show clearly that the load
® bres into the binder to improve performance and reduce bearing capacity of the ® bre modi® ed shells is much higher
shell production times is proving successful in foundry trials. on the vulnerable edges where the majority of autoclave
The most common ® bre additions are nylon with a diameter cracking takes place. In the green state the shells have a 65%
of 20 mm and a length of 1 mm, giving an aspect ratio of higher load bearing capacity. In the ® red state, the strength
50 6. These ® bres are designed to act as a composite rein- increase is equivalent to 86% higher than the polymer
forcing agent, increasing green strength to prevent cracking system. Thus, due to the increased shell build on the edges
during autoclave wax removal and burning out during ® ring of components, the ® bre systems have a much higher load
to leave increased porosity within the ceramic structure. bearing capacity in both the green and ® red states. The
The use of ® bres increases the shell thickness by a factor increase is suf® cient for the shells to be produced with fewer
of 15% over that of a comparable system using a latex coats and still maintain an equivalent, if not superior,
modi® ed binder. This equates to the ability to reduce the strength and performance both green and ® red.
number of coats applied to the wax, which has signi® cant The hot strength for both shell systems is approxi-
production time and material cost implications for foun- mately equivalent when measured using standard ¯ at bar
dries. A comparison of strengths at room temperature, test methods, suggesting no difference in mechanical proper-
measured in three point bend mode, suggests that the ® bre ties during the casting cycle. The AFL for the ® bre modi® ed
addition produces a shell system with a slightly lower load shell is slightly higher, giving a slightly higher load bearing
bearing capacity in the green state but still high enough to capacity for the ® red shell. Therefore, it will be possible to
make the shell a viable alternative. reduce the number of coats applied to shells and still maintain
The apparent loss of green strength, a factor which is not an equivalent hot strength. The strengths of the two systems
seen in industrial practice, may be a result of the test system are approximately equivalent at casting temperatures. How-
chosen as most autoclave cracking occurs on the edges ever, the strain accommodated by the ® bre modi® ed shell is
of components where shell build is an important factor. 0 .1% higher at casting temperatures suggesting that the shell
Another factor responsible for this may be the smooth nature would be less likely to fail during the casting process.
of the nylon ® bre surface, appearing to result in very easy Another signi® cant advantage of the ® bre additions
® bre pull-out from the ceramic matrix. With little resistance appears to be the increased porosity and ultimately per-
there will be little extra energy expended for fracture and the meability of the ceramic shell after ® ring. The addition of
shell will remain weak. Though the maximum green strength ® bre increases the permeability at casting temperatures by a
of the ® bre system in ¯ at bar testing is lower, the actual strain factor of 3, which will be an enormous bene® t when reducing
accommodated by the system is only slightly less. This defects associated with air entrapment. The permeability is
suggests, as does the AFL, that the ® bre modi® cation will unaffected by time at temperature and remains consistent
have adequate mechanical properties in the green state. throughout the test. This is of great bene® t to the casting
The edge, or wedge test,9 was developed at The Univer- process, allowing increased transport of air displaced from
sity of Birmingham in conjunction with Rolls-Royce plc. the mould cavity and providing a consistent data input
The test is speci® cally designed as a quality control proce- for computer simulations of the casting cycle. These results
dure to determine the actual strength of the ceramic shell suggest that the ® bre system will produce a ceramic shell
at the most vulnerable edges where strength and shell build with greatly increased ® red permeability from the polymer
are vital to prevent cracking. In terms of edge strength, modi® ed system making it a enhanced alternative.

Materials Science and Technology December 2002 Vol. 18

 Jones et al. Organic ® bre modi® ed ceramic shell moulding for investment casting 1573

Further work willbe needed to quantify the completebene® ts 2. p. r. taylor: Proc. 17th Annual BICTA Conf., September
of this development in shell technology but all the initial results 1983; 1983, Birmingham, BICTA.
are suggesting that the use of organic ® bre reinforcement will 3. e. l. kotzin: `Metalcasting and molding processes’ ; 1981, Des
Plaines, IL, American Foundrymen s Society.
have enormous time, cost, and weight saving implications for
4. s. o. barnett: Foundry Trade J. Int., 1988, 11, (3), 33 ± 37.
the standard investment casting foundry. 5. s. jones and p.m. marquis: Br. Ceram. Trans., 1995, 94, 68 ± 73.
6. s. jones: `Improved sol based ceramic moulds for use in invest-
ment casting’ , PhD thesis, The University of Birmingham, UK,
Acknowledgements May 1993.
7. s. jones and s. p. leyland: Proc. 22nd BICTA Tech. Conf.
on `Investment casting’ , Bath, UK, September 1995; 1995,
The authors wish to gratefully acknowledge the ® nancial Birmingham, BICTA.
support of the EPSRC through the grant GR/M60101 8. d. duffey: Proc. 25th BICTA Conference, Cheltenham, UK,
(FOCAST) and the industrial partners of this project June 2001; 2001, Birmingham, BICTA.
(irc.bham.ac.uk/epsrc/focast/Who_we_are.html). 9. r. hyde, p. withey, and s. p. leyland: in Proc. 9th World
Conference on `Investment casting’ , San Francisco, USA, Sept
1996; 1996, Dallas, TX, Investment Casting Institute.
10. j. vandermeer: Proc. 10th World Conf. on `Investment
References casting’ , Monte Carlo, Monaco, May 2000, Paper 3.
11. s. p. leyland, r. hyde, and p. a. withey: Precast 95, Proc. 8th
1. p. r. beeley and r. f. smart: `Investment casting’ , 1st edn; 1995, Int. Symp. on `Investment casting’ , Brno, Czech Republic, May
London, The Institute of Materials. 1995.
Downloaded by [McMaster University] at 13:24 06 April 2016

Materials Science and Technology December 2002 Vol. 18