Available online at www.sciencedirect.

com

Journal of the European Ceramic Society 33 (2013) 3397–3402

Investigation on cristobalite crystallization in silica-based ceramic cores for

investment casting
A. Kazemi a,∗ , M.A. Faghihi-Sani b , H.R. Alizadeh c
a Department of Materials Engineering, Science & Research Branch, Islamic Azad University, Tehran, Iran
b Department of Materials Science and Engineering, Sharif University of Technology, Tehran, Iran
c MAPNA Turbine Blade Eng. & Mfg. Co., PARTO, Iran

Received 26 February 2013; received in revised form 26 June 2013; accepted 29 June 2013
Available online 2 August 2013

Abstract
In this work, cristobalite crystallization and its effects on mechanical and chemical behaviour of injection moulded silica-based ceramic cores were
investigated. In order to simulate casting process condition, the sintered samples at 1220 ◦ C were also heated up to 1430 ◦ C. Flexural strength test
was carried out on both sintered and heat treated samples. Chemical resistance of the cores was evaluated by leaching the samples inside 43 wt%
KOH solution at its boiling point. Phase evolution and microstructure were investigated by thermal analyses (DTA and DSC), X-ray diffraction
(XRD), scanning electron microscopy (SEM) and optical microscopy (OM). Results showed that cristobalite was crystallized on the surface of
fused silica grains at about 1380 ◦ C. Flexural strength of the sintered cores was decreased after simulated casting heat treatment due to cristobalite
phase transformation. The formed cristobalite on the surface of fused silica grains dramatically decreased the leachability of ceramic cores.
© 2013 Elsevier Ltd. All rights reserved.

Keywords: In situ cristobalite; Crystallization; Leachability; Fused silica; Ceramic core

1. Introduction In this study, crystallization of fused silica and its effects on

Ceramic cores with very precise and often complex shapes are
manufactured by injection moulding to achieve a near net shape.
They are extensively used in turbine blade casting. Proper match
between the ceramic core, investment casting ceramic mould and
metal blade is essential for the proper performance of the cast-
ing process.1 Silica-based ceramic cores are extensively used for
casting nickel based blades with casting temperature of lower
than 1500 ◦ C due to their high dimensional stability, thermal
shock resistance, and leachability.2 These cores are usually com-
posed of silica, zircon, and negligible amounts of cristobalite
and alumina.3,4 The ␤-cristobalite, as a high temperature poly-
morph of silica, undergoes ␤ → ␣ phase transformation with a
volume contraction of about 5 vol% when it is cooled down to
200–270 ◦ C.5,6
∗ Corresponding author. Tel.: +98 912 225 0115; fax: +98 2144869783.
E-mail addresses: arg.kazemi@gmail.com, a cercis@yahoo.com,
a.kazemi@srbiau.ac.ir (A. Kazemi).
the most important properties of injection moulded silica-based
ceramic cores, including flexural strength and leachability, have
been investigated.
2. Experimental procedure
The silica-based ceramic cores containing 65.2 wt% fused
silica (Remet, purity 97.3%; d50 = 27.7 ␮m), 16 wt% zircon
(Cookson, purity 88%; d50 = 20.4 ␮m), 2.8 wt% cristobalite
(Sibelco, purity 99.4%; d50 = 14.1 ␮m) and 1.2 wt% alumina
(Ceramatco, purity 96.9%; d50 = 9.4 ␮m) were fabricated
through injection moulding.
To investigate crystallization temperature of fused silica
in the prepared ceramic cores, Differential Thermal Analysis
(DTA) was carried out on a 7 mm × 7 mm × 5 mm LWH sam-
ple from room temperature up to 1430 ◦ C with a heating rate
of 10 ◦ C min−1 . In addition, differential scanning calorimetry
(DSC) analysis was also conducted to confirm in situ cristo-
balite formation. The fused silica powders with various grain
sizes (as shown in Table 1) were also heated at various temper-
atures (1200, 1300 and 1400 ◦ C) for 1 h and then were analyzed

0955-2219/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jeurceramsoc.2013.06.025
3398 A. Kazemi et al. / Journal of the European Ceramic Society 33 (2013) 3397–3402

Fig. 1. DTA and DDTA graphs of 7 mm × 7 mm× 5 mm ceramic core from 25 to 1430 ◦ C as well as DSC results of the sintered sample, the sintered sample after
DTA test and the simulated casting heat treated sample (all with heating rate of 10 ◦ C min−1 ).

by an X-ray powder diffractometer with Ni-filtered CuK␣ radi- with span distance of 100 mm and 6 N s−1 loading rate) accord-
ation, (PW1800-Philips) with 0.04 step size and 4 s detecting ing to ASTM C674. Each flexural strength result was an average
time, in order to study the fused silica crystallization. Polished of five bending tests.
cross sections and fractured surfaces of the prepared samples The sintered specimens, before and after simulated cast-
were also monitored by optical microscope (Olympus BH-2) and ing heat treatment, were cut into 22 mm × 10 mm × 5 mm LWH
scanning electron microscope (SEM, VEGA TESCAN, Czech, pieces and were leached inside a 43% KOH solution at its boil-
15 kV). ing point for 15 and 30 min, respectively. Then, leaching rate
To study the effects of alumina addition on fused silica was calculated according to Eq. (1), where W was the total
crystallization, two series of compacts containing fused silica weight loss (g), A was the initial surface area (cm2 ), t was the
particles with and without 10 wt% fine alumina particles were leaching time (h), and ρ was the specimen density (g cm−3 ).
sintered at 1220 ◦ C for 6 h, and some of them were subsequently
heat treated at 1400 ◦ C for 1 h. For semi-quantitative analysis, W
K= (1)
10 wt% zircon powder was added to the heat treated samples Aρt 1/2
before XRD test. Then, intensity ratio of cristobalite peak (1 0 1)
to zircon peak (2 0 0) was calculated for each sample. This equation provides leaching rate independently of the sam-
In order to investigate the effect of in situ formed ple size and shape, which fits well with diffusion controlled
cristobalite on mechanical and chemical behaviour of the model for leaching process.7
silica-based ceramic core, the injection moulded test bars of
150 mm × 22 mm × 5 mm LWH were first sintered at 1220 ◦ C
for 6 h, and were subsequently subjected to a simulated casting 3. Result and discussion
heat treatment at 1430 ◦ C for 30 min. Then, flexural strengths of
the sintered samples, before and after heat treatment, were mea- 3.1. Cristobalite crystallization
sured by a three-point bending test (SANTAM STM-5 machine
The derived differential thermal analysis (DDTA) graph of
the sintered ceramic core at 1220 ◦ C in Fig. 1 showed two peaks
Table 1 at 230 and 1380 ◦ C. The first one indicates ␣ → ␤ polymorphic
Specifications of the used fused silica powders. transformation of primary cristobalite, and the second one is
Powder PSD Specific surface related to the crystallization of cristobalite.
area (m2 g−1 ) DSC thermal analyses of the sintered sample, the sintered
D10 (␮m) D50 (␮m) D90 (␮m) sample after DTA test (heated up to 1430 ◦ C with a rate of
Fused silica (1 0 0) 4.37 49.6 177.36 0.513
10 ◦ C min−1 ), and the simulated casting heat treated sample, in
Fused silica (2 3 0) 2.74 27.7 105 0.777 Fig. 1, also show that intensity of the cristobalite polymorphic
Fused silica (4 0 0) 1.45 6.58 79.73 1.830 transformation peak (around 200 ◦ C) increases after the DTA
test or after the simulated casting heat treatment, demonstrating
 A. Kazemi et al. / Journal of the European Ceramic Society 33 (2013) 3397–3402
Fig. 2. XRD patterns of the fused silica after heat treatment at 1200, 1300 and
1400 ◦ C.
Fig. 3. XRD patterns of various fused silica grains with different sizes after heat
treatment at 1400 ◦ C.
Fig. 4. Optical image of a crystallized silica grain after heating at 1430 ◦ C for
30 min.
3399
Fig. 5. SEM micrograph of a crystallized silica grain after heating at 1430 ◦ C
for 30 min at different magnifications (a) 2000× and (b) 3000×.
that the second peak in DTA graph (at 1380 ◦ C) is related to
in situ cristobalite crystallization.
XRD patterns of the fused silica powder, after heating at
various temperatures in Fig. 2, confirm the DDTA results. Cristo-
balite crystallization was started at 1300 ◦ C and was obviously
occurred at 1400 ◦ C.
XRD patterns of the fused silica powders with various grain
sizes, after heating at 1400 ◦ C for 1 h in Fig. 3, indicate that
cristobalite peaks intensity increases by decreasing the particle
3400 A. Kazemi et al. / Journal of the European Ceramic Society 33 (2013) 3397–3402

Fig. 6. XRD patterns of the fused silica compacts with and without 10 wt% alumina after sintering at 1220 ◦ C for 6 h and subsequent heating at 1400 ◦ C for 1 h.

size or increasing the specific surface area of the fused silica.
These results demonstrate that crystallization occurs on the sur-
face of fused silica grains without the need for the presence of
primary cristobalite as a seed.
Optical image from polished section of the heat treated silica-
based ceramic core in Fig. 4 clearly shows the crystallized layer
on the surface of fused silica grains.
Electron microscope image of the crystallized layer on the
surface of fused silica grains, after heat treating at 1430 ◦ C for
30 min, is also shown in Fig. 5(a). Fig. 5(b) also illustrates the
formed microcracks near the surface of fused silica grains, which
grow toward the centre of particles due to phase transformation
(␤ → ␣) of the in situ formed cristobalite during cooling, and
consequently the introduction of tensile stresses on the surface
of silica grains.
3.2. The effect of alumina addition on the crystallization of
fused silica
3.3. The effects of cristobalite crystallization on the
properties of ceramic core
Cristobalite crystallization on the surface of silica grains
during simulated casting heat treatment can affect the most
important behaviours of silica-based ceramic cores. In this
study, mechanical and chemical properties of the sintered
ceramic cores, before and after heat treatment, have been
compared to investigate the effects of the in situ formed cristo-
balite.
As shown in Fig. 8, flexural strength of the simulated casting
heat treated samples is 25% lower than that of the sintered ones
due to cristobalite crystallization at high temperatures, and con-
sequently its phase transformation from ␤ to ␣ during cooling
below 250 ◦ C. This phase transformation introduces consider-
able microcracks into ceramic core microstructure (as illustrated
in Fig. 9), consequently decreasing the flexural strength.9,10
Moreover, leaching test results in Fig. 10 indicate a tremen-
dous difference in leaching rate of the sintered samples, before

Fig. 6 presents XRD patterns of the heated fused silica com-

pacts with and without 10 wt% alumina addition. As it seems,
the cristobalite peaks intensity increases in the presence of alu-
mina. The peak intensity ratio of cristobalite to zircon for two
series of samples in Fig. 7 confirms the above results.
Alumina, as a component of ceramic core, can affect the
crystallization rate of fused silica through its solution inside
amorphous silica grains. In silicate glasses containing alkali and
alkaline earth oxide as well as Al2 O3 , when the molar ratio
of alumina to alkali and alkaline oxide is less than unit, the
Al3+ is believed to substitute Si in SiO4 tetrahedra. In this case
the addition of Al2 O3 introduces only 1.5 oxygen per network-
forming cation; therefore nonbridging oxygen are used up and
convert to bridging oxygen. On the contrary, in the absence of
alkali and alkaline oxides, the dissolved alumina acts as a mod-
ifier, increases the nonbridge oxygen atoms and consequently
decreases the viscosity of silica glass, increases the mobility of
SiO4 tetrahedrons and leads to higher crystallization tendency Fig. 7. The effect of alumina addition on crystallization of fused silica after
of the silicate network.8 sintering at 1220 ◦ C for 6 h and subsequent heating at 1400 ◦ C for 1 h.
 A. Kazemi et al. / Journal of the European Ceramic Society 33 (2013) 3397–3402
Fig. 8. Flexural strength results of the sintered and simulated casting heat treated
samples.
and after heat treatment. Since cristobalite has higher chemical
resistance to caustic solution than fused silica, the formation of
in situ cristobalite layer on the surface of fused silica grains,
especially after simulated casting heat treatment, can protect
them from corrosion, consequently decreasing the leachabil-
ity of ceramic core intensively. At the same time, the primary
cristobalite particles can increase the leachability by introduc-
ing several microcracks, opening leaching solution diffusion
path into the ceramic core body, and increasing the chemical
interaction between them.
Fig. 9. SEM micrograph of the polished section of ceramic core after simulated
casting heat treatment at 1430 ◦ C.
3401
Fig. 10. Leaching rate of the sintered samples, before and after simulated casting
heat treatment.
4. Conclusion
In this work, the fused silica crystallization and its
effects on mechanical and chemical behaviour of injection
moulded silica-based ceramic cores have been investi-
gated. Results demonstrated that cristobalite crystallization
occurred on the surface of fused silica grains at about
1380 ◦ C. It was also found that the primary cristobalite had
no considerable effect on the fused silica crystallization.
However, alumina, as a network modifier, decreased the melt-
ing point and viscosity of amorphous silica, subsequently
providing better conditions for fused silica crystalliza-
tion.
Cristobalite in situ crystallization on the surface of fused sil-
ica grains decreased the flexural strength and leachability of the
ceramic core intensively. However, in the presence of primary
cristobalite particles, the leachability of fused silica ceramic core
was improved by introducing several microcracks, opening the
KOH solution diffusion path into the ceramic core body, and
increasing their chemical interaction.
Acknowledgments
The authors gratefully acknowledge the MAPNA tur-
bine Blade Eng. & Mfg. Co. (PARTO) and Research and
Development Department of MAPNA and PARTO for their
funding.
References
1 Rotger JC. What is a good ceramic core? INCAST 2008;6:21.
2 Huang L, Duffrène L, Kieffer J. Structural transitions in silica glass:
thermo-mechanical anomalies and polyamorphism. J Non-Cryst Solids
2004;349:1–9.
3 Edirisinghe MJ, Evans JRG. Review: fabrication of engineering ceramics
by injection molding. I. Materials selection. Int J High Technol Ceram
1986;2:1–31.
3402 A. Kazemi et al. / Journal of the European Ceramic Society 33 (2013) 3397–3402
4 Pattnaik S, Karunakar DB, Jha PK. Developments in investment casting
process – a review. J Mater Process Technol 2012;212:2332–48.
5 Krug S, Evans JRG, ter Maat JH. Residual stresses and cracking in large
ceramic injection mouldings subjected to different solidification schedules.
J Eur Ceram Soc 2000;20:2535–41.
6 Wilson PJ, Blackburn S, Greenwood RW, Prajapti B, Smalley K. The
role of zircon particle size distribution, surface area and contamination
on the properties of silica–zircon ceramic materials. J Eur Ceram Soc
2011;31:1849–55.
7 N. D. Sangeeta, Method of dissolving or leaching ceramic cores in airfoils,
US patent, Jul 1998, 779–809.
8 Kingery WD. Introduction to ceramics. New York: J Wiley; 1975.
9 Zawrah M, Hamzawy E. Effect of cristobalite formation on sinterabil-
ity, microstructure and properties of glass/ceramic composites. Ceram Int
2002;28:123–30.
10 Svoboda R, Málek J. Interpretation of crystallization kinetics results pro-
vided by DSC. Thermochim Acta 2011;526:237–51.