Science of Sintering, 50 (2018) 205-215

________________________________________________________________________

doi:https://doi.org/10.2298/SOS1802205C

UDK 549.632; 539.4

Effect of Al(OH)3 Content on the Microstructure and Strength
of Porous Cordierite-Mullite Ceramics Prepared by an In-Situ
Pore Forming Technique

Qingjie Chen, Wen Yan*), Nan Li, Xiaoli Lin, Zhenyan Zhang, Bingqiang
Han, Yaowu Wei
The State Key Laboratory of Refractories and Metallurgy, Wuhan University of
Science and Technology, Wuhan, China

Abstract:
Five porous cordierite-mullite ceramics with similar porosity and different neck
characteristics were prepared from Al(OH)3, magnesite, silica and clay using an in-situ pore-
forming technique. The phase composition, pore and neck characteristics and strength of the
porous ceramics were investigated by an X-ray diffractometer (XRD), a scanning electron
microscopy (SEM) and a microscopy measured method, etc. The experimental results showed
that Al(OH)3 content had a significant effect on the pore size distribution and neck
characteristics (neck size distribution, total value of neck size and phase composition) and
then affecting the strength. With an increase in Al(OH)3 content, the median pore size
decreased, the total length of necks and the uniformity of neck size increased, also the mullite
content of necks increased, resulting in the increase of strength of the porous cordierite-
mullite ceramics. When the Al(OH)3 content was 64.9 wt%, the porous cordierite-mullite
ceramics had the best performance of high apparent porosity of 45.1 % and high compressive
strength of 55.9 MPa.
Keywords: Porous cordierite-mullite ceramics; Neck characteristics; Pore size distribution;
Phase composition; Strength.

1. Introduction

Porous cordierite ceramic has been widely used in insulating refractories due to its
low thermal conductivity coefficient and high thermal-shock resistances at elevated
temperatures [1-3]. However, some disadvantages of cordierite, such as a narrow range of
firing temperature, a low refractoriness, and poor mechanical properties, limited its
application in high-temperature industry. Mullite has a high strength, a high refractoriness, as
well as a good thermal-shock resistance [4, 5], therefore, the properties of porous cordierite
can be improved by adding mullite as a reinforcing phase [6-8].
In recent years, considerable attention had been devoted towards the preparation
methods of porous ceramics. Usually, a porous structure could be achieved by adding some
pore-forming agents [9, 10], by gel-casting [11, 12] or by foaming method [13], etc.
Benhammou [10] fabricated porous cordierite ceramics using oil shale as a natural pore-
forming agent. Zhang [14] prepared the cordierite-mullite-alumina ceramic foams by
impregnation of polymer performs with an optimized stock suspension. Suzuki [15] fabricated
_____________________________
*)
Corresponding author: yanwen@wust.edu.cn
206 Q. Chen et al. /Science of Sintering, 50 (2018) 205-215
___________________________________________________________________________

porous mullite ceramics by gel-casting. Obviously, the disadvantages of the above methods
including the CO2 emission, the complex process and the high cost limited their application in
the refractory industry. The in-situ pore-forming technique [3, 7, 16-19], which utilizes the
release of gases (such as H2O) caused from the decomposition of solid materials (such as
Al(OH)3) to form pores in-situ, is an environmentally friendly way to produce porous
ceramics with well-distributed pores.
As we know, porous ceramic consists of pore and solid phase. Yan [1, 7] prepared the
pore porous ceramics by the in-situ pore forming technique and found that the pore structure
was interconnected [1, 7] and the solid phase was isolated by pores. Therefore, the strength of
porous ceramics mainly depended on the strength of necks, accordant with the sintering
theory of M. Randall [20]. In our previous work [3, 7, 16], the porous cordierite-mullite
ceramics had been prepared by the in-situ pore-forming technique and the influences of pore
characteristics and phase composition on the strength of the porous ceramics were fully
discussed. Unfortunately, until now, the reaction sintering process of the porous cordierite-
mullite ceramics prepared by the in-situ pore-forming technique and the effects of the neck
characteristics on the strength has not been understood. In present work, the porous cordierite-
mullite ceramics with similar apparent porosity and different neck characteristics were
prepared by changing Al(OH)3 content and the effects of neck characteristics on the strength
of the porous ceramics were investigated.

2. Experimental procedure

Al(OH)3 (d50, 99.8μm), magnesite (d50, 41.6μm), silica (d50, 35.4μm) were used as raw
materials and clay (d50, 1.01μm) as binder in this study. The chemical compositions of the raw
materials are listed in Tab. I. Five powders named as A ~ E were mixed according to Tab. II.
Then the powders were compressed into cylindrical samples with a height of 36 mm and a
diameter of 36 mm at a pressure of 50 MPa. The green samples were dried at 110 °C for 24 h
and heated at 1410 °C for 3 h in an electric furnace and then cooled down to room
temperature.

Tab. I Chemical compositions of raw materials (wt%).

SiO2 Al2O3 Fe2O3 CaO MgO K2 O Na2O TiO2 Ignition loss
Magnesite 2.37 2.62 0.032 1.78 43.9 0.005 0.01 0.036 48.95
Al(OH)3 0.13 67.93 0.03 0.01 0.01 0.01 0.016 0.01 31.59
Silica 98.7 - - - - - - - -
clay 49.94 32.59 1.22 0.07 0.15 0.29 0.79 2.07 12.58

Tab. II Formulae of the experiment (wt%).

A B C D E
Magnesite 16.3 13.9 11.6 9.3 6.9
Al(OH)3 48.7 52.7 56.9 60.9 64.9
Silica 32.0 30.3 28.6 26.8 25.1
Clay 3 3 3 3 3

The apparent porosity was detected by Archimedes’ Principle with water as a

medium. The compressive strength of sintered samples was obtained according to Chinese
 Q. Chen et al./Science of Sintering, 50 (2018) 205-215 207
___________________________________________________________________________

standard GB/B 3997.2-1998. Phase analysis was carried out by X-ray diffractometry (XRD,
X'Pert Pro, Philips, Netherlands) with a scanning speed of 2°/minute. In addition, the relative
contents of the identified phases were calculated by the semi-quantitative analysis in
HighScore 3.0 works on the basis of the RIR (=Reference Intensity Ratio) values. The
microstructure was observed by a scanning electron microscope (SEM, JSM-6610, JEOL
Company, Japan) equipped with energy dispersive X-ray spectroscopy (EDS,
QUANTAX200-30, BRUKER Company, Germany). Pore and neck characterizations were
obtained by a microscopy measurement method based on Micrographs using image Analysis
and Process System (MIAPS) software. And the neck parameters in an area of 1.19 mm2 of
each sample were counted using ten Micrographs. The content and chemical composition of
the liquid phase at 1410 ºC were calculated from the SiO2-Al2O3-Fe2O3-CaO-MgO-K2O-
Na2O-TiO2 system using the Equilibrium Mode of the Factsage 6.2 thermochemical software.
In this calculation, FACT 53 and FToxid databases were chosen, and the liquid phase came
from FToxid-SLAG. The viscosity of liquid phase was calculated by the Viscosity Mode of
Factsage 6.2 using the glass database.

3. Results and Discussion

3.1. Phase identification

XRD patterns of samples are shown in Fig. 1 and the relative phase contents are listed
in Tab. III. It can be seen that the five samples consist of major cordierite (Ref code:
010841221) and mullite (Ref code: 969005503) and minor corundum (Ref code: 010711123)
and spinel (Ref code: 969002098). With an increase in Al(OH)3 content from 48.7 wt% to
64.9 wt%, the content of mullite increases from 30 wt% to 62 wt% and the content of
cordierite decreases from 55 wt% to 24 wt%, while the contents of corundum and spinel
change little.

Fig. 1. XRD patterns of samples.

Tab. III Relative phase contents of samples (wt%).

Samples A B C D E
Cordierite 55 52 40 32 24
Mullite 30 38 47 55 62
Corundum 9 7 9 10 11
Spinel 5 3 4 3 3
208 Q. Chen et al. /Science of Sintering, 50 (2018) 205-215
___________________________________________________________________________

3.2. Pore characterization

Fig. 2. Apparent porosities and median pore sizes of samples.

Apparent porosities and median pore sizes of the samples are shown in Fig. 2. With
an increase in Al(OH)3 content, viz. from sample A to E, the apparent porosity changes little
(the error range is ±1% according to Chinese standard GB/T 2998-2001), but the median pore
sizes of the samples are rather different. From sample A to C, the median pore size of samples
decreases slightly from 134.6μm to 120.0 μm, but from sample C to E, it decreases sharply
from 120.0μm to 54.1μm. It confirms that the Al(OH)3 content of samples gives little effect
on the apparent porosity but gives significant effect on the median pore size of the porous
ceramics.

Fig. 3. Pore size distributions of samples.

The pore size distributions of the samples are shown in Fig. 3. Unimodal pore size
distributions are observed in all samples. With an increase in Al(OH)3 content, the curves
shift towards left obviously. The pore size of sample A is in the range of 20 μm to 300 μm
while the pore size of sample E is in the range of 10 μm to 100μm. The median pore sizes
(d50) of sample A to E are 134.6 μm, 114.6 μm, 103.0 μm, 96.61 μm and 62.60 μm,
respectively.
 Q. Chen et al./Science of Sintering, 50 (2018) 205-215 209
___________________________________________________________________________

Fig. 4. Micrographs of samples sintered at 1410 ºC, a: sample A, b: sample E.

In order to investigate the change of the pore size distribution, the micrographs of
samples A and E are given in Fig. 4. It can be seen that the distribution of solid phase and
pores of sample E is more homogeneous than sample A. And the pore size of sample E is
much smaller than sample A. Correspondingly, the neck size and solid particle size of sample
E are much smaller than sample A, but the total number of necks of sample A are much larger
than sample E.

3.3. Compressive Strength

Fig. 5 shows the compressive strength of samples. From sample A to C, the

compressive strength slightly increases from 18.8 MPa to 28.2 MPa, but from sample C to E,
the compressive strength significantly increases from 28.2 MPa to 55.9 MPa. Comparing with
a porous cordierite-mullite ceramic with a strength of 30.6 MPa and an apparent porosity of
44 % [3] and a porous cordierite-mullite ceramic with a strength of 37.7 MPa and an apparent
porosity of 42 % [7] in our previous work, samples D and E have similar apparent porosity of
45 % but higher compressive strength of 50.3 MPa and 55.9 MPa.

Fig. 5. Compressive strengths of samples.

210 Q. Chen et al. /Science of Sintering, 50 (2018) 205-215
___________________________________________________________________________

3.4 Discussion

The Al(OH)3 content strongly affected the phase compositions, microstructures and
strengths of the porous ceramics through changing the microstructure evolution during
sintering process.

Fig. 6. Micrograph of sample E sintered at 900 ºC.

The early work [17] proved that Al(OH)3 and magnesite powders decomposed into
pseudomorphs with micropores at 300 ºC and 670 ºC, respectively. When the sample in
present work was heated at the temperature above 670 ºC, a skeletal structure which formed
from the Al(OH)3 pseudomorphs (AlPs) would filled with the small powders including
magnesite pseudomorphs (MaPs), silica and clay, such as the microstructure of sample E
sintered at 900 ºC (Fig. 6). When the temperature continued to rise, the reaction sintering
process taking place in the samples could be divided into four stages, as illustrated in Fig. 7.
(1) Formation of local liquid phase.

Fig.7. Schematic diagram of reaction sintering process of samples.

According to our early works [21], MgO decomposed from MgCO3 firstly reacted
with the production of the kaolinite gangue to form local liquid phase with high viscosity at
 Q. Chen et al./Science of Sintering, 50 (2018) 205-215 211
___________________________________________________________________________

1200 ºC in the Al(OH)3-kaolinite gangue-MgCO3 system; so, similarly, the magnesite

pseudomorphs also firstly reacted with the production of the clay to form local liquid phase in
present work before the sintering temperature of 1410 ºC. When the liquid phase contacted
with the AlPs and silica, both Al2O3 in the AlPs and the SiO2 in silica began to dissolve in this
liquid (Fig. 7a). The higher the magnesite content, the higher the liquid phase content, so the
liquid phase content of sample A was higher than that of sample E (Fig. 8).
(2) Spread and penetration of liquid phase.

Fig. 8. Calculated content and viscosity of liquid phase of samples at different temperatures.

Once the local liquid phases were formed, they would begin to spread into the AlPs
and silica particles and then penetrate into the micropores of the AlPs due to the capillarity
(Fig. 7b). The solubility of Al2O3 and SiO2 into the liquid phase aided the spread and
penetration. As a result, the small pores in the AlPs were filled with liquid phase, and the
number of small pores decreased greatly (Fig. 3). When the contents of Mg2+, Al3+ and SiO44-
in liquid phase reached the equilibrium concentration, cordierite precipitated out (Fig. 7b).
(3) Reaction sintering and particle rearrangement. With the penetration of liquid
phase in the AlPs, the dissolution rate of Al2O3 into liquid phase increased because of the
increasing contact area between the liquid phase and the AlPs, and then the Al3+ content in the
penetrated liquid phase increased. This would generate the chemical potential gradient of
Al3+, SiO44- and Mg2+ between the exterior and interior of the Al(OH)3 pseudomorph. As a
consequence, mullite mainly precipitated in the interior of the Al(OH)3 pseudomorph (Fig. 7c)
and the cordierite mainly precipitated in the exterior. Additionally, the penetration of liquid
phase also led to the segregation of the AlPs. When the Al(OH)3 content was low, the liquid
phase content was high and the viscosity of liquid phase was low (Fig. 8, sample A), the
fragments of the AlPs were connected into a large particle through rearrangement, and then
the micropores content in pseudomorphs decreased and the pores between the pseudomorph
and silica particles became larger, as shown in Fig. 7c; but, when the Al(OH)3 content was
high, the liquid phase content was low and the viscosity of liquid phase was high (Fig. 8,
sample E), and then the rearrangements of the fragments were inhibited, leaving a lot of pores
212 Q. Chen et al. /Science of Sintering, 50 (2018) 205-215
___________________________________________________________________________

between the fragments (Fig. 7e); this was the reason why the pore size decreased with an
increase in Al(OH)3 content.

Fig. 9. Micrographs of phase distributions of (a) sample A and (b) sample E.

(4) Neck formation and particles densification. With the development of the reaction
sintering process, the particles became dense and they were connected by necks. When the
Al(OH)3 content was low, the necks between the particles consisted of major interconnected
cordierite and minor isolated mullite, whereas, when the Al(OH)3 content was high, the necks
consisted of major interconnected mullite and minor isolated cordierite; which could be
verified by the results in Fig. 9.
The compressive strength of the porous ceramics depended on the characteristics of
the necks, including neck size distribution and neck phase composition.

Fig. 10. Neck size distributions of samples in an area of 1.19 mm2.

Tab. IV The total number of necks, average neck size, total length of necks and standard
deviation of neck size of samples.
A B C D E
Total number of necks 26 30 42 61 113
Average neck size (μm) 86.2 86.4 76.2 56.2 46.6
Total length of necks1 (μm) 2241.2 2592.0 3200.4 3428.2 5265.8
Standard deviation2 of neck
48.83 43.34 36.81 26.92 20.82
size(μm)
1
Total length of necks are equal to the total number of necks multiply by the average neck size.
2
Standard deviation refers to the fluctuation of data, the bigger the standard deviation, the greater the
fluctuation, the less uniformly distributed the data.
 Q. Chen et al./Science of Sintering, 50 (2018) 205-215 213
___________________________________________________________________________

On one hand, the neck size distributions of the samples are shown in Fig. 10, and the
total numbers, average sizes, total lengths and standard deviation of size of the necks are
listed in Tab. IV. It can be seen that with an increase in Al(OH)3 content, the total number of
necks increased but the average neck size decreased because the decrease of liquid phase
content in samples (Fig. 8) inhibited the rearrangement of particles and the growth of necks
(Fig. 7e). However, the increases of total length of necks and the uniformity of neck size were
conducive to enhancing the strength of necks.

Fig. 11. Micrographs of necks of samples, (a) sample A, (b) sample C and (c) sample E.

On the other hand, the necks mainly consisted of mullite and cordierite, and the
mullite content of necks increased with an increase in Al(OH)3 content, as illustrated in Fig. 7,
Fig. 11 and Tab. V. As reported in Ref. [22], the strengths of mullite and cordierite were 200
MPa and 120 MPa respectively, so the strength of necks increased with an increase in
Al(OH)3 content without considering the neck size characteristics.
When taking the above two hands into consideration, the increasing neck strength
increased the strength of the porous ceramics with an increase in Al(OH)3 content (Fig. 5).

Tab. V EDS results of points in Fig. 11 (at%).

Samples Points Mg Al Si O Possible phase
1 5.56 8.68 22.4 63.77 Major cordierite,
2 3.44 5.64 26.35 64.58 Minor silica
A
3 9.32 21.98 8.81 59.9 Major cordierite,
4 10.29 23.46 6.92 59.33 Minor spinel
5 5.31 9.98 21.48 63.23 Major cordierite,
6 6.13 12.55 18.78 62.53 Minor silica
C 7 0.00 28.71 9.41 61.88 Mullite
Major mullite,
8 1.48 23.79 12.52 62.21
minor cordierite
Major mullite, minor
9 3.09 16.15 17.82 62.95
cordierite
E 10 0.00 27.23 10.64 62.13 Mullite
11 1.83 22.85 13.08 62.25 Major mullite,
12 1.21 17.71 17.77 63.61 Minor cordierite

4. Conclusions

Five porous cordierite-mullite ceramics with different pore characteristics and neck
characteristics were prepared by changing Al(OH)3 content through an in-situ pore-forming
technique. The Al(OH)3 content had little effect on the apparent porosity but had a significant
214 Q. Chen et al. /Science of Sintering, 50 (2018) 205-215
___________________________________________________________________________

effect on the phase composition, pore size distribution, neck characteristics and strength.
With an increase in Al(OH)3 content, the median pore size decreased, the total length
of necks and the uniformity of neck size increased, also the mullite content of necks
increased, resulting in the increase of strength of the porous cordierite-mullite ceramics.
When the Al(OH)3 content was 64.9 wt%, the porous cordierite-mullite ceramics consisting of
62 wt% mullite and 24 wt% cordierite had the best performance of high apparent porosity of
45.1 % and high compressive strength of 55.9 MPa.

Acknowledgements

This work was supported by the Key Project of National Natural Science of Hubei
Province of China [grant number: 2015CFA122] and the Youth Dawn Plan of Science and
Technology in Wuhan City of China [grant number: 2016070204010108] for financially
supporting this work.

5. Reference

1. W. Yan, N. Li, B. Q. Han, J. Liu, G.P. Liu, Trans. Ind. Ceram. Soc., 70 (2011) 65.
2. H. M. Alves, G. Tarı̀, A. T. Fonseca, J. M. F. Ferreira, 33 (1998) 1439.
3. W. Yan, H. Luo, J. Tong, N. Li, Int. J. Mater. Res., 103 (2012) 1239.
4. J. E. Fenstermacher, F. A. Hummel, J. Am. Ceram. Soc., 44 (1961) 284.
5. X. Dong, G. Sui, Z. Yun, M. Wang, A. Guo, J. Zhang, Mater. Design, 90 (2016) 942.
6. J. Takahashi, M. Natsuisaka, S. Shimada, J. Eur. Ceram. Soc., 22 (2002) 479.
7. W. Yan, N. Li, Y. Li, J. Tong, H. Luo, J. Ceram. Process. Res., 14 (2013) 109.
8. E. Z. Hayati, O. M. Moradi, M. G. Kakroudi, Mater. Design, 45 (2013) 571.
9. J. Liu, Y. Li, Y. Li, S. Sang, S. Li, Ceram. Int., 42 (2016) 8221.
10. A. Benhammou, Y. El Hafiane, A. Abourriche, Y. Abouliatim, L. Nibou, A.
Yaacoubi, Ceram. Int., 40 (2014) 8937.
11. Z. Hou, H. Du, J. Liu, R. Hao, X. Dong, M. Liu, J. Eur. Ceram. Soc., 33 (2013) 717.
12. C. Zou, C. Zhang, B. Li, S. Wang, F. Cao, Mater. Design, 44 (2013) 114.
13. X. He, X. Zhou, B. Su, Mater. Lett., 63 (2009) 830.
14. L. Zhang, J.M.F. Ferreira, S. Olhero, L. Courtois, T. Zhang, E. Maire, et al, Acta.
Mater., 60 (2012) 4235.
15. H. Suzuki, H. Miyazaki, J. Takahashi, S. Shimada, T. Ota, M. Takahashi, Y. Hikichi,
Key Eng. Mater., 206-213 (2001) 2021.
16. W, Yan. N Li. N., T. Jong, G. Liu, J. Xu, Sci. Sinter., (2013) 165.
17. W. Yan, J. Chen, N. Li, W. Qiu, Y. Wei, B. Han, Ceram. Int., 41 (2015) 515.
18. Z. Miao, N. Li, W. Yan, Ceram. Int., 40 (2014) 15795.
19. W. Yan, X. Lin, J. Chen, N. Li, Y. Wei, B. Han, J. Alloy Compd., (2015) 287.
20. R. M. German, Sintering Theory and Practice, John & Sons, 1996.
21. W. Yan, Reach on impact factors and related applications of porous ceramics via in-
situ decomposition pore-forming technique, Wuhan Unvisity of Science and
Technology, (2008) pp. 33-67.
22. H. Schneider, R. X. Fischer, J. Schreuer, J. Am. Ceram Soc., 98 (2015) 2948.
23. Wen Yan, X. Lin, J. Chen, N. Li, Y. Wei, B. Han, Effect of TiO2 addition on
microstructure and strength of porous spinel (MgAl2O4) ceramics prepared from
magnesite and Al(OH)3, J. Alloy Compd., (2015) 287-291.
24. M. Randall, German, Sintering Theory and Practice, John & Sons, (1996) pp. 1-30.
 Q. Chen et al./Science of Sintering, 50 (2018) 205-215 215
___________________________________________________________________________

25. W. Yan, Reach on impact factors and related applications of porous ceramics via in-
situ decomposition pore-forming technique, WuHan Unvisity of Science and
Technology, (2008) pp. 33-67.
26. H. Schneider, R. X. Fischer, J. Schreuer, Mullite: Crystal Structure and Related
Properties, J. Am. Ceram Soc., 98 (2015) 2948-2967.

Садржај: Пет порозних керамика кордијерит-малита сличне порозности и

различитих карактеристика „врата“ је припремљено од Al(OH)3, магнезита, силике и
глине, користећи in-situ технику за формирање пора. Фазни састав, карактеристике
пора и врата и чврстоћа порозне керамике су испитани дифрактометријом Х-зрака,
скенирајућом електронском микроскопијом и мерењем димензија оптичком
микроскопијом. Резултати показују да садржај Al(OH)3 има значајан утицај на
дистрибуцију величине пора и карактеристике врата (дистрибуцију величине, укупну
вредност величине и фазни састав), као и на чврстоћу керамике. Повећање садржаја
Al(OH)3 доводи до смањења средњег пречника пора, док се укупна дужина врата,
униформност величине врата и удео малита у врату повећавају, што има за резултат
повећану чврстоћу порозне керамике кордијерит-малита..Најбоља својства керамике
су утврђена при садржају Al(OH)3 од 64,9 % - порозност од 45,1 % и велика
компресивна чврстоћа од 55,9 МРа.
Кључне речи: Порозна керамика кордијерит-малита; карактеристике врата;
дистрибуција величине пора; фазни састав; чврстоћа.

© 2016 Authors. Published by the International Institute for the Science of Sintering. This
article is an open access article distributed under the terms and conditions of the Creative
Commons — Attribution 4.0 International license
(https://creativecommons.org/licenses/by/4.0/).