Journal of the European Ceramic Society 35 (2015) 41154120

Contents lists available at www.sciencedirect.com

Journal of the European Ceramic Society

journal homepage: www.elsevier.com/locate/jeurceramsoc

Effects of pore shape and porosity on the dielectric constant of porous

-SiAlON ceramics
Zhaoping Hou a, , Feng Ye b , Limeng Liu b
a
b

College of Materials Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, PR China
School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, PR China

a r t i c l e

i n f o

Article history:
Received 9 January 2015
Received in revised form 1 July 2015
Accepted 2 July 2015
Available online 26 July 2015
Keywords:
-SiAlON
Dielectric constant
Porosity
Porous microstructure
Freeze casting

a b s t r a c t
Gradient porous and homogeneous porous -SiAlON (Z = 2) ceramics were successfully fabricated
through the unidirectional freezing of camphene-based suspensions at 0 C and 196 C, respectively. The
dielectric constant of the porous -SiAlON ceramics was investigated as a function of the pore shape and
porosity. The dielectric constant decreased with increasing porosity, and the gradient porous -SiAlON
ceramics with channel and dendritic pores exhibited lower dielectric constants than homogeneous samples with irregular pores. The dielectric constants of the gradient porous -SiAlON ceramics were found
to be in good agreement with the prediction of the modied cubes model. The dielectric constant of
the porous -SiAlON ceramics was associated not only with porosity and pore shape, but also with pore
anisotropy. The value of dielectric loss was not sensitive to porosity. These results have practical value
for fabricating and optimizing porous ceramics on specic applications.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Porous silicon nitride-based materials or SiAlON ceramics are
recognized as a promising candidate for a wide range of applications that involve high temperature and corrosive environments,
such as molten-metal ltrations, sensors, catalyst supports, separation membranes, light weight structural components and radome,
due to their good thermal shock resistance, high strain tolerance,
high strength and corrosion resistance [14].
Porous SiAlON ceramics are conventionally fabricated based on
a straight forward processing route through the partial sintering of
powder mixtures [5]. In this process, it is challenging to control the
porosity and pore shape. However, for porous -SiAlON ceramics,
it is important to tailor the porosity and the pore structures to meet
the specic application demands, such as good permeability, large
surface area and low dielectric constant.
Over the last few years, the freeze casting process has attracted
considerable focus in a wide variety of ceramic materials such as
Al2 O3 , Si3 N4 , SiC, and ZrO2 [6]. The popularity of this process has
grown because the microstructures are easily tunable by varying
the solvent and processing parameters, which results in complex

Corresponding author. Tel.: +86 351 6111121.

E-mail addresses: houzhaoping@tyut.edu.cn, zhaopinghou@hotmail.com
(Z. Hou).
http://dx.doi.org/10.1016/j.jeurceramsoc.2015.07.002
0955-2219/ 2015 Elsevier Ltd. All rights reserved.

pore structures and component geometries where the pores are a

replica of the solvent crystals.
Due to the unique crystal growth characteristics of different solvents, the pore microstructure is lamellar when water
is used as a solvent [7]; a dendritic structure is formed in the
case of camphene [8]; and, long straight prisms are obtained
when using TBA (tert-butyl alcohol) as a solvent [9]. Additionally,
the pore microstructure can be tailored by properly controlling
the freezing conditions to form either homogeneous or directional microstructures [10]. When the freezing conditions are
controlled to achieve a minimum temperature gradient throughout the sample and the particles offset particle expulsion from
the freezing front, the directional growth of solidied solvent
is inhibited and homogenous microstructures with interconnected pores can be fabricated. When the anisotropic growth
of a solidifying solvent is favored by the temperature gradient, a well-dened directional microstructure is also able to be
fabricated.
To date, the nucleation and crystal growth of aqueous systems
have been studied extensively. However, a clear understanding
of the kinetics and mechanisms of the camphene growth process
during freeze casting is lacking. Conversely, most papers concentrate on the mechanical properties of porous ceramics fabricated
by freeze casting [11,12]. Nevertheless, there are relatively few
results concerning about the effects of pore shape on the dielectric
properties of porous Si3 N4 -based ceramics [3].

4116

Z. Hou et al. / Journal of the European Ceramic Society 35 (2015) 41154120

In this article, porous -SiAlON ceramics with homogenous and

gradient pore structures were fabricated through the unidirectional
freezing of camphene-based suspensions at different temperatures. The objectives of the present study were to demonstrate
the difference in both porosity and pore morphology between
porous -SiAlON ceramics frozen at different temperatures, and
to evaluate the dependence of the dielectric constant on the pore
microstructure.

2. Experimental procedure
-Si3 N4 (5 wt% , d50 = 0.50 m, Junyu Ceramic Co., Ltd.,
Shanghai, China), Al2 O3 (grade A16SG, Sinopharm Chemical
Reagent Co., Ltd., Shanghai, China) and AlN (grade C, H. C. Starck,
Berlin, Germany) were used as the starting materials to fabricate the porous -SiAlON (Si6z Alz Oz N82 , Z = 2) ceramics. Freeze
casting of ceramic green bodies was performed using camphene
(C10 H16 ) without any further purication (95% purity, Guangzhou
Huanpu Chemical Factory, Guangzhou, China) as the solvent, and
HAOFAST923 (Shanghai Haoyang Co., Ltd., Shanghai, China) as
the dispersant (1 wt% based on the total weight of the ceramic
powders).
Different initial solid loading content slurries (10 vol%, 20 vol%,
30 vol% and 40 vol%) containing the -SiAlON precursors were ball
milled at 60 C for 24 h in sealed bottles. The prepared slurries
exhibited stable and excellent owability for casting.
Unidirectional solidication was performed by pouring
the resultant warm slurries into polyethylene molds (60 mm
20 mm) placed on a copper plate which maintained at 0 C
using an icewater mixture or at 196 C using liquid nitrogen.
To eliminate the inuence of the temperature gradient in the
radial direction, the polyethylene molds were packed by a heat
barrier layer on the surrounding side and were pre-warmed
at 60 C prior to placing them on the cold copper plate. Therefore, the slurry froze from the cold bottom with camphene
crystals.
Thereafter, the green bodies were placed on a polyurethane
sponge in open air at room temperature for approximately 4 days
to eliminate the solidied camphene. Sintering was performed in a
conventional graphite resistance furnace at 1800 C for 2 h under a
nitrogen gas pressure of 0.4 MPa. A BNSi3 N4 powder bed (50 wt%
BN, 50 wt% Si3 N4 ) was used. The heating and cooling rates were
both 10 C/min.
The densities of the sintered samples were calculated using
mass and volume measurements. Pore size distribution was measured using mercury porosimetry (Autopore9500, Micromeritics
Co., UAS). The crystalline phases of the sintered porous ceramics
were characterized by X-ray diffraction (XRD). The microstructures
of the sintered samples were observed using scanning electron
micro scope (SEM). The pore size in different regions of the specimens was also determined by measuring the average size of the
pores from the SEM micrographs.
The dielectric properties of the obtained porous -SiAlON
ceramics were measured using the wave guide method with a vector network analyzer (Agilent E8362B) over 12.418 GHz. The sizes
of the measured samples were 7.9 mm 15.8 mm 2.0 mm. The
measuring direction was parallel to the solidication direction of
the samples. The test specimens were taken from the central zone of
the sintered samples. Additionally, in order to reduce measurement
errors to minimum, the measured samples are nely processed in
our experiment to reduce the gap between sample and waveguide
wall. In this study, on account of the big discrepancy among samples with different porosity and pore shape, the systemic errors are
not strong enough to impact the analysis of inuence of porosity
and pore shape on dielectric constants.

Fig. 1. XRD patterns of the sintered porous -SiAlON (Z = 2) ceramics with different
initial solid loadings frozen at 196 C.

3. Results and discussion

3.1. Phase formation
Fig. 1 shows the XRD patterns of the sintered porous -SiAlON
(Z = 2) ceramics with different initial solid loading that were frozen
at 196 C (liquid N2 ). Only -SiAlON phase was detected in the
samples with initial solid loadings of 20 vol%, 30 vol%, and 40 vol%,
which indicates that the raw ceramic Si3 N4 , Al2 O3 , and AlN powders
reacted completely to form the target phase, -Si4 Al2 O2 N6 .
When the samples were prepared by freezing at 0 C (icewater
mixture), all of the samples showed very similar patterns (Ref. [13]
(Fig. 1)) [13], which means that all of the samples consisted of the
-SiAlON phase and no other secondary phases, regardless of the
freeze temperature, predicating the freezing temperature did not
inuence the reaction of the raw ceramic powders and the formation of the -SiAlON (Z = 2) phase.
3.2. Porosity and porous microstructure
The porosities of the sintered porous -SiAlON (Z = 2) ceramics frozen at 0 C or 196 C with different initial solid loadings
are listed in Tables 1 and 2. The porosity of the sintered porous
-SiAlON ceramic decreased proportionally with the increase in
the initial solid loading. Compared with the samples frozen at
196 C, the resultant porosity of samples frozen at 0 C was lower
given the same initial solid loading. Additionally, the porosities
Table 1
Porosity of porous -SiAlON (Z = 2) ceramic frozen at 196 C.
The solid loading content (vol%)

Porositya (%)

Porosityb (%)

20
30
40

52.9
38.1
27.5

45.4
32.7
23.6

a
b

measured by Archimedes s method.

measured by mercury intrusion porosimetry.

Table 2
Porosity of porous -SiAlON (Z = 2) ceramic frozen at 0 C.
The solid loading content (vol%)

Porositya (%)

Porosityb (%)

10
20
30

71.3
47.0
36.1

63.5
44.8
32.1

a
b

Measured by Archimedes s method.

Measured by mercury intrusion porosimetry.

Z. Hou et al. / Journal of the European Ceramic Society 35 (2015) 41154120

4117

Fig. 2. SEM images of a cross-section parallel to the freezing direction of the sintered porous -SiAlON ceramic with initial solid loading of 10 vol% frozen at 0 C.

measured by mercury porosimetry are almost the same as the

porosities measured using Archimedes method, which implies that
the pore channels were almost completely interconnected.
Fig. 2 shows the SEM images of a cross-section parallel to the
freezing direction of the sintered porous -SiAlON ceramic with
initial solid loading of 10 vol% frozen at 0 C. The SEM observation
reveals that the sample has a gradient porous structure with a long
range (over 5 mm). The pore morphology and pore size varied as a
function of the distance from the bottom of the directionally solidied sample, which is where the solidication initiated. In Zone 1
(Fig. 2a), close to the initial cold nger, the material is characterized
by aligned elongated pores parallel to the direction of solidication.
In Zone 2, the pore morphologies have a columnar pore structure, which was formed by the growth of camphene dendrites. It is

obvious that the channel size and the secondary arm dendrite
spacing (SDAS) increase with an increase in the distance from the
bottom of the sample.
The underlying reason for this architecture was described in our
previous report [13]. In unidirectional freeze casting, the temperature gradient and the heat transfer are critical factors that inuence
the pore microstructure. Camphene kinetically favors the formation of directional microstructures in the 0 0 1 or c-direction under
a steep thermal gradient, which leads to the growth of primary
dendrites down the temperature gradient. Consequently, aligned
elongated pores were formed in Zone 1. With a decrease in the temperature gradient, the primary crystals grow with side-branching
or secondary dendritic growth, which leads to the accumulation
of particles between the secondary dendrites emanating from the

4118

Z. Hou et al. / Journal of the European Ceramic Society 35 (2015) 41154120

Fig. 3. SEM micrographs of the sintered porous -SiAlON ceramics (in Zone 2) with different initial solid loadings frozen at 0 C: (a) 10 vol%; (b) 20 vol%; (c) 30 vol%.

Fig. 4. SEM micrographs of the sintered porous -SiAlON ceramics with different initial solid loadings frozen at 196 C: (a) 20 vol%; (b) 30 vol%; (c) 40 vol%.

primary dendrites. Therefore, a rough surface of the oriented pore

is observed in the aligned-to-dendritic transition zone (Fig. 2b), and
dendritic pores were formed in Zone 2 (Fig. 2c). Additionally, with
an increase in the distance from the bottom, the solidication velocity decreased, which led to an increase in the pore size and the
secondary dendrite arm spacing (SDAS).
Fig. 3 shows the microstructure of the sintered porous -SiAlON
ceramics (in Zone 2) with different initial solid loadings that were
frozen at 0 C. Compared with the samples with 10 vol% solid
loading, similar morphology was observed in Zone 2 of the sample fabricated with 20 vol% solid loading. The channel size of the
primary and dendritic pores decreased when the solid loading
increased from 10 vol% to 20 vol%. When the solid loading further
increased to 30 vol%, the samples did not form dendrites and channel pores but instead had a round shape.
SEM micrographs of the sintered porous -SiAlON ceramics
with different initial solid loadings that were frozen at 196 C are
shown in Fig. 4. In each case, the pores were distributed uniformly
throughout the entire sample. The -SiAlON grains were connected
to one another to form a space skeleton with an irregular porous
structure. With an increase in the initial solid loading from 20 vol%
to 40 vol%, the porosity and pore size decreased.
The pore size distribution was also analyzed by mercury
porosimetry (Fig. 5). By increasing the solid loading from 20 vol%
to 40 vol%, the pore size range decreased from 28 m to 13 m.
Compared to the samples prepared by freezing at 196 C, the pore
size ranges of the samples frozen at 0 C with an initial solid loading of 20 vol% and 30 vol% were 615 m and 35 m, respectively
(Ref. [13] (Fig. 4)) [13], which are larger than the pore size ranges
in samples frozen at 196 C.
In unidirectional freeze casting, the freezing temperature, freezing time, and freezing rate are also signicant factors affecting the
crystal nucleation and growth and the particle arrangement and,
subsequently, the morphology of the porous microstructure. At a
relatively lower freezing temperature (196 C), i.e., high supercooling, the nucleation rate is greater than the crystal growth rate,

Fig. 5. Pore size distribution of the obtained porous -SiAlON ceramics frozen at
196 C.

so camphene nucleation is more favorable than camphene crystal

growth. Consequently, a large number of small camphene crystals
formed, and the homogenous porous microstructure consists of
uniformly small pores.
Overall, different freezing temperatures lead to different camphene solidication behaviors and, ultimately, different pore
morphologies. In the present work, unidirectional freezing on a cold
substrate at 0 C or 196 C was used to fabricate gradient porous
and homogeneous porous -SiAlON ceramics.
3.3. Dielectric properties
The dielectric constants of -SiAlON ceramic with different initial solid loadings are measured at 12.416 GHz, and the results are
shown in Fig. 6. All of the materials exhibited reasonably stable
dielectric constant values. The dielectric constant of these porous

Z. Hou et al. / Journal of the European Ceramic Society 35 (2015) 41154120

4119

determined to be = 1 and = 1 for the serial and parallel mixing model, respectively. Nevertheless, the equation is not realistic,
and as a result, these models do not t the data well. As shown in
Fig. 7, the measured dielectric constant data points of the porous
-SiAlON ceramics fall between the predictions of the serial and
parallel models. For the two data points from the homogeneous
porous -SiAlON ceramics at <40% porosity, in which pore size is
relatively smaller, the dielectric constants measured are closer to
the prediction of the parallel model. While the pore size increased
at a higher porosity regardless of the shape of pore, the measured
dielectric constants are found to change toward the prediction of
the serial model.
For porous ceramics, the shape of the pores is also a signicant factor affecting the dielectric properties. Considering the pore
shape, an efcient modied cube model has been proposed to predict the relative dielectric constant of porous ceramics [16]:

r = 0r

Fig. 6. The dielectric constants of porous -SiAlON ceramics with different initial
solid loadings.

materials varied between 2.3 and 5.6, which is less than that of the
dense -SiAlON (Z 2) ceramic, which ranged from 7.2 to 7.67 [14].
The dielectric constant ( ) of porous -SiAlON ceramics measured at 15 GHz as a function of the porosity is shown in Fig. 7, and
the theoretically predicted data are shown with lines. The dielectric
constant value decreased with an increase porosity over the range
of 27 to 71% due to the volume fraction effect, which is consistent
with previously reported results [4]. However, in the whole range of
porosities, the homogeneous porous -SiAlON ceramics frozen at
196 C exhibited greater dielectric constants than gradient porous
ceramics frozen at 0 C.
Porous ceramics can be regarded as a gassolid composites consisting of an active ceramic phase and a passive phase, which is air.
The dielectric properties of the composite could be predicted using
theoretical models. The following well-known general empirical
equation can be used to predict the dielectric constants of porous
ceramics [15]:

r =

Vi
ri

where r is the relative dielectric constant of the porous material,

ri and Vi are the dielectric constant and volume fraction of the
ith materials, respectively, and is a constant. The value of is

Fig. 7. The dielectric constant ( ) of porous -SiAlON ceramics measured at 15 GHz

as a function of the porosity.

1+

 P 2/3

1
2/3

P 1/3 (0r 1)KS

+1

KS

 P 2/3

KS

where r and 0r are the relative dielectric constants of porous

ceramics and the solid matrix, respectively. P is the porosity and Ks
is a parameter determined by the pore shape. For a spherical pore,
Ks = 1; for an oval pore, Ks = 0.5. The dielectric constant of gradient
porous -SiAlON ceramics are more similar to the prediction made
by the modied cubes model with Ks = 1, which is consistent with
the SEM observation of channel pores for the 10 vol% and 20 vol%
initial solid loading samples, with round pores for the 30 vol%, as
shown in Fig. 3.
As shown in Fig. 7, the dielectric constant of gradient porous
ceramic with 20 vol% initial solid loading was less than the predicted value. Such a discrepancy can also be rationalized by
considering the pore microstructure of the samples. Note that the
measurement direction was parallel to the solidication direction
for the samples, i.e., parallel to the direction of primary dendritic as
shown in Fig. 3a and b. Additionally, when the initial solid loading
increased from 10 vol% to 20 vol%, the angle between the secondary
dendritic and solidication direction increased; namely, the secondary dendritic tilted perpendicular to the solidication direction.
On the basis of the anisotropy of secondary dendritic pore in the
10 vol% and 20 vol% samples, it is reasonable to assume that the
alignment of the pores in the measurement direction results in
an increase in the dielectric constant of the porous ceramics. The
inuence of the pore anisotropy on the properties of the porous
piezoelectric composites has been researched by Bowen [17], and

Fig. 8. The dielectric loss (tan ) of porous -SiAlON ceramics with different initial
solid loadings.

4120

Z. Hou et al. / Journal of the European Ceramic Society 35 (2015) 41154120

these researchers found that the alignment pores in the direction

of poling results in an increasing permittivity, , due to an increase
in the volume fraction of the active material connected in the 3direction.
Fig. 8 shows the dielectric loss (tan ) of -SiAlON ceramic with
different initial solid loadings measured at 12.416 GHz. In the case
of samples with homogeneous pore structure, the value of tan is
not sensitive to porosity and pore size. The overall appearance of the
curves was similar. And, the homogeneous porous ceramics have
lower dielectric loss than the gradient porous ceramics over the
range of 12.416 GHz. It is unclear why the homogeneous porous
ceramics exhibited lower dielectric loss than the gradient porous
ceramics, but the possible reasons may be related not only to the
pore shape, but also to the grain morphology and the secondary
phases.
4. Conclusions
In this paper, porous -SiAlON ceramics were fabricated using
a method based on the unidirectional freezing of camphene-based
suspensions on a cold substrate (196 C or 0 C). When the freezing
temperature was relatively high (0 C), the gradient pore structure, including aligned channels and dendritic pores, was formed
due to the thermal gradient throughout the samples. However,
at a low temperature (196 C), homogeneous porous -SiAlON
ceramics with irregular pores were prepared. The dielectric constant decreased with an increase in the porosity in the range of
27 to 71% for porous -SiAlON ceramics. Except for the porosity,
the shape of pores in the porous ceramics is also a signicant factor affecting the dielectric constant. Compared with homogeneous
porous -SiAlON ceramics with irregular pores, the gradient porous
ceramics with channel pores exhibited lower dielectric constant.
The variation in the dielectric constant of gradient porous -SiAlON
ceramics with increasing porosity was well predicted using a modied cubes model. Furthermore, the homogeneous porous ceramics
exhibited lower dielectric loss than the gradient porous ceramics
over the range of 12.416 GHz.

References
[1] J.M. Wu, X.Y. Zhang, J.L. Yang, Novel porous Si3 N4 ceramics prepared by aqueous
gelcasting using Si3 N4 poly-hollow microspheres as pore-forming agent, J. Eur.
Ceram. Soc. 34 (5) (2014) 10891096.
[2] G.J. Zhang, J.F. Yang, T. Ohji, Fabrication of porous ceramics with unidirectionally aligned continuous pores, J. Am. Ceram. Soc. 84 (6) (2001) 13951397.
[3] B. Li, K. Liu, C.R. Zhang, et al., Fabrication and properties of borazine derived
boron nitride bonded porous silicon aluminum oxynitride wave-transparent
composite, J. Eur. Ceram. Soc. 34 (15) (2014) 35913595.
[4] Z.P. Hou, F. Ye, L.M. Liu, et al., Effects of solid content on the phase assemblages,
mechanical and dielectric properties of porous -SiAlON ceramics fabricated
by freeze casting, Ceram. Int. 39 (2) (2013) 10751079.
[5] J.F. Yang, Y. Beppu, G.J. Zhang, T. Ohji, S. Kanzaki, Synthesis and properties
of porous single-phase  -SiAlON ceramics, J. Am. Ceram. Soc. 85 (7) (2002)
18791881.
[6] S. Deville, Freeze-casting of porous ceramics: a review of current achievements
and issues, Adv. Eng. Mater. 10 (3) (2008) 155169.
[7] S. Deville, E. Saiz, A.P. Tomsia, Ice-templated porous alumina structures, Acta
Mater. 55 (6) (2007) 19651974.
[8] K. Araki, J.W. Halloran, New freeze-casting technique for ceramics with sublimable vehicles, J. Am. Ceram. Soc. 87 (10) (2004) 18591863.
[9] R. Chen, C.A. Wang, Y. Huang, et al., Ceramics with special porous structures
fabricated by freeze-gelcasting: using tert-butyl alcohol as a template, J. Am.
Ceram. Soc. 90 (11) (2007) 34783484.
[10] W.L. Li, K. Lu, J.Y. Walz, Freeze casting of porous materials: review of
critical factors in microstructure evolution, Int. Mater. Rev. 57 (1) (2012)
3760.
[11] X. Liu, M.N. Rahaman, Q. Fu, Oriented bioactive glass (1393) scaffolds with
controllable pore size by unidirectional freezing of camphene-based suspensions: microstructure and mechanical response, Acta Biomater. 7 (1) (2011)
406416.
[12] B.H. Yoon, W.Y. Choi, H.E. Kim, et al., Aligned porous alumina ceramics with high
compressive strengths for bone tissue engineering, Scr. Mater. 58 (7) (2008)
537540.
[13] Z.P. Hou, F. Ye, L.M. Liu, Q. Liu, Fabrication of gradient porous -SiAlON ceramics
via a camphene-based freeze casting process, Mater. Sci. Eng. A 558 (2012)
742746.
[14] I. Ganesh, Development of -SiAlON based ceramics for radome applications,
Process. Appl. Ceram. 5 (3) (2011) 113138.
[15] K. Wakino, T. Okada, N. Yoshida, et al., A new equation for predicting
the dielectric constant of a mixture, J. Am. Ceram. Soc. 76 (10) (1993)
25882594.
[16] H. Banno, S. Saito, Piezoelectric and dielectric properties of composites of
synthetic rubber and PbTiO3 or PZT, Jpn. J. Appl. Phys. 22 (S2) (1983)
67.
[17] C.R. Bowen, H. Kara, Pore anisotropy in 33 piezoelectric composites, Mater.
Chem. Phys. 75 (1) (2002) 4549.