G Model

JECS-10919; No. of Pages 8 ARTICLE IN PRESS

Journal of the European Ceramic Society xxx (2016) xxxxxx

Contents lists available at www.sciencedirect.com

Journal of the European Ceramic Society

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

Processing and properties of glass-bonded silicon carbide membrane

supports
Su Chang Kim a , Young-Wook Kim a, , In-Hyuck Song b
a
Functional Ceramics Laboratory, Department of Materials Science and Engineering, The University of Seoul, Seoul 02504, Republic of Korea
b
Powder and Ceramics Division, Korea Institute of Materials Science, 797 Changwondaero, Changwon, Gyeongnam 51508, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: Porous SiC membrane supports were fabricated from SiC and glass frit at a temperature as low as 850 C
Received 1 August 2016 in air by a simple pressing and heat-treatment process. The effects of the initial SiC particle size and frit
Received in revised form 4 November 2016 content on the porosity, exural strength, and air permeation of the membrane supports were investi-
Accepted 8 November 2016
gated. During heat-treatment, the glass frit transformed to a viscous glass phase, which acted as a bonding
Available online xxx
material between SiC particles and as a protecting layer for severe oxidation of SiC particles. The porosity
of the porous SiC membrane supports could be controlled within a range of 3746% with the present
Keywords:
set of processing conditions. The typical exural strength, permeability, and specic air ow rate of the
Silicon carbide
Membranes
porous membrane supports fabricated using 23 m SiC particles with 15 wt% glass frit were 75 MPa,
Porosity 4.2 1013 m2 , and 32.4 L/min/cm2 , respectively.
Flexural strength 2016 Elsevier Ltd. All rights reserved.
Specic air ow rate

1. Introduction (mostly from LiqTech International A/S (Ballerup, Denmark)) [8,16],

Water is a key element in our daily lives, and membrane l-
tration for water treatment is one of the most promising methods
for cleaning and recycling water. Among the membranes commer-
cially available and/or under development, SiC membranes have
attracted attention due to their set of unique properties: low ther-
mal expansion coefcient, high thermal conductivity, excellent
thermochemical and thermomechanical stability, and biocompat-
ibility [110]. The structure of the SiC microlters consist of
membrane supports (porous media) and a surface layer (so-called
membranes). The quality of SiC membranes, which is hampered by
the presence of pinholes and microcracks, is strongly dependent on
the porosity and microstructure (pore size, pore shape, pore size
distribution of SiC grains) of the membrane supports [11]. Thus,
preparation of membrane supports with tailored microstructural
characteristics is a key factor for producing high quality lters.
Most of the previous research on SiC membranes focused on (1)
the processing of a disk or tube-type SiC support by a partial sinter-
ing method and subsequent coating of SiC membranes on the SiC
supports with polymeric precursors (mostly allylhydrido polycar-
bosilane) [13,1215], (2) processing of ultraltration membranes
on commercial SiC microltration or macroporous supports
Corresponding author.
E-mail address: ywkim@uos.ac.kr (Y.-W. Kim).
and (3) processing of SiC hollow ber membranes and characteri-
zation of their mechanical properties and permeability [10].
Various processing methods, including partial sintering
[1,5,10,17], recrystallization [18], and bonding techniques [1927],
have been reported for processing SiC membrane supports.
Partial sintering and recrystallization techniques require high
temperatures 1500 C for processing if SiC powder is used as a
starting material because of its strong covalent bonding [1,5,17,18].
Possible routes to reduce the processing temperature are to use
preceramic polymers as a precursor for SiC [5,6,28] and the use
of bonding materials [1927]. Mullite (3Al2 O3 2SiO2 ) [19,21,29],
silica (SiO2 ) [22,23], silicon oxycarbide (SiOC) [24,25], sodium
borate [26], cordierate [27], and glass frit [30,31] were investigated
as bonding materials for porous SiC ceramics. However, open-cell
SiC supports prepared from preceramic polymers showed poor
mechanical strength compared to powder-processed SiC mem-
brane supports [28,32]. In contrast, porous SiC supports prepared
with bonding materials showed relatively better mechanical
properties [22,24,25]. Among the above bonding methods, the
glass-bonding technique has not been investigated extensively
although it provides several benets for processing SiC membrane
supports, such as ease of porosity and pore size control, and
cost-effectiveness. Only a few previous works were found in the
literature and are summarized as follows:

http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
0955-2219/ 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
G Model
JECS-10919; No. of Pages 8 ARTICLE IN PRESS
2 S.C. Kim et al. / Journal of the European Ceramic Society xxx (2016) xxxxxx
(1) Chae et al. [30] prepared glass-bonded porous SiC using SiC
powders, a glass frit, and glass microbeads as raw materials and
characterized the exural strength of the porous ceramics. The
typical exural strengths of glass-bonded porous SiC ceramics
were 1923 MPa at 4649% porosity.
(2) Wang et al. [31] prepared glass-bonded porous SiC for vacuum
chuck applications. By adjusting the particle size of SiC powders
and glass frit content, the porosity of the porous SiC ceramics
could be controlled within the range of 3265%. The air perme-
ability and adsorbability were investigated for the glass-bonded
porous SiC ceramics.
However, mechanical properties and chemical durability of
glass-bonded porous SiC membrane supports have been rarely
investigated. This paper investigated the effects of SiC particle size
and glass frit content on porosity, exural strength, and air perme-
ation of glass-bonded porous SiC membrane supports. The chemical
durability of glass-bonded porous SiC ceramics in acid and alkali
solutions was also investigated.
2. Experimental procedure
Commercially-available SiC powders with particle sizes of
10 m, 23 m, 35 m (Sinxing Advanced Materials, Hong Kong),
and 65 m (Showa Denko K.K., Tokyo, Japan), and a glass frit bond-
ing material (Korea Ceramics Co. Ltd., Gwangju-shi, Gyeonggi-do,
Korea) were used as the starting powders. The chemical compo-
sition of the glass frit material was as follows: 73.2% SiO2 , 10.6%
B2 O3 , 6.4% Al2 O3 , 6.3% Na2 O and 3.5% other materials. The weight
fraction of the glass frit was 10, 15, 20, and 25 wt%. To prepare a
batch consisting of SiC and glass frit, the SiC powder, glass frit, and
organic binders (polyethylene glycol and polyvinyl alcohol, 5 wt%
with respect to ceramic powders) were mixed by ball milling using
SiC media in distilled water in a polypropylene jar for 24 h. The mix-
ture was dried, sieved, and uniaxially pressed at an applied pressure
of 40 MPa. The samples were sintered at 850 C for 2 h with a heat-
ing rate of 3 C/min in air without an applied pressure. The number
after SC denotes the particle sizes of the SiC powders, e.g., SC10 and
SC65 represent the SiC membrane supports prepared from 10 m
and 65 m SiC powders, respectively.
The bulk densities of the SiC membrane supports were cal-
culated from the weight-to-volume ratios of the samples. The
porosities of the SiC membrane supports were measured using
Archimedes method. The crystalline phases in the sintered spec-
imens were characterized by X-ray diffraction (XRD; D8 Discover,
Bruker AXS GmbH, Karlsruhe, Germany) of the crushed powders
using Cu K radiation.
The fracture surface morphology was observed using scanning
electron microscopy (SEM, S4300, Hitachi Ltd., Hitachi, Japan). For
the exural strength measurements, bar-shaped samples were cut
to a size of 3 4 35 mm3 and bending tests were performed at a
constant crosshead speed of 0.5 mm/min using a three-point bend-
ing xture with a span of 30 mm.
The open porosity and pore size distribution of SiC membrane
supports were measured using mercury porosimetry (AutoPore
IV 9500, Micromeritics, Norcross, GA, USA). The air ux of some
selected specimens was measured by capillary ow porosimetry
(CFP-1200-AEL, Porous Materials Inc., Ithaca, NY, USA). The speci-
mens (25 mm in diameter and 3.5 mm thick) were tted between
O-rings in the bottom of the chamber and the bottom of the cham-
ber was inserted for capillary ow porosimetry. The ux was then
measured automatically by sensors while incrementally changing
the pressure of the regulator. The gas permeability () was com-
Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
Fig. 1. XRD patterns of the glass frit and SiC membrane supports sintered at 850 C
for 2 h in air: SC10, SC23, SC35, and SC65 denote membrane supports prepared from
SiC powders with particle sizes of 10 m, 23 m, 35 m, and 65 m, respectively,
containing 15 wt% glass frit.
puted from the measured ow rate and pressure difference using
Darcys law,
p/L = Q /A (1)
where p is the pressure drop from the entrance to the exit of the
sample, L is the thickness of the sample, Q is the ow rate of air
through the sample, is the viscosity of air, A is the cross-sectional
area of the sample, and is the permeability.
The chemical stability of the glass-bonded SiC membrane sup-
ports (SC10) with 15 wt% glass frit was evaluated by placing
samples in glass bottles lled with a pH solution of 3 or 11, respec-
tively. The pH 3 and 11 solutions were prepared using nitric acid
and ammonia solution with distilled water, respectively. The bot-
tles were stored at room temperature for 63 days. The strength
of the samples subjected to acid or base solution was compared
with an identical unexposed sample (SC10). The glass frit compo-
sition in the neck areas after 63-days exposure in an acid or base
solution was analyzed using energy dispersive spectroscopy (EDS,
JSM-6010PLUS/LA, JEOL, Tokyo, Japan) and compared to that of the
identical, unexposed sample.
3. Results and discussion
XRD analysis was performed to identify the phases present in
the SiC membrane supports sintered with 1025 wt% glass frits at
850 C for 2 h in air. XRD patterns of the SiC membrane supports sin-
tered with 15 wt% glass frit are shown in (Fig. 1). Those of the other
supports with 10, 20, and 25 wt% glass frit were almost the same for
corresponding specimens in Fig. 1. The glass frit was an amorphous
material with no crystallization product after heat-treatment at
850 C for 2 h and furnace cooling. All membrane supports consisted
of -SiC (6H) and a trace of SiO2 (-cristobalite), which is an oxi-
dation product of SiC powders from sintering in air. The intensities
of the -cristobalite peaks increased with decreasing SiC particle
size because of the accelerated oxidation of ner powders at the
sintering temperature.
Fig. 2 shows the typical microstructures of the SC10 spec-
imens with 1025 wt% glass frit. The microstructures show
well-developed pore structures in all specimens. Most of the pores
originated from the inter-particle spaces in the green compact.
Comparison of Fig. 2(a) and (d) clearly shows that the addition
of more glass frit produced thick bonding areas and the lling of
smaller pores. The glass-bonding phase in the present specimen is
G Model
JECS-10919; No. of Pages 8 ARTICLE IN PRESS
S.C. Kim et al. / Journal of the European Ceramic Society xxx (2016) xxxxxx 3

Fig. 2. Typical microstructures of SiC membrane supports prepared from 10 m SiC powders with variation of glass frit content: (a) 10 wt%, (b) 15 wt%, (c) 20 wt%, and (d)
25 wt%.

Fig. 3. Pore size distributions of SC10 specimens with various glass frit content: (a) 10 wt%, (b) 15 wt%, (c) 20 wt%, and (d) 25 wt%.

formed by the following procedure: (1) the glass frit transforms
into a viscous glass phase during sintering at 850 C; (2) oxida-
tion of SiC during sintering in air leads to the formation of the
cristobalite phase (as conrmed by XRD (Fig. 1)); (3) the oxidation-
derived SiO2 partially dissolves into the viscous glass phase; (4)
the glass phase results in bonding between SiC particles and par-
tially lls smaller pores in the membrane supports. The pore size
of the SiC membrane supports increased with increasing glass frit
content (Fig. 3). The SC10 with 10 wt% and 15 wt% samples showed
unimodal pore size distributions (Fig. 3(a) and (b)). In contrast, the
SC10 with 20 wt% and 25 wt% samples showed bimodal pore size
distributions (Fig. 3(c) and (d)). The average pore size of the SC10
supports with 10, 15, 20, and 25 wt% was 2.5, 3.0, 3.6, and 4.2 m,
respectively. Thus, the addition of glass frit in excess of 15 wt% led
to partial lling of the pores (Fig. 2(c) and (d)) and resulted in the
bimodal pore size distribution (Fig. 3(c) and (d)), and an increase in
average pore size.
Fig. 4 shows the effect of SiC particle size on the microstructure
of the SiC membrane supports with 15 wt% glass frit. As shown
in Fig. 4, the increase in SiC particle size increased the pore size

Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
G Model
JECS-10919; No. of Pages 8 ARTICLE IN PRESS
4 S.C. Kim et al. / Journal of the European Ceramic Society xxx (2016) xxxxxx

Fig. 4. Typical microstructures of SiC membrane supports prepared from different SiC powders with 15 wt% glass frit: (a) 10 m, (b) 23 m, (c) 35 m, and (d) 65 m.

Fig. 5. Porosity of SiC membrane supports as a function of glass frit content. The SiC Fig. 6. Flexural strength of SiC membrane supports as a function of glass frit content.
membrane supports were sintered at 850 C for 2 h in air. The SiC membrane supports were sintered at 850 C for 2 h in air.

because of the particle size dependence of inter-particle space in

the glass-bonded SiC supports. The average pore sizes of SC10,
SC23, SC35, and SC65 with 15 wt% glass frit were 3.0, 6.3, 6.9, and
8.2 m, respectively. However, the pore morphology was insensi-
tive to the particle size of SiC when the glass frit content was xed at
15 wt%. The results suggest that (1) the pore size can be controlled
by adjusting the initial particle size of the SiC powders and (2) thick-
ness of bonding areas can be adjusted by controlling the glass frit
content in the present glass-bonded SiC membrane supports.
Fig. 5 shows the porosity of the SiC membrane supports with dif-
ferent SiC particle sizes and sintered with various glass frit content.
The porosity of the SiC membrane supports ranged from 42.2% to
46.3% when 10 m SiC powders were used, and from 37.0% to 38.7%
when 65 m SiC powders were used. The SC10 with 10 wt% glass
frit specimen showed the highest porosity of 46.3%, whereas SC65
with 25 wt% glass frit showed the lowest porosity of 37% among the
compositions examined. The difference in porosity was attributed
to the difference in green density of the pressed bars and the degree
of partial densication during sintering. Generally, the porosity
decreased with increasing glass frit content when the same SiC par-
ticles were used because of enhanced densication by viscous ow
of a glass bonding phase. In contrast, the porosity increased with
decreasing SiC particle size when the same amount of glass frit was
added, due to the lower packing density of the specimen prepared
with ner SiC particles.
Fig. 6 shows the exural strength of the SiC membrane sup-
ports as a function of glass frit content. The exural strength of all
specimens increased with increasing frit content from 10 to 15%,
followed by a decrease in strength when more glass frit was added
for SC23, SC35, and SC65. In contrast, SC10 maintained its strength
up to 25 wt% glass frit addition. In order to understand the rela-
tionship between the exural strength and porosity, the exural
strength of all specimens were plotted as a function of the porosity
(Fig. 7). Some important ndings from the plot can be summarized
as follows:

Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
G Model
JECS-10919; No. of Pages 8 ARTICLE IN PRESS
S.C. Kim et al. / Journal of the European Ceramic Society xxx (2016) xxxxxx
Fig. 7. Flexural strength of SiC membrane supports sintered at 850 C for 2 h in air
as a function of porosity.
(1) The exural strength generally increased with increasing
porosity, except the SiC membrane supports with 10 wt% glass
frit (data in a circle). This trend is very unusual in porous SiC
ceramics because the exural strength of porous SiC ceramics
generally decreases with increasing porosity [21,22,25,33].
(2) The exural strength increased in the order of
SC10 > SC23 > SC35 > SC65, indicating that the exural strength
of glass-bonded SiC ceramics is primarily dependent on pore
size rather than the porosity in the porosity range of 3746%.
(3) The deviation from the trend, i.e., low strength of SiC mem-
branes with 10 wt% glass frit (data in a circle), was due to
incomplete bonding of SiC grains caused by insufcient glass
frit content.
According to models proposed by Gibson and Ashby [34] and
by Rice [35], the exural strength of porous SiC ceramics decreases
with increasing porosity. Both models do not expect any depen-
dence of the strength on pore size, but a variation only with the
porosity of the porous ceramics. However, our results (Fig. 7)
showed that the strength of glass-bonded SiC supports did not fol-
low the above models. The average pore sizes of the supports with
15 wt% glass frit were 3.0 m for SC10, 6.3 m for SC23, 7.9 m for
SC35, and 12.1 m for SC65, respectively. The exural strength and
porosity of the SC10, SC23, SC35, and SC65 supports with 15 wt%
glass frit were 81 MPa and 45%, 75 MPa and 42%, 71 MPa and 41%,
and 65 MPa and 38%, respectively. The pore size of glass-bonded
SiC supports increased with increasing the SiC particle size when
the same amount of glass frit was added. In contrast, the porosity
decreased with increasing the SiC particle size.
Our results showed that the strength of glass-bonded SiC sup-
ports increased with decreasing the SiC particle size, i.e., pore size
at the same amount of glass frit. These results can be explained
in terms of (i) an increase in strut strength with decreasing the
pore-to-pore distance when smaller particles were used for a given
amount of frit and (ii) a decrease in the critical aw size when
smaller particles were used [36]. Thus, the exural strength of
glass-bonded SiC supports was inuenced by pore size as well as
porosity. However, the exural strength of glass-bonded SiC ceram-
ics is primarily dependent on pore size rather than the porosity in
the porosity range of 3746%. Similar trends were also observed in
porous liquid phase sintered SiC ceramics [36].
There are several reports showing the mechanical properties of
SiC membrane supports. The exural strengths of mullite-bonded
SiC ceramics were 314 MPa at 4654% porosity [21], 24 MPa at
43% porosity [19], and 37 MPa at 48% porosity [29]. The ex-
ural strengths of silica-bonded SiC ceramics were 630 MPa at
Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
5
Fig. 8. Air ux of SiC membrane supports prepared from different SiC powders with
15 wt% glass frit as a function of the average pressure across the membrane supports.
3655% porosity [23] and 65 MPa at 40% porosity [22]. The ex-
ural strength of sodium borate-bonded SiC ceramics was 37 MPa
at 52% porosity [26]. The exural strength of cordierite-bonded
porous SiC membrane supports was 49 MPa at 38% porosity [27].
The exural strength of recrystallized SiC ceramics was 45 MPa at
47% porosity [18]. The exural strengths of SiC hollow ber mem-
branes with inner and outer diameters of 0.8 and 2.0 mm were
3040 MPa [10]. The exural strength of glass-bonded SiC ceramics
with 5 wt% glass frit was 23 MPa at 46% porosity [30]. The ex-
ural strengths obtained in SC10 and SC23 specimens with 15 wt%
glass frit were 81 and 75 MPa, respectively. The superior strength of
these glass-bonded SiC membrane supports was attributed to the
homogeneous microstructure of the membrane supports (Fig. 4)
and formation of strong bonding between SiC and the present glass
frit, which was obtained by optimizing the glass frit content and
glass frit composition.
Fig. 8 shows the specic ow rate of SiC membrane supports sin-
tered at 850 C for 2 h in air. The specic ow rate of the SC10, SC23,
SC35, and SC65 samples with 15 wt% glass frit at a P of 15 psi were
7.2, 13.4, 16.7, and 34.0 L/min/cm2 , respectively. The ow rate of
the glass-bonded SiC membrane supports increased with increas-
ing SiC particle size in the initial composition. The porosities of
SC10, SC23, SC35, and SC65 samples were 45%, 42%, 42%, and 38%,
respectively. In contrast, the average pore sizes of SC10, SC23, SC35,
and SC65 samples were 3.0, 6.3, 6.9, and 8.2 m, respectively. Thus,
the SC65 has a lower porosity (38%) and larger pore size (8.2 m)
than SC10 (45%, 3.0 m). SC65 showed a higher specic ow rate
(34.0 L/min/cm2 and 63.1 L/min/cm2 at a P of 15 psi and 30 psi,
respectively) than SC10 (7.2 L/min/cm2 and 21.0 L/min/cm2 at a P
of 15 psi and 30 psi, respectively). These ndings indicate that, in
the porosity range of 3845%, the specic ow rate of the glass-
bonded SiC supports is primarily dependent on pore size rather
than porosity.
The permeabilities of SC10, SC23, SC35, and SC65 were
0.24 1012 m2 , 0.42 1012 m2 , 0.52 1012 m2 , and
1.04 1012 m2 , respectively. Song et al. [37] reported permeabil-
ity values in the range of 0.13-0.82 1012 m2 for macroporous SiC
ceramics with a 6875% porosity range. Eom et al. [38] reported
permeability values in the range of 0.411.89 1012 m2 for macro-
porous SiC ceramics with a porosity of 55.658.4%. Ding et al. [20]
reported permeability values in the range of 0.20.9 1012 m2
for macroporous mullite-bonded SiC ceramics with a porosity
of 4557%. Although the processing temperature of the present
glass-bonded SiC membrane supports (850 C) was much lower
that the processing temperatures of both macroporous SiC ceram-
ics (17501950 C) [37,38] and the mullite-bonded SiC ceramics
G Model
JECS-10919; No. of Pages 8 ARTICLE IN PRESS
6 S.C. Kim et al. / Journal of the European Ceramic Society xxx (2016) xxxxxx

(13501500 C) [20], the permeability values of the current glass-

bonded SiC ceramics (0.241.04 1012 m2 ) were comparable
to both macroporous SiC ceramics (0.131.89 1012 m2 ) and
the mullite-bonded SiC ceramics (0.20.9 1012 m2 ). In con-
trast, Wang et al. [31] reported a very high permeability value
(6.53 1012 m2 ) in porous SiC ceramics with a porosity of 45% and
an average pore size of 35 m. The porosity and pore size of SC65
were 38% and 8.2 m, respectively. Thus, the lower permeability
of SC65 (1.04 1012 m2 ) compared to that reported by Wang
et al. [31] (6.53 1012 m2 ) was primarily due to the smaller pore
size of SC65 and secondarily, the lower porosity. This comparison
also conrms that the permeability of the porous SiC membrane
supports could be controlled by adjusting the average pore size of
the porous ceramics.
Ceramic membranes for cleaning pool water and treating waste
water are often subjected to acid and base environments during
operation. Thus, to characterize the chemical durability of SC10 Fig. 9. Flexural strength variation of the SiC membrane supports (SC10) sintered
samples with 15 wt% glass frit, 60 samples were exposed to acid with 15 wt% glass frit at 850 C for 2 h in air after exposure to acidic and basic
and base solutions up to 63 days at room temperature and their solutions for different times.

strengths were evaluated as shown in Fig. 9. The strength of vir-

gin glass-bonded SiC membranes (hereafter, SC10-15) gradually Table 1
EDS results on the bonding area of glass-bonded SiC membrane supports (an average
decreased with exposure. However, the degree of degradation in of 10 samples each): SC10 specimen with 15 wt% glass frit (SC10-15), SC10-15 after
exural strength slowed with prolonged exposure in both solu- 63 days of exposure in a pH 3 solution (SC10-15-pH3), and SC10-15 after 63 days of
tions. The exural strength of the virgin SC10 samples decreased exposure in a pH 11 solution (SC10-15-pH11).
from 80.5 MPa to 57.0 MPa in the pH 3 solution (SC10-15-pH3) and
Sample Element (Atom%)
to 48.4 MPa in the pH 11 solution (SC10-15-pH11) after 63 days of
exposure. Thus, the exural strength of virgin SC10-15 showed 29% Si O C Na Al Ca K Ba

and 40% degradation of its strength after 63 days of exposure in pH 3 SC10-15 28.74 30.36 38.92 1.18 0.53 0.18 0.07 0.02
and 11 solutions, respectively. To understand the cause of degrada- SC10-15-pH3 29.65 27.08 42.23 0.64 0.26 0.10 0.03 0.01
SC10-15-pH11 27.51 28.99 42.62 0.54 0.24 0.06 0.03 0.01
tion on exural strength, the compositions of bonding areas in the
fracture surfaces were analyzed using EDS after 63 days of exposure
in pH 3 and 11 solutions (Fig. 10) and the results are summarized in
modier in the bonding glass. The weight losses of the SC10-15-pH3
Table 1. As shown in Table 1, the content of alkali (Na, K) and alka-
and SC10-15-pH11 specimens were 0.29 wt% and 0.97 wt%, respec-
line earth (Ca, Ba) elements decreased after 63 days of exposure in
tively, after 63 days of exposure in both pH 3 and pH 11 solutions.
both pH 3 and 11 solutions. Those elements play a role as a glass
The weight loss in the pH 11 solution was greater than that in pH

Fig. 10. Scanning electron micrograph images of glass-bonded SiC membrane supports: (a) SC10 specimen with 15 wt% glass frit (SC10-15), (b) SC10-15 after 63 days of
exposure in a pH 3 solution (SC10-15-pH3), and (c) SC10-15 after 63 days of exposure in a pH 11 solution (SC10-15-pH11). The cross denotes the spot analyzed using EDS
with a spatial resolution of 1 m.

Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
G Model
JECS-10919; No. of Pages 8 ARTICLE IN PRESS
S.C. Kim et al. / Journal of the European Ceramic Society xxx (2016) xxxxxx
3. EDS results were also consistent with more loss of alkali (Na, K)
and alkaline earth (Ca, Ba) elements in the SC10-15-pH11 specimen.
The total alkali (Na, K) and alkaline earth (Ca, Ba) element content
decreased from 1.45 atom% for SC10-15 to 0.78 atom% for SC10-
15-pH3 and to 0.64 atom% for SC10-15-pH11. Thus, the 29% and
40% strength degradation of the SC10-15 specimen after 63 days of
exposure in pH 3 and 11 solutions, respectively, were attributed to
the partial loss of glass modiers (Na, K, Ca, Ba), which usually lead
to the degradation of mechanical properties in glasses [39,40]. Thus,
degradation of the mechanical strength in glass-bonded SiC mem-
brane supports exposed to acid or base solutions is unavoidable.
However, the glass-bonded SiC membrane supports maintained its
virgin strength up to 71% and 60% after 63 days of exposure in pH
3 and pH 11 solutions. The strength values (57 MPa and 48 MPa)
after 63-days exposure in pH 3 and 11 solutions are still sufcient
for the ltration of pool and waste waters.
4. Conclusions
Porous silicon carbide membrane supports were successfully
fabricated from SiC and glass frit at a temperature as low as 850 C
in air by a simple pressing and heat-treatment process. During
heat treatment, the glass frit transformed to a viscous glass phase,
which acted as a bonding material between SiC particles and a
protecting layer for severe oxidation of SiC particles during heat
treatment in air. The porosity of the porous SiC membrane supports
decreased with increasing initial particle size of SiC. In contrast,
the pore size increased with increasing initial particle size of SiC.
The exural strength of glass-bonded porous SiC membrane sup-
ports increased with decreasing SiC particle size, indicating that the
exural strength of glass-bonded SiC ceramics is primarily depen-
dent on pore size rather than the porosity in the porosity range of
3746%.
The specic ow rate of the glass-bonded SiC membrane sup-
ports increased with increasing SiC particle size in the initial
composition. This result indicates that the specic ow rate, in
the porosity range of 3746%, is primarily dependent on the
pore size rather than the porosity. The strength retention of the
glass-bonded SiC membrane supports was 71% and 60%, respec-
tively, after 63 days of exposure in pH 3 and 11 solutions. Typical
exural strength, permeability, and specic ow rate of the
porous SiC membrane supports fabricated using 65 m SiC par-
ticles and 15 wt% glass frit were 65 MPa, 1.04 1012 m2 , and
63 L/min/cm2 at a p of 30 psi, respectively.
Acknowledgement
This study was supported nancially by the Fundamental
Research Program of the Korea Institute of Materials Science
(KIMS).
References
[1] M. Fukushima, Y. Zhou, H. Miyazaki, Y.I. Yoshizawa, K. Hirao, Y. Iwamoto, S.
Yamazaki, T. Nagano, Microstructural characterization of porous silicon
carbide membrane support with and without alumina additive, J. Am. Ceram.
Soc. 89 (2006) 15231529.
[2] B. Elyassi, M. Sahimi, T.T. Tsotsis, Silicon carbide membranes for gas
separation applications, J. Membr. Sci. 288 (2007) 290297.
[3] B. Elyassi, M. Sahimi, T.T. Tsotsis, A novel sacricial interlayer-based method
for the preparation of silicon carbide membranes, J. Membr. Sci. 316 (2008)
7379.
[4] M. Fukushima, Y. Zhou, Y.I. Yoshizawa, K. Hirao, Water vapor corrosion
behavior of porous silicon carbide membrane support, J. Eur. Ceram. Soc. 28
(2008) 10431048.
[5] Y.W. Kim, J.H. Eom, C. Wang, C.B. Park, Processing of porous silicon carbide
ceramics from carbon-lled polysiloxane by extrusion and carbothermal
reduction, J. Am. Ceram. Soc. 91 (2008) 13611364.
Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019
7
[6] S.H. Chae, Y.W. Kim, I.H. Song, H.D. Kim, M. Narisawa, Porosity control of
porous silicon carbide ceramics, J. Eur. Ceram. Soc. 29 (2009) 28672872.
[7] M.A. Bautista, J.Q. Cancapa, J. Martinez Fernandez, M.A. Rodriguez, M. Singh,
Microstructural and mechanical evaluation of porous biomorphic silicon
carbide for high temperature ltering applications, J. Eur. Ceram. Soc. 31
(2011) 13251332.
[8] K. Konig, V. Boffa, B. Buchbjerg, A. Farsi, M.L. Christensen, G. Magnacca, Y. Yue,
One-step deposition of ultraltration SiC membranes on macroporous SiC
supports, J. Membr. Sci. 472 (2014) 232240.
[9] S. Naserifar, T.T. Tsotsis, W.A. Goddard III, M. Sahimi, Toward a process-based
molecular model of SiC membranes: III. Prediction of transport and
separation of binary gaseous mixtures based on the atomistic reactive force
eld, J. Membr. Sci. 473 (2015) 8593.
[10] P. de Wit, E.J. Kappert, T. Lohaus, M. Wessling, A. Nijmeijer, N.E. Benes, Highly
permeable and mechanically robust silicon carbide hollow ber membranes,
J. Membr. Sci. 475 (2015) 480487.
[11] V. Suwanmethanond, E. Goo, P.K.T. Liu, G. Johnston, M. Sahimi, T.T. Tsotsis,
Porous silicon carbide sintered substrates for high-temperature membranes,
Ind. Eng. Chem. Res. 39 (2000) 32643271.
[12] M. Fukushima, Y. Zhou, Y.I. Yoshizawa, Fabrication and microstructural
characterization of porous SiC membrane supports with Al2 O3 -Y2 O3
additives, J. Membr. Sci. 339 (2009) 7884.
[13] R. Mourhatch, T.T. Tostsis, M. Sahimi, Network model for the evolution of the
pore structure of silicon-carbide membranes during their fabrication, J.
Membr. Sci. 356 (2010) 138146.
[14] W. Deng, X. Yu, M. Hahimi, T.T. Tsotsis, Highly permeable porous silicon
carbide support tubes for the preparation of nanoporous inorganic
membranes, J. Membr. Sci. 451 (2014) 192204.
[15] D. Song, M. Jung, Gas permeation of Y2 O3 -SiC composite membrane, J. Korean
Ceram. Soc. 52 (2015) 234236.
[16] M. Facciotti, V. Boffa, G. Magnacca, L.B. Jorgensen, P.K. Kristensen, A. Farsi, K.
Konig, M.L. Christensen, Y. Yue, Deposition of thin ultraltration membranes
on commercial SiC microltration tubes, Ceram. Int. 40 (2014) 32773285.
[17] P.K. Lin, D.S. Tsai, Preparation and analysis of a silicon carbide composite
membrane, J. Am. Ceram. Soc. 80 (1997) 365372.
[18] Y. Kim, K. Min, J. Shim, D.J. Kim, Formation of porous SiC ceramics via
recrystallization, J. Eur. Ceram. Soc. 32 (2012) 36113615.
[19] S. Ding, S. Zhu, Y.P. Zeng, D. Jiang, Fabrication of mullite-bonded porous
silicon carbide ceramics by in situ reaction bonding, J. Eur. Ceram. Soc. 27
(2007) 20952102.
[20] S. Ding, Y.P. Zeng, D. Jiang, Gas permeability behavior of mullite-bonded
porous silicon carbide ceramics, J. Mater. Sci. 42 (2007) 71717175.
[21] B.V.M. Kumar, J.H. Eom, Y.W. Kim, I.H. Song, H.D. Kim, Effect of aluminum
hydroxide content on porosity and strength of porous mullite-bonded silicon
carbide ceramics, J. Ceram. Soc. Jpn. 119 (2011) 367370.
[22] Y.S. Chun, Y.W. Kim, Processing and mechanical properties of porous
silica-bonded silicon carbide ceramics, Met. Mater. Int. 11 (2005) 351355.
[23] A. Dey, N. Kayal, O. Chakrabarti, M.D.M. Innocentini, W.S. Chacon, Evaluation
of air permeation behavior of porous SiC ceramics synthesized by
oxidation-bonding technique, Int. J. Appl. Ceram. Technol. 10 (2013)
10231033.
[24] J.H. Eom, Y.W. Kim, Low-temperature processing of silicon
oxycarbide-bonded silicon carbide, J. Am. Ceram. Soc. 93 (2010) 24632466.
[25] K.Y. Lim, Y.W. Kim, I.H. Song, Low-temperature processing of porous SiC
ceramics, J. Mater. Sci. 48 (2013) 19731979.
[26] K.Y. Lim, Y.W. Kim, I.H. Song, Porous sodium borate-bonded SiC ceramics,
Ceram. Int. 39 (2013) 68276834.
[27] W. Shi, B. Liu, X. Deng, J. Li, Y. Yang, In-situ synthesis and properties of
cordierite-bonded porous SiC membrane supports using diatomite as silicon
source, J. Eur. Ceram. Soc. 36 (2016) 34653472.
[28] J.H. Eom, Y.W. Kim, I.H. Song, Effect of SiC ller content on microstructure and
exural strength of highly porous SiC ceramics fabricated from carbon-lled
polysiloxane, J. Korean Ceram. Soc. 49 (2012) 625630.
[29] C.Y. Bai, Y. Li, Z.M. Liu, P.W. Liu, X.Y. Deng, J.B. Li, J. Yang, Fabrication and
properties of mullite-bonded porous SiC membrane supports using bauxite as
aluminum source, Ceram. Int. 41 (2015) 43914400.
[30] S.H. Chae, Y.W. Kim, I.H. Song, H.D. Kim, J.S. Bae, S.M. Na, S.I. Kim, Low
temperature processing and properties of porous frit-bonded SiC ceramics, J.
Korean Ceram. Soc. 46 (2009) 488492.
[31] B. Wang, H. Zhang, H.T. Phuong, F. Jin, J.F. Yang, K. Ishizaki, Gas permeability
and adsorbability of the glass-bonded porous silicon carbide ceramics with
controlled pore size, Ceram. Int. 41 (2015) 22792285.
[32] J.H. Eom, Y.W. Kim, C.B. Park, C. Wang, Effect of forming methods on porosity
and compressive strength of polysiloxane-derived porous silicon carbide
ceramics, J. Ceram. Soc. Jpn. 120 (2012) 199203.
[33] J.H. Eom, Y.W. Kim, S. Raju, Processing and properties of macroporous silicon
carbide ceramics: a review, J. Am. Ceram. Soc. 1 (2013) 220242.
[34] L.J. Gibson, M.F. Ashby, Cellular Soilds, Cambridge University Press,
Cambridge, 1999, pp. 210214.
[35] R.W. Rice, Comparison of stress concentration versus minimum solid area
based mechanical property-porosity relations, J. Mater. Sci. 28 (1993)
21872190.
[36] J.H. Eom, Y.W. Kim, Effect of template size on microstructure and strength of
porous silicon carbide ceramics, J. Ceram. Soc. Jpn. 116 (2008) 11591163.
G Model
JECS-10919; No. of Pages 8 ARTICLE IN PRESS
8 S.C. Kim et al. / Journal of the European Ceramic Society xxx (2016) xxxxxx

[37] I.H. Song, I.M. Kwon, H.D. Kim, Y.W. Kim, Processing of microcellular silicon
carbide ceramics with a duplex pore structure, J. Eur. Ceram. Soc. 30 (2010)
26712676.
[38] J.H. Eom, Y.W. Kim, I.H. Song, Effects of the initial -SiC content on the
microstructure, mechanical properties, and permeability of macroporous
silicon carbide ceramics, J. Eur. Ceram. Soc. 32 (2012) 12831290.
[39] E. Metwalli, R.K. Brow, Modier effects on the properties and structures of
aluminophosphate glasses, J. Non-Cryst. Solids 289 (2001) 113122.
[40] Y. Makhkhas, E.H. Sayouty, Chemical durability and characterization of the
phosphate glasses containing iron, sodium and chromium, Int. J. Adv. Chem. 2
(2016) 17.

Please cite this article in press as: S.C. Kim, et al., Processing and properties of glass-bonded silicon carbide membrane supports, J Eur
Ceram Soc (2016), http://dx.doi.org/10.1016/j.jeurceramsoc.2016.11.019