Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Silicon carbide ceramic membranes
(USC Thesis Other)
Silicon carbide ceramic membranes
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
SILICON CARBIDE CERAMIC MEMBRANES
by
Varaporn Suwanmethanond
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMICAL ENGINEERING)
August 2012
Copyright 2012 Varaporn Suwanmethanond
ii
Dedication
To me,
my husband, Radu,
my daughters, Jasmine and Daisy
iii
Acknowledgments
With gratitude,
My friend, Dr. Ali Mehrabi for his valuable help.
My aunt, Nalinee Esariyakuntham for her support.
All my family, living and deceased.
Finally, this dissertation could not have been published without the help of
Shayna Kessel.
iv
Table of Contents
Dedication .............................................................................................................ii
Acknowledgments ................................................................................................ iii
List of Figures .......................................................................................................vi
Abstract ............................................................................................................ix
Chapter 1: Dissertation Objective and Outline ...................................................... 1
Chapter 1 References ......................................................................... 5
Chapter 2: Porous SiC Sintered Substrates for High-Temperature Membranes .. 6
Abstract ............................................................................................... 6
2.1 Introduction ................................................................................... 7
2.2 Experimental Section .................................................................. 11
2.2.1 SiC Sample Preparation .................................................... 11
2.2.2 Sample Characterization ................................................... 12
2.3 Results and Discussion ............................................................... 15
2.3.1 The Effect of Powder Type ................................................. 15
2.3.2 The Effect of Sintering Aid ................................................. 25
2.4 Conclusions................................................................................. 38
Chapter 2 References ....................................................................... 39
Chapter 3: Silicon Carbide Ceramic Membranes Prepared By A Sol Gel
Technique ......................................................................................... 42
Abstract ............................................................................................. 42
3.1 Introduction ................................................................................. 43
v
3.2 Experimental Procedure ............................................................. 47
3.2.1 The Sol-gel Synthesis of Silicon Carbide ............................ 47
3.2.2 Membrane Preparation Using the Sol-Gel Method .............. 48
3.2.3 The Characterization of SiC Membranes ............................. 49
3.3 Results and Discussion .............................................................. 51
3.3.1 The SiC Powders ................................................................ 51
Table 3.1 The surface area of SiC samples after undergoing
various treatments. ........................................................................ 54
3.3.2 The SiC Membranes ............................................................ 57
3.4 Conclusions ................................................................................ 60
Chapter 3 References ....................................................................... 72
Bibliography ........................................................................................................ 74
vi
List of Figures
Figure 2.1: Particle size distributions for powders P1 and P2. ........................ 16
Figure 2.2: Effect of phenolic resin content on the porosity at constant
content of 0.1 wt% boron. ............................................................. 17
Figure 2.3: Cumulative particle and pore sizes distributions of substrates.
(a) substrate made from the 0.6 µm (P2) average particle size
powder. (b) substrate made from the 1 µm (P1) average
particle size powder. ..................................................................... 18
Figure 2.4: The permeability and separation factor of the P1 and P2
substrates. .................................................................................... 19
Figure 2.5: TEM picture of the SiC powder with an average particle size
of 0.03 µm. ..................................................................................... 19
Figure 2.6: Pore volume distribution (dv/dlogD) of the sintered substrate,
prepared from the powder with an average particle size of
0.03 µm. ......................................................................................... 21
Figure 2.7: (a) SEM of P13 substrate, prepared from mixing powders with
average particle sizes of 1 µ and 0.03 µ in the ratio of 2:1,
(b) SEM for the P1 substrate prepared from the powder with an
average particle size of 1 µm. ........................................................ 23
Figure 2.8: Effect of alumina content on the sample porosity. ......................... 28
Figure 2.9: Effect of boron content on the porosity at a constant phenolic
resin content of 0.05 wt%. ............................................................. 30
Figure 2.10: Effect of the sintering aid on the cumulative pore size
distribution of samples prepared from the P1 powder. .................. 31
vii
Figure 2.11: Effect of the amount of sintering aid on the permeability and
ideal separation factor of samples prepared from the P1
powder. ......................................................................................... 32
Figure 2.12: Effect of the sintering aid on the permeability and cumulative
pore size distribution of samples prepared from the P2 powder. .. 33
Figure 2.13: Effect of the amount of sintering aid on the permeability and
ideal separation factor of samples prepared from the P2
powder. ......................................................................................... 34
Figure 2.14: AFM image of the surface of the sintered substrate, prepared
from β-SiC powder with an average particle size of 0.03 µm,
containing 0.1 wt% boron and 4wt% phenolic resin. ..................... 36
Figure 2.15: Pore size distribution of the SiC unsupported film, prepared by
sol-gel technique. .......................................................................... 37
Figure 3.1: The TEM picture of the organo-silica sol type IPAST with
particle size 8-11 nm, provided by Nissan Chemical
Industries, Ltd................................................................................ 61
Figure 3.2: The XRD pattern of a SiC powder treated with HF, air, and
steam; *signifies the peaks corresponding to the (6H) phase. ...... 62
Figure 3.3: The pore size distribution of the SiC powders after they
have been subjected to various treatments ................................... 63
Figure 3.4: The dV/dlogD of the SiC substrate utilized in the preparation
of the Sol-Gel membranes. ........................................................... 64
Figure 3.5: The argon permeance of the membrane as a function of the
pressure gradient and the number of coatings. ............................. 65
Figure 3.6: The separation factor of the membrane as a function of the
pressure gradient and the number of coatings. ............................. 66
Figure 3.7: AFM image of the SiC sol-gel membrane ..................................... 67
Figure 3.8: The XRD patterns of the SiC membrane and the unsupported
film (powder) prepared by the same techniques. .......................... 68
Figure 3.9: The membrane argon permeance during exposure to steam. ...... 69
viii
Figure 3.10: SEM picture of the cross section of the SiC membrane
prepared by Sol-Gel technique...................................................... 70
Figure 3.11: The XPS spectrum of the SiC powder sample; the
superimposed peaks correspond to SiC at 100.4-101.0 eV
and SiO
2
at 103.0-103.3 eV, respectively. .................................... 71
ix
Abstract
This dissertation focuses on the preparation of silicon carbide (SiC) ceramic
membranes on SiC substrates. An original technique of SiC porous substrate
preparation using sintering methods was developed during the work for the com-
pletion of the dissertation. The resulting SiC substrates have demonstrated high
porosity, high internal surface area, well interconnected surface pore network
and, at the same time, good thermal, chemical and mechanical stability. In a fur-
ther development, sol -gel techniques were used to deposit micro -porous SiC
membranes on these SiC porous substrates. The SiC membranes were charac-
terized by a variety of techniques: ideal gas selectivity (He and N
2
), XRD, BET,
SEM, XPS, and AFM. The characterization results confirmed that the asymmetric
sol-gel SiC membranes were of high quality, with no cracks or pinholes , and ex-
hibiting high resistance to corrosion and high hydro- thermal stability. In conclu-
sion, the SiC ceramic membrane work was successfully completed. Two public a-
tions in international peer reviewed journals resulted out of this work.
1
Chapter 1
Dissertation Objective and Outline
Critical scientific and industrial applications requiring separation of gases rely on
the availability of nano -porous materials. An important class of such materials is
the inorganic separators. Among the inorganic separators, SiC is a bioinert large
band-gap semiconductor that has been receiving sustained interest due a set of
unique properties: low thermal expansion coefficient (4 x 10
-6
o
C
-1
), stability at
high temperatures (>1500
o
C), high thermal conductivity (120 W m
-1
K
-1
), resi s-
tance in acidic and alkali harsh chemical environments and chemical inertness,
thermal shock resistance and high mechanical strength, biocompatibility. SiC
membranes have been shown to exhibit excellent permselectivity at high tem-
peratures in the presence of steam. These unique properties made SiC me m-
branes the component of choice for a wide spectrum of applications, especially in
membrane reactors requiring hydrogen selectivity for increased efficiency power
generation [1].
It is now well known that SiC membranes can be prepared by two different tec h-
niques: a) chemical -vapor deposition (CVD)/chemical -vapor infiltration (CVI);
2
b) pyrolysis of polymeric precursors. While several research gr oups focused their
efforts to develop these techniques, especially during the last decade [2,3], some
of the most important contributions to this field have been carried out at the D e-
partment of Chemical Engineering, USC. This research has been dedicated t o
both new experimental methods for preparing SiC membranes [4,5] as well as
new theoretical models for a fundamental understanding of the mechanisms g o-
verning the formation of the pore space in SiC membranes [6,7]. Notable
achievements at USC in developing novel techniques for fabricating SiC mem-
branes include: improvements in the preparation of amorphous SiC -based mem-
branes by the pyrolysis of allyl -hydridopolycarbosilane (AHPCS) films coated on
tubular SiC porous supports [4], using a combination of slip casting and dip-
coating techniques; further progress in the development of AHPCS -derived
amorphous SiC membranes by a novel sacrificial interlayer-based method [5].
The work presented in this dissertation was completed in August 2001 and was
dedicated to the development of sintering methods for SiC porous substrates as
well as membranes for separation of gaseous mixtures. The findings in this work
were considered to set the foundation for further progress in SiC membrane de-
velopment, as no prior experimentation with SiC had been conducted at Depar t-
ment of Chemical Engineering, USC, at that time. The original results obtained
during the work for this thesis have been reported in two publications [8,9]; in
3
particular , the results presented in reference 8, have been cited in several of the
references [1-7] mentioned above.
For this thesis, SiC ceramic membranes on SiC substrates were fabricated to
yield the desirable specifications such as crack free SiC porous thin films exhibi t-
ing good hydrothermal stability as well as high porosity and surface area with i n-
terconnected pore network in the SiC substrates. SiC porous substrates were
prepared by pressureless sintering of SiC powders under an inert atmosphere of
argon. Three different starting powders and four different sintering aids, Al
2
O
3
,
B
4
C, carbon black, and phenolic resin, either by themselves or in combination,
were investigated in terms of their ability to prepare good quality substrates. This
study confirmed that the porosity, pore size distribution, and transport charact e-
ristics of the resulting SiC sintered bodies depend on the nature of the original
powder, the particle size in the SiC powder, and the type and molar ratio of si n-
tering aid utilized. Depending on the preparation technique, both mesoporous
and macroporous materials could be prepared. These supports were utilized for
the preparation of microporous membranes. The approach to prepare SiC thin
films was the sol -gel technique. SiC thin films were prepared by sol -gel tec h-
niques using silica sol precursors as the source of silicon and phenolic resin as
the source of carbon. The carbothermal reduction reaction between the silicon
and the carbon sources was conducted at 1550
o
C under an argon atmosphere in
order to convert the precursors to SiC. The resulting SiC materials were sub-
4
jected to a hydrothermal stability test, which was carried out by exposing the cal-
cined gel to 50- mol% steam and 50- mol% argon at 350
o
C under atmospheric
pressure for 24 hours. The results show that the final SiC contains micropores
and that they satisfy the criteria of hydrothermal stability. The separation charac-
teristics of the membrane were tested using single gas He and N
2
permeation
tests. The ideal gas s electivity was up to 2.8, indicative of the fact that the thin
SiC films have no cracks or pinholes . Investigations by, XRD, BET, SEM, XPS
and AFM techniques were conducted on the same films and the results sup-
ported the above claim.
5
Chapter 1 References
1. Zhou, Y.; Fukushima, M.; Miyazaki, H.; Yoshizawa, Y.; Hirao, K.; Iwamoto,
Y.; Sato, K., Preparation and characterization of tubular porous silicon
carbide membrane supports. Journal of Membrane Science, 2011, 369,
112.
2. Takeda, Y.; Shibata, N.; Kubo, Y., SiC coating on porous γ-Al
2
O
3
using
alternative supply CVI method. Journal of the Ceramical Society of Japan,
2001, 109, 305
3. Pages, X.; Rouessac, V.; Cot, D.; Nabias, G.; Durand, J., Gas permeation
of PECVD membranes inside alumina substrate tubes. Separation and
Purification Technology, 2001, 25, 399.
4. Elyassi, B.; Sahimi, M.; Tsotsis, T. T, Silicon carbide membranes for gas
separation applications. Journal of Membrane Science, 2007, 288, 290.
5. Elyassi, B.; Sahimi, M.; Tsotsis, T. T, Silicon carbide membranes for gas
separation applications. Journal of Membrane Science, 2008, 316, 73.
6. Rajabbeigi, N.; Elyassi, B .; Tsotsis, T. T.; Sahimi, M., Molecular pore-
network model for nanoporous materials, I: Application to adsorption in sil-
icon-carbide membranes. Journal of Membrane Science, 2009, 335, 5.
7. Mourhatch, R., Tsotsis, T. T.; Sahimi, M., Network model for the evolution
of the pore structure of silicon-carbide membranes during their fabrication.
Journal of Membrane Science, 2010, 356, 138.
8. Suwanmethanond, V.; Goo, E.; Johnston, G.; Liu, P.; Sahimi, M.; Tsotsis,
T.T.; Porous SiC Sintered Substrates for High Temperature Membranes
for Gas Separations. Journal of Industrial & Engineering Chemistry R e-
search. 2000, 39[9], 3264.
9. Ciora, R. J.; Fayyaz, B.; Liu, P. K. T.; Suwanmethanond, V.; Mallada, R .;
Sahimi, M.; Tsotsis, T. T., Preparation and reactive applications of nano-
porous silicon carbide membranes. Chemical Engineering Science, 2004,
59, 4957.
6
Chapter 2
Porous SiC Sintered Substrates for
High-Temperature Membranes
Abstract
Silicon carbide (SiC) porous substrates are prepared by pressureless sintering of
SiC powders under an inert atmosphere of argon. The porous SiC substrates
were characterized by measuring their porosity, pore size distribution, surface
characteristics and structure. Their transport characteristics were investigated
using N
2
and He as the test gases. Three different starting powders and four di f-
ferent sintering aids, Al
2
O
3
, B
4
C, carbon black, and phenolic resin, either by
themselves or in combination, were investigated in terms of their ability to pr e-
pare good quality substrates. It was found that the porosity, pore size distribution,
and transport characteristics of the resulting SiC sintered bodies depend on the
nature of the original powder, the particle size in the green SiC samples, and the
type and molar ratio of sintering aid utilized. Depending on the preparation tec h-
7
nique, both mesoporous and macroporous materials could be prepared. These
supports are currently utilized for the preparation of microporous membranes.
2.1 Introduction
Silicon carbide (SiC) is a promising material for high temperature membrane a p-
plications, because it is capable of withstanding high temperatures and mechani-
cal stresses, and corrosive environments. In addition, it has a high thermal co n-
ductivity and a low thermal expansion coefficient, potentially advantageous cha-
racteristics during high temperature membrane separations. Relatively few st u-
dies have reported the use of SiC for membrane applications [1,2]. Sea et al. [1]
prepared SiC membranes by chemical vapor deposition (CVD) on the top of α-
alumina substrates at 700-800
o
C using triisopropylsilane as the precursor. After
the deposition process the membranes were heat -treated in deoxidized argon at
1000
o
C for 1 h. The resulting membranes were tested for the separation of
H
2
/H
2
O mixtures. The permeation experiments were performed at temperatures
between 200 and 400
o
C, and have shown a H
2
/H
2
O permselectivity in the range
of 3-5. Using α-alumina as the support for the preparation of SiC membranes
presents a number of challenges for the type of applications these membranes
may be useful for. Porous α-aluminas are not very resistant, relative to SiC, to
chemically corrosive environments. In addition, their thermal expansion coeff i-
cient is different from that of silicon carbide, raising concerns about the mechani-
8
cal stability of the resulting composite membrane system. Lee and Tsai [2] pre-
pared asymmetric mesoporous SiC membranes by low pressure CVD (LPCVD)
of SiH
4
, C
2
H
2
, and Ar mixtures at 800
o
C on the surface of Al
2
O
3
-dopped SiC ma-
croporous supports. The alumina content of the macroporous SiC support de-
creased from the center to the surface of the support.
The preparation of good quality porous SiC substrates, to be used for the prep a-
ration of meso- and microporous membranes, still remains a challenging technic-
al problem. This study is among the first that systematically focuses on the tec h-
nical issues involved in the preparation of such materials. In our investigations
we have drawn upon the knowledge that has been acquired through the years in
the preparation of dense SiC materials. Extensive technical literature is available
on preparing SiC dense ceramics for use as SiC heating elements, seals, tools,
and high temperature gas turbine elements. The general consensus here is that
good quality sintered samples cannot be prepared without the addition of an ap-
propriate sintering aid [3-5]. This is because SiC is basically a non- sinterable
material, because of its highly covalent bonding. However, the addition to the
SiC powders of components such as carbon, boron, and alumina as sintering
aids, either by themselves or in combination, has been shown to result in signif i-
cant densification [5]. Prochazka [3], as an example, used different ceramic
forming processes to produce SiC green samples from a sub- micron β-SiC
powder. He sintered the samples at 1950- 2100
o
C in the presence of boron and
9
carbon, as sintering aids, to obtain dense samples with a density as high as 95-
98% of the theoretical value for dense SiC. Coppola et al. [6], in order to obtain
higher densification, proposed the use of a graphite crucible, which had been
previously saturated with boron, in addition to the boron and carbon added to the
SiC powder as sintering aids. Providing the carbon in the form of free carbon
and carbonaceous additiv es [7] together with boron also results in the prepar a-
tion of dense SiC samples with densities higher than 85% of the theoretical de n-
sity. The particle size of the original powder also plays an important role, as r e-
ported by Owens and Ruppel [8]. They used boron carbide and phenolic resin as
the sintering aids, in order to prepare sintered SiC bodies with porosity in the
range 3-25% by utilizing a variety of powders with different particle size distribu-
tions. The use of alumina as the sintering aid and its effects on the final density
of the resulting SiC samples prepared by pressureless sintering have been st u-
died by a number of authors [9-15]. Dense hot -pressed SiC, for example, has
been prepared by Lange [11] by addition of alumina as the sintering ai d into SiC
powder.
It should be obvious from the previous discussion that substantial work has al-
ready been carried out with the goal of preparing either very low porosity or com-
pletely dense SiC parts. However, such SiC materials are of little interest in the
area of high temperature membranes. For SiC materials to be appropriate as
starting supports in the preparation of high temperature microporous me m-
10
branes, they must be highly porous (porosity preferably in the range 30- 45%),
and must have an interconnected, sample- spanning porous structure. Little e f-
fort in the open literature has gone into producing materials of this kind to be
used as optimal membrane supports for gas separation. In fact, the study of Lee
and Tsai [2], previously referred to, is the only one we are aware of that makes
use of macroporous SiC substrates for SiC membrane preparation. To prepare
materials, which are appropriate for use as supports for the preparation of high
temperature microporous membranes, one must first try to systematically under-
stand the effects that the preparation conditions have on their final porosity, pore
size and pore size distribution, surface characteristics, and internal surface area.
The particle size and particle size distribution, the type of SiC powders one uses,
the type and the amount of additives and the way one homogeneously blends
these additives with the major components, the fabrication method of the green
samples, and the sintering temperature and duration all contribute to the proper-
ties of the final SiC part.
This study is among the first efforts in this area. We aim here to study the effect
of the nature of the starting powder material (the particle size and type of SiC
powder, i.e., α or β), and of the use of a number of conventional sintering aids in
the preparation of supports appropriate for the preparation of high temperature
microporous membranes. In the experiments we have used the pressureless
sintering technique. We have found it to be a practical and efficient method that
11
has allowed us to prepare samples with controlled porosity, shape and mechani-
cal properties. To reiterate, the purpose of this study is to produce SiC porous
substrates with an interconnected porous structure with a relatively high (30-
45%) porosity, which are suitable for further modification by sol -gel or CVD tech-
niques for the production of microporous membrane materials.
2.2 Experimental Section
2.2.1 SiC Sample Preparation
For the preparation of the green samples we first mix the SiC powder(s) together
with the appropriate sintering aids using acetone as the dispersing medium.
Oleic acid is then added as the pressing aid. The resulting slurry materials are
then mixed thoroughly to ensure complete homog enization using an ultrasonic
device for 20 min, and are then dried in an oven. The dried powder is then
pressed into a cylindrical flat mold utilizing fitted stainless steel disks with a pres-
sure of up to 117.15 MPa to obtain disk -shaped green samples 3.2 cm (1.25 in)
in diameter. The SiC green samples are then heated at a rate of 3
o
C/min in a
graphite resistance furnace (Thermal Technology, Inc.), until they reach the ap-
propriate sintering temperature, and are kept there for a predetermined period of
time under an inert atmosphere of argon gas. Upon completion of the sintering
process, the samples are cooled down to room temperature at a cooling rate of
12
6
o
C/min. A calibrated optical pyrometer is used for temperature measurement
and control. The resulting sintered porous SiC substrates are characterized by
measurements of their overall porosity (using the Archimedes method), their pore
size distribution (using the bubble- point test and Brunauer -Emmett-Teller (BET)
analysis techniques), scanning electron microscopy (SEM), and atomic force m i-
croscopy (AFM). These characterization techniques are complimented by trans-
port investigations using single gases.
2.2.2 Sample Characterization
For measuring the porosity of the sintered SiC samples, we have used the Ar c-
himedes method. One advantage of this method is that it measures only the ac-
cessible pores that are of relevance to membrane transport. Another advantage
is that it is not a destructive technique. In our measurements acetone was used
as the buoyant liquid instead of water, because it wets the surface of the SiC bet-
ter (our measurements indicate a contact angle = 0). The key assumption of this
method is that acetone penetrates into all the accessible pores of the SiC sa m-
ple. The porosity (in percent) is calculated using the following formula.
ε
ρ
= − { }} 1 100
{
W - W
V
air a
a s
, (2.1)
13
where ε is the percent accessible porosity of the substrate, W
air
is the weight of
the substrate in air, W
a
is the corresponding weight of the substrate in acetone,
V
S
is the bulk volume of the substrate, and ρ
a
is the density of acetone at 25
o
C.
A digital linear gauge (EG -133 ± 0.0001% accuracy, Mitutoyo) is used to meas-
ure the thickness and diameter of each substrate. The repeatability in the porosi-
ty measurements was shown to be better than 0.5%.
Our membranes have been characterized for their transport properties by expe-
riments conducted at room temperature using two relatively inert gases, Helium
(He) and Nitrogen (N
2
), to obtain the ideal separation factor (defined as the ratio
of permeabilities of the two individual gases). The diffusion cell for measuring
the properties of flat membranes has been previously described [16]. The abso-
lute values of permeability and the separation factor are important in determining
the appropriateness of the resulting samples as substrates for the preparation of
microporous membranes. The permeability of each species is calculated from
the following equation.
K
QL
A P
i
i
i
=
∆
, (2.2)
where K
i
is the permeability of each gas in cm
3
/(cm min psi), Q
i
is the gas vol u-
metric flow rate (cm
3
/min) through the membrane (measured by a soap- bubble
14
flow meter), A is the cross -sectional area (cm
2
), L is the thickness of the me m-
brane (cm), and ∆P
i
is the pressure difference across the membrane disk (psi).
For mesoporous systems, the experimentally measured separation factor is
compared to the ideal Knudsen separation factor, which is given by the following
relationship:
S
MW
MW
k
N
He
=
2
,
(2.3)
where MW are the molecular weights of the two gases.
For the measurement of the pore size distribution, we have used two techniques.
For the pores smaller than 100 nm, the pore size distribution is determined by N
2
adsorption-desorption experiments using an ASAP2010 (Micromeritics) appar a-
tus and the BJH and Horvath-Kawazoe models. The pore size distribution for the
pores greater than 100 nm is calculated by the bubble- point test technique de-
scribed in ASTM 2499.
15
2.3 Results and Discussion
2.3.1 The Effect of Powder Type
We have used two commercially available SiC powders with similar average par-
ticle size (in the micron range) and pore size distributions. These are an α-SiC
powder distributed by HC Starck (hereinafter referred to as the powder P1), with
a nominal average particle size of 1 µm, and a β-SiC powder (HSC059) distr i-
buted by Superior Graphite Co. (hereinafter referred to as the powder P2), with a
nominal average particle size of 0.6 µm. The particle size distributions for both
the P1 and P2 powders (as provided by their manufacturers) are shown in Figure
2.1. Though the reported nominal average diameters are different, in reality the
two powders appear to have very similar size characteristics. The goal in using
these two powders is to compare the effect that the nature of the starting powder
( α or β) has on the properties of the final sintered samples, because as already
noted their average particle sizes and distributions are fairly similar. The thinking
here was that the final sintered samples made from the P1 powder would have
higher porosity than the corresponding samples made from the P2 powder. The
reason for that, we believe, is because the grain growth for P1 is reported [7] to
be faster than that of P2, resulting in anisotropic grain growth and final sintered
samples with a lower degree of densification.
16
Figure 2.1 Particle size distributions for powders P1 and P2.
Experimental results on sintered sample porosity, shown in Figure 2.2, are co n-
sistent with this hypothesis. Samples shown in this figure were prepared with
powders P1 and P2 (and a third powder P3 to be discussed further later). Each
contains as sintering aids 0.1wt% B (added in the form of B
4
C) and phenolic r e-
sin in the range of 0- 4wt% (see further discussion in the Effect of Sintering Aids
Section).
17
Figure 2.2 Effect of phenolic resin content on the porosity at constant con-
tent of 0.1 wt% boron.
There is scatter in the experimental porosity data, but clearly the P1 samples
have on the average a higher porosity than the P2 samples. The cumulative
pore size distributions (measured by the bubble flow technique) for two samples,
one made from powder P1 and the other from powder P2, each containing 0.1
wt% B and 4 wt% phenolic resin are shown in Figures 2.3a and 2.3b, together
with the cumulative particle size distributions for the original powders.
18
Figure 2.3 Cumulative particle and pore sizes distributions of substrates.
(a) substrate made from the 0.6 µm (P2) average particle size powder.
(b) substrate made from the 1 µm (P1) average particle size powder.
The samples are macroporous and no pinholes or cracks were detected during
the bubble-point test, an important result in terms of the further utilization of such
substrates as supports for the preparation of microporous membranes. The
transport characteristics of these two samples, in terms of the corresponding He
permeabilities and the ideal He/N
2
separation factor, are shown in Figure 2.4.
19
Figure 2.4 The permeability and separation factor of the P1 and P2 substrates.
Figure 2.5 TEM picture of the SiC powder with an average particle size of
0.03 µm.
20
Sample P1, which has the higher porosity and the more open structure, also has
permeability that is 6 times higher than that of the P2 sample. This is indicative of
the significant effect that the choice of the type of the starting powder ( α vs. β)
has on the properties of the resulting substrates.
We have, in addition, also used a third β-SiC powder, kindly provided to us by
Sumitomo Osaka Cement Co., Ltd. (hereinafter referred to as the powder P3)
with an average particle size in the sub- micron range, i.e., 0.03 µm, as reported
by the manufacturer. Closer inspection of a TEM image of the powder (provided
by the manufacturer; see Figure 2.5) indicates that, for the most part, the par-
ticles are uniform in diameter and close to the reported 0.03 µ m values. There
are, however, a number of smaller particles present, which may tend to explain
the broadness of the pore size distribution that is observed for the sintered sub-
strates (see discussion). The purpose for using this third powder is to investigate
the effect of particle size on the properties of the resulting final samples. The po-
rosities of samples prepared with the P3 powder, with 0.10wt% B and 0- 4wt%
phenolic resin as sintering aids, are also shown in Figure 2.2. The smaller size
β-SiC powder results in samples with a higher porosity than the larger particle β-
SiC powder. In addition, phenolic resin seems to have a smaller effect on the po-
rosity of these samples. This observation is somewhat counterintuitive, and still
remains the topic of ongoing investigations. It is known, for example, that particle
size and size distribution have a complex effect on the porous characteristics of
21
the compacted green samples [17], and the effect of initial pore size characteri s-
tics of the green samples on the density and pore size distribution of the sintered
substrates is also complex, and not well understood (see further discussion and
references). The pore size distribution (measured by BET) for one of the sintered
samples made with the P3 powder with 0.1 wt% B and 4 wt% phenolic resin as
sintering aids, is shown in Figure 2.6. Clearly, this sample contains pores in the
mesoporous region. Of note, in particular, is the bimodal pore size distribution
with a small shoulder centered around 20 Å.
Figure 2.6 Pore volume distribution (dv/dlogD) of the sintered substrate,
prepared from the powder with an average particle size of 0.03 µm.
22
BET analysis of the original P3 powder also indicates the presence of a small
shoulder in this micro/mesoporous range. The pores in the lower region of pore
diameters occupy a small fraction of the total pore volume, with the pore diam e-
ters in the meso/macro porous region determining the transport characteristics.
We have also prepared a number of samples by combining the P3 powder either
with the P1 or the P2 powder. We have tried three different approaches in the
preparation of such substrates. The first two methods were aimed at developing
asymmetric membranes, in which the smaller particle powder forms a thin layer
on the top of the macroporous support prepared by the larger particle powder.
This may be achieved by preparation of a thin layer of the mesoporous substrate
on the top of the macroporous support. The first method involves modifying the
surface of a previously sintered substrate (made from the larger particle powder
P1 or P2) by sli p-casting using a suspension of the P3 powder. The resulting
composite substrate is then sintered in the graphite furnace at 1950
o
C under an
inert atmosphere of argon using the sintering protocol previously described. The
second technique involved modifying the surface of a green sample made of the
larger powder by pressing on the top of it a layer of the smaller particle P3 pow d-
er. The resulting composite green samples are then sintered at 1950
o
C, as pre-
viously described. We have found it difficult to prepare good quality, crack -free
substrates using these two techniques.
23
(a)
(b)
Figure 2.7 (a) SEM of P13 substrate, prepared from mixing powders with aver-
age particle sizes of 1 µ and 0.03 µ in the ratio of 2:1, (b) SEM for the P1 substrate
prepared from the powder with an average particle size of 1 µm.
24
We attribute this (at least for the samples made by the second technique above)
to the different shrinkage characteristics during sintering. To verify this, we have
measured the shrinkage of each SiC sample from its green to its final sintered
form. We have found that the shrinkage of the P1 substrates is about 5%, that of
the P2 substrates is about 10%, and that of the P3 substrates is about 7- 9% at
the processing conditions. Though small, the differences turn out to be signif i-
cant in preventing the two powders from sintering together perfectly in a compo-
site two-layer form without the formation of cracks.
In the third approach we have tried two different combinations of powders (P1
and P3 or P2 and P3), mixing them in different proportions with boron carbide
and phenolic resin as the sintering aids in order to prepare the green samples.
As expected, the sintered SiC samples have a bimodal pore size distribution,
which is characterized by a region consistent with the presence of the large par-
ticle powder, and a substantial region, consistent with the presence of the smaller
particle size powder. The SEM picture (see Figure 2.7a) of one of the P13 sam-
ples (made from 67% P1 and 33% P3 powders, with 0.1 wt% B and 4 wt% phe-
nolic resin as sintering aids), for example, shows the visible surface pore sizes to
be in the range of 2- 5 µm, and the particle sizes to be in the range of 5- 10 µm.
Both the observable pore size and the particle sizes are substantially larger than
the corresponding pore and particle sizes of samples made with the powder P1
alone sample (see Figure 2.7b). Clearly, the presence of the P3 powder aids the
25
sintering process of the P1 powder. BET analysis also indicates the presence of
a porous region, with similar, in general characteristic of the P3 powder. There
are, however, also subtle differences in terms of the position and the size of the
various peaks. Similar observations apply for the samples made from mixing the
P2 and P3 powders. Mixing powders together provides one with an additional
degree of freedom in the preparation of supports with characteristics appropriate
for further modification for the preparation of microporous membranes.
2.3.2 The Effect of Sintering Aid
In addition to the type of the starting powder, we have found that sintering aids
(type and amount) are key in determining the quality of the final produced sub-
strate. To study their effects, we have utilized four different sintering aids to pr e-
pare SiC macro- and mesoporous substrates. These are: (i) B
4
C (99.7% pure,
Alfa Chemicals) with a particle size in the range of 1- 7 µm commonly used as a
boron source; (ii) carbon black (R.T. Vanderbilt Company , Inc.) with a particle
size in the range of 1- 10 µm as a carbon source; (iii) phenolic resin (resole type,
Monomer-Polymer Dajac Labs, Inc.) often used as a carbon source, but also as a
temporary binder, and (iv) alumina oxide powder (Buehler) with a reported aver-
age particle size of 0.03 µm. These sintering aids have been previously studied
during the preparation of dense SiC specimens, and have been reported to have
a variety of effects during the sintering process of the green samples. Some of
26
these effects are still not very well understood. Carbon is reported to participate
in a number of chemical reactions. It is thought, for example, to react with the
omnipresent SiO
2
impurities to produce SiC according to the following reaction:
SiO
2
+ 3 C → SiC + 2 CO (2.4)
In addition, carbon at the sintering temperatures of ~ 1950
o
C that is utilized here
will readily react with any free Si to also form SiC:
Si + C → SiC (2.5)
Boron and carbon are both thought to aid the overall densification process by sol-
id phase sintering [3,5,7,18]. It is believed that in pure SiC the grain boundary
energy is very high and, therefore, prohibits the sintering process. Boron reduc-
es the ratio of the grain boundary energy to the surface energy. In its pres ence,
the formation of grain boundaries becomes, as a result, more favorable than the
formation of free surface in SiC particles. Consequently, particles reduce their
free surface and sinter together to reach a lower energy state [5,18]. Carbon is
thought to increase the overall mobility of SiC. Both carbon black and the phe-
nolic resin are reported in the literature to be good sources of carbon. The phe-
nolic resin, however, is expected to be a more effective densification agent, be-
cause it may be made more readily and uniformly available to all SiC particles.
27
Alumina is thought to participate in sample densification through a liquid- phase
sintering mechanism [9-15]. For the range of temperatures we use for the sinter-
ing of our samples it has been reported that an alumina- rich phase is created, in
which the β-SiC dissolves partially [9-15,19]. The SiC may then re- precipitate in
its crystalline α form [9,15,19], a process which is believed to promote densific a-
tion [11,13]. Anisotropic grain growth and formation of plate-like grains is usually
observed due to the β α transformation [15, 20-22]. This confirms the pres-
ence of an interface- reaction-controlled coarsening step [19], which usually oc-
curs during liquid- phase sintering. The grain size increases with sintering time
[14,19] (usually proportionally to the square root of time) [10],
but is independent
of the amount of the liquid phase present [10,19]. It is believed that at the begi n-
ning of the process of coarsening, Al
2
O
3
reacts and dissolves the oxide layer
(SiO
2
) at the surface of the SiC particles [13,15], resulting in the formation of an
alumina-rich alumina-silicate liquid phase [11,13]. It is this liquid phase that is
responsible for the SiC dissolution and reprecipitation processes. Upon cooling,
the liquid phase is crystallized to α-alumina and occupies the space between the
SiC grains [14]. The sintered samples, upon prolonged heating above 1700
o
C,
may end up containing less alumina than that added to the original green sam-
ples.
28
Figure 2.8 Effect of alumina content on the sample porosity.
The reduction of the alumina content occurs by the following reaction [12], which
results in the formation of volatile compounds:
SiC(s) + Al
2
O
3
(s) Al
2
O(g) + SiO(g) + CO(g) (2.6)
Our observations in the preparation of porous SiC samples are, generally, in line
with the data reported for their dense counterparts. We have found alumina to
be an effective sintering aid. On the other hand, as can be seen in Figure 2.8 ,
the exact amount of alumina added in the original green samples does not seem
29
to be a determining factor of the porosity of the final sintered samples. This o b-
servation is in agreement with the observation by others that above a critical
alumina content the grain size is independent of the amount of the alumina liquid
phase but is, instead, dependent on the time and temperature of sintering. Add-
ing to the green samples phenolic resin, in addition to alumina, as a sintering aid
seems to result in sintered samples with a somewhat reduced porosity. Again, in
the presence of phenolic resin, changing the alumina content has little noticeable
effect on the final porosity. One of the downsides of the use of alumina as a si n-
tering aid is the fact that the final sintered samples contain a second phase, i.e.,
alumina. This is likely to potentially weaken the substrate's resistance to corr o-
sive (acidic or caustic) environments [23].
We have found it difficult to prepare samples with consistent porosity and trans-
port characteristics using carbon black as the carbon source. This may be attr i-
buted to the inability to distribute the carbon uniformly and evenly to all the SiC
grains and particles throughout the mass of the green samples. Boron carbide
and phenolic resin, on the other hand, have been shown to be effective sintering
aids in the preparation of SiC samples with consistent and good overall porosity
and transport characteristics. As can be seen in Figure 2.2, increasing the con-
tent of the phenolic resin in the green sample, while keeping the boron carbide
content constant, decreases the porosity of the resulting membranes. The effect
becomes less noticeable at higher boron contents.
30
Figure 2.9 Effect of boron content on the porosity at a constant phenolic resin
content of 0.05 wt%.
In Figure 2.9 we plot the porosity of the sintered samples resulting from the three
different powders prepared with a constant content of phenolic resin and varying
amounts of boron. As can be seen in Figure 2.9, increasing the boron content of
the green samples, while keeping the phenolic resin content constant, again r e-
sults in a decrease in the porosity of the resulting membranes. Once more the
effect becomes less noticeable at higher phenolic resin contents. Of the four sin-
tering aids we have tested to produce porous SiC supports, phenolic resin and
boron carbide are the more appropriate and promising candidates. By appr o-
priately manipulating the content of these two sintering aids in the original
31
Figure 2.10 Effect of the sintering aid on the cumulative pore size distribution of
samples prepared from the P1 powder.
green samples, one can predictably and consistently prepare SiC supports with
the proper porosity. We have been able to prepare well -sintered substrates with
a B content as low as 0.025 wt%. In fact, adding a larger amount than that may
turn out to be counterproductive. The difference in behavior that we have ob-
served between the phenolic resin and the carbon black systems is, likely, due to
the fact that the phenolic resin is in the liquid phase. As a result, it most likely,
disperses better within the SiC powder and mixes together more effectively with
the other sintering additives, for example, boron carbide.
32
Figure 2.11 Effect of the amount of sintering aid on the permeability and ideal
separation factor of samples prepared from the P1 powder.
As previously noted, the beneficial effect of the B
4
C-phenolic resin system is,
most likely, due to solid-phase sintering as a result of the reduction in the ratio of
the grain boundary energy to the surface energy [18]. Carbon, in addition, has a
number of other beneficial effects, which also result in enhanced sintering of the
SiC [3]. As previously noted, it will reduce or remove the oxide film at the SiC
powder grain surface according to the reaction SiO
2
+ 3C SiC + 2CO and will
react with free silicon (that may form by the reaction SiO
2
+ 2SiC 3Si + 2CO)
in the starting powder according to the reaction Si + C SiC.
33
Figure 2.12 Effect of the sintering aid on the permeability and cumulative pore
size distribution of samples prepared from the P2 powder.
34
Figure 2.13 Effect of the amount of sintering aid on the permeability and ideal
separation factor of samples prepared from the P2 powder.
The type and amount of sintering aid also has a significant effect on the pore size
distribution and transport characteristics of the resulting SiC substrates. Figure
2.10 shows the pore size distribution of two samples both made with the P1
powder. The first sample was made by adding 0.1 wt% B and 0 wt% phenolic
resin. The second sample was made by adding 0.1 wt% B and 4 wt% phenolic
resin. It is clear from these figures that the additional amount of phenolic resin as
sintering aid shifts the pore size distribution toward the smaller pore sizes. The
He permeability and the He/N
2
ideal separation factor for both samples are
shown in Figure 2.11. In agreement with the pore size distribution results, the
35
sample made with 0.1 wt% B and 4 wt% phenolic resin as sintering aids has a
smaller permeability and a higher ideal separation factor. Similar trends are o b-
served with samples prepared from powder P2 as can be seen in Figures 2.12
and 2.13. The effects of sintering aids on the pore size distribution and transport
properties are more complex than the above four figures would imply and the
corresponding effect of these aids on total porosity. Such phenomena still r e-
main the subject of ongoing investigations by our group. Because the sintering
occurs in the solid phase, many parameters including the grain and volume diff u-
sion affect the pore size dynamics. The percentage of the sintering aids added,
as well as the degree to which they homogeneously disperse throughout the
samples, determines the pore sizes. The local packing imperfections also contr i-
bute to a phenomenon called pore channeling, which actually increases the pore
sizes in the initial and intermediate steps of sintering [24-29]. In fact, simulations
[30-32] have confirmed that heterogeneous sintering increases the size of the
pores by introducing different shrinkage rates at different spatial locations.
36
Figure 2.14 AFM image of the surface of the sintered substrate, prepared from β-
SiC powder with an average particle size of 0.03 µm, containing 0.1 wt% boron
and 4wt% phenolic resin.
In addition to being crack -free, and having a highly porous and interconnected
structure, supports to be used in the preparation of microporous membranes
must be sufficiently smooth to allow for the deposition of microporous and mes o-
porous sol-gel films. We have been able to prepare by the appropriate choice of
initial powder, sintering aid, and sintering conditions SiC supports that are suff i-
ciently smooth for the deposition of SiC sol -gel films. Fig ure 2.14, for example,
shows an AFM image of the surface of one of these supports with an average
roughness of about 400 Å, which is suitable for coating on its surface of sol -gel
layers. The pore size distribution of an unsupported microporous SiC sol -gel
37
layer, prepared by our group, is shown in Figure 2.15. Further details for the
preparation and characterization of sol-gel, thin SiC supported films will be pr e-
sented in future publications.
Figure 2.15 Pore size distribution of the SiC unsupported film, prepared by sol-
gel technique.
38
2.4 Conclusions
In this work we have reported the preparation of pinhole- and crack-free, smooth
porous SiC substrates with interconnected porous structures. It has been shown
that the porosity, average pore size and the pore size distribution of these su b-
strates can be varied over a broad range by the appropriate choice of the initial
powders, sintering aids, and preparation conditions. Of the sintering aids that we
have used, boron carbide and phenolic resin (as the boron and carbon sources,
respectively) appear to be the most promising for obtaining appropriate SiC sub-
strates. The ability to homogeneously disperse the sintering aids has a signif i-
cant effect on the final shape and size of the pores after sintering. The proper-
ties of the resulting membranes strongly depend on the type of the SiC powders
utilized ( α or β), their particle sizes and distributions, and the mixing ratios. We
have been unable, so far, to prepare asymmetric membranes by slip- casting
powders with smaller size particles on underlying macroporous support. We
attribute this to our inability, so far, to perfectly match the shrinkage of the pow d-
ers that we have utilized.
39
Chapter 2 References
1. Sea, B.K.; Ando, K; Kusakabe, K; Morooka, S. Separation of Hydrogen
From Steam Using a SiC-Based Membrane Formed by Chemical Vapor
Deposition of Trisopropylsilane. Journal of Membrane Science. 1998,
146[1], 73.
2. Lee, L.; Tsai, D.S. Silicon Carbide Membranes Modified by Chemical Va-
por Deposition Using Species of Low-Sticking Coefficients in a Si-
lane/Acetylene Reaction System. Journal of American Ceramics Society .
1998, 81[1], 159.
3. Prochazka, S. Sintering of Silicon Carbide. In Ceramics for High Perfor-
mance Applications; Burke, J.J., Gorum, A.E., Katz, R.N., Eds.; Published
in cooperation with the Metals and Ceramics Information Center, Colum-
bus,: OH, 1973.
4. Hojo, J. Sintering Behavior of Ultrafine Silicon Carbide Powder. In Silicon
Carbide Ceramics-1; Somiya, S., Inomata, Y., Eds., Elsevier Applied
Science,: New York, NY, 1991.
5. Tanaka, H. Sintering of Silicon Carbide. In Silicon Carbide Ceramics-1;
Somiya, S., Inomata, Y., Eds.; Elsevier Applied Science, New York, NY,
1991.
6. Coppola, J.A.; Hailey, L.N.; McMurtry, C.H. Sintered Alpha Silicon Carbide
Ceramic Body Having Equiaxed Microstructure. US. Patent
4,179,299,1979.
7. Prochazka, S. Sintered Dense Silicon Carbide. US. Patent
4,004,934,1977.
8. Owens, D.P.; Ruppel, I.B. Porous Silicon Carbide. US. Patent
5,441,799,1995.
9. Suzuki, K. Pressureless-Sintered Silicon Carbide with Addition of Alumi-
num Oxide. In Silicon Carbide Ceramics-2; Somiya, S., Inomata, Y., Eds.;
Elsevier Applied Science.: New York, NY, 1991.
10. Ye, H.; Pujar, V.V.; Padture, N.P. Coarsening Mechanism in Liquid-Phase-
Sintered SiC. American Ceramic Society Annual Meeting, Ceramograhic
Exhibit, 1997.
40
11. Lange, F.F. Hot-Pressing Behavior of Silicon Carbide Powders with Addi-
tions of Alumina Oxide. Journal of Materials Science. 1975, 10, 314.
12. Grande, T.; Sommerset, H.; Hagan, E.; Wiik, K.; Einarsrud, M.A. Effect of
Weight Loss on liquid-Phase-Sintered Silicon Carbide. Journal of Amer i-
can Ceramics Society. 1997, 80[4], 1047.
13. Hirata, Y.; Hidaka, K.; Matsumura, H.; Fukushige, Y.; Sameshima, S. Col-
loidal Processing and Mechanical Properties of Silicon Carbide with Alu-
mina. Journal of Materials Research. 1997, 12[11], 3146.
14. Sasaki, M.; Suzuki, K.; Nishimura, H.; Noshiro, M. Microstructure of Pres-
sureless Sintered Silicon Carbide with Additions of Aluminum Oxide. In
High Tech Ceramics. Vincenzini, P., Eds.; Elsevier Applied Science.: Ams-
terdam, Netherlands, 1987.
15. Bill, J.; Aldinger, F. Progress in Materials Synthesis. Zeitschrift Fur Metall-
kunde. 1996, 87[11], 827.
16. Nourbakhsh, N. Anodic Alumina Films: Preparation, Characterization, and
Investigation of Reaction and Transport Properties. Ph.D. Dissertation,
University of Southern California, Los Angeles, CA, 1990.
17. Zheng, J; Carlson, W.B.; Reed, J.S. Dependence of Compaction Efficien-
cy in Dry Pressing on the Particle Size Distribution. Journal of America
Ceramic Society. 1995, 78(9), 2527.
18. Prochazka, S. Sintering of Silicon Carbide. Materials Science Research.
1975, 9, 421.
19. Ye, H.; Pujar, V.V.; Padture, N.P. Coarsening in Liquid-Phase-Sintered
Alpha-SiC. Acta. Materialia. 1999, 47[2], 481.
20. Moberlychan, W.J.; Cao, J.J.; Dejonghe, L.C. The Roles of Amorphous
Grain Boundaries and the β −α Transformation in Toughening SiC. Acta.
Materialia. 1998, 46[5], 1625.
21. Kim, Y.W.; Mitomo, M.; Emoto, H.; Lee, J.G. Effect of Initial α-Phase Con-
tent on Microstructure and Mechanical Properties of Sintered Silicon Car-
bide Using Carbon Black as the Sintering Aid. Journal of American C e-
ramics Society. 1998, 81[12], 3136.
41
22. Lee, J.K.; Kim, Y.J.; Kim, H. Formation of Self-Reinforced Microstructure
by the Control of Starting Phase in Liquid-Phase Sintered Silicon Carbide
Ceramics. Journal of Materials Science Letter. 1997, 16, 1958.
23. Korenaga, S.; Izumi, J.; Kani, A.; Matsumoto, S.; Koga, T. Corrosion Be-
havior of Sliding Materials in NH
3
-H
2
O Solutions at High-Temperature. Ce-
ramic Society of Japan. 1992, 100[12], 1458.
24. Soppe, W.J.; Janssen, G.J.M.; Bonekamp, B.C.; Correia, L.A.; V eringa,
H.J. A Computer Simulation Method for Sintering in Three-Dimensional
Powder Compacts. Journal of Materials Science. 1994, 29[3], 754.
25. Shiau, F.S.; Fang, T.T.; Leu, T.H. Effect of Particle-Size Distribution on the
Microstructural Evolution in the Intermediate Stage of Sintering. Journal of
American Ceramics Society. 1997, 80[2], 286.
26. Weiser, M.W.; Jonghe, L.C.D. Rearrangement During Sintering in Two-
Dimensional Arrays. Journal of American Ceramics Society . 1986, 69[11],
822.
27. Leu, J.J.; T. Hare, T.; Scattergood, R.O. A Computer Simulation Method
For Particle Sintering. Acta Metallurgica. 1988, 36[8], 1977.
28. Parhami, F.; McMeeking, R.M. A Network Model for Initial Stage Sintering.
Mechanics of Materials. 1998, 27[2], 111.
29. Ikegami, T. Early-stage Sintering in a Powder Compact of Polyhedral Par-
ticles I. Models. Ceramics International. 1999, 25, 415.
30. Arzt, E. The Influence of an Increasing Particle Coordination on the Densi-
fication of Spherical Powders. Acta Metallurgica. 1982, 30, 1883.
31. Missiaen, J.M.; Roure, S. A General Morphological Approach of Sintering
Kinetics: Application to WC -Co Solid Phase Sintering. Acta Materialia.
1998, 46[11], 3985.
32. Ross, J.W.; Miller, W.A.; Weatherly, G.C. Computer Simulation of Sinter-
ing in Powder Compacts. Acta Mettalurgica. 1982, 30, 203.
42
Chapter 3
Silicon Carbide Ceramic Membranes Prepared
By A Sol-Gel Technique
Abstract
Powders and supported thin silicon carbide films (membranes) were prepared by
a sol-gel technique using silica sol precursors as the source of silicon, and phe-
nolic resin as the source of carbon. The powders and films were prepared by the
carbothermal reduction reaction between the silica and the carbon source, which
was conducted at 1550
o
C under an inert atmosphere of argon. The powders
and films were characterized by XRD, and BET for their surface area and pore
size distribution. The SiC membranes were also characterized by permeation
tests using single gases (He and Ar) to determine their separation characteri s-
tics. The XRD analysis indicates that the powders and films consist of SiC, while
BET shows that they contain micropores. The membranes show ideal Knudsen
diffusivities, indicative that they do not have a significant number of cracks or
pinholes. SEM and AFM studies of the same films also validate this observation.
The powders and membranes were also subjected to treatment at different cor-
43
rosive and harsh environments, including immersing them in HF, burning in air,
and reacting with steam at high temperatures. The effects of these different
treatments on the internal surface area, pore size distribution, and transport
properties, were studied for both the powders and the membranes using the
aforementioned techniques and XPS. The materials are shown to have satisfac-
tory hydrothermal stability.
3.1 Introduction
Sol-gel processing techniques are receiving attention recently in the preparation
of ceramic membranes [1-5]. Ceramic membranes are better suited than the
conventional polymeric membranes for high temperature separation applications,
due to their good resistance to high temperatures [6-7]. Silicon carbide (SiC) is a
particularly promising material for the preparation of high temperature mem-
branes, since it is capable of withstanding high temperatures and mechanical
stresses, and it is resistant to highly corrosive environments.
Relatively few studies have reported on the preparation of SiC membranes by
any technique [8,9]. Morooka, et al. [8] prepared SiC membranes by chemical
vapor deposition on γ-alumina substrates at 700- 800
o
C using tri- isopropylsilane
as the precursor. After the deposition process, the membranes are heat -treated
in argon at 1000
o
C for 1 h. The resulting membranes are tested for the separ a-
44
tion of H
2
/H
2
O mixtures. The permeation experiments are performed at tempera-
tures between 200
o
C and 400
o
C, and have shown a H
2
/H
2
O selectivity in the
range of 3-5 [8]. Using γ-alumina as the support for the preparation of silicon
carbide membranes presents a number of challenges for the type of applications
these membranes may be useful for. Porous γ-alumina is not very resistant to
corrosive environments. In addition, its thermal expansion coefficient is different
from that of silicon carbide, raising concerns about the mechanical stability of the
resulting composite membrane system. Lee and Tsai [9] have prepared asy m-
metric SiC membranes by low -pressure chemical vapor deposition (LPCVD) of
SiH
4
, C
2
H
2
and argon mixtures at 800
o
C on the surface of Al
2
O
3
-doped SiC ma-
croporous supports. The macroporous SiC support, itself, had an asymmetric
structure, i.e., the alumina content decreased from the center to the surface of
the support. The CVD process reduced the pore size of the mem brane from 297
nm to 14 nm [9]. However, this pore size reduction is achieved at the cost of a
large reduction in the permeance of the membranes. Our group has devoted at-
tention to the preparation of SiC mesoporous and macroporous substrates for
use as supports in the preparation of asymmetric SiC membranes. We have
used a number of different starting powders and sintering aids to prepare crack -
free substrates, which have high porosity and are sufficiently smooth, to be used
for further processing by sol -gel, CVD and other techniques. Our study was r e-
ported in a recent publication by Suwanmethanond et al [10].
45
In the present chapter we describe the results of our investigations on the use of
these substrates for the preparation of asymmetric SiC membranes using a sol -
gel method. In our study we have utilized a number of commercial silica sols
with different sol particle sizes. These are then used as the silicon source for
production of SiC. Phenolic resin is utilized as the carbon source. The S iC sol
precursors are then coated on the SiC substrates, which are prepared by our
group [10] using dry-pressing techniques, and sintering at 1950
o
C. The SiC sol
precursors undergo carbothermal reduction at high temperatures to produce the
final SiC thin films. The sol -gel step is an important intermediate stage in the
preparation of asymmetric microporous SiC membranes (the final stage can be a
CVD or a polymeric precursor pyrolysis step), which are capable of withstanding
high service temperatures and are thermally stable to the presence of steam.
Mesoporous sol -gel membranes, in addition, can potentially find application in
nano- and ultrafiltration applications of corrosive liquid mixtures.
We are not aware of other studies, which report on the us e of a sol-gel technique
for the preparation of SiC membranes. A number of prior studies in recent years
have, however, focused on the preparation of SiC powders using sol -gel tech-
niques. Cerovic et al.,[11]
for example, have prepared silicon carbide powders by
heating a mixture of a silica sol, prepared by an ion exchange method, with sac-
charose or activated carbon (as the carbon sources), and boric acid; the latter is
reported to act as the catalyst for the carbothermal reduction of silica with carbon
46
at 1550
o
C. They report that the optimum molar ratio of C/Si for the production of
SiC is 3 for activated carbon and 4 for saccharose. The diameter of the spherical
SiC particles prepared with activated carbon was twice that of the particles pr e-
pared from saccharose. Using phenyltrimethoxysilane (PTMS), and/or tetraethy-
lorthosilicate (TEOS) as the silica source Seog and Kim [12] have prepared sil i-
con carbide powders, which are made of spherical particles. Mono- dispersed
spherical powders were produced using a base catalyzed route, while poly -
dispersed powders were obtained with the aid of an acid-base catalyzed route. A
number of different carbon sources such as ethycellulose, polyacrylonitrile
(PAN), and starch were utilized by Raman et al. [13] for reaction with a silicon
source at 1550
o
C to produce SiC powders. It was shown that the type of carbon
source utilized determines the crystalline form of SiC that is produced and the
grain size.
As noted previously, there are no published studies, we are aware of, that specif-
ically focus on preparing SiC membranes using a sol -gel technique. The prep a-
ration of such membranes presents unique challenges, since the thin gel film that
is generated on the porous substrate, typically using a dip- coating method, must
remain crack-free during the SiC forming process at high temperatures.
47
3.2 Experimental Procedure
3.2.1 The Sol-gel Synthesis of Silicon Carbide
A variety of silica sols (both water and alcohol based) have been utilized in the
investigation. For the materials, whose preparation is described here, a partic u-
lar organo-silica sol type IPAST, with a particle size in the range of 8- 10 nm (see
Figure 3.1) is used as the silica source, as received from the manufacturer (Ni s-
san Chemical Industries, Ltd). A phenolic resole type resin (Occidental Chem i-
cal) is used as the carbon source. A 50 wt % solution of phenolic resin in ethanol
is mixed with the silica sol in a quantity that is adjusted to obtain a molar ratio of
Si/C =1:3. The choice of this Si/C molar ratio is based on the stochiometric r e-
quirement for the completion of the overall carbothermal reduction reaction,
which is thought to proceed as follows: [14]
SiO
2
+ 3C SiC + 2CO (3.1)
The mixture of silica sol and phenolic resin is then sonicated using an ultrasonic
device for 10 min in order to obtain a homogeneous gel -polymer mixture. Sub-
sequently, the procedure one follows depends on whether one prepares powders
or membranes. The membrane preparation is described below. For the pow d-
ers, the sample is dried at room temperature overnight and then transferred into
48
an alumina crucible. The crucible containing the dried gel is then placed in a t u-
bular furnace. The samples are heated first up to 800
o
C at a heating rate of 75
o
C/hr and then up to 1550
o
C at a heating rate of 50
o
C/hr. The gel is calcined at
1550
o
C for 3 hrs; upon calcination it is then cooled down to room temperature at
a rate of 85
o
C/hr. During the calcination process the sample is constantly
purged with ultra-high purity argon.
3.2.2 Membrane Preparation Using the Sol-Gel Method
The supported SiC sol -gel membranes are prepared utilizing macroporous SiC
substrates. These substrates are prepared as follows [10]. A silicon carbide fine
powder (Sumitomo Osaka Cements, Japan) with an average particle size of
0.03 µ is mixed with a more course silicon carbide powder (Superior Graphite
Co.,) with an average particle size of 0.6 µ in the ratio of 1:2. Boron carbide (in
the amount of 0.1 wt %) and phenolic resin (in the amount 4 wt %) are used as
the sintering aids. The mixture of powders and various sintering aids are
pressed into disks, which are then calcined at 1950
o
C to obtain the macroporous
SiC membrane substrates (more details about the preparation of such s ubstrates
can be found in our prior publication) [10]. These SiC substrates are then used
for further surface modification via the sol-gel technique.
The silicon carbide membranes are prepared by conventional dip- coating tech-
niques. One first prepares the solution containing the appropriate amounts of the
49
silica and carbon sources and sonicates it for 10 min in order to obtain a hom o-
geneous blend. The macroporous SiC substrate to be dip- coated is wrapped
with Teflon tape on one side, and is dipped in the sol solution for a period 30 s; it
is then withdrawn out of the solution at a withdrawal rate of 0.001m/s. After the
coating process is completed, the coated substrate disk is placed in an alumina
crucible and is then inserted in the tubular furnace. The sample is heated to
1550
o
C with a heating rate of 50
o
C/hr, where it is calcined for 3 hr under an ar-
gon atmosphere; upon calcination it is cooled to room temperature with a cooling
rate of 50
o
C/hr. If additional coatings are required they are applied after the first
calcination step using the same dip-coating/calcination procedure.
3.2.3 The Characterization of SiC Membranes
The permeation characteristics of the membrane are measured using argon and
helium as the test gases. The transport properties of each membrane are r e-
ported in terms of the permeance of the two individual gases and the ideal sepa-
ration factor (defined as the ratio of permeances of these two individual gases).
The absolute values of permeance and the separation factor are both important
in terms of determining the usefulness of membrane for further applications and
processing. The permeances of the membranes were measured and separation
factor were calculate using the formula described in chapter 2.
50
For both the powders and the membranes X -ray diffraction (XRD) analysis is
used in order to identify the crystalline compounds that are present. The XRD
technique is complemented with X -ray photoelectron spectroscopy (XPS) mea-
surements, which determine the surface composition of the membranes. XPS
can potentially identify amorphous compounds, which may not be detected by
XRD. The pore size distributions of both powders and films are measured using
an ASAP2010 (Micromeritics) BET apparatus. For BET data analysis we utili ze
the BJH and Horvath- Kawazoe models. Therefore, the pore size distributions
(PSD) generated must only be viewed as qualitative measures of the pore struc-
ture (two SiC membranes with the same PSD’s are likely to have the same pore
structure characteristics) rather than as quantitative indicators of the pore size.
Scanning electron microscopy (SEM) and atomic force microscopy (AFM) are
also used to study the surface morphology of the membranes. The SEM tec h-
nique is used to locate the position of the film, and in order to determine the de-
gree of its adhesion to the underlying macroporous substrate. AFM is utilized to
determine the degree of surface roughness, which is important when determining
whether these materials are appropriate as substrates for the further deposition
of microporous films by other techniques.
The hydrothermal stability of the powders and the membranes is tested by treat-
ing them at 350
o
C
in flowing ultra high purity argon containing 50% mole of
steam for a pre-determined period of time. Upon completion of the hydrothermal
51
tests powders and membranes are tested again by a variety of techniques in or-
der to determine the effect of the treatment on surface and transport properties.
3.3 Results and Discussion
3.3.1 The SiC Powders
Sol-gel production of silicon carbide involves the reaction between silica and a
carbon source at relatively high temperatures. This calcination formation process
is also called the carbothermal reduction reaction. This is because during the
process silica is reduced (loses its oxygens) by a series of reactions, and is
eventually converted to silicon carbide. The carbothermal reaction and the for-
mation of silicon carbide can be easily traced using X -ray diffraction analys is.
The bottom line in Figure 3 .2, for example, shows the XRD pattern of a silicon
carbide powder prepared by us by the sol -gel method previously described and
after the calcination process. The XRD pattern indicates that the carbothermal
reduction reaction has resulted in a material, which is pure silicon carbide, its
peaks corresponding to that of crystalline β-SiC. The following set of reactions
have been reported to occur during the carbothermal reduction reaction [15-16]:
SiO
2
(s) + C (s) SiO (g) + CO (g) (3.2)
SiO
2
(s) + CO(g) SiO (g) + CO
2
(g) (3.3)
52
SiO (g) + 2C(s) SiC (s) + CO (g) (3.4)
SiO (g) + 3CO(g) SiC (s) + 2CO
2
(g) (3.5)
CO
2
(g) + C(s) 2CO(g) (3.6)
Here (s) and (g) refer to solid and gaseous states, respectively. The main reac-
tions for the formation of SiC are reactions (3.2) and (3.4). Reaction (3.2) occurs
rapidly at high temperatures; the SiO(g) that is produced reacts with C(s) that is
in close proximity to form SiC(s). When enough CO and CO
2
are present reac-
tions (3.3), (3.5), and (3.6) may also occur. It is clear from the above mechan-
ism that the miscibility of the carbon source (e.g., the phenolic resin solution) with
the silica sol plays an important role in the formation of the final SiC microstruc-
ture. The silica sol and phenolic resin need to be homogeneously and well mixed
in order for the SiO
2
and C to be in intimate contact and for crystalline SiC to be
produced.
Though β-SiC appears to be the form of SiC that forms immediately treatment of
such powders under various conditions may convert this form of SiC into other
forms. As can be seen in Figure 3.2 treatment of the powder in flowing air at 420
o
C for 3 hr leaves the β-SiC phase unchanged. Treatment in a strong hydrofluo-
ric acid (HF) solutio n (Aldrich, 40 wt% HF) for 5 hr also seems to have a minor
effect on the SiC powders. Treatment in a flowing mixture of 50- mol% water and
50-mol% argon at 350
o
C for 24 hr seems to convert part of the β-SiC into α-SiC
53
(see top of Figure 3.2). Bootsman et al. [17]
have
studied the phase transforma-
tion of different SiC phases. They concluded, in agreement with our own obser-
vations, that the 6H ( α-SiC) type is generally the most stable phase. However,
whether the other crystal polytypes (including β-SiC) convert into α-SiC depends
on the type and pressure of the atmosphere that prevails, and the type and
amount of dopants and impurities that are present in the sample.
To understand the nature of the materials that are formed by the sol -gel tech-
nique, and in order to evaluate their performance in corrosive and oxidative env i-
ronments the as prepared powder samples were subjected to further tests.
Three different types of treatments were applied in various sequences (see Table
3.1). They include heating of the silicon carbide samples in air at 420
o
C for vary-
ing periods of time (as indicated in Table 3.1) in order to burn away any un-
reacted carbon; washing in a strong HF solution (Aldrich, 40 wt% HF), which has
been reported by other investigators to etch away any residual silica that may be
left behind from the carbothermal reduction reaction [18]; and subjecting the
samples to a treatment at 350
o
C
in flowing ultra high purity argon containing
50% mole of steam for a pre- determined period of time, a hydrothermal stability
test which is important in terms of the eventual application of such membranes.
Table 3.1 shows a number of different sequences of tests performed on the sol -
gel SiC samples in order to study the effect of each step individually and in com-
bination. We have taken the internal surface area of the powder samples (as
54
measured by BET) as an indicator of the changes in the pore structure brought
upon by the various treatments.
Table 3.1 The surface area of SiC samples after undergoing various treatments.
It is clear from the results shown in Table 3.1 that etching with HF does not affect
the internal surface area of the SiC sample significantly. This is true for the as
received samples (less than a 0.6 % change) and the samples after they have
been subjected to a 3 hr oxidation treatment in air (less than a 3% change). This
result taken together with the XRD analysis results of the same sample, which
BET Surface Area
366.56 m2/g
Steam Treatment
for 24 hours
BET Surface Area
359.98 m2/g
Air Treatment
for 3 hours
BET Surface Area
221.95 m2/g
HF Treatment
BET Surface Area
323.98 m2/g
HF Treatment
BET Surface Area
360.65 m2/g
Steam Treatment
for 24 hours
BET Surface Area
333.78 m2/g
Air Treatment
for 3 hours
BET Surface Area
363.29 m2/g
Air Treatment
for 14 hours
BET Surface Area
362.90 m2/g
Air Treatment
for 7 hours
SiC Powder
BET Surface Area
220.65 m2/g
sol + C
calcined at 1550 oC
55
show no evidence for the presence of SiO
2
, confirms that no significant amounts
of residual silica are left in the samples after the carbothermal reduc tion reaction.
The fact that the sample that is oxidized in air before the HF treatment, shows a
slightly higher change in the surface area may be due to the fact that the HF is
removing the passive oxide layer that may form as the result of burning the s am-
ple in air.
The BET test results indicate that the surface area of the SiC samples increases
considerably after the air treatment at 420
o
C for 3 hr. This is likely to be due to
the burning away of some of the carbon left behind after the calcination process.
That little, if any, residual SiO
2
is present and that carbon is left behind after the
calcination process may be indicative that the other reactions (3.3, 3.5, and 3.6),
in addition to (3.2) and (3.4), participate in the formation of SiC. The excess ca r-
bon that is left behind may also be an indicator of some loss of volatile Si co m-
ponents (e.g., SiO), which are created at the beginning of the carbothermal r e-
duction process by solid- solid reactions at the interface between silica and car-
bon. Based on the proposed reaction mechanism, it is likely that the carbon m a-
trix formed by the pyrolysis of the phenolic resin, strongly influences the final sil i-
con carbide structure. Homogeneous mixing of the original precursors, as stated
earlier, is essential in order to assure that a final SiC sample is formed, which
has uniform properties including porosity and pore structure. The size and type
of the silica sols one uses determines the ability of homogeneously distributing
56
the sol with the carbon source. The finer the initial silica sol particles are, for ex-
ample, the more homogeneously they distribute in the phenolic resin, and activ e-
ly participate in the carbothermal reaction.
It should also be noted after a certain period of air treatment the surface area
does not change significantly (note, for example, the surface areas after the 7
and 14 hr treatments in air in column 3 of Table 3.1) indicative that most carbon
inclusions are removed in the first few hours of burning in air; further heating in
air does not appear to change the properties of the silicon carbide, and the sa m-
ple appears to be stable.
Steam treatment is equally effective with air oxidation in removing the residual
carbon. Note, furthermore, that after the carbon has been removed with air ox i-
dation steam treatment has little effect on the surface area (compare the two val-
ues at the bottom of columns 1 and 4 in Table 3.1, which are more or less the
same). In terms of the surface area but also of the pore size distribution (see
Figure 3.3) steam treatment does not affect the structure of the SiC powder after
the residual carbon is removed. On the other hand, as shown in Figure 3.2, the
steam treatment seems to convert some of the β-SiC into α-SiC; surface area
and pore size distribution appear not be sensitive indicators of this change, how-
ever. Interestingly, steam, HF, and air treatments all have little effect on the pore
size distribution of the final SiC powders (see Figure 3.3). Based on the above
57
tests, the SiC powders prepared by the sol-gel process all exhibit a good hydr o-
thermal stability, and corrosion and oxidation resistance.
3.3.2 The SiC Membranes
As previously noted the SiC membranes are prepared by dip- coating of sub-
strates made of mixtures of powders which are pressed into disks and then cal-
cined at high temperatures. These substrates are highly porous with a bimodal
pore size distribution with one sharp peak centered around 40 Å (corresponding
to the finer particle size powder) together with a much broader mac roporous
peak centered around 1500 Å corresponding to the larger size powder (see Fi g-
ure 3.4). The He/Ar separation factor for this substrate is close to unity and
He/Ar permeance is ~7.2*10
-7
m
3
/(m
2
.Pa.sec). The effect on Ar permeance and
the corresponding separation factor of coating a number SiC layers is shown in
Figures 3.5 and 3.6. The Ar permeance changes during the first coatings but
remains unchanged after that. There is no dependence of the permeance on the
transmembrane pressure gradient, which is also a good indicator of the lack of
any substantial convective flow contributions to the membrane transport. These
observations are also validated by the behavior of the separation factor with the
number of coatings (Figure 3.6). After the third coating the measured separation
factors are very close to the ideal Knudsen value of 3. AFM has also been used
to study the surface morphology of the sol-gel coated samples. Figure 3.7 shows
58
the AFM image of the SiC sol -gel membrane. There are no cracks and/or pin-
holes observed in the membrane layer, and the surface appears to be relatively
smooth. Figure 3.8 shows the XRD pattern of the SiC membrane together with
the XRD pattern of one of the powders prepared under similar conditions. The
two XRD patterns are indistinguishable with the peaks corresponding to β-SiC
and very small quantity of α-SiC phase.
The SiC membranes were also subjected to the same hydrothermal test the var i-
ous powders went through. The test, as previously, was carried out by exposing
the membranes at 350
o
C to a flowing mixture consisting of 50- mol% of water
and 50-mol% of argon for a period of 30 hr (the test was terminated after this p e-
riod because there was no noticeable change in the membrane properties). Dur-
ing the test the argon permeation rate of through the membrane was constantly
monitored, and t he results are shown in Figure 3.9. The argon permeance r e-
mained constant throughout the whole period the membrane was exposed to
steam. After the hydrothermal test the membrane separation characteristics
were also studied by measuring the permeation rate of He and Ar. No noticeable
changes were observed in the membrane transport characteristics, as c an be
seen in Figures. 3.5 and 3.6 which show the Ar permeance and the He/Ar s epa-
ration factor as a function of transmembrane pressure gradient for the membrane
before and after the hydrothermal test. Figure 3.10 shows an SEM picture of a
cross section of the SiC membrane after the hydrothermal stability test. The sol -
59
gel SiC thin film (~4-5 µm thick) lies on the top of the macroporous SiC disk. It
appears to be strongly adhering without any is visible cracks or pinholes develop-
ing after the hydrothermal stability test.
To further investigate the ability of the membranes to withstand the various cor-
rosive environments, the powder is analyzed by XPS. The XPS data are col-
lected by a Perkin-Elmer /Physical Electronics Division model 5100 X-ray photoe-
lectron spectrometer with a non-monochromatic Al K
α
1486.6 eV radiation source
(15 kV, 300 W). Data acquisition and instrument control is performed using an
RBD model 147 controller with Augerscan software. The powder sample, after
being subjected to the various treatments, are placed into the analysis chamber
and allowed to outgas until a vacuum of < 1 X 10
-7
torr had been restored (typical
analysis pressure was in the range of 1 - 8 X 10
-8
torr). Figure 3 .11 shows the
XPS results of a powder sample, which after being treated by HF for 5 hr, air at
420
o
C for 3 hr was subjected to a steam treatment. There is a strong peak at
100.4-101.0eV corresponding to SiC, which is indicative that the sample is silicon
carbide (deconvolution of the spectra indicates the SiC content to be in excess of
94%). A small side shoulder at 103.0-103.3eV corresponds to a trace impurity of
SiO
2
, which may be either a remnant of the original sol, or most likely due to the
surface oxidation of the SiC as a result of the air oxidation and steam treatments.
The XPS results seem to be consistent with the results of the XRD analysis.
60
3.4 Conclusions
Silicon carbide asymmetric membranes and powders have been prepared by a
sol-gel processing step followed by a carbothermal reduction reaction between
the silica and carbon precursors. The resulting materials consist mostly of SiC
and some residual carbon. This carbon is easily removed by an air or steam
treatment without any negative impact on the mechanical properties of the mem-
brane. Permeation studies of the SiC membranes show ideal separation factors
for single gases, which are close to the Knudsen values. This means that the
resulting films have no substantial fraction of macroporous cracks and pinholes.
They are, therefore, promising substrates for the preparation of permselective
microporous membranes by either CVD or polymeric precursor pyrolysis tec h-
niques or for ultrafiltration or nanofiltration applications involving corrosive liquids.
The membranes are resistant to treatment by HF, oxidation by air, and prolonged
exposure to steam at high temperatures.
61
Figure 3.1 The TEM picture of the organo-silica sol type IPAST with particle size
8-11 nm, provided by Nissan Chemical Industries, Ltd.
62
Figure 3.2 The XRD pattern of a SiC powder treated with HF, air, and steam;
*signifies the peaks corresponding to the (6H) phase.
63
Figure 3.3 The pore size distribution of the SiC powders after they have been
subjected to various treatments
64
Figure 3.4 The dV/dlogD of the SiC substrate utilized in the preparation of the
Sol-Gel membranes.
65
Figure 3.5 The argon permeance of the membrane as a function of the pressure
gradient and the number of coatings.
66
Figure 3.6 The separation factor of the membrane as a function of the pressure
gradient and the number of coatings.
67
Figure 3.7 AFM image of the SiC sol-gel membrane
68
Figure 3.8 The XRD patterns of the SiC membrane and the unsupported film
(powder) prepared by the same techniques.
69
Figure 3.9 The membrane argon permeance during exposure to steam.
70
Figure 3.10 SEM picture of the cross section of the SiC membrane prepared by
Sol-Gel technique
71
Figure 3.11 The XPS spectrum of the SiC powder sample; the superimposed
peaks correspond to SiC at 100.4-101.0 eV and SiO
2
at 103.0-103.3 eV, respectively.
72
Chapter 3 References
1. Puhlfurss, P.; Voigt, A.; Weber, R.; Morbe, M.; Microporous TiO2 Mem-
branes with a cut off < 500 Da. Journal of Membrane Science . 2000,
174[1], 123.
2. Wu, L.Q.; Huang, P.; Xu, N.P.; Shi, J.; Effects of Sol Properties and Calci-
nation on the Performance of Titanium Tubular Membranes. Journal of
Membrane Science. 2000, 173[2], 263.
3. De Vos, R.M.; Maier, W.F.; Verweij, H.; Hydrophobic silica membranes for
gas separation. Journal of Membrane Science. 1999, 158[1-2], 277.
4. Kim, J.; Lin, Y.S.; Sol-Gel synthesis and Characterization of Yitria-
Stabilized Zirconia Membranes. Journal of Membrane Science . 1998,
139, 75.
5. Delange, R.S.A.; Hekkink, J.H.A.; Keizer, K.; Burggraaf, A.J.; Microstruc-
tural Properties of Non-Supported Microporous. Journal of Non-
Crystalline Solids. 1996, 195, 203.
6. Kim, K.S.; Jung, I.H.; The permeability characteristics of non-porous
membrane by C
7
H
5
F
3
/SiH
4
plasma polymeric membrane. Korean Journal
of Chemical Engineering. 2000, 17[2] 149.
7. Drzal, P.L.; Halasa, A.F.; Kofinas, P.; Microstructure orientation and nano-
porous gas transport in semicrystalline block copolymer membranes Po-
lymer. 2000, 41[12] 4671.
8. Sea, B.K.; Ando, K.; Kusakabe, K.; Morooka, S.; Separation of Hydrogen
from Steam Using a SiC-Based Membrane Formed by Chemical Vapor
Deposition of Tri-isopropylsilane. Journal of Membrane Science, 1998,
146, 73.
9. Lee, L.; Tsai, D.S.; Silicon Carbide Membranes Modified by Chemical Va-
por Deposition Using Species of Low-Sticking Coefficients in a Si-
lane/Acetylene Reaction System. Journal of American Ceramic Society.
1998, 81[1] 159.
73
10. Suwanmethanond, V.; Goo, E.; Johnston, G.; Liu, P.; Sahimi, M.; Tsotsis,
T.T.; Porous SiC Sintered Substrates for High Temperature Membranes
for Gas Separations. Journal of Industrial & Engineering Chemistry R e-
search. 2000, 39[9], 3264.
11. Cerovic, L.J.; Milonjic, S.K.; Zec, S.P.; A Comparison of Sol-Gel Derived
Silicon Carbide Powders from Saccharose and Activated Carbon. Ceram-
ics International. 1995, 21, 271.
12. Seog I.S.; Kim, C.H.; Preparation of Monodispersed Spherical Silicon
Carbide by the Sol-Gel Method. Journal of Materials Science. 1993, 28,
3277.
13. Raman, V.; Bahl, O.P.; Dhawan, U.; Synthesis of Silicon Carbide through
the Sol-Gel Process from Different Precursors. Journal of Material
Science. 1995, 30, 2686.
14. Schubert, U.; Husing, N.; Synthesis of Inorganic Materials, WILEY -VCH,
Weinheim 2000, ch.2.
15. Weimer, A.W.; Nilsen, K.J.; G.A. Cochran, G.A.; Roach, R.P.; Kinetics of
Carbothermal Reduction of Beta Silicon Carbide. AIChE Journal. 1993,
39, 493.
16. Meng, G.W.; Zhang, L.D.; Mo, C.M.; Zhang, S.Y.; Qin, Y.; Feng, S.P.; Li,
H.J.; Synthesis of “A β-SiC Nanorod with SiO
2
Nanorod” One Dimensional
Composite Nanostructures. Solid State Communications . 1998, 106[4],
215.
17. Bootsma, G.A.; Knippenberg, W.F.; and Verspui, G.; Journal of Crystal
Growth. 1971, 8, 341.
18. Kraft, T.; Nickel, K.G.; Hydrothermal Degradation of Chemical Vapour De-
posited SiC Fibres. Journal of Materials Science, 1998, 33[17], 4357.
74
Bibliography
Arzt, E. The Influence of an Increasing Particle Coordination on the Densification
of Spherical Powders. Acta Metallurgica. 1982, 30, 1883.
Bill, J.; Aldinger, F. Progress in Materials Synthesis. Zeitschrift Fur Metallkunde .
1996, 87[11], 827.
Bootsma, G.A.; Knippenberg, W.F.; and Verspui, G.; Journal of Crystal Growth.
1971, 8, 341.
Cerovic, L.J.; Milonjic, S.K.; Zec, S.P.; A Comparison of Sol-Gel Derived Silicon
Carbide Powders from Saccharose and Activated Carbon. Ceramics Internation-
al. 1995, 21, 271.
Ciora, R. J.; Fayyaz, B.; Liu, P. K. T.; Suwanmethanond, V.; Mallada, R .; Sahimi,
M.; Tsotsis, T. T., Preparation and reactive applications of nanoporous silicon
carbide membranes. Chemical Engineering Science, 2004, 59, 4957.
Coppola, J.A.; Hailey, L.N.; McMurtry, C.H. Sintered Alpha Silicon Carbide Ce-
ramic Body Having Equiaxed Microstructure. US. Patent 4,179,299,1979.
De Vos, R .M.; Maier, W.F.; Verweij, H.; Hydrophobic silica membranes for gas
separation. Journal of Membrane Science. 1999, 158[1-2], 277.
Delange, R.S.A.; Hekkink, J.H.A.; Keizer, K.; Burggraaf, A.J.; Microstructural
Properties of Non-Supported Microporous. Journal of Non-Crystalline Solids .
1996, 195, 203.
Drzal, P.L.; Halasa, A.F.; Kofinas, P.; Microstructure orientation and nanoporous
gas transport in semicrystalline block copolymer membranes Polymer. 2000,
41[12] 4671.
Elyassi, B.; Sahimi, M.; Tsotsis, T. T, Silicon carbide membranes for gas separa-
tion applications. Journal of Membrane Science, 2007, 288, 290.
Elyassi, B.; Sahimi, M.; Tsotsis, T. T, Silicon carbide membranes for gas separa-
tion applications. Journal of Membrane Science, 2008, 316, 73.
75
Grande, T.; Sommerset, H.; Hagan, E.; Wiik, K.; Einarsrud, M.A. Effect of Weight
Loss on liquid-Phase-Sintered Silicon Carbide. Journal of American Ceramics
Society. 1997, 80[4], 1047.
Hirata, Y.; Hidaka, K.; Matsumura, H.; Fukushige, Y.; Sameshima, S. Colloidal
Processing and Mechanical Properties of Silicon Carbide with Alumina. Journal
of Materials Research. 1997, 12[11], 3146.
Hojo, J. Sintering Behavior of Ultrafine Silicon Carbide Powder. In Silicon Car-
bide Ceramics-1; Somiya, S., Inomata, Y., Eds., Elsevie r Applied Science,: New
York, NY, 1991.
Ikegami, T. Early-stage Sintering in a Powder Compact of Polyhedral Particles I.
Models. Ceramics International. 1999, 25, 415.
Kim, J.; Lin, Y.S.; Sol-Gel synthesis and Characterization of Yitria-Stabilized Zir-
conia Membranes. Journal of Membrane Science. 1998, 139, 75.
Kim, K.S.; Jung, I.H.; The permeability characteristics of non-porous membrane
by C
7
H
5
F
3
/SiH
4
plasma polymeric membrane. Korean Journal of Chemical Eng i-
neering. 2000, 17[2] 149.
Kim, Y.W.; Mitomo, M.; Emoto, H.; Lee, J.G. Effect of Initial α-Phase Content on
Microstructure and Mechanical Properties of Sintered Silicon Carbide Using Car-
bon Black as the Sintering Aid. Journal of American Ceramics Society . 1998,
81[12], 3136.
Korenaga, S.; Izumi, J.; Kani, A.; Matsumoto, S.; Koga, T. Corrosion Behavior of
Sliding Materials in NH
3
-H
2
O Solutions at High-Temperature. Ceramic Society of
Japan. 1992, 100[12], 1458.
Kraft, T.; Nickel, K.G.; Hydrothermal Degradation of Chemical Vapour Deposited
SiC Fibres. Journal of Materials Science, 1998, 33[17], 4357.
Lange, F.F. Hot-Pressing Behavior of Silicon Carbide Powders with Additions of
Alumina Oxide. Journal of Materials Science. 1975, 10, 314.
Lee, J.K.; Kim, Y.J.; Kim, H. Formation of Self-Reinforced Microstructure by the
Control of Starting Phase in Liquid-Phase Sintered Silicon Carbide Ceramics.
Journal of Materials Science Letter. 1997, 16, 1958.
Lee, L.; Tsai, D.S. Silicon Carbide Membranes Modified by Chemical Vapor De-
position Using Species of Low-Sticking Coefficients in a Silane/Acetylene Reac-
tion System. Journal of American Ceramics Society. 1998, 81[1], 159.
76
Leu, J.J.; T. Hare, T.; Scattergood, R.O. A Computer Simulation Method For Par-
ticle Sintering. Acta Metallurgica. 1988, 36[8], 1977.
Meng, G.W.; Zhang, L.D.; Mo, C.M.; Zhang, S.Y.; Qin, Y.; Feng, S.P.; Li, H.J.;
Missiaen, J.M.; Roure, S. A General Morphological Approach of Sintering Kinet-
ics: Application to WC -Co Solid Phase Sintering. Acta Materialia. 1998, 46[11],
3985.
Moberlychan, W.J.; Cao, J.J.; Dejonghe, L.C. The Roles of Amorphous Grain
Boundaries and the β− α Transformation in Toughening SiC. Acta. Materialia .
1998, 46[5], 1625.
Mourhatch, R., Tsotsis, T. T.; Sahimi, M., Network model for the evolution of the
pore structure of silicon-carbide membranes during their fabrication. Journal of
Membrane Science, 2010, 356, 138.
Nourbakhsh, N. Anodic Alumina Films: Preparation, Characterization, and Inves-
tigation of Reaction and Transport Properties. Ph.D. Dissertation, University of
Southern California, Los Angeles, CA, 1990.
Owens, D.P.; Ruppel, I.B. Porous Silicon Carbide. US. Patent 5,441,799,1995.
Pages, X.; Rouessac, V.; Cot, D.; Nabias, G.; Durand, J., Gas permeation of
PECVD membranes inside alumina substrate tubes. Separation and Purification
Technology, 2001, 25, 399.
Parhami, F.; McMeeking, R.M. A Network Model for Initial Stage Sintering. Me-
chanics of Materials. 1998, 27[2], 111.
Prochazka, S. Sintered Dense Silicon Carbide. US. Patent 4,004,934,1977.
Prochazka, S. Sintering of Silicon Carbide. In Ceramics for High Performance
Applications; Burke, J.J., Gorum, A.E., Katz, R.N., Eds.; Published in cooper a-
tion with the Metals and Ceramics Information Center, Columbus,: OH, 1973.
Prochazka, S. Sintering of Silicon Carbide. Materials Science Research. 1975, 9,
421.
Puhlfurss, P.; Vo igt, A.; Weber, R.; Morbe, M.; Microporous TiO2 Membranes
with a cut off < 500 Da. Journal of Membrane Science. 2000, 174[1], 123.
Rajabbeigi, N.; Elyassi, B .; Tsotsis, T. T.; Sahimi, M., Molecular pore-network
model for nanoporous materials, I: Application to adsorption in silicon-carbide
membranes. Journal of Membrane Science, 2009, 335, 5.
77
Raman, V.; Bahl, O.P.; Dhawan, U.; Synthesis of Silicon Carbide through the
Sol-Gel Process from Different Precursors. Journal of Material Science. 1995,
30, 2686.
Ross, J.W.; Miller, W.A.; Weatherly, G.C. Computer Simulation of Sintering in
Powder Compacts. Acta Mettalurgica. 1982, 30, 203.
Sasaki, M.; Suzuki, K.; Nishimura, H.; Noshiro, M. Microstructure of Pressureless
Sintered Silicon Carbide with Additions of Aluminum Oxide. In High Tech Ceram-
ics. Vincenzini, P., Eds.; Elsevier Applied Science.: Amsterdam, Netherlands,
1987.
Schubert, U.; Husing, N.; Synthesis of Inorganic Materials, WILEY -VCH, Wein-
heim 2000, ch.2.
Sea, B.K.; Ando, K; Kusakabe, K; Morooka, S. Separation of Hydrogen From
Steam Using a SiC-Based Membrane Formed by Chemical Vapor Deposition of
Trisopropylsilane. Journal of Membrane Science. 1998, 146[1], 73.
Seog I.S.; Kim, C.H.; Preparation of Monodispersed Spherical Silicon Carbide by
the Sol-Gel Method. Journal of Materials Science. 1993, 28, 3277.
Shiau, F.S.; Fang, T.T.; Leu, T.H. Effect of Particle-Size Distribution on the Mi-
crostructural Evolution in the Intermediate Stage of Sintering. Journal of Amer i-
can Ceramics Society. 1997, 80[2], 286.
Soppe, W.J.; Janssen, G.J.M.; Bonekamp, B.C.; Correia, L.A.; Veringa, H.J. A
Computer Simulation Method for Sintering in Three-Dimensional Powder Com-
pacts. Journal of Materials Science. 1994, 29[3], 754.
Suwanmethanond, V.; Goo, E.; Johnston, G.; Liu, P.; Sahimi, M.; Tsotsis, T.T.;
Porous SiC Sintered Substrates for High Temperature Membranes for Gas Sepa-
rations. Journal of Industrial & Engineering Chemistry Research. 2000, 39[9],
3264.
Suzuki, K. Pressureless-Sintered Silicon Carbide with Addition of Aluminum
Oxide. In Silicon Carbide Ceramics -2; Somiya, S., Inomata, Y., Eds.; Elsevier
Applied Science.: New York, NY, 1991.
Synthesis of “A β-SiC Nanorod with SiO
2
Nanorod” One Dimensional Composite
Nanostructures. Solid State Communications. 1998, 106[4], 215.
78
Takeda, Y.; Shibata, N.; Kubo, Y., SiC coating on porous γ-Al
2
O
3
using alterna-
tive supply CVI method. Journal of the Ceramical Society of Japan, 2001, 109,
305.
Tanaka, H. Sintering of Silicon Carbide. In Silicon Carbide Ceramics-1; Somiya,
S., Inomata, Y., Eds.; Elsevier Applied Science, New York, NY, 1991.
Weimer, A.W.; Nilsen, K.J.; G.A. Cochran, G.A.; Roach, R.P.; Kinetics of Carbo-
thermal Reduction of Beta Silicon Carbide. AIChE Journal. 1993, 39, 493.
Weiser, M.W.; Jonghe, L.C.D. Rearrangement During Sintering in Two-
Dimensional Arrays. Journal of American Ceramics Society. 1986, 69[11], 822.
Wu, L.Q.; Huang, P.; Xu, N.P.; Shi, J.; Effects of Sol Properties and Calcination
on the Performance of Titanium Tubular Membranes. Journal of Membrane
Science. 2000, 173[2], 263.
Ye, H.; Pujar, V.V.; Padture, N.P. Coarsening in Liquid-Phase-Sintered Alpha-
SiC. Acta. Materialia. 1999, 47[2], 481.
Ye, H.; Pujar, V.V.; Padture, N.P. Coarsening Mechanism in Liquid-Phase-
Sintered SiC. American Ceramic Society Annual Meeting, Ceramograhic Exhibit,
1997.
Zheng, J; Carlson, W.B.; Reed, J.S. Dependence of Compaction Efficiency in Dry
Pressing on the Particle Size Distribution. Journal of America Ceramic Society .
1995, 78(9), 2527.
Zhou, Y.; Fukushima, M.; Miyazaki, H.; Yoshizawa, Y.; Hirao, K.; Iwamoto, Y.;
Sato, K., Preparation and characterization of tubular porous silicon carbide
membrane supports. Journal of Membrane Science, 2011, 369, 112.
Abstract (if available)
Abstract
This dissertation focuses on the preparation of silicon carbide (SiC) ceramic membranes on SiC substrates. An original technique of SiC porous substrate preparation using sintering methods was developed during the work for the completion of the dissertation. The resulting SiC substrates have demonstrated high porosity, high internal surface area, well interconnected surface pore network and, at the same time, good thermal, chemical and mechanical stability. In a further development, sol-gel techniques were used to deposit micro-porous SiC membranes on these SiC porous substrates. The SiC membranes were characterized by a variety of techniques: ideal gas selectivity (He and N2), XRD, BET, SEM, XPS, and AFM. The characterization results confirmed that the asymmetric sol-gel SiC membranes were of high quality, with no cracks or pinholes, and exhibiting high resistance to corrosion and high hydro-thermal stability. In conclusion, the SiC ceramic membrane work was successfully completed. Two publications in international peer reviewed journals resulted out of this work.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Fabrication of nanoporous silicon carbide membranes for gas separation applications
PDF
A process-based molecular model of nano-porous silicon carbide membranes
PDF
Experimental studies and computer simulation of the preparation of nanoporous silicon-carbide membranes by chemical-vapor infiltration/chemical-vapor deposition techniques
PDF
Fabrication of silicon carbide sintered supports and silicon carbide membranes
PDF
Continuum and pore netwok modeling of preparation of silicon-carbide membranes by chemical-vapor deposition and chemical-vapor infiltration
PDF
Molecular modeling of silicon carbide nanoporous membranes and transport and adsorption of gaseous mixtures therein
PDF
Hydrogen storage in carbon and silicon carbide nanotubes
PDF
Thermal properties of silicon carbide and combustion mechanisms of aluminum nanoparticle
PDF
Fabrication of nanoporous silicon oxycarbide materials via a sacrificial template technique
PDF
Exploring properties of silicon-carbide nanotubes and their composites with polymers
PDF
The use of carbon molecule sieve and Pd membranes for conventional and reactive applications
PDF
Biogas reforming: conventional and reactive separation processes and the preparation and characterization of related materials
PDF
Fabrication of silicon-based membranes via vapor-phase deposition and pyrolysis of organosilicon polymers
PDF
In situ studies of the thermal evolution of the structure and sorption properties of Mg-Al-CO3 layered double hydroxide
PDF
Preparation of polyetherimide nanoparticles by electrospray drying, and their use in the preparation of mixed-matrix carbon molecular-sieve (CMS) membranes
PDF
Methanol synthesis in a membrane reactor
PDF
A flow-through membrane reactor for destruction of a chemical warfare simulant
PDF
Development of carbon molecular-sieve membranes with tunable properties: modification of the pore size and surface affinity
PDF
Studies of transport phenomena in hydrotalcite membranes, and their use in direct methanol fuel cells
PDF
Silica sol-gel thin film coatings for integrated photonic devices
Asset Metadata
Creator
Suwanmethanond, Varaporn
(author)
Core Title
Silicon carbide ceramic membranes
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Chemical Engineering
Publication Date
05/21/2014
Defense Date
05/21/2012
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
ceramic membranes,gas separation,hydro-thermal atability,OAI-PMH Harvest,SiC,SiC membranes,silicon carbide,sintering,sol-gel
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Goo, Edward K. (
committee member
), Sahimi, Muhammad (
committee member
), Tsotsis, Theodore T. (
committee member
)
Creator Email
jenny_suwan@yahoo.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-42243
Unique identifier
UC11289333
Identifier
usctheses-c3-42243 (legacy record id)
Legacy Identifier
etd-Suwanmetha-857.pdf
Dmrecord
42243
Document Type
Dissertation
Rights
Suwanmethanond, Varaporn
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
ceramic membranes
gas separation
hydro-thermal atability
SiC
SiC membranes
silicon carbide
sintering
sol-gel