Journal of Membrane Science 275 (2006) 1728

Effects of novel silane modification of zeolite surface on polymer chain

rigidification and partial pore blockage in polyethersulfone (PES)zeolite
A mixed matrix membranes
Yi Li a , Huai-Min Guan a , Tai-Shung Chung a, , Santi Kulprathipanja b
a

Department of Chemical and Biomolecular Engineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260, Singapore
b UOP LLC, 50 East Algonquin Road, Des Plaines, IL 60017-5016, USA
Received 1 July 2005; received in revised form 28 August 2005; accepted 31 August 2005
Available online 30 September 2005

Abstract
A novel silane coupling agent, (3-aminopropyl)-diethoxymethyl silane (APDEMS) was used in this work to modify zeolite surface for mixed
matrix membranes (MMMs). Both elementary analysis and XPS spectra confirm the chemical modification, while BET measurements show no
changes in zeolite surface area and total pore volume after the modification. Polyethersulfone (PES)zeolite 3A, 4A and 5A MMMs were fabricated
at high processing temperatures using unmodified and chemical modified zeolite. SEM images of these MMMs indicate the interface between
polymer and zeolite phases becomes better if modified zeolite is used. The effects of chemical modification of zeolite surface and zeolite loadings
on the gas separation performance of these MMMs were investigated. Both permeability and selectivity of MMMs made from APDEMS modified
zeolite are higher than those of MMMs made from unmodified zeolite at 20 wt% zeolite loading because of a decrease in the degree of partial pore
blockage of zeolites. The permeability of all studied gases decreases with increasing zeolite content for PESzeolite 4A-NH2 MMMs, while for
PESzeolite 5A-NH2 MMMs, the gas permeability decreases and then increases with an increase in zeolite loadings. This unique phenomenon
implies that using large pore-size zeolite for MMMs would potentially offset the negative effects of partial pore blockage and polymer chain
rigidification on permeability. A modified Maxwell model with adjusted parameters was applied to study the PESzeolite 4A-NH2 MMM system.
The predicted permeability and selectivity show very good agreement with experimental data, indicating the modified Maxwell model is fully
capable of predicting the gas separation performance of MMMs made from both unmodified and modified zeolite.
2005 Elsevier B.V. All rights reserved.
Keywords: Mixed matrix membrane; Silane modification of zeolite surface; Polymer chain rigidification; Partial pore blockage; Maxwell model

1. Introduction
During the last 2 decades, polymer-based organicinorganic
composites have received world-wide attention in the field of
material science. This is due to the fact that the resultant materials may offer superior performance in terms of mechanical
toughness for engineering resins, permeability and selectivity
for gas/liquid separation, and photoconductivity for electronics
[13]. This new concept has also been applied to the membranebased gas/liquid separation by combining the easy processability
of organic polymers with the excellent separation properties
of inorganic materials [2,410]. However, it has been found

Corresponding author. Tel.: +65 6874 6645; fax: +65 6779 1936.
E-mail address: chencts@nus.edu.sg (T.-S. Chung).

0376-7388/$ see front matter 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.memsci.2005.08.015

that there is an obstacle to the successful introduction of inorganic molecular sieve materials into an organic polymer matrix
because of the poor compatibility between molecular sieve and
polymer matrix, especially for the case of rigid glassy polymer
materials [913].
Therefore, some methods have been proposed to improve
the interfacial strength in order to enhance the separation
performance [1318]. One of them is the chemical modification of the surface of molecular sieve with silane coupling agents, such as (-aminopropyl)-triethoxy silane, N-(aminoethyl)--aminopropyltrimethoxy silane, (-glycidyloxypropyl)-trimethoxy silane and (3-aminopropyl)-dimethylethoxy
silane [1417]. By means of FTIR and 29 Si MAS NMR, Matsumoto et al. have identified that there are three types of silanol
groups on the surface of zeolite [19], i.e., single ((SiO)3 SiOH);
hydrogen-bonded ((SiO)3 SiOH OHSi(SiO)3 ) and geminal

18

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

((SiO)2 Si(OH)2 ). Among these silanol groups, free SiOH groups

are the most reactive groups and may provide the sites for the
physical and chemical adsorption of silane coupling agents. With
appropriate silane coupling agents, one may not only modify
surface properties of zeolite from hydrophilic to hydrophobic,
but also increase zeolite affinity to the functional groups of the
polymer matrix as illustrated in references [2022].
Duval et al. [16] demonstrated that the adhesion between
zeolite and polymer matrix phases could be improved by modifying zeolite surface with some silane coupling agents. However,
there was no significant improvement on permselectivity though
SEM micrographs indicated good coupling between silane and
zeolite. Mahajan and Koros [14,23] also reported similar phenomena. Although the improvement on interface was clearly
observed by SEM, their mixed matrix membranes (MMMs)
made of polymer-modified zeolite exhibited worse performance
(i.e., decreases in both permeability and selectivity) compared
with those made of polymer-unmodified zeolite. This may be due
to the fact that the addition of silane coupling agents may reduce
the voids between unmodified zeolite and polymer phases, but
cannot eliminated them completely. The size of these voids may
be in the range of nanometer or sub-nanometer [14,15,23], which
is still larger than the sizes of gas molecules. As a consequence,
the addition of silane coupling agents on zeolite surface may help
reduce voids and result in an increase in gas transport resistance
across the membrane, but there is almost no much improvement
on permselectivity. Unless a proper silane coupling agent is chosen, it is very difficult to completely prevent the gas penetrants
from going through these voids.
Therefore, the first purpose of this study is to introduce a
novel silane coupling agent, (3-aminopropyl)-diethoxymethyl
silane (APDEMS), on zeolite surface and to investigate its effects
on the separation performance of MMMs. To our best knowledge, so far there is no academic literature available on using
APDEMS to modify the zeolite surface for MMMs. Zeolites
with different pore sizes are used in order to identify if the
effectiveness of silane modification is pore-size dependent. In
addition, although the interphase occupies an extremely small
volume fraction (i.e., less than 1010 %), it appears to have a significant effect on the separation performance of mixed matrix
membranes [10,1315,2325]. The origins of the imperfect
interphase are complicated. Poor compatibility between molecular sieve and polymer matrix, uneven shrinkages and stresses of
these two components during the membrane formation may be
some of the possible causes. As a result, polymer chain rigidification and partial pore blockage of zeolites have been hypothesized
and partially confirmed from the lower gas permeation data of
mixed matrix membranes [10,13,15,2325]. Therefore, the second purpose of this paper is to investigate if and how the novel
silane coupling agent affects polymer chain rigidification and
partial pore blockage of zeolites and to study if we can reduce
the nanometric interphase.
For easy comparison with our previous work [25], zeolite 3A, 4A and 5A were used as the dispersed phase in this
work, and polyethersulfone (PES) was chosen as the continuous polymer matrix due to its appropriate Tg of 215 C and
various applications in gas separation [26]. Zeolite and its

surface were characterized by elementary analysis, XPS and

BrunaerEmmettTeller (BET). A high processing temperature
was employed in order to maintain the flexibility of polymer
chains during membrane formation [1315,23,25]. The gas permeation rates of the PESzeolite A MMMs were measured as a
function of zeolite loading and zeolite pore sizes. SEM and DSC
were utilized to study the morphology and Tg of these developed
MMMs, respectively.
2. Experimental
2.1. Materials
A commercial Radel A-300 PES, purchased from Amoco
Performance Products Inc., OH, USA, was selected as the continuous phase. It has a weight-average molecular weight of about
15,000. N-Methyl-2-pyrrolidone (NMP) was purchased from
Merck; toluene was from Fisher Scientific UK Ltd.; methanol
was from J.T. Baker. NMP and toluene were dried using the
activated molecular sieve 4A beads with a diameter of 35 mm
supplied by Research Chemicals Ltd., and then filtered through
a 0.2 m Teflon filter before use. The silane coupling agent,
APDEMS, was from SigmaAldrich and used without further
purification. The molecular sieves involved were zeolite 4A supplied by UOP LLC, Des Plaines, IL, USA, zeolite 3A and 5A
resulted from K+ and Ca2+ exchange treatment of zeolite 4A,
respectively. The ion exchange treatment was done by Dr. Huang
in our research group based on some related literatures [2729].
Their particle sizes vary from 1 to 2 m tested by SEM. In order
to remove the adsorbed water vapor or other organic vapors,
all modified zeolites were dehydrated at 110 C for 2 h under
vacuum before use.
2.2. Chemical modication method of zeolite surface
Fig. 1 shows the flowchart of chemical modification used in
this study. This procedure was derived with modifications from
the Plueddemanns method [20]. It consists of three steps: (1)
stir the mixture of toluene (500 ml) + APDEMS (20 ml) + zeolite
(2.5 g) for 24 h at room temperature under N2 ; (2) filter and wash
the zeolite with 500 ml toluene and then 250 ml methanol to
remove the unreacted silane, respectively; (3) dry the modified
zeolite for 1 h at 110 C under vacuum.
2.3. Membrane preparation procedure
Because the fabrication method developed in our previous
paper achieves good properties for mixed matrix membranes
[25], the same method was employed in this work. We applied
high processing temperatures during membrane formation to
eliminate the void between polymer and zeolite phases. The
resultant dried MMMs have thicknesses varying from 60 to
90 m for permeability measurements.
2.4. Gas permeability measurement
The gas permeation properties of every mixed matrix membrane were measured using variable-pressure constant-volume

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

19

Fig. 1. Flowchart of the chemical modification of zeolite surface.

method with a pre-calibrated permeation cell described elsewhere [30]. Each pure gas was tested in the sequence of He, H2 ,
O2 , N2 , CH4 and CO2 , and measured three times for each membrane. The measurement of H2 gas for MMMs was performed
at 35 C and 3.546375 105 Pa (3.5 atm); the measurement
of other gases was conducted at 35 C and 10.1325 105 Pa
(10 atm).

Table 1
Comparison of elementary analysis data of zeolites before and after the chemical
modification

2.5. Other characterizations

Two analysis methods were used to determine whether
the chemical modification had taken place. The first one
was elementary analysis carried out on a Perkin-Elmer PE
2400 Series II CHNS/O analyzer. The second analysis method
is X-ray photoelectron spectroscopy (XPS). XPS data were
obtained on a Shimadzu ESCAKL spectrometer using a nonmonochromatic Mg K photon source. Subsequently, nitrogen
adsorptiondesorption measurements were carried out at 77 K
on a Quantachrome Autosorb-1 apparatus to compare the BET
surface area and total pore volume between the unmodified and
modified zeolites.
The glass transition temperature (Tg ) of MMMs was analyzed with differential scanning calorimetry (DSC) on a MettlerToledo DSC 822e. Tg of the sample was determined as the
midpoint temperature of the transition region in the second heating cycle [31,32]. Tg provides a qualitative estimation of the
flexibility of polymer chains. It would be very useful to compare the rigidity of polymer chains in MMMs at different zeolite
types and loadings with that of neat polymer membrane [25]. The
electron micrographs of MMM morphology were examined by
SEM on a JEOL JSM-5600LV and JSM-6700F to determine if
the zeolite particles were dispersed homogenously and if there
existed voids between polymer and zeolite phases.
3. Results and discussion
3.1. Characterization and comparison of unmodied and
modied zeolites
In order to determine if the chemical modification of zeolite
surface is actually successful, Table 1 lists the results of ele-

Sample name

N (%)

C (%)

H (%)

Zeolite 3 A
Zeolite 3A-NH2 a

0.0
2.47

0.16
7.42

2.16
2.44

Zeolite 4A
Zeolite 4A-NH2 a

0.0
2.42

0.16
7.30

2.12
2.30

Zeolite 5A
Zeolite 5A-NH2 a

0.0
2.43

0.15
7.28

2.10
2.28

NH2 means the zeolite after the modification with APDEMS.

mentary analysis before and after the chemical modification.

Zeolite 3A, 4A and 5A mean the zeolite before the chemical modification, while zeolite 3A-NH2 , 4A-NH2 and 5A-NH2
denote the zeolite after the chemical modification. Results show
that more nitrogen and carbon elements are detected after the
modification, thus confirming that the silane coupling agent has
been attached to the zeolite surface. XPS spectra of zeolite 3A
and 4A before and after the chemical modification are exhibited in Fig. 2 . After modification, a peak clearly appears at
the electronic binding energy of about 397.2 eV that attributes
to the nitrogen atom of the silane coupling agent containing
amino group. This indicates that some molecules of the silane
coupling agent have been grafted onto the external surface of
zeolite, which is in good agreement with the elementary analysis
results.
That APDEMS is selected as the silane coupling agent
because it has two ethoxy groups (CH3 CH2 O) and an additional hydrophobic group (CH3 ). Compared to 3-aminopropyl
triethoxy silane with three ethoxy groups, APDEMS may lower
the number of coupling points with zeolite surface during the
silanization reaction and thus avoid blocking the micropore of
zeolites. Table 2 lists the results of the multipoint BET surface
area and total pore volume of zeolites before and after the chemical modification. There are no sharp changes in the surface area
and total pore volume of zeolites after the modification. These
results imply that the addition of APDEMS does not alter the
micropore of zeolites.

20

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

Fig. 2. Comparison of XPS spectra of the zeolite before and after the chemical modification (A: zeolite 3A; B: zeolite 3A-NH2 ; C: zeolite 4A; D: zeolite 4A-NH2 ).
Table 2
Comparison of total pore volume and multipoint BET surface area of zeolites before and after the chemical modification
Sample name

Zeolite 3 A
Zeolite 4A
Zeolite 5A

Total pore volume (cm3 /g)

Multipoint BET surface area (m2 /g)

Before modifying

After modifying

Before modifying

After modifying

0.197
0.200
0.198

0.196
0.198
0.196

557.6
567.0
561.1

553.5
561.6
556.7

3.2. Effect of chemical modication of zeolite surface on

gas separation performance
PESzeolite 3A, PESzeolite 4A and PESzeolite 5A
MMMs before and after the chemical modification of zeolite
surface were fabricated with a zeolite loading of 20 wt%. Fig. 3
shows their permeability for four fast gases, while Fig. 4 summarizes their selectivity for four gas pairs. Interestingly, different from the previous studies [14,16], the permeability of
MMMs with modified zeolite is higher than those MMMs with
unmodified zeolite. This increasing trend in permeability seems
counter-intuitive at first because the contact between polymer
and zeolite phases becomes better as illustrated in Fig. 5 and
the leakage through the interface becomes impossible after the
modification.
Perhaps we can explain this increasing trend from an opposite perspective. Two possible mechanisms have been proposed
in the previous work [2325] in order to explain the significant
decrease in gas permeability for MMMs; namely, the inhibition
of polymer chain mobility near the polymerzeolite interface
and the partial pore blockage of zeolites induced by polymer
chains. The first mechanism can be easily detected by the Tg
shift. Table 3 shows a comparison of Tg values of PESzeolite A
MMMs made of modified and unmodified zeolite. Their Tg values are almost the same, indicating the degree of polymer chain

rigidification for these MMMs are almost the same. Therefore,

we may be able to rule out a decrease in polymer chain rigidification as a possible cause for the permeability increase.
Previous researchers have measured the chain length
of (3-aminopropyl)-triethoxy silane molecules of around
[33,34] by means of spectroscopic
(59) 1010 m (59 A)
ellipsometry. Because APDEMS and (3-aminopropyl)-triethoxy
silane have similar chemical structures, APDEMS may have
a molecular chain length of approximately (59) 1010 m
Table 3
Change of glass transition temperatures of MMMs over neat PES dense film
before and after the chemical modification of zeolite surface
Membrane name

Glass transition temperature ( C)

Before modifying

After modifying

215 1

Neat PES dense film

PESzeolite 4A (20 wt% loading)
PESzeolite 4A (30 wt% loading)
PESzeolite 4A (40 wt% loading)
PESzeolite 4A (50 wt% loading)

216
217
218
219

1
1
1
1

216
217
218
219

1
1
1
1

PESzeolite 5A (20 wt% loading)

PESzeolite 5A (30 wt% loading)
PESzeolite 5A (40 wt% loading)
PESzeolite 5A (50 wt% loading)

218
219
220
221

1
1
l
1

217
218
220
221

1
1
l
1

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

21

Fig. 3. Comparison of gas permeability of PESzeolite A MMMs before and after the chemical modification of zeolite surface (A: He permeability; B: H2 permeability;
C: O2 permeability; D: CO2 permeability).

Fig. 4. Comparison of gas pair selectivity of PESzeolite A MMMs before and after the chemical modification of zeolite surface (A: He/N2 selectivity; B: H2 /N2
selectivity; C: O2 /N2 selectivity; D: CO2 /CH4 selectivity).

22

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

Fig. 5. Comparison of cross-section SEM images of MMMs before and after the chemical modification of zeolite surface (AC: PESzeolite 3A, 4A and 5A MMMs,
respectively; DF: PESzeolite 3A-NH2 , 4A-NH2 and 5A-NH2 MMMs, respectively).

If the attachment of the polymer matrix upon zeolite

(59 A).
surface is the main cause of partial pore blockage of zeo thick
lites [2325], the addition of (59) 1010 m (59 A)
APDEMS on top of zeolite surface may effectively reduce the
direct partial pore blockage induced by the polymer matrix. As
a result, the gas permeability of MMMs increases if APDEMS
modified zeolite is used.
The same reasons may be applicable to explain the permselectivity increase if MMMs are made of APDEMS modified zeolite
as illustrated in Fig. 4. Previous works have demonstrated that,
owing to partial pore blockage, the selectivity obtained from the
experiments is far below from the prediction calculated from
the Maxwell model [2325]. Because of the unique structure of
APDEMS, the thin APDEMS layer on zeolite surface not only
can reduce the pore blockage as discussed in the previous section, but also can diminish the transport of gas penetrants via
the interphase due to better adhesion between APDEMS and the
polymer matrix and the existence of various functional groups
in this space (i.e., CH2 and CH3 ). Therefore, the originally high
selectivity nature of zeolite remains almost intact and the resultant mixed matrix membranes have a higher permselectivity.
3.3. Effect of zeolite loadings on gas permeability
PESzeolite 4A-NH2 and PESzeolite 5A-NH2 mixed
matrix membranes with different zeolite loadings were fabricated. Neat PES dense films were also prepared with the
same procedure for comparison. The permeability of all gases
and selectivity of four gas pairs at different zeolite loadings
are shown in Figs. 6 and 7 , respectively. The permeability

of all gases decreases with an increase in zeolite content for

PESzeolite 4A-NH2 MMMs, while the relationship between
gas permeability and zeolite loading for PESzeolite 5A-NH2
MMMs is not so straightforward. Their gas permeability shows
a decrease then an increase with increasing zeolite loading. The
maximum percentages of permeability increments of He, H2 ,
O2 and N2 for PESzeolite 5A-NH2 MMMs are approximately
73%, 67%, 35% and 11%, respectively, at 50 wt% zeolite loading.
The decreasing trend of gas permeability of PESzeolite 4ANH2 MMMs with zeolite loading may be easily understandable
because it has been previously demonstrated that both polymer chain rigidification and partial pore blockage of zeolites
may lead to a decrease in the permeability of MMMs [2325].
However, for PESzeolite 4A-NH2 MMMs, the polymer chain
rigidification may play a much more important role than the
partial pore blockage. This is due to the fact that the surface
modification by APDEMS has lowered the degree of pore blockage in zeolite. In addition, the slight increase in Tg with zeolite
loading as shown in Table 3 implies greater chain rigidification with increasing zeolite loading. As a result, permeability
of PESzeolite 4A-NH2 MMMs decreases with an increase
in zeolite loading. A modified Maxwell model [25] will be
used to quantitatively estimate the effect of both polymer chain
rigidification and partial pore blockage of zeolites on the O2 permeability of PESzeolite 4A-NH2 MMMs in the next section.
However, why the permeability of PESzeolite 5A-NH2
MMMs decreases and then increases with increasing zeolite content as shown in Fig. 6? This trend is different from our previous
results [25] where the permeability of PESzeolite 5A MMMs

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

23

Fig. 6. Effect of zeolite loadings on gas permeability of PESzeolite 4A-NH2 and PESzeolite 5A-NH2 MMMs (A: He and H2 permeability; B: CO2 permeability;
C: O2 permeability; D: N2 and CH4 permeability).

monotonously decreases with an increase in the zeolite loading.

This difference may arise from three factors: (1) the pore size of
is larger than the molecular
zeolite 5A, 4.8 1010 m (4.8 A),
kinetic diameter of all tested gases; therefore, the permeability of PESzeolite 5A MMMs should increase with an increase
in zeolite loadings, (2) the effect of polymer chain rigidification may still play an important role when the zeolite loading is
small. The benefit of using large pore-size zeolite only magnifies when its content is greater than approximately 30% and (3)
even though the APDEMS surface modification can reduce the
degree of partial pore blockage, the partial pore blockage cannot be eliminated completely. A part of zeolite 5A pores may be
thus leading to a
narrowed to approximately 4 1010 m (4 A),
permeability decrease in PESzeolite 5A MMMs. Once the zeolite loading is further increased, the positive effect of using large
pore-size zeolite gradually offsets the negative effect of partial
pore blockage on permeability. As a consequence, permeability
becomes increase with an increase in zeolite loading.
The above three factors compete one another and result in
different influence on the relationship between gas permeability
and zeolite loading for different gases. For He and H2 gases with
smallest kinetic diameters, their permeability starts to surpass
that of neat PES dense film from 20 wt% zeolite loading, while
for O2 gas, it begins from 40 wt% loading; for N2 , it starts from
50 wt% loading; for CH4 gas, it is always lower than that of
neat PES dense film in our experimental range. The information

hints different gases exhibit different responses to polymer chain

rigidification and partial pore blockage of zeolites.
The above phenomena and arguments may not be applicable
to the case of PESzeolite 4A-NH2 MMMs. This arises from
the fact that the pore size of zeolite 4A is about 3.8 1010 m
which is very approximate with the kinetic diameter of
(3.8 A),
tested gases. Therefore, the APDEMS induced chemical modification cannot bring much positive effects on the permeability
versus zeolite content relationship.

3.4. Effect of zeolite loadings on gas permselectivity

Fig. 7 exhibits that the selectivity of both PESzeolite
4A-NH2 and PESzeolite 5A-NH2 MMMs increases with an
increase in zeolite loadings. The maximum increment occurs
at 50 wt% zeolite loading. The maximum percentages of selectivity increments of He/N2 , H2 /N2 , O2 /N2 and CO2 /CH4 for
PESzeolite 4A-NH2 MMMs are approximately 33%, 34%,
32% and 44%, respectively, and for PESzeolite 5A-NH2
MMMs are around 56%, 51%, 22% and 15%, respectively.
Amazingly, PESzeolite 5A-NH2 MMMs not only increase the
permeability of He, H2 and O2 , but also enhance the selectivity of He/N2 , H2 /N2 and O2 /N2 , especially at 50 wt% loading.
This simultaneous advance in both permeability and selectivity
is very attractive to potential industrial applications.

24

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

Fig. 7. Effect of zeolite loadings on gas pair selectivity of PESzeolite 4A-NH2 and PESzeolite 5A-NH2 MMMs (A: He/N2 selectivity; B: H2 /N2 selectivity; C:
O2 /N2 selectivity; D: CO2 /CH4 selectivity).

The increasing trend in gas pair selectivity of both

PESzeolite 4A-NH2 and PESzeolite 5A-NH2 MMMs with
an increase in zeolite loadings is easily explainable due to the
influence of molecular sieving mechanism via zeolite. Although
the magnitude of increment is still below from the calculation
of the Maxwell model, the gas separation performance of the
newly developed MMMs made from APDEMS modified zeolite is much higher than those made from unmodified zeolite
[25].
3.5. Applicability of the modied Maxwell model to predict
gas separation performance
In our previous work [25], a modified Maxwell model was
proposed to predict the gas separation performance. It takes the
combined effects of the polymer chain rigidification and partial pore blockage of zeolites into calculation. This modified
Maxwell model needs intrinsic gas permeation properties of zeolites to predict the gas separation performance of MMMs in the
calculation. So far only O2 and N2 permeability data of zeolite
4A can be found in the literatures and relatively consistent, while
there are no consistent CO2 and CH4 permeability data of zeolite 4A available. For other zeolites, such as 5A, silicalite-1, no
consistent data on intrinsic gas permeability can be found in the
literatures. Therefore, the PESzeolite 4A-NH2 MMM is chosen
as a sample to predict O2 permeability and O2 /N2 selectivity by
using this modified Maxwell model in this study. This exercise

may provide convincing information to quantitatively estimate

the effect of chemical modification on the polymer chain rigidification and partial pore blockage of zeolites.
Firstly, we assume the whole system of mixed matrix membranes as a pseudo three-phase composite, as illustrated in Fig. 8
. The first phase is the polymer matrix; the second phase is composed of the third phase and rigidified polymer region near the
zeolite; and the third phase comprises the bulk of zeolite and the
zeolite skin affected by the partial pore blockage. Secondly, the
permeability of the third phase can be calculated with the aid of
the Maxwell equation [35] by assuming the bulk of zeolite as
the dispersed phase and the zeolite skin affected by the partial
pore blockage as the continuous phase



PD + 2Pblo 2v3 (Pblo PD )
P3rd = Pblo
(1)
PD + 2Pblo + v3 (Pblo PD )
where P3rd is the composite permeability of the third phase, PD
the permeability of the bulk of zeolite, Pblo the permeability of
the zeolite skin affected by the partial pore blockage and v3 is
the volume fraction of the bulk of zeolite in the third phase and
can be determined by the following equation:
vD
(2)
v3 =
vD + vblo
where vD is the volume fraction of the bulk of zeolite in the
total membrane and vblo is the volume fraction of zeolite skin
affected by the partial pore blockage in the total membrane.

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

25

Fig. 8. Schematic diagram for the modified Maxwell model.

Thirdly, the permeability of the second phase can be determined with the aid of Maxwell equation a second time by
regarding the rigidified polymer region as the continuous phase
and the third phase as the dispersed phase



P3rd + 2Prig 2v2 (Prig P3rd )
P2nd = Prig
(3)
P3rd + 2Prig + v2 (Prig P3rd )
where P2nd is the composite permeability of the second phase,
Prig the permeability of the rigidified polymer region and v2 is
the volume fraction of the third phase in the second phase and
can be estimated by the following equation:
v2 =

vD + vblo
vD + vblo + vrig

(4)

where vrig is the volume fraction of the rigidified polymer region

in the total membrane.
Finally, one can obtain the permeability of the whole mixed
matrix membrane with the aid of Maxwell equation a third time
by considering the polymer matrix as the continuous phase (i.e.,
the first phase) and the second phase as the dispersed phase



P2nd + 2PC 2(vD + vblo + vrig )(PC P2nd )
PMMM = PC
P2nd + 2PC + (vD + vblo + vrig )(PC P2nd )
(5)

where PMMM is the composite permeability of mixed matrix

membranes and PC is the permeability of the polymer matrix
(i.e., the first phase).
Zeolite 4A Linde crystals have an O2 permeability
of around 5.775385 1018 m2 s1 Pa1 (0.77 Barrer) and
an O2 /N2 selectivity of around 37 at 35 C [36]. The
neat PES dense film has an O2 permeability of around
3.5927395 1018 m2 s1 Pa1 (0.479 Barrer) and an O2 /N2
selectivity of 5.81 at 35 C according to our experimental results.
Therefore, four unknowns in Eqs. (1)(5) need to be set to predict
the resultant permeability of mixed matrix membranes. They are
Pblo , v3 , Prig and v2 , respectively. Prig and Pblo can be assumed
to be the permeability of the polymer matrix divided by a chain
immobilization factor and the permeability of the bulk of zeolite divided by an average decline factor  induced by the partial
pore blockage, respectively [2325,37]. v2 and v3 can be calculated if the volume fraction of the rigidified polymer region and
affected zeolite skin in the total membrane are known.
The polymer chain rigidification near the zeolite surface is
probably resulted from the inhibited isotropic contraction of the
polymer. Zeolite with a smaller particle size may have a less
effect on the inhibition of the polymer contraction, thus leading to a smaller thickness of rigidified polymer region [24]. The
thickness of rigidified polymer region near the zeolite, r, has

26

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

Table 4
Change of O2 and N2 permeability in different regions of the PESzeolite 4A-NH2 MMM
Permeability in the
PES matrix (Pc )a

Permeability in the rigidified

PES region (Prig = Pc /b )

Permeability in the zeolite 4A skin affected

with partial pore blockage (Pblo = PD /c )

Permeability in the bulk

of zeolite 4 A (PD )d

O2
0.479

0.160

0.00308

0.770

N2
0.0825

0.0206

0.00208

0.0208

a
b
c
d

Gas permeability data in the PES matrix come from experiment results of neat PES dense film.
is 3 for O2 gas, while is 4 for N2 gas.
 is 250 for O2 gas, while  is 10 for N2 gas.
Gas permeability data in the bulk of zeolite 4A come from reference [36].

Table 5
Calculated volume fraction data of dispersed phase in different phases of PESzeolite 4A-NH2 MMMs in the modified Maxwell model which simultaneously
considers both polymer chain rigidification and partial pore blockage of zeolites
Calculated volume fraction of the bulk of zeolite 4A (considered as the
dispersed phase) in the third phase (v3 ) (defined in Fig. 8)
Calculated volume fraction of the third phase (considered as the dispersed
phase) in the second phase (v2 ) (defined in Fig. 8)
Calculated volume fraction of the second phase (considered as the
dispersed phase) in the whole mixed matrix membrane (defined in Fig. 8)

been assumed to be 0.66 m [24], 0.3 m [25] and 0.066 m

[24] as the average particle size of zeolite 4A or carbon molecular sieve is 5 m [24], 3 m [25] and 1 m [24], respectively.
Therefore, r is assumed to be 0.12 m in this work because the
average particle size of used zeolite 4A is 1.5 m. This inhibited
isotropic contraction is assumed to affect only the thickness of
rigidified polymer region, but no effect on the extent of polymer
chain rigidification [24]. Therefore, the chain immobilization
parameter is still assumed to be 3 for O2 gas and 4 for N2 in
this work which are the same as our previous work [25]. The
partial pore blockage of zeolites is resulted from the attachment
of polymer chains on the zeolite surface. The thickness of this
affected zeolite skin, r , is assumed to be almost unchanged with
zeolite particle size [2325]. However, the chemical modification of zeolite surface may reduce the extent of the partial pore
blockage, so in this work, r is assumed to be 70 1010 m
or more) in
lower than that (i.e., about 108 m (100 A)
(70 A)
other related literatures [2325]. To simplify the calculation, we
presume neither the particle size of zeolites nor the chemical
modification has any influence on the magnitude of permeability reduction in this zeolite skin affected by the partial pore
blockage; therefore,  is still assumed to 250 for O2 and 10
for N2 in this work. The reason that and  have different
values for different gases has been explained in our previous
work [25]. O2 and N2 permeability data in different regions of
PESzeolite 4A-NH2 MMMs are shown in Table 4. All calculated volume fraction data of dispersed phase defined in Fig. 8 in
the third phase, second phase and whole MMM in this modified
Maxwell model are listed in Table 5. The calculation method
to obtain these volume fractions has been discussed extensively
by Moore et al. [24], so it is not necessary to duplicate it in this
work.

0.972
0.641
20 wt% Zeolite loading (18.6 vol% zeolite loading)
30 wt% Zeolite loading (28.1 vol% zeolite loading)
40 wt% Zeolite loading (37.8 vol% zeolite loading)
50 wt% Zeolite loading (47.7 vol% zeolite loading)

0.290
0.439
0.590
0.745

Fig. 9. Comparison of O2 permeability of PESzeolite 4A-NH2 MMMs

between experimental data and predictions from the Maxwell model and the
modified Maxwell model.

Fig. 10. Comparison of O2 /N2 selectivity of PESzeolite 4A-NH2 MMMs

between experimental data and predictions from the Maxwell model and the
modified Maxwell model.

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

Figs. 9 and 10 illustrate the ultimate prediction of O2 permeability and O2 /N2 selectivity of PESzeolite 4A-NH2 MMMs,
respectively. It can be found that the prediction from the modified Maxwell model is in very good agreement with experimental
data. Therefore, this demonstrates our revised Maxwell model
can be utilized for MMMs made from both unmodified and modified zeolite.
4. Conclusions
The following conclusions can be drawn from this work:
(1) Results from elementary analysis and XPS demonstrate that
the silane coupling agent (APDEMS) has been successfully
grafted onto the external surface of zeolite after the chemical
modification. BET results show that the surface area and
total pore volume of zeolites before and after modification
are almost the same, indicating the micropore of zeolites is
almost not influenced by the addition of APDEMS. These
characterizations assure the zeolite modified by APDEMS
as a good candidate which can be applied in the fabrication
of MMMs.
(2) Both gas permeability and gas pair selectivity of
PESzeolite A-NH2 MMMs are higher than those of
PESzeolite A MMMs at 20 wt% zeolite loading. This may
be mainly due to the fact that the presence of APDEMS

introduces a distance of around (59) 1010 m (59 A)

between polymer chains and zeolite surface, thus reduces
the extent of the partial pore blockage of zeolites induced
by polymer chains.
(3) The permeability of all studied gases decreases with increasing zeolite content for PESzeolite 4A-NH2 MMMs, while
for PESzeolite 5A-NH2 MMMs, the gas permeability
decreases and then increases with an increase in zeolite loadings. Clearly, a high loading of zeolite 5A-NH2 offsets the
effects of partial pore blockage and polymer chain rigidification on permeability. The selectivity of both PESzeolite
4A-NH2 and PESzeolite 5A-NH2 MMMs increases with
an increase in zeolite loadings due to the influence of molecular sieving mechanism.
(4) A modified Maxwell model proposed in our previous work
[25] was applied to predict the gas separation performance
of the newly developed PESzeolite 4A-NH2 MMMs with
adjusted parameters. The predicted permeability and selectivity show very good agreement with experimental data.
Acknowledgements
The authors would like to thank UOP LLC and NUS for
funding this research with the grant numbers of R-279-000-140592, R-279-000-140-112 and R-279-000-184-112. Dr. Huang is
thanked for aiding in the ion exchange treatment of zeolite 4A.
References
[1] G.W. Peng, F. Qiu, V.V. Ginzburg, D. Jasnow, A.C. Balazs, Forming
supramolecular networks from nanoscale rods in binary, phase-separating
mixtures, Science 288 (2000) 1802.

27

[2] T.C. Merkel, B.D. Freeman, R.J. Spontak, Z. He, I. Pinnau, P. Meakin,
A.J. Hill, Ultrapermeable, reverse-selective nanocomposite membranes,
Science 296 (2002) 519.
[3] Y. Wang, N. Herron, X-ray photoconductive nanocomposites, Science
273 (1996) 632.
[4] S. Kulprathipanja, R.W. Neuzil, N.N. Li, Separation of fluids by means
of mixed matrix membranes in gas permeation, US Patent No. 4,740,219
(1988).
[5] H.J.C. Te Hennepe, Zeolite Filled Polymeric Membranes: A New Concept in Separation Science, Thesis, University of Twente, 1988.
[6] M. Jia, K.V. Peinemann, R.D. Behling, Molecular sieving effect of zeolite filled silicone rubber membranes, J. Membr. Sci. 57 (1991) 289.
[7] J.M. Duval, B. Folkers, M.H.V. Mulder, G. Desgrandchamps, C.A. Smolders, Adsorbent-filled membranes for gas separation. Part 1. Improvement of the gas separation properties of polymeric membranes by
incorporation of microporous adsorbents, J. Membr. Sci. 80 (1993) 189.
[8] D.R. Paul, D.R. Kemp, The diffusion time lag in polymer membranes
containing adsorptive fillers, J. Polym. Sci. Symp. 41 (1973) 79.
[9] M.G. Suer, N. Bac, L. Yilmaz, Gas permeation characteristics of
polymerzeolite mixed matrix membranes, J. Membr. Sci. 91 (1994)
77.
[10] R. Mahajan, W.J. Koros, Factors controlling successful formation of
mixed-matrix gas separation materials, Ind. Eng. Chem. Res. 39 (2000)
2692.
[11] A. Erdem-Senatalar, M. Tatler, S.B. Tantekin-Ersolmaz, Estimation of
the interphase thickness and permeability in polymerzeolite mixed
matrix membrane, in: Proceedings of the 13th International Zeolite Conference, Montpellier, France, 2001.
[12] S.B. Tantekin-Ersolmaz, L. Senorkyan, N. Kalaonra, M. Tatler, A.
Erdem-Senatalar, n-Pentane/i-pentane separation by using zeolitePDMS
mixed matrix membranes, J. Membr. Sci. 189 (2001) 59.
[13] R. Mahajan, R. Burns, M. Schaeffer, W.J. Koros, Challenges in forming
successful mixed matrix membranes with rigid polymeric materials, J.
Appl. Polym. Sci. 86 (2002) 881.
[14] R. Mahajan, W.J. Koros, Mixed matrix membrane materials with glassy
polymers. Part 1, Polym. Eng. Sci. 42 (7) (2002) 1420.
[15] R. Mahajan, W.J. Koros, Mixed matrix membrane materials with glassy
polymers. Part 2, Polym. Eng. Sci. 42 (7) (2002) 1432.
[16] J.M. Duval, A.J.B. Kemperman, B. Folkers, M.H.V. Mulder, G. Desgrandchamps, C.A. Smolders, Preparation of zeolite filled glassy polymer membrane, J. Appl. Polym. Sci. 54 (1994) 409.
[17] Y.L. Liu, Y.H. Su, K.R. Lee, J.Y. Lai, Crosslinked organicinorganic
hybrid chitosan membranes for pervaporation dehydration of
isopropanolwater mixtures with a long-term stability, J. Membr. Sci.
251 (2005) 233.
[18] H.H. Yong, H.C. Park, Y.S. Kang, J. Won, W.N. Kim, Zeolite-filled polyimide membrane containing 2,4,6-triaminopyrimidine, J. Membr. Sci.
188 (2001) 151.
[19] A. Matsumoto, K. Tsutsumi, K. Schumacher, K.K. Unger, Surface
functionalization and stabilization of mesoporous silica spheres by
silanization and their adsorption characteristics, Langmuir 18 (2002)
4014.
[20] E.P. Plueddemann, Silane Coupling Agents, second ed., Plenum Press,
New York, 1991.
[21] D.H. Flinn, D.A. Guzonas, R.H. Yoon, Characterization of silica surfaces
hydrophobized by octadecyltrichlorosilane, Colloids Surf. A: Physicochem. Eng. Aspects 87 (1994) 163.
[22] K.C. Vrancken, K. Possemiers, P. Van Der Voort, E.F. Vansant, Surface
modification of silica gels with aminoorganosilanes, Colloids Surf. A:
Physicochem. Eng. Aspects 98 (1995) 235.
[23] R. Mahajan, Formation, characterization and modeling of mixed matrix
membrane materials, Ph.D. Dissertation, The University of Texas,
Austin, 2000.
[24] T.T. Moore, R. Mahajan, De Q. Vu, W.J. Koros, Hybrid membrane materials comprising organic polymers with rigid dispersed phases, AICHE
J. 50 (2) (2004) 311.
[25] Y. Li, T.S. Chung, C. Cao, S. Kulprathipanja, The effects of polymer chain rigidification, zeolite pore size and pore blockage on

28

[26]
[27]

[28]
[29]

[30]

[31]

Y. Li et al. / Journal of Membrane Science 275 (2006) 1728

polyethersulfone (PES)zeolite A mixed matrix membranes, J. Membr.
Sci. 260 (2005) 45.
J.S. Chiou, Y. Maeda, D.R. Paul, Gas permeation in polyethersulfone,
J. Appl. Polym. Sci. 33 (1987) 1823.
H.M. Guan, T.S. Chung, Z. Huang, M.L. Chng, S. Kulprathipanja,
Poly(vinyl alcohol) multilayer mixed matrix membranes for the
dehydration of ethanolwater mixture, J. Membr. Sci., doi:10.1016/
j.memsci.2005.05.032.
K. Taizo, N. Tooru, T. Isao, I. Wataru, Manufacture of zeolite 3A from
zeolite 4A, European Patent No. JP5147926 (1993).
W. Heide, Production of acid-modified potassium zeolite 3A by cation
exchange of sodium zeolite 4A, European Patent No. DE10107819
(2002).
W.H. Lin, R.H. Vora, T.S. Chung, The gas transport properties of
6FDA-durene/1,4-phenylenediamine (pPDA) copolyimides, J. Polym.
Sci. Polym. Phys. 38 (2000) 2703.
R. Sirnha, R.F. Bayer, General relation involving the glass temperature
and coefficients of expansion of polymers, J. Chem. Phys. 37 (1962)
1003.

[32] A. Eisenberg, B. Hird, R.B. Moore, A new multiplet-cluster model

for the morphology of random ionomers, Macromolecules 23 (1990)
4098.
[33] V.V. Tsukruk, V.N. Bliznyuk, Adhesive and friction forces between
chemically modified silicon and silicon nitride surfaces, Langmuir 14
(1998) 446.
[34] D.F. Siqueira Petri, G. Wenz, P. Schunk, T. Schimmel, An improved
method for the assembly of amino-terminated monolayers on SiO2
and the vapor deposition of gold layers, Langmuir 15 (1999)
4520.
[35] K. Mizoguchi, K. Terada, Y. Naito, Y. Kamiya, S. Tsuchida, S.
Yano, Miscibility blends and gas permeability of poly(ethylene-co-3,5,5trimethylhexyl methacrylate)polydimethylsiloxane, Colloid Polym. Sci.
86 (1997) 275.
[36] C.M. Zimmerman, A. Singh, W.J. Koros, Tailoring mixed matrix composite membranes for gas separation, J. Membr. Sci. 137 (1997)
145.
[37] A.S. Michaels, R.B. Parker, Sorption and flow of gases in polyethylene,
J. Polym. Sci. 41 (138) (1959) 53.