Advances in Chemical Engi neering and Science , 2011, 1, 305-312
doi:10.4236/aces.2011.14042 Published Online October 2011 (http://www.SciRP.org/journal/aces)
Separation Performance of Sodium Alginate/Poly(Vinyl
Pyrrolidone) Membranes for Aqueous/Dimethylformamide
Mixtures by Vapor Permeation and Vapor Permeation with
Temperature Difference Methods
Ebru Kondolot Solak1, Oya Şanlı2*
1Department of Cemistry and Chemical Processing Technology, Atatürk Vocational High School,
Gazi University, Teknikokullar, Ankara, Turkey
2Faculty of Science, Department of Chemistry, Gazi University, Teknikokullar, Ankara, Turkey
Recieved July 18, 2011; revised September 7, 2011; accepted September 19, 2011
Abstract
In this study sodium alginate (NaAlg)/poly(vinyl pyrrolidone) (PVP) blend membranes were prepared and
crosslinked with CaCl2 (0.1 Molarity (M)) for the separation of aqueous/dimethylformamide (DMF) mix-
tures. Membranes were characterized by Fourier transform infrared spectroscopy (FTIR), scanning electron
microscopy (SEM) and their performance was examined by varying experimental parameters such as feed
composition (0 - 100 wt%), operating temperature (30˚C - 50˚C) and membrane thickness (30 - 90 microme-
ter (m)). Blending NaAlg with PVP, decreased separation factor whereas increased the permeation rate as
the permeation temperature was increased in Vapor Permeation (VP) and Vapor Permeation with Tempera-
ture Difference (TDVP) methods. In the TDVP method, the separation factors increased and the permeation
rates decreased as the temperature of the membrane surrounding is decreased. The highest separation factor
of 60 was obtained in TDVP method for 90 wt% DMF concentration in the feed.
Keywords: Vapor Permeation, Sodium Alginate, Dimethylformamide, Membrane, Poly(Vinyl Pyrrolidone)
1. Introduction
Vapor permeation and vapor permeation with tempera-
ture difference methods were proposed by Uragami and
coworkers [1-2]. In these methods, membrane is in con-
tact with the vapor of feed mixture (Figure 1). Hence,
the swelling or shrinking of the membranes due to the
feed mixtures can be largely prevented and consequently
improvement of membrane performance may be ex-
pected. Furthermore, a temperature difference between
the membrane surrounding and the feed mixture was
established in the TDVP method (Figure 2).
DMF is an important solvent, it is primarily used as a
solvent in the production of acrylic fibers and plastics. It
is also used as a solvent in peptide coupling for pharma-
ceutics, in the development and production of pesticides
and in the manufacture of adhesives, synthetic leathers,
films and surface coatings [3]. Its separation from water
is important and essential because of it is carcinogenic to
human beings and other animals.
Alginic acid is a highly hydrophilic polymer used in
biotechnology, pharmaceutical, and cosmetic industries.
(a) (b)
Figure 1. Schematic diagram of (a) the VP method, (b) TDVP
method.

E. K. SOLAK ET AL.
306
Figure 2. Schematic diagram of the vapor permeation and vapor permeation with temperature difference apparatus used in
this study: 1 vacuum pump, 2-4, 6 permeation traps; 5 Mc Leod manometer; 7 vent; 8 permeation cell; 9 constant tempera-
ture water bath; 10 peristaltic pump; 11 temperature indicator; 12 feed tank.
It has widespread applications as a membrane material
because of its high hydrophilicity [4-7]. Although alginic
acid can hardly dissolve in commercially available
solvents, its alkali metal salt form (alginate), obtained by
neutralizing the acidic functional groups with strong
alkalis, is well soluble in water. Thus, a membrane can
be easily prepared from an alginate aqueous solution.
When a highly permeable polymer material is preferred,
this membrane material should be modified to have suit-
able combination of permeation rate and separation fac-
tor. In recent years alginic acid based membranes and
their modified forms are used in VP and TDVP methods
[8-10]. For this purpose, alginate and PVP semi-IPN
membranes were prepared and successfully crosslinked
with CaCl2 in this study (Scheme 1).
PVP is soluble in water and other polar solvents.
When dry it is a light flaky powder, which readily
absorbs up to 18% of its weight in atmospheric water. In
solution, it has excellent wetting properties and readily
forms films. The monomer is carcinogenic and is extre-
mely toxic to aquatic life. However, the polymer PVP in
its pure form is so safe that not only it is edible by
humans, but also it was used as a blood plasma expander
for trauma victims after the first half of 20th century.
There are a few articles about the DMF and water
mixtures in the literature [8-14]. Shah and coworkers [12]
prepared hydrophilic zeolite NaA membranes for the PV
separation of DMF-water mixtures. They have reported
that the water flux for the DMF-water system decreases
rapidly with an increase in feed DMF concentration.
Aminabhavi and Naik [13] grafted poly(vinyl alcohol)
(PVA) with acrylamide for the separation of water/DMF
mixtures. It was found that these membranes are more
selective to water than DMF. Separation factors in-
creased with grafting, but permeation flux did not con-
siderable change with grafting.
In our previous study [10] we have used NaAlg/PVP
membrane for the pervaporation separation of aque-
ous/DMF mixtures. As a confirmation of the use of
NaAlg/PVP membranes, in this research we have aimed
to investigate separation characteristics of aqueous/DMF
mixtures by VP and TDVP methods.
2. Experimental
2.1. Materials
DMF (C3H7NO, purity; 99.9%), calcium chloride (CaCl2,
purity; 90%) and PVP ((C6H9NO)n, purity; 99.9%) were
obtained from Merck and used as supplied. NaAlg
((C6H7NaO6)n, medium viscosity) was provided from
Sigma.
2.2. Preparation of Blend Membranes
PVP (8 wt%) and NaAlg (2 wt%) were dissolved in wa-
ter, mixed in different ratios (w/w), stirred and then
casted onto rimmed round glass dishes [10].
Solvent was evaporated at 60˚C to form the membrane.
The dried membrane was crosslinked with calcium chlo-
ride (0.1 M) for 24 h. The thickness of the membranes
thus prepared was 70 (±10) μm. Membranes prepared in
this research were used at least 10 times without any
deformation during the VP and TDVP processes.
2.3. Swelling Study of the Blend Membranes
Dried membranes were immersed in different concentra-
tions of DMF/water mixtures at 40˚C for 48 h. Then
these membranes wiped with cleansing tissue to remove

E. K. SOLAK ET AL.
307
Scheme 1. Schematic representation of the polymer.
the excess solvent mixture. These samples were dried at
60 ˚C until a constant weight and water uptake was cal-
culated as;

Water uptake(%)=100
SD
D
MM
M
 (1)
where MS is the mass of the swollen membrane in the
feed solution and MD is the mass of the dried membrane.
2.4. Vapor Permeation and Vapor Permeation
with Temperature Difference Experiments
In VP and TDVP methods, the capacity of permeation
cell was about 150 mL [8]. The effective membrane area
was 16.5 cm2 and pressure was kept at 0.6 mbar with a
vacuum pump (Edwards). The mixture of DMF and wa-
ter that was used as a feed solution placed into the lower
part of the permeation cell, permeation side of the cell
(upper part) was kept under vacuum. The feed mixture
was circulated between the permeation cell and feed tank
at constant temperature and permeate was collected in
liquid nitrogen traps. In TDVP method while the tem-
perature of the feed solution was kept constant 40˚C
temperature of the membrane surroundings (0˚C - 50˚C)
was controlled by a cold medium in a permeation cell of
a jacket type. Upon reaching steady state conditions
permeate vapor was collected in liquid nitrogen traps and
weighed. The composition of permeates was found by

308 E. K. SOLAK ET AL.
means of refractive index values with Atago DD-5 type
digital refractometer. Then these indexes were translated
to concentration by using a calibration curve.
Membrane performance was expressed by separation
factor (α) and permeation rate (J). The separation factor
was defined as follows [15]:
.
WDMF
sep WDMF
WDMF
PP
FF
  (2)
where PW and PDMF, FW and FDMF and are the mass frac-
tions of water and DMF components in the permeate and
feed vapor (measured by isoteniscope method) respec-
tively. The permeation rate was calculated using Equa-
tion 3.
.
W
J
A
t
 (3)
where W is the mass of permeate (kg), A is membrane
surface area (m2), t is the experiment time (h).
3. Results and Dicussion
3.1. Blend Ratio Selection and Membrane
Characterization
In our previous study (10) we have studied NaAlg/PVP
blend ratios of 100/0, 95/5, 90/10, 85/15, 80/20, 75/25
(w/w) for 20 wt% DMF/water mixtures at 40˚C and de-
cided to use 75/25 ratio due to its acceptable flux and
separation factor. For this reason in this study we have
used this ratio for the membrane preparation. It was
found in the previous study that, the permeation rate in-
creases whereas separation factor decreases as the PVP
content of the membrane increases [16-18]. In this study,
the results were supported by the crosslink density and
molecular mass between crosslinks (MC) (Table 1).
MC of the polymer was determined by using Flory-
Rehner Equation [19] as given below:

1
13 2
ln 1
CpS
MV  


 
 
(4)

is the volume fraction of the polymer in the swollen
state and can be calculated as:
Table 1. Molecular mass between crosslinks (MC) values
and crosslink density (δx) of the NaAlg/PVP blend mem-
branes.
1
1a
P
Sb S
M
M



P



 

 




 (5)
where
P

and S

are the densities of the polymer and
solvent, respectively. Ma and Mb are the mass of the
polymer before and after swelling, respectively. Vs is the
molar volume fraction of the polymer in the swollen
state.
Interaction parameter
 can be calculated from fol-
lowing equation.
 

1
1
1
221
1ln1
2
NN
NT T
 
 






 




 


(6)
where

1
23 13
3232 3N 

 

and temperature
is taken as Kelvin.
For the NaAlg/PVP membrane, the MC value was
found to be 1680.
Finally crosslink density,
x

;
x
pc
M

  (7)
In our study the crosslink density of NaAlg/PVP
(75/25, w/w) membrane was found as 5.30 × 10–4
mol/cm3 taking the

(aqueous/DMF) as 0.7238.
We have studied with 75/25 (NaAlg/PVP, w/w) ratio
in the rest of the study due to acceptable flux and separa-
tion factor. The prepared membrane was characterized
with Fourier transform infrared spectroscopy (Figure 3)
and scanning electron microscopy (Figure 4).
Crosslinked membrane was scanned with Mattson
1000 Fourier Transfer Infrared Spectroscopy (FTIR)
(Figure 3). In the FTIR spectrum of NaAlg/PVP and
NaAlg, the peak at 3000 - 3500 cm–1 area presences the
stretching vibration of –OH band. In the FTIR spectrum
of NaAlg/PVP these stretching vibrations appear as a
wider band than the spectrum of NaAlg. This peak ap-
pear at 3445 cm–1 in the spectrum of PVP. The peak at
1625 cm–1 in the spectrum of NaAlg is due to the
stretching band of C=O. The spectrum of PVP appears
strong absorption band at 1640 cm–1, due to the presence
of the C=C-N group. In the FTIR spectrum of
NaAlg/PVP, these bands were seen together. The spec-
trums of PVP, NaAlg and NaAlg/PVP appear stretching
bands of C-H group at 2964 cm–1, 2946 cm–1 and 2954
cm–1, respectively.
The morphology of the NaAlg and NaAlg/PVP mem-
branes was observed using Scanning Electron Micros-
copy (SEM, JEOL JSM-6400) (Figures 4(a) and 4(b)). It
was seen from the SEM results that the NaAlg membrane
surface (Figure 4(a)) had a smoother appearance than
the blend membrane [10].
NaAlg/PVP (w/w) δx (mol/cm3) × 10–4 MC (g/mol) × 103
100/0 10.52 0.565
95/5 9.81 0.670
90/10 8.23 0.711
85/15 7.01 0.956
80/20 6.94 1.143
75/25 5.30 1.680

E. K. SOLAK ET AL.
309
Figure 3. IR spectra of NaAlg/PVP, PVP and NaAlg mem-
branes.
(a)
(b)
Figure 4. (a) Scanning electron microscopic picture of
NaAlg membrane; (b) Scanning electron microscopic pic-
ture of NaAlg/PVP membrane.
3.2. Effect of Temperature in VP
The effect of operating temperature on the permeation
rate and separation factor was studied for the NaAlg/PVP
semi-IPN membrane using 20 wt% dimethylformamide
solutions. Results were shown in Figure 5. Permeation
rate increased whereas the separation factors decreased
as the operating temperature increased. The temperature
increases membrane becomes more swollen and both
DMF and water molecules diffuse easily through the
membrane. As a result, the permeation rate increases
whereas the separation factor decreases with increasing
temperature. Similar results were reported in the litera-
ture [20-23].
Sommer and Melin [22] studied influence of operation
parameters on the separation of mixtures by pervapora-
tion and vapor permeation with silica membranes. They
have reported that increase in temperature improved flux
rates for alcohol/water mixtures.
3.3. Effect of Feed Composition in VP
Figure 6 illustrates the permeation performance of
semi-IPN membranes in VP at 40˚C. As can be seen
from the figure, the best separation factors were obtained
at high feed compositions. As the amount of water in the
feed vapor increases membrane material becomes more
swollen (Figure 7) and DMF molecules that have larger
moleculer size than that of water molecules diffuse easily
through the swollen membrane. As a result, the permea-
tion rate increases whereas the separation factor de-
creases with increasing water content. Similar results
were reported in the literature [8,12].
3.4. Effect of the Membrane Surrounding
Temperature in TDVP
The effect of temperature of the membrane surroundings
Figure 5. Change in the permeation rate and the separation
factor with the temperature in VP.

310 E. K. SOLAK ET AL.
Figure 6. Effect of the feed composition in VP. The permea-
tion conditions; membrane thickness: 70 μm, operating tem-
perature: 40˚C, pressure: 0.6 mbar.
Figure 7. Change in the water uptake with the feed compo-
sition in VP.
on the permeation rate and the separation factor in TDVP
was studied and results were shown in Figure 8. In the
study the temperature of the feed solution was kept con-
stant at 40˚C and the temperature of the membrane sur-
rounding was changed in the range of 0˚C - 50˚C. In-
crease in the temperature of the membrane surroundings
increased the permeation rate and decreased the separa-
tion factor. Similar results were reported in the literature
[8].
Kondolot Solak and Şanlı [8] studied the separation cha-
racteristics of dimethylformamide/water mixtures through
alginate membranes by pervaporation, vapor permeation
and vapor permeation with temperature difference me-
thods. They have observed that the permeation rate in-
creased with the increase in temperature of the mem-
brane surroundings so the separation factor decreased.
The comparison of both methods was shown in Fig-
ures 9 and 10. As it is from the figures that the highest
separation factors were obtained in TDEV method. This may
Figure 8. Effect of the temperature of the membrane sur-
roundings on the permeation rate and separation factor.
The permeation conditions; membrane thickness: 70 μm,
temperature of the feed solution: 40˚C, pressure: 0.6 mbar.
Figure 9. Change in the separation factor in VP (■) and
TDVP (♦) methods. The permeation conditions; membrane
thickness: 70 μm, operating temperature: 40˚C, pressure:
0.6 mbar, membrane surrounding temperature: 10˚C.
Figure 10. Change in the permeation rate in VP (■) and
TDVP (♦) methods. The permeation conditions were as
follows; membrane thickness: 70 μm, operating tempera-
ture: 40˚C, pressure: 0.6 mbar, membrane surrounding
temperature: 10˚C.

E. K. SOLAK ET AL.
311
be attributed to the temperature difference between the
feed mixture and the membrane surrounding. When the
dimethylformamide and water molecules were vaporized,
these vaporized molecules came close the membrane
surrounding kept at a lower temperature, the DMF
molecules were liable to be aggregated more than the
water molecules (the freezing point of DMF (–61˚C) is
lower than that of water (0˚C)). This aggregation of
DMF was responsible for the increase of separation fac-
tor of water.
4. Conclusions
NaAlg/PVP blend membrane have been prepared and
used in the separation of aqueous/DMF mixtures by va-
por permeation and vapor permeation with temperature
difference methods. It was shown experimentally that
membranes could be used to separate aqueous/DMF
mixtures with acceptable permeation rates and separation
factors. The effects of feed composition, permeation
temperature and membrane surrounding temperature on
the permeation rate and the separation factor were invest-
tigated. Increase in the operating temperature in VP and
TDVP methods increased the permeation rate whereas
decreased the separation factor. Permeation rate de-
creased whereas separation factor increased as the DMF
content of the feed increased in both methods. The high-
est separation factor (60) was found in TDVP method
whereas highest permeation rate (0.986 kg/m2h) was
observed in VP method.
5. Acknowledgements
The authors are grateful to Gazi University Research
Fund for the support of this study.
6. References
[1] T. Uragami and K. Takigawa, “Permeation and Separa-
tion Characteristics of Ethanol-Water Mixtures through
Chitosan Derivative Membranes by Pervaporation and
Evapomeation,” Polymer, Vol. 31, No. 4, 1990, pp. 668-
672. doi:10.1016/0032-3861(90)90287-9
[2] T. Uragami, S. Kato and T. Miyata, “Structure of N-alkyl
Chitosan Membranes on Water-Permselectivity for Aque-
ous Ethanol Solutions,” Journal of Membrane Science,
Vol. 124, No. 2, 1997, pp. 203-211.
doi:10.1016/S0376-7388(96)00238-4
[3] C. A. Redich, W. S. Beckett, J. Sparer, K. W. Barwich, C.
A. Riely, H. Miller, S. L. Sigal, S. L. Shalat and M. R.
Cullen, “Liver Disease Associated with Occupational
Exposure to the Solvent Dimethylformamide,” Annals of
Internal Medicine, Vol. 108, No. 5, 1998, pp. 680-686.
[4] X. P. Wang, “Preparation of Crosslinked Alginate Com-
posite Membrane for Dehydration of Ethanol-Water Mix-
tures”, Journal of Applied Polymer Science, Vol. 77, No.
14, 2000, pp. 3054-3061.
doi:10.1002/1097-4628(20000929)77:14<3054::AID-AP
P50>3.0.CO;2-L
[5] C. K. Yeom and K. H. Lee, “Characterization of Sodium
Alginate and Poly (Vinyl Alcohol) Blend Membranes in
Pervaporation Separation,” Journal of Applied Polymer
Science, Vol. 67, No. 5, 1998, pp. 949-959.
doi:10.1002/(SICI)1097-4628(19980131)67:5<949::AID-
APP20>3.0.CO;2-H
[6] Y. Shi, X. Wang, G. Chen, G. Golemme and S. Zhang,
“Preparation and Characterization of High-Performance
Dehydrating Pervaporation Alginate Membranes,” Jour-
nal of Applied Polymer Science, Vol. 68, No. 6, 1998, pp.
959-968.
doi:10.1002/(SICI)1097-4628(19980509)68:6<959::AID-
APP9>3.0.CO;2-G
[7] C. K. Yeom and K. H. Lee, “Characterization of Sodium
Alginate Membrane Crosslinked with Glutaraldehyde in
Pervaporation Separation”, Journal of Applied Polymer
Science, Vol. 67, No. 2, 1998, pp. 209-219.
doi:10.1002/(SICI)1097-4628(19980110)67:2<209::AID-
APP3>3.0.CO;2-Y
[8] E. Kondolot Solak and O. Şanlı, “Separation Characteri-
stics of Dimethylformamide/Water Mixtures through Al-
ginate Membranes by Pervaporation, Vapor Permeation
and Vapor Permeation with Temperature Difference Me-
thods,” Separation Science and Technology, Vol. 41,
2006, pp. 627-646. doi:10.1080/01496390500526789
[9] E. Kondolot Solak and O. Şanlı, “Separation Characteri-
stics of Dimethylformamide/Water Mixtures Using Sodi-
um Alginate-g-N-vinyl-2-Pyrrolidone Membranes by Perva-
poration Method,” Chemical Engineering and Prosessing,
Vol. 47, No. 4, 2008, pp. 633-641.
doi:10.1016/j.cep.2006.12.001
[10] E. Kondolot Solak, G. Asman, P. Çamurlu and O. Şanlı,
“Sorption, Diffusion, and Pervaporation Characteristics
of Dimethylformamide/Water Mixtures Using Sodium
Alginate/Polyvinyl Pyrrolidone Blend Membranes,” Va-
cuum, Vol. 82, No. 6, 2008, pp. 579-587.
doi:10.1016/j.vacuum.2007.08.012
[11] D. A. Devi, B. Smitha, S. Sridhar and T. M. Aminabhavi,
“Pervaporation Separation of Dimethylformamide/Water
Mixtures through Poly(Vinyl Alcohol)/Poly(Acrylic Acid)
Blend Membranes,” Separation and Purification Techno-
logy, Vol. 51, No. 1, 2006, pp. 104-111.
doi:10.1016/j.seppur.2006.01.006
[12] D. Shah, K. Kissick, A. Ghorpade, R. Hannah amd D.
Bhattacharyya, “Pervaporation of Alcohol-Water and Di-
methylformamide-Water Mixtures Using Hydrophilic Zeo-
lite NaA Membranes: Mechanisms and Experimental Re-
sults,” Journal of Membrane Science, Vol. 179, No. 1-2,
2000, pp. 185-205. doi:10.1016/S0376-7388(00)00515-9
[13] T. M. Aminabhavi and H. G. Naik, “Pervaporation Sepa-
ration of Water/Dimethylformamide Mixtures Using
Poly(Vinyl Alcohol)-g-polyacrylamide Copolymeric Mem-
branes,” Journal of Applied Polymer Science, Vol. 83, No.
2, 2002, pp. 273-282. doi:10.1002/app.2242
[14] M. D. Kurkuri and T. M. Aminabhavi, “Polyacryloni-

E. K. SOLAK ET AL.
312
Trile-g-Poly(Vinyl Alcohol) Membranes for the Perva-
poration Separation of Dimethyl Formamide and Water
Mixtures,” Journal of Applied Polymer Science, Vol. 91,
No. 6, 2004, pp. 4091-4097. doi:10.1002/app.13640
[15] G. Asman and O. Şanlı, “Characteristics of Permeation
and Separation for Acetic Acid-Water Mixtures Through
Poly(Vinyl Alcohol) Membranes Modified with Poly(Ac-
rylic Acid),” Separation Science and Technology, Vol. 38,
No. 9, 2003, pp. 1963-1980.
doi:10.1081/SS-120020129
[16] H. Wu, X. Fang, X. Zhang, Z. Jiang, B. Li and X. Ma,
“Cellulose Acetate-Poly(N-Vinyl-2-pyrrolidone) Blend Mem-
brane for Pervaporation Separation of Methanol/ MTBE
Mixtures,” Separation and Purification Techno- logy, Vol.
64, No. 2, 2008, pp. 183-191.
doi:10.1016/j.seppur.2008.09.013
[17] X. H. Zhang, Q. L. Liu, Y. Xiong, A. Zhu, M. Y. Chen
and Q. G. Zhang, “Pervaporation Dehydration of Ethyl
Acetate/Ethanol/Water Azeotrope Using Chitosan/Poly(Vin-
yl Pyrrolidone) Blend Membranes,” Journal of Membrane
Science, Vol. 327, No. 1-2, 2009, pp. 274-280.
doi:10.1016/j.memsci.2008.11.034
[18] J. Lu, Q. Nguyen, L. Zhou, B. Xu and Z. Ping, “Study of
the Role of Water in the Transport of Water and THF
through Hydrophilic Membranes by Pervaporation,”
Journal of Membrane Science, Vol. 226, No. 1-2, 2003,
pp. 135-143. doi:10.1016/j.memsci.2003.08.016
[19] P. J. Flory, “Principles of Polymer Chemistry,” Cornell
University, Ithaca, New York, 1953.
[20] D. A. Devi, B. Smitha, S. Sridhar and T. M. Aminabhavi,
“Pervaporation Separation of Dimethylformamide/Water
Mixtures through Poly(Vinyl Alcohol)/Poly(Acrylic Acid)
Blend Membranes,” Separation and Purufication Tech-
nology, Vol. 51, No. 1, 2006, pp. 104-111.
doi:10.1016/j.seppur.2006.01.006
[21] S. Sommer and T. Melin, “Influence of Operation Pa-
rameters on the Separation of Mixtures by Pervaporation
and Vapor Permeation with Inorganic Membranes. Part 2:
Purely Organic Systems,” Chemical Engineering and
Science, Vol. 60, No. 16, 2005, pp. 4525-4533.
doi:10.1016/j.ces.2005.02.070
[22] S. Sommer and T. Melin, “Influence of Operation Pa-
rameters on the Separation of Mixtures by Pervaporation
and Vapor Permeation with Inorganic Membranes. Part 1:
Dehydration of Solvents,” Chemical Engineering and
Science, Vol. 60, No. 16, 2005, pp. 4509-4523.
doi:10.1016/j.ces.2005.02.059
[23] J. E. Elshof, C. R. Abadal, J. Sekulic, S. R. Chrowdhury
and D. H. A. Blank, “Transport Mechanisms of Water
and Organic Solvents through Microporous Silica in the
Pervaporation of Binary Liquids,” Microporous Meso-
porous Materials, Vol. 65, No. 2-3, 2003, pp. 197-208.
doi:10.1016/j.micromeso.2003.08.010