Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Chemical Engineering and Processing 50 (2011) 227–235

Contents lists available at ScienceDirect

Chemical Engineering and Processing:


Process Intensification
journal homepage: www.elsevier.com/locate/cep

Review

A review of membrane selection for the dehydration of aqueous ethanol by


pervaporation
Brian Bolto ∗ , Manh Hoang, Zongli Xie
CSIRO Materials Science and Engineering, Private Bag 33, Clayton South MDC, Victoria 3169, Australia

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

Article history: Four broad types of membranes are categorised: organic polymers generally, crosslinked poly(vinyl alco-
Received 12 July 2010 hol), organic–inorganic hybrids and charged polymers. The best performers in terms of flux, which reaches
Received in revised form a maximum of 5 kg/m2 h, are anionic or cationic polymers, including polysalts. Polyanion and polysalt
19 December 2010
membranes are superior. Two examples are thin layers of the active polysalt membrane on a support-
Accepted 11 January 2011
ing membrane. The best combination for flux and selectivity is a polyethyleneimine/poly (acrylic acid)
Available online 28 January 2011
polysalt deposited on a reverse osmosis membrane, at 4 kg/m2 h and 1075 respectively. It is noticeable
that hybrid poly(vinyl alcohol)/inorganic membranes do not show enhanced fluxes. Very high separation
Keywords:
Pervaporation
factors were observed, covering a range of polymers, of neutral, anionic or cationic character. The top
Polar polymers results (>10,000) were for charged membranes, either cationic or anionic, but not polysalts. The fluxes
Polyanions encountered here were miniscule, the best being caesium alginate at about 1 kg/m2 h. The ideal structure
Polycations for high fluxes would appear to be one containing discrete domains of oppositely charged species of
Polysalts optimal size. Fresh approaches are being actively studied, such as layer-by-layer deposition of oppositely
charged polyelectrolytes, with due attention to appropriate separation of the sites of opposite character.

Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
2. Pervaporation membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
2.1. Membranes based on organic polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
2.2. Crosslinked PVA membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
2.3. Hybrid PVA/inorganic polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
2.4. Modified chitosan membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
2.5. Other polymers containing charged groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
2.6. Membranes formed from polysalts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
3. Best performing membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

1. Introduction
tion on one side, while permeate is removed as a vapour from the
Membranes can be used for the selective removal of water from other side [1]. Transport through the membrane is driven by the
aqueous organic mixtures. The knowledge gained from this area vapour pressure difference between the feed solution and the per-
can influence the design of membranes for other water treatment meate vapour. The vapour pressure difference can be maintained
processes. Pervaporation (PV), aimed at the separation of liquid by applying a vacuum on the permeate side, or by cooling the per-
mixtures, involves a membrane that is in contact with the feed solu- meate vapour so that it condenses, thus creating a partial vacuum.
Commercial systems for the dehydration of concentrated alcohol
and other solutions have been developed since the 1980s, much of
∗ Corresponding author. Tel.: +61 3 9545 2037; fax: +61 3 9545 1128. the push coming from interest in the production of pure ethanol
E-mail address: brian.bolto@csiro.au (B. Bolto). as an alternative liquid fuel, where PV can be used to dehydrate

0255-2701/$ – see front matter. Crown Copyright © 2011 Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.cep.2011.01.003
228 B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235

Table 1
PV dehydration of ethanol through various polymeric membranes [4,5].

Polymer Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference

Regenerated cellulose 50 45 5.0 2060 [4,47]


Cellulose acetate 4 60 5.9 200 [4]
Teflon-g-polyvinylpyrrolidone 4 25 2.9 2200 [50]
Perfluorinated polymer on PAN support 1.3 50 387 1650 [5]
Nafion-H+ 4 70 2.5 5000 [36]
Polyacrylonitrile-polyvinylpyrrolidone 4 20 3.2 2200 [51]
Poly(maleimide-co-acrylonitrile) 15 15 33 8 [52]
Poly(acrylic acid-co-acrylonitrile) 18 15 877 13 [53]
Polystyrene 4 40 101 5 [54]
Poly(vinyl chloride) 4 40 63 3 [55]
Alginic acid 4 40 8.8 48 [56]
5 60 13 2800 [37]

Chitosan 4 40 2208 4 [16]


Chitosan acetate salt 4 40 2556 2 [21]
Chitosan/glutaraldehyde 4 40 202 7 [21]
PVA/25% TEOS, annealed at 160 ◦ C 15 40 329 5 [4]
PVA/25% TEOS, annealed at 130 ◦ C 15 40 893 4 [4]

azeotropic ethanol/water mixtures [2]. The 200–500-fold separa- these membranes may be highly crosslinked. This survey of the
tion achieved is due entirely to the selectivity of the membranes different polymer types is somewhat crude as there are crucial
used, which are much more permeable to water than to ethanol. variables that are not taken into account. Only one publication
More than 100 plants for the dehydration of ethanol have been has been found where further variables are addressed, with the
installed, the largest processing 5000 kg/h [1]. data being normalised by the authors to remove the effects of
PV data can be expressed in a number of ways [1]. The bulk different membrane thicknesses and an applied pressure or vac-
of the literature expresses them in terms of the total flux through uum [6]. Crosslinked membranes exhibited less swelling, had a
the membrane and a separation factor. The flux of a component i higher separation factor, but lower water permeability, measured
through a PV membrane can be expressed in terms of the partial as g␮m/m2 h kPa. The results are included in Table 2 and will be
vapour pressures pio and pit on either side of the membrane by discussed further under the heading of charged polymers.
Of course the physical format of the membranes is crucial, but
PG
i there are no data to tell whether any nanoporosity or the like is
Ji = (pio − pit )
l present in the membranes mentioned so far. It may be that the
where Ji is the flux, l is the membrane thickness and PG different heat treatment on crosslinking PVA results in enhanced
i is the
gas separation permeability coefficient. crystallinity, with more pronounced channels on the molecular
The separation factor ˛sep can be defined for a two-component scale being formed between the crystalline regions, thus facilitat-
liquid system as the ratio of the two components on the permeate ing transport. In the choice of appropriate membranes, crosslinked
side divided by the ratio of the two components on the feed side of PVA is pertinent because of its very hydrophilic nature [7]. A wide
the membrane. Thus range of crosslinking agents has been employed in the fabrication
of PVA membranes.
PH2 O /PEtOH
˛sep =
FH2 O /FEtOH

where PH2 O and PEtOH are the mass fractions of water and ethanol 2.2. Crosslinked PVA membranes
in the permeate, and FH2 O and FEtOH are the mass fractions of water
and ethanol in the feed [3]. An amic acid has been used as a crosslinker for PVA [8].
Imidisation at 150 ◦ C gave an improved membrane for the PV of
2. Pervaporation membranes aqueous ethanol when there was 12 wt% crosslinker present. More
crosslinker showed the reverse effect because of the dispersion
2.1. Membranes based on organic polymers of unreacted crosslinker within the membrane. Separation factors
ranged from 70 to 380 and permeation rates from 30 to 1600 g/m2 h
A general survey of the permeation characteristics of a variety at 30–75 ◦ C, depending on the operating temperature and feed
of organic polymers in the treatment of aqueous ethanol solutions mixture composition. At 45 ◦ C the separation factor was 100 and
of low water content was published by an early reviewer in the the permeation rate 200 g/m2 h when the feed contained 30 wt%
field, and is given in Table 1 [4]. It is difficult to get a consis- water (Table 3). As the water concentration increased, the separa-
tent picture, but it would seem that the non-polar polystyrene and tion factor decreased and the permeation rate increased. This was
poly(vinyl chloride) membranes are responsible for moderate sep- explained by the plasticising effect of water, making the amorphous
aration factors and low fluxes. A recent example of a hydrophobic regions of the PVA membrane more swollen, as the polymer chains
active layer of a commercial amorphous perfluorinated polymer on become more flexible. At 75 ◦ C the permeation rate was 500 g/m2 h,
a polyacrylonitrile (PAN) supporting membrane gives, a moderate with the separation factor little changed.
separation factor, but with a much higher flux [5]. However, the Crosslinking of PVA with maleic acid or citric acid has been
very hydrophilic poly(vinyl alcohol)/tetraethoxysilane (PVA/TEOS) used to prepare membranes for permeation studies of water and
and chitosan products also show low flux behaviour. The highest lower aliphatic alcohol mixtures [9]. Citric acid gave more selec-
fluxes are those for the hydrophilic membranes based on cellulose tive behaviour than maleic acid, with the selectivity being highest
and Nafion, and grafts of hydrophilic poly(vinyl pyrrolidone) on for water–isopropanol and water–isobutanol systems. For the
Teflon and polyacrylonitrile. The PVA/TEOS membranes are excep- water–ethanol system the selectivity was 80 and the flux 495 g/m2 h
tions in that they are hydrophilic, but exhibit low fluxes. However, for a 30% aqueous solution at 30 ◦ C.
B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235 229

Table 2
Normalised separation data for PV dehydration of ethanol with modified PVA membranes [6].

Membrane synthesis Feed, wt% Temp., ◦ C Water permeability, Permselectivity Reference


water g␮m/m2 h kPa

Poly(ethylene glycol) blend and TEOS 15 50 0.17 218 [17]


Poly(acrylic acid) 15 40 18 17* [16]
copolymer and TEOS 250
Sulphated zirconia crosslinker 10 50 5.6 80 [7]
Acrylic acid graft 4.4 60 3.0 3.5 [8]
␥-aminopropyl-triethoxysilane 15 50 3.0 63 [3]
Quaternised PVA uncrosslinked 15 50 13 58 [6]
Quaternised PVA and glutaraldeyde 15 50 8.0 75 [6]
*
No poly(ethylene glycol).

Table 3
PV dehydration of ethanol as a function of PVA crosslinking agent.

Crosslinker Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference

None 4.6 75 721 247 [32]


Amic acid to imide 30 45 100 200 [8]
30 75 95 500 [8]

Citric acid 30 30 80 495 [9]


PAA 50 75 60 2800 [11]
PVA/GA containing PAA/EG IPNs 5 50 50 260 [12]
Fumaric acid 20 100 211 1511 [13]
20 80 411 771 [13]
20 60 779 217 [13]

Glutaraldehyde 10 30 180 50 [14]


20 30 150 130 [15]
10 40 104 280 [15]
20 30 – 93 [34]

PVA (MW 50 kDa) has been crosslinked with poly(acrylic acid) or PV of 10 wt% water in ethanol at 40 ◦ C, the fluxes ranging from
PAA (MW 2 kDa) by heat treatment at 150 ◦ C for 1 h [10]. The prod- 249 to 313 g/m2 h and the separation factors from 69 to 108. The
ucts have been used for the PV of methanol/methyl tert-butyl ether details are given in Table 4. The more crosslinked membranes are
mixtures, and later ethanol–water mixtures [11]. For 50 wt% water less hydrophilic as OH groups are consumed in the crosslinking
ethanol solutions the separation factor was 60 at 75 ◦ C and the flux reaction, and the membranes are more rigid, making for greater
2800 g/m2 h. Interpenetrating networks (IPNs) of PVA crosslinked obstruction to diffusion within.
with glutaraldehyde (GA) containing PAA crosslinked with ethy-
lene glycol (EG) have similarly low permeation rates to PVA/GA 2.3. Hybrid PVA/inorganic polymers
membranes, with semi-IPNs giving the better results [12].
Fumaric acid, trans-HO2 C CH CH CO2 H, has been employed Membranes made from PVA crosslinked with 25 wt% of TEOS
as the crosslinking agent at 0.05 mole per mole of PVA in multilayer have been prepared for the PV of aqueous ethanol, with the aim
membranes formed on a polyacrylonitrile support membrane [13]. of minimising the swelling of the PVA [4]. Annealing of the mem-
In the PV of 20 wt% water in ethanol, the separation factors varied branes under nitrogen at temperatures of 100, 130 and 160 ◦ C was
from 779 to 211 over the temperature range 60–100 ◦ C, and fluxes needed to complete the condensation reaction that introduced
from 217 to 1511 g/m2 h. bridging and crosslinking, when higher selectivity resulted. As
The dehydration of ethanol and other alcohol/water mixtures shown in Table 5, the best result, for a 15 wt% water mix in ethanol
has been explored with PVA membranes crosslinked with GA [14]. at 40 ◦ C, was a separation factor of 893 and a flux of 40 g/m2 h when
At 30 ◦ C the flux was 50 g/m2 h for ethanol of 10 wt% water content, the annealing temperature was 130 ◦ C, versus 329 and 50 g/m2 h
and 130 g/m2 h when the water content was 20 wt%. The corre- respectively when annealing was at 160 ◦ C. It was postulated that
sponding separation factors were 180 and 150. The flux increased the crosslinking reaction took place in the non-crystalline parts of
with increasing water content, while the reverse was the case with the PVA membrane, forming denser non-crystalline regions.
the separation factor. The flux also increased with the feed temper- Annealing also improved the selectivity of similar membranes
ature, but the separation factor was lowered. made from poly(vinyl alcohol-co-acrylic acid) [16]. In one example
From the summary of the results obtained for variously the separation factor was 250 for a 15 wt% aqueous ethanol at 40 ◦ C,
crosslinked PVA membranes given in Table 3 it can be seen that and a flux of 18 g/m2 h. Blends of PVA and poly(ethylene glycol) or
the highest fluxes are obtained when the crosslinking agent is a PEG have been crosslinked with TEOS and annealed at 130 ◦ C to
carboxylic acid, which could arise from the ionic component in the
polymer network. The charge is dependent on the pH of the sys- Table 4
tem, so it would be illuminating to see how performance varies with Influence of degree of GA crosslinking of PVA on PV of 10% aqueous ethanol at 40 ◦ C
pH. There is no comparable performance with alginic acid though [15].

(Table 1). Likewise with the ethylenediamine salt of Nafion, or chi- Swelling degree, % Flux, g/m2 h Water selectivity
tosan, except for the GA crosslinked material, the latter having a 28.4 313 69
high flux. 27.9 302 84
Similar GA crosslinked membranes have been reported and 27.1 289 93
their swelling properties explored [15]. The more crosslinked, less 26.5 279 107
24.8 249 108
swollen membranes had lower fluxes but higher selectivities in the
230 B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235

Table 5
PV dehydration of ethanol using PVA/inorganic hybrid membranes.

Crosslinker Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference

TEOS (160 C) 15 40 329 50 [4]
TEOS (130 ◦ C) 15 40 893 40 [4]
PEG blend and TEOS 15 50 300 46 [17]
No PEG 15 50 160 500 [17]
Poly(acrylic acid) copolymer and TEOS 15 40 250 18 [16]
␥-aminopropyl-triethoxysilane 5 50 537 36 [3]
Sulphated zirconia 5 50 263 10 [18]
10 50 142 105 [18]
20 50 86 183 [18]
30 50 61 1036 [18]

produce a membrane with a separation factor of 300 for a 15 wt% Chemical modification of chitosan has been carried out by
water mix in ethanol at 50 ◦ C [17]. The presence of PEG decreased carboxylation, sulphonation or phosphorylation to produce PV
the flux from 500 to 46 g/m2 h, but almost doubled the separation membranes [22]. The phosphorylated products [23] gave the high-
factor (Table 5). est flux in dehydrating ethanol, with a selectivity of 541 and a flux
When ␥-aminopropyltriethoxysilane was used for crosslink- of 180 g/m2 h for mixtures containing 10 wt% water at 70 ◦ C. Other
ing PVA, the hydrophilicity of the membranes increased when the variants such a sulphonation or carboxymethylation had much
silane content was <5%, and the permeation increased remarkably higher selectivities, but lower fluxes. It was proposed that phos-
while the selectivity increased at the same time, thus breaking the phorylation formed phosphates at the C-6 position in the glucose
trade-off between the two [3]. With a feed containing 5 wt% water units, reducing the chain relaxation markedly because of the bulky
the separation factor was 537 and the flux 36 g/m2 h at 50 ◦ C. side chain and crosslinking. A less packed structure would make the
PVA has been crosslinked with the solid acid of sulphated zir- membrane more relaxed and mobile, allowing passage of ethanol
conia by an acid-catalysed reaction which affected the degree of and resulting in a lower selectivity.
swelling and the crosslinking density of the membrane [18]. For the Blends of chitosan with hydroxyethylcellulose that have been
same duty, both flux and selectivity were lower than for the silane crosslinked with a urea–formaldehyde–sulphuric acid mixture pro-
products (Table 5). The flux increased with feed water content, duce PV membranes with the very high selectivity of 10,491 when
while the selectivity decreased. the hydroxyethylcellulose content is 50 wt% [24]. The addition of
Overall, by comparison with the results for organically the blending agent increases the hydrophilicity of the membranes,
crosslinked PVA membranes (Table 4), the data in Table 5 show and since the hydroxyethyl group is larger than the amino group in
that there is no advantage in introducing inorganic particulates into chitosan, there is an increased flexibility and openness in the poly-
PVA. The story is quite different for totally inorganic membranes mer network. The fluxes measured were many times those of the
such as those based on zeolites, which have been reviewed recently unmodified chitosan membranes.
[19]. They are beyond the scope of the present paper. An LTA zeo- Likewise, in the preparation of PV membranes for
lite membrane can give extremely high fluxes of 12, 15, 18 and ethanol dehydration, chitosan has been crosslinked with 3-
22 kg/m2 h at 100,110, 120 and 130 ◦ C respectively in dehydrating aminopropyltriethoxysilane to form chitosan–silica hybrid
a 93 wt% ethanol mixture. membranes [25]. The hydrophilicity of the membranes increased
with increasing amounts of the crosslinker, reaching a maximum
when the content was 10 wt%. The flux and separation factor
2.4. Modified chitosan membranes
increased as a result, compared to the unmodified chitosan mem-
brane. The crosslinking limited the swelling of the polymer. It was
Another highly hydroxylated polymer, chitosan, or poly(2-
claimed that there were strong hydrogen bonds as well as covalent
amino-2-deoxy-d-glucose), has been used to make PV membranes.
bonds formed, along with a lowering of crystallinity. Thus the
It has a cationic charge when the amino group is protonated. At neu-
FTIR spectrum showed an intense absorption band at ∼1072 cm−1 ,
tral pH levels it is slightly charged with 17% of the amino groups
indicating that esterification took place between the OH groups
being protonated [20]. A chitosan acetate membrane has more
in the polymer and the silanol groups in the crosslinker, to form
than 10 times the separation factor of chitosan itself, as does a GA
C O Si bonds. If the second ethoxy group reacts with the OH
crosslinked chitosan [21], as detailed in Table 6. Very low flux lev-
in another chain, bridges can occur. The crosslinking makes the
els were obtained with all these membranes for feeds of low water
polymer chains in the amorphous regions more compact, resulting
contents.

Table 6
PV dehydration of ethanol using chitosan-based membranes.

Chitosan product Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference

GA crosslinked 4 40 2208 4 [21]


Uncrosslinked 202 7 [21]
Acetate salt 2556 2 [21]
Phosphorylated 10 70 541 180 [22,23]
Acetate salt 242 142 [22,23]
Sulphonated & GA 1560 52 [22,23]
Carboxymethylated 1294 36 [22,23]
Carboxyethylated 301 30 [22,23]
Cyanoethylated 52 80 [22,23]
Hydroxyethylcellulose 50% blend 10 60 10,491 112 [24]
10% 3-aminopropyl-triethoxysilane 15 50 597 887 [25]
Uncrosslinked 15 50 200 275 [25]
B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235 231

Table 7
PV dehydration of aqueous ethanol with chitosan membranes in various forms [26].

Chitosan format Feed, wt% water Temp., ◦ C Sepn. Factor Flux, g/m2 h

99% deacetylated 5 70 5.4 840


99% deacetylated 4.4 40 17 65
73% deacetylated 8 35 1240 6
Crosslinked with GA 10–50 40 >2000 –
98% deacetylated-HOAc 10 60 123 400
98% deacetylated-HCl 10 60 593 210
98% deacetylated-H2 SO4 10 60 1040 160
98% deacetylated-non-ionised 10 60 36 350

Table 8
PV dehydration of 20% aqueous ethanol at 70 ◦ C using graft polymer membranes having different charges [27].

Host polymer Grafted polymer Graft site charge Separation factor Flux, g/m2 h

Polyvinylidene fluoride 4-vinylpyridine Neutral 9 200


Polyvinylidene fluoride N-vinyl-imidazole Neutral 10 1600
Polyvinyl fluoride N-vinylmethyl- acetamide Neutral 4 2900
Polyvinyl fluoride N-vinyl-pyrrolidone Neutral 7 1000
Polyacrylonitrile Acrylic acid Neutral 10 1800
Polyacrylonitrile K+ acrylate Anionic 500 3000
Polyvinyl fluoride K+ acrylate Anionic 156 4700
Polyvinylidene fluoride Quaternised 4-vinylpyridine Cationic 175 400
Polyvinylidene fluoride Quaternised N-vinylimidazole Cationic 61 1800
Polyvinylidene fluoride 4-vinylpyridine/BrCH2 COOH Zwitterionic 76 200
Polyvinyl fluoride Vinylimidazole/BrCH2 COOH Zwitterionic 63 1250

in less space for species to permeate through the membrane and acid [32]. Performance depended on the degree of substitution
making the resistance to flow higher for the larger species. (DS), the separation factor increasing with an increase in the DS for
Research into various forms of chitosan has shown that a lighter the cationic membrane, to 1540 for the higher DS, but decreasing
level of deacetylation of the original chitin is beneficial as regards in the case of the anionic membrane. For the cationic membrane
the separation factor, as is crosslinking of chitosan with GA, or hav- both selectivity and flux were improved by a higher DS, which
ing the amino groups in the salt form when the anion involved is is unusual. The highest flux was still not impressive though, and
divalent, as shown in Table 7 [26]. However, the fluxes obtained levels were higher for the anionic than the cationic membrane.
were all low. Cationic and anionic PVA membranes have been made also via
copolymers with acrylamidopropyltrimethylammonium chloride
2.5. Other polymers containing charged groups or with itaconic acid [15]. The degree of swelling decreased from
neutral to cationic to anionic PVA with values of 26.5%, 15.7% and
The use of hydrophilic membranes for the dehydration of 14.0% respectively. Here the anionic PVA gave better performance
organic solvents in general by PV has been the subject of compre- than the cationic derivative, in contrast with the other examples in
hensive reviews [27,28]. The emphasis is on selectivity, postulated Table 9. The anionic membrane had 2.5 times the separation factor
to be determined by selective sorption and selective diffusion. of the neutral PVA, at 837 versus 335 for one duty. The flux was
Selective sorption is governed by the presence in the membrane correspondingly lower (86 g/m2 h versus 189 g/m2 h).
of active centres such as charged sites which are capable of spe- The plasma modification of GA crosslinked PVA membranes
cific interaction with water, while selective diffusion is governed by grafting acrylic acid onto them gave membranes which had
by the rigidity and regularity of the polymer structure and the enhanced hydrophilicity [33]. The fluxes ranged from 122 to
nature of the polymer interspace, exemplified by the degree of 144 g/m2 h for 96 wt% ethanol at 25–60 ◦ C, and the separation fac-
swelling and the frequency of the crosslinks. The results for a series tors from 23 to 8. The respective results for unmodified membranes
of membranes made by grafting neutral or charged polymers onto were 75–120 g/m2 h and 19–10. Other work on PVA crosslinked
supporting membranes are given in Table 8. A neutral amide graft with GA reports fluxes of up to 93 g/m2 h for the PV of ethanol
and an anionic acrylate graft produced the highest moderate fluxes. containing 20 wt% water at 30 ◦ C [34]. Incorporating sericin, the
Charged membranes formed by crosslinking PVA with 7 wt% of gelatinous protein that binds the fibres in silk, into PVA that is
a sulphosuccinic acid yields membranes that in the separation of crosslinked with dimethylolurea gives a membrane that is only
10/90 wt% water/ethanol at 60 ◦ C have a separation factor of 240 moderately selective to water [35]. The crosslinker reacts with the
and a flux of 180 g/m2 h [29]. At 70 ◦ C the results are 175 and OH groups of the polymers leaving the amino groups in sericin free
300 g/m2 h respectively. When the lithium salt of the crosslinker to interact with water.
was prepared, it showed lower results of 44 and 59 g/m2 h, but at As shown in Table 2, a quaternised PVA, made by grafting
50 ◦ C and with 5 wt% crosslinker, as shown in Table 9 [30]. 2,3-epoxypropyltrimethylammmonium chloride onto the polymer
Quaternised poly(acrylonitrile-co-3-N ,N -dimethylaminopro- and then crosslinking with GA has been tested [6]. The quater-
pylacrylamide) has been used as a PV membrane for dehydrating nary ammonium groups reduced the crystallinity of the polymer
ethanol [31]. The selectivity factor increased with decreasing lev- and enhanced the hydrophilicity and water permselectivity. The
els of water in the feed solution, reaching over 15,000 for the fully results are still modest: a membrane with 4% quaternisation
quaternised polymer at 30 ◦ C when the water content was 10 wt%. had a water permselectivity of 58 and a water permeability of
However, the flux was then only 10 g/m2 h. 13 g␮m/m2 h kPa for 15% aqueous ethanol at 50 ◦ C. The data here
Cationic and anionic PVA membranes have been have been normalised by the authors to remove the effects of
made by reacting PVA with a catalyst and 3-chloro-2- different membrane thickness and applied pressure or vacuum.
hydroxypropyltrimethylammonium chloride or phosphoric Crosslinked membranes exhibited less swelling, and had a higher
232 B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235

Table 9
PV dehydration of ethanol with membranes containing charged groups.

Polymer and charged species Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference

PVA/7 wt% sulphosuccinic acid 10 70 175 300 [29]


10 60 240 180 [29]

PVA/5 wt% Li+ sulphosuccinate 10 50 44 59 [30]


Quaternised poly(acrylo-nitrile-co-3-N ,N - 10 30 >15,000 10 [31]
dimethyl-aminopropyl
acrylamide)
Cationic PVA* 4.6 75 424a 216 [32]
4.6 75 1540b 249 [32]

Anionic PVA* 4.6 75 398c 632 [32]


4.6 75 72d 407 [32]

Unmodified PVA 4.6 75 721 247 [32]


Cationic PVA/GA† 4 40 709(216) 89(123) [15]
Anionic PVA/GA† 4 40 837(228) 86(119) [15]
PVA/GA† 4 40 335(107) 189(279) [15]
PVA/GA acrylic acid 4 40 14 135 [33]
Plasma graft 4 60 8.4 144 [33]
Unmodified PVA 4 40 15 91 [33]
4 60 10 120 [33]

PVA/9% acrylic acid graft 10 15 800 7 [27]


20 15 700 10 [27]

PVA/sericin blend 8.5 50 90 70 [35]


*
DS: a, 2.9%; b, 5.2%; c, 2.3%; d, 5.0%.

The data in parentheses are for a down stream pressure of 0.4 kPa; the rest are for a down stream pressure of 1.4 kPa.

water permselectivity of 75, but a lower water permeability of on selectivity (Table 10). Mixing two polymers of opposite charge
8.0 g␮m/m2 h kPa. like this forms a polysalt.
Nafion sulphonic acid membranes have been tested for the influ-
ence of the counterion on PV performance [36]. The data (Table 10) 2.6. Membranes formed from polysalts
show that the acid form has a much greater flux than the various salt
forms, no doubt because of the less space occupied by the proton It has been speculated that polysalts, formed from anionic and
versus the cations studied. cationic polyelectrolytes, would be appropriate for obtaining both
Alginic acid has been tested for PV behaviour [37]. It was found highly permeable and highly selective membranes [27]. Earlier
that the selectivity factor increased enormously when the acid work on PV with membranes formed from sodium polyacrylate and
was neutralised with alkali, from 19 to over 10,000, while the various cationic polyelectrolytes [42,43] is summarised in Table 11.
flux decreased markedly, with the acid form again exhibiting by Having a quaternary ammonium structure in the backbone
far the highest flux. The counter ion influenced the flux in the of the cationic polymer appears to be advantageous as regards
order Na+ < K+ < Rb+ < Cs+ < H+ , which was ascribed to a change in selectivity, but less so for the epichlorohydrin/dimethylamine
the conformation of the alginate molecule so that its mobility was (ECH/DMA) polymer, where the presence of an OH group may
increased, as was the mobility of the ion hydration shell and the be a handicap. The flux levels are again not impressive. Several
crystallinity of the membrane. The conformational change was ver- membranes based on polysalts have been explored, as detailed in
ified by 13 C NMR. A blend of sodium alginate with PVA crosslinked Table 12. The highest flux results were from structures based on
with 1.5 wt% maleic acid had a lower separation factor than the cellulose sulphate as the polyanion [44], plus one based on sodium
membrane without the PVA [38]. carboxymethyl cellulose, or Na+ CMC, and chitosan [45] or poly(N-
Ion exchange membranes in the form of styrene-based qua- ethyl-4-vinylpyridinium bromide) [46].
ternary ammonium and sulphonated polyethylene containing Other physical structures affecting membrane behaviour have
different counter ions have been tested, to give the results shown in been examined as ways to enhance the flux. Thus a porous poly-
Table 10 [2]. The sulphonic acid membrane had a much higher flux sulphone membrane has been coated with a polysalt formed from
than the sodium salt, by a factor of several hundred, somewhat sim- chitosan and poly(acrylic acid), and the composite membrane
ilar to the Nafion and alginic acid experiences. The selectivity was tested for PV of aqueous ethanol [47]. For dehydration of 5 wt%
correspondingly lower. Diffusion of liquid water through the con- aqueous ethanol at 30 ◦ C, when compared with the dense homoge-
tinuous pathway of water shells around charged groups is claimed neous membrane that is unsupported, the composite version had a
to be faster than diffusion through clusters of free water in the flux of 132 g/m2 h versus 33 g/m2 h and a separation factor of 1008
membrane [39]. versus 2216 (Table 12). The composite membrane was only supe-
Membranes have been made by modifying poly(vinyl rior for solutions of high ethanol content. For all membranes both
chloride) or PVC membranes by reacting them with parameters increased with heating for feeds of high water con-
sodium N-methyldithiocarbamate or sodium N-methyl-N- tent (>30 wt%), which is unusual. This was ascribed to the strong
carboxydithiocarbamate, and then with metal ions such as Li+ , Ca++ sorption contribution to the activation energy for PV with these
and Cu++ [40]. The Cu++ complex showed an extremely high selec- membranes. For optimal performance the ratio of membrane thick-
tivity of >5000 in the dehydration of ethanol at high water contents ness to the permeation characteristics of the dense layer and the
(80 wt%). Various salt forms of sulphonated polyamide membranes porous supporting membrane had to be appropriate. One way of
have been tested [41]. There was little variation in selectivity, but achieving this is to use a layer-by-layer deposition method, which
the flux decreased in the order PEI+ > K+ > NEt4 + > Cs+ > Na+ , where allows one to choose which thickness of active membrane skin is
PEI is polyethyleneimine, which had by far the greatest influence the best.
B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235 233

Table 10
Effect of counter ion on PV dehydration of ethanol with membranes containing charged groups.

Polymer and charged species Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference
+
Nafion-H 4 70 2.5 5000 [36]
Nafion-Na+ 4 70 5.0 500 [36]
Nafion-K+ 4 70 9.8 200 [36]
Nafion-N(CH3 )3 + 4 70 10.5 340 [36]
Alginic acid 10 60 19 1300 [37]
Li+ alginate 10 60 11,400 100 [37]
Na+ alginate 10 60 11,600 210 [37]
K+ alginate 10 60 11,600 550 [37]
Rb+ alginate 10 60 11,600 750 [37]
Cs+ alginate 10 60 10,000 1030 [37]
Na+ alginate-PVA blend 10 45 380 380 [38]
Anion exchanger-SCN− 15 60 38 470 [2]
Sulphonated PE-H+ 16 26 2.6 1360 [2]
Sulphonated PE-Na+ 16 26 671 80 [2]
PVC/dithiocarbamate-Cu++ 40 25 63 20 [40]
80 25 >5000 103 [40]

Polyamide sulphonate-Na+ 20 60 1.3 160 [41]


Polyamide sulphonate-K+ 20 60 2.1 230 [41]
Polyamide sulphonate-Cs+ 20 60 1.4 11 [41]
Polyamide sulphonate-NEt4 + 20 60 1.5 18 [41]
Polyamide sulphonate-PEI+ 20 60 11 14 [41]

Table 11
PV dehydration of 5% aqueous ethanol at 70 ◦ C with membranes based on polysalts where the polyanion is sodium polyacrylate [42,43].

Polycation Carbon backbone Charged group Charged site Separation factor Flux, g/m2 h
+
Polyethyleneimine –CH2 CH2 – –NH2 – Backbone 220 830
Polyallylamine –CH2 CH2 – –CH2 NH3 + Pendant 750 510
Ethylene/phenylene dimethylammonium Ditto & –CH2 C6 H4 CH2 – –N+ (CH3 )2 – Backbone 1710 790
Propylenedimethyl-ammonium –(CH2 )3 – –N+ (CH3 )2 – Backbone 1940 820
ECH/DMA polymer –CH2 CH(OH)CH2 – –N+ (CH3 )2 – Backbone 830 340
Trimethylammoniumethyl methacrylate –CH2 C(CH3 )– –N+ (CH3 )3 Pendant 380 220

Table 12
PV dehydration of aqueous ethanol with membranes based on various polysalts.

Polyanion Polycation Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference

Poly(acrylonitrile-co-acrylic acid) Poly(acrylonitrile-co-vinyl pyridine) 10 – 5000 400 [44]


Cellulose-SO3 − Na+ Polyethyleneimine 16 50 295 1900 [45]
Cellulose-SO3 − Na+ PolyDADMAC, linear 16 50 140 3200 [45]
Cellulose-SO3 − Na+ Same, but branched 16 50 123 4900 [45]
Cellulose-SO3 − Na+ Poly-N,N-dimethyl-3,5- 16 50 123 2700 [45]
dimethylenepiperidine
chloride
Aromatic polyamide sulphonate Polyethyleneimine 20 60 15 300 [41]
Poly(acrylic acid) Chitosan
On polysulphone Supporting membrane 5 30 1008 132 [47]
No supporting Membrane 5 30 2216 33 [47]

Na+ polystyrene sulphonate Polyallylamine.HCl 6.2 70 70 230 [57]


Na+ CMC Chitosan 10 70 1062 1140 [45]
Na+ CMC N-ethyl-4-vinyl-pyridinium bromide 10 70 782 1320 [46]
Anionic PVA, Cationic PVA,
DS 2.3% DS 2.9% 4.6 75 2250 378 [32]
DS 5.0% DS 5.2% 4.6 75 1910 284 [32]

3. Best performing membranes hybrid PVA/inorganic membranes do not show enhanced fluxes,
the best being PVA/sulphated zirconia, at just over 1 kg/m2 h, so
The best performers in terms of flux, which at a maximum of there is no great advantage in incorporating inorganic fillers within
5 kg/m2 h never achieve high values, are charged polymers of one the membrane. External factors enhancing the flux include heat-
type or another, including polysalts. Table 13 shows that anionic ing, and an increased water content, which follows from greater
and polysalt membranes are superior. For anionic polymers, the swelling and a low level of crosslinking. However, crosslinking can
proton form has a significantly higher flux than the metal or qua- occur while maintaining the swollen state. It is very difficult to com-
ternary ammonium salt versions, owing to the greater free space pare different membranes in terms of fluxes and separation factors
within the polymer network. Two of the polysalt examples involve as these parameters are not only a function of the intrinsic proper-
thin layers of the active membrane on an RO or UF supporting mem- ties of the membranes themselves, but rely also on the operating
brane. They also exhibit the highest separation factors. The best conditions such as temperature permeate pressure and the feed
combination for flux and selectivity is PEI/PAA on an RO membrane, concentration [48,49]. So for membranes tested under different
at 4 kg/m2 h and 1075 respectively. It is interesting to note that conditions, caution needs to be applied in attempting comparisons.
234 B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235

Table 13
Highest fluxes for PV dehydration of aqueous ethanol.

Membrane polymer Mem. type Feed, wt% water Temp., ◦ C Flux, g/m2 h Separation factor Reference
+
Nafion-H Anionic 4 70 5000 2.5 [36]
Cellulose-SO3 − Na+ and polyDADMAC, branched Polysalt 16 50 4900 125 [44]
K+ acrylate graft on poly(vinyl fluoride) Anionic 20 70 4700 156 [27]
PEI/PAA on RO membrane Polysalt 10* 70 4050 1075 [58]
Cellulose-SO3 − Na+ and polyDADMAC, linear Polysalt 16 50 3200 140 [44]
K+ acrylate graft on PAN Anionic 20 70 3000 500 [27]
PolyDADMAC/Na+ CMC on UF membrane Polysalt 10* 75 3000 960 [59]
*
Aqueous isopropanol, not ethanol.

Table 14
Highest separation factors for PV dehydration of aqueous ethanol.

Membrane polymer Mem. type Feed, wt% water Temp., ◦ C Separation factor Flux, g/m2 h Reference

Quaternised poly(acrylo-nitrile-co-3-N ,N - Cationic amide 10 30 >15,000 10 [31]


dimethyl-aminopropyl
acrylamide)
Metal alginate salts (Na+ , K+ , Rb+ , Li+ or Cs+ ) Anionic 10 60 11,600–10,000 100–1030 [37]
Chitosan/50% blend with hydroxyethylcellulose Cationic 10 60 10,491 112 [24]
PVC/dithiocarbamate-Cu++ Anionic 80 25 >5000 103 [40]
Chitosan acetate Cationic 4 40 2556 2 [57]
PVA crosslinked with GA Neutral 10–50 40 >2000 – [26]

A range of very high separation factors is shown in Table 14 Hence expenditure on thermal energy will improve the water
for the top selective performers. A variety of polymer types are flux. Cheaper thermal sources that could be exploited are solar,
included, of neutral, anionic or cationic character. The top results geothermal and industrial waste heat.
(>10,000) are for charged membranes again, either cationic or A related process is vapour permeation, which employs a poly-
anionic, but not polysalts. The fluxes encountered here are minis- meric a membrane as a semi-permeable barrier between a feed
cule, the best being caesium alginate at about 1 kg/m2 h. External side under high pressure and a permeate side under low pressure
factors that improve the flux, such as heating and increased water [60]. It differs from PV in that the feed is a vapour and not a liq-
content, have detrimental effects on the separation factor. uid, so that there is no phase change or significant temperature
difference across the membrane. The membranes work according
4. Conclusions to a solution–diffusion mechanism. Separation is achieved by the
different degrees to which components are dissolved in and dif-
The ideal structure for good fluxes would appear to be one fuse through the membrane. The material used in the membrane
containing discrete domains of oppositely charged species of opti- will hence depend on the types of compounds being separated – a
mal size, such that internal neutralisation is minimised. Possible variety of organic and inorganic membranes can be used [61,62].
solutions could be a tightly crosslinked non-polar host polymer The driving force can be approximated to the difference in partial
containing charged nanoparticles, or a crosslinked block copolymer pressures of the components in the feed.
that has dual substitution. The influence of a number of parameters
needs to be determined, such as the minimum desirable size of the References
charge domain, the distance between oppositely charged sites, and
[1] R.W. Baker, Pervaporation Membrane Technology and Applications, second ed.,
the role of counter ions. Fresh approaches are needed; one such Wiley, New York, 2004, pp. 355–392.
could be layer-by-layer deposition of oppositely charged polyelec- [2] G.H. Koops, C.A. Smolders, Estimation and evaluation of polymeric materials for
trolytes, with due attention to appropriate separation of the sites of pervaporation membranes, in: R.Y.M. Huang (Ed.), Pervaporation Membrane
Separation Processes, Elsevier, Amsterdam, 1991, pp. 253–278.
opposite character. Imperfect matching of the oppositely charged [3] Q.G. Zhang, Q.L. Liu, Z.Y. Jiang, Y. Chen, Anti-trade-off in dehydration of ethanol
centres would create an open structure. A greater understanding of by novel PVA/APTEOS hybrid membranes, Journal of Membrane Science 287
the free volume within charge mosaic membranes, as provided for (2007) 237–245.
[4] T. Uragami, K. Okazaki, H. Matsugi, T. Miyata, Structure and permeation charac-
example by positron annihilation lifetime spectroscopy, is needed. teristics of an aqueous ethanol solution of organic-inorganic hybrid membranes
Better insight into the detailed crosslinked structure of the mem- composed of poly(vinyl alcohol) and tetraethoxysilane, Macromolecules 35
branes could be provided by using a synchrotron light source. (2002) 9156–9163.
[5] V. Smuleac, J. Wu, S. Nemser, S. Majumdar, D. Bhattacharyya, Novel
As well as the water content of the aqueous ethanol feed, tem-
perfluorinated polymer-based pervaporation membranes for the separa-
perature is a crucial parameter, the benefit of which is attributed tion of solvent/water mixtures, Journal of Membrane Science 352 (2010)
to the increase in diffusivity and reduction in flow viscosity that 41–49.
[6] Q.G. Zhang, Q.L. Liu, A.M. Zhu, Y. Xiong, L. Ren, Pervaporation performance of
occurs on heating. Mechanical energy in the form of extra applied
quaternized poly(vinyl alcohol) and its crosslinked membranes for the dehy-
pressure or vacuum can likewise be called upon to enhance the flux. dration of ethanol, Journal of Membrane Science 335 (2009) 68–75.
The thickness of the membrane is another vital factor: the active [7] B.A. Bolto, M. Hoang, T. Tran, Crosslinked poly(vinyl alcohol) membranes,
layer should be as thin as possible. The inherent permeability of the Progress in Polymer Science 34 (2009) 969–981.
[8] R.Y.M. Huang, C.K. Yeom, Pervaporation separation of aqueous mixtures using
membrane polymer is similarly important. The main parameters crosslinked poly(vinyl alcohol). II. Permeation of ethanol-water mixtures, Jour-
for better water flux are therefore nal of Membrane Science 51 (1990) 273–292.
[9] M.C. Burshe, S.B. Sawant, J.B. Joshi, V.G. Pangarkar, Sorption and permeation
of binary water-alcohol systems through PVA membranes crosslinked with
• Higher feed water temperature multifunctional crosslinking agents, Separation and Purification Technology
• Increased pressure/vacuum 12 (1997) 145–156.
• [10] J.-W. Rhim, Y.-K. Kim, Pervaporation separation of MTBE-methanol mixtures
Thinner membranes
using crosslinked PVA membranes, Journal of Applied Polymer Science 7 (2000)
• Improved membrane permeability 1699–1707.
B. Bolto et al. / Chemical Engineering and Processing 50 (2011) 227–235 235

[11] K.-H. Lee, H.-K. Kim, J.-W. Rhim, Pervaporation separation of binary organic- Proceedings of the 5th International Conference on Pervaporation Processes in
aqueous liquid mixtures using crosslinked PVA membranes. III. Ethanol-water the Chemical Industry, Bakish Materials Corporation, Engelwood, New Jersey,
mixtures, Journal of Applied Polymer Science 58 (2003) 1707–1712. 1991, pp. 98–108.
[12] E. Ruckenstein, L. Liang, Poly(acrylic acid)-poly(vinyl alcohol) semi- and inter- [41] Y.E. Kirsh, P.A. Vdonin, Y.A. Fedotov, S.I. Semenova, Pervaporation mem-
penetrating polymer network pervaporation membranes, Journal of Applied branes for dehydration of organic solvents based on sulfonate containing
Polymer Science 62 (1996) 973–987. aromatic polyamides: effects of configurational structure and counterion
[13] Z. Huang, H.-M. Guan, W.-L. Tan, X.-Y. Qiao, S. Kulprathipanja, Pervaporation nature on their selectivity and permeability, in: Proceedings of the Interna-
study of aqueous ethanol solution through zeolite-incorporated multilayer tional Congress on Membranes and Membrane Processes, Yokohama, 1996,
poly(vinyl alcohol) membranes: effect of zeolites, Journal of Membrane Science pp. 432–433.
276 (2006) 260–271. [42] H. Karakane, M. Tsuyumoto, Y. Maeda, K. Satoh, Z. Honda, Separation of water-
[14] C.-K. Yeom, S.H. Lee, J.M. Lee, Pervaporative permeations of homologous series ethanol by pervaporation through polyelectrolyte complex membrane, in: R.
of alcohol aqueous mixtures through a hydrophilic membrane, Journal of Bakish (Ed.), Proceedings of the 3rd International Conference on Pervaporation
Applied Polymer Science 79 (2001) 703–713. Processes in the Chemical Industry, Bakish Materials Corporation, Engelwood,
[15] V.S. Praptowidodo, Influence of swelling on water transport through PVA-based New Jersey, 1988, pp. 194–202.
membrane, Journal of Molecular Structure 739 (2005) 207–212. [43] H. Karakane, M. Tsuyumoto, Y. Maeda, Z. Honda, Separation of water-ethanol
[16] T. Uragami, H. Matsugi, T. Miyata, Pervaporation characteristics of organic- by pervaporation through polyion complex composite membrane, Journal of
inorganic hybrid membranes composed of poly(vinyl alcohol-co-acrylic acid) Applied Polymer Science 42 (1991) 3229–3239.
and tetraethoxysilane for water ethanol separation, Macromolecules 38 (2005) [44] W.H. Jo, Y.S. Kang, H.J. Kim, Association behaviour in copolymer blend mem-
8440–8446. branes and the pervaporation of water/ethanol mixtures, Journal of Membrane
[17] L.Y. Ye, Q.L. Liu, Q.G. Zhang, A.M. Zhu, G.B. Zhou, Pervaporation characteris- Science 85 (1993) 81–88.
tics and structure of poly(vinyl alcohol/poly(ethylene glycol)/tetraethoxysilane [45] Q. Zhao, J. Qian, Q. An, C. Gao, Z. Gui, H. Gin, Synthesis and characterisation
hybrid membranes, Journal of Applied Polymer Science 105 (2007) 3640–3648. of soluble chitosan/sodium carboxymethylcellulose polyelectrolyte complexes
[18] K.-J. Kim, S.-H. Park, W.-W. So, S.-J. Moon, Pervaporation separation of aque- and the pervaporation dehydration of their homogeneous membranes, Journal
ous organic mixtures through sulfated zirconia-poly(vinyl alcohol) membrane, of Membrane Science 333 (2009) 68–78.
Journal of Applied Polymer Science 79 (2001) 1450–1455. [46] H. Jin, Q. An, G. Zhao, J. Qian, M. Zhu, Pervaporation dehydration of ethanol
[19] J. Caro, M. Noack, Zeolite membranes – recent developments and progress, by using polyelectrolyte complex membranes based on poly(N-ethyl-4-
Microporous and Mesoporous Materials 115 (2008) 215–634. vinylpyridinium bromide) and sodium carboxymethyl cellulose, Journal of
[20] R.A.A. Muzzarelli, Chitin, Pergamon Press, Oxford, 1977, p. 95. Membrane Science 347 (2010) 183–192.
[21] T. Uragami, K. Takigawa, Permeation and separation characteristics of ethanol- [47] J.-J. Shieh, R.Y.M. Huang, Pervaporation with chitosan membranes II. Blend
water mixtures through chitosan derivative membranes by pervaporation and membranes of chitosan and poly(acrylic acid) and comparison of homogeneous
evapomeation, Polymer 31 (1990) 668–672. and composite membrane based on polyelectrolyte complexes of chitosan
[22] Y.M. Lee, Modified chitosan membranes for pervaporation, Desalination 90 and poly(acrylic acid) for the separation of ethanol-water mixtures, Journal
(1993) 277–290. of Membrane Science 127 (1997) 185–202.
[23] Y.M. Lee, E.M. Shin, Pervaporation separation of water-ethanol through modi- [48] R.W. Baker, J.G. Wijmans, Y. Huang, Permeability, permeance and selectiv-
fied chitosan membranes. IV. Phosphorylated chitosan membranes, Journal of ity: a preferred way of reporting pervaporation performance data, Journal of
Membrane Science 64 (1991) 145–152. Membrane Science 348 (2010) 346–352.
[24] A. Chanachai, R. Jiraratananon, D. Uttapap, G.Y. Moon, W.A. Anderson, R.Y.M. [49] E.G Heisler, A.S. Hunter, J. Siciliano, R.H. Treadway, Solvent and temperature
Huang, Pervaporation with chitosan/hydroxyethylcellulose blended mem- effects in the pervaporation of aqueous alcoholic solution, Science 124 (1956)
branes, Journal of Membrane Science 166 (2000) 271–280. 77–79.
[25] J.H. Chen, Q.L. Liu, X.H. Zhang, Q.G. Zhang, Pervaporation and characterisation [50] P. Aptel, N. Challard, J. Cuny, J. Neel, Application of the pervaporation pro-
of chitosan membranes cross-linked by 3-aminopropyltriethoxysilane, Journal cess to separate azeotropic mixtures, Journal of Membrane Science 1 (1976)
of Membrane Science 292 (2007) 125–132. 271–287.
[26] Y. Maeda, M. Kai, Recent progress in pervaporation membranes for [51] Q.T. Nguyen, L. Le Blanc, J. Neel, Preparation of membranes from
water/ethanol separation, in: R.Y.M. Huang (Ed.), Pervaporation Membrane polyacrylonitrile-polyvinylpyrrolidone blends and the study of their behaviour
Separation Processes, Elsevier, Amsterdam, 1991, Chapter 9. in the pervaporation of water-organic liquid mixtures, Journal of Membrane
[27] S.I. Semenova, H. Ohya, K. Soontarapa, Hydrophilic membranes for pervapora- Science 22 (1985) 245–255.
tion: an analytical review, Desalination 110 (1997) 251–286. [52] M. Yoshikawa, H. Yokoi, K. Sanui, N. Ogata, Selective separation of water-alcohol
[28] P.D. Chapman, T. Oliveira, A.G. Livingston, K. Li, Membranes for the dehydration binary mixture through poly(maleimide-co-acrylonitrile) membrane, Journal
of solvents by pervaporation, Journal of Membrane Science 318 (2008) 5–37. of Polymer Science 22 (1984) 2159–2168.
[29] J.-W. Rhim, C.-K. Yeom, S.-Y. Kim, Modification of poly(vinyl alcohol) [53] M. Yoshikawa, T. Yukoshi, K. Sanui, N. Ogata, Separation of water and ethanol by
membranes using sulfur-succinic acid and its application to pervaporation pervaporation through poly(acrylic acid-co-acrylonitrile) membrane, Journal
separation of water-alcohol mixtures, Journal of Applied Polymer Science 68 of Polymer Science: Polymer Letters Edition 22 (1984) 473–475.
(1998) 1717–1723. [54] T. Uragami, T. Morikawa, Studies on synthesis and permeabilities of special
[30] J.-W. Rhim, S.-W. Lee, Y.-W. Kim, Pervaporation separation of water-ethanol polymer membranes, 70. Permeation and separation characteristics for aque-
mixtures using metal ion-exchanged poly(vinyl alcohol/sulfosuccinic acid ous alcohol solutions by evapomeation and pervaporation through polystyrene
membranes, Journal of Applied Polymer Science 85 (2002) 1867–1873. membranes, Makromolecular Chemie 190 (1989) 399–404.
[31] M. Yoshikawa, S. Ochiai, M. Tanigaki, W. Eguchi, Application and develop- [55] T. Uragami, T. Morikawa, H. Okuno, Characteristics of permeation and separa-
ment of synthetic polymer membranes, 4. Relation between permselectivity for tion of aqueous alcohol solutions through hydrophobic polymer membranes,
aqueous ethanol solution and state of water in synthetic polymer membranes, Polymer 30 (1989) 1117–1122.
Makromolecular Chemie, Rapid Communications 9 (1998) 559–564. [56] T. Uragami, M. Saito, Studies on synthesis and permeabilties of special
[32] B. Sun, J. Zou, Poly(vinyl alcohol)-based polyelectrolyte pervaporation mem- polymer membranes. 68. Analysis of permeation and separation charac-
branes, Annals of New York Academy of Science 984 (2003) 386–400. teristics and new technique for separation of aqueous alcoholic solutions
[33] M. Rafik, A. Mas, M.-F. Guimon, C. Guimon, A. Elharfi, F. Schué, Plasma-modified through alginic acid membranes, Separation Science and Technology 24 (1989)
poly(vinyl alcohol) membranes for the dehydration of ethanol, Polymer Inter- 541–554.
national 52 (2003) 1222–1229. [57] L. Krasemann, B. Tieke, Ultrathin self-assembled polyelectrolyte mem-
[34] G. Li, W. Zhang, J. Yang, X. Wang, Time-dependence of pervaporation per- branes for pervaporation, Journal of Membrane Science 150 (1998)
formance for the separation of ethanol/water mixtures through poly(vinyl 23–30.
alcohol) membrane, Journal of Colloid Interface Science 306 (2007) 337–344. [58] Q.G. Zhang, J. Qian, Y. Yang, Q. An, X. Liu, Z. Gui, Polyelectrolyte layer-by-
[35] M.L. Gimenes, L. Liu, X. Feng, Sericin/poly(vinyl alcohol) blend membranes layer self-assembly enhanced by electric field and their multilayer membranes
for pervaporation separation of ethanol/water mixtures, Journal of Membrane for separating isopropanol-water mixtures, Journal of Membrane Science 320
Science 295 (2007) 71–79. (2008) 73–77.
[36] T. Ishikawa, Membranes for dehydrating aqueous ethanol by pervaporation, [59] Q. Zhao, J.-W. Qian, Q.-F. An, Q. Yang, P. Zang, A facile route for fabricat-
Kemikaru Enjiniyaringu 29 (1984) 19–25. ing polyelectrolyte complex membrane with high pervaporation performance
[37] A. Mochizuki, S. Amiya, Y. Sato, H. Ogawara, S. Yamashita, Pervaporation sepa- in isopropanol dehydration, Journal of Membrane Science 320 (2008)
ration of water/ethanol mixtures through polysaccharide membranes. IV. The 8–12.
relationship between the permselectivity of alginic acid membrane and its solid [60] T. Brinkmann, M. Dijkstra, K. Ebert, K. Ohlrogge, Improved simulation of a
state structure, Journal of Applied Polymer Science 40 (1990) 385–400. vapour permeation module, Journal of Chemical Technology and Biotechnology
[38] Y.Q. Dong, L. Zhang, J.N. Shen, M.Y. Song, H.L. Chen, Preparation of poly(vinyl 78 (2003) 332–337.
alcohol)-sodium alginate hollow-fibre composite membranes and pervapora- [61] S.J. Metz, J. Potreck, M.H.V. Mulder, M. Wessling, Water vapour and gas
tion dehydration characterisation of aqueous alcohol mixtures, Desalination transport through a poly(butylene terephthalate) poly(ethylene oxide) block
193 (2006) 202–210. copolymer, Desalination 148 (2002) 303–307.
[39] I. Cabasso, E. Korngold, Z.-Z. Liu, On the separation of alcohol/water mixtures [62] W. Kujawski, Application of pervaporation and vapour permeation in envi-
by polyethylene ion exchange membranes, Polymer Science: Polymer Letters ronmental protection, Polish Journal of Environmental Studies 9 (2000)
Edition 23 (1985) 577–581. 13–26.
[40] S. Yamada, T. Nagawa, Effect of metal ion on separation of alcohol-water mix-
ture with dithiocarbamated poly(vinyl chloride) membrane, in: R. Bakish (Ed.),

You might also like