Membranes For Bioethanol Production by Pervaporation
Membranes For Bioethanol Production by Pervaporation
Membranes For Bioethanol Production by Pervaporation
Abstract
Background: Bioethanol as a renewable energy resource plays an important role in alleviating energy crisis and
environmental protection. Pervaporation has achieved increasing attention because of its potential to be a useful way
to separate ethanol from the biomass fermentation process.
Results: This overview of ethanol separation via pervaporation primarily concentrates on transport mechanisms, fab-
rication methods, and membrane materials. The research and development of polymeric, inorganic, and mixed matrix
membranes are reviewed from the perspective of membrane materials as well as modification methods. The recovery
performance of the existing pervaporation membranes for ethanol solutions is compared, and the approaches to
further improve the pervaporation performance are also discussed.
Conclusions: Overall, exploring the possibility and limitation of the separation performance of PV membranes
for ethanol extraction is a long-standing topic. Collectively, the quest is to break the trade-off between membrane
permeability and selectivity. Based on the facilitated transport mechanism, further exploration of ethanol-selective
membranes may focus on constructing a well-designed microstructure, providing active sites for facilitating the fast
transport of ethanol molecules, hence achieving both high selectivity and permeability simultaneously. Finally, it is
expected that more and more successful research could be realized into commercial products and this separation
process will be deployed in industrial practices in the near future.
Keywords: Pervaporation, Membrane, Ethanol, Polymer
© The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and
the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material
in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material
is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the
permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativeco
mmons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/
zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 2 of 33
most commonly used vehicle fuels and are considered as attempts to provide a review on the research progress in
promising alternatives to conventional petroleum [7, 8]. ethanol-selective PV membranes from the perspectives
Theoretically, any of alcohols including methanol, etha- of transport mechanisms, common fabrication methods,
nol, propanol, and butanol can be used for petrol engines and membrane materials. It will benefit a lot for taking
due to their oxygen enrichment, octane enhancer, and inspiration from the previous works, and finding new
reduction of carbon monoxide and unburned hydrocar- ideas and strategies for developing a new generation of
bons emissions that increases engine efficiency and per- high-performance and high-stability PV membranes for
formance [9]. Only methanol and ethanol fuels, however, ethanol recovery.
are economically and technically feasible for internal
combustion engines. Unfortunately, methanol has tox- Transport mechanisms of PV
icity. The use of bioethanol–gasoline-blended fuels for In the PV process, the desired component in feed pref-
automobiles is beneficial to not only greatly reducing erentially permeates through a membrane, evaporates
the consumption of petroleum but also the emissions of into vapor, and enriches at the permeate side of the
greenhouse gases such as CO2, CO, SO2, HC, and NOx membrane [17]. Due to the complex interaction between
[8, 10]. From the viewpoint of the life cycle assessment, membrane materials and permeating components, the
bioethanol produced from corn through the fermenta- mass transfer in the membranes is considerably compli-
tion process raises the energy by 20–30% than that of cated. So far, there is no such a universal model to char-
fossil fuel used to produce it, while bioethanol produced acterize every detail of mass transfer in membranes [18].
from cellulosic and sugarcane yields nine times as much The permeation flux is significantly dependent on the
energy as fossil energy consumed [7]. Cellulose is the physical chemistry properties of the membranes, such as
most abundant natural plant resource in the world, so thickness [19], affinity for permeate components [20, 21],
the production of fuel ethanol by fermenting cellulose and diffusion coefficient of permeate components across
has become a guarantee of sustainable development [11]. membranes [22, 23]. For example, the permeation flux is
Unfortunately, fermentation broths typically contain inversely proportional to the membrane thickness. This
ethanol at less than 10 wt. % since a higher concentration is principally because that the thinner the membrane is,
of ethanol would have a suppression effect on microor- the lower the overall mass transfer resistance of permea-
ganisms for bioethanol fermentation, thereby leading to tion components through the membrane, therefore, the
the stop of the fermentation process [12]. Therefore, an greater the permeation flux [24]. To prepare membranes
important step in the cellulose-to-ethanol conversion with high-PV performance, it is critical to understand the
process is the extraction of ethanol from fermentation transport mechanism [25].
broths. A conventional separation process is intermittent,
at first, a batch of fermenting liquid is distilled [13], and Facilitated transport mechanism
then the obtained azeotrope products (95.6 wt. % etha- The facilitated transport mechanism is proposed to
nol) are separated and dehydrated to meet fuel specifi- explain the preferential permeation of a species (either
cations. Distillation requires a tremendous amount of gas or liquid) through a mixed matrix membrane. Fillers
energy, meaning high capital costs [14]. Apparently, such in the membrane act as facilitated transport carriers, to
intermittent production process is low-efficiency and accelerate the transfer of desired species. According to
high-energy consumption, which does not accord with mobility, they can be classified into three categories: free-
requirements of sustainable development and circular moving carriers, semi-mobile carriers, and fixed-site car-
economy. riers [26]. In most cases, nanoparticle-filled membranes
Pervaporation (PV), as an emerging membrane sepa- fall into the last category due to they are covalently
ration technology, has received increasing attention as trapped into matrices. In this situation for ethanol PV
Kober first proposed the concept of PV at the begin- separation, a reversible chemical reaction (C + E ⇔ CE)
ning of the last century [15]. Nowadays, PV has broadly occurs between each carrier (C) and ethanol molecule (E)
enlisted by researchers for the separation and purifica- in the membrane, whereby ethanol preferentially diffuses
tion of biofuels by means of its advantages including through the membrane, whereas C and CE only exist in
highly efficient separation, simple equipment, low cost, the inside of the membrane due to the mobile constraint
low pollution, and low-energy consumption since the of C. This is the reason that the hybrid membranes are
separation of the liquid mixture achieved by PV refers selective for ethanol. At present, almost only the dual-
to the difference in the dissolution rate [16]. The PV in mode sorption model and resistor–capacitor (RC) cir-
conjunction with a fermentor cannot only achieve con- cuit model are put forward to systematically unravel this
tinuous biomass-ethanol production but also reduce the facilitated transport mechanism for fixed-site hybrid
energy, economic, and environmental costs. This article membranes.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 3 of 33
Solution–diffusion mechanism
The solution–diffusion mechanism, first developed by
Thomas Graham to elucidate gas penetrant across mem- Fig. 1 Schematic of the solution–diffusion mechanism [32]
branes, is the currently most widely accepted transport
mechanism. Up to now, it is extended to other separa-
tions technologies, such as PV, nanofiltration, reverse
osmosis as well as dialysis. According to the solution– by elevating either solubility selectivity or diffusivity
diffusion model, the transport of permeating species selectivity. The former is thermodynamically controlled,
through a membrane involves three distinct and con- whereas the latter is a kinetically favorable process. The
secutive steps: solution (also known as sorption), diffu- membrane solubility selectivity favors more conden-
sion, and desorption, as pictorially shown in Fig. 1 [32]. sable molecules or molecules with special interactions
When reaching a solution equilibrium, the overall rate with membrane materials [35]. In practice, this interac-
is determined by the first two steps since molecules des- tion is usually assessed by means of the Hildebrand and/
orb generally with extreme rapidity. Therefore, the activ- or Hanson solubility parameter [36]. Typically, the closer
ity coefficient in the solution process can be theoretically the solubility parameters of permeation component and
calculated based on the Henry’s Law and Flory–Huggins membrane material, the higher the perm-selectivity of
theory in which the simplifying assumptions are no and the permeation component would be [37]. The mem-
intense interactions between permeating species and brane diffusivity selectivity is primarily associated with
membrane material, corresponding to two extreme cases. properties of permeation components (i.e., shape and
In contrast, the diffusion process is an irregular diffusion, size) and membrane materials (i.e., flexibility and inter-
which is not completely governed by the Fick’s first law chain spacing), as well as the intermolecular interaction
[33]. In light of this situation, Fujita developed the free between the permeating components themselves and
volume theory to illustrate the diffusion behavior of per- the interaction between the permeating components
meates in polymeric membranes. The permeating mol- and the membrane material. Therefore, it is possible to
ecules were assumed to diffuse essentially along with the determine whether the membrane is water-selective or
polymeric interchain spacing and internal micropores of organic-selective by the choice of membrane material
the membranes [34]. Currently, this theory is generally and the control of membrane morphology [38]. Interest-
applicable to explain the diffusion of penetrants through ingly, incorporation of appropriate inorganic materials or
mixed matrix membranes. hybrids into polymer matrices is capable of consumedly
From the solution–diffusion mechanism standpoint, improving both the solubility selectivity and the diffusiv-
the PV performance of membranes would be enhanced ity selectivity of the hybrid membranes [39]. For ethanol
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 4 of 33
perm-selective separation from water, it is confirmed that pisat is the saturated vapor pressure of species i. Besides,
the addition of a hydrophobic material with a wide range the membrane selectivity for a binary mixture containing
of hydrophobic groups into membranes could improve species i and j, αi,j, can be expressed as:
both the solubility selectivity and the diffusivity selectiv-
ity. On the one hand, the hydrophobic fillers tend to pro- Pi Di ci,membr ci,feed
αi,j = = , (4)
vide a higher affinity of the membranes for ethanol rather Pj Dj cj,membr cj,feed
than water, which can be responsible for the increase in
solubility selectivity [40]. On the other hand, with regard where Pi and Pj represent the aforementioned perme-
to the diffusivity selectivity, it can be categorized into two abilities of species i and j; Di and Dj are the diffusion
cases: one is porous fillers; and the other is nonporous coefficients of species i and j; ci,membr and cj,membr are the
fillers. Porous fillers behave as molecular sieves attribut- concentrations of species i and j in the membrane sur-
ing to their specific pore structure; and nonporous fill- face; ci,feed and cj,feed are the concentrations of species i
ers create a good number of channels at the filler/matrix and j in the feed. D can be related to the cmembr of a par-
interface for ethanol to preferentially traverse the mem- ticular species by the Fick’s first law as follows:
branes, which results in the improvement of diffusivity dcmembr
selectivity. In addition, the incorporation of fillers is also J = −D , (5)
dδ
capable of severely disrupting the inherent polymer chain
packing, thereby improving the free volume inside the where δ is the position variable. To decoupled the
membranes, subsequently increasing the membrane per- impact of “effective” membrane thickness, especially if
meability [41]. the thickness of selective layer is unknown or the mem-
brane is asymmetric, the selectivity αi,j of species i and j
Membrane performance can be further described as:
The PV separation performance of membranes is usu-
Pi l
ally evaluated by the productivity and the separation αi,j = . (6)
Pj l
ability for targeted components from mixtures that are
expressed in terms of permeation flux and separation
factor or permeability (or permeance) and selectivity,
respectively [42]. In the case of a binary mixture, the per- Fabrication of PV membranes
meation flux J and separation factor β can be experimen- According to the membrane structure, PV membranes
tally obtained with the following equations as follows: can be classified into dense membranes and asymmetric
membranes. The dense membranes with the nonporous
Q
J= , (1) structure are relatively thick, commonly over 100 μm,
A·t
and tend to deliver limited permeation flux since mass
transport becomes limited, accordingly they are inap-
Yj Yi propriate for large-scale industrial PV separation. Alter-
β= , (2)
Xj Xi natively, the asymmetric membranes are constructed of
thin separation layers and micropores support layers. The
where Q stands for the total weight of permeates porous support materials, such as polyvinylidene fluoride
through the effective surface area A in time t. In addition, (PVDF) [43, 44] and cellulose acetate (CA) microfiltra-
Y and X represent the mass fractions of species i and j on tion membranes [45, 46], ceramic materials [23, 47], and
the permeate and feed sides, respectively. other various customized porous materials [48–54], pro-
The permeability P, permeance P/l, and selectivity α vide structural strength for fragile thin separation layers
take into account the impact of the driving force. The and enhance permeability without affecting the intrinsic
permeation flux of species i, Ji, is a strong function of its selectivity because of the reduction of the effective thick-
partial pressure gradient and inversely proportional to ness of separation layers. Membranes can be divided into
the membrane thickness l as follows: three types in terms of their shapes: the flat-sheet [55–
Pi Pi 58], tubular [59–61], and hollow-fiber membranes [62,
γi χi pisat − pi,p , (3)
Ji = pi,f − pi,p = 63].
l l
The flat-sheet membranes have been employed in bio-
where pi,f and pi,p are the partial pressures of species i reactor systems for recent decades [31]. In addition, most
in feed and permeate vapor, respectively, whereas γi and of the currently and commercially available membranes
χi denote the activity coefficient and the mass fraction of for recovering ethanol from water via PV are in the flat-
species i in the feed solution, respectively. In addition, sheet configuration on account of easy fabrication and
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 5 of 33
convenient assembly [64–66]. Considering that the eth- suspension and the polymeric solution are mixed [73, 79].
anol-selective membranes are almost polymer-based It is important to highlight that mechanical stirring and/
materials, only the fabrication of polymeric membranes or ultrasonication using an ultrasonic bath or a probe-
(both dense and composite membranes) was discussed type ultrasonic homogenizer is performed for a specific
herein. period of time to prevent the fillers from agglomeration
Dense flat-sheet membranes are frequently prepared or sedimentation [14]. Besides, the remaining fabricated
by solution casting followed by solvent evaporation [32]. steps are essentially the same as those of aforementioned
Typically, an artificial polymer is first and completely dis- pure polymer membranes.
solved in an appropriate solvent. After degasification, the In some cases, to ameliorate the interfacial compatibil-
obtained dope solution is poured on a flat plate, allowed ity of two phases and/or prevent fillers from agglomerat-
to spread, and subsequently left to evaporate the solvent ing, the fillers should be modified prior to addition. For
[67]. Membranes with multiple selective layers, in sand- instance, coupling agents are utilized as intermediates to
wich-like structure, are constructed by a layer-by-layer form durable covalent bonds between inorganic fillers
solution casting method in much the same procedures as and polymer matrices [80]. It is worth mentioning that
mentioned above [68]. Surface pre-treatment for a pre- apart from the loading amount, the physical and chemi-
viously formed layer, if necessary, is applied prior to the cal properties of fillers, such as size, shape, and functional
coating of the next one to enhance the interlayer inter- groups on the surface, are deemed to dramatically impact
action [69]. In another case, the polymeric solutions can membrane performance [19, 81–91]. In practice, a good
be coated on various porous supports consisting of either number of nanofillers have been employed in conjunc-
organic or inorganic substances to fabricate compos- tion with polymers to enhance their PV performance for
ite membranes [23, 45, 70]. Apart from the regular flat ethanol recovery including zeolites (e.g., silicalite-1 and
plates, the supports could be in hollow fiber or tubular ZSM-5) [76, 92], fumed silica [93], metal–organic frame-
format [50, 63, 71]. To prevent polymer solutions from works (MOFs) [94], zeolitic imidazolate frameworks
penetrating and blocking the micropores of supports (ZIFs) [95–97], carbon nanotubes [98], carbon blacks [99,
during the coating process, an effective approach is infil- 100], graphene-like fillers (graphene and graphene oxide
trating the supports by high-volatile solvents which are derivatives) [101, 102], and polyhedral oligomeric silses-
immiscible with the solvents of dope solutions before quioxanes (POSSs) [103].
coating [67]. Taking an organophilic polymer dissolved Compared to flat-sheet membranes, hollow-fiber
in an organic solvent as an example, the supports can be membranes used in PV separation offer some attractive
first immersed in water to thoroughly intrude into their advantages, i.e., higher space efficiency and self-support-
pores. The as-fabricated composite membranes are then ing structure. A hollow-fiber membrane is shaped like a
obtained after evaporation of coating solvents followed soda straw with a thin dense selective layer on the outside
by the removal of the solvents in the supports [72]. surface along with a microporous structure in the lumen
The improvement in separation performance and/ side. Hollow-fiber membranes are commonly fabricated
or robustness of membranes can be achieved through via an extrusion process, referred to as “spinning” includ-
physicochemical modifications. A most common and ing solution spinning and melt spinning [104]. Until now,
facile technique is the incorporation of either inorganic few hollow-fiber membranes are, until recently, available
or hybrid nanofillers into polymeric matrices as mixed for ethanol recovery application by PV. At present, the
matrix membranes including physical blending and gain of an ideal multi-phase membrane-forming system
chemical grafting method. Technically, physical blending and the exploitation of a corresponding well-controlled
is the most frequently used method in which fillers can membrane fabrication procedure to achieve a high-per-
be simply and directly introduced into polymer matrices forming hollow-fiber membrane have become a rising
with vigorous stirring and/or sonication to suppress their research area in hollow-fiber spinning [105].
agglomeration and to facilitate their uniform distribution
[73, 74]. Membrane materials for ethanol recovery
The dispersion of nanofillers into polymeric matrices Theoretically, the ethanol recovery from water by PV
is generally carried out by three different procedures: (i) involves three effects: concentration polarization, cou-
nanofillers are dispersed into a solvent first, and subse- pling between mixed components as well as membrane
quently polymer is added into the suspension [75, 76]; (ii) swelling [106]. Concentration polarization, as an inher-
polymer is dissolved in a solvent, afterward, nanofillers ent phenomenon in the membrane separation process,
are added into the polymeric solution [77, 78]; (iii) nano- denotes the formation of concentration gradients in
fillers are dispersed into a solvent and polymer is dis- the vicinity of membrane surface and bulk feed liquid
solved in the same solvent separately, and then the filler due to the enrichment of rejected species (water) at the
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 6 of 33
upstream membrane surface but also the decline of per- including polydimethylsiloxane (PDMS), poly(1-tri-
meating species (ethanol). If ethanol molecules in the methylsilyl-1-propyne) (PTMSP) and so on.
bulk feed are to permeate through the membrane, they
have to diffuse through the concentration polarization PDMS membranes
region, referred to as the boundary layer, to the mem- PDMS, usually referred to as ‘‘silicone rubber’’, is the most
brane surface. It definitely impacts the permeation of representative, well-studied, and widely used membrane
ethanol in a negative way. Higher feed flow velocity can material [110]. PDMS delivers an ethanol selectivity as a
effectively promote turbulence for a better mixing, bring result of its backbone comprised entirely of Si–O bonds.
ethanol molecules to the surface, and subsequently pre- It is considered as the benchmark material to recover
vent water from further excessive accumulation, result- ethanol from aqueous solutions. In addition, for now,
ing in reducing the thickness of the diffusion boundary from the aspect of the development of membranes selec-
layer, thereby restraining this effect as much as possible tively recovered ethanol in recent decades, it appears
[107]. The coupling effect means that the diffusion rates that PDMS would not be edged out of the top spot in
of ethanol and water influence each other during the the short term. In addition, PDMS also possesses a good
PV progress. Ethanol and water molecules are bound membrane-forming property which is more prone to fab-
together by hydrogen bonds. These combined molecules ricate more practical configurations, such as common flat
cause ethanol permeation resistance in the membrane, sheet and hollow fiber. A comprehensive review reported
and are partially restricted and not allowed to leave the by O’Brien group on the economic analysis of ethanol
membrane. An increase in the affinity of the membrane production by a PV process coupled with a fermentor
towards ethanol over water would foster ethanol per- has concluded that a membrane with a permeation flux
meation, thus mitigating the coupling effect [108]. The of over 150 g·m−2·h−1 as well as a separation factor of no
swelling effect refers to the increase in the volume of less than 10.3 would be cost-competitive with distillation
membranes resulted from the absorption and accumu- alone on a commercial scale [111]. In fact, PDMS-based
lation of ethanol. Especially, many of polymeric mem- membrane products in the ethanol removal applica-
branes suffer from excessive swelling which breaks the tion have been so far marketed, e.g., Pervap, Pervatech,
regular arrangement of molecular chains, inducing a PolyAn, and SolSep manufactured by Sulzer Chemtech,
terrible change in the membrane structure. As a result, Pervatech BV, PolyAn GmbH, and SolSep BV, respec-
the membrane selectivity and the service life are deterio- tively. Unfortunately, Kujawski et al. found that the mem-
rated. Some strategies such as crosslinking and addition brane Pervap 4060 presented the separation efficiency
of fillers can prevent the excessive build-up of ethanol with the ethanol flux of 276 g·m−2·h−1, the water flux of
so as to suppress the ethanol-induced swelling behavior 969 g·m−2·h−1, and the lower separation factor of 6 in a
[109]. 4 wt. % of ethanol in water at 60 °C [64]. Similar lower
Given these discussions provided herein, it is obvi- separation factors were obtained for other commercial
ous that the approaches for circumventing the coupling membranes under different test conditions [56, 65, 66], as
and swelling effect are all related to membranes. Hence, summarized in Table 1.
membrane materials play crucial roles not only in per- Besides, Beaumelle reviewed that pristine PDMS mem-
meability and selectivity but also in overall properties branes generally delivered lower ethanol/water separa-
including swelling resistance and mechanical strength. tion factors of less than 10 along with a broad range of
Till now, a wide variety of membrane materials have permeation fluxes ranging from 1 to 1000 g·m−2·h−1
been exploited for the removal of ethanol from aqueous [112]. Similarly, O’Brien also indicated that ethanol/water
solutions by PV. Generally, ethanol perm-selective PV separation factors for ‘pure’ PDMS membranes were in
membranes can be classified into polymeric membranes, a range of 4.4–10.8. This significant variation in values
inorganic membranes, and mixed matrix membranes. is mainly related to the molecular weight, crosslinking
density, membrane thickness, potential support layer,
operating condition, etc. From these analyses, it is clear
Polymeric membranes that the ethanol recovery performance of pure PDMS are
Hydrophobic polymers are identified as the most ver- expected to be elevated.
satile and prospective membrane materials in PV appli- To this end, many efforts have been spent on the
cation for ethanol extraction from water on account of modification of PDMS membranes. A simpler way is to
their better malleability and processability as well as prepare composite membranes by coating thin PDMS
lower production costs. In addition, they have attracted active layers on highly porous supports. Through reduc-
widespread research interest. To date, many ethanol- ing the effective thickness, the permeation flux can be
selective polymeric membranes have been reported, improved, as listed in Table 1. For example, permeation
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 7 of 33
Table 1 (continued)
PSF polysulfone, PES polyethersulfone, PA polyamide, CA cellulose acetate, PI polyimide, PVDF polyvinylidene fluoride, IPAA-FA copoly(N-
isopropylacrylamide/1H,1H,2H,2H-perfluorododecyl acrylate), HMDSO hexamethyldisiloxane, PPO polyphenylene oxide, PSSQ phenylsilsesquioxane, SDS
polystyrene-b-PDMS-b-polystyrene, PS polystyrene, PTFE polytetrafluoroethylene, F-PBZ fluorinated polybenzoxazine, SBS styrene–butadiene block copolymer, TMVS
trimethoxyvinylsilane, PEI polyetherimide.
fluxes of composite membranes supported on polysul- (labeled as PIS6, PIS11, and PIS19, respectively) by poly-
fone (PSF) [113], polyamide (PA) [113], and cellulose ace- condensation, and used them to make the tough self-
tate (CA) [45] were achieved as high as 1600, 1850, and standing membranes. It was observed that all of these
2800 g·m−2·h−1, respectively. membranes exhibited perm-selectivity toward ethanol as
In addition, PDMS modification by blending, well as the other organics along with stable and incred-
blocking, or grafting with other polymers has been ibly high permeability. In particular, the PIS19 offered the
attracting considerable attention. A blend mem- best PV performance with the moderate separation fac-
brane consisting of PDMS and 5.3 wt. % copoly(N- tor of around 9.5 and the very advantageous permeation
isopropylacrylamide/1H,1H,2H,2H-perfluorododecyl flux of 5720 g·m−2·h−1 which was at least more than six-
acrylate) (IPAA-FA) showed a higher separation factor fold that of PDMS alone [119].
(19.7) for a 2.5 wt. % ethanol/water mixture [114]. Chang Similar to graft copolymerization, surface modifica-
et al. blended PDMS with hexamethyldisiloxane tion with a silane compound is also an effective approach
(HMDSO) cured by phenyl triethoxysilane, then coated to maximize the membrane ethanol-selectivity. Kashi-
the silicone layer on the microporous PVDF support on wagi and co-workers treated PDMS membranes utilizing
which PDMS was plasma-induced grafted for enhancing plasma grafting with the silane compound octadecyldi-
the adhesion. Although the as-prepared loose network ethoxymethylsilane containing a long alkyl chain. The
structural composite membrane displayed a lower sepa- treated membrane yielded a maximum separation factor
ration factor of 5.1, the corresponding permeation flux of 16.3 at 25 °C [120].
was up to approximately 1300 g·m−2·h−1 [115]. Other PDMS-based membranes reported in the lit-
For block copolymerization, Liu reported a PDMS- erature have primarily been concentrating on incorpo-
b-polyphenylene oxide (PPO) block copolymer mem- rating fillers to fabricate mixed matrix membranes. The
brane. Its permeation flux could reach a high of related studies will be discussed later in “Mixed matrix
3817 g·m−2·h−1, along with a separation factor of 8.5 membranes”. Meanwhile, significant efforts have been
when concentrating a 5 wt. % ethanol aqueous solu- positioned toward exploiting novel fabrication meth-
tion [116]. Guo et al. prepared membranes with a novel ods. A PDMS membrane was fabricated by Qin’s group
silicone block copolymer which was synthesized by in the water phase instead of the conventional organic
the condensation of rubbery PDMS with glassy lad- phase with dodecylbenzenesulfonic acid (DBSA) as sur-
der like phenylsilsesquioxane (PSSQ). The unique cage- factant and PVDF as support for the recovery of ethanol
type structure of the PSSQ mitigated the swelling of the from model ethanol aqueous solution and fermenta-
membrane and enhanced the affinity of the membrane tion–PV integrated process, respectively. The permeation
towards ethanol. The PV results showed that the maxi- fluxes fell in the range of 396–664 g·m−2·h−1 and 332–
mum separation factor afforded by the prepared mem- 548 g·m−2·h−1, and the separation factors were in the
branes was 11 at a 5 wt. % ethanol solution [117]. Indeed, range of 8.6–11.7 and 8–11.6 in the fed-batch and contin-
it was improved as compared with that of pristine PDMS uous fermentation–PV system, respectively [121]. Wang
membranes. In addition, a microphase-separated triblock et al. unveiled a facile and scale-up roll-coating method
copolymer–polystyrene (PS) -b-PDMS-b-PS (SDS) was by which multi-layer PDMS was easily and controllably
reported to prepare membranes for recovery of vola- assembled on top of the PSF support. The as-fabricated
tile organic compounds (VOCs) from aqueous mixtures pilot-scale composite membranes were adopted to sepa-
by PV. It was proven that both the permeation flux and rate ethanol by PV in a lab scale and a pilot plant. The PV
separation factor offered by these SDS membranes were results showed a high and stable PV selectivity towards
closely dependent on ethanol concentration in feed mix- ethanol over water with a separation factor of 11.6 as
tures. When separating a 5 wt. % ethanol binary aqueous well as a permeation flux of 1493 g·m−2·h−1, which was
solution, a permeation flux of 540 g·m−2·h−1 and a sepa- excellent as compared with the corresponding values
ration factor of 8.8 were observed [118]. for the PDMS/PSF membranes prepare by other meth-
With regard to graft copolymerization, Nagase et al. ods. It was also concluded that this method was instruc-
successfully synthesized three kinds of PDMS-graft tive in obtaining defect-free membranes and scaling up
copolyimides with different PDMS segment lengths to larger systems [48]. A membrane reported by Mori
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 9 of 33
et al., impregnated PDMS into the pores of polytetra- limitation of the optimal combination of strength and
fluoroethylene (PTFE) support, can be thinned to 5 μm performance result in limited offerings.
in thickness. In this case, it gave the permeation flux and
separation factor corresponding to 760 g·m−2·h−1 and PTMSP membranes
8.3, respectively [85]. Another widely studied polymeric membrane material
Moreover, constructing an architectural multilayer for this application is poly [1- (trimethylsilyl) -1-propyne]
structure is also an ingenious way. More recently, Figoli (PTMSP). It is a super glassy polymer with better pro-
and co-workers explored a three-layer asymmetric cessability and scalability than rubbery ones like PDMS.
flat-sheet composite membrane, in sandwich-like con- In addition, it possesses an extra high free volume which
figuration, which consists of a dense top layer of styrene– is associated with a higher flux of permeates. The data of
butadiene block copolymer (SBS) and an intermediate ethanol/water PV performance reported in the literature
layer of PDMS or polyurethane (PU) layer followed by a references for pristine PTMSP membranes are tabulated
support layer of fluoroplast F-42 with high porosity via in Table 2. It is observed that the ethanol separation fac-
layer-by-layer assembly. An increase in ethanol selectiv- tor, shown in the fourth column, is ranged from 9 to 26,
ity of the composite membranes with PDMS as the inter- which is much larger than that of pure PDMS. Unfor-
mediate layer was observed with a maximum of 8.3 in tunately, a large number of experimental studies have
an aqueous solution containing 3 wt. % ethanol at 30 °C. proved that such super glassy polymers suffer from rapid
On the other hand, the membranes also showed a strong physical aging which is a ubiquitous phenomenon since
increase in permeation flux as a result of the lower mem- polymer chains undergo physical relaxation with time
brane thickness of selective layers [68]. Chang and Chang and are prone to converge towards a thermodynamic
synthesized a copolymer of PDMS and phosphate ester equilibrium [133], resulting in the reduction in free vol-
as the selective membrane material [122], and designed a ume [134]. Under these circumstances, the separation
multi-layer membrane configuration by employing alter- performance of PTMSP membranes in terms of both
nating layers of dense copoly(PDMS-phosphate ester) sil- permeability and selectivity decay continuously over
icone and porous PVDF on which trimethoxyvinylsilane time. To complicate matters, this decline becomes more
(TMVS) was previously plasma-polymerized to form a serious for thinner membranes, which restricts the incre-
thin and loose polymeric layer. It was confirmed that the ment in permeation flux. Gonzalez-Velasco et al. calcu-
multiple-layer composite membranes exhibited excellent lated the decreasing tendency of ethanol selectivity with
PV performance, especially at lower ethanol concentra- time based on the experimental data. They demonstrated
tion. Within a specific reference, the more the silicone that, for a PTMSP membrane of 100-μm thickness, the
layers were, the higher the separation factor, but the separation factor fell to a level of approximately 8 from
lower the permeation flux was due to the trade-off effect. the initial value of around 10.7 after continuous exposure
In particular, the four-layer PDMS membrane delivered to a 10 wt. % ethanol aqueous solution at 75 °C for 450 h.
the optimal separation capability with an exceptional Though the separation factor maintained practically con-
separation factor of 31 and a high permeation flux of stant over a 40-h period, the permeation flux dropped
900 g·m−2·h−1 in a 10 wt. % ethanol-containing aqueous from 540 to 350 g·m−2·h−1 [135]. A similar result that
solution [123]. the selectivity for a 30-μm PTMSP membrane could only
Several researchers have reported the PDMS mem- remain at about the initial level for just 40 h was attained
branes in hollow-fiber configuration. The commer- by Masuda [136]. This severe long-term instability is fur-
cial PDMS hollow-fiber membranes as well as the PVC ther aggravated for concentrated solutions and higher
membrane module manufactured by PAM-Membranas temperatures due to the higher mobility of polymer
Selectivas in Brazilian were chosen by Marangoni et al. chains. Gonzalez-Velasco assessed the deterioration of
to selectively remove ethanol from actual fermentation PTMSP over a long operation time of 572 h in a higher
broth with 3 wt. % of ethanol. Notably, the permeation feed concentration of 50 wt. % at higher temperatures of
flux was only about 6 g·m−2·h−1 [62]. Apparently, this 50 and 75 °C. Experimentally, it was observed that both
value was incredibly lower although it was essentially in the separation factor and permeation flux of PTMSP
accordance with the datum provided by the supplier. In membranes in various thickness diminished rapidly in a
contrast, the PDMS hollow-fiber membranes with sup- short time, then reduced slowly for test time longer than
porting materials, such as PSF and polyetherimide (PEI) 250 h. Eventually, when operating at 50 °C, the separa-
as gathered in Table 1, achieved relatively higher permea- tion factor and permeation flux for the 52-μm mem-
tion fluxes. Nevertheless, these data are still lower than brane decreased to 3 and 800 g·m−2·h−1 from the initial
those of PDMS-base flat-sheet membranes. The com- value of around 4.5 and 1200 g·m−2·h−1, respectively
plexity of the hollow-fiber fabrication process and the [137]. It is because of this unfavorable characteristic, the
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 10 of 33
commercial application of PTMSP remains limited at this PFA at the other end of PFA-g-PDMS was mainly local-
time. ized on the membrane surface just like the grass on the
In this regard, several methods have been put forth to ground. The surface-modified PTMSP membranes dis-
stabilize the PTMSP membranes and to further improve played higher water repellency, leading to lower solubility
the ethanol-selective separation performance for bright- of water, therefore, higher ethanol permselectivity [141].
ening the application prospects of this material. At the Subsequently, Kang et al. also employed PDMS to cir-
very beginning, Nagase’s group successively attempted cumvent the intrinsic issue without sacrificing perme-
to optimize PTMSP performance by integrating short ability if possible. Three preparation methods, semi-IPN,
PDMS chains [138], trialkylsilyl groups [139], and fluoro- PDMS sorption, and PDMS sorption and crosslink-
alkyl groups [140]. In particular, all PTMSP membranes ing, were carried out. Although the initial permeation
grafted with PDMS in any proportion delivered higher flux could be up to 600 g·m−2·h−1 along with a modest
selectivity to ethanol than PTMSP and even the most separation factor of around 13.7 by semi-IPN method, it
outstanding PDMS alone. The separation factor reached quickly reduced to 200 g·m−2·h−1 after a 7-day run along
a maximum of 28.3 for the copolymer membrane con- with a strikingly reduced separation factor of 7.3. In con-
taining 12 mol % of PDMS, while the corresponding trast, the membranes prepared by PDMS sorption as
permeation flux was mere 61 g·m−2·h−1 which was not well as PDMS sorption and crosslinking presented more
ideal or even discouraged [138]. From these researches, stable permeable behavior with time, but their separa-
it is inferred that polymer design by block or graft copo- tion properties were lower [142]. As reviewed herein,
lymerization would probably be an effective and flexible it appears that future research focusing on developing
way. Furthermore, the blend method is also a promis- novel preparation methods of PTMSP-based membranes
ing strategy, including polymer blend and integration is a reasonable approach to retard the physical aging.
inorganic or hybrid nanofillers reviewed in “Mixed Since Masuda et al. systematically reported the impact
matrix membranes”. Uragami and co-workers added the of polymerization conditions on properties of PTMSP
poly(fluoroacrylate) (PFA)-g-PDMS graft copolymer into in the 1980s [143–145], almost all the PTMSP has
the dope solution to enhance the surface hydrophobic- been synthesized according to their technique, using
ity of PTMSP membranes. The PDMS chains of PFA-g- metal catalytic systems. As confirmed by Volkov, over-
PDMS were trapped in the PTMSP matrix, while the all properties of PTMSP were strongly correlated with
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 11 of 33
catalyst. Despite a combination of high permeation flux the separation factors multiplied even more than two
(over 300 g·m−2·h−1) and high separation factor (over times (up to 10.5 and 11.7 from 5.3) when operating at
15) for all PTMSP membranes synthesized with three 50 °C [154, 155]. Meanwhile, Li et al. prepared polyphe-
catalytic systems (TaCl5/n-BuLi, TaCl5/Al(i-Bu)3, and nylmethylsiloxane (PPMS) membranes with the same
NbCl5), the permeation flux and separation factor for backbone as PDMS. The resultant membranes yielded a
TaCl5/n-BuLi-catalyzed PTMSP membranes deteriorated higher permeation flux, but a lower separation factor as
significantly with time; conversely, those for the mem- compared with PDMS prepared in the same way attrib-
branes synthesized by the other two catalysts, TaCl5/ uting to the presence of the more hydrophobic and rigid
Al(i-Bu)3 and NbCl5, could remain stable in either syn- phenyl groups in PPMS [46].
thetic acetic acid-containing aqueous solutions or actual Krea et al. synthesized polysiloxaneimides (PSI) block
fermentation broths [146]. More than 3 decades later, the copolymers with high contents of siloxane block in the
most common catalysts for PTMSP synthesis nowadays 70–95 wt. % range. They found that the higher the silox-
are still T
aCl5 [135, 147], NbCl5 [148], and TaCl5/Al(i- ane content was, the better the PV performance. At the
Bu)3 [58, 149]. Hence, the exploitation of active catalysts 94 wt. % PDMS content and 1.5:2:0.5 equivalents of
should deserve more attention. aminopropyl siloxane (ODMS):1,2,4,5-benzenetetracar-
boxylic dianhydride (PMDA):1,3-bis(3-aminopropyl)
Other polymeric membranes tetramethyldisolxane (MDMS), the as-fabricated PSI
Apart from widely known PDMS and PTMSP, a consid- copolymer membranes displayed the optimum PV per-
erable amount of effort has been placed in searching for formance (separation factor of 10.6 and permeation flux
polymeric materials with higher ethanol perm-selectivity, of 560 g·m−2·h−1) when carrying out in a 10 wt. % etha-
higher permeation fluxes, and better physical properties. nol solution at 40 °C [156].
Unfortunately, a few membrane materials are available, Besides, the GKKS Research Center in Germany devel-
and the separation performance of most of them is not as oped a hydrophobic polyoctylmethyl siloxane (POMS)
good as expected. A list of other polymeric materials and membrane on a polyacrylonitrile (PAN) support. Several
respective separation data from the literature for ethanol groups employed their POMS membranes to separate
removal from water is assembled in Table 3. binary ethanol aqueous solutions. Garcia et al. observed
In summary, the reported ethanol perm-selective that the POMS membrane displayed a permeation flux of
materials are mostly based on homopolymers or copoly- approximately 670 g·m−2·h−1 accompanied with a mod-
mers of siloxanes with hydrophobic Si–O–Si backbone erate separation factor of about 8.3 at 53 °C with a feed
and the separation performance is strongly correlated to containing 4.7 wt. % ethanol in water. In addition, the
siloxane which provides excellent separation potential. At authors indicated that the membrane permeability was
the very beginning, Kashiwagi et al. synthesized a series related to the preferential sorption, whereas the selec-
of ethanol-permeable membranes by plasma polymeri- tivity was related to the solubility rather than diffusivity
zation with various silicone and silane monomers. These [157]. However, the result carried out by Lazarova et al.
membranes had PDMS-like structures and their separa- showed that the permeation flux and separation factor
tion factors and permeation fluxes were in the range of for the POMS membranes were 120 g·m−2·h−1 and 3.95,
1.5–5.2 and 180–380 g·m−2·h−1, respectively. Meanwhile, respectively, when exposing in a 5 wt. % ethanol feed at
they utilized silane compounds with long-chain alkyl 50 °C [158]. In contrast, Straathof and co-works obtained
groups to prepare silane membranes. Despite showing intermediate data (the permeation flux of 235 g·m−2·h−1
gratifying separation factors ranging from 13.2 to 16.9, and the separation factor of 5.7) in the same concentra-
permeation fluxes afforded by the polymerized silane tion of ethanol aqueous solution [159].
membranes were relatively low in the 5.3–21 g·m−2·h−1 More recently, Zhang et al. prepared novel thin
range. Similarly, their plasma-polymerized silicone oil poly(vinyltriethoxysilane) (PVTES) membranes to
membranes or the ones further treated with octadecyldi- recover ethanol. They unveiled that the resulting mem-
ethoxymethylsilane all showed discouraging separation branes delivered consumedly high permeation fluxes
results [120]. ranging from 6000 to 10,000 g·m−2·h−1, which clearly
Thereafter, two room-temperature vulcanizing-type transcended the upper limit for PDMS membranes as
(RTV) silicone rubbers, polymethylethoxysiloxane shown in Table 1, while maintaining more constant
(PMES) and polymethylphenylsiloxane (PMPS), were separation factors of around 5 over feed concentrations
employed by Chen et al. for the fabrication of ethanol- range from 3 to 13 wt. % at 35 °C [160]. Subsequently,
selective membranes. Simultaneous increments in they continued their work on modifying the rigid struc-
both permeation flux and separation factor for them ture of the PVTES membrane by copolymerization with
were achieved by comparison with PDMS. Especially, dimethyldiethoxysilane (DMDES), oligomer of hydroxy
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 12 of 33
silicone oil (HSO), and PDMS to optimize its separation PVTES-DMDES, and PVTES-HSO membranes showed
performance. It was proved that the separation perfor- great potential for in situ recovery of ethanol [161].
mance had the direct proportion with the chain flexibil- It should be noted that the siloxane-based membranes
ity and the amount of hydrophobic groups. Hence, the prepared by conventional methods were all dense, where
PVTES-HSO membrane showed the best PV perfor- only intermolecular spacing between the chains, referred
mance which was the exceptionally high permeation to as accessible free volume, provided the pathway for
flux of 8160 g·m−2·h−1 and moderate separation fac- permeate transport. A synthetic membrane composed
tor of 6.6 in a 9 wt. % ethanol aqueous solution at 35 °C. of the monomers of octamethyl cyclotetrasiloxane, vinyl
As summarized in Table 3, it appeared that PVTES, heptamethyl cyclotetrasiloxane, styrene, and divinyl
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 13 of 33
benzene was constructed by Shi et al. via the concen- the membrane materials and enough high diffusivity
trated emulsion polymerization method, in which the across the membranes were indispensable for the ideal
gaps among the latex region and inside the dense latexes membranes.
provided diffusional pathways for permeates. However, Abu-Saied’s group fabricated poly(acrylonitrile-co-
the structured membrane was much thicker than tradi- methyl acrylate) membranes for ethanol extraction from
tional dense ones, and its thickness was 750 μm, more aqueous mixtures and bioethanol purification from fer-
than ten times as thick as the one with a uniform struc- mentation broth originating from cellulosic fiber wastes
ture, and even more than that with a porous support. both in the laboratory scale [167]. It was found that the
Although the separation performance of the resultant polymeric membrane was hydrophobic and ethanol-
membranes was not as good as expected, displaying a selective. Under a nitrogen pressure value of 40 psi, the
permeation flux of 70 g·m−2·h−1 and a separation factor separation factor of the resultant polymeric membrane
of 7.4 with a feed mixture of 30 wt. % ethanol at 16 °C, it was found to reach a maximum value of 11.03 with an
has provided a new opportunity for the development of appreciably poor permeation flux of 1.3 g·m−2·h−1 for
new material membranes [162]. a mixture of 25 wt. % ethanol. On the other hand, the
Other polymeric membrane materials have been inves- membrane exhibited an optimal separation factor of 9.27
tigated by several research groups. For instance, Ishihara and a similarly low permeation flux of 1.4 g·m−2·h−1 with
et al. successfully fabricated a composite membrane com- a 33% fermentation system.
posed of a 20-μm styrene (St)-fluoroalkyl acrylate (FAA) More recently, glassy PIM-1, one of the so-called poly-
graft copolymer skin layer and a 100-μm PDMS layer. The mers of intrinsic microporosity (PIMs), was employed in
membrane presented the separation factor value as high this particular application in terms of its organophilicity
as 45.9. They suggested that this was attributable to the and intrinsic microporosity. Adymkanov and co-workers
relatively low affinity between the hydrophobic St-FAA reported that the PIM-1 membranes with thicknesses
graft copolymer and ethanol molecules, effectively sup- ranging from 25 to 40 µm exhibited a maximum selec-
pressing the membrane swelling and resisting the disso- tivity of 10.7 and a permeation flux of 470 g·m−2·h−1 in
lution of water. Due to the trade-off relationship between a 10 wt. % dilute aqueous mixture at 30 °C. In addition,
permeation flux and separation factor, the permeation the permeation flux could be as high as 1400 g·m−2·h−1
flux of the St-FAA/PDMS composite membrane was only corresponding to a marginally lower separation factor of
5 g·m−2·h−1 [163, 164]. Later, Masuoka and co-workers 9.3 when the operating temperature rose to 60 °C [168].
fabricated plasma-polymerized perfluoropropane (PFP) However, a recent study by Alberto et al. showed that a
membranes on porous PSF supports, and observed a per- 60 ± 9-µm-thick PIM-1 membrane possessed an appre-
meation flux of 300 g·m−2·h−1 with a separation factor of ciably lower separation factor (3.1 ± 1.7) and a relatively
7 [165]. higher permeation flux (circa 700 g·m−2·h−1) than the
Liu et al. employed poly(ether block amide) (PEBA pure PTMSP membrane [20]. It seemed that such mem-
2533) membranes for pervaporative recovery of etha- brane was not very selective for ethanol. Unfortunately,
nol from water. Depressingly, their observation showed glassy polymers, such as PIM-1 and PTMSP, have been
that the separation performance was poor, especially the proven to suffer from physical aging that is the long-term
separation factor was as low as 2.5, owing to the weak operating instability. Furthermore, the polymeric PIM-1
affinity between ethanol and PEBA. What is worse, the membrane also has serious excessive swelling in the pres-
permeation flux reduced to 37 from 118 g·m−2·h−1, more ence of ethanol [169]. These problems would severely sty-
than three times, when thickening up to 100 μm in thick- mie its application in ethanol extraction.
ness from 30 μm [166]. In addition to the aforementioned nonporous poly-
A series of polyphosphazene heteropolymers with dif- meric membranes, the porous membranes also have been
ferent hydrophobic pendant groups were produced by proposed to recover ethanol via the PV process. Nakao
Huang et al. to fabricate membranes for ethanol/water et al. applied a commercially available microporous PTFE
separation. They found that the polyphosphazene mem- membrane (0.2 μm pore diameter) into a continuous
brane with –OCH2CF3 substituting groups showed ethanolic fermentation system. It was demonstrated that
the highest separation performance both in permea- the PTFE membrane displayed an attractive permeation
tion flux and separation factor, which were, respectively, flux with a value of 5700 g·m−2·h−1, whilst the extracted
260 g·m−2·h−1 and 6.1, owing to its highest affinity to ethanol concentration was 6–8 times higher than that
ethanol, highest diffusivity, as well as highest diffusion in feed [12]. Aroujalian et al. also applied the micropo-
selectivity in comparison with the other two membranes rous PTFE membrane with the same pore diameter to
with –OC2H5 and –OCH2CF2CF2CF2CF2H groups [127]. separate ethanol/water mixtures. The maximum values
It was concluded that enough affinity of ethanol toward of 10,592 g·m−2·h−1 for permeation flux was achieved
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 14 of 33
at 60 °C, but with a rather poor separation factor of 2.25 As reviewed above, reported ethanol/water separation
[170]. On the other hand, Sukitpaneenit and co-workers factors for polymeric materials other than PDMS and
explored and reported an asymmetric PVDF hollow- PTMSP cover a fairly broad range, from 1.5 to 113, with
fiber membrane produced by the dry-jet wet spinning reversely changed permeation fluxes ranging from 5 to
for ethanol/water separation. It possessed a porous inner over 14,000 g·m−2·h−1 attributing to the intrinsic trade-
surface, a thicker finger-like macrovoid cross-section, off effect. Collectively, new film-forming materials with
and a rough outer surface with small voids or defects, higher permeability and ethanol selectivity, as well as bet-
thus yielding a superior permeation flux of roughly ter chemical, thermal, and long-term operation stability,
8000 g·m−2·h−1 [171]. Moreover, the membrane perfor- should be actively explored.
mance was strongly restricted by morphology, pore size,
as well as pore distribution which could be controlled Inorganic membranes
by altering spinning conditions. Overall, the membrane It is known that inorganic membranes, fabricated from
with the long finger-like macrovoid structure displayed ceramics or zeolites, are widely used today for dehydrat-
a satisfactory permeation flux (8795 g·m−2·h−1) with an ing organic compounds via PV [176, 177]. Over the past
associated separation factor of near 5; while the mem- 2 decades, however, some researchers have committed
brane with the nearly macrovoid-free morphology pre- themselves to exploring inorganic membranes for recov-
sented an improved separation factor of around 7.8 and ery of ethanol from aqueous solutions because of their
an inevitably and considerably reduced permeation flux special and competitive superiority in separation per-
(3961 g·m−2·h−1) [172]. In general, porous membranes formance (both separation factor and permeation flux),
exhibited low selectivity, but high permeability due to mechanical properties, chemical resistance, thermal sta-
their porosity. bility, anti-fouling ability, and long-term durability over
In contrast, microporous hydrophobic membranes most polymeric ones.
could be employed as supports to immobilize liquids Hydrophobic zeolites have attracted tremendous atten-
with high boiling points so as to form liquid membranes tion from the researchers for ethanol recovery applica-
to achieve high ethanol-selectivity. Thongsukmak and tion via PV. Zeolite consisting of hydrated aluminosilicate
Sirkar immobilized trioctylamine into the pores of a is an inorganic crystalline structure with a uniform pore
polypropylene (PP) hollow-fiber membrane and coated a diameter ranged in 0.3–1.3 nm. Zeolite membranes are
nanoporous fluorosilicone coating on the outside. Their most often fabricated by depositing a polycrystalline
study has demonstrated the liquid membrane had long- zeolite layer onto a porous inorganic support layer (like
term operational stability over 300 h. For a binary mix- tubular or discoid alumina or stainless steel, and even
ture containing 5 wt. % ethanol, it had a high separation ceramics) via hydrothermal synthesis method [178].
factor (maximum 32), but an extremely low permea- Generally, whether zeolite membranes are water-selec-
tion flux of 5.1 g·m−2·h−1 [173]. When adding 2.5 wt. % tive or organic-selective relies on their hydrophilicity/
n-butanol into a feed of ~ 10 wt. % ethanol solution, the hydrophobicity. Hydrophilic zeolite membranes, such as
permeation flux slightly went up to 17 g·m−2·h−1. How- A- and X-type zeolites, allow preferential permeation of
ever, the separation factor increased conspicuously to as water; whereas hydrophobic zeolite membranes, such as
much as 113, namely, the concentration of organic com- silicalite-1 and ZSM-5, preferentially permeate organics.
pounds in the permeate exceeded 95%. In addition, they During PV separation, permeating molecules first adsorb
evaluated the performance of a thinner trioctylamine liq- into the zeolite pores due to intermolecular attractive
uid membrane, and their results confirmed that the per- forces and then diffuse through the zeolite membrane,
meation flux and separation factor were approximately driven by the chemical potential gradient. It is obvious
65 g·m−2·h−1 and 100, respectively, for separation of a that their separation performance is closely related to
dilute aqueous ethanol (~ 10 wt. %) and butanol mixture the zeolite framework structure and pore size which can
(2 wt. %) [51]. Notably, the major problem restricting the be adjusted by means of the change of the content of Al
widespread application of liquid membranes is stability and other metals substituted into the framework (if any)
because of various losses. Further, the lost liquid mem- as well as preparation conditions during the membrane
brane compounds to the fermentation broth are most fabrication process.
probably toxic to the yeast cells [174]. Accordingly, only The MFI zeolite structure is most commonly deployed
perfectly ensure the stability of liquid membranes and to prepare membranes to date, as compared with many
hence prevent the fermentation broths from contamina- other zeolite structures (more than 14, including MEL
tion can liquid membrane-based PV become a potential [179], MOR [180, 181], CHA [182, 183] and so on). This
separation technology in separating ethanol from aque- is not only due to its medium pore size (nearly 0.55 nm),
ous solutions. but also because this structure is relatively easy to
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 15 of 33
prepare. The MFI structure includes silicalite-1 consist- dry seeds [191, 199]), and vacuum seeding [193] have
ing of pure silica and ZSM-5 in which some Si atoms been reported.
are substituted by Al. The PV performance of MFI-type The dip coating is the simplest and most broadly avail-
membranes for recovering ethanol from aqueous solu- able technique. Shu et al. utilized the dip-coating method
tion is tabulated in Table 4. On account of the presence of to fabricate MFI membranes on yttria-stabilized zir-
hydrophilic silanol groups in structural defects and inter- conia (YSZ) hollow-fiber supports that displayed an
crystalline boundaries in the chemical structure of hydro- extremely high permeation flux of 7400 g·m−2·h−1 cou-
phobic zeolite membranes, their intrinsic hydrophobicity pled with a reasonable high separation factor of 47 at
is not so high as the hydrophilicity of the hydrophilic 60 °C [54]. Wang et al. prepared silicalite-1 membranes
zeolite membranes [184, 185], hydrophilic A-type zeo- on α-alumina hollow fibers by means of the dip-coating
lite membranes in particular. As a result, their separation method. Their membranes possessed both the high per-
performance in the ethanol removal from water is much meance (permeation fluxes of 2900 and 5400 g·m−2·h−1)
worse than that of hydrophilic ones in the dehydration of and the high selectivity (separation factor of 66 and 54)
ethanol [178]. As for this, considerable researches have for the pervaporative recovery of ethanol from aqueous
been conducted to enhance the separation performance solutions (3 wt. % at 60 °C and 5 wt. % at 75 °C) [195].
of hydrophobic zeolite membranes through developing Besides, Kita and co-workers reported intergrown
fabrication methods [186–188], porous supports [47, 53, silicalite membranes on tubular mullite supports syn-
189], membrane post-treatment [190], etc. thesized in ultradilute precursor solutions by this fac-
The fabrication methods of membranes for the PV ile method. The observation showed that the silicalite
application in the recovery of ethanol were tabulated membranes, prepared with either the classical templates
by Elyassi et al., summarizing the works published from (tetrapropylammonium bromide (TPABr) and tetrapro-
1994 to 2015 [191]. As reviewed [53, 187, 192], zeo- pylammonium hydroxide (TPAOH)) or the inexpensive
lite membranes are typically synthesized by two meth- pure TPABr templates, exhibited the high-PV perfor-
ods, namely, direct in situ crystallization and secondary mance. The highest permeation flux and separation fac-
(seeded) growth [60, 193]. The in situ method is that a tor were 1910 g·m−2·h−1 and 66 for the former case, and
porous support is immersed into a precursor growth 1770 g·m−2·h−1 and 63 for the latter case, respectively,
sol or gel that placed in an autoclave, and the zeolite towards a 5 wt. % ethanol/water feed at 60 °C [61]. Yet,
membrane is directly crystallized on the support. Com- despite its advantages, this method is usually applicable
paratively, this method is simple and easy to imple- to the synthesis of zeolite membranes on smooth and
ment, and therefore, is suitable for preparation in a large uniform supports with small pore sizes (generally less
scale. Nevertheless, it is limited by the fact that a dense than 1 μm). It is possibly because, as noted by Wang et al.,
and compact zeolite membrane is challenging because seeding on large-pore (over 1 μm) supports would prob-
of nucleating and growing in the bulk solution simulta- ably lead to a low membrane reproducibility [200]. How-
neously. Moreover, the thickness and orientation of the ever, this is bound to make an augment in not only the
crystal layer can hardly be controlled [53]. capital costs of supports but also the support transport
The secondary growth method, also referred to as two- resistance and sometimes limiting for PV separations.
step crystallization, involves the crystal nucleation and In the rub-coating seeding method, supports are simply
growth steps which were totally separated and carried rubbed with zeolite particles. Kita’s group demonstrated
out independently. In this method, zeolite crystals (typi- that high-PV performance silicalite-1 membranes with
cally nanocrystals) with uniform size were pre-seeded high reproducibility could be achievable by this method
onto the supports prior to the secondary hydrothermal on tubular mullite and alumina supports [53, 198]. The
growth, which is conducive to control the membrane resultant membranes offered an incredibly high separa-
microstructure and orientation of crystal growth, result- tion factor up to 106 with a moderate permeation flux of
ing in a dense zeolite membrane layer and a higher repro- 930 g·m−2·h−1 for ethanol/water separation (at 50 °C).
ducibility [194]. Unfortunately, this method is relatively Precious few works reported in the literature have, so far,
complicated, and cannot be available for large-scale been superior to this one. They also identified that the
application. The quality of the membranes synthesized by rub coating is a straightforward, effective, and reproduci-
this method is extremely correlated with the seed layer. ble seeding method [53]. However, the main vulnerability
Therefore, the seeding is a crucial procedure to obtain a of this method is that the preparation process of zeolite
continuous and uniform seed layer, and thus a homoge- seed particles is too complicated.
neous defect-free zeolite membrane. At present, many More recently, Ueno et al. have reported a novel seed-
seeding techniques such as dip coating [54, 60, 61, 194– ing method to build the silicalite-1 selective layer. A
197], rub coating (with either wetting slurry [53, 198] or zeolite-dispersed polymer film was applied to deposit a
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 16 of 33
Table 4 (continued)
Active layer Support Feed Feed Separation Flux (g·m−2·h−1) References
concentration temperature factor
(wt. %) (°C)
uniform and continuous seed film on a porous tubular rubbing provides an effective strategy to reproducibly
silica support. The superiority of this seeding approach fabricate zeolite membranes on defective low-cost sup-
was not subject to the constraints of the support pore ports, which is industrially attractive.
size and seed crystal size. It would thus make highly Based on the above, the membrane PV performance is
reproducible zeolite membranes with high separation dependent on not only the membrane synthesis condi-
performance even be scaled up. A maximal separa- tions (e.g., time, temperature, and gel composition) and
tion factor of 120 with an associated permeation flux operating conditions (e.g., time, temperature, and com-
of 3160 g·m−2·h−1 was achieved at the operating tem- position of the mixture to be separated), but the sup-
perature of 50 °C for a 10 wt. % ethanol/water mixture. ports as well. For clarity, the supports of MFI membranes
Overall, that is the best MFI membrane performance reported in the literature are also assembled in Table 4.
reported to date for the pervaporative removal of etha- Giaya et al. demonstrated that the hydrophobicity of
nol. In addition, the authors stated that the polymer film- the dealuminated Y (DAY) zeolite enhanced with the
seeding method, given the simplicity and effectiveness, augment of the Si/Al ratio in the zeolite framework [201].
had immense potential in seeding other particles except The dependence of hydrophobicity on the Si/Al ratio was
for zeolite [59]. also presented for other zeolite structures, such as zeolite
To break through the limitations of pore sizes and capi- X, Y, BEA, and mordenite [202–205]. Kovo proved that
tal costs of supports, Wang and co-workers have worked A- and X-type zeolite membranes were both selective
on the preparation of zeolite membranes on inexpen- towards water because of their hydrophilicity occasioned
sive and defective large-pore supports. They developed a by the high Al content, whereas the ZSM-5 membrane
novel seeding method known as wetting–rubbing which showed selectivity towards ethanol due to its hydropho-
involved wetting and rubbing steps (corresponding to bicity associated with the low Al content [204]. As com-
dip coating supports with wetting agents, and then rub pared with the ZSM-5, silicalite-1 is inherently more
coating with dry crystal seeds, respectively). They found hydrophobic attributed to the absence of Al atom in its
that in the case of n-butanol served as the wetting agent, framework. Accordingly, silicalite-1 is the most com-
high-performance membranes (separation factor of 62 monly-used membrane material so far for the separation
and permeation flux of 1820 g·m−2·h−1) were attained on of ethanol from water, as shown in Table 4. Besides, if
tubular α-alumina supports (1–3 μm in average pore size) an alumina support is employed, silicalite-1 membranes
[199]. On the same operating condition, the performance would show a reduction of hydrophobicity owing to the
was as much as 30% higher than that of membranes on dissolution of alumina during hydrothermal synthesis
similar supports by the vacuum-seeding method [193]. [186, 206]. From these views, it appears that Al-free sup-
The method of organic solvent wetting followed by ports, including stainless steel, all-silica, titania, and YSZ,
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 18 of 33
are expected to be the most promising supports for pre- dissolution of ceramic alumina supports, enhancing the
paring pure-silica MFI membranes [190]. For instance, membrane hydrophobicity, and improving the selectivity
Chen et al. prepared a silicalite-1 membrane on a novel for ethanol. However, the high surface to volume ratio of
porous silica tube support instead of an alumina support. capillaries and the rapid growth assisted by microwave
It exhibited high-PV performance with a permeation flux heating tended to form a thinner and probably more
of 560 g·m−2·h−1 and a separation factor of 84 with a feed defective selective layer, indicating that the membranes
of 3 wt. % ethanol/water solution at 60 °C. In addition, were expected to deliver a high permeation flux with a
they indicated that high-performance silicalite-1 mem- sacrifice in separation factor [194].
branes were easier to be prepared with the silica supports More recently, inspired by “like dissolves like” prin-
than others [186]. Interestingly, even less hydrophobic ciple, Huang et al. confirmed that hydrophobic ZIF-8
ZSM-5 membranes, when being deposited on the top of membranes preferred to grow on hydrophobic support
the support which was composed of a titania tube and surface which conformed to a principle as “like grows
three depositing intermediate ceramic titania layers, were like” [208]. In their subsequent work, instead of growing
found to exhibit exceptionally good separation perfor- the crystal layer directly onto the supports, they modi-
mance (permeation flux: 810–11,300 g·m−2·h−1 and sep- fied the hydrophilic porous α-alumina and titania sup-
aration factor: 49.9–97.7) for ethanol enrichment from ports with 1H,1H,2H,2H-perfluoroalkyltriethoxysilanes
round 5 wt. % of aqueous solutions over a temperature (POTS), thereby converting them into superhydropho-
range from 40 to 120 °C [47]. bic ones. Then, through in situ hydrothermal synthe-
In addition to the above-mentioned common fabrica- sis, dense, well-intergrown, and phase-pure MFI zeolite
tion methods, other approaches have also been proposed membranes were obtained. As mentioned previously,
in recent years. For example, a two-step in situ hydro- because of without the alumina dissolution, it was proved
thermal synthesis approach was applied to fabricate sili- that their silicalite-1 membranes on POTS-modified tita-
calite-1 membranes. Before each hydrothermal synthesis, nia disks had an attractive separation factor/permeation
a “solution-filling (SF)” method, filling porous supports flux combination (as high as 103 and 2560 g·m−2·h−1,
with a viscous mixture composed of water and glycerol, respectively), which was almost a doubling of the per-
was also performed to protect the supports from invad- formance of the ZSM-5 membranes on POTS-modified
ing synthesis solutions. It was found that the permeation α-alumina disks. They suggested that this principle could
flux with SF pre-treatment was approximately twice that be liable to scale-up in the form of tubular configuration.
without SF pre-treatment, and the silicalite-1 membranes In addition, from the industrial application of view, sur-
supported on silica tubes displayed high-reproducibility face modification could be effective for batch processing
and consistently high separation performance (permea- of supports, irrespective of length, size, and shape [209].
tion flux: 870 g·m−2·h−1 and separation factor: 69) for Different from Kita et al.’s observation that the separa-
ethanol extraction from water at 60 °C [187]. tion factor for randomly oriented silicalite-1 membranes
Another interesting work by Soydas and co-work- was higher than that for oriented ones [53], Elyassi et al.
ers revealed a method to synthesize MFI membranes developed b-oriented silicalite-1 membranes supported
by recirculating synthesis solutions through tubu- on silica discs by gel-free secondary growth, and obtained
lar α-alumina supports. Their membranes had a a very high ethanol/water separation factor of 85 and a
200 g·m−2·h−1 permeation flux and a 43 separation factor sustainable high permeation flux of 2100 g·m−2·h−1 at
in 5 wt. % ethanol feed concentration when operating at 60 °C [191]. Likewise, both c-oriented [210] and c-/h0h-
25 °C. The authors also noted that the membrane formed out-of-plane-oriented [211] silicalite-1 membranes
in the recirculating flow system showed a higher ethanol/ showed good separation performance, particularly the
water separation factor than that prepared in the conven- permeation flux of the former as high as 9800 g·m−2·h−1
tional batch system, despite the similarity in permeation at 60 °C for a 5 wt. % ethanol/water feed [210]. It is sug-
flux. They attributed this to the utilization of recirculat- gested that oriented zeolite membranes appear to be
ing flow synthesis system which provided a more uni- promising and look worthy of further investigation.
form fabrication environment around the supports [207]. Furthermore, metals substituted into the zeolite frame-
Meanwhile, microwave-assisted hydrothermal synthe- work would change its pore size and intrinsic hydro-
sis was pioneered by Sebastian et al. They fabricated sili- phobicity, thereby affecting diffusion and adsorption
calite-1 membranes on ceramic α-alumina capillaries by properties. Some researchers have attempted to intro-
seeded secondary growth under microwave irradiation. duce some metallic elements, such as boron (B), germa-
By integrating the microwave-heating technique, the syn- nium (Ge), iron (Fe), titanium (Ti), and zirconium (Zr),
thesis duration was greatly shorted down to 2 h from a into MFI structure for the application of PV in the recov-
typical period of 8 h up to 3 days, thereby preventing the ery of ethanol.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 19 of 33
Noble et al. incorporated Al, Fe, B, and Ge into the simple modification technique to eliminate hydrophilic
ZSM-5 framework by isomorphous substitution of Si to silanol groups in zeolite structures, enhancing its hydro-
prepare zeolite membranes on asymmetric stainless steel phobic nature, hence augmenting its membrane perfor-
tubes. The Ge-ZSM-5 membranes had both the highest mance. For example, Sano and co-workers conducted the
permeation flux and ethanol separation factor, whereas silylation of silicalite membranes using silylating agents
the other Al, Fe, and B-substituted ZSM-5 membranes (octyltrichlorosilane and octadecyltrichlorosilane). It
all showed lower PV performance than silicalite-1 mem- was interesting to find that the separation factors of
branes as PV at the same condition [212–214]. Tetrava- silylated silicalite membranes were markedly improved
lent Ge is chemically similar to Si, but the Ge atom is to 20–44, which is around fourfold to ninefold higher
larger than the Si atom, resulting in smaller pores than than that without silylation. Unfortunately, they found
silicalite-1 [215] In all probability, the change of the zeo- that there was a sacrifice in permeation flux, declining
lite structure influences the hydrophobicity [178]. On noticeably from 843 to 53.4 g·m−2·h−1 [220, 221]. More
the other hand, trivalent metal (Al, Fe, and B) substitu- recently, Wu et al. opened up a novel method of masking
tion destructs the charge balance of zeolite. The result- the silanol groups on the zeolite surface. They modified
ing negatively charged zeolite framework involves an the MFI zeolite membranes with dopamine. After dopa-
increase in the local polarity in the pores, thereby making mine modification, the initial ethanol separation factor
it hydrophilic. It seems that ZSM-5 zeolites substituted was similar to that of unmodified membranes; whereas
by trivalent metals are not desirable materials for the the initial permeation flux reduced by about 50%. How-
separation of ethanol from water. In reality, the B-ZSM-5 ever, it was proved that the long-term PV stability of the
zeolite membranes on alumina-coated SiC multi-channel dopamine-modified membranes was improved appar-
monolith supports were prepared by Bowen et al. from ently. Over 180 h of exposure to the feed at 60 °C, the
a gel with a H2O/SiO2 ratio of about 22, and exhibited separation factor for unmodified membranes was down
an optimal separation factor of 31 and a permeation to 1 from 43, meanwhile the permeation flux to 500 from
flux of 160 g·m−2·h−1 for a 5 wt.% ethanol feed [216]. 4900 g·m−2·h−1; but the separation factor and permea-
Saboor et al. reported that B-ZSM-5 membranes fab- tion flux for the modified ones maintained almost con-
ricated from clear synthesis solutions (H2O/SiO2 = 64) stant values of 44 and 2600 g·m−2·h−1, respectively [190].
on seeded α-alumina discs preferentially permeated Besides MFI-type zeolite membranes, other type zeo-
ethanol from water, showing a relatively high permea- lite membranes have also been investigated for ethanol
tion flux of 1110 g·m−2·h−1 but a low separation factor recovery. Li et al. incorporated B and Al atoms into MEL-
of 13.9. Recently, Chai and co-workers observed a high type zeolite ZSM-11 framework structure and prepared
permeation flux of 2600 g·m−2·h−1 and a moderate sepa- B and Al-substituted ZSM-11 membranes on porous
ration factor of 55 using a B-ZSM-5 membrane synthe- tubular supports. The separation factors through the Al-
sized from a dilute solution (H2O/SiO2 = 600) on seeded ZSM-11 membranes were less than 6 with a wide range
inexpensive macroporous α-alumina tube [217]. It sug- of permeation fluxes (210–1800 g·m−2·h−1); whereas
gests that the PV performance for B-ZSM-5 membranes the maximal separation factor through the B-ZSM-11
is greatly dependent on their preparation conditions. membranes reached a value of 42, coupled with a per-
Nevertheless, it remains lower than that of the most sili- meation flux of 930 g·m−2·h−1 at 5 wt. % of ethanol in
calite-1 membranes. the feed at 60 °C [222]. Similarly, Chai et al. also reported
On the other hand, Chen et al. incorporated Ti atoms the B-ZSM-11 membranes for this application, and they
into silicalite-1. They found that the resultant Ti-sili- obtained a similar observation for the PV performance
calite-1 membranes had a higher ethanol separation per- (highest separation factor of 35 with a permeation flux of
formance than the similar silicalite-1 membranes. For 1510 g·m−2·h−1) [179]. Other work on organic/water sep-
separating 5 wt. % ethanol/water feed at 60 °C, the sepa- aration by mesoporous silica MCM-48 membranes was
ration factor and permeation flux reached the highest carried out by Kim et al. However, the separation factor
values, 127 and 770 g·m−2·h−1, respectively [218]. Nota- for pure MCM-48 membranes was lower than 1. After
bly, the Ti-silicalite-1 membrane is currently the best silylation, the selectivity was facilitated, but the permea-
performing inorganic membrane for ethanol recovery. tion flux descended considerably as a result of water flux
Later, they continued their work on Zr-substituted sili- sharp decline [223].
calite-1 membranes, and observed a high ethanol/water Other than zeolite-based membranes, researchers have
separation factor of 73 with a corresponding permeation also attempted to develop new inorganic membranes.
flux of 1010 g·m−2·h−1 [219]. For instance, the boehmite–sol-coated membranes were
Post-treatment has also exerted a positive impact on reported by Song et al. for the separation of ethanol/
membrane performance. Silanization is an effective and water mixtures. The results, disappointingly, were far
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 20 of 33
from being satisfactory. The separation factor and per- application, several critical issues have to be overcome
meation flux were as low as 1.14 and 31.5 g·m−2·h−1 for as yet, such as membrane reproducibility, long-term sta-
a fivefold coated-membrane, and 1.76 and 12 g·m−2·h−1 bility, manufacture cost—10–50 times that of polymeric
for a tenfold coated-membrane [224]. In addition, MOFs membranes [228], and even the permeation flux which
including its subclass ZIFs have been utilized in the fab- is generally low and should be improved further. More
rication of organophilic PV membranes. Dong and Lin efforts are needed to make the inorganic membranes
successfully fabricated integrated ZIF-71 membranes on industrialized practically in clean bioethanol production.
ZnO disk supports by the reactive method, and demon-
strated the feasibility of ZIF membranes for PV separa-
tion of organic solvent. For PV of a 5 wt. % ethanol/water Mixed matrix membranes
solution at 25 °C, a similar observation for the separa- Mixed matrix membranes (MMMs), whose concept
tion performance to the pure PDMS membranes was can be traced back to the mid-1980s [242, 243], are also
obtained, showing a separation factor of 6.07 and a lim- known as hybrid membranes. MMMs are generally fab-
ited permeation flux of 322 g·m−2·h−1 [225]. Huang et al. ricated by dispersing inorganic particles (so-called fill-
also prepared ZIF-71 membranes on ceramic α-alumina ers) into continuous polymer matrices. MMMs integrate
hollow-fiber supports with a modified contra-diffusion their strengths, such as high separation performance as
method. It was observed that the ZIF-71 membranes well as high stability of inorganic membranes, and easy
were of very high integrity and exhibited a very competi- fabrication as well as low costs of polymeric membranes.
tive permeation flux of 2601 g·m−2·h−1, combined with a MMMs provide a flexible and cost-effective way to over-
separation factor of 6.88 as PV at the same temperature come the trade-off behavior in polymeric membranes,
and ethanol concentration in feed with Dong’s work. obtaining high permeation flux and high separation fac-
They explained that the inspiring result could be attribut- tor simultaneously. MMMs have been considered as a
able to the low resistance of ceramic hollow-fiber support promising and competitive candidate for ethanol recov-
as well as the thin selective layer (approximate thickness ery by PV. Over the past 3 decades, they have gained
2.5 µm) [226]. More recently, the Zr-based MOF struc- increasing interest and become a research focus in mem-
ture of UiO-66 was made into membranes. The effective brane separation fields. In addition, the rapid develop-
membrane thickness was merely 0.5–1 μm. The thin- ment of MMMs provides a new thought for the design
ner membranes displayed highly stable PV performance and preparation of new PV membranes.
and had elevated permeation fluxes of 1490 g·m−2·h−1 Based on the literature reports, the most common
at 50 °C and 3150 g·m−2·h−1 at 60 °C, respectively, with MMMs for ethanol-selective PV are still based on the
separation factors slightly less than 5 in a 10/90 etha- PDMS matrix. Its source including components and
nol/water mixture [227]. Despite unique advantages of composition is important in determining the perfor-
MOFs, such as high porosity, ordered structures, and mance of MMMs [244]. Most PDMS used for ethanol-
tunable chemical functionality, a few MOFs membranes selective MMMs are two-component vinyl-terminated
have been investigated into PV to date as a result of their General Electric RTV 615 [76, 77, 82, 245] and hydroxyl-
poor chemical stability. In fact, MOFs have mainly been terminated PDMS produced in China [21, 72, 126, 246].
used to fabricated mixed matrix membranes [21, 87]. Generally, it seems that vinyl end-capped PDMS MMMs
Anyway, all these studies are significative attempts for are capable of delivering higher ethanol selectivity. The
designing and fabricating new inorganic membranes, and vinyl-based PDMS involves the hydrosilylation reac-
further research works are expected. tion between the vinyl groups of prepolymer (RTV 615
In summary, the separation performance of inorganic part A) and methyl-hydride groups of silicone copoly-
membranes typically outperforms those of polymeric mer crosslinking agent (RTV 615 part B). This system is
ones. As shown in Table 4, the reported ethanol/water highly advantageous to improve the viscosity of the mem-
separation factors for MFI zeolite-based membranes brane casting suspension, leading to the increase in shear
range from 1.3 to 127, with permeation fluxes rang- stress to break down particle aggregations which makes
ing from 0.1 to 51,600 g·m−2·h−1, which is superior to the particles more likely to be surrounded individually
the performance of all polymeric membranes including by polymer chains, thus increasing the maximum prac-
PDMS, PTMSP and so on. However, inorganic materials ticable particle loadings which results in an increase in
have some serious intrinsic drawbacks, such as poor film- selectivity [19]. Other polymer matrices reported include
forming capability and high brittleness, and hence are PTMSP [55, 90, 151, 153], PEBA [103, 247], PMPS [97],
very difficult to be made into large-area defect-free mem- PIM [20], etc. A list of literature performance of various
branes, still under studying in the laboratory for ethanol polymer-based MMMs for ethanol recovery from water
recovery. Moreover, from the perspective of industrial is assembled in Table 5.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 21 of 33
PDMS/silicalite-1
PDMS/PEI Silicalite-1 (77 wt. %) 7 22 59 71 [19]
PDMS Silicalite (60 wt. %) 5 50 21 ~ 105 [76]
PDMS Silicalite-1 (40 wt. %) 5 50 17.9 ~ 60 [73]
PDMS/alumina Silicalite-1 (30 wt. %) 4 25 16.5 ~ 125 [75]
PDMS Silicalite-1 (30 wt. %) 6 40 14.9 51 [84]
PDMS/PTFE Silicalite-1 (30 wt. %) 5 50 13 39 [132]
PDMS/PES Silicalite-1 (2 wt. %) 5 40 10.9 635 [126]
PDMS Silicalite-1 (30 wt. %) 6 35 ~ 10 ~ 70a [86]
PDMS/PSF Silicalite-1 (50 wt. %) 4.8 60 7.5 231 [262]
PDMS (commercial MMM) Silicalite-1 (50 wt. %) 6 35 ~ 7.2 ~ 55 [86]
PDMS/PI Silicalite-1 (15 wt. %) 3 41 4.8 170 [74]
PDMS/modified silicalite-1
PDMS VTES-modified silicalite-1 (67 wt. 5 50 34.3 176 [80]
%)
PDMS Acid- and steam-treated silicalite-1 4.4 50 29.3 120 [155]
(50 wt. %)
F-PBZ modified PDMS/PTFE VTMS-modified silicalite-1 (30 5 50 28.7 207 [132]
wt. %)
PDMS/PVDF VTES-modified silicalite-1 (67 wt. 5 50 26.3 1940 [77]
%)
PDMS/PVDF Dodecyltrichlorosilane-modified 5 40 19.9 66.3 [81]
silicalite-1 (50 wt. %)
PDMS Silylated nano-sized silicalite-1 (40 6 35 16.4 ~ 86 [125]
wt. %)
PDMS/PSF TMDS-modified silicalite-1 (50 5 60 14.7 400 [78]
wt. %)
PDMS/ZSM-5
PDMS ZSM-5 (60 wt. %) 5 50 ~ 37 ~ 380 [82]
PDMS/PVDF ZSM-5 (30 wt. %) 5 50 13.7 821 [263]
PDMS/PVDF/polyester nonwo- ZSM-5 (20 wt. %) 5 40 10.4 ~ 210 [257]
ven fabric
PDMS ZSM-5 (30 wt. %) 6 35 ~ 6.3 ~ 55a [86]
PDMS ZSM-5 (CBV 28,014) (60 wt. %) 5 50 39 66 [248]
PDMS ZSM-5 (CBV-28014) (50 wt. %) 5 50 37 195 [244]
PDMS with a pure PDMS top ZSM-5 (CBV-28014) (65 wt. %) 5 50 18 ~ 520 [44]
coat/PVDF
PDMS/PI ZSM-5 (CBV-3002) (30 wt. %) 3 41 5.5 151 [74]
PDMS/modified ZSM-5
PDMS/PVDF HF etched ZSM-5(30 wt. %) 5 50 16.7 134 [83]
PDMS/PVDF/polyester nonwo- Dodecyltrichlorosilane-modified 5 40 15.8 203 [257]
ven fabric ZSM-5 (30 wt. %)
PDMS/CA APTS-modified ZSM-5 (20 wt. %) 10 40 14.1 348 [79]
PDMS/ceramic ZSM-5 first grafted with OTES and 5 40 14 408 [260]
then coated with a thin PDMS
layer (40 wt. %)
PDMS/other zeolites
PDMS Ultrastable zeolite type Y (50 wt. –- 30 16.1 610 L·m−2·h−1 [92]
%)
PDMS Hollow spheres with silicalite-1 6 40 15.3 72 [84]
shell (30 wt. %)
PDMS Zeolite (TZP-9023) (30 wt. %) 10 25 12.5 332 [261]
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 22 of 33
Table 5 (continued)
Polymer membrane material/ Filler (loading) Feed Feed Separation Flux (g·m−2·h−1) Reference
support concentration temperature factor
(wt. %) (°C)
Table 5 (continued)
Polymer membrane material/ Filler (loading) Feed Feed Separation Flux (g·m−2·h−1) Reference
support concentration temperature factor
(wt. %) (°C)
After dozens of years of development, a good number solution (or pre-dispersed in a solvent) with thoroughly
of fillers have been incorporated into polymer matrices stirring and ultrasonication [21, 78, 81]. The sonication
for ethanol recovery. The most commonly used fillers are process can be carried out using an ultrasonic bath or
zeolites including silicalite-1 [19, 73, 75, 245] and ZSM-5 a probe-type sonicator. The probe-type sonicator has a
[82, 92, 248, 249]. Other fillers reported are MOFs [97, high localized intensity compared to the tank-type and
250] and its subclass ZIFs [87, 251, 252], fumed silica [55, hence bigger crushing force, resulting in finer particles
88, 93], carbon nanotubes [98, 175], carbon blacks [99, and more uniform particle dispersion. Vane et al. con-
100], POSSs [103, 246], etc. Taking into account the filler firmed that sonication was effective to disrupt particle
particles, filler type, particle size, and particle dispersion agglomerations. Due to defects triggered by substantial
status in MMMs are crucial for the PV separation prop- particle agglomerates for membranes fabricated without
erties. To form desirable thin, smooth, homogeneous, sonication, more instances of poor separation perfor-
and high-performing MMMs, works could focus on fol- mance or failure appeared. Their study also verified that
lowing approaches: (a) achieving a uniform dispersion of the effectiveness of probe-type sonicators at particle dis-
filler particles; (b) eliminating undesirable organic–inor- persal was obviously better than that of ultrasonic baths
ganic interfacial defects; (c) increasing filler loadings (but [244]. Zhang and co-workers proposed in situ ultrasonic
do prevent excessive loading); and (d) optimizing particle strengthening assembly in which sonication was applied
sizes. not only during mixing but during assembly to avoid
The uniform dispersion of inorganic particles in the secondary aggregation on supporting surface. It was
polymer matrix is one of the greatest challenges. Usu- found that the dispersion of nanofillers (particle size of
ally, most of inorganic particles are physically mixed with 7–40 nm for silica nanoparticles and about 100 nm for
polymer matrix. Since the differences in physicochemical ZIF-8 nanoparticles) in PDMS was markedly improved
properties between them, the particles have a tendency [93].
to agglomerate, and the agglomeration is more severe for Minimizing particle size is theoretically possible to
nanoparticles on account of their higher surface energy make much thinner defect-free membranes, which is
[253]. The particle aggregation may induce the formation advantageous to improve the membrane separation per-
of nonideal-defects (e.g., nonselective voids [254]), which formance. However, the tendency of particles to agglom-
often leads to a decrease of selectivity. To restrict the erate which is detrimental to the performance has an
agglomeration and enhance the homogeneous distribu- inverse relationship to the particle size [244]. Besides,
tion of particles, various methods have been developed, the impact of agglomeration caused by particle load-
such as sonication [73, 75, 97, 248], optimization of par- ings should be taken into consideration [259]. Li et al.
ticle size [244, 255, 256], surface chemistry modification embedded micron- and nano-sized ZIF-71 particles
of particles [76, 80, 83, 257], and pre-polymerization of into PDMS to prepare MMMs. They demonstrated that
polymer [19, 77]. the effect of particle size in MMMs was remarkable.
It is often difficult to achieve a homogeneous disper- Micron-sized ZIF-71 particles were more prone to cre-
sion only by simple mechanical stirring [258]. In prac- ate defects in MMMs, especially at higher loadings. In
tice, the fillers are usually dispersed in a polymer dope contrast, nano-sized ZIF-71 would be more preferable
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 24 of 33
for preparing MMMs with higher loadings (more than 20 avoid the polymer chains from infiltrating into the pores
wt. %) but fewer defects [259]. Similar to Li et al.’s obser- of carbon blacks [100]. However, when the viscosity of
vation, Wee et al. found that submicrometer-sized ZIF- membrane solutions was too high after pre-polymeriza-
71 particles were evenly dispersed in PDMS membranes tion, it would render membrane fabrication more diffi-
with no interfacial voids nor ZIF-71 aggregates, while cult because of shorter working time available for casting
MMMs containing micrometer-sized ZIF-71 were rough [77, 244].
and uneven with large interstitial voids. Correspond- Incompatibility between fillers and polymer matrix is
ingly, the submicrometer-sized (290 nm) ZIF-71 filled another critical issue. To enhance their compatibility, a
PDMS MMMs delivered a maximal separation factor of variety of approaches have been proposed and demon-
10.1 which outperformed that of 7.6 for the ones filled strated: (i) surface chemistry modification of inorganic
with micrometer-sized (1–2 µm) ZIF-71 [255]. The study particles with coupling agents [77, 80, 132, 257]; (ii) sur-
carried out by Yin et al., however, showed that micron- face coating of inorganic particles with a thin polymer
sized (1 µm) ZIF-71 particles had less agglomeration and layer [260]; and (iii) employment of hybrid particles (i.e.,
better dispersion in PDMS MMMs than smaller ZIF-71 MOFs [21, 87, 250, 251] and POSS [103, 246]).
particles (~ 150 and 500 nm). The PV performance of Silylation is the most widely used surface modification
resultant MMMs containing micron-sized ZIF-71 out- method of inorganic particles to ameliorate the interfa-
performed those prepared with smaller particles in terms cial compatibility with polymeric matrix as well as to
of both selectivity and permeability due to the less non- facilitate the particle dispersion. Several groups con-
selective and tortuous pathways through the MMMs that ducted silylation of zeolites by employing silane-cou-
provided less mass transfer resistance [256]. Similarly, pling agents to prepare modified zeolite/PDMS MMMs
Vane et al. also observed that membranes made with [78–81, 257, 260]. The silane chains immobilized onto
micron-sized (2.4 µm) commercial ZSM-5 (Zeolyst CBV- the zeolites through hydrolysis and condensation reac-
28014) exhibited higher ethanol separation performance tion could entangle with PDMS chains [80, 260], which
than those made with submicrometer-sized (0.35 and offered good interaction between the zeolite particles and
0.70 μm) silicalite-1 because of their tendency to form the PDMS matrix. Moreover, vinyl silanes, such as vinyl-
silicalite-1 aggregates, especially for particle loadings of triethoxysilane (VTES) [80] and vinyltrimethoxysilane
50 wt. % or higher [244]. (VTMS) [132], were the representative of silane-coupling
Pre-polymerization or pre-crosslinking of polymer has agents since strong chemical bonds were formed by the
also been developed to assist fillers to disperse homog- reaction of their terminal vinyl groups with Si–H groups
enously in polymer network. It not only prevented fillers on the short cross-linker (RTV 615 B). That intuitively
from agglomeration or sedimentation in casting suspen- suppressed the formation of nonideal micro-voids at
sions, but also inhibited infiltration of polymer chains the particle–polymer interface and mitigated membrane
into filler pores. Jia et al. performed pre-crosslinking to swelling [76]. Zhuang et al. modified silicalite-1 parti-
partially polymerize two components of PDMS with sili- cles with four alkoxysilanes (VTES, ethyltriethoxysilane,
calite-1 fillers present. Such a pre-crosslinking process octyltriethoxysilane, and octadecyltriethoxysilane) and
increased viscosity of casting suspensions and improved investigated their effect on the performance of PDMS-
their stabilization. The resultant membranes displayed a based MMMs. It was confirmed that the VTES PDMS
maximum separation factor of 59 at 22 °C. This perfor- membranes embedded with modified silicalite-1 dis-
mance is the highest of any MMMs reported so far. It, played the best ethanol selectivity as a consequence of
however, combined with a particularly low permeation the best compatibility with PDMS just because of the
flux of 71 g·m−2·h−1 [19]. Zhou et al. employed the pre- chemical linking between VTES-modified silicalite-1 and
polymerization of PDMS polymer network to facilitate PDMS. The resulting MMMs coupled with VTES at 67
the dispersion of silicalite-1. It was demonstrated that 67 wt. % silicalite-1 loading had a separation factor higher
wt. % loading modified silicalite-1 maintained a uniform than 34.3, while the unmodified membranes at 60 wt.
dispersal in PDMS solution. A 5-μm-thick modified sili- % silicalite-1 loading had a separation factor of 23 [80].
calite-1 filled PDMS membrane had a high permeation Similar results were observed by Yi et al., and they found
flux of 5520 g·m−2·h−1 with a separation factor of 15.5 that silylation modification of silicalite-1 with VTES rose
for the pervaporative recovery of ethanol from aqueous the maximum loading to 67 from 60 wt. %, accompanied
solutions (5 wt. %) at 50 °C [77]. Instead of adding fill- with an obvious increment in separation factor from 21
ers into polymer solutions before pre-crosslinking as Jia to 32 [76].
et al. and Zhou et al. did, Vankelecom et al. incorporated The increase of particle loadings in MMMs generally
carbon black particles after pre-crosslinking of PDMS to has a positive effect on separation performance in terms
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 25 of 33
of separation factor, while permeation flux may increase permeation flux and separation factor was observed.
simultaneously [75, 251, 261] or reduce due to intrin- Despite the monotonous increase in permeation flux,
sic trade-off effect [132, 260]. However, excessive addi- the separation factor declined, which was mainly attrib-
tion would definitely lead to particle agglomeration and uted to the change of MMM hydrophobicity as well as
formation of nonselective pores and defects, thus dete- the formation of defects due to severe particle agglom-
riorating the performance. Overall, increasing filler load- eration [21]. Another work from Khan and co-workers
ings is based on good dispersion and good compatibility. confirmed that the ZIF-67/PDMS membranes with 20
Although the highest zeolite particle loading in ethanol- wt. % of loading showed a simultaneous increment in
selective MMMs could reach up to 77 wt. % [19], the both permeation flux and separation factor, exhibiting a
optimal loadings of zeolites reported are usually limited good anti-trade-off phenomenon. Their values were up to
to 50 wt. % [92, 244, 262] or less [73, 86, 263]. After the 2780 g·m−2·h−1 and 15.4—triple and double that of pris-
silylation modification of zeolites, the maximum loadings tine PDMS membranes, respectively [251].
have been obtained to be as high as 67 wt. % [76, 77, 80]. In summary, the separation performance of MMMs
MOFs, developed rapidly in recent years, have garnered for ethanol recovery is generally superior to that of
extensive attention for their porous structure and good polymeric membranes but worse than that of inorganic
compatibility with polymer matrix. They are introduced membranes alone. Moreover, some MMMs still display
into matrix ranging in 15–40 wt. % [87, 95, 250, 251]. The a trade-off relationship in separation performance. Till
optimal silica loadings range from 5 [93, 264] to 30 wt. now, only a minority of MMMs have been commercial-
% [130]. A higher amount of silica (50 wt. %) was doped ized, such as silicalite-1/PDMS membranes marketed
into PTMSP by Claes et al. via surface hydrophobic treat- by Sulzer Chemtech [86]. However, MMMs have their
ment [151]. Other particle loadings are generally no more comprehensive advantages, such as desirable mechani-
than 10 wt. % [98, 103, 246, 265], even lower than 0.5 wt. cal and long-term operation stability, anti-swelling prop-
% [20, 175]. erty, economical processability, and easy to large-scale
From the summary of pervaporative separation data production. They are a promising strategy for ethanol
for MMMs listed in Table 5, the ethanol/water separa- recovery from the bulk feed by PV. Further efforts should
tion factors ranging from 3.6 to 59 can be observed. This be devoted to design new nanoscale filler particles with
range is between those of pure polymeric membranes superhydrophobic pores as well as excellent interfacial
and pure inorganic membranes. On the other hand, the compatibility with polymers and explore new prepara-
permeation fluxes vary over a fairly wide range, from tion processes, breaking the trade-off effect, thus obtain-
39 to 9500 g·m−2·h−1, which is still dependent on the ing desirable competitive MMMs with judiciously higher
polymers used (continuous phase) to a larger extent. permeation flux and separation factor.
For representative zeolite/PDMS systems, the ethanol/
water separation factors are much higher than that of Conclusions and perspectives
pure PDMS, while the permeation fluxes are still lim- This review summarizes the research progresses on
ited within 1000 g·m−2·h−1. For instance, Zhuang et al. ethanol-selective pervaporative membranes from the
embedded VTES-modified silicalite-1 into PDMS, and perspectives of transport mechanisms, fabrication meth-
the resultant MMMs with 67 wt. % modified silicalite-1 ods, and membrane materials. The facilitated transport
loading exhibited a high separation factor of 34.3 in com- mechanism and solution–diffusion mechanism have
bination with a permeation flux of 176 g·m−2·h−1 at 50 °C been mentioned to understand transport and separation
for an ethanol/water mixture at 5 wt. % feed concentra- fundamentals so as to better guide membrane design,
tion [80]. In comparison, the doping of MOFs in MMMs and therefore, govern membrane performance. The
is more beneficial to acquire high permeation flux. Zhang membrane materials (polymeric, inorganic, and hybrid
et al. incorporated materials institute Lavoisier-53 (MIL- materials) with various modification approaches and
53) particles into PDMS to form MMMs. They proved the fabrication methods including solution casting and
that the permeation flux showed a considerable increase physical blending to prepare PV membranes have been
from 1667 g·m−2·h−1 for neat PDMS dense membranes overviewed. As reviewed here, although some encour-
to 5467 g·m−2·h−1 for PDMS MMMs containing 40 aging results have been achieved in recovering ethanol
wt. % MIL-53, with a slight increase in separation fac- from aqueous solution via PV after many decades of
tor from 7.6 to 11.1. The gain in permeation flux was extensive research worldwide, it is still rather far from
largely accounted for the water-repellency surface and industrial applications because of the lack of industri-
ethanol-affinity channels of MIL-53. When the MIL-53 ally practical membrane modules. Therefore, further
content exceeded 40 wt. %, a trade-off behavior between research is required to explore high-performing, stable,
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 26 of 33
and economic ethanol-selective membranes for large limited. The current research status is that most of the PV
applications. Advances in the following aspects have the experiments reported have been conducted at laboratory
potential to impel the development and deployment of scales and performed with simple and ideal binary aque-
PV membranes. ous solutions, thus resulting in a lack of understanding
Developing new high-performance membrane materi- on membrane performance in real fermentation broths.
als is an effective strategy. Fabricating membranes with Long-term trial evaluation of membranes is preferable
organic–inorganic nanohybrids is probably the most to be implemented at actual industrial conditions where
efficient approach to improve the performance stabil- PV integrates with fermentor. In addition, more efforts
ity and to break the permeability-selectivity trade-off at should be focused on enhancing the long-term operation
this stage. Future research in this area may focus on the stability of membranes including selectivity, chemical
exploration of new nanometer-sized inorganic particles and temperature resistance, and robustness, in particular
(or hybrid particles) with ultra-hydrophobic pores and polymeric and mixed matrix membranes. Future stud-
extremely high compatibility with polymers. In addition, ies on ethanol perm-selective PV membranes should be
intelligent materials may be a prospective membrane positioned toward increasing membrane permeability
material for ethanol recovery in the future. By adjust- and maintaining a satisfactory selectivity so as to reduce
ing the intensity of external light, electric field, magnetic required membrane area.
field, etc., the microstructure of membranes could be Overall, exploring the possibility and limitation of
regulated, and therefore, improving membrane recovery the separation performance of PV membranes for etha-
performance. In the meantime of exploring new mem- nol extraction is a long-standing topic. Collectively, the
brane materials, the modification of existing ones may quest is to break the trade-off between membrane per-
serve as a promising alternative method. meability and selectivity. Based on the facilitated trans-
Besides, exploring novel membrane fabrication pro- port mechanism, further exploration of ethanol-selective
cesses is another strategy. Efforts should be focused membranes may focus on constructing a well-designed
on preparing a large-surface-area, defect-free, and microstructure, providing active sites for facilitating the
ultrathin selective skin layer. In situ synthesis and fast transport of ethanol molecules, hence achieving both
layer-by-layer self-assembly method may be further high selectivity and permeability simultaneously. Finally,
extended. Meanwhile, establishing a hierarchical mem- it is expected that more and more successful research
brane configuration can also be considered. Construct- could be realized into commercial products and this sep-
ing a thin protective layer on top of the selective layer aration process will be deployed in industrial practices in
may be a viable way to not only suppress the swell- the near future.
ing of the selective layer but also prevent or mitigate
Acknowledgements
membrane fouling. In practice, the development of Not applicable.
manufacturing methods for environmentally friendly,
high-efficient, and economic synthetic membranes Authors’ contributions
PP and YL conceived the study. Ping Peng wrote the draft of the manuscript.
is encouraged. More attention should be paid to bio- LL and KJ collated the data. PP critically reviewed the full manuscript content.
mimetic methods such as bioinspired mineralization All the authors read and approved the final manuscript.
in which the dispersed particles in situ generate from
Funding
their precursors under a mild condition and have a The authors acknowledge the financial supported by the Key Laboratory
fine dispersion. Innovative methods to manipulate the of Bio-based Material Science & Technology (Northeast Forestry University),
microstructure of organic–inorganic nanohybrid mate- Ministry of Education (No. SWZ-MS201907), the National Natural Science
Foundation of China (No. 51903138), the Natural Science Foundation of Fujian
rials for forming membranes with high-loading and Province (No. 2020J01380 and 2020J01381), and the Education Scientific
uniformly dispersed filler particles as well as super- Research Project of Youth Teacher in the Education Department of Fujian
hydrophobic surfaces are expected to be developed Province (No. JAT190691-B201904).
actively. Other constructive combination between the Availability of data and materials
polymer and inorganic particles, rather than direct All data generated or analyzed during this study are included in this published
dispersion, should be tried. In addition, much effort article.
should be carried out to promote the reform of the Ethics approval and consent to participate
existing fabrication procedures. For example, optimiz- This study is based on a survey answered by people anonymously, this decla-
ing post-modification processes could be employed to ration is not applicable.
introduce active groups on the membrane surface to Consent for publication
increase the affinity of ethanol toward it. Not applicable.
From the standpoint of industrial application, ethanol
Competing interests
removal from fermentation broths via PV is still rather The authors declare that they have no competing interests.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 27 of 33
45. Mohammadi T, Aroujalian A, Bakhshi A. Pervaporation of dilute alco- 67. Lan Y, Yan N, Wang W. Application of PDMS pervaporation membranes
holic mixtures using PDMS membrane. Chem Eng Sci. 2005;60:1875–80. filled with tree bark biochar for ethanol/water separation. RSC Adv.
46. Luo Y, Tan S, Wang H, Wu F, Liu X, Li L, Zhang Z. PPMS composite mem- 2016;6:47637–45.
branes for the concentration of organics from aqueous solutions by 68. Santoro S, Galiano F, Jansen JC, Figoli A. Strategy for scale-up of SBS
pervaporation. Chem Eng J. 2008;137:496–502. pervaporation membranes for ethanol recovery from diluted aqueous
47. Weyd M, Richter H, Puhlfürß P, Voigt I, Hamel C, Seidel-Morgenstern solutions. Sep Purif Technol. 2017;176:252–61.
A. Transport of binary water–ethanol mixtures through a multilayer 69. Li J, Ji S, Zhang G, Guo H. Surface-modification of
hydrophobic zeolite membrane. J Membr Sci. 2008;307:239–48. poly(dimethylsiloxane) membrane with self-assembled monolayers for
48. Li J, Wang N, Yan H, Zhang G, Qin Z, Ji S, Zhang G. Roll-coating of alcohol permselective pervaporation. Langmuir. 2013;29:8093–102.
defect-free membranes with thin selective layer for alcohol permselec- 70. Wei W, Xia S, Liu G, Dong X, Jin W, Xu N. Effects of polydimethylsi-
tive pervaporation: From laboratory scale to pilot scale. Chem Eng J. loxane (PDMS) molecular weight on performance of PDMS/ceramic
2016;289:106–13. composite membranes. J Membr Sci. 2011;375:334–44.
49. Takegami S, Yamada H, Tsujii S. Pervaporation of ethanol/water mixtures 71. Xiangli F, Chen Y, Jin W, Xu N. Polydimethylsiloxane (PDMS)/ceramic
using novel hydrophobic membranes containing polydimethylsiloxane. composite membrane with high flux for pervaporation of ethanol−
J Membr Sci. 1992;75:93–105. water mixtures. Ind Eng Chem Res. 2007;46:2224–30.
50. Hongjoo L, Eunjin C, Younggyu K, Inseong C, Hyeunjong B. Per- 72. Peng P, Shi B, Lan Y. Preparation of PDMS—silica nanocomposite
vaporative separation of bioethanol using a polydimethylsiloxane/ membranes with silane coupling for recovering ethanol by pervapo-
polyetherimide composite hollow-fiber membrane. Bioresour Technol. ration. Sep Sci Technol. 2011;46:420–7.
2012;109:110–5. 73. Liu X, Hu D, Li M, Zhang J, Zhu Z, Zeng G, Zhang Y, Sun Y. Preparation
51. Thongsukmak A, Sirkar KK. Extractive pervaporation to separate ethanol and characterization of Silicalite-1/PDMS surface sieving pervapo-
from its dilute aqueous solutions characteristic of ethanol-producing ration membrane for separation of ethanol/water mixture. J Appl
fermentation processes. J Membr Sci. 2009;329:119–29. Polym Sci. 2015;132:42460.
52. Xu C, Lu X, Wang Z. Effects of sodium ions on the separation 74. Dobrak A, Figoli A, Chovau S, Galiano F, Simone S, Vankelecom IFJ,
performance of pure-silica MFI zeolite membranes. J Membr Sci. Drioli E, Bruggen BVD. Performance of PDMS membranes in pervapo-
2017;524:124–31. ration: Effect of silicalite fillers and comparison with SBS membranes.
53. Lin X, Chen X, Kita H, Okamoto K. Synthesis of silicalite tubular mem- J Colloid Interface Sci. 2010;346:254–64.
branes by in situ crystallization. AlChE J. 2003;49:237–47. 75. Yadav A, Lind ML, Ma X, Lin YS. Nanocomposite silicalite-1/polydi-
54. Shu X, Wang X, Kong Q, Gu X, Xu N. High-flux MFI zeolite membrane methylsiloxane membranes for pervaporation of ethanol from dilute
supported on YSZ hollow fiber for separation of ethanol/water. Ind Eng aqueous solutions. Ind Eng Chem Res. 2013;52:5207–12.
Chem Res. 2012;51:12073–80. 76. Yi S, Su Y, Wan Y. Preparation and characterization of vinyltriethoxysi-
55. Claes S, Vandezande P, Mullens S, De Sitter K, Peeters R, Van Bael MK. lane (VTES) modified silicalite-1/PDMS hybrid pervaporation mem-
Preparation and benchmarking of thin film supported PTMSP-silica brane and its application in ethanol separation from dilute aqueous
pervaporation membranes. J Membr Sci. 2012;389:265–71. solution. J Membr Sci. 2010;360:341–51.
56. Rozicka A, Niemistö J, Keiski RL, Kujawski W. Apparent and intrinsic 77. Zhou H, Zhang J, Wan Y, Jin W. Fabrication of high silicalite-1 content
properties of commercial PDMS based membranes in pervaporative filled PDMS thin composite pervaporation membrane for the separa-
removal of acetone, butanol and ethanol from binary aqueous mix- tion of ethanol from aqueous solutions. J Membr Sci. 2017;524:1–11.
tures. J Membr Sci. 2014;453:108–18. 78. Wang N, Liu J, Li J, Gao J, Ji S, Li J-R. Tuning properties of silicalite-1
57. Baker RW, Athayde AL, Daniels R, Le M, Pinnau I, Ly J, Wijmans JG, for enhanced ethanol/water pervaporation separation in its PDMS
Kaschemekat JH, Helm V: Development of pervaporation to recover hybrid membrane. Microporous Mesoporous Mater. 2015;201:35–42.
and reuse volatile organic compounds from industrial waste Streams. 79. Ji L, Shi B, Wang L. Pervaporation separation of ethanol/water
DOE/AL/98769–1 edn: Report to US Department of Energy; 1997. mixture using modified zeolite filled PDMS membranes. J Appl Polym
58. Fadeev AG, Kelley SS, McMillan JD, Selinskaya YA, Khotimsky VS, Volkov Sci. 2015;132:41897.
VV. Effect of yeast fermentation by-products on poly[1-(trimethylsilyl)- 80. Zhuang X, Chen X, Su Y, Luo J, Feng S, Zhou H, Wan Y. Surface modi-
1-propyne] pervaporative performance. J Membr Sci. 2003;214:229–38. fication of silicalite-1 with alkoxysilanes to improve the performance
59. Ueno K, Horiguchi Y, Negishi H, Miyamoto M, Uemiya S, Takeno A, of PDMS/silicalite-1 pervaporation membranes: Preparation, charac-
Sawada Y, Oumi Y. Fabrication of high-performance silicalite-1 mem- terization and modeling. J Membr Sci. 2016;499:386–95.
brane by a novel seeding method using zeolite-dispersed polymer film. 81. Zhan X, Li J, Fan C, Han X. Pervaporation separation of ethanol/water
Microporous Mesoporous Mater. 2018;261:58–62. mixtures with chlorosilane modified silicalite-1/PDMS hybrid mem-
60. Lin X, Kita H. Okamoto Ki: Silicalite membrane preparation, characteriza- branes. Chin J Polym Sci. 2010;28:625–35.
tion, and separation performance. Ind Eng Chem Res. 2001;40:4069–78. 82. Offeman R, Ludvik C. Poisoning of mixed matrix membranes by
61. Zhang XL, Zhu MH, Zhou RF, Chen XS, Kita H. Synthesis of a silicalite fermentation components in pervaporation of ethanol. J Membr Sci.
zeolite membrane in ultradilute solution and its highly selective separa- 2011;367:288–95.
tion of organic/water mixtures. Ind Eng Chem Res. 2012;51:11499–508. 83. Xia Z, Lu J, Tan T, Li J. Mixed matrix membranes with HF acid etched
62. Bello RH, Linzmeyer P, Franco CM, Souza O, Sellin N, Medeiros SH, ZSM-5 for ethanol/water separation: Preparation and pervaporation
Marangoni C. Pervaporation of ethanol produced from banana waste. performance. Appl Surf Sci. 2012;259:547–56.
Waste Manage (Oxford). 2014;34:1501–9. 84. Naik PV, Kerkhofs S, Martens JA, Vankelecom IFJ. PDMS mixed matrix
63. Guo J, Zhang G, Wu W, Ji S, Qin Z, Liu Z. Dynamically formed inner skin membranes containing hollow silicalite sphere for ethanol / water
hollow fiber polydimethylsiloxane/polysulfone composite membrane separation by pervaporation. J Membr Sci. 2016;502:48–56.
for alcohol permselective pervaporation. Chem Eng J. 2010;158:558–65. 85. Mori Y, Inaba T. Ethanol production from starch in a pervaporation
64. Knozowska K, Kujawska A, Kujawa J, Kujawski W, Bryjak M, Chrzanowska membrane bioreactor using Clostridium thermohydrosulfuricum.
E, Kujawski JK. Performance of commercial composite hydrophobic Biotechnol Bioeng. 1990;36:849–53.
membranes applied for pervaporative reclamation of acetone, butanol, 86. Vankelecom IFJ, Depre D, Beukelaer SD, Uytterhoeven JB. Influence
and ethanol from aqueous solutions: binary mixtures. Sep Purif Technol. of zeolites in PDMS membranes: pervaporation of water/alcohol
2017;188:512–22. mixtures. J Phys Chem. 1995;99:13193–7.
65. Gaykawad SS, Zha Y, Punt PJ, van Groenestijn JW, La VDW, Straathof 87. Yin H, Lau CY, Rozowski M, Howard C, Xu Y, Lai T, Dose ME, Lively RP,
AJ. Pervaporation of ethanol from lignocellulosic fermentation broth. Lind ML. Free-standing ZIF-71/PDMS nanocomposite membranes for
Bioresour Technol. 2013;129:469–76. the recovery of ethanol and 1-butanol from water through pervapo-
66. Verhoef A, Figoli A, Leen B, Bettens B, Drioli E, Van der Bruggen B. ration. J Membr Sci. 2017;529:286–92.
Performance of a nanofiltration membrane for removal of ethanol from 88. Tang X, Wang R, Xiao Z, Shi E, Yang J. Preparation and pervaporation
aqueous solutions by pervaporation. Sep Purif Technol. 2008;60:54–63. performances of fumed-silica-filled polydimethylsiloxane-polyamide
(PA) composite membranes. J Appl Polym Sci. 2007;105:3132–7.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 29 of 33
89. Takamizawa S, Kachi-Terajima C, Kohbara MA, Akatsuka T, Jin T. 110. Li S, Qin F, Qin P, Karim MN, Tan T. Preparation of PDMS membrane
Alcohol-vapor inclusion in single-crystal adsorbents [MII2(bza)4(pyz)]n using water as solvent for pervaporation separation of butanol–water
(M=Rh, Cu): structural study and application to separation mem- mixture. Green Chem. 2013;15:2180–90.
branes. Chemistry An Asian Journal. 2007;2:837–48. 111. O’Brien DJ, Roth LH, McAloon AJ. Ethanol production by continuous
90. Ulutan S, Nakagawa T. Separability of ethanol and water mixtures fermentation–pervaporation: a preliminary economic analysis. J Membr
through PTMSP–silica membranes in pervaporation. J Membr Sci. Sci. 2000;166:105–11.
1998;143:275–84. 112. Beaumelle D, Marin M, Gibert H. Pervaporation with organophilic mem-
91. Sukitpaneenit P, Chung T-S. PVDF/nanosilica dual-layer hollow fibers branes: state of the art. Food Bioprod Process. 1993;71:77–89.
with enhanced selectivity and flux as novel membranes for ethanol 113. Shi E, Huang W, Xiao Z, Li D, Tang M. Influence of binding interface
recovery. Ind Eng Chem Res. 2012;51:978–93. between active and support layers in composite PDMS membranes on
92. Adnadjević B, Jovanović J, Gajinov S. Effect of different physicochemical permeation performance. J Appl Polym Sci. 2010;104:2468–77.
properties of hydrophobic zeolites on the pervaporation properties of 114. Aoki T, Yamagiwa K, Yoshino E, Oikawa E. Temperature-sensitive ethanol
PDMS-membranes. J Membr Sci. 1997;136:173–9. permselectivity of poly(dimethylsiloxane) membrane by the modifica-
93. Yan H, Li J, Fan H, Ji S, Zhang G, Zhang Z. Sonication-enhanced in situ tion of its surface with copoly(N-isopropylacrylamide1H,1H,2H,2H-
assembly of organic/inorganic hybrid membranes: Evolution of perfluorododecyl acrylate). Polymer. 1993;34:1538–40.
nanoparticle distribution and pervaporation performance. J Membr Sci. 115. Chang CL, Chang PY. Performance enhancement of silicone/PVDF com-
2015;481:94–105. posite membranes for pervaporation by reducing cross-linking density
94. Liu Q, Li Y, Li Q, Liu G, Liu G, Jin W. Mixed-matrix hollow fiber composite of the active silicone layer. Desalination. 2006;192:241–5.
membranes comprising of PEBA and MOF for pervaporation separation 116. Liu W. A new PDMS-b-PPO block copolymer membrane with novel
of ethanol/water mixtures. Sep Purif Technol. 2019;214:2–10. non-perforated structure towards high flux for alcohol permselective
95. Mao H, Zhen H, Ahmad A, Li S, Liang Y, Ding J, Wu Y, Li L, Zhao Z. Highly pervaporation. Express Polym Lett. 2015;9:372–83.
selective and robust PDMS mixed matrix membranes by embedding 117. Guo Z, Hu C. Pervaporation of organic liquid/ water mixtures through a
two-dimensional ZIF-L for alcohol permselective pervaporation. J novel silicone copolymer membrane. Chin Sci Bull. 1998;43:487–90.
Membr Sci. 2019;582:307–21. 118. Greer DR, Ozcam AE, Balsara NP. Pervaporation of organic compounds
96. Naik PV, Wee LH, Meledina M, Turner S, Li Y, Tendeloo GV, Martens JA, from aqueous mixtures using polydimethylsiloxane-containing block
Vankelecom IFJ. PDMS membranes containing ZIF-coated mesoporous copolymer membranes. AlChE J. 2015;61:2789–94.
silica spheres for efficient ethanol recovery via pervaporation. J Mater 119. Yun CM, Akiyama E, Yamanobe T, Uehara H, Nagase Y. Characterizations
Chem A. 2016;4:12790–8. of PDMS-graft copolyimide membrane and the permselectivity of
97. Liu X, Li Y, Zhu G, Ban Y, Xu L, Yang W. An organophilic pervapora- gases and aqueous organic mixtures. Polymer. 2016;103:214–23.
tion membrane derived from metal–organic framework nanopar- 120. Kashiwagi T, Okabe K, Okita K. Separation of ethanol from ethanol/
ticles for efficient recovery of bio-alcohols. Angew Chem Int Ed. water mixtures by plasma-polymerized membranes from silicone
2011;50:10636–9. compounds. J Membr Sci. 1988;36:353–62.
98. Xue C, Wang Z, Du G, Fan L, Mu Y, Ren J, Bai F. Integration of ethanol 121. Fu C, Cai D, Hu S, Miao Q, Wang Y, Qin P, Wang Z, Tan T. Ethanol fer-
removal using carbon nanotube (CNT)-mixed membrane and ethanol mentation integrated with PDMS composite membrane: An effective
fermentation by self-flocculating yeast for antifouling ethanol recovery. process. Bioresour Technol. 2016;200:648–57.
Process Biochem. 2016;51:1140–6. 122. Chang CL, Chang MS. Preparation of composite membranes of func-
99. Shi S, Du Z, Ye H, Zhang C, Li H. A novel carbon black/polydimethylsilox- tionalised silicone polymers and PVDF for pervaporation of ethanol-
ane composite membrane with high flux for the separation of ethanol water mixture. Desalination. 2002;148:39–42.
from water by pervaporation. Polym J. 2006;38:949–55. 123. Chang CL, Chang MS. Preparation of multi-layer silicone/PVDF com-
100. Vankelecom IFJ, Kinderen JD, Anda BMD, Uytterhoeven JB. Incorpora- posite membranes for pervaporation of ethanol aqueous solutions. J
tion of hydrophobic porous fillers in PDMS membranes for use in Membr Sci. 2004;238:117–22.
pervaporation. J Phys Chem B. 1997;101:5182–5. 124. Slater CS, Hickey PJ, Juricic FP. Pervaporation of aqueous ethanol
101. Liu G, Jin W, Xu N. Graphene-based membranes. Chem Soc Rev. mixtures through poly(dimethyl siloxane) membranes. Sep Sci Technol.
2015;44:5016–30. 1990;25:1063–77.
102. Dharupaneedi SP, Anjanapura RV, Han JM, Aminabhavi TM. Functional- 125. Moermans B, Beuckelaer WD, Vankelecom IFJ, Ravishankar R, Martens
ized graphene sheets embedded in chitosan nanocomposite mem- JA, Jacobs PA. Incorporation of nano-sized zeolites in membranes.
branes for ethanol and isopropanol dehydration via pervaporation. Ind Chem Commun. 2000;24:2467–8.
Eng Chem Res. 2014;53:14474–84. 126. Chen J, Zhang H, Wei P, Zhang L, Huang H. Pervaporation behavior
103. Le NL, Wang Y, Chung T-S. Pebax/POSS mixed matrix membranes for and integrated process for concentrating lignocellulosic ethanol
ethanol recovery from aqueous solutions via pervaporation. J Membr through polydimethylsiloxane (PDMS) membrane. Bioproc Biosyst Eng.
Sci. 2011;379:174–83. 2014;37:183–91.
104. Du CH, Xu YY, Zhu BK. Structure formation and characterization of PVDF 127. Huang Y, Fu J, Pan Y, Huang X, Tang X. Pervaporation of ethanol aque-
hollow fiber membranes by melt-spinning and stretching method. J ous solution by polyphosphazene membranes: Effect of pendant
Appl Polym Sci. 2007;106:1793–9. groups. Sep Purif Technol. 2009;66:504–9.
105. Jiang LY, Song ZW. Interfacial resistance of dual-layer asymmetric hol- 128. Rezakazemi M, Shahidi K, Mohammadi T. Synthetic PDMS composite
low fiber pervaporation membranes formed by co-extrusion. J Polym membranes for pervaporation dehydration of ethanol. Desalin Water
Res. 2011;18:2505–14. Treat. 2015;54:1542–9.
106. Kargupta K, Datta S, Sanyal SK. Quantitative approach for the prediction 129. Wang R, Shan L, Zhang G, Ji S. Multiple sprayed composite membranes
of preferential sorption in the case of pervaporation of a physico-chem- with high flux for alcohol permselective pervaporation. J Membr Sci.
ically similar binary mixture. J Membr Sci. 1997;124:253–62. 2013;432:33–41.
107. Nagy E. Coupled effect of the membrane properties and concentration 130. Okamoto K-I, Butsuen A, Tsuru S, Nishioka S, Tanaka K, Kita H, Asakawa
polarization in pervaporation: Unified mass transport model. Sep Purif S. Pervaporation of water-ethanol mixtures through poly-dimethylsilox-
Technol. 2010;73:194–201. ane block-copolymer membranes. Polym J. 1987;19:747–56.
108. Heintz A, Stephan W. A generalized solution–diffusion model of the 131. Nagase Y, Mori S, Matsui K. Chemical modification of poly(substituted-
pervaporation process through composite membranes Part I Prediction acetylene). IV. Pervaporation of organic liquid/water mixture through
of mixture solubilities in the dense active layer using the UNIQUAC poly(1-phenyl-1-propyne)/polydimethylsiloxane graft copolymer
model. J Membr Sci. 1994;89:143–51. membrane. J Appl Polym Sci. 1989;37:1259–67.
109. Gref R, Nguyen QT, Néel J. Influence of membrane properties on system 132. Zhu Z, Hu D, Liu Y, Xu Y, Zeng G, Wang W, Zhang Y, Cui F. Three-
performances in pervaporation under concentration polarization component mixed matrix organic/inorganic hybrid membranes for
regime. Sep Sci Technol. 1992;27:467–91. pervaporation separation of ethanol-water mixture. J Appl Polym Sci.
2017;134:44753.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 30 of 33
133. Kurchan J. and out of equilibrium. Nature. 2005;433:222–5. 153. Cheng XQ, Konstas K, Doherty CM, Wood CD, Mulet X, Xie Z, Ng D, Hill
134. Yampol’Skii YP, Shishatskii SM, Shantorovich VP, Antipov EM, Kuzmin NN, MR, Lau CH, Shao L. Organic microporous nanofillers with unique alco-
Rykov SV, Khodjaeva VL, Platé NA. Transport characteristics and other hol affinity for superior ethanol recovery toward sustainable biofuels.
physicochemical properties of aged poly(1-(trimethylsilyl)-1-propyne). J Chemsuschem. 2017;10:1887–91.
Appl Polym Sci. 1993;48:1935–44. 154. Long YC, Chen X, Ping ZH, Fu SK, Sun YJ: MFI-type zeolite filled silicone
135. González-Velasco JR, López-Dehesa C, González-Marcos JA. Pervapora- rubber membranes: preparation, composition, and performance. In
tion performance of PTMSP membranes at high temperatures. J Appl Stud Surf Sci Catal. Volume 84. Edited by Weitkamp J, Karge HG, Pfeifer H,
Polym Sci. 2003;90:2255–9. Hölderich W: Elsevier; 1994: 1083–1090
136. Masuda T, Takatsuka M, Tang BZ, Higashimura T. Pervaporation of 155. Chen X, Ping Z, Long Y. Separation properties of alcohol-water mixture
organic liquid-water mixtures through substituted polyacetylene mem- through silicalite-1-filled silicone rubber membranes by pervaporation.
branes. J Membr Sci. 1990;49:69–83. J Appl Polym Sci. 1998;67:629–36.
137. López-Dehesa C, González-Marcos JA, González-Velasco JR. Pervapo- 156. Krea M, Roizard D, Moulai-Mostefa N, Sacco D. New copolyimide mem-
ration of 50 wt % ethanol–water mixtures with poly(1-trimethylsilyl- branes with high siloxane content designed to remove polar organics
1-propyne) membranes at high temperatures. J Appl Polym Sci. from water by pervaporation. J Membr Sci. 2004;241:55–64.
2007;103:2843–8. 157. García M, Sanz MT, Beltrán S. Separation by pervaporation of ethanol
138. Yu N, Ishihara K, Matsui K. Chemical modification of poly(substituted- from aqueous solutions and effect of other components present in
acetylene): II. Pervaporation of ethanol/water mixture through poly(1- fermentation broths. J Chem Technol Biotechnol. 2009;84:1873–82.
trimethylsilyl-1-propyne)/poly(dimethylsiloxane) graft copolymer 158. Lazarova M, Bösch P, Friedl A. POMS membrane for selective separation
membrane. J Polym Sci Pol Phys. 1990;28:377–86. of ethanol from dilute alcohol-aqueous solutions by pervaporation. Sep
139. Yu N, Takamura Y, Matsui K. Chemical modification of poly(substituted- Sci Technol. 2012;47:1709–14.
acetylene). V. Alkylsilylation of poly(1-trimethylsilyl-1-propyne) and 159. Gaykawad SS, Wielen LA, Straathof AJJ. Effects of yeast-originating
improved liquid separating property at pervaporation. J Appl Polym Sci. polymeric compounds on ethanol pervaporation. Bioresour Technol.
1991;42:185–90. 2012;116:9–14.
140. Nagase Y, Sugimoto K, Takamura Y, Matsui K: Chemical modification of 160. Zhang W, Xia C, Li L, Ren Z, Liu J, Yang X. Preparation and application
poly(substituted-acetylene). VI. Introduction of fluoroalkyl group into of a novel ethanol permselective poly(vinyltriethoxysilane) membrane.
poly(1-trimethylsilyl-1-propyne) and the improved ethanol permselec- RSC Adv. 2014;4:14592–6.
tivity at pervaporation. J Appl Polym Sci. 1991;43:1227–1232. 161. Jia W, Sun W, Xia C, Yang X, Cao Z, Zhang W. Copolymerization modifi-
141. Uragami T, Doi T, Miyata T: Control of permselectivity with surface cation of poly(vinyltriethoxysilane) membranes for ethanol recovery by
modifications of poly[1-(trimethylsilyl)-1-propyne] membranes. Int J pervaporation. RSC Adv. 2017;7:54441–52.
Adhes Adhes. 1999;19:405–409. 162. Shi S, Du Z, Zhang C, Li H. Pervaporation of ethanol/water mixtures
142. Kang YS, Shin EM, Jung B, Kim JJ. Composite membranes of poly(1- using novel hydrophobic membranes based on concentrated emulsion
trimethylsilyl-1-propyne) and poly(dimethyl siloxane) and their polymerization. Polym J. 2005;37:925.
pervaporation properties for ethanol-water mixture. J Appl Polym Sci. 163. Ishihara K, Matsui K. Pervaporation of ethanol-water mixture through
1994;53:317–23. composite membranes composed of styrene-fluoroalkyl acrylate graft
143. Masuda T, Isobe E, Higashimura T, Takada K. Poly[1-(trimethylsilyl)-1-pro- copolymers and cross-linked polydimethylsiloxane membrane. J Appl
pyne]: a new high polymer synthesized with transition-metal catalysts Polym Sci. 1987;34:437–40.
and characterized by extremely high gas permeability. J Am Chem Soc. 164. Matsui K, Ishihara K. Fluorine-containing graft copolymer and adhesive
1983;105:7473–4. and composite membrane made thereof. United States: Sagami Chemi-
144. Masuda T, Niki A, Isobe E, Higashimura T. Effect of organometallic cal Research Center (Tokyo, JP); 1987.
cocatalysts on the polymerization of 1-phenyl-1-propyne by tantalum 165. Masuok T, Iwatsubo T, Mizoguchi K. Pervaporation membranes for
pentachloride ( TaCl5) and niobium pentachloride ( NbCl5). Macromol- ethanol-water mixture prepared by plasma polymerization of fluorocar-
ecules. 1985;18:2109–13. bons. II Perfluoropropane membranes J Membr Sci. 1992;69:109–20.
145. Masuda T, Higashimura T. Polyacetylenes with substituents: Their syn- 166. Liu F, Liu L, Feng X. Separation of acetone–butanol–ethanol (ABE)
thesis and properties. Adv Polym Sci. 1986;1:121–65. from dilute aqueous solutions by pervaporation. Sep Purif Technol.
146. Volkov VV, Fadeev AG, Khotimsky VS, Litvinova EG, Selinskaya YA, Mcmil- 2005;42:273–82.
lan JD, Kelley SS. Effects of synthesis conditions on the pervaporation 167. Abu-Saied MA, Taha TH, Elnaggar EM, Amer RA, Mansy AE, Elkady GM.
properties of poly[1-(trimethylsilyl)-1-propyne] useful for membrane Green production of bio-ethanol from cellulosic fiber waste and its
bioreactors. J Appl Polym Sci. 2004;91:2271–7. separation using polyacrylonitrile-co-poly methyl acrylate membrane.
147. González-Marcos JA, López-Dehesa C, González-Velasco JR. Effect of Cellulose. 2018;25:6621–44.
operation conditions in the pervaporation of ethanol–water mixtures 168. Adymkanov SV. Yampol’skii YP, Polyakov AM, Budd PM, Reynolds KJ,
with poly(1-trimethylsilyl-1-propyne) membranes. J Appl Polym Sci. McKeown NB, Msayib KJ: Pervaporation of alcohols through highly
2004;94:1395–403. permeable PIM-1 polymer films. Polym Sci Ser A+. 2008;50:444–50.
148. Dibrov GA, Volkov V, Vasilevsky VP, Shutova AA, Bazhenov S, Khotimsky 169. Fritsch D, Merten P, Heinrich K, Lazar M, Priske M. High performance
V, van de Runstraat A, Goetheer E, Volkov A. Robust high-permeance organic solvent nanofiltration membranes: Development and thorough
PTMSP composite membranes for C O2 membrane gas desorption at testing of thin film composite membranes made of polymers of intrin-
elevated temperatures and pressures. J Membr Sci. 2014;470:439–50. sic microporosity (PIMs). J Membr Sci. 2012;401–402:222–31.
149. Borisov IL, Malakhov AO, Khotimsky VS, Litvinova EG, Finkelshtein ES, 170. Aroujalian A, Raisi A. Pervaporation as a means of recovering ethanol
Ushakov NV, Volkov VV. Novel PTMSP-based membranes containing from lignocellulosic bioconversions. Desalination. 2009;247:509–17.
elastomeric fillers: Enhanced 1-butanol/water pervaporation selectivity 171. Sukitpaneenit P, Chung TS, Jiang LY. Modified pore-flow model for
and permeability. J Membr Sci. 2014;466:322–30. pervaporation mass transport in PVDF hollow fiber membranes for
150. Schmidt SL, Myers MD, Kelley SS, McMillan JD, Padukone N. Evaluation ethanol–water separation. J Membr Sci. 2010;362:393–406.
of PTMSP membranes in achieving enhanced ethanol removal from fer- 172. Sukitpaneenit P, Chung TS. Molecular design of the morphology
mentations by pervaporation. Appl Biochem Biotechnol. 1997;63:469. and pore size of PVDF hollow fiber membranes for ethanol–water
151. Claes S, Vandezande P, Mullens S, Leysen R, Sitter KD, Andersson A, separation employing the modified pore-flow concept. J Membr Sci.
Maurer FHJ, Rul HVD, Peeters R, Bael MKV. High flux composite PTMSP- 2011;374:67–82.
silica nanohybrid membranes for the pervaporation of ethanol/water 173. Thongsukmak A, Sirkar KK. Pervaporation membranes highly
mixtures. J Membr Sci. 2010;351:160–7. selective for solvents present in fermentation broths. J Membr Sci.
152. Hickey PJ, Juricic FP, Slater CS. The effect of process parameters on the 2007;302:45–58.
pervaporation of alcohols through organophilic membranes. Sep Sci 174. Qin Y, Sheth JP. Pervaporation membranes that are highly selective for
Technol. 1992;27:843–61. acetic acid over water. Ind Eng Chem Res. 2003;42:582–95.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 31 of 33
175. Marjani A, Mohammadi M, Pelalak R, Moradi S. Ethanol purification 197. Xia S, Peng Y, Lu H, Wang Z. The influence of nanoseeds on the per-
using polyamide-carbon nanotube composite membranes. Polym Eng vaporation performance of MFI-type zeolite membranes on hollow
Sci. 2014;54:961–8. fibers. Microporous Mesoporous Mater. 2016;222:128–37.
176. Kujawski W, Krajewska S, Kujawski M, Gazagnes L, Larbot A, Persin M. 198. Lin X, Kita H, Okamoto KI. A novel method for the synthesis of high
Pervaporation properties of fluoroalkylsilane (FAS) grafted ceramic performance silicalite membranes. Chem Commun. 2000;19:1889–90.
membranes. Desalination. 2007;205:75–86. 199. Peng Y, Zhan Z, Shan L, Li X, Wang Z, Yan Y. Preparation of zeolite MFI
177. Subaer A, Haris A, Irhamsyah A, Permatasari AD, Desa SS, Irfanita R, Wah- membranes on defective macroporous alumina supports by a novel
yuni S. Pervaporation membrane based on laterite zeolite-geopolymer wetting–rubbing seeding method: Role of wetting agent. J Membr Sci.
for ethanol-water separation. J Clean Prod. 2020;249:119413. 2013;444:60–9.
178. Bowen TC, Noble RD, Falconer JL. Fundamentals and applica- 200. Wang Z, Ge Q, Gao J, Shao J, Liu C, Yan Y. High-performance zeolite
tions of pervaporation through zeolite membranes. J Membr Sci. membranes on inexpensive large-pore supports: highly reproducible
2004;245:1–33. synthesis using a seed paste. Chemsuschem. 2011;4:1570–3.
179. Chai L, Li H, Zheng X, Wang J, Yang J, Lu J, Yin D, Zhang Y. Pervapora- 201. Giaya A, Thompson RW, Denkewicz R Jr. Liquid and vapor phase
tion separation of ethanol–water mixtures through B-ZSM-11 zeolite adsorption of chlorinated volatile organic compounds on hydrophobic
membranes on macroporous supports. J Membr Sci. 2015;491:168–75. molecular sieves. Microporous Mesoporous Mater. 2000;40:205–18.
180. Chen C, Cheng Y, Peng L, Zhang C, Wu Z, Gu X, Wang X, Murad S. 202. Chen NY. Hydrophobic properties of zeolites. J Phys Chem.
Fabrication and stability exploration of hollow fiber mordenite zeolite 1976;80:60–4.
membranes for isopropanol/water mixture separation. Microporous 203. Machado MS, Pérez-Pariente J, Sastre E, Cardoso D, de Guereñu AM.
Mesoporous Mater. 2019;274:347–55. Selective synthesis of glycerol monolaurate with zeolitic molecular
181. Navajas A, Mallada R, Téllez C, Coronas J, Menéndez M, Santamaría J. sieves. Appl Catal A-Gen. 2000;203:321–8.
Preparation of mordenite membranes for pervaporation of water- 204. Kovo AS: Development of zeolites and zeolite membranes from Ahoko
ethanol mixtures. Desalination. 2002;148:25–9. Nigerian Kaolin. The University of Manchester, 2011.
182. RoyChoudhury P, Majumdar S, Sarkar S, Kundu B, Sahoo GC. Perfor- 205. Gabla JJ, Mistry SR, Maheria KC. An efficient green protocol for the syn-
mance investigation of Pb(II) removal by synthesized hydroxyapatite thesis of tetra-substituted imidazoles catalyzed by zeolite BEA: effect of
based ceramic ultrafiltration membrane: Bench scale study. Chem Eng surface acidity and polarity of zeolite. Catal Sci Technol. 2017;7:5154–67.
J. 2019;355:510–9. 206. Sano T, Hasegawa M, Kawakami Y, Kiyozumi Y, Yanagishita H, Kitamoto
183. Kalipcilar H, Bowen TC, Noble RD, Falconer JL. Synthesis and separation D, Mizukami F: Potentials of silicalite membranes for the separation
performance of SSZ-13 zeolite membranes on tubular supports. Chem of alcohol/water mixtures. In Stud Surf Sci Catal. Volume 84. Edited by
Mater. 2002;14:3458–64. Weitkamp J, Karge HG, Pfeifer H, Hölderich W: Elsevier; 1994: 1175–1182
184. Takaba H, Koyama A, Nakao S-I. Dual ensemble Monte Carlo simula- 207. Soydaş B, Dede Ö, Çulfaz A, Kalıpçılar H. Separation of gas and organic/
tion of pervaporation of an ethanol/water binary mixture in silicalite water mixtures by MFI type zeolite membranes synthesized in a flow
membrane based on a Lennard-Jones interaction model. J Phys Chem system. Microporous Mesoporous Mater. 2010;127:96–103.
B. 2000;104:6353–9. 208. Pan F, Wang M, He D, Song Y, Li W, Hong W, Jiang Z, Wang B, Cao X.
185. Oumi Y, Miyajima A, Miyamoto J, Sano T. Binary mixture adsorption of Embedding Ag+ @COFs within Pebax membrane to confer mass trans-
water and ethanol on silicalite. Stud Surf Sci Catal. 2002;142:1595–602. port channels and facilitated transport sites for elevated desulfurization
186. Chen H, Li Y, Liu J, Yang W. Preparation and pervaporation performance performance. J Membr Sci. 2018;552:1–12.
of high-quality silicalite-1 membranes. Sci China Ser B. 2007;50:70–4. 209. Ma X, Wu X, Caro J, Huang A. Seeding-free synthesis of high-per-
187. Chen H, Li Y, Yang W. Preparation of silicalite-1 membrane by solution- formance MFI zeolite membranes on superhydrophobic supports
filling method and its alcohol extraction properties. J Membr Sci. inspired by “like grows like” principle. Microporous Mesoporous Mater.
2007;296:122–30. 2019;288:109589.
188. Chen H, Song C, Yang W. Effects of aging on the synthesis and 210. Xia S, Yong P, Wang Z. Microstructure manipulation of MFI-type zeolite
performance of silicalite membranes on silica tubes without seeding. membranes on hollow fibers for ethanol–water separation. J Membr
Microporous Mesoporous Mater. 2007;102:249–57. Sci. 2016;498:324–35.
189. Matsuda H, Yanagishita H, Kitamoto D, Haraya K, Nakane T, Takada T, 211. Stoeger JA, Choi J, Tsapatsis M. Rapid thermal processing and separa-
Idemoto Y, Koura N, Sano T. Preparation of silicalite pervaporation mem- tion performance of columnar MFI membranes on porous stainless
brane with ethanol permselectivity by a 2-step hydrothermal synthesis. steel tubes. Energ Environ Sci. 2011;4:3479–86.
Sep Sci Technol. 2001;36:3305–10. 212. Tuan VA, Li S, Falconer JL, Noble RD. Separating organics from water by
190. Wu Z, Zhang C, Peng L, Wang X, Kong Q, Gu X. Enhanced stability of pervaporation with isomorphously-substituted MFI zeolite membranes.
MFI zeolite membranes for separation of ethanol/water by eliminating J Membr Sci. 2002;196:111–23.
surface Si-OH groups. ACS Appl Mater Inter. 2018;10:3175–80. 213. Li S, Tuan VA, Falconer JL, Noble RD. Properties and separation perfor-
191. Elyassi B, Jeon MY, Tsapatsis M, Narasimharao K, Basahel SN, Al-Thabaiti mance of Ge-ZSM-5 membranes. Microporous Mesoporous Mater.
S. Ethanol/water mixture pervaporation performance of b-oriented 2003;58:137–54.
silicalite-1 membranes made by gel-free secondary growth. AlChE J. 214. Bowen TC, Li S, Tuan VA, Falconer JL, Noble RD. Pervaporation of aque-
2016;62:556–63. ous organic mixtures through Ge-ZSM-5 zeolite membranes. Desalina-
192. Sano T, Yanagishita H, Kiyozumi Y, Mizukami F, Haraya K. Separation tion. 2002;147:327–9.
of ethanol/water mixture by silicalite membrane on pervaporation. J 215. Kosslick H, Tuan VA, Fricke R, Peuker C, Pilz W, Storek W. Synthesis and
Membr Sci. 1994;95:221–8. characterization of Ge-ZSM-5 zeolites. J Phys Chem. 1993;97:5678–84.
193. Shen D, Xiao W, Yang J, Chu N, Lu J, Yin D, Wang J. Synthesis of 216. Bowen TC, Kalipcilar H, Falconer JL, Noble RD. Pervaporation of organic/
silicalite-1 membrane with two silicon source by secondary growth water mixtures through B-ZSM-5 zeolite membranes on monolith sup-
method and its pervaporation performance. Sep Purif Technol. ports. J Membr Sci. 2003;215:235–47.
2011;76:308–15. 217. Chai L, Yang J, Lu J, Yin D, Zhang Y, Wang J. Ethanol perm-selective
194. Sebastian V, Mallada R, Coronas J, Julbe A, Terpstra RA, Dirrix RWJ. B-ZSM-5 zeolite membranes from dilute solutions. AlChE J.
Microwave-assisted hydrothermal rapid synthesis of capillary MFI-type 2016;62:2447–58.
zeolite–ceramic membranes for pervaporation application. J Membr 218. Chen X, Chen P, Kita H. Pervaporation through TS-1 membrane.
Sci. 2010;355:28–35. Microporous Mesoporous Mater. 2008;115:164–9.
195. Shan L, Shao J, Wang Z, Yan Y. Preparation of zeolite MFI membranes 219. Chen P, Chen X, Chen X, An Z, Kita H. Pervaporation separation and
on alumina hollow fibers with high flux for pervaporation. J Membr Sci. catalysis activity of novel zirconium silicalite-1 zeolite membrane. Chin J
2011;378:319–29. Chem. 2009;27:1692–6.
196. Zhang X, Zhu M, Zhou R, Chen X, Kita H. Synthesis of silicalite-1 mem- 220. Sano T, Hasegawa M, Ejiri S, Kawakami Y, Yanagishita H. Improvement of
branes with high ethanol permeation in ultradilute solution containing the pervaporation performance of silicalite membranes by modifica-
fluoride. Sep Purif Technol. 2011;81:480–4. tion with a silane coupling reagent. Microporous Mater. 1995;5:179–84.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 32 of 33
221. Sano T, Yamada K, Ejiri S, Hasegawa M, Kawakami Y, Yanagishita H. Silyla- 244. Vane LM, Namboodiri VV, Bowen TC. Hydrophobic zeolite–silicone rub-
tion of silicalite membrane and its pervaporation performance. Stud ber mixed matrix membranes for ethanol–water separation: Effect of
Surf Sci Catal. 1997;105:2179–86. zeolite and silicone component selection on pervaporation perfor-
222. Li S, Tuan VA, Noble RD, Falconer JL. ZSM-11 membranes: Characteriza- mance. J Membr Sci. 2008;308:230–41.
tion and pervaporation performance. AICHE J. 2002;48:269–78. 245. te Hennepe HJC, Bargeman D, Mulder MHV, Smolders CA: Zeolite-
223. Kim H-J, Jang K-S, Galebach P, Gilbert C, Tompsett G, Conner WC, Jones filled silicone rubber membranes: Part 1. Membrane preparation and
CW, Nair S. Seeded growth, silylation, and organic/water separation pervaporation results. J Membr Sci. 1987;35:39–55.
properties of MCM-48 membranes. J Membr Sci. 2013;427:293–302. 246. Zhan X, Lu J, Xu H, Liu J, Liu X, Cao X, Li J. Enhanced pervapora-
224. Song KH, Kim HY, Kim BM, Lee KR. Boehmite sol-coated membrane tion performance of PDMS membranes based on nano-sized
preparation for the separation of aqueous alcohol. Desalination. Octa[(trimethoxysilyl)ethyl]-POSS as macro-crosslinker. Appl Surf Sci.
2008;234:244–51. 2019;473:785–98.
225. Dong X, Lin YS. Synthesis of an organophilic ZIF-71 membrane for 247. Gu J, Shi X, Bai Y, Zhang H, Zhang L, Huang H. Silicalite-filled PEBA mem-
pervaporation solvent separation. Chem Commun. 2013;49:1196–8. branes for recovering ethanol from aqueous solution by pervaporation.
226. Huang K, Li Q, Liu G, Shen J, Guan K, Jin W. A ZIF-71 hollow fiber Chem Eng Technol. 2009;32:155–60.
membrane fabricated by contra-diffusion. ACS Appl Mater Inter. 248. Offeman RD, Ludvik CN. A novel method to fabricate high perme-
2015;7:16157–60. ance, high selectivity thin-film composite membranes. J Membr Sci.
227. Miyamoto M, Hori K, Goshima T, Takaya N, Oumi Y, Uemiya S. An 2011;380:163–70.
organoselective zirconium-based metal–organic-framework UiO-66 249. Vane LM, Namboodiri VV, Meier RG. Factors affecting alcohol–water
membrane for pervaporation. Eur J Inorg Chem. 2017;2017:2094–9. pervaporation performance of hydrophobic zeolite–silicone rubber
228. Wee SL, Tye CT, Bhatia S. Membrane separation process—Pervapora- mixed matrix membranes. J Membr Sci. 2010;364:102–10.
tion through zeolite membrane. Sep Purif Technol. 2008;63:500–16. 250. Li Q, Cheng L, Shen J, Shi J, Chen G, Zhao J, Duan J, Liu G, Jin W.
229. Yu L, Korelskiy D, Grahn M, Hedlund J. Very high flux MFI membranes for Improved ethanol recovery through mixed-matrix membrane with
alcohol recovery via pervaporation at high temperature and pressure. hydrophobic MAF-6 as filler. Sep Purif Technol. 2017;178:105–12.
Sep Purif Technol. 2015;153:138–45. 251. Khan A, Ali M, Ilyas A, Naik P, Vankelecom IFJ, Gilani MA, Bilad MR, Sajjad
230. Korelskiy D, Leppäjärvi T, Zhou H, Grahn M, Tanskanen J, Hed- Z, Khan AL. ZIF-67 filled PDMS mixed matrix membranes for recovery of
lund J. High flux MFI membranes for pervaporation. J Membr Sci. ethanol via pervaporation. Sep Purif Technol. 2018;206:50–8.
2013;427:381–9. 252. Xu S, Zhang H, Yu F, Zhao X, Wang Y. Enhanced ethanol recovery of
231. Sakaki K, Habe H, Negishi H, Ikegami T. Pervaporation of aqueous dilute PDMS mixed matrix membranes with hydrophobically modified ZIF-90.
1-butanol, 2-propanol, ethanol and acetone using a tubular silicalite Sep Purif Technol. 2018;206:80–9.
membrane. Desalin Water Treat. 2011;34:290–4. 253. Samanta HS, Ray SK, Das P, Singha NR. Separation of acid–water
232. Matsuda H, Yanagishita H, Negishi H, Kitamoto D, Ikegami T, Haraya K, mixtures by pervaporation using nanoparticle filled mixed matrix
Nakane T, Idemoto Y, Koura N, Sano T. Improvement of ethanol selectiv- copolymer membranes. J Chem Technol Biotechnol. 2012;87:608–22.
ity of silicalite membrane in pervaporation by silicone rubber coating. J 254. Jia Z, Hao S, Jiang M. PAF-11/poly (dimethylsiloxane) mixed matrix
Membr Sci. 2002;210:433–7. pervaporation membranes for dealcoholization of aqueous solutions. J
233. Ikegami T, Yanagishita H, Kitamoto D, Haraya K, Nakane T, Matsuda H, Chem Technol Biotechnol. 2018;93:3276–83.
Koura N, Sano T. Production of highly concentrated ethanol in a cou- 255. Wee LH, Li Y, Zhang K, Davit P, Bordiga S, Jiang J, Vankelecom IFJ, Mar-
pled fermentation/pervaporation process using silicalite membranes. tens JA. Submicrometer-sized ZIF-71 filled organophilic membranes for
Biotechnol Tech. 1997;11:921–4. improved bioethanol recovery: mechanistic insights by Monte Carlo
234. Nomura M, Bin T. Nakao Si: Selective ethanol extraction from simulation and FTIR spectroscopy. Adv Funct Mater. 2015;25:516–25.
fermentation broth using a silicalite membrane. Sep Purif Technol. 256. Yin H, Khosravi A, O’Connor L, Tagaban AQ, Wilson L, Houck B, Liu Q,
2002;27:59–66. Lind ML. Effect of ZIF-71 particle size on free-standing ZIF-71/PDMS
235. Kuhn J, Sutanto S, Gascon J, Gross J, Kapteijn F. Performance and composite membrane performances for ethanol and 1-butanol
stability of multi-channel MFI zeolite membranes detemplated by removal from water through pervaporation. Ind Eng Chem Res.
calcination and ozonication in ethanol/water pervaporation. J Membr 2017;56:9167–76.
Sci. 2009;339:261–74. 257. Han X, Zhang X, Ma X, Li J. Modified ZSM-5/polydimethylsiloxane mixed
236. Ikegami T, Yanagishita H, Kitamoto D, Negishi H, Haraya K, Sano T. matrix membranes for ethanol/water separation via pervaporation.
Concentration of fermented ethanol by pervaporation using silicalite Polym Compos. 2016;37:1282–91.
membranes coated with silicone rubber. Desalination. 2002;149:49–54. 258. Samanta HS, Ray SK. Separation of ethanol from water by pervapora-
237. Hasegawa Y, Kimura K, Nemoto Y, Nagase T, Kiyozumi Y, Nishide T, tion using mixed matrix copolymer membranes. Sep Purif Technol.
Mizukami F. Real-time monitoring of permeation properties through 2015;146:176–86.
polycrystalline MFI-type zeolite membranes during pervaporation 259. Li Y, Wee L, Martens J, Vankelecom I. ZIF-71 as a potential filler to pre-
using mass-spectrometry. Sep Purif Technol. 2008;58:386–92. pare pervaporation membranes for bio-alcohol recovery. J Mater Chem.
238. Zou X, Bazin P, Zhang F, Zhu G, Valtchev V, Mintova S. Ethanol recovery 2014;2:10034–40.
from water using silicalite-1 membrane: an operando infrared spectro- 260. Liu G, Xiangli F, Wei W, Liu S, Jin W. Improved performance of PDMS/
scopic study. ChemPlusChem. 2012;77:437–44. ceramic composite pervaporation membranes by ZSM-5 homogene-
239. Algieri C, Bernardo P, Golemme G, Barbieri G, Drioli E. Permea- ously dispersed in PDMS via a surface graft/coating approach. Chem
tion properties of a thin silicalite-1 (MFI) membrane. J Membr Sci. Eng J. 2011;174:495–503.
2003;222:181–90. 261. Lue SJ, Chien C-F, Mahesh KPO. Pervaporative concentration of etha-
240. Saboor FH, Ashrafizadeh SN, Kazemian H. Synthesis of BZSM-5 mem- nol–water mixtures using heterogeneous polydimethylsiloxane (PDMS)
branes using nano-zeolitic seeds: characterization and separation mixed matrix membranes. J Membr Sci. 2011;384:17–26.
performance. Chem Eng Technol. 2012;35:743–53. 262. Li X, Wang S. Some characteristics of pervaporation for dilute ethanol-
241. Sebastian V, Motuzas J, Dirrix RWJ, Terpstra RA, Mallada R, Julbe A. water mixtures by alcohol-permselective composite membrane. Sep
Synthesis of capillary titanosilicalite TS-1 ceramic membranes by MW- Sci Technol. 1996;31:2867–73.
assisted hydrothermal heating for pervaporation application. Sep Purif 263. Zhan X, Li J, Chen J, Huang J. Pervaporation of ethanol/water mixtures
Technol. 2010;75:249–56. with high flux by zeolite-filled PDMS/PVDF nanofiltration membranes.
242. S K, RW N, NN L: Separation of fluids by means of mixed matrix mem- Chin J Polym Sci. 2009;27:771–80.
branes. US; 1988. 264. Sun D, Li B, Xu Z. Pervaporation of ethanol/water mixture by organo-
243. Chung T-S, Jiang LY, Li Y, Kulprathipanja S. Mixed matrix membranes philic nano-silica filled PDMS composite membranes. Desalination.
(MMMs) comprising organic polymers with dispersed inorganic fillers 2013;322:159–66.
for gas separation. Prog Polym Sci. 2007;32:483–507.
Peng et al. Biotechnol Biofuels (2021) 14:10 Page 33 of 33
265. Huang Y, Zhang P, Fu J, Zhou Y, Huang X, Tang X. Pervaporation of 269. Zhan X, Wang M, Gao T, Lu J, He Y, Li J. A highly selective sorption pro-
ethanol aqueous solution by polydimethylsiloxane/polyphosphazene cess in POSS-g-PDMS mixed matrix membranes for ethanol recovery
nanotube nanocomposite membranes. J Membr Sci. 2009;339:85–92. via pervaporation. Sep Purif Technol. 2020;236:116238.
266. Pan Y, Yu X. Preparation of Zeolitic Imidazolate Framework-91 and its 270. Naik PV, Verlooy PLH, Smet S, Martens JA, Vankelecom IFJ. PDMS mixed
modeling for pervaporation separation of water/ethanol mixtures. Sep matrix membranes filled with novel PSS-2 nanoparticles for ethanol/
Purif Technol. 2020;237:116330. water separation by pervaporation. RSC Adv. 2016;6:78648–51.
267. Wang N, Shi G, Gao J, Li J, Wang L, Guo H, Zhang G, Ji S. MCM-41@ZIF-8/
PDMS hybrid membranes with micro- and nanoscaled hierarchical
structure for alcohol permselective pervaporation. Sep Purif Technol. Publisher’s Note
2015;153:146–55. Springer Nature remains neutral with regard to jurisdictional claims in pub-
268. Zhou H, Shi R, Jin W. Novel organic–inorganic pervaporation mem- lished maps and institutional affiliations.
brane with a superhydrophobic surface for the separation of ethanol
from an aqueous solution. Sep Purif Technol. 2014;127:61–9.