This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
5/14/2018 Copy of Per Vapor at Ion Study of Aqueous Ethanol Solution Through Zeolite-t...
Pervaporation study of aqueous ethanol solution through zeolite-incorporatedmultilayer poly(vinyl alcohol) membranes: Effect of zeolites
Zhen Huang a,∗, Huai-min Guan b, Wee lee Tan b, Xiang-Yi Qiao b, Santi Kulprathipanja c
a Department of Packaging Engineering, Tianjin University of Commerce, Tianjin 300134, PR Chinab Department of Chemical & Biomolecular Engineering, National University of Singapore, Singapore 119260, Singapore
c UOP LLC, 50 East Algonquin Road, Des Plaines, IL 60017-5016, USA
Received 1 July 2005; received in revised form 26 September 2005; accepted 29 September 2005
Available online 2 November 2005
Abstract
In this study, a series of three-layer zeolite-embedded poly(vinyl alcohol) (PVA) composite membranes have been successfully fabricated with
a casting machine. Zeolites, examined with a loading of 20 wt%, include 3A, 4A, 5A, NaX, NaY, silicalite and beta. These hydrophilic composite
membranes have been evaluated in the dehydration of ethanol aqueous solution by means of pervaporation. The unfilled PVA membrane is observed
to exhibit much higher separation factor than two commercial PERVAP 2210 and PERVAP 2510 membranes. After adding zeolites into the PVA
matrix, higher separation factor and higher fluxes or higher selectivity and higher penetrant permeances are both achieved by these resultant zeolite-
incorporated membranes, indicating that ethanol/water separation has been enhanced with the aid of incorporated zeolites. Through evaluating the
pervaporation performance in terms of water permeance, ethanol permeance and selectivity, we have revealed that the separation performances of
zeolite-filled membranes are strongly related to the zeolite pore dimension, its hydrophilic/hydrophobic nature as well as its crystal framework.
The temperature dependence of the pervaporation behaviors like the penetrant fluxes and permeances has been discussed in detail in terms of
Arrhenius activation energy. The evaluated results have revealed that the permeance and selectivity (i.e., the membrane intrinsic properties) are less
dependent on the operating temperature than the flux and separation factor. Zeolite addition has led to decreased activation energies for water and
ethanol, and more considerable drop of the water activation energy has subsequently resulted in the increased selectivity in ethanol dehydration.
264 Z. Huang et al. / Journal of Membrane Science 276 (2006) 260–271
Fig. 2. FESEM photographs of filled and unfilled multilayer PVA-based membranes: (a–c) unfilled; (d–f) filled with UOP 4A; (a and d) top view; (b and e)
cross-sectional view at low magnification; (c and f) cross-sectional view at high magnification.
(NaX and NaY) chosen may elucidate better the attribute of the
zeolite hydrophilicity for pervaporation.
Hydrophobic/hydrophilic nature of zeolites also appears
to depend on their framework structure [8]. Pure siliceous
zeolite beta has been reported to be much more hydropho-
bic than silicalite-1 and the other siliceous 12-numbered ring
zeolites even though they contain almost no aluminum [32].
In addition, silicalite-1 and beta both possess intricate three-
dimensional channel systems, and may discriminate competing
molecules on the basis of a difference in molecular shape.
In this work, we have selected high-aluminum beta and low-
aluminum silicalite-1 to investigate their effect on the membraneperformance.
3.2. Morphologies of fabricated membranes
Fig. 2 shows the FESEM images of the zeolite-incorporated
and unfilled membranes. From the cross-sectional view of both
membranes (Fig. 2b and e), the multi-layered structure of fabri-
cated membranes can be clearly observed, namely, a very dense
top selective layerof PVA or PVA-zeolite,a porous backing layer
of PAN and a support layer of non-woven fabric (PET RS21).
Fig. 2c and f presents FESEM images at high magnification for
two-layers cast. The top layer is seen to be very dense and thin
with a thickness of less than 10m. The backing layer possessesa cave-like structure and is much thicker (∼70m) than the top
selective layer. It is certain that this highly porous layer only
provides mechanic support to the selective layer and contributes
little to ethanol/water separation. The same can be applied to the
most porous and thickest (∼120m) non-woven fabric layer.
The thickness of the layers can be readily approximated from
the scale bar given at the bottom of the FESEM picture. The total
thickness of the multi-layered composite membrane is approxi-
mately 200m.
The non-porous selective layer is expected to be responsible
for ethanol/water separation. The cross-sectional and top views
(Fig. 2d and f) of the selective layer show that the zeolite parti-
cles are well distributed within the polymeric matrix and form a
good contact with polymer with no visible macroscopic voids.
This suggests that the selective layer is possibly defect-free, and
hence able to be effective in ethanol/water separation. In con-
trast, some literature reports have revealed that the addition of
zeolites caused microporous cave-like structures for cellulose
acetate–zeolite and layered PAN–zeolite composite membranes
which resulted in low separation factor and high permeant flux
[15,16]. This is, however, not the case for our multi-layered
PVA-zeolite composite membrane as confirmed by the FESEM
pictures. Hence, these zeolite-filled three-layer PVA membranes
are expected to achieve good performances in ethanol/water per-vaporation separation.
3.3. Pervaporation results
3.3.1. Comparison with commercial membranes
The pervaporation performance evaluation of the unfilled
multilayer membrane has started from a comparison made with
two commercial PVA/PAN membranes of PERVAP 2210 and
PERVAP 2510 (Sulzer Chemtech, Germany). The pervapora-
tion studies of PERVAP 2510 membrane have been previously
reported for water removal from high concentrated IPA and
butanol systems [27,29]. Their results show that this two-layer
composite membrane has achieved very high separation fac-tor for water. For example, separation factor obtained for water
over IPA has ranged from 300 to 1400 with feed water concen-
tration of 2–15 wt%, pervaporation temperature of 60–100 ◦C
and the downstream pressure of less than 100 Pa (1 mbar). The
separation factors are even much higher for butanol isomer sys-
tems. However, this membrane gives very poor performance for
ethanol system as reflected by an invariant separation factor of
15 throughout the test temperature range (see Table 3). The most
possible reason is that the linear ethanol molecule is (1) much
smaller than IPA and butanol isomers [8] and (2) able to form
stronger interaction with water [33] and thus lead to a consider-
able mutual-dragging effect.
5/14/2018 Copy of Per Vapor at Ion Study of Aqueous Ethanol Solution Through Zeolite-t...
266 Z. Huang et al. / Journal of Membrane Science 276 (2006) 260–271
It can be seen from Fig. 3 that all these zeolite-incorporated
multilayer PVA membranes, except silicallite-1, have consid-
erably higher total pervaporation fluxes as compared with the
unfilled PVA membrane. In terms of the pervaporation separa-
tion factor, all these the zeolite-incorporated membranes have
higher values than the unfilled one except the zeolite NaY-
filled membrane that exhibits much lower separation factor.
These results indicate that at least some penetrant molecules
transport across the membrane through the zeolite pores and
thus the interaction between the penetrant and zeolite pore sur-
face has an important role in affecting the membrane perfor-
mance. The different performances observed for these zeolites
are believably related to their characteristic features: pore size,
its composition and structure. These properties strongly affect
zeolitic molecular sieve effect and the hydrophilic/hydrophobic
nature.
After incorporating zeolitic molecular sieves into the PVA
polymer matrix, the intrinsic properties of the membrane mate-
rials have varied. As highlighted in recent works [27,29], the
membrane pervaporation flux and separation factor are heavilydependent on the operating conditions, which make a meaning-
ful comparison of data nearly impossible and obscure the effect
of the driving force in the pervaporation process. Therefore, we
have evaluated the pervaporation results of the ethanol–water
mixture in terms of permeance and selectivity for clearly under-
standing the effects of incorporated zeolites on membrane sep-
aration performance.
For the polymer-based pervaporation separation, the
solution-diffusion model may be applied. Thus, the permeation
flux ( J ) can be written as:
J water = Qwater(pfeedwater − p
permeatewater ) (3)
J ethanol = Qethanol(pfeedethanol − p
permeateethanol ) (4)
where p is the partial vapor pressure of each component and Q
is the membrane permeance. The partial vapor pressure of water
and ethanol on the membrane feed side can be calculated by
using the Wilson’s equation, as described in our preceding work
[30]. The membrane selectivity (β) is defined as the ratio of the
water permeance over the ethanol permeance.
β =Qwater
Qethanol(5)
Fig. 5a–c shows the pervaporation performances in terms of per-
meance and selectivity of the A-type zeolite-filled and unfilled
membranes for an ethanol–water mixture containing 20% of
water at various temperatures. Also included in Fig. 5a–c are
those pervaporation data obtained for the UOP 3A-filled mem-
branes [30]. It can be found that the unfilled three-layer mem-brane either manually cast [30] or machine-cast has similar
performance to each other. Similar to the pervaporation flux and
separation factor shown in Figs. 3 and 4, these membranes all
exhibit increased water and ethanol permeances but decreased
selectivities as the temperature arises. The temperature effect
can be explained by the variations of polymer free volume and
mobility of penetrants. At high temperature, the PVA free vol-
ume increases remarkably as a consequence of random thermal
motion of the polymer chains [33]. Furthermore, the theoretical
Fig. 5. Pervaporation performance in terms of (a) water permeance, (b) ethanol permeance and (c) selectivity of the A-type zeolite-filled three-layer PVA membranes
for dehydrating ethanol aqueous solution (20 wt% water). (*) From Ref. [30].
5/14/2018 Copy of Per Vapor at Ion Study of Aqueous Ethanol Solution Through Zeolite-t...
Z. Huang et al. / Journal of Membrane Science 276 (2006) 260–271 267
diffusivity of the permeating molecules can increase exponen-
tially as temperature increases [34]. Therefore, more penetrant
molecules can be transported through the membranes at high
temperatures, resulting in enhanced permeances for both water
and ethanol. The temperature effect is more pronounced on the
transport rate of ethanol than water as reflected by more signifi-
cantly increased ethanol permeance (Fig. 5a and b). As a result,
the dehydration selectivity reduces at higher operation temper-
ature (Fig. 5c).
Comparing with unfilled membranes, the A-type zeolite-
incorporated membranes all have higher ethanol dehydration
selectivities. However, Gao et al. and Okumus et al. have
reported that a decrease in separation factor were obtained after
the incorporation of zeolite A, which may be due to their mem-
brane casting techniques as revealed by the formation of porous
cave-like structures [14–16]. The multilayer A-type zeolite-
filled membranes also exhibit higher water permeances and
ethanol permeances except the permeances of UOP 3A-filled
membrane. The higher extent of separation can be explained by
the molecular sieving effect and zeolite hydrophilicity. Both fac-tors tend to increase the water selectivity. These A-type zeolites
all have pore sizes (see Table 2) larger than the kinetic diameter
of water molecule (0.264 nm) but smaller than or close to that of
ethanol molecule (0.430 nm) [8]. This distinction may possibly
induce the molecular sieving effect of the zeolite A-based mem-
branes. The Si/Al ratio of zeolite A is 1.0 which makes it one
of the most hydrophilic zeolites. The hydrophilicity, introduced
by zeolite A along with the polymer itself, can make the pre-
pared membrane able to form more specific interactions, such
as hydrogen bonding between the membrane functional groups
and water molecules, more preferably attract water molecules
and transport through the membrane. This effect has also been
reflected by higher water permeances for these zeolite-filled
membranes as seen from Fig. 5a. Furthermore, zeolite particles
are more resistant to swelling caused by water, and then reduce
possibility of loosening of polymeric chains within the zeolite-
based membrane, thus are able to achieve high selectivity.
Among the A-type zeolite-incorporated membranes, it can be
seen that the pervaporation permeance of either water or ethano
is the highest for COM 5A, then for UOP 4A, COM 3A and
lowest for UOP 3A. But the membrane selectivity follows the
opposite order;this is thesameorderthat their pore size decrease
As described earlier, these zeolites are all hydrophilic due to the
high aluminum content. Ion exchange changes the local polarity
in the pores and therefore the adsorption, but may not affect the
hydrophilic nature. Thus, the pore size variation arising from
the ion exchange treatment seems to be the main reason for the
ethanol dehydration results.To better examine the effect of zeolite hydrophobic
hydrophilic nature on the membrane separation performance
we have chosen two FAU zeolites, i.e., NaX and NaY, to study
their influences in separating water from ethanol–water solu-
tion. These two zeolites have the same zeolitic framework bu
with different aluminum contents. The experimental pervapo-
ration results for zeolite NaX and NaY-filled membranes are
shown in Fig. 6a–c. It can be observed that NaX-filled mem-
Fig. 6. Pervaporation performance in terms of (a) water permeance, (b) ethanol permeance and (c) selectivity of the NaX and NaY zeolite-filled three-layer PVA
membranes for dehydrating ethanol aqueous solution (20 wt% water).
5/14/2018 Copy of Per Vapor at Ion Study of Aqueous Ethanol Solution Through Zeolite-t...
268 Z. Huang et al. / Journal of Membrane Science 276 (2006) 260–271
brane shows much higher separation selectivity than the unfilled
membrane, while NaY-filled membrane produces significantly
lower degree of separation performance than the other two.
Zeolite NaX and NaY basically have the same pore size of
7.4 A but the former is more hydrophilic due to a lower Si/Al
ratio of 1.3 than NaY of 2.5. The large pore size and rela-
tively hydrophobic nature of NaY may explain the rather low
separation factor obtained by NaY-incorporated PVA mem-
brane for ethanol/water system. The aperture size of 0.74 nm
is significantly larger than kinetic diameters of both water and
ethanol molecules. Thus, NaY zeolite could not perform any
molecular sieving effect to water and ethanol or exhibit any
preferred attractive forces to water than ethanol, instead, may
provide the preferable passage for ethanol penetrant to trans-
port through the membrane. Subsequently, greater ethanol per-
meance has obtained for the NaY-incorporated membrane as
shown Fig. 6b. The less hydrophilic (or more hydrophobic)
nature of zeolite Y seems to agree with that previously reported,
where Y-type zeolite–PDMS membranes were employed for
removal ethanol from water via pervaporation, and resultedin an increase in both separation factor and flux [18,19].
NaX zeolite, being more hydrophilic, has more trivalent atoms
(e.g., Al) substituted for Si atoms and thus possesses more
charge-balancing cations which are occluded in the zeolitic
framework. The electrostatic forces formed by the negatively
charged framework and positively charged cations have ren-
dered it more hydrophilic and are able to selectively attract water
molecules than ethanol molecules, thereby producing higher
selectivity for water. Therefore, higher separation selectivity has
been obtained after adding NaX zeolite despite the large pore
dimension.
Shown in Fig. 7a–c is pervaporation performance results
for the zeolite silicalite-1 and beta-filled membranes. To our
knowledge, this is the first time to apply zeolite beta to fab-
ricate polymer-based composite membranes for pervapora-
tions. Very interestingly, the beta-incorporated membrane yields
much better separation performance than the unfilled membrane
whereas the silicalite-1-filled membrane produces compara-
ble water selectivities. It must be noted that the water selec-
tivity for zeolite materials can be affected not only by their
hydrophilic/hydrophobic property (due to the Al content) but
also by their surface properties and shape selectivities to water
and ethanol [31]. Since silicalite-1 is the most hydrophobic one
amongst the tested zeolites (which is qualitatively determined
from their Si/Al ratios), its incorporation into PVA membrane
likely results in the decease of the water permeance as compared
to the unfilled membrane. In addition, silicalite-1 has sinusoidal
channels that possibly drag the transport of the penetrants, lead-
ing to the deceased ethanol permeance. However, the presence
of silanol groups on the zeolite surfaces stemming from the
intracrystalline boundaries and defects or the aluminum increasethe local hydrophilicity to water. The unique asymmetrical aper-
ture (5.2× 5.7) and sinuous channels of silicalite-1 may very
likely produce additional shape selectivity to small molecular-
sized water. As a sequence, the membrane selectivity has no
much variation.
Similar to silicalite-1, beta zeolite also possesses sinuous
three-dimensional channel systems and a specified asymmet-
rical aperture (7.1× 7.3). Compared to silicalite-1, beta zeolite
generates better pervaporation results in terms of higher water
selectivity and higher permeances. The Si/Al ratio of zeolite
beta of 16.0 is very much lower than zeolite silicalite-1 of 196,
Fig. 7. Pervaporation performance in terms of (a) water permeance, (b) ethanol permeance and (c) selectivity of the silicalite-1 and beta zeolite-filled three-layer
PVA membranes for dehydrating ethanol aqueous solution (20 wt% water).
5/14/2018 Copy of Per Vapor at Ion Study of Aqueous Ethanol Solution Through Zeolite-t...