Thin film composite membranes on ceramic for pervaporation dehydration of isopropanol

Journal of Membrane Science 448 (2013) 34–43

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

Thin film composite membranes on ceramic for pervaporation dehydration of isopropanol Gui Min Shi, Tai-Shung Chung n Department of Chemical and Biomolecular Engineering, National University of Singapore, Singapore 117576, Singapore

art ic l e i nf o

a b s t r a c t

Article history: Received 2 April 2013 Received in revised form 21 July 2013 Accepted 22 July 2013 Available online 7 August 2013

We have developed novel thin film composite (TFC) membranes on ceramic substrates (referred as ceramic TFC membranes thereafter) for pervaporation dehydration of isopropanol (IPA) by using interfacial polymerization (IP) of m-phenylenediamine (MPD) and trymesoylchoride (TMC) to create ultra-thin polyamide films as selective layers. Prior to IP, a gutter layer of polydopamine (PDA) or polyethyleneimine (PEI) was deposited on α-alumina ceramic surfaces to improve the adhesion of polyamide thin films onto the ceramic. It was found that the PEI-coated ceramic substrate provided better adhesion with the polyamide thin film than the original and PDA-coated ones owing to its smoother surface. Polydimethylsiloxane (PDMS) was employed to seal defects on ceramic TFC membranes and significantly improved the performance of pervaporation dehydration of IPA with a flux of 6.05 kg/m2 h and a separation factor 1396 at 80 1C. Moreover, the developed ceramic TFC membrane shows high stability after 120 h operation. Positron annihilation spectroscopy (PAS) analyses revealed that the newly developed membrane has an ultrathin polyamide layer of 79 nm and a PDMS coating of
26 nm. This study demonstrates the promising application of ceramic TFC membranes for pervaporation dehydration of organic solutions. & 2013 Elsevier B.V. All rights reserved.

Keywords: Ceramic Thin film composite Interfacial polymerization Pervaporation PAS

1. Introduction Interfacial polymerization (IP), developed by Morgan and Kwolek [1] and Cadotte [2], is a powerful technique for the fabrication of thin film composite (TFC) membranes. TFC membranes have become the most dominant reverse osmosis (RO) membranes for desalination because of their superior permeability and selectivity [2,3]. IP has also been applied to fabricate membranes for other applications such as nanofiltration [4–6], forward osmosis (FO) [7,8], and gas separation [9,10]. Several studies have shown that TFC membranes are promising for pervaporation separation [11–14]. Owing to the ultrathin polyamide selective layer, a very high pervaporation flux can be achieved at relatively low operating temperatures [12,14]. TFC membranes generally consist of a polyamide layer and a support substrate. While it is important to have an ultra-thin functional selective layer for high pervaporation performance, the porous support also plays an important role for membrane applications under harsh feed conditions like aggressive solvents and high temperatures [15]. Typical polymer supports for commercially available TFC membranes are made of polymeric materials such as polysulfone (PS) [2] and polyethersulfone (PES) [16].

n

Corresponding author. Tel.: +65 6516 6645; fax: +65 6779 1936. E-mail address: [email protected] (T.-S. Chung).

0376-7388/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2013.07.049

Although these support materials are suitable for water purification applications, their chemical or thermal resistance may not be sufficient when being exposed to harsh chemical and high temperature environments in pervaporation applications. Swelling of the polymeric substrates causes damage to the ultra-thin skin layer and may deteriorate the pervaporation performance. Ceramic materials are promising because of their excellent properties such as strong mechanical strength, high chemical resistance and thermal stability which are essential for a substrate material to be used in pervaporation [17]. The rigidity of ceramic materials maintains the stability of their shape during aggressive conditions, thus protects the top selective layer. This unique feature is usually not possessed by most polymeric supports. To the best of our knowledge, there is no previous publication of TFC membranes fabricated by means of IP upon ceramic supports. Conventional ceramic composite membranes and methods of creating thereof have been previously reviewed comprehensively [17]. A common method to modify ceramic membranes is by dip coating, which is relatively easy to employ and has been widely studied [18,19]. However, the throughput of a dip coated membrane is low due to a thick selective layer (Z1 mm) that gives large transport resistance. The flux of a dip coated polyvinyl alcohol (PVA)/ceramic is 1.01 kg/m2 h for dehydration of ethyl acetate at 60 1C [18], whereas the flux of polyimide P84/γ-ceramic is
1.4 kg/m2 h for dehydration of t-butanol at 95 1C [19], suggesting that this method cannot produce an ultrathin selective layer.

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

Another alternative of fabrication is by directly grafting a monomer onto a ceramic substrate to create a selective skin layer. It has been reported that a hydrophilic ceramic/polymeric pervaporation membrane was prepared by free-radical graft polymerization of acrylic acid onto the ceramic membrane [20]. The membrane had a separation factor of 1442 for dehydration of ethanol, but possessed a low flux of 0.54 kg/m2 h which was attributed to the pore blocking effect by the grafted polymer. Another method is to deposit self-assembled polyelectrolytes on ceramic substrates [21]. A 60-layer pair of polyethylenimine/poly (vinylsulfate) (PEI/PVS) showed a high flux of 18.4 kg/m2 h but a low separation factor for dehydration of ethanol, suggesting possible defects on the membrane. Table 1 summarizes some recent studies of ceramic composite membranes [18–23] and TFC membranes with polymeric supports [12–14] for pervaporation dehydration applications. We found that there are only few studies on ceramic composite membranes despite the promising advantages they may exhibit. This may be due to the difficulty of fabricating defect-free membranes, especially on the surface of low-cost microfiltration ceramic membranes that have been commercially applied for years [21]. In this study, thin film composite membranes by interfacial polymerization on top of ceramic membrane supports are prepared for pervaporation dehydration of isopropanol (IPA). IPA is an important solvent and cleaning agent with wide applications in semiconductor, microelectronic and pharmaceutical industries. Due to the formation of azeotropic mixture with water, the purification of IPA by pervaporation is more economical than conventional distillation process [24–30]. Thus we have chosen IPA to be studied by our newly developed membranes. On the surface of the ceramic support, we propose a coating of polydopamine (PDA) or polyethyleneimine (PEI) to improve the surface properties. PDA is formed from the polymerization of dopamine under the presence of oxygen [31]. As reported by earlier studies, the coating of PDA on any surface can reach as high as
40 nm [31,32]. PEI, a branched chain polymer with amine functional groups, has also been extensively used to modify membranes [33–36]. In view of significant advantages of a ceramic support over a polymeric support especially in some extreme conditions such as high temperature and pressure, it is important to explore the

35

feasibility of using α-alumina ceramic tubes as porous supports for thin film composite membranes via interfacial polymerization. The material engineering of the top selective layer and the manipulation of interfacial properties between the selective layer and the ceramic substrate are parts of our foci to improve the productivity and durability of pervaporation membranes. By creating an ultrathin selective layer through interfacial polymerization and minimizing the substrate resistance with a porous ceramic substrate, we aim to increase flux substantially without much sacrificing selectivity. Therefore, the objectives of this study are to: (1) conceptually demonstrate the possibility of fabricating high performance ceramic TFC pervaporation membranes with interfacial polymerization; (2) systematically investigate the surface morphology and topology change of the PDA- and PEI-modified ceramic TFC membranes and their intrinsic pervaporation performance under different testing temperatures; (3) examine the effects of polydimethylsiloxane (PDMS) coating on the ceramic TFC membranes and deduce the layer information of the membranes with positron annihilation spectroscopy; and (4) study the long term performance and stability of the prepared ceramic composite membranes.

2. Experimental 2.1. Materials The Inoceps α-alumina tube M20 (outside diameter 3.7 mm, inside diameter 2.7 70.1 mm, and average pore size 20 nm, supplied by Hyflux SIP Pte Ltd.) was utilized as the porous support for the fabrication of ceramic TFC membranes. Hyperbranched polyethyleneimine (PEI, molecular weight 60,000 g/mol), dopamine hydrochloride, m-phenylenediamine (MPD) (499%) and trimesoyl chloride (TMC) (4 98%), were supplied by SigmaAldrich. For the IP process, TMC was dissolved in n-hexane (499%, Merck). Isopropanol (IPA, 499.5%) for pervaporation experiments was purchased from Merck. The chemical structures of dopamine and PEI are shown in Fig. 1. Sylgards184 silicone elastomer and its curing agent were purchased from Dow Corning Singapore Pte Ltd. (Singapore). All chemicals were used as received.

Table 1 Literature comparison of ceramic composite membranes for pervaporation dehydration applications. Membrane

Feed compositions (wt%/wt%)

Temperature (1C)

Total flux (kg/m2 h)

Separation factor

Reference

IP/polyacrylonitrile IP/Torlon IP/polyetherimide Polyvinyl alcohol (PVA)/alumina Polyimide P84/γ-alumina PVA–chitosan/alumina PVA/γ-alumina poly(acrylic acid)/alumina polyelectrolyte/alumina IP/α-alumina

Ethanol/water (90/10) IPA/water (85/15) IPA/water (85/15) Ethyl acetate/water (95/5) n-butanol/water (95/5) Ethyl acetate/water (92/8) IPA/water (95/5) Ethanol/water (95/5) Ethanol/water (94/6) IPA/water (85/15)

25 50 50 60 95 50 80 30 65 80


1.75 1.28 3.5 1.05
1.20 2.22
1.00 0.54 18.40 6.05


600 624 278 633
1000 500
160 1168 8.2 1396

[12] [13] [14] [18] [19] [22] [23] [20] [21] This study

Fig. 1. Chemical structures of (a) dopamine and (b) polyethyleneimine.

36

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

2.2. Membrane preparation The ceramic tubes were cut into pieces with a length of
15 cm. These pieces were ultrasonically washed in deionized (DI) water for 20 min before further coating with PDA or PEI. Prior to the coating process, the ceramic tubes were wrapped with Teflon tape, in order to ensure that PDA or PEI solutions will only exclusively coat on the inner surface of the tubes. The PDA coating solution (2 g/L) was prepared by dissolving dopamine-hydrochloride in tris-buffer at pH ¼8.5. The ceramic tubes were immersed in the PDA coating solution for 24 h without agitation at ambient conditions. On the other hand, PEI was dissolved in DI water to obtain a 10 g/L solution. The PEI coating was performed by dip coating the ceramic tube in the PEI solution for 1 min. After dip coating, excess PEI solution was removed by blowing the tubes with air gun, and finally the tubes were dried in a vacuum oven at 100 1C for 10 min. The dip coating process was repeated once to ensure the formation of a PEI film on the inner surface of ceramic tubes. The PDA- and PEI-coated ceramic tubes were then used as substrates for formation of TFC membranes. The interfacial polymerization (IP) was carried out by firstly immersing PDA- and PEI-coated ceramic substrates in a 2 wt% MPD solution for 10 min, with the excess solution removed by blowing with air gun. Subsequently, these ceramic tubes were immersed in 0.1 wt% TMC solution for 2 min, and subjected to heat treatment at 65 1C for 10 min. The membranes after the IP process were designated as PDA-IP and PEI-IP. PDA-IP and PEI-IP ceramic membranes were then immersed in a 3 wt% PDMS/n-hexane (Sylgads 184) solution for 3 min and allowed the PDMS layer to cure for at least 24 h before pervaporation experiments. The PDAIP and PEI-IP coated with PDMS were designated as PDA-IP–PDMS and PEI-IP–PDMS, respectively. The prepared ceramic membranes were then potted into a stainless steel module comprising two tees and a tube for pervaporation tests. Each module consisted of one ceramic tube. Cotton wool was stuffed at the end of each module to improve sealing. Epoxy (KS Bond EP231, Bondtec) was then added to cure the ends of modules overnight. The effective fiber length in each module for pervaporation tests was 11 cm with an effective surface area of 9.40 cm2. 2.3. Pervaporation experiments A laboratory scale pervaporation unit was employed with the details described elsewhere [37]. A feed solution of IPA/water

Cross section

(85/15 wt%) was used and the feed flow rate was maintained at 20 l/h for each module. The permeate pressure was maintained o2 mbar by a vacuum pump. Retentate and permeate samples were collected after the membrane being conditioned for at least 2 h. The flux J was determined by Eq. (1): J¼

Q AT

ð1Þ

where Q is the total mass transferred over a time t and A is the membrane area. The mass of permeate was weighed by using a Mettler Toledo balance. The separation factor, β was defined by Eq. (2): βi=j ¼

ywi =ywj xwi =xwj

ð2Þ

where subscripts i and j refer to water and IPA, respectively. yw and xw are the weight fractions of a component in the permeate and feed, respectively. They were analyzed using a Hewlett-Packard GC 7890 with a HP-INNOWAX column and a TCD detector. The permeance Pi/l of the membranes was defined as follows: Pi p ¼ J i =ðxi γ i P sat i yi P Þ l

ð3Þ

where Pi is the membrane permeability of component i and is a product of diffusivity and solubility coefficient, l is the membrane thickness, xi and yi are the feed and permeate mole fractions of component i respectively, Pp is the total pressure at the permeate side. P sat is the partial vapor pressure of component i at the feed i side and γi, is the activity coefficient, and were obtained from the AspenTech Process Modeling V7.2. The permeance of component i, expressed in GPU (1  10-6 cm3 (STP)/cm2 s cm Hg), can be determined if other terms are known. The membrane selectivity αi/j was defined by Eq. (4): αi=j ¼

P i =l P j =l

ð4Þ

2.4. Characterizations The pore size distribution was determined using a 1500 AE Capillary Flow Porometer (Porous Materials Inc., USA). The membrane pores were filled by Galwick wetting liquid with a surface tension of 15.9 dynes cm  1. Detailed procedures have already been described elsewhere [38]. The ceramic tube membrane was cut open by a blade for preparation of SEM and AFM samples.

Inner cross section

Overall cross section 100 µm Inner surface

100 nm Outer surface

1 mm

10 µm Fig. 2. FE-SEM images of an original ceramic tube.

10 µm

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

The membrane morphology was observed by using a JSM-6700F field emission scanning electron microscope (FESEM). To understand the roughness and topology of the surface of modified membranes, an atomic force microscope (AFM, Agilent Technology, USA) with the contact mode was employed. The mean roughness Rms was calculated based on the AFM measurement. An X-ray photoelectron spectrometer (XPS, Kratos XPS SystemAXIS His-165 Ultra) was utilized to measure the surface chemical compositions of the ceramic membranes. The PAS experiments were carried out using a variable monoenergy beam (0–30 keV). Membranes were prepared as per Section 2.2 on the outer surface of the ceramic tubes. A number of ceramic tubes were laid neatly on an aluminum holder with a width and a length larger than 1.5 cm. An HP Ge detector (EG&G Ortec. with 35% efficiency and energy resolution of 1.5 keV at 511 keV peak) was used to measure the Doppler broadening energy spectra (DBES) with a counting rate of approximately 800 cps. The S parameter, defined as a ratio of integrated counts between 510.3 and 511.7 keV to total counts after subtracting the background noise counts, measured the relative value of the low momentum part of positron electron annihilation. The S parameters of DBES were measured as a function of positron incident energy (mean depth) at ambient temperature. A million counts for each DBES were conducted. The positron incident energy E (keV) was correlated to the mean depth Z (nm) of the polymeric material

35 25 20 15 10 5 0

Diameter (µm)

Z¼

40 1:6 E ρ

ð5Þ

where ρ is the density of the polymer material in g/cm3.

3. Results and discussion 3.1. Effects of coating on surface of ceramic tubes by PDA or PEI As shown in Fig. 2, the inner surface of an asymmetric ceramic tube is smoother than the outer surface and is selected as the surface to be further modified. The pore size of the ceramic tube is mostly in the range of 20 nm with some large pore components which are probably attributed to defects from the ceramic preparation process (Fig. 3). In order to create an ultrathin selective layer on the porous ceramic inner surface, a gutter layer with good adhesion to the ceramic surface may be required to further reduce the large pore components. A typical approach is to deposit another thin layer of γ-alumina, but this process is not economical [40]. Another thin organic layer such as polydopamine (PDA) and polyethyleneimine (PEI) (Fig. 1) may serve as a good gutter layer between the organic and inorganic materials. PDA has been reported to possess excellent adhesive capability on all kind of surfaces [31], while PEI has a high free energy of adhesion of 191 J/mol [41]. In addition, PDA may react with TMC which has been reported by earlier study [42] while PEI has been utilized as an amine monomer to react with TMC for TFC membranes [13]. As shown in Fig. 4, some larger pores in the original surface have been reduced by either PDA or PEI coating. Moreover, a change in color from white to black after coating with PDA, as observed by bare eyes, confirms the formation of a PDA layer on the ceramic surface. Further characterizations by XPS (Table 2) suggest that a gutter layer of PDA or PEI has been formed on top of the alumina surface. The existence of Al signal in both PDA- and PEI-coated surfaces may imply that (1) the gutter layer is thinner than the XPS probing depth of 9 nm; (2) the gutter layer is not continuous owing to rough surface of original ceramic tubes. Moreover, the Al signal of the PDA-coated ceramic membrane is 8.68% which is greatly reduced from that of the original ceramic tube (30.83%), suggesting that PDA strongly adheres on the ceramic substrate.

x1000

Fig. 3. Pore size distribution of the original ceramic tube.

according to following equation [39]:

10 µm

x50000

Frequency (%)

30

37

100 nm Fig. 4. FE-SEM images of (a) the original ceramic tube, and the ceramic tubes coated with (b) PEI and (c) PDA at different magnifications.

38

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

3.2. Morphology of ceramic TFC membranes by interfacial polymerization Fig. 5 shows the AFM pictures of the original, PDA- and PEIcoated, and final ceramic TFC membranes. Contrary to our expectation, the roughness of the PDA-coated ceramic tube surface is larger than that of the original ceramic membrane. This is less desirable because more defects may form on the thin polyamide film during the interfacial polymerization. The larger roughness is probably contributed by particle generation during the PDA deposition on the ceramic surface which has been reported in some earlier studies [32,43] because the PDA coating was not under agitation. The PEI coating gives a much smoother surface which is more desirable for the fabrication of defect free thin TFC films through IP. After IP, the surface of TFC membranes on PEI-coated ceramic substrates has a smoother surface than that of PDA-coated ceramic substrates (i.e., roughness: 17 vs. 21.0 nm). Figs. 6 and 7 show the membrane morphology of the ceramic TFC membranes where the ceramic substrates were coated with PDA and

Table 2 A comparison of elements of PDA- and PEI-coated ceramic surfaces measured by XPS. Membrane

Original ceramic surface

PDA-coated ceramic PEI-coated ceramic surface surface

O (atomic conc%) C (atomic conc%) N (atomic conc%) Al (atomic conc%)

53.7

24.4

25.5

14.9

60.8

39.7

0.55 30.8

6.10

16.1

8.68

18.6

PEI, respectively. A continuous polyamide film with a thickness of about 100 nm can be seen on top of both PDA- and PEI-coated ceramic surface while both PDA and PEI layers are too thin to be observed. The TFC membrane synthesized on the PEI-coated tube (thereafter referred to as PEI-IP) seems to be less rough compared to the one on the PDA-coated ceramic tube (referred to as PDA-IP).

3.3. Pervaporation performance of ceramic TFC membranes 3.3.1. Effects of PDA or PEI coating on ceramic TFC membranes Table 3 shows the pervaporation performance of ceramic TFC membranes for dehydration of IPA at 50 1C. Both PDA-IP and PEI-IP have higher separation factors than the TFC membrane synthesized on the uncoated ceramic substrate (referred to as ceramicIP). The improved separation performance of PDA-IP and PEI-IP may be attributed to higher hydrophilicity [42] and smaller pores of PDA- and PEI-coated ceramic surfaces. Therefore, comparing to the uncoated ceramic substrate, PDA- and PEI-coated ceramic surfaces may absorb more MPD in the aqueous phase for the interfacial polymerization reaction with TMC in the organic phase. Since the PDA-coated ceramic surface has a higher roughness than the PEI-coated ceramic surface, PEI-IP has a better pervaporation performance than the PDA-PI. In addition, the amine functional group of PEI may react with TMC and further improve the separation performance. Since the membranes developed have some minor defects, a common approach to seal the defects is by applying the PDMS coating [44]. While PDMS is a dominant recovery material, its separation factors towards organics decrease to nearly unity for organic feed concentrations close to 80 wt% [45,46]. The PDMS coating on dehydration membranes has been reported to improve separation performance for pervaporation dehydration of IPA and acetic acid [26,47]. Moreover, an additional layer on top of the

Rms: 13.9±1.40 nm Rms: 21.0±1.90 nm

Rms:13.1±1.10 nm

Rms:10.7±0.70 nm

Rms:17.2±2.10 nm

Fig. 5. AFM images of (a) the original ceramic surface; (b) PDA-coated ceramic surface; (c) TFC surface on PDA-coated ceramic substrates; (d) PEI-coated ceramic surface and (e) TFC surface on PEI-coated ceramic substrates.

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

Inner surface

Inner surface

1 µm

Cross section

39

100 nm

Cross section

~100nm 1 µm

100 nm

Fig. 6. FE-SEM images of the ceramic TFC membrane where the ceramic substrate was coated with PDA (referred to as PDA-IP).

Inner surface

1 µm

Cross section

Inner surface

100 nm

Cross section ~100nm

1 µm

100 nm

Fig. 7. FE-SEM images of the ceramic TFC membrane where the ceramic substrate was coated with PEI (referred to as PEI-IP).

selective layer may prevent the selective layer from physical damage. As shown in Table 3, the PEI-IP membrane coated with PDMS exhibits a significant improvement in separation factor, with a less than 20% drop in the flux. Since the separation factor of PEI-IP–PDMS is 2800 which corresponds to the permeate of 99.7% water, suggesting defects have been effectively sealed. The PDA-IP membrane coated with PDMS does not exhibit much improvement possible due to too many surface defects which cannot be effectively sealed by PDMS [44].

3.3.2. Temperature effect Feed temperature is an important parameter for pervaporation tests. So far, most studies of TFC membranes for pervaporation separation were conducted at low temperatures from room temperature to 50 1C [11–13]. It is important to show that the developed TFC membrane can withstand a higher feed temperature. Table 4 summarizes the pervaporation dehydration of IPA as a function of feed temperature for both PDA-coated and PEI-coated TFC membranes. As temperature increases, separation factor

40

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

decreases while flux increases significantly for both ceramic composite membranes. As the operation temperature increases to 80 1C, fluxes of PEI- and PDA-coated ceramic composite membranes increase to more than 6 kg/m2 h which exceeds most of the fluxes of TFC membranes synthesized on polymeric supports (Table 1). This superior performance can be attributed to large pore sizes of ceramic substrates as well as no solvent-induced swelling in pore size and substrate morphology comparing to most polymeric substrates. Table 4 also tabulates permeance and selectivity as a function of feed temperature in order to examine the evolution of intrinsic membrane properties with temperature. Water permeances for both PDA-IP–PDMS and PEI-IP–PDMS decrease slightly when feed temperatures are above 60 1C. The decrease may be due to two reasons. Firstly, an increase in feed temperature reduces water sorption and thus lowers water permeance [48,49]. Secondly, it may be due to annealing effect at temperatures above 60 1C. Since the newly developed membranes were previously thermally treated only at 65 1C, the annealing would reduce free volume and permeability [12]. The decline of selectivity with increasing temperature is probably due to the reduced sorption abilities of the membrane towards water at higher temperatures [50]. Nevertheless, PEI-coated ceramic TFC membrane has a negligible IPA permeance at all feed temperatures. 3.4. PAS study of ceramic TFC membranes

3.5. Long term stability of ceramic-thin film membranes Fig. 10 shows the separation performance of the newly developed ceramic TFC membranes at 60 1C over 120 h. Small variations

Mean depth (μm) 0.49

0 Polymeric layer

0.47

0.283

0.093

0.540

Transition layer

1.224

Ceramic PEI-IP-PDMS PEI-IP

Ceramic support

0.45

0.857

0.43

S

In order to determine various layer thicknesses of the ceramic TFC membranes, positron annihilation spectroscopy (PAS) was carried out. PAS is an important characterization technique for probing free volume properties of membranes [39,51,52]. By analyzing the free volume depth profile of a composite membrane, various layer thicknesses of the membrane could be determined quantitatively by a computer program, VEPFIT [52]. Fig. 8 shows the variations of S parameter of the original ceramic tube, PEI-IP and PEI-IP–PDMS ceramic TFC membranes as a function of positron incident energy. The original ceramic membrane shows a monotonically decreasing S parameter as the membrane thickness increases, while each of other composite membranes have a peak value of S parameter (both at
1 keV). The decreasing S

parameter of the original ceramic membrane is due to the positron quenching effect of aluminum oxide [39], while commonly dense layers of polymeric membranes have the trend of increasing S parameter. The PEI-IP–PDMS has a two-step increase of S parameter before reaching its maximum value at
1 keV, suggesting a two-layer polymer structure on top of the ceramic substrate. To obtain the total thickness of both polyamide selective layer and PDMS coating layer on the ceramic TFC membranes, both 3layer and 4-layer models were employed in the VEPFIT analyses. Since PDMS and polyamide layer have very close S parameter values (
0.455 vs. 0.46), the 3-layer model fits the S parameter values remarkably better. This model describes that the top surface is mainly composed by three layers. The first layer is a polymer layer where the S parameter increases. The second layer is a transition layer which could be a diffusion layer where the polyamide and PEI could have been formed in the pores of ceramic. The third layer is basically the ceramic substrate. Fig. 8 plots the fitted S parameter values from VEPFIT analyses and shows a very good agreement with experimental data, while Table 5 summarizes the fitted values and the corresponding layer thicknesses. The thicknesses of polymer layers on PEI-IP and PEI-IP–PDMS are 79 and 105 nm, respectively; which are close to the thicknesses of
100 nm and
150 nm from SEM images (shown in Figs. 7 and 9, correspondingly). From their difference, the PDMS layer on top of the polyamide layer is estimated to be
26 nm.

0.41

Table 3 Pervaporation performance of ceramic thin film composite membranes (Feed: IPA/ Water 85/15 wt/wt% at 50 1C). 2

Membrane

Total flux (kg/m h)

Permeate water (wt%)

β(water/IPA)

Ceramic-IP PDA-IP PEI-IP PDA-IP–PDMS PEI-IP–PDMS

3.40 2.85 2.70 2.55 2.19

76.0 95.4 97.5 97.6 99.7

27.8 110 220 233 2800

0.39 0.37 0.35 0

2

4

6

8

10

Positron incident energy (keV) Fig. 8. S parameter vs. positron incident energy (or membrane depth) of the outer layers of uncoated ceramic, PEI-IP and PEI-IP–PDMS membranes. (The lines are from VEPFIT results in a 3-layer model, which are polymeric, transition and ceramic layers.)

Table 4 The effect of temperature on ceramic TFC membranes for pervaporation dehydration of IPA. (Feed: IPA/water
85/15 wt%). Membrane

Feed temp. (1C)

Total flux (kg/m2 h)

PDA-IP–PDMS

40 50 60 80

1.68 2.55 3.70 6.40

PEI-IP–PDMS

40 50 60 80

1.31 2.19 3.10 6.05

Note: 1 GPU¼ 1  10  6 cm3 (STP)/cm2 s cm Hg.

Separation factor (βwater/IPA)

Water permeance (GPU)

IPA permeance (GPU)

414 233 90 71

13,416 12,289 10,810 9755

29.5 47.0 105.3 118.1

3350 2800 2000 1396

10,561 10,768 9595 9917

2.87 3.43 4.20 6.10

Selectivity (αwater/IPA) 454.6 261.3 102.7 82.62 3678 3140 2282 1625

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

41

Table 5 Layer analyses for S data in interfacial polymerized polyamide/ceramic composite membranes. Membrane

S1

S2

S3

L1 (nm)

L2 (nm)

D1 (nm)

D2 (nm)

D3 (nm)

PEI-IP PEI-IP–PDMS

0.46807 0.0022 0.46497 0.0013

0.3959 70.0012 0.4091 70.0014

0.37247 0.0012 0.37757 0.0007

797 12 1057 13

2707 45 2197 21

387 6 317 7

54 722 52 714

107 723 182 786

(1) L1 and L2 are the boundary lengths from fitted S data using VEPFIT software. (2) Diffusion lengths D1, D2 and D3 are the lengths that positronium and positron diffuse in the polymer layer, transition layer and ceramic, respectively.

Surface

Cross section

~150nm

100 nm Fig. 9. Membrane morphology of the PEI-IP–PDMS ceramic composite membrane.

strong to maintain the mechanical integrity under the testing conditions. Clearly, the ceramic TFC is a promising pervaporation dehydration membrane with robustness suitable for applications in harsh feed conditions.

5

Flux (kg/m2h)

4 3 PDA-IP-PDMS PEI-IP-PDMS

2

4. Conclusions

1 0 0

20

40

60

80

100

120

140

Permeate water (wt%)

100

95

90

85

80 0

20

40

60

80

100

120

140

Time (h) Fig. 10. Long term pervaporation performance for PDA- and PEI-coated ceramic thin film composite membranes (Feed: IPA/water
85/15 wt%, temperature: 60 1C).

of separation performance for both PDA-IP–PDMS and PEI-IP– PDMS are mainly due to some variations of feed compositions. Nevertheless, the permeate water for PDA-IP–PDMS and PEI-IP– PDMS composite are stable at about 94 wt% and 99.5 wt% throughout the testing period. The sustainable high separation performance for the developed ceramic composite membranes indicates that the thin polyamide selective layer is not damaged and the adhesion between the thin film and ceramic support is sufficiently

In this study, we have developed novel ceramic TFC membranes for pervaporation dehydration of IPA. A gutter layer of PDA or PEI was coated on the α-alumina ceramic surface to make the ceramic surface to be more suitable for interfacial polymerization reaction. A number of conclusions can be drawn as following: (1) The PEI-coated ceramic substrate is more desirable for interfacial polymerization than the PDA-coated ceramic substrate because the former provides a smoother surface than the latter and the original ceramic substrates. As a result, a less defective polyamide layer can be formed. (2) The PDMS coating on PEI-IP is effective in sealing defects of the TFC membrane and improves the separation performance for IPA dehydration with only a modest decrease of separation flux. (3) The PEI-coated ceramic TFC membrane has achieved high separation performance (J: 6.05 kg/m2 h, β: 1396) for pervaporation dehydration of IPA at 80 1C. (4) Based on PAS analyses, the newly developed ceramic TFC membrane consists of an ultrathin polyamide layer of
79 nm and a PDMS coating layer of
26 nm. (5) The newly developed ceramic TFC membrane exhibits stable separation performance for more than 120 h for pervaporation dehydration of IPA.

Acknowledgments The authors would like to acknowledge the financial support provided by the National Research Foundation of Singapore (NRF)

42

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

through the Competitive Research Program for the project entitled, “New Biotechnology for Processing Metropolitan Organic Wastes into Value-Added Products” (Grant no. R-279-000-311281). Thanks are due to Dr. Dave Mangindaan for proofreading the manuscript.

References [1] P.W. Morgan, S.L. Kwolek, Interfacial polycondensation. 2. Fundamentals of polymer formation at liquid interfaces, Journal of Polymer Science 40 (1959) 299–327. [2] J.E. Cadotte, Interfacially synthesized reverse osmosis membrane, US Patent 4,277,344, FilmTec Corp., 1981. [3] L.T. Rozelle, J.E. Cadotte, K.E. Cobian, C.V. Kopp Jr., Nonpolysaccharide membranes for reverse osmosis: NS-100 membranes, in: S. Sourirajan (Ed.), Reverse Osmosis and Synthetic Membranes, National Research Council of Canada, Ottawa, 1977(Chapter 12). [4] R.J. Petersen, Composite reverse-osmosis and nanofiltration membranes, Journal of Membrane Science 83 (1993) 81–150. [5] M.F.J. Solomon, Y. Bhole, A.G. Livingston, High flux membranes for organic solvent nanofiltration (OSN)-interfacial polymerization with solvent activation, Journal of Membrane Science 423 (2012) 371–382. [6] S.P. Sun, T.A. Hatton, S.Y. Chan, T.S. Chung, Novel thin-film composite nanofiltration hollow fiber membranes with double repulsion for effective removal of emerging organic matters from water, Journal of Membrane Science 401–402 (2012) 152–162. [7] N.Y. Yip, A. Tiraferri, W.A. Phillip, J.D. Schiffman, M. Elimelech, High performance thin-film composite forward osmosis membrane, Environmental Science and Technology 44 (2010) 3812–3818. [8] N. Widjojo, T.S. Chung, M. Weber, C. Maletzko, V. Warzelhan, The role of sulphonated polymer and macrovoid-free structure in the support layer for thin-film composite (TFC) forward osmosis (FO) membranes, Journal of Membrane Science 383 (2011) 214–223. [9] Y.T. Chern, L.W. Chen, Preparation of composite membranes via interfacial polyfunctional condensation for gas separation applications, Journal of Applied Polymer Science 44 (1992) 1087–1093. [10] F. Yuan, Z. Wang, S.C. Li, J.X. Wang, S.C. Wang, Formation–structure–performance correlation of thin film composite membranes prepared by interfacial polymerization for gas separation, Journal of Membrane Science 421 (2012) 327–341. [11] S.H. Huang, C.J. Hsu, D.J. Liaw, C.C. Hu, K.R. Lee, J.Y. Lai, Effect of chemical structures of amines on physicochemical properties of active layers and dehydration of isopropanol through interfacially polymerized thin-film composite membranes, Journal of Membrane Science 307 (2008) 73–81. [12] S.H. Huang, W.S. Hung, D.J. Liaw, C.H. Lo, W.C. Chao, C.C. Hu, C.L. Li, K.R. Lee, J.Y. Lai, Interfacially polymerized thin-film composite polyamide membranes: effects of annealing processes on pervaporative dehydration of aqueous alcohol solutions, Separation and Purification Technology 72 (2010) 40–47. [13] J. Zuo, Y. Wang, S.P. Sun, T.S. Chung, Molecular design of thin film composite (TFC) hollow fiber membranes for isopropanol dehydration via pervaporation, Journal of Membrane Science 405–406 (2012) 123–133. [14] J. Zuo, Y. Wang, T.S. Chung, Novel organic-inorganic thin film composite membranes with separation performance surpassing ceramic membranes for isopropanol dehydration, Journal of Membrane Science 433 (2013) 60–71. [15] A.P. Korikov, P.B. Kosaraju, K.K. Sirkar, Interfacially polymerized hydrophilic microporous thin film composite membranes on porous polypropylene hollow fibers and flat films, Journal of Membrane Science 279 (2006) 588–600. [16] S.C.J.M. van Hoof, J.G. Minnery, B. Mack, Dead-end ultrafiltration as alternative pre-treatment to reverse osmosis in seawater desalination: a case study, Desalination 139 (2001) 161–168. [17] G.P. Liu, W. Wei, W.Q. Jin, N.P. Xu, Polymer/ceramic composite membranes and their application in pervaporation process, Chinese Journal of Chemical Engineering 20 (2012) 62–70. [18] S.S. Xia, X.L. Dong, Y.X. Zhu, W. Wei, F.J. Xiangli, W.Q. Jin, Dehydration of ethyl acetate–water mixtures using PVA/ceramic composite pervaporation membrane, Separation and Purification Technology 77 (2011) 53–59. [19] R. Kreiter, D.P. Wolfs, C.W.R. Engelen, H.M. van Veen, J.F. Vente, Hightemperature pervaporation performance of ceramic-supported polyimide membranes in the dehydration of alcohols, Journal of Membrane Science 319 (2008) 126–132. [20] X.Z. Cao, T.Z. Zhang, Q.T. Nguyen, Y.Y. Zhang, Z.H. Ping, A novel hydrophilic polymer–ceramic composite membrane 1: Acrylic acid grafting membrane, Journal of Membrane Science 312 (2008) 15–22. [21] Y.W. Chen, F.J. Xiangli, W.Q. Jin, N.P. Xu, Organic–inorganic composite pervaporation membranes prepared by self-assembly of polyelectrolyte multilayers on macroporous ceramic supports, Journal of Membrane Science 302 (2007) 78–86. [22] Y.X. Zhu, S.S. Xia, G.P. Liu, W.Q. Jin, Preparation of ceramic-supported poly (vinyl alcohol)–chitosan composite membranes and their applications in pervaporation dehydration of organic/water mixtures, Journal of Membrane Science 349 (2010) 341–348.

[23] T.A. Peters, C.H.S. Poeth, N.E. Benes, H.C.W.M. Buijs, F.F. Vercauteren, J.T. F. Keurentjes, Ceramic-supported thin PVA pervaporation membranes combining high flux and high selectivity; contradicting the flux-selectivity paradigm, Journal of Membrane Science 276 (2006) 42–50. [24] P.D. Chapman, T. Oliveira, A.G. Livingston, K. Li, Membranes for the dehydration of solvents by pervaporation, Journal of Membrane Science 318 (2008) 5–37. [25] L.Y. Jiang, T.S. Chung, R. Rajagopalan, Dehydration of alcohols by pervaporation through polyimide Matrimid (R) asymmetric hollow fibers with various modifications, Chemical Engineering Science 63 (2008) 204–216. [26] R.X. Liu, X.Y. Qiao, T.S. Chung, The development of high performance P84 copolyimide hollowfibers for pervaporation dehydration of isopropanol, Chemical Engineering Science 60 (2005) 6674–6686. [27] X.Y. Qiao, T.S. Chung, K.P. Pramoda, Fabrication and characterization of BTDATDI/MDI (P84) co-polyimide membranes for the pervaporation dehydration of isopropanol, Journal of Membrane Science 264 (2005) 176–189. [28] X.S. Feng, R.Y.M. Huang, Preparation and performance of asymmetric polyetherimide membranes for isopropanol dehydration by pervaporation, Journal of Membrane Science 109 (1996) 165–172. [29] H.A. Tsai, W.H. Chen, C.Y. Kuo, K.R. Lee, J.Y. Lai, Study on the pervaporation performance and long-term stability of aqueous iso-propanol solution through chitosan/polyacrylonitrile hollow fiber membrane, Journal of Membrane Science 309 (2008) 146–155. [30] V.V. Namboodiri, R. Ponangi, L.M. Vane, A novel hydrophilic polymer membrane for the dehydration of organic solvents, European Polymer Journal 42 (2006) 3390–3393. [31] H. Lee, S.M. Dellatore, W.M. Miller, P.B. Messersmith, Mussel-inspired surface chemistry for multifunctional coatings, Science 318 (2007) 426–430. [32] V. Ball, D.D. Frari, V. Toniazzo, D. Ruch, Kinetics of polydopamine film deposition as a function of pH and dopamine concentration: insights in the polydopamine deposition mechanism, Journal of Colloid and Interface Science 386 (2012) 366–372. [33] Z.A. Chen, M.C. Deng, C. Yong, G.H. He, M. Wu, J.D. Wang, Preparation and performance of cellulose acetate/polyethyleneimine blend microfiltration membranes and their applications, Journal of Membrane Science 235 (2004) 73–86. [34] P. Kumar, S. Kim, J. Ida, V.V. Guliants, Polyethyleneimine-modified MCM-48 membranes: effect of water vapor and feed concentration on N2/ CO2 selectivity, Industrial and Engineering Chemistry Research 47 (2008) 201–208. [35] S.P. Sun, T.A. Hatton, T.S. Chung, Hyperbranched polyethyleneimine induced cross-linking of polyamide-imide nanofiltration hollow fiber membranes for effective removal of ciprofloxacin, Environmental Science and Technology 45 (2011) 4003–4009. [36] L. Setiawan, R. Wang, K. Li, A.G. Fane, Fabrication of novel poly(amide-imide) forward osmosis hollow fiber membranes with a positively charged nanofiltration-like selective layer, Journal of Membrane Science 369 (2011) 196–205. [37] G.M. Shi, Y. Wang, T.S. Chung, Dual-layer PBI/P84 hollow fibers for pervaporation dehydration of acetone, AIChE Journal 58 (2012) 1133–1145. [38] P. Sukitpaneenit, T.S. Chung, Molecular design of the morphology and pore size of PVDF hollow fiber membranes for ethanol–water separation employing the modified pore-flow concept, Journal of Membrane Science 374 (2011) 67–82. [39] Y.C. Jean, P.E. Mallon, D.M. Schrader, Principles and Applications of Positron and Positronium Chemistry, World Scientific Publishing Co Pte Ltd., Singapore, 2003. [40] S. Sommer, T. Melin, Performance evaluation of microporous inorganic membranes in the dehydration of industrial solvents, Chemical Engineering and Processing 44 (2005) 1138–1156. [41] J.E. Mark, Polymer Data Handbook, Oxford University Press, New York, 1999. [42] G. Han, S. Zhang, X. Li, N. Widjojo, T.S. Chung, Thin film composite forward osmosis membranes based on polydopamine modified polysulfone substrates with enhancements in both water flux and salt rejection, Chemical Engineering Science 80 (2012) 219–231. [43] X. Li, S. Zhang, F.J. Fu, T.S. Chung, Deformation and reinforcement of thin-film composite (TFC) polyamide-imide (PAI) membranes for osmotic power generation, Journal of Membrane Science 434 (2013) 204–217. [44] J.M.S. Henis, M.K. Tripodi, Composite hollow fiber membranes for gas separation: the resistance model approach, Journal of Membrane Science 8 (1981) 233–246. [45] T. Uragami, T. Morikawa, Permeation and separation characteristics of alcohol–water mixtures through poly (dimethyl siloxane) membrane by pervaporation and evapomeation, Journal of Applied Polymer Science 44 (1992) 2009–2018. [46] Y.K. Hong, W.H. Hong, Influence of ceramic support on pervaporation characteristics of IPA/water mixtures using PDMS/ceramic composite membrane, Journal of Membrane Science 29–39 (1999) 159. [47] J. Bai, A.E. Fouda, T. Matsuura, J.D. Hazlett, A study on the preparation and performance of polydimethylsiloxane-coated polyetherimide membranes in pervaporation, Journal of Applied Polymer Science 48 (1993) 999–1008. [48] G.M. Shi, T.X. Yang, T.S. Chung, Polybenzimidazole (PBI)/zeolitic imidazolate frameworks (ZIF-8) mixed matrix membranes for pervaporation dehydration of alcohols, Journal of Membrane Science 415–416 (2012) 577–586. [49] X.Y. Qiao, T.S. Chung, Fundamental characteristics of sorption, swelling, and permeation of P84 co-polyimide membranes for pervaporation dehydration of alcohols, Industrial and Engineering Chemistry Reserach 44 (2005) 8938–8943.

G.M. Shi, T.-S. Chung / Journal of Membrane Science 448 (2013) 34–43

[50] Y. Wang, T.S. Chung, M. Gruender, Sulfonated polybenzimidazole membranes for pervaporation dehydration of acetic acid, Journal of Membrane Science 415–416 (2012) 486–495. [51] G.M. Shi, H.M. Chen, Y.C. Jean, T.S. Chung, Sorption, swelling, and free volume of polybenzimidazole (PBI) and PBI/zeolitic imidazolate framework (ZIF-8) nano-composite membranes for pervaporation, Polymer 54 (2013) 774–783.

43

[52] H.M. Chen, W.S. Hung, C.H. Lo, S.H. Huang, M.L. Cheng, G. Liu, K.R. Lee, J.Y. Lai, Y.M. Sun, C.C. Hu, R. Suzuki, T. Ohdaira, N. Oshima, Y.C. Jean, Free-volume depth profile of polymeric membranes studied by positron annihilation spectroscopy: layer structure from interfacial polymerization, Macromolecules 40 (2007) 7542–7557.