- Email: [email protected]
TPU blends asymmetric hollow fiber membranes prepared with the use of hydrophilic additive PVP (K30)
Desalination 223 (2008) 438–447
Porous PVDF/TPU blends asymmetric hollow fiber membranes prepared with the use of hydrophilic additive PVP (K30) Zhou Yuan, Xi Dan-Li* Department of Environmental Science and Engineering, Donghua University, Shanghai, 201620, People’s Republic of China Tel. +86 21 6779 2545; Fax +86 21 6219 2522; email: [email protected] Received 19 December 2006; accepted 3 January 2007
Abstract Asymmetric blend hollow fiber membranes had been made from a new casting dope containing poly(vinylidene fluoride) (PVDF)/Thermoplastic polyurethane (TPU)/Polyvinylpyrrolidone (PVP) N, N-dimethylacetamide (DMAc). The effect of hydrophilic additive, polyvinylpyrrolidone, on the morphology and crystal structure of PVDF/TPU blends membranes by phase inversion process was studied. The separation property, microstructure and crystalline phase of membranes were characterized by bovine serum albumin (BSA) retention experiments, scanning electron microscopy (SEM), Fourier transform infrared spectroscopy-attenuated total reflection (FTIR-ATR) and differential scanning calorimetric (DSC), respectively. The results showed that adding low concentration (less than 3 wt%) of PVP can reduce the hydrophobicity of PVDF and increase its hydrophilicity. With further increment of PVP, the solution demixing is delayed and the kinetic hindrance due to the increase of viscosity. By SEM it was observed that with the increase of PVP the pyriform voids were replaced by the ‘finger-like’ voids, which became longer and widespread. In further increment of PVP (5 wt%), PVP contributes to the suppression rather than the enlargement of macrovoid structure in the membranes. For formed membranes by additive-free casting dope, FTIR–ATR indicated that PVDF crystallized including α and β phase. TPU and low concentration of PVP both can reduce the crystallinity of PVDF membranes. When PVDF/TPU was precipitated from increasing amount of PVP in dopes (e.g. ≥3 wt%), some α phase disappeared and the water flux decreased little by little, the enthalpy of fusion and crystallinity of membranes also declined. When the addition of PVP increased as high as 10 wt%, the heat of fusion and crystallinity of membranes increased slightly compared with membranes adding 3 wt% PVP. In addition the water flux dropped significantly rather than improved. Keywords: Poly(vinylidene fluoride); Thermoplastic polyurethane; Polyvinyl pyrrolidone; Blend; Membrane; Crystallization
*Corresponding author. Presented at the conference on Desalination and the Environment. Sponsored by the European Desalination Society and Center for Research and Technology Hellas (CERTH), Sani Resort, Halkidiki, Greece, April 22–25, 2007. 0011-9164/06/$– See front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.desal.2007.01.184
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
1. Introduction Polymeric membranes can be prepared by phase inversion process. The demixing of a polymer solution can be induced by compositional changes of constituents due to mass transfer in a system. One important goal in membrane technology is to control membrane structure and membrane performance. Due to its excellent chemical resistance and good physical and thermal stability, PVDF industrial waste treatment including oily emulsion, organic/water separations, gas absorption and stripping, membrane distillation and ultrafiltration. PVDF membranes prepared were not of practical interest because of their low permeate fluxes [1]. In addition, they are nowadays too expensive to consider for environmental depolluting applications used widely. Blending is frequently used for improving the properties of polymeric membranes [2–4]. Addition of a second polymer often leads to a significant change in the membrane morphology. Thermoplastic polyurethane, with properties covering from a high performance elastomer to tough thermoplastic and low price, has been extensively used due to its superior physical properties (e.g. high tensile strength, abrasion and tear resistance, low temperature flexibility, etc.) and high versatility in chemical structures [5]. TPU membrane has been used for clinical applications due to its superior cytocompatibility. Addition of organic hydrophilic polymer in casting dope to prepare porous membranes by phase inversion method has been an effective practice to improve the permeability and selectivity of ultrafiltration membranes. As far as we know, the additives are widely used for structure control of membranes such as polyvinylpyrrolidone (PVP) and Polyethylene glycol (PEG) [1,6]. Low rejection and large pore size are caused by the fact that large molecular weight PVP tends to form thicker skin layer containing bigger pores [1,7]. Xu et al. studied the effect of PVP for different molecular weight on morphology of
439
polyetherimide hollow fiber membrane. They found that the higher molecular weight of PVP was added, the bigger pore was made [8] on the other hand, Wienk et al. draw that PVP is an agent for suppressing macropore formation in the phase inversion membranes [9]. Much less studies on PVP not only has nonsolvent characteristics (demixing enhancer) but also property of suppressing macropore formation (demixing hindrance). In the present research, PVP are used as an additive in PVDF/TPU/DMAc solution systems for the development of improved performance of hollow fiber membranes. The effect of PVP on the final structure and performance of the hollow fiber membrane will be analyzed. PVP is thermodynamic enhancement and kinetic hindrance during phase inversion in the membrane preparation. Crystal forms of PVDF changes for dopes with addition of PVP, as revealed by FTIR–ATR, which is due to the increase of viscosity causing stress variation of nascent fiber. Variation of the heat of fusion (ΔHf) and melting temperature of PVDF by DSC analysis of these membranes reveals crystallinity of PVDF in membranes changes, which is coherent relative with crystal structure variation. Furthermore, addition of PVP in the dopes is found that morphology of the membranes’s macrovoids varies with different PVP content in the dope, which examines using scanning electron microscope to observe. These morphological features are consistent with the measured water flux and retention data. 2. Experimental 2.1. Material The polyvinylidene fluoride used was a commercial product (FR904), thermoplastic polyurethane obtained from Townsand Corporation (RE-FLEX585-XU), was an amorphous polymer. N, N-dimethylacetamide (DMAc, >99% reagent) was employed as the solvent and distilled water was used as nonsolvent for the polymers
440
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
precipitation, polyvinylpyrrolidone (PVP) (K30) (MW 25,000–40,000) was added to the casting dopes to modify the membrane structure. BSA (MW 67,000) was used to characterize the separation performance of hollow fiber membranes. 2.2. Preparation of PVDF/TPU blend membrane The membranes were made by the phase inversion method. Casting dopes (80°C) were prepared by dissolving PVDF and TPU in the solvent and adding PVP to casting dopes under stirring. Dopes consisted 16 wt% PVDF/TPU and 84 wt% PVP/ DMAc mixture. Their compositions and viscosities with different amount of PVP are shown in Table 1. Casting solutions were deposited at 50°C in a no sun light place for one day for removing air bubbles from it and the degassed dope was transferred to a stainless steel reservoir and pressurized to 0.1 MPa using nitrogen. In general, water as internal coagulant adjusted at 3 mL/mi, the air gap was kept at 14 cm for the spinning runs. The spinneret with an orifice diameter/inner diameter of 0.8/0.2 mm was used. The nascent fiber was guided through water baths at 25°C at a take-up velocity, carefully adjusted to match the free falling velocity to complete the solidification process. The detailed experimental procedures can be found elsewhere [10]. The formed membranes
were washed in water bath for at least 24 h to remove the traces of PVP and DMAc. 2.3. Characterization of membrane The membranes prepared were characterized by the methods showed below: (1) The viscosity of the prepared casting dope was measure using digital rheometer (Model NDJ-9 S, Shanghai Precision Instruments Co. Ltd.) at the constant temperature of 25°C. The cloud points of the casting dopes were obtained by measuring the amount of precipitant, distilled water, added to the dopes until the solutions revealed the cloudy feature. (2) Morphologies of the membranes were observed in cross-section views by SEM (JSM5600 LV, JEOL Ltd.). (3) The thermal behaviors of membrane were observed using a differential scanning calorimeter (PerkinElmer DSC-7, PerkinElmer, Inc.). The temperature was raised from 25 to 250°C at a rate of 5°C/min in a nitrogen atmosphere. The melting temperature and the heat of fusion (ΔHf) were determined from the obtained thermograms. (4) FTIR-ATR spectra were obtained by a NEXUS670 (Nicolet Instruments Corp.) spectrometer with 4 cm−1 resolution. It was employed to examine the residual PVP content
Table 1 Cloud points and viscosity data of dopes with different PVP contentsa Composition (wt%) PVP
DMAc
PVDF/TPU
0 1 3 5 7 10
84 83 81 79 77 74
12.8/3.2 12.8/3.2 12.8/3.2 12.8/3.2 12.8/3.2 12.8/3.2
a
Polymer concentration is 16 wt% .
Water content at cloud point (wt%)
Viscosity (cps)
16.67 15.25 13.04 10.71 9.09 7.41
3520 4653 5507 7020 7947 8687
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
in the formed membranes, the transition of crystal type of PVDF was observed. (5) Membrane porosity was measured in the method of dry-wet weight. (6) Water permeation of the membrane was measured in a died-end ultrafiltration cup. The permeation performances of BSA through various membranes at transmembrane pressure of 0.1 Mpa were measured to see the solute retention capabilities of the membranes.
441
molecular weight. The diffusion can be inhibited by PVP in the dope due to the increase of the kinetical hindrance or a delayed exchange between solvent (DMAc) and nonsolvent (water) during the phase inversion process. Namely, it is concluded that PVP addition could function as the thermodynamical enhancer and kinetic hindrance in PVDF/TPU casting dope demixing. Thermodynamics and kinetic properties are correlative in the phase inversion system in the preparation of PVDF/TPU hollow fiber membranes.
3. Results and discussion 3.1. Thermodynamic and dynamic effect on membrane-forming system by PVP As an additive PVP can induce a dual effect on a casting dope. One is a thermodynamic variation due to the nonsolvent effect and the other a rheological varation due to high molecular weight of PVP. The cloud point data was shown in Table 1. The original binary dope of 16 wt% polymer (PVDF/TPU) in solvent (DMAc) was phase separated as water added into the dope reached to 16.67 wt%. With the addition of PVP, in the thermodynamic equilibrium casting dope for phase inversion needed much less amount of water at a demixing point. The water content to demix PVDF/TPU dope decreased with the increasing PVP concentration. With 3 wt% PVP the water concentration causing demixing was reduced to 13.04 wt% and with 5 wt% PVP to 10.71 wt%. The cloud point data suggested PVP introduced to the casting dope reduce the miscibility of dope with water and PVP can work in favor of the enhancement of the demixing of casting dope thermodynamically. The rheological (dynamic) behavior was estimated by measuring the viscosity of dopes. The addition of PVP to casting dope increased its viscosity as shown in Table 1. For 3 wt% additive the viscosity increased up to about 56% and the addition of 10 wt% increased up to about 150%. This effect was the type of additive, PVP’s high
3.2. Crystal structure and vibrational spectra Both infrared and Raman spectra distinguish the different crystalline forms of PVDF and the presence of polymer chain defects, as shown by many studies [11,12]. Studies of the crystallization process and its dependence on solvent nature, temperature and time indicate that form α type is thermodynamically favored, while form β type is kinetically favored [12]. However, an actual PVDF sample, depending on the preparation condition, may present one or more of the different crystalline structures [12,13]. In Fig. 1, FTIR-ATR spectra of membranes prepared by immersing a dope containing free-additive, 3 wt%
Fig. 1. FTIR-ATR spectra of membranes with various amount of PVP.
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
3.3. Thermal behavior of the PVDF/TPU/PVP membrane Melting temperatures and the heat of fusion (ΔHf) of PVDF in various membranes were determined by means of DSC thermal analysis. In Fig. 2 the thermograms of different PVP content were presented, for which the scanning rate was 5°C/min in all experiments. It can be seen that these membranes exhibited a melting behavior with a progressive decrease of the average melting temperatures with the increase of PVP. The reason for this may be that the crystal structure was different during the processes of membranes formation. The similar results were obtained by Prest and Luca when they studied the melting behavior of melt-crystallized PVDF samples [15]. Membranes with PVDF crystallites in different crystal types (e.g. α, β type) also exhibit different melting temperatures. Laroche et al. obtained that α and β type of PVDF exhibits two melting
6
5 (a) 4 Power (mW)
3 (b) 2 (c) 1
0
5 13 0 13 5 14 0 14 5 15 0 15 5 16 0 16 5 17 0 17 5 18 0 18 5 19 0 19 5 20 0
0
12
and 10 wt% PVP respectively was shown that when the dope was free of PVP, the membrane precipitated into α and β type. Well known peaks of the α type of PVDF located at 795 and 976 cm−1 could be recognized very distinctly. The peak at 1275 cm−1 of the β type is seen clearly as well. With PVP addition partial α type of PVDF in the membrane disappeared. When PVP content was 3 wt%, α type located at 795 and 976 cm−1 disappeared. When PVP content added to 10 wt%, this phenomenon remained. That revealed that crystallinity of PVDF in membranes reduction with the increase of PVP content. The reason is that during the spinning procedure, PVP addition makes the solution viscosity increase (Table 1), the spinning speed slows down, which causes stress variation of nascent fiber. While, this stress is a main factor bring to crystalline structure transitions [14]. It is also concluded that with the increase of PVP concentration, the effect of thermodynamic variation is much less than kinetic variation on membrane formation.
12
442
Temperature (°C)
Fig. 2. DSC thermograms of membranes with various PVP concentrations (a) no additive; (b) 3 wt%; (c) 10 wt%.
peak which are close to and it appears that the β type has a lower melting point than α type [16]. In Fig. 2 it also observed that when PVP concentration added to 10 wt%, two melting peaks which were close to were obtained obviously and the average melting temperature lowered. DSC analysis revealed α type which displays higher melting temperature had a trend of reduction resulting in lower melting temperature (Table 2) with increasing PVP content. The result was also supported by crystal structure variation obtained by the FTIR-ATR spectra. The crystallinity of each membrane prepared with different PVP content was calculated based on the heat of fusion data of ideal PVDF crystal (ΔH f0 = 105 J/g) [17]. As the peak area revealed the crystallinity had obvious changes. ΔHf of membranes obtained directly from the DSC thermograms, ΔHf of PVDF obtained by dividing the ΔHf of membrane by the weight fraction of PVDF in the membrane [18]. The results were shown in Table 2. With addition of PVP up to 3 wt% in the casting dope, the crystallinity dropped to 30%, while progressive increment PVP content in the
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
443
Table 2 The enthalpy of fusion and crystallinity of PVDF/TPU/PVP membranes Composition (wt%)
DSC
PVDF/TPU
PVP
DMAc
PVDF/TPU
Tm (°C)
ΔHf (membrane) (J/g)
ΔHf (PVDF) (J/g)
Crystallinity (%)
12.8/3.2 12.8/3.2 12.8/3.2
0 3 10
84 81 74
12.8/3.2 12.8/3.2 12.8/3.2
163.29 162.84 161.04
26.06 18.35 22.74
32.58 22.94 28.43
31.02 21.85 27.07
dope resulted in increasing the crystallinity of PVDF in membrane, which impelled the crystallization process. DSC measurement revealed that crystallization is more favorable in a high concentration of PVP. It is beneficial to suppress the polymer crystallinity in the membrane formation process [19]. Growth of polymer crystallinity has a detrimental effect to the final membrane transport properties as it both decreases the free volume of amorphous region available for species transport and increases the membrane tortuosity [20]. 3.4. Morphologies of PVDF/TPU membranes Fig. 3 showed SEM micrographs of the crosssection of membranes prepared with increasing PVP content. All membranes demonstrated the so-called asymmetric morphology being characterized by a thin skin and a porous bulk that comprises fully developed macropores extending to the central or even toward the bottom region of the membrane, irrespective of the amount of PVP addition in the casting dope. With increase of PVP content the asymmetric morphology exhibited evident. With a low content of PVP (3 wt%) in the casting dope, the macropores enlarged obviously (Fig. 3(b)), while as PVP content in the dope increased, the size of macrovoid decreased, their structure being stretched down like a finger from the underneath of membrane skin to the bottom of a sublayer.
Thermodynamics and kinetics, correlating each other during phase inversion, are two dominating factors controlling the morphologies of membranes. At a low level of PVP in the casting dope, the thermodynamic driving force played a major role on solution demixing, inducing the demixing enhancement, corresponding to the acceleration of phase separation due to the PVP’s nonsolvent effect. Thus the PVP acts as a phase separation enhancer, resulting in macropore enlargement (Fig. 3(b)). At a high level of PVP, Water and DMAc mutual diffusion was inhibited with PVP in solution due to the increase of the dope viscosity, which affected the exchange between water and DMAc during the phase inversion and phase separation time increased and demixing delayed. Table 3 also reveled that crystallization is more favorable in a high concentration of PVP. Crystallization favors to take place during the precipitation process, it takes a considerably long time for nonsolvent to reach the bottom region to induce liquid–liquid demixing [21]. Under the delayed demixing condition, membranes can be expected low porosity and permeability, with suppressing formation of macrovoid [22]. PVP in the PVDF/ TPU dope works as an agent for suppressing macrovoid formation (kinetic factor) at high concentration, while at a low concentration for enhancing a macrovoid formation (thermodynamic effect). PVP can not come to terms with the demixing enhancement, it is not because the PVP fails to induce thermodynamic enhancement but because it is overwhelmed by kinetic hindrance [23].
444
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447 (a)
(b)
(c)
(d)
Fig. 3. SEM micrographs of membranes added into different concentration of PVP (a) no additive; (b) 3 wt% PVP; (c) 5 wt% PVP; (d) 10 wt%.
Table 3 Properties of membranes added different PVP contenta PVDF:TPU
PVP (wt%)
Water flux (L m−2 h−1)
Retention (%)
Porosity (%)
Take-up speed (m min−1)
100:0 80:20 80:20 80:20 80:20 80:20 80:20
0 0 1 2 3 5 10
9.25 28.25 55.41 321.55 440.43 346.73 32.47
83.34 82.54 80 79 77.24 86.88 80.22
56 69.52 79.05 82.00 83.60 81.40 69.22
13.00 11.30 10.00 8.00 7.50 6.50 5.90
a
Polymer concentration is 16 wt%.
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
Additionally, variation of macrovoid can be explained that the competitive thermodynamic and kinetic effect of PVP on the phase inversion process [24]. During the formation of hollow fiber membranes, due to an air gap exited, the mass transfer happened much earlier between the internal coagulant (water) and solvent (DMAc) preceding which between the external coagulant (water) and solvent (DMAc). Smolders et al. suggested that macrovoid growth is controlled by solvent diffusional flows and the macrovoids are formed at distances further away from the membrane toplayer [25].Thus macrovoids generally grew near internal surface, varied significantly the morphology of the membrane toplayer. In the case of adding 10 wt% PVP, owing to the increased viscosity of dope, the take-up speed slowed down and the time of solvent diffusion exposing in air prolonged before fibers reached the coagulation bath. Macrovoids expanded by solvent diffusional flows [25], hence the length of macrovoids became longer while the macropores were smaller in diameter, which resulted in more prominent asymmetric morphology. 3.5. Effect of the addition of PVP on membrane flux and molecular weight cut-off Membrane flux and retention were affected by the addition of PVP. Marchese et al. suggested that when PVP is entrapped by membrane materials, PVP should increase hydrophilicity, and results in important change in the performance of ultrafiltration membrane such as solute retention and fouling [26]. Below 3 wt% of PVP added the water flux was increased, after this point, however, the more the concentration of PVP in the casting dope, the less the flux. Although adding 3 wt% PVP water flux was max, retention had much loss. Even for 10 wt% of PVP added the water flux was close to fluxes of membranes without PVP. There has still been argued that the entrapped PVP offers hydrophilic to hydrophobic membranes and reduces the water flux due to the
445
swelling of the pore-filled PVP [27]. The water flux of membranes increases when higher amount of PVP was bleached out [28]. Growth of polymer crystallinity has a detrimental effect to the final membrane transport properties [20]. Without TPU and PVP, PVDF hollow fiber membrane was low permeate fluxes (Table 3). Addition of TPU made crystallinity of PVDF declined and flux raised. When PVP added into PVDF/TPU membranes, with the increase of PVP content, the crystallinity declined firstly and then with the sharp increase of viscosity (10 wt%) leaded to the demixing time of dope prolonged and the crystallinity of PVDF raised. Simultaneously the permeate fluxes enhanced first and then lessened (Table 3). A downhill of retention was found with PVP concentration below 3 wt% of PVP added, when the amount of PVP was larger than 3 wt%, the retention was increased up to a max. While PVP concentration exceeded 5 wt%, although the retention slightly declines, compared with additive-free PVDF/TPU membranes, the changes of retention was inconspicuous. PVP operates as a permeate flux enhancer at a low content (<3 wt%), but there is a sign that a high content PVP (10 wt%) is less effective for casting dope to improve permeate flux. 4. Conclusion Thermodynamic and kinetic variation is a key to understand the phase inversion process with addition of PVP into PVDF/TPU casting dope. Thermodynamically, PVP works as a demixing enhancer that accelerates the phase inversion reduces α crystal structure resulting in reduction of the crystallinity of PVDF in membranes, contributing to the enlargement macrovoid structure in the membranes, which improves the membrane flux. However, in further increment of PVP, kinetic impact is dominant. The demixing of the casting dope delayed and the crystallinity of PVDF increases, resulting in macrovoids being stretched down like a finger from the underneath
446
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
of membrane skin to the bottom of a sublayer. PVP in the PVDF/TPU dope suppress macrovoid formation at high concentration. The flux is decreased rather than increased. It reveals that kinetic effect takes over thermodynamic enhancement.
[9]
[10]
Acknowledgement The research was supported by the major science project of Shanghai Science and Technology Committee (NO. 012312032).
[11]
[12]
References [1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
D. Wang and K. Li, Preparation and characterization of polyvinylidene fluoride (PVDF) hollow fiber membranes, J. Membr. Sci., 163 (1999) 211–220. D.L. Wang, K. Li and W.K. Teo, Porous PVDF asymmetric hollow fiber membranes prepared with the use of small molecular additives, J. Membr. Sci., 178 (2000) 13–23. X.L. Yin, H.B. Cheng, X. Wang and Y.X. Yao, Morphology and properties of hollow-fiber membrane made by PAN mixing with small amount of PVDF, J. Membr. Sci., 146 (1998) 179–184. D.J. Lin, C.L. Chang, C.K. Lee and L.P. Cheng, Preparation and characterization of microporous PVDF/PMMA composite membranes by phase inversion in water/DMSO solutions, J. Eur. Polym., 42 (2006) 2407–2418. Q.W. Lu, W. Christopher and Macosko, Comparing the compatibility of various functionalized polypropylenes with thermoplastic polyurethane (TPU), J. Polym., 45 (2004) 1981–1991. J.H. Kim and K.H. Lee, Effect of PEG additive on membrane formation by phase inversion, J. Membr. Sci., 138 (1998) 153–163. B. Jung, Effect of molecular weight of polymeric additives on formation, permeation properties and hypochlorite treatment of asymmetric polyacrylonitrile membranes, J. Membr. Sci., 163 (2004) 45–57. Z.L. Xu, T.S. Chung and Y. Huang, Effect of polyvinylpyrrolidone molecular weights on morphology, oil/water separation, mechanical and thermal properties of polyetherimide/polyvinylpyrrolidone
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
hollow fiber membranes, J. Appl. Polym. Sci., 74 (1999) 2220–2233. S.P. Deshmukh and K. Li, Effect of ethanol composition in water coagulation bath on morphology of PVDF hollow fibre membranes, J. Membr. Sci., 150 (1998) 75–85. Y. Bormashenko, R. Pogreb, O. Stanevsky and E. Bormashenko, Vibrational spectrum of PVDF and its interpretation, J. Polym. Test., 23 (2004) 791–796. T. Boccaccio, A. Bottino, G. Capannelli and P. Piaggio, Characterization of PVDF membranes by vibrational spectroscopy, J. Membr. Sci., 210 (2002) 315–329. B.E.E. Mohajir and N. Heymans, Changes in structural and mechanical behaviour of PVDF with processing and thermomechanical treatments. 1. Change in structure, J. Polym., 42 (2001) 5661–5667. P. Ren, H. Zhang and G.F. Zhang, Study of the crystalline of the melt-spinning fiber of PVDF, Journal of Tianjin Polytechnic University, 22 (2003) 8–13. W.M. Prest and D.J. Luca, The morphology and thermal response of high-temperature crystallized poly(vinylidence fluoride), J. Appl. Phys., 46 (1975) 4136–4143. G. Laroche, Y. Marois and R. Guidoin, polyvinylidene fluorid (PVDF) as a biomaterial: From polymeric raw material to monofilament vascular suture, J. Biomed. Mat. Res., 29 (1995) 1525–1536. K. Nakagawa and Y. Isahida, Annealing effects in poly(vinylidene fluoride) as revealed by specific volume measurements, J. Polym. Sci. Polym. Phys. Edn., 11 (1973) 2153–2171. Q.P. Guo, Completely miscible ternary blends — III. Poly(vinylidenefluoride)-poly(methyl methacrylate)-poly(vinyl acetate), Eur. Polym. J., 32 (1996) 1409–1413. M.L. Yeow, Y. Liu and K. Li, Preparation of porous PVDF hollow fibre membrane via a phase inversion method using lithium perchlorate (LiClO4) as an additive, J. Membr. Sci., 258 (2005) 16–22. R.E. Kesting, Synthetic Polymeric Membranes: A Structural Perspective, 2nd edn., Wiley, New York, 1985, p. 348. L.P. Cheng, D.J. Lin, C.H. Shih, A.W. Dwan and C.C. Gryte, J. Polym. Sci. B Polym. Phys., 37 (1995) 2079–2092. P. van de Witte, P.J. Dijkstra, J.W.A. van den Berg and J. Feijen, Phase separation processes in
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447 polymer solutions in relation to membrane formation, J. Membr. Sci., 117 (1996) 1–31. [22] K.W. Lee, B.K. Seo, S.T. Nam and M.J. Han, Trade-off between thermodynamic enhancement and kinetic hindrance during phase inversion in the preparation of polysulfone membranes, Desalination, 159 (2003) 289–296. [23] E. Fontananova, J.C. Jansen, A. Cristiano, E. Curcio and E. Drioli, Effect of additives in the casting solution on the formation of PVDF membranes, Desalination, 192 (2006) 190–197. [24] C.A. Smolder, A.J. Reuver, R.M. Boom and I.M. Wienk, Microstructures in phase-inversion membranes. Part I. Formation of macrovoids, J. Membr. Sci., 73 (1992) 259–275.
447
[25] J. Marchese, M. Ponce and N.A. Ochoa, Fouling behaviour of polyethersulfone UF membranes made with different PVP, J. Membr Sci., 211 (2003) 1–11. [26] I.M. Wienk, E.E.B. Meuleman, Z. Borneman, T.V.D. Boomgaard and C.A. Smolders, Chemical treatment of membranes of a polymer blend: Mechanism of the reaction of hypochlorite with poly(vinylpyrrolidone), J. Polym. Sci. A Polym. Chem., 33 (1995) 49–54. [27] B. Jung, Effect of molecular weight of polymeric additives on formation, permeation properties and hypochlorite treatment of asymmetric polyacrylonitrile membranes, J. Membr. Sci., 243 (2004) 45–57.
Porous PVDF/TPU blends asymmetric hollow fiber membranes prepared with the use of hydrophilic additive PVP (K30) Zhou Yuan, Xi Dan-Li* Department of Environmental Science and Engineering, Donghua University, Shanghai, 201620, People’s Republic of China Tel. +86 21 6779 2545; Fax +86 21 6219 2522; email: [email protected] Received 19 December 2006; accepted 3 January 2007
Abstract Asymmetric blend hollow fiber membranes had been made from a new casting dope containing poly(vinylidene fluoride) (PVDF)/Thermoplastic polyurethane (TPU)/Polyvinylpyrrolidone (PVP) N, N-dimethylacetamide (DMAc). The effect of hydrophilic additive, polyvinylpyrrolidone, on the morphology and crystal structure of PVDF/TPU blends membranes by phase inversion process was studied. The separation property, microstructure and crystalline phase of membranes were characterized by bovine serum albumin (BSA) retention experiments, scanning electron microscopy (SEM), Fourier transform infrared spectroscopy-attenuated total reflection (FTIR-ATR) and differential scanning calorimetric (DSC), respectively. The results showed that adding low concentration (less than 3 wt%) of PVP can reduce the hydrophobicity of PVDF and increase its hydrophilicity. With further increment of PVP, the solution demixing is delayed and the kinetic hindrance due to the increase of viscosity. By SEM it was observed that with the increase of PVP the pyriform voids were replaced by the ‘finger-like’ voids, which became longer and widespread. In further increment of PVP (5 wt%), PVP contributes to the suppression rather than the enlargement of macrovoid structure in the membranes. For formed membranes by additive-free casting dope, FTIR–ATR indicated that PVDF crystallized including α and β phase. TPU and low concentration of PVP both can reduce the crystallinity of PVDF membranes. When PVDF/TPU was precipitated from increasing amount of PVP in dopes (e.g. ≥3 wt%), some α phase disappeared and the water flux decreased little by little, the enthalpy of fusion and crystallinity of membranes also declined. When the addition of PVP increased as high as 10 wt%, the heat of fusion and crystallinity of membranes increased slightly compared with membranes adding 3 wt% PVP. In addition the water flux dropped significantly rather than improved. Keywords: Poly(vinylidene fluoride); Thermoplastic polyurethane; Polyvinyl pyrrolidone; Blend; Membrane; Crystallization
*Corresponding author. Presented at the conference on Desalination and the Environment. Sponsored by the European Desalination Society and Center for Research and Technology Hellas (CERTH), Sani Resort, Halkidiki, Greece, April 22–25, 2007. 0011-9164/06/$– See front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.desal.2007.01.184
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
1. Introduction Polymeric membranes can be prepared by phase inversion process. The demixing of a polymer solution can be induced by compositional changes of constituents due to mass transfer in a system. One important goal in membrane technology is to control membrane structure and membrane performance. Due to its excellent chemical resistance and good physical and thermal stability, PVDF industrial waste treatment including oily emulsion, organic/water separations, gas absorption and stripping, membrane distillation and ultrafiltration. PVDF membranes prepared were not of practical interest because of their low permeate fluxes [1]. In addition, they are nowadays too expensive to consider for environmental depolluting applications used widely. Blending is frequently used for improving the properties of polymeric membranes [2–4]. Addition of a second polymer often leads to a significant change in the membrane morphology. Thermoplastic polyurethane, with properties covering from a high performance elastomer to tough thermoplastic and low price, has been extensively used due to its superior physical properties (e.g. high tensile strength, abrasion and tear resistance, low temperature flexibility, etc.) and high versatility in chemical structures [5]. TPU membrane has been used for clinical applications due to its superior cytocompatibility. Addition of organic hydrophilic polymer in casting dope to prepare porous membranes by phase inversion method has been an effective practice to improve the permeability and selectivity of ultrafiltration membranes. As far as we know, the additives are widely used for structure control of membranes such as polyvinylpyrrolidone (PVP) and Polyethylene glycol (PEG) [1,6]. Low rejection and large pore size are caused by the fact that large molecular weight PVP tends to form thicker skin layer containing bigger pores [1,7]. Xu et al. studied the effect of PVP for different molecular weight on morphology of
439
polyetherimide hollow fiber membrane. They found that the higher molecular weight of PVP was added, the bigger pore was made [8] on the other hand, Wienk et al. draw that PVP is an agent for suppressing macropore formation in the phase inversion membranes [9]. Much less studies on PVP not only has nonsolvent characteristics (demixing enhancer) but also property of suppressing macropore formation (demixing hindrance). In the present research, PVP are used as an additive in PVDF/TPU/DMAc solution systems for the development of improved performance of hollow fiber membranes. The effect of PVP on the final structure and performance of the hollow fiber membrane will be analyzed. PVP is thermodynamic enhancement and kinetic hindrance during phase inversion in the membrane preparation. Crystal forms of PVDF changes for dopes with addition of PVP, as revealed by FTIR–ATR, which is due to the increase of viscosity causing stress variation of nascent fiber. Variation of the heat of fusion (ΔHf) and melting temperature of PVDF by DSC analysis of these membranes reveals crystallinity of PVDF in membranes changes, which is coherent relative with crystal structure variation. Furthermore, addition of PVP in the dopes is found that morphology of the membranes’s macrovoids varies with different PVP content in the dope, which examines using scanning electron microscope to observe. These morphological features are consistent with the measured water flux and retention data. 2. Experimental 2.1. Material The polyvinylidene fluoride used was a commercial product (FR904), thermoplastic polyurethane obtained from Townsand Corporation (RE-FLEX585-XU), was an amorphous polymer. N, N-dimethylacetamide (DMAc, >99% reagent) was employed as the solvent and distilled water was used as nonsolvent for the polymers
440
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
precipitation, polyvinylpyrrolidone (PVP) (K30) (MW 25,000–40,000) was added to the casting dopes to modify the membrane structure. BSA (MW 67,000) was used to characterize the separation performance of hollow fiber membranes. 2.2. Preparation of PVDF/TPU blend membrane The membranes were made by the phase inversion method. Casting dopes (80°C) were prepared by dissolving PVDF and TPU in the solvent and adding PVP to casting dopes under stirring. Dopes consisted 16 wt% PVDF/TPU and 84 wt% PVP/ DMAc mixture. Their compositions and viscosities with different amount of PVP are shown in Table 1. Casting solutions were deposited at 50°C in a no sun light place for one day for removing air bubbles from it and the degassed dope was transferred to a stainless steel reservoir and pressurized to 0.1 MPa using nitrogen. In general, water as internal coagulant adjusted at 3 mL/mi, the air gap was kept at 14 cm for the spinning runs. The spinneret with an orifice diameter/inner diameter of 0.8/0.2 mm was used. The nascent fiber was guided through water baths at 25°C at a take-up velocity, carefully adjusted to match the free falling velocity to complete the solidification process. The detailed experimental procedures can be found elsewhere [10]. The formed membranes
were washed in water bath for at least 24 h to remove the traces of PVP and DMAc. 2.3. Characterization of membrane The membranes prepared were characterized by the methods showed below: (1) The viscosity of the prepared casting dope was measure using digital rheometer (Model NDJ-9 S, Shanghai Precision Instruments Co. Ltd.) at the constant temperature of 25°C. The cloud points of the casting dopes were obtained by measuring the amount of precipitant, distilled water, added to the dopes until the solutions revealed the cloudy feature. (2) Morphologies of the membranes were observed in cross-section views by SEM (JSM5600 LV, JEOL Ltd.). (3) The thermal behaviors of membrane were observed using a differential scanning calorimeter (PerkinElmer DSC-7, PerkinElmer, Inc.). The temperature was raised from 25 to 250°C at a rate of 5°C/min in a nitrogen atmosphere. The melting temperature and the heat of fusion (ΔHf) were determined from the obtained thermograms. (4) FTIR-ATR spectra were obtained by a NEXUS670 (Nicolet Instruments Corp.) spectrometer with 4 cm−1 resolution. It was employed to examine the residual PVP content
Table 1 Cloud points and viscosity data of dopes with different PVP contentsa Composition (wt%) PVP
DMAc
PVDF/TPU
0 1 3 5 7 10
84 83 81 79 77 74
12.8/3.2 12.8/3.2 12.8/3.2 12.8/3.2 12.8/3.2 12.8/3.2
a
Polymer concentration is 16 wt% .
Water content at cloud point (wt%)
Viscosity (cps)
16.67 15.25 13.04 10.71 9.09 7.41
3520 4653 5507 7020 7947 8687
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
in the formed membranes, the transition of crystal type of PVDF was observed. (5) Membrane porosity was measured in the method of dry-wet weight. (6) Water permeation of the membrane was measured in a died-end ultrafiltration cup. The permeation performances of BSA through various membranes at transmembrane pressure of 0.1 Mpa were measured to see the solute retention capabilities of the membranes.
441
molecular weight. The diffusion can be inhibited by PVP in the dope due to the increase of the kinetical hindrance or a delayed exchange between solvent (DMAc) and nonsolvent (water) during the phase inversion process. Namely, it is concluded that PVP addition could function as the thermodynamical enhancer and kinetic hindrance in PVDF/TPU casting dope demixing. Thermodynamics and kinetic properties are correlative in the phase inversion system in the preparation of PVDF/TPU hollow fiber membranes.
3. Results and discussion 3.1. Thermodynamic and dynamic effect on membrane-forming system by PVP As an additive PVP can induce a dual effect on a casting dope. One is a thermodynamic variation due to the nonsolvent effect and the other a rheological varation due to high molecular weight of PVP. The cloud point data was shown in Table 1. The original binary dope of 16 wt% polymer (PVDF/TPU) in solvent (DMAc) was phase separated as water added into the dope reached to 16.67 wt%. With the addition of PVP, in the thermodynamic equilibrium casting dope for phase inversion needed much less amount of water at a demixing point. The water content to demix PVDF/TPU dope decreased with the increasing PVP concentration. With 3 wt% PVP the water concentration causing demixing was reduced to 13.04 wt% and with 5 wt% PVP to 10.71 wt%. The cloud point data suggested PVP introduced to the casting dope reduce the miscibility of dope with water and PVP can work in favor of the enhancement of the demixing of casting dope thermodynamically. The rheological (dynamic) behavior was estimated by measuring the viscosity of dopes. The addition of PVP to casting dope increased its viscosity as shown in Table 1. For 3 wt% additive the viscosity increased up to about 56% and the addition of 10 wt% increased up to about 150%. This effect was the type of additive, PVP’s high
3.2. Crystal structure and vibrational spectra Both infrared and Raman spectra distinguish the different crystalline forms of PVDF and the presence of polymer chain defects, as shown by many studies [11,12]. Studies of the crystallization process and its dependence on solvent nature, temperature and time indicate that form α type is thermodynamically favored, while form β type is kinetically favored [12]. However, an actual PVDF sample, depending on the preparation condition, may present one or more of the different crystalline structures [12,13]. In Fig. 1, FTIR-ATR spectra of membranes prepared by immersing a dope containing free-additive, 3 wt%
Fig. 1. FTIR-ATR spectra of membranes with various amount of PVP.
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
3.3. Thermal behavior of the PVDF/TPU/PVP membrane Melting temperatures and the heat of fusion (ΔHf) of PVDF in various membranes were determined by means of DSC thermal analysis. In Fig. 2 the thermograms of different PVP content were presented, for which the scanning rate was 5°C/min in all experiments. It can be seen that these membranes exhibited a melting behavior with a progressive decrease of the average melting temperatures with the increase of PVP. The reason for this may be that the crystal structure was different during the processes of membranes formation. The similar results were obtained by Prest and Luca when they studied the melting behavior of melt-crystallized PVDF samples [15]. Membranes with PVDF crystallites in different crystal types (e.g. α, β type) also exhibit different melting temperatures. Laroche et al. obtained that α and β type of PVDF exhibits two melting
6
5 (a) 4 Power (mW)
3 (b) 2 (c) 1
0
5 13 0 13 5 14 0 14 5 15 0 15 5 16 0 16 5 17 0 17 5 18 0 18 5 19 0 19 5 20 0
0
12
and 10 wt% PVP respectively was shown that when the dope was free of PVP, the membrane precipitated into α and β type. Well known peaks of the α type of PVDF located at 795 and 976 cm−1 could be recognized very distinctly. The peak at 1275 cm−1 of the β type is seen clearly as well. With PVP addition partial α type of PVDF in the membrane disappeared. When PVP content was 3 wt%, α type located at 795 and 976 cm−1 disappeared. When PVP content added to 10 wt%, this phenomenon remained. That revealed that crystallinity of PVDF in membranes reduction with the increase of PVP content. The reason is that during the spinning procedure, PVP addition makes the solution viscosity increase (Table 1), the spinning speed slows down, which causes stress variation of nascent fiber. While, this stress is a main factor bring to crystalline structure transitions [14]. It is also concluded that with the increase of PVP concentration, the effect of thermodynamic variation is much less than kinetic variation on membrane formation.
12
442
Temperature (°C)
Fig. 2. DSC thermograms of membranes with various PVP concentrations (a) no additive; (b) 3 wt%; (c) 10 wt%.
peak which are close to and it appears that the β type has a lower melting point than α type [16]. In Fig. 2 it also observed that when PVP concentration added to 10 wt%, two melting peaks which were close to were obtained obviously and the average melting temperature lowered. DSC analysis revealed α type which displays higher melting temperature had a trend of reduction resulting in lower melting temperature (Table 2) with increasing PVP content. The result was also supported by crystal structure variation obtained by the FTIR-ATR spectra. The crystallinity of each membrane prepared with different PVP content was calculated based on the heat of fusion data of ideal PVDF crystal (ΔH f0 = 105 J/g) [17]. As the peak area revealed the crystallinity had obvious changes. ΔHf of membranes obtained directly from the DSC thermograms, ΔHf of PVDF obtained by dividing the ΔHf of membrane by the weight fraction of PVDF in the membrane [18]. The results were shown in Table 2. With addition of PVP up to 3 wt% in the casting dope, the crystallinity dropped to 30%, while progressive increment PVP content in the
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
443
Table 2 The enthalpy of fusion and crystallinity of PVDF/TPU/PVP membranes Composition (wt%)
DSC
PVDF/TPU
PVP
DMAc
PVDF/TPU
Tm (°C)
ΔHf (membrane) (J/g)
ΔHf (PVDF) (J/g)
Crystallinity (%)
12.8/3.2 12.8/3.2 12.8/3.2
0 3 10
84 81 74
12.8/3.2 12.8/3.2 12.8/3.2
163.29 162.84 161.04
26.06 18.35 22.74
32.58 22.94 28.43
31.02 21.85 27.07
dope resulted in increasing the crystallinity of PVDF in membrane, which impelled the crystallization process. DSC measurement revealed that crystallization is more favorable in a high concentration of PVP. It is beneficial to suppress the polymer crystallinity in the membrane formation process [19]. Growth of polymer crystallinity has a detrimental effect to the final membrane transport properties as it both decreases the free volume of amorphous region available for species transport and increases the membrane tortuosity [20]. 3.4. Morphologies of PVDF/TPU membranes Fig. 3 showed SEM micrographs of the crosssection of membranes prepared with increasing PVP content. All membranes demonstrated the so-called asymmetric morphology being characterized by a thin skin and a porous bulk that comprises fully developed macropores extending to the central or even toward the bottom region of the membrane, irrespective of the amount of PVP addition in the casting dope. With increase of PVP content the asymmetric morphology exhibited evident. With a low content of PVP (3 wt%) in the casting dope, the macropores enlarged obviously (Fig. 3(b)), while as PVP content in the dope increased, the size of macrovoid decreased, their structure being stretched down like a finger from the underneath of membrane skin to the bottom of a sublayer.
Thermodynamics and kinetics, correlating each other during phase inversion, are two dominating factors controlling the morphologies of membranes. At a low level of PVP in the casting dope, the thermodynamic driving force played a major role on solution demixing, inducing the demixing enhancement, corresponding to the acceleration of phase separation due to the PVP’s nonsolvent effect. Thus the PVP acts as a phase separation enhancer, resulting in macropore enlargement (Fig. 3(b)). At a high level of PVP, Water and DMAc mutual diffusion was inhibited with PVP in solution due to the increase of the dope viscosity, which affected the exchange between water and DMAc during the phase inversion and phase separation time increased and demixing delayed. Table 3 also reveled that crystallization is more favorable in a high concentration of PVP. Crystallization favors to take place during the precipitation process, it takes a considerably long time for nonsolvent to reach the bottom region to induce liquid–liquid demixing [21]. Under the delayed demixing condition, membranes can be expected low porosity and permeability, with suppressing formation of macrovoid [22]. PVP in the PVDF/ TPU dope works as an agent for suppressing macrovoid formation (kinetic factor) at high concentration, while at a low concentration for enhancing a macrovoid formation (thermodynamic effect). PVP can not come to terms with the demixing enhancement, it is not because the PVP fails to induce thermodynamic enhancement but because it is overwhelmed by kinetic hindrance [23].
444
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447 (a)
(b)
(c)
(d)
Fig. 3. SEM micrographs of membranes added into different concentration of PVP (a) no additive; (b) 3 wt% PVP; (c) 5 wt% PVP; (d) 10 wt%.
Table 3 Properties of membranes added different PVP contenta PVDF:TPU
PVP (wt%)
Water flux (L m−2 h−1)
Retention (%)
Porosity (%)
Take-up speed (m min−1)
100:0 80:20 80:20 80:20 80:20 80:20 80:20
0 0 1 2 3 5 10
9.25 28.25 55.41 321.55 440.43 346.73 32.47
83.34 82.54 80 79 77.24 86.88 80.22
56 69.52 79.05 82.00 83.60 81.40 69.22
13.00 11.30 10.00 8.00 7.50 6.50 5.90
a
Polymer concentration is 16 wt%.
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
Additionally, variation of macrovoid can be explained that the competitive thermodynamic and kinetic effect of PVP on the phase inversion process [24]. During the formation of hollow fiber membranes, due to an air gap exited, the mass transfer happened much earlier between the internal coagulant (water) and solvent (DMAc) preceding which between the external coagulant (water) and solvent (DMAc). Smolders et al. suggested that macrovoid growth is controlled by solvent diffusional flows and the macrovoids are formed at distances further away from the membrane toplayer [25].Thus macrovoids generally grew near internal surface, varied significantly the morphology of the membrane toplayer. In the case of adding 10 wt% PVP, owing to the increased viscosity of dope, the take-up speed slowed down and the time of solvent diffusion exposing in air prolonged before fibers reached the coagulation bath. Macrovoids expanded by solvent diffusional flows [25], hence the length of macrovoids became longer while the macropores were smaller in diameter, which resulted in more prominent asymmetric morphology. 3.5. Effect of the addition of PVP on membrane flux and molecular weight cut-off Membrane flux and retention were affected by the addition of PVP. Marchese et al. suggested that when PVP is entrapped by membrane materials, PVP should increase hydrophilicity, and results in important change in the performance of ultrafiltration membrane such as solute retention and fouling [26]. Below 3 wt% of PVP added the water flux was increased, after this point, however, the more the concentration of PVP in the casting dope, the less the flux. Although adding 3 wt% PVP water flux was max, retention had much loss. Even for 10 wt% of PVP added the water flux was close to fluxes of membranes without PVP. There has still been argued that the entrapped PVP offers hydrophilic to hydrophobic membranes and reduces the water flux due to the
445
swelling of the pore-filled PVP [27]. The water flux of membranes increases when higher amount of PVP was bleached out [28]. Growth of polymer crystallinity has a detrimental effect to the final membrane transport properties [20]. Without TPU and PVP, PVDF hollow fiber membrane was low permeate fluxes (Table 3). Addition of TPU made crystallinity of PVDF declined and flux raised. When PVP added into PVDF/TPU membranes, with the increase of PVP content, the crystallinity declined firstly and then with the sharp increase of viscosity (10 wt%) leaded to the demixing time of dope prolonged and the crystallinity of PVDF raised. Simultaneously the permeate fluxes enhanced first and then lessened (Table 3). A downhill of retention was found with PVP concentration below 3 wt% of PVP added, when the amount of PVP was larger than 3 wt%, the retention was increased up to a max. While PVP concentration exceeded 5 wt%, although the retention slightly declines, compared with additive-free PVDF/TPU membranes, the changes of retention was inconspicuous. PVP operates as a permeate flux enhancer at a low content (<3 wt%), but there is a sign that a high content PVP (10 wt%) is less effective for casting dope to improve permeate flux. 4. Conclusion Thermodynamic and kinetic variation is a key to understand the phase inversion process with addition of PVP into PVDF/TPU casting dope. Thermodynamically, PVP works as a demixing enhancer that accelerates the phase inversion reduces α crystal structure resulting in reduction of the crystallinity of PVDF in membranes, contributing to the enlargement macrovoid structure in the membranes, which improves the membrane flux. However, in further increment of PVP, kinetic impact is dominant. The demixing of the casting dope delayed and the crystallinity of PVDF increases, resulting in macrovoids being stretched down like a finger from the underneath
446
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447
of membrane skin to the bottom of a sublayer. PVP in the PVDF/TPU dope suppress macrovoid formation at high concentration. The flux is decreased rather than increased. It reveals that kinetic effect takes over thermodynamic enhancement.
[9]
[10]
Acknowledgement The research was supported by the major science project of Shanghai Science and Technology Committee (NO. 012312032).
[11]
[12]
References [1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
D. Wang and K. Li, Preparation and characterization of polyvinylidene fluoride (PVDF) hollow fiber membranes, J. Membr. Sci., 163 (1999) 211–220. D.L. Wang, K. Li and W.K. Teo, Porous PVDF asymmetric hollow fiber membranes prepared with the use of small molecular additives, J. Membr. Sci., 178 (2000) 13–23. X.L. Yin, H.B. Cheng, X. Wang and Y.X. Yao, Morphology and properties of hollow-fiber membrane made by PAN mixing with small amount of PVDF, J. Membr. Sci., 146 (1998) 179–184. D.J. Lin, C.L. Chang, C.K. Lee and L.P. Cheng, Preparation and characterization of microporous PVDF/PMMA composite membranes by phase inversion in water/DMSO solutions, J. Eur. Polym., 42 (2006) 2407–2418. Q.W. Lu, W. Christopher and Macosko, Comparing the compatibility of various functionalized polypropylenes with thermoplastic polyurethane (TPU), J. Polym., 45 (2004) 1981–1991. J.H. Kim and K.H. Lee, Effect of PEG additive on membrane formation by phase inversion, J. Membr. Sci., 138 (1998) 153–163. B. Jung, Effect of molecular weight of polymeric additives on formation, permeation properties and hypochlorite treatment of asymmetric polyacrylonitrile membranes, J. Membr. Sci., 163 (2004) 45–57. Z.L. Xu, T.S. Chung and Y. Huang, Effect of polyvinylpyrrolidone molecular weights on morphology, oil/water separation, mechanical and thermal properties of polyetherimide/polyvinylpyrrolidone
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
hollow fiber membranes, J. Appl. Polym. Sci., 74 (1999) 2220–2233. S.P. Deshmukh and K. Li, Effect of ethanol composition in water coagulation bath on morphology of PVDF hollow fibre membranes, J. Membr. Sci., 150 (1998) 75–85. Y. Bormashenko, R. Pogreb, O. Stanevsky and E. Bormashenko, Vibrational spectrum of PVDF and its interpretation, J. Polym. Test., 23 (2004) 791–796. T. Boccaccio, A. Bottino, G. Capannelli and P. Piaggio, Characterization of PVDF membranes by vibrational spectroscopy, J. Membr. Sci., 210 (2002) 315–329. B.E.E. Mohajir and N. Heymans, Changes in structural and mechanical behaviour of PVDF with processing and thermomechanical treatments. 1. Change in structure, J. Polym., 42 (2001) 5661–5667. P. Ren, H. Zhang and G.F. Zhang, Study of the crystalline of the melt-spinning fiber of PVDF, Journal of Tianjin Polytechnic University, 22 (2003) 8–13. W.M. Prest and D.J. Luca, The morphology and thermal response of high-temperature crystallized poly(vinylidence fluoride), J. Appl. Phys., 46 (1975) 4136–4143. G. Laroche, Y. Marois and R. Guidoin, polyvinylidene fluorid (PVDF) as a biomaterial: From polymeric raw material to monofilament vascular suture, J. Biomed. Mat. Res., 29 (1995) 1525–1536. K. Nakagawa and Y. Isahida, Annealing effects in poly(vinylidene fluoride) as revealed by specific volume measurements, J. Polym. Sci. Polym. Phys. Edn., 11 (1973) 2153–2171. Q.P. Guo, Completely miscible ternary blends — III. Poly(vinylidenefluoride)-poly(methyl methacrylate)-poly(vinyl acetate), Eur. Polym. J., 32 (1996) 1409–1413. M.L. Yeow, Y. Liu and K. Li, Preparation of porous PVDF hollow fibre membrane via a phase inversion method using lithium perchlorate (LiClO4) as an additive, J. Membr. Sci., 258 (2005) 16–22. R.E. Kesting, Synthetic Polymeric Membranes: A Structural Perspective, 2nd edn., Wiley, New York, 1985, p. 348. L.P. Cheng, D.J. Lin, C.H. Shih, A.W. Dwan and C.C. Gryte, J. Polym. Sci. B Polym. Phys., 37 (1995) 2079–2092. P. van de Witte, P.J. Dijkstra, J.W.A. van den Berg and J. Feijen, Phase separation processes in
Z. Yuan, X. Dan-Li / Desalination 223 (2008) 438–447 polymer solutions in relation to membrane formation, J. Membr. Sci., 117 (1996) 1–31. [22] K.W. Lee, B.K. Seo, S.T. Nam and M.J. Han, Trade-off between thermodynamic enhancement and kinetic hindrance during phase inversion in the preparation of polysulfone membranes, Desalination, 159 (2003) 289–296. [23] E. Fontananova, J.C. Jansen, A. Cristiano, E. Curcio and E. Drioli, Effect of additives in the casting solution on the formation of PVDF membranes, Desalination, 192 (2006) 190–197. [24] C.A. Smolder, A.J. Reuver, R.M. Boom and I.M. Wienk, Microstructures in phase-inversion membranes. Part I. Formation of macrovoids, J. Membr. Sci., 73 (1992) 259–275.
447
[25] J. Marchese, M. Ponce and N.A. Ochoa, Fouling behaviour of polyethersulfone UF membranes made with different PVP, J. Membr Sci., 211 (2003) 1–11. [26] I.M. Wienk, E.E.B. Meuleman, Z. Borneman, T.V.D. Boomgaard and C.A. Smolders, Chemical treatment of membranes of a polymer blend: Mechanism of the reaction of hypochlorite with poly(vinylpyrrolidone), J. Polym. Sci. A Polym. Chem., 33 (1995) 49–54. [27] B. Jung, Effect of molecular weight of polymeric additives on formation, permeation properties and hypochlorite treatment of asymmetric polyacrylonitrile membranes, J. Membr. Sci., 243 (2004) 45–57.