Structuring and Characterization of a Novel Microporous PVDF Membrane with Semi-Interpenetrating Polymer Networks for Vacuum Membrane Distillation.
In membrane distillation (MD), only vapor state molecules can pass through the porous polymer membrane, the driving force of the separation process is the partial pressure difference of a component on the two sides of a membrane . There are four kinds of membrane distillation based on how the driving force is created, they are sweeping gas membrane distillation (SGMD), direct contact membrane distillation (DCMD), vacuum membrane distillation (VMD) and air gap membrane distillation (AGMD) , Compared with the other separation processes, the advantages of MD are high non-volatile rejection, low feed temperature and low operation pressure. Currently membrane distillation is widely used in environmental protection , water treatment [4-6], pharmaceutical industry  and food processing [8-10].
For MD separation, the porous polymer membrane is very crucial even though it is only used as a physical support, there-for it is studied widely. The membrane must be hydrophobic and its performances can be affected by many factors, such as the chemical properties of the polymer material, the hydrophobicity of the membrane, and the morphology-related characteristics such as the porosity, pore size and pore tortuosity of the membranes. At present, the MD membrane materials which have been used widely are polypropylene (PP), polytetrafluoro-ethylene (PTFE) and polyvinylidenefluoride (PVDF). Among them, PVDF is the most popular membrane material and has been studied by many researchers in recent years .
With the repeated unit of--[(C[H.sub.2]C[F.sub.2]).sub.n]--, PVDF is a kind of semicrystalline polymer, it has the advantages of good chemical resistance, high mechanical strength, good aging resistance as well as high thermal stability. These characteristics are very crucial for its actual applications in industry. In addition, PVDF has the merit of good processability, so it can be used to make hollow fiber membranes, flat sheet membranes and tubular membranes easily. PVDF can also be dissolved in a lot of common solvents such as dimethyl formamide (DMF), N,N-dimethylacetamide (DMAc) and N-methyl-2-pyrrolidone (NMP), so we can use the conventional method of non-solvent induced phase separation (NIPS) process to make the PVDF membrane. However, for PVDF membrane, water contact angle is about 70-90[degrees], which means that the hydrophobicity of PVDF membrane is not strong [12, 13], For this reason, a lot of research has been done to improve the hydrophobicity of the PVDF membranes, for example, by depositing the nanoparticles of Ti[O.sub.2] on the surface of the commercial PVDF membrane, Razmjou et al. developed the superhydrophobic PVDF membranes ; by mixing LiCl (used as the addictive) into the casting solution, Tomaszewska made the hydrophobic PVDF flat-sheet membrane  and they also studied the effect of LiCl concentration on the permeate flux of MD; by using the double-layer casting method of vapor-induced phase separation, Fan et al. fabricated the hydrophobic PVDF flat-sheet membrane for VMD ; by adding polyvinyl pyrrolidone) (PVP) and water as the additives of pore forming to the PVDF casting solution, Simone et al.  made the microporous hydrophobic VMD hollow fiber membranes of PVDF. In addition, grafting and blending a low energy material to PVDF membranes are also popular ways to acquire superhydrophobic membranes [18-20], For example, Wu et al.  made the PVDF/FEP(poly(tetrafluoroethylene-co-hexafluoropropylene)) flat-sheet membrane by immerged phase-inversion method. Teoh and Chung  increased the hydrophobicity of the PVDF-PTFE hollow fibers with a macrovoid free structure. Using the electrospun PVDF-PTFE nanofibrous scaffold, Dong et al.  fabricated the superhydrophobic nanofibrous membranes. Zhan et al.  developed the multi-layer composite membrane of PDMS/PVDF/nonwoven-fiber/PVDF/PDMS. In addition, blending is a simple way of membrane modification and is more possible to commercialize. Except the mechanical blending and membrane polymerization, there is a new way of membrane modification, which is the interpenetrating of polymer networks (IPNs).
In interpenetrating polymer networks (IPNs), when a multifunctional monomer reacts with another polymer, two distinct multi-functional polymers entangle with each other at the molecular level, then the polymeric structure forms [25-28]. According to fabricating methods, IPNs are divided into three types, they are sequential IPN, simultaneous IPN and semi-IPNs. In sequential IPN, crosslinked polymer I is swollen by a monomer II under the existence of the activating agent, cross-linking agent and polymerizing monomer II in situ. Simultaneous IPN includes two independent noninterfering polymerizations and the crosslinking reactions which happen simultaneously. In semi-IPNs, only one of the two polymers is crosslinked and the other polymer is linear .
As a kind of IPN, semi-IPNs usually exemplify good electrochemical stability, high compatibility, as well as strong mechanical property , Due to these good natures of the semi-IPN network, over the last decade, this topic has been studied widely. Cznotka et al.  made the solid polymer semi-IPNs electrolyte membrane through solution casting technique, they used poly(methyl methacrylate) (PMMA) as the polymeric host matrix, polysiloxane-combpropyloxymethoxytriglycol (PSx) as the ion conducting component and lithium bis(trifluoromethane)sulfonimide (LiTFSI) as the conducting salt. Cheng et al.  fabricated the semi-IPNs anion exchange membrane (AEMs) by immobilizing polyvinyl chloride (PVC) in divinylbenzene (DVB) copolymer (P(DMAM-co-DVB)) and dimethylaminoethyl methacrylate (DMAM), and the membrane was also tested in the process of acid recovery. Soldani et al.  developed a kind of semi-IPN biomedical material of PEtU-PDMS (poly(ether)urethane-polydimethylsiloxane).
Usually, the above mentioned semi-IPNs polymers were prepared through casting method and were used as polymer electrolytes in lithium batteries [28-30], However, the semi-IPNs membrane prepared through NIPS process for VMD has been studied scarcely. In our study, VMD membranes with semi-IPN structure were fabricated by producing the network between immobilized linear PVDF and crosslinked PDMS using NIPS process. A series of semi-IPNs membranes with different mass ratio of PDMS/TEOS were obtained, the membranes were characterized by ATR-FTIR, water contact angle (WCA), SEM, XRD, TGA, mechanical properties and pore structure. In addition, the semi-IPNs PDMS-PVDF membranes were tested in vacuum membrane distillation experiment to explore the performance of the membranes.
Polydimethylsiloxane (PDMS) (Silicone Rubber 107, 20000 cp) and PVDF (SOLEF[R]6020/1001) were obtained from Shanghai resin Company (China) and Solvay Solexis, Inc. (France), respectively. Solvent N,N-dimethylacetamide (DMAC, >99.5%), solvent tetrahydrofuran (THF) and additive Polyvinylpyrrolidone K30 (PVP, Mw30000) were purchased from Tianjin Tiantai Fine Chemical Co., Ltd (China), and solvent Triethyl phosphate (TEP) were bought from Shanghai Sinopharm Chemical Reagent Company (China). Catalyst Dibutyltin dilaurate (DBTDL) and crosslinking agent Ethyl orthosilicate four (TEOS) were provided by Tianjin Fuchen Chemical Reagent Factory (China). Ethanol (GR grade, 99.9%) and NaCl (GR grade, 99.5%) were produced by Beijing Chemical works (China). Deionized water was self-made in the lab. All chemicals used in this study were analytical grade without further purification.
Measurement of Casting Solution
Viscosity. The viscosity of casting solution can severely influence the exchange rate of the solvent and the non-solvent during a phase inversion process, and therefore, it can be used as an important parameter to regulate the precipitation kinetics and membrane morphologies. Viscosity of the casting solution was measured with rotary viscometer (NDJ-1, Shanghai An De Equipment Co., Ltd, China) under environmental temperature.
Light Transmittance. During the phase inversion process, light transmittance experiment can be used to show the precipitation kinetic of casting solution [32, 33], According to Kuo et al , the appearance of the optical inhomogeneities due to the demixing will lead light transmittance to decrease. The device in this experiment had been described in previous literatures [35, 36], From this experiment, light transmittance curve as a function of time was acquired and reorganized.
Semi-IPNs microporous PDMS-PVDF membranes were prepared through NIPS process and the procedures were as follows. First, to make PVDF solution, PVDF powder was dried at 60[degrees]C in an Electro-Thermostatic Blast Oven (DHG-9143BS-III, Shanghai CIMO Medical Instrument Manufacturing Co., LTD) to remove moistures; then a certain amount of dried PVP powder and PVDF powder were dissolved in a mixed solvent of TEP and DMAC under mechanical stirring for 2 h at 60[degrees]C to form the PVDF solution; after that, the solution was put into the Electro-Thermostatic Blast Oven at 60[degrees]C and stay there for 24 h to ensure the completely dissolving of the polymers; after totally dissolved, PVDF solution was taken out and cooled at room temperature. At the same time, PDMS solution was prepared by dissolving a certain amount of TEOS and PDMS in THF at room temperature. Then the above two solutions were mixed under stirring at 80 rpm for 4 h at room temperature to obtain the homogeneous PDMS-PVDF solution. Finally, certain amount of DBTDL was added to the above PDMS-PVDF solution and was stirred for 0.5 h, and then the homogeneous casting solution was obtained. After the deaeration of 14 h, with a glass rod, the casting solution was cast on a glass plate under 60% relative humidity and 25[degrees]C. Exposed in air for 40 s, the film was immersed into the coagulation bath (deionized water) together with the glass plate. Taken out from the coagulation bath, the precipitated film was peeled off from the glass plate and was soaked in fresh deionized water for 3 days to remove the residues of the solvent. The deionized water was changed twice a day. At last, the membranes were soaked in a 33.3% ethanol solution for a day, and then were taken out and were dried in air. The semi-IPNs structures of the resulting membrane are shown in Figs. 1 and , from which we can see the cross linking reaction happened between PDMS and TEOS.
The preparation conditions of the semi-IPNs PDMS-PVDF membranes are listed in Table 1 which also showed the mass ratios of PDMS/TEOS.
SEM-EDX. Membrane morphologies were investigated using scanning electron microscopy (SEM, JSM-5600LV, Japan). First, membrane samples were frozen in liquid nitrogen, then being fractured to get fragments and sputtered with gold by Ion Sputtering device (KYKY SBC-12), after that the samples were observed with SEM to get their morphologies. The relative element compositions of the bottom surface and the top surface of the produced membranes were recorded by EDX detector (EDAX-Falcon).
Membrane Porosity and Thickness. Membrane porosity [epsilon] is defined as the volume of the pores in the membrane divided by the total volume of the membrane. It can be obtained by gravimetric method and calculated according to the following equation:
[epsilon] = ([m.sub.w] - [m.sub.D])/[rho] x A x [delta] (1)
Where mw is the weight of the wet membrane which has been soaked in a low wetting liquid (surface tension 16dyn/cm) (Profil supported by IB-FT GmbH Germany), g; [m.sub.D] is the weight of a dry membrane, g; A is membrane effective area, [m.sup.2]; [delta] is membrane thickness, cm; [rho] is Profil density (1.87 g/[cm.sup.3]).
Membrane thickness was measured by an electronic outside micrometer (0-150 mm, Shanghai constant measuring tool Co., Ltd). For one membrane, five measurements on different locations of this membrane surface were conducted and were averaged.
Pore Size Distribution. To characterize the pore size distribution (PSD) and the mean pore radius ([r.sub.m]) of a membrane, a capillary flow porometer (POROLUX 500, Germany) was employed using the wet/dry flow method. Firstly, membrane sample was wetted with the Profil (a low surface tension wetting liquid), and then it was placed into a sealed chamber through which the Nitrogen gas flowed. The gas was forced to pass the flat membrane from the downside to the upside, and the pressure of the flowing gas was gradually increased and was recorded its through flow simultaneously. When the pressure of the gas was high enough for the liquid to be removed from the largest pore, then the size of the pore was recorded, which was the largest pore size (bubble point). With the increasing of gas pressure, the smaller pores became unblocked and caused the increase in the gas flow, when the whole membrane became dried, the pressure of the gas got to the cumulative. This highest pressure was used to calculate the pore size distribution and the average pore size. Non-cylindrical pores (such as the pores in PVDF membranes) may have many diameters, but the wet/dry porometry method only measures one diameter for per pore, that is the smallest diameter (throat).
Water Contact Angle. Water contact angles (WCA) of both sides of the membrane were tested at room temperature using the HARKE-SPCA contact angle meter (Beijing Hakko test instrument factory). A certain amount of droplets of water were carefully placed on the surface of a membrane sample using a tight syringe. The final result was the average value of 10 measurements at different position of the surface of three membrane samples.
XRD. The X-ray diffraction (XRD) spectra of the untreated PDMS-PVDF and the crosslinked PDMS-PVDF membranes were acquired at ambient temperature using an X-ray diffractometer (D/MAX 2000/PC, Japan; 60 kV, 300 mA). The diffractograms were measured at a scanning speed of 5[degrees]/min in a 20 range of 5-50[degrees].
TGA. Membrane thermal stability was checked out by using PerkinElmer simultaneous thermal Analyzer (STA 6000) under the temperature ranging from 30 to 800[degrees]C with a continuous heating at a rate of 10[degrees]C [min.sup.-1] and with a continuous flushing of inert gas N2 at a flowing rate of 20 ml*[min.sup.-1].
ATR-FTIR. FTIR spectra of the prepared membranes were gained through attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR, Nicolet iS10, Thermo Scientific). Small pieces of membranes were pressed against an ATR crystal. All FTIR spectra were recorded in the range of 4,000400 [cm.sup.-1].
Mechanical Properties. Membrane mechanical properties were obtained by a tensile tester (3365, Instron) with the loading velocity 20 mm/min at room temperature. To get an average value, at least three samples were tested.
VMD experiments were performed to obtain the permeate performances of the semi-IPNs PDMS-PVDF membranes. The schematic process of the vacuum membrane distillation was illustrated in our previous work . A prepared membrane was put into a stainless steel membrane module which had a contact area of 50 [cm.sup.2]. The bottom surface of the membrane was in contact with the feed stream. Feed solution was heated to certain temperature in a temperature-controlled water bath. Feed flow rate was regulated and controlled by a rotameter (LZB-10). The permeate side was cooled by a chiller (DLSB) using the mixture of water and ethanol as the refrigerant. NaCl solution of certain concentration was prepared through dissolving NaCl in deionized water and was used as the feed.
Penetrant was collected at 1 h intervals, the conductivity of the permeate was analyzed by the conductivity meter (DDS307A) to determine the salt content of the penetrant. Permeate flux and salt rejection were calculated from the given equation:
J = Q/AT (2)
R = (1 - [C.sub.p]/[C.sub.f]) x 100% (3)
Where J is permeate flux, kg/([m.sup.2] x h); A is membrane effective area, [m.sup.2]; T is time interval, h; R is salt rejection; [C.sub.f] and [C.sub.p] are salt concentrations in the feed side and in the permeate side, g/ L.
RESULTS AND DISCUSSION
Viscosity of Casting Solutions
Generally speaking, the viscosity of the casting solution is a very important parameter to the successful forming of a membrane [12, 15, 17, 22], it affects the coagulation kinetics of the casting solutions through impacting the interdiffusions of the non-solvent and the solvent. In semi-IPNs, because of the existence of crosslinking agent, the linear pre-crosslinking polymers can participate in the reactions of polymerization or crosslinking, which make it possible to create semi-IPNs structure. With the increase in crosslinking agent, the concentration of crosslinks increases , so the liquidity of casting solution decreases, and the viscosity of the casting solution increases accordingly .
From Fig. 3, with an increase in TEOS in the casting solution, the viscosity of casting solutions increased slowly. This could be attributed to the changing of additive: with the increasing of the additive, due to the polycondensation reaction as shown in Fig. 2, hydroxyl-terminated polydimethylsiloxane (PDMS) changed from flexible linear structure to the three-dimensional network structure, as a result the concentration of crosslinks increased and viscosity increased too. The change in the viscosity of the casting solutions would affect the demixing process, and further affected membrane morphology and VMD performance as a result.
Phase Inversion Kinetics of Casting Solution
It is well known that for the membrane making process of NIPS, the rheological variation (i.e., viscosity) and the thermodynamic variation (i.e., the interaction between polymer and non-solvent) influence both the de-mixing rate and the demixing way of both solvent and non-solvent in casting solutions , and when the viscosity of casting solution is low, the precipitation rate is fast [40, 41], and the lower the viscosity is, the faster the precipitation will be.
As we can see from Fig. 4, the precipitation of the casting solution demonstrated a delayed demixing process, and it included two stages. In the first stage, light transmittance slowly decreased; in the second stage, light transmittance first reduced rapidly and then decreased slowly to a steady state. Increasing TEOS dosage, the length of both stages increased. For example, when PDMS/TEOS was increased from 3:0 to 3:1, the length of the first stage prolonged from 104 s to 156 s; the time for light transmittance to reach a steady state, in the second stage, extended from 988 s to 1,092 s. This means that, with the adding of TEOS, precipitation rate decreased. The main reason for this was that, with the increase in TEOS dosage, the concentration of the crosslinks in the semi-IPNs system increased, so the viscosity of casting solution increased, as analyzed in the Viscosity of casting solutions section, which in turn caused the decreasing of the precipitation rate [40, 41]. Ultimately, the morphology of the membranes will be affected by phase inversion kinetics.
Membrane morphologies of the top surface, bottom surface and cross section were tested by SEM images as shown in Fig. 5.
For all membranes there were two layers which have different structures, the upper layer is finger-like and the bottom layer is sponge-like; the top surfaces is thin and dense; but the bottom surfaces had voids. As TEOS dosage in casting solutions increased, the finger-like macrovoids of the upper layer became shorter and narrower, and the proportion of sponge-like pores of the bottom layer increased. Based on the kinetics theory of precipitation as stated in the Phase inversion kinetics of casting solution section, high precipitation rate of the casing solution could result in the finger-like macrovoid structure whereas low precipitation rate of the casing solution could result in the sponge-like structure [17, 42]. In addition, all magnified pictures showed that there were loose cellular micropores in the sponge-like bottom layer, which is the most common phenomenon of liquid-liquid phase separation [36, 41], this kind of microstructure can increase the porosity of the membranes. According to the mechanism of macrovoid formation , dope solutions with lower viscosity are more favorable for the pore initiation than higher viscosity dope solutions. From M1 to M5, with the adding of TEOS, the viscosity of the casting solutions increased due to the formation of the semi-IPNs structure. As a result, for the bottom surfaces, from Ml to M5, pore size decreased and the surfaces became even and even. In addition, when the exposure was 40 s, a skin layer started to be formed, and with the extension of the exposure, the dense and smooth top membrane surface formed .
According to Table 3, with the gradually addition of TEOS, membrane thickness increased warmly. This may be caused by the changing of the viscosity of dope solutions showed in the Viscosity of casting solutions section: with the adding of TEOS, the semi-IPNs structure formed, and the semi-IPNs structure resulted in an increase in the viscosity of the casting solutions .
The distribution of elements F and Si on membrane surface was obvious from the EDX analysis shown in Table 2. The distribution would affect membrane hydrophobicity. As TEOS added, the ratio of F/Si increased on both the top surface and the bottom surface. This was associated with the content of the uncrosslinking linear molecular PDMS. With the adding of TEOS, more and more PDMS would take part in the crosslinking reactions, as shown in Fig. 2, so the concentration of the crosslinks increased, and then, during the phase inversion process, the generated semi-IPNs dragged PDMS chains into the inside of the membrane, as a result, the ratio of F/Si increased.
To verify the presence of PDMS and PVDF in PDMS-PVDF IPNs, ATR-FTIR spectroscopy is used widely in present studies  and so in our research. Figure 6 showed the FTIR spectra of (a) pure PVDF, (b) pure PDMS and (c) the resulting semi-IPNs PDMS-PVDF membranes. The peaks of pure PVDF and pure PDMS showed in Fig. 6a and b can also be found elsewhere [46, 47]. As exhibited in Fig 6c, all semi-IPNs membranes had similar spectra peaks at the waves ranging from 660 [cm.sup.-1] to 4,000 [cm.sup.-1]. The characteristic peaks at around 795-871 [cm.sup.-1] represented the vibration of the PVDF crystalline phase. The absorption peak at around 1,178 [cm.sup.-1] represented the stretching vibration of C[F.sub.2] in PVDF . The peak at 1,404 cnT1 was assigned to dissymmetry deformation vibration of the two methyls which linked with Si. The characteristic absorption peak at about 1,066 [cm.sup.-1] was corresponded to the Si-OH stretching , which means that PDMS had been incorporated and presented in PVDF matrix during the process of semi-IPN synthesis. Furthermore, with the keep adding of TEOS, the absorption peaks of Si--OH at about 1,066 [cm.sup.-1] became weakened for the reason that Si--OH had been consumed by the crosslinking reaction, which also means that the crosslinking reaction had been occurred.
Generally speaking, the crystallinity of a polymer is very crucial for the mechanical properties and the resistance of a membrane . As illustrated in Fig. 7, PVDF and all the resulting semi-IPNs PDMS-PVDF membranes had the similar crystalline structure and we can find the typical crystalline peaks of PVDF at about 2[theta] = 17.5[degrees], 20.5[degrees] and 27.5[degrees] , but the PDMS amorphous peak at around 2[theta] = 12[degrees]  was not found. It was possibly because that the PDMS cross linking chains might have evenly distributed among the PVDF polymer chains and the semi-IPNs structures did not change the crystal structure of PVDF, fortunately, this was good for the mechanical properties of the membranes.
Thermal Property Analysis
The TGA curves which showed the thermal stability of the resulting membranes are shown in Fig. 8. Generally, material thermal stability can be indicated by its initial decomposition temperature. In addition, according to literature , the thermal stability of hybrid materials could be decided by the temperature of 10% weight loss.
From Fig. 8, with the increase in TEOS content, the thermal stability of resulting membranes increased. For membranes Ml, M2, M3, M4 and M5, the initial decomposition temperature were around 196[degrees]C, 234[degrees]C, 296[degrees]C, 326[degrees]C and 340[degrees]C; the temperatures of 10% weight loss of them were around 363[degrees]C, 377[degrees]C, 368[degrees]C, 373[degrees]C and 380[degrees]C, respectively. For the semiIPNs, which were produced by the mutual entanglement of PVDF molecule and PDMS network, the thermal stability of the membranes was improved, and with the addition of TEOS, this improvement was particularly evident.
Membrane Pore Structure
It has been generally accepted that the VMD performances of the microporous membranes can be greatly affected by the structure of membrane pores [41-43, 52-55], Table 3 shows the porosity values of the Membranes and Fig. 9 shows the pore size distributions of the prepared membranes.
The uncrosslinked membrane (Ml) exhibited a pore size distribution of dominantly 19.45-235.00 nm. After the application of crosslinking agent TEOS, from M2 to M5, the pore size distribution of the membranes narrowed, and were 37.2-183 nm, 29.71-156.00 nm, 17.98-124.80 nm and 26.00-124.80 nm, respectively. This is because the crosslinking reaction in the semi-IPNs system caused a close structure . In addition, with the gradually adding of TEOS, from Ml to M5, the total porosity value increased linearly from 73.65% to 79.05%, and the mean pore radius declined as shown in Table 3, which was in agreement with the analysis of membrane morphology in the Membrane morphologies section. As shown in Fig. 5, the finger-like pores in the sublayer of the membranes were gradually replaced by the sponge-like pores.
The porous membranes used in the membrane distillation processes must be hydrophobic so that the vapor molecules can pass through the porous membrane and the bulk water cannot , for this reason the hydrophobicity of the membranes is a crucial factor for a MD process, therefore, it is important to study the hydrophobic property of membranes, and our results are listed in Table 4.
With the increase in TEOS, water contact angle of the top surface decreased from 78.18[degrees] to 73.38[degrees] and water contact angle of the bottom surface decreased from 128.85[degrees] to 103.33[degrees] due to the increase in F/Si on the two surfaces as shown in Table 2. According to literatures [50, 56], the hydrophobicity of PDMS is higher than that of PVDF, water contact angle is 100-110[degrees] for PDMS and is 70-90o for PVDF [12, 13].
Mechanical Property Test
Membrane mechanical properties have big effects on the real applications of the membranes. As a result, the tensile strength, elongation at break and young's modulus of the prepared membranes were measured and the testing results are listed in Table 5.
From M1 to M5, the tensile strength, elongation at break and young's modulus all showed a rising trend. This phenomenon might be caused by the changing of the membrane morphologies as shown in Fig. 5. According to the literatures [57-59], the sponge-like cross-section structure can improve the mechanical properties of the membrane, while the finger-like macrovoids may decrease the mechanical properties of the membrane seriously. With the increase in TEOS content, the proportions of the sponge-like pores increased. In addition, the semi-IPN structure worked as a bridging point and improved the actuation strains of the membranes and also increased the elongation at break .
In this study, using the 30 g/L NaCl aqueous solution, the VMD performances of the semi-IPNs PDMS-PVDF membranes were carried out at feed flow rate 80 L/h and permeate-side vacuum 77 KPa. As shown in Fig. 10a, the permeate fluxes increased exponentially with an increase in feed temperature, for example, for M5, permeate flux increased nearly four times as the feed temperature varied from 54[degrees]C to 80[degrees]C, and reached 21.52 kg/([m.sup.2] x h) at 80[degrees]C. In MD process, feed temperature is an important process parameter that can affect permeate flux significantly, so the effects of feed temperature on membrane properties has been studied widely in recent years [4, 58, 60, 61]. For MD process, the driving force is the vapor pressure difference, according to Antoine Equation, with the change of feed temperature, vapor pressure varies exponentially [59, 62, 63], with the increase in feed temperature, vapor pressure increases and the vapor pressure in the trans-membrane increased too. In addition, under the same temperature, the changing of flux was decided by the thickness, average pore radius, porosity (showed in Table 3) and morphologies of the membranes (showed in Fig. 5). As it is known, for the membranes, small thickness, high average pore radius and big porosity are favorable to permeate flux. With the increase in TEOS dosage, the fluxes of all the membranes increased and M5 had a highest flux for the reason that M5 had the largest porosity which generated from the semi-IPNs.
In addition, Fig. 10b presented that, for all the resulting membranes, the permeate conductivities were less than 20 [micro]s/ cm and salt rejections were higher than 99.9%; with an increase in feed temperature, the conductivity of the permeate declined due to the narrow pore size distribution as showed in Fig. 9.
To assess the VMD stability of the PDMS-PVDF membrane, PDMS-PVDF (M5) were tested for a 12 h continuous desalination. During the 12 h test of the membrane performance, distillate water was returned to the feed tank to maintain the concentration of the feed solution (NaCl aqueous) at 30 g/L, 80[degrees]C and 77 kPa. The dependences of VMD performances on operating time including flux and permeate conductivity are shown in Fig. 11. For M5, flux stabilized at about 20.952-1.54 kg/([m.sup.2] x h); The permeate conductivity was quite stable for the whole operating time; the rejections to salt were higher than 99.9% during the 12 h test. These results revealed that there was hardly any wetting during the operating for M5 and it was prospective to be employ in the VMD process.
Comparison with Other M D Membranes
Table 6 illustrates a comparison of our study with other investigations on MD membranes; the data were obtained in the present work or reported in the literature. Where [T.sub.f] is feed temperature, [T.sub.p] is permeate temperature and [P.sub.p] is the vacuum in the permeate side. The present membrane achieved a high salt rejection of 99.9% in addition to a very high flux of 21.5 kg/ ([m.sup.2] x h), which is compared with or even better than the performances of the reported PVDF membranes. The synthesized PDMS-PVDF microporous membrane with semi-IPN has superior performance, which is promising as an attractive membrane for the desalination of NaCl aqueous.
In this study, a novel semi-IPNs microporous membrane of PDMS-PVDF which had different dosage of TEOS was fabricated by NIPS process. The effects of TEOS dosage on membrane structures were investigated. In addition, membrane distillation experiments were carried out to test the performances of the resulting membranes.
When adding TEOS into casting solutions, polycondensation reaction occurred and the structure of PDMS changed from the flexible linear structure to the three-dimensional network structure, and the semi-IPNs structure appeared. As a result, with the adding of TEOS, the viscosity of casting solutions increased, for all membranes, the precipitation process presented a delayed demixing process; all membranes had a finger-like upper layer and a sponge-like bottom layer and the length of the finger-like pore structure became shorter and narrower; the thickness of membranes increased warmly; pore size distribution became narrower; porosity value increased; mean pore radius decreased; water contact angles of the top surfaces decreased from 78.18[degrees] to 73.38[degrees] and the bottom surfaces decreased from 128.85[degrees] to 103.33[degrees]; Young's modulus, tensile strength and elongation ratio increased and the biggest values of them for M5 were 45.73 MPa, 2.29 MPa and 126.06%, respectively. Moreover, TEOS content had obvious effects on membrane properties for the VMD separation of NaCl aqueous. When TEOS dosage in casting solution was increased, it had a big impact on membrane average pore radius, porosity and morphologies which in turn caused the increase in permeate flux, for all membranes, the salt rejection were all higher than 99.9%.
[1.] A. Alkhudhiri, N. Darwish, and N. Hilal, Desalination, 287, 2 (2012).
[2.] C.K. Chiam and R. Sarbatly, Chem. Eng. Process., 74, 27 (2013).
[3.] R.D. Ambashta and M.E.T. Sillanpaa, J. Environ. Radioact., 105, 76 (2012).
[4.] A.M. Alklaibi and N. Lior, Desalination, 171, 111 (2005).
[5.] F. Lagan, G. Barbieri, and E. Drioli, J. Membr. Sci., 166, 1 (2000).
[6.] P.P. Zolotarev, V.V. Ugrozov, I.B. Volkina, and V.M. Nikulin, J. Hazard. Mater., 37, 77 (1994).
[7.] F. Zaviska, P. Drogui, A. Grasmick, A. Azais, and M. Heran, J. Membr. Sci., 429, 121 (2013).
[8.] K.S. Bahceci, H.G. Akillioglu, and V. Gokmen, Innov. Food Sci. Emerg. Technol., 131, 31 (2015).
[9.] V.D. Alves and I.M. Coelhoso, J. Food Eng., 74, 125 (2006).
[10.] C. Charcosset, J. Food Eng., 92, 241 (2009).
[11.] F. Liu, N.A. Hashim, Y. Liu, M.R.M. Abed, and K. Li, J. Membr. Sci., 375, 1 (2011).
[12.] D. Hou, H. Fan, Q. Jiang, J. Wang, and X. Zhang, Sep. Purif. Technol., 135, 211 (2014).
[13.] G. Kang and Y. Cao, J. Membr. Sci., 463, 145 (2014).
[14.] A. Razmjou, E. Arifin, G. Dong, J. Mansouri, and V. Chen, J. Membr.Sci., 415-416, 850 (2012).
[15.] M. Tomaszewska, Desalination, 104, 1 (1996).
[16.] H. Fan, Y. Peng, Z. Li, P. Chen, Q. Jiang, and S. Wang, J. Polym. Res., 20, 1 (2013).
[17.] S. Simone, A. Figoli, A. Criscuoli, M.C. Carnevale, A. Rosselli, and E. Drioli, J. Membr. Sci., 364, 219 (2010).
[18.] M.S. Islam, N. Akter, and M.R. Karim, Colloids Surf. A: Physicochem. Eng. Aspects, 362, 117 (2010).
[19.] Y. Chen and H. Kim, Appl. Suif. Sci., 255, 7073 (2009).
[20.] S. Wang, Y. Li, X. Fei, M. Sun, C. Zhang, Y. Li, Q. Yang, and X. Hong, J. Colloid Interface Sci., 359, 380 (2011).
[21.] Y. Wu, Q. Huang, C. Xiao, K. Chen, X. Li, and N. Li, Desalination, 353, 118 (2014).
[22.] M.M. Teoh and T.S. Chung, Sep. Purif. Technol., 66, 229 (2009).
[23.] Z.Q. Dong, X. Ma, Z.L. Xu, W.T. You, and F. Li, Desalination, 347, 175 (2014).
[24.] X. Zhan, J. Li, J. Huang, and C. Chen, Appl. Biochem. Biotechnoi, 160, 632 (2010).
[25.] B. Suthar, H.X. Xiao, D. Klempner, and K.C. Frisch, Polym. Adv. Technol., 7, 221 (1996).
[26.] J.R. Nowers and B. Narasimhan, Polymer, 47, 1108 (2006).
[27.] X. Hou and K.S. Siow, Polymer, 42, 4181 (2001).
[28.] L. Chikh, V. Delhorbe, and O. Fichet, J. Membr. Sci., 368, 1 (2011).
[29.] E. Cznotka, S. Jeschke, P. Vettikuzha, and H.D. Wiemhofer, Solid State Ionics, 274, 55 (2015).
[30.] C. Cheng, Z. Yang, Y. He, A.N. Mondai, E. Bakangura, and T. Xu, J. Membr. Sci., 493, 645 (2015).
[31.] G. Soldani, P. Losi, M. Bernabei, S. Burchielli. D. Chiappino, S. Kull, E. Briganti, and D. Spiller, Biomaterials, 31, 2592 (2010).
[32.] M. Liu, Y.M. Wei, Z.L. Xu, R.Q. Guo, and L.B. Zhao, J. Membr. Sci., 437, 169 (2013).
[33.] A.J. Reuvers and C.A. Smolders, J. Membr. Sci., 34, 67 (1987).
[34.] C.Y. Kuo, H.N. Lin, H.A. Tsai, D.M. Wang, and J.Y. Lai, Desalination, 233, 40 (2008).
[35.] J.F. Li, Z.L. Xu, and H. Yang, Polym. Adv. Technol., 19, 251 (2008).
[36.] P.Y. Zhang, Y.L. Wang, Z.L. Xu, and H. Yang, Desalination, 278, 186 (2011).
[37.] D. Sun, M.Q. Liu, J.H. Guo, J.Y. Zhang, B.B. Li, and D.Y. Li, Desalination, 370, 63 (2015).
[38.] M. Palencia, M. Vera, and B.L. Rivas, J. Membr. Sci., 466, 192 (2014).
[39.] D. Li, F. Xu, Z. Liu, J. Zhu, Q. Zhang, and L. Shao, Appl. Surf. Sci., 266, 368 (2013).
[40.] D.M. Wang and J.Y. Lai, Curr. Opin. Chem. Eng., 2,229 (2013).
[41.] S. Devi, P. Ray, K. Singh, and P.S. Singh, Desalination, 346, 9 (2014).
[42.] L.B. Zhao, Z.L. Xu, M. Liu, and Y.M. Wei, J. Membr. Sci., 454, 184 (2014).
[43.] A. Figoli, S. Simone, A. Criscuoli, S.A. AL-Jlil, F.S. Al Shabouna, H.S. Al-Romaih, E. Di Nicolo, O.A. Al-Harbi, and E. Drioli, Polymer, 55, 1296 (2014).
[44.] Q. Li, Z.L. Xu, and L.Y. Yu, J. Appl. Polym. Sci., 115, 2277 (2010).
[45.] H. Hillborg and U.W. Gedde, IEEE Trans. Dielectr. Electr. Insul., 6, 703 (1999).
[46.] R.D. Simoes, M.A. Rodriguez-Perez, J.A. De Saja, and C.J.L. Constantino, Polym. Eng. Sci., 49, 2150 (2009).
[47.] H. Sertchook, H. Elimelech, and D. Avnir, Chem. Mater., 17, 4711 (2005).
[48.] N.F.C. Pa, I. Ahmed, M.G.M. Nawawi, and W.A.W.A. Rahman, Sep. Sci. Technol., 47, 562 (2012).
[49.] P.Y. Zhang, H. Yang, Z.L. Xu, Y.M. Wei, J.L. Guo, and D.G. Chen, J. Polym. Res., 66, 1 (2013).
[50.] D. Sun, B.B. Li, and Z.L. Xu, Desalination, 322. 159 (2013).
[51.] D. Yang, S. Yang, Z. Jiang, S. Yu, J. Zhang, F. Pan, X. Cao, B. Wang, and J. Yang, J. Membr. Sci., 487, 152 (2015).
[52.] M. Essalhi and M. Khayet, J. Membr. Sci., 454, 133 (2014).
[53.] E. Drioli, A. Ali, S. Simone, F. Macedonio, S.A. AL-Jlil, F.S.A. Shabonah, H.S. Al-Romaih, O. Al-Harbi, A. Figoli, and A. Criscuoli, Sep. Purif. Technol., 115, 27 (2013).
[54.] M.M. Teoh, T.S. Chung, and Y.S. Yeo, Chem. Eng. J., 171, 684 (2011).
[55.] C. Feng, B. Shi, G. Li, and Y. Wu, J. Membr. Sci., 237, 15 (2004).
[56.] M.T. Khorasani and H. Mirzadeh, J. Appl. Polym. Sci., 91, 2042 (2004).
[57.] P. van de Witte, P.J. Dijkstra, J.W.A. van den Berg, and J. Feijen, J. Membr. Sci., 117, 1 (1996).
[58.] J. Zhang, J.D. Li, M. Duke, Z. Xie, and S. Gray, J. Membr. Sci., 362, 517 (2010).
[59.] J. Zhang, N. Dow, M. Duke, E. Ostarcevic, J.D. Li, and S. Gray, J. Membr. Sci., 349, 295 (2010).
[60.] M. Khayet, M.P. Godino, and J.I. Mengual, J. Non-Equilib. Therm., 26, 1 (2001).
[61.] M. Khayet, P. Godino, and J.I. Mengual, J. Membr. Sci., 165, 261 (2000).
[62.] M.N. Chemyshov, G.W. Meindersma, and A.B. de Haan, Desalination, 157, 315 (2003).
[63.] T.Y. Cath, V.D. Adams, and A.E. Childress, J. Membr. Sci., 228, 5 (2004).
[64.] Z.L. Chen, D. Rana, T. Matsuura, Y.F. Yang, and C.Q. Lan, Sep. Purif Technol., 133, 303 (2014).
[65.] D.E. Suk, T. Matsuura, H.B. Park, and Y.M. Lee, Desalination, 261, 300 (2010).
[66.] Z. Chen, D. Rana, T. Matsuura, D. Meng, and C.Q. Lan, Chem. Eng. J., 276, 174 (2015).
[67.] J.E. Efome, M. Baghbanzadeh, D. Rana, T. Matsuura, and C.Q. Lan, Desalination, 373, 47 (2015).
[68.] H. Abdallah, A.F. Moustafa, A.A.H. AlAnezi, and H.E.M. El-Sayed, Desalination, 346, 30 (2014).
[69.] Y. Liao, R. Wang, M. Tian, C.Q. Qiu, and A.G. Fane, J. Membr. Sci., 425, 30 (2013).
[70.] Y. Liao, C.H. Loh, R. Wang, and A.G. Fane, ACS Appl. Mater. Inteifaces, 6, 16035 (2014).
[71.] J.A. Prince, D. Rana, G. Singh, T. Matsuura, T.J. Kai, and T.S. Shanmugasundaram, Chem. Eng. J., 242, 387 (2014).
[72.] J.A. Prince, G. Singh, D. Rana, T. Matsuura, V. Anbharasi, and T.S. Shanmugasundaram, J. Membr. Sci., 397-398, 80 (2012).
[73.] Y. Liao, R. Wang, and A.G. Fane, Environ. Sci. Technol., 48, 6335 (2014).
De Sun, (1) Zhao-Shan Zheng, (1) Mei-Qin Liu, (1) Bing-Bing Li, (1) Fei Huang, (1) Da-Yong Li (2)
(1) Department of Chemical Engineering, Changchun University of Technology, 2055 Yanan Street, Changchun 130012, People's Republic of China
(2) COFCO Bio-Chemical Energy (Yushu) Co, Ltd. Economic Development Wukeshu, 1 Dongfeng Street, Changchun 130033, People's Republic of China
Correspondence to: D. Sun; e-mail: firstname.lastname@example.org Contract grant sponsor: Jilin Provincial Science & Technology Department; contract grant number: 20150204077GX; contract grant sponsor: The Foundation of Key Programs for Science and Technology of Changchun; contract grant number: 2013058; contract grant sponsor: Science & Technology Research Program of Changchun University of Technology; contract grant number: 2012010.
Published online in Wiley Online Library (wileyonlinelibrary.com).
Caption: FIG. 1. Schematic diagram of PDMS-PVDF semi-IPNs.
Caption: FIG. 2. PDMS cross linking reaction.
Caption: FIG. 3. Viscosity of casting solutions with TEOS dosage.
Caption: FIG. 4. Precipitation rates of casting solutions with TEOS dosage.
Caption: FIG. 5. SEM images of the resulting membranes, (a) (bottom surface) magnification 2,000X, (b) (top surface) magnification 2,000X, (c) (cross-section) magnification 350X, (d) (cross-section) magnification 2,000X.
Caption: FIG. 6. FTIR spectra of (a) pure PVDF, (b) pure PDMS and (c) the resulting membrane.
Caption: FIG. 7. XRD patterns of the resulting membrane with TEOS dosage.
Caption: FIG. 8. TGA curves of the resulting membrane with TEOS dosage.
Caption: FIG. 9. Pore size distribution of membranes.
Caption: FIG. 10. VMD performances for membranes with TEOS dosage ([P.sub.v] = 77 kPa, [C.sub.f] = 30 g/L and [Q.sub.c] = 80 L/h).
Caption: FIG. 11. VMD performances of PDMS-PVDF membrane (M5) with operating time.
TABLE 1. Preparation conditions of the semi-IPNs PDMS-PVDF membranes. PVDF/PVP/DMAc/ PVDF/PDMS THF/PDMS TEP (mass ratio) (mass ratio) (mass ratio) M1 15/1.5/33.5/50 M2 M3 30:1 10/1 M4 M5 PDMS/TEOS/DBTDL Evaporation Coagulation (mass ratio) time (s) bath M1 3/0/0 M2 3/0.2/0.1 M3 3/0.3/0.15 40s Deionized water M4 3/0.6/0.3 M5 3/1/0.5 TABLE 2. Concentration distribution of F and Si elements on the top(T) and the bottom(B) surfaces of the membranes. Membrane Element [F.sub.T] wt% Element [F.sub.B] wt% M1 92.77 [+ or -] 0.35 93.85 [+ or -] 0.43 M2 93.46 [+ or -] 0.42 94.01 [+ or -] 0.46 M3 94.53 [+ or -] 0.43 94.82 [+ or -] 0.48 M4 96.88 [+ or -] 0.51 96.75 [+ or -] 0.53 M5 97.00 [+ or -] 0.62 96.75 [+ or -] 0.54 Membrane Element [Si.sub.T] wt% Element [Si.sub.B] wt% M1 7.23 [+ or -] 0.05 6.15 [+ or -] 0.04 M2 6.54 [+ or -] 0.04 5.99 [+ or -] 0.04 M3 5.47 [+ or -] 0.02 5.18 [+ or -] 0.03 M4 3.12 [+ or -] 0.01 3.25 [+ or -] 0.02 M5 3.00 [+ or -] 0.01 3.25 [+ or -] 0.01 Membrane Top F/Si Bottom F/Si M1 12.83 [+ or -] 0.13 15.26 [+ or -] 0.17 M2 14.29 [+ or -] 0.15 15.69 [+ or -] 0.19 M3 17.28 [+ or -] 0.14 18.31 [+ or -] 0.19 M4 31.05 [+ or -] 0.27 29.77 [+ or -] 0.35 M5 32.33 [+ or -] 0.32 29.77 [+ or -] 0.26 TABLE 3. Membrane thickness ([delta]), mean pore size ([r.sub.m]) and porosity ([epsilon]) of the resulting membranes. Membrane [delta] [r.sub.m] [epsilon] ([micro]m) ([micro]m) (%) M1 167 [+ or -] 3 0.062 [+ or -] 0.008 73.65 [+ or -] 2.41 M2 175 [+ or -] 2 0.060 [+ or -] 0.007 76.34 [+ or -] 3.11 M3 175 [+ or -] 2 0.043 [+ or -] 0.005 78.22 [+ or -] 3.24 M4 175 [+ or -] 2 0.026 [+ or -] 0.004 78.88 [+ or -] 3.32 M5 191 [+ or -] 3 0.037 [+ or -] 0.006 79.05 [+ or -] 3.86 TABLE 4. Water contact angles of membranes. Contact angle of Contact angle of Membrane top surface ([degrees]) bottom surface ([degrees]) M1 78.18 [+ or -] 1.93 128.85 [+ or -] 3.11 M2 75.96 [+ or -] 1.32 108.44 [+ or -] 2.35 M3 75.60 [+ or -] 1.21 105.57 [+ or -] 2.23 M4 74.72 [+ or -] 0.98 104.85 [+ or -] 2.11 M5 73.38 [+ or -] 0.52 103.33 [+ or -] 2.03 TABLE 5. Mechanical properties of membranes. Breaking Elongation at Membrane strength (MPa) break (%) M1 1.78 [+ or -] 0.12 85.70 [+ or -] 2.6 M2 1.81 [+ or -] 0.17 91.83 [+ or -] 3.3 M3 1.84 [+ or -] 0.16 90.62 [+ or -] 3.2 M4 2.05 [+ or -] 0.21 94.56 [+ or -] 3.7 M5 2.29 [+ or -] 0.25 126.06 [+ or -] 3.8 Young's modulus Membrane (MPa) M1 28.76 [+ or -] 2.53 M2 29.28 [+ or -] 2.68 M3 31.63 [+ or -] 3.35 M4 32.12 [+ or -] 3.64 M5 45.73 [+ or -] 3.75 TABLE 6. Comparison of experimental data obtained in this work with the membrane performances of various membranes reported in the literatures. Membrane type MD type Refs. PVDF membrane supported on VMD  non-woven polyester fabric PVDF membranes with high and VMD  low molecular weights SMMa/PES VMD  PDMS/PVDF VMD  membrane with intermediate VMD  molecular weight of PVDF PVDF/SiO2 flat sheet VMD [671 composite membranes PESATNTs VMD  Electrospun PVDF DCMD  nanofiber membranes Silica-PVDF/PVDF dual-layer DCMD  membranes PVDF-PVP-hydrophobic SMM DCMD  composite membranes PVDF-Clay nanofiber membranes DCMD  Si[O.sub.2]-PVDF Flat DCMD  sheet/nanofiber PDMS-PVDF microporous VMD This work membrane with semi-IPN Membrane type Solution PVDF membrane supported on [T.sub.f] = 75[degrees]C; non-woven polyester fabric 30,000 ppm NaCl PVDF membranes with high and [T.sub.f] = 27[degrees]C; low molecular weights [P.sub.P] = 98 kPa; deionized water as feed SMMa/PES [T.sub.f] = 26[degrees]C; [P.sub.P] = 400 Pa: distilled water as feed PDMS/PVDF [T.sub.f] = 50[degrees]C; [P.sub.P] = 80 kPa;20 g/L NaCl membrane with intermediate [T.sub.f] = 50[degrees]C; molecular weight of PVDF [P.sub.P] = 98 kPa; deionized water as feed PVDF/SiO2 flat sheet [T.sub.f] = 27[degrees]C; composite membranes [P.sub.P] = 94.8 kPa; 35g/L NaCl PESATNTs [T.sub.f] = 65[degrees]C; [I.sub.P] = 30 kPa;7,000 ppm NaCl Electrospun PVDF [T.sub.f] = 60[degrees]C; nanofiber membranes [T.sub.P] = 20[degrees]C; 35g/L NaCl; Silica-PVDF/PVDF dual-layer [T.sub.f] = 60[degrees]C; membranes [T.sub.P] = 20[degrees]C; 35g/L NaCl PVDF-PVP-hydrophobic SMM [T.sub.f] = 80[degrees]C; composite membranes [T.sub.P] = 20[degrees]C; 35g/L NaCl PVDF-Clay nanofiber membranes [T.sub.f] = 80[degrees]C; [T.sub.P] = 17[degrees]C; 3.5 wt% NaCl solution Si[O.sub.2]-PVDF Flat [T.sub.f] = 60[degrees]C; sheet/nanofiber [T.sub.P] = 20[degrees]C; 35g/L NaCl PDMS-PVDF microporous [T.sub.f] = 80[degrees]C; membrane with semi-IPN [P.sub.p] = 77 kPa;30g/L NaCl Flux kg/ Salt Membrane type ([m.sup.2] * h) rejection (%) PVDF membrane supported on 12.5 >99 non-woven polyester fabric PVDF membranes with high and 0.7 -- low molecular weights SMMa/PES 1.9 -- PDMS/PVDF 15.4 >99.9 membrane with intermediate 0.3 -- molecular weight of PVDF PVDF/SiO2 flat sheet 2.9 >99.98 composite membranes PESATNTs 15 >98 Electrospun PVDF 20.6 -- nanofiber membranes Silica-PVDF/PVDF dual-layer 21 99.99 membranes PVDF-PVP-hydrophobic SMM 14 <99.98 composite membranes PVDF-Clay nanofiber membranes 5.7 <99.99 Si[O.sub.2]-PVDF Flat 18.9 -- sheet/nanofiber PDMS-PVDF microporous 21.5 >99.99 membrane with semi-IPN SMM, surface modifying macromolecules.
|Printer friendly Cite/link Email Feedback|
|Author:||Sun, De; Zheng, Zhao-Shan; Liu, Mei-Qin; Li, Bing-Bing; Huang, Fei; Li, Da-Yong|
|Publication:||Polymer Engineering and Science|
|Date:||Dec 1, 2017|
|Previous Article:||Synergistic Effects of Silica Nanoparticles and Reactive Compatibilizer on the Compatibilization of Polystyrene/ Polyamide 6 Blends.|
|Next Article:||Preparation of a New Filament Based on Polyamide-6 for Three-Dimensional Printing.|