Shadi Hasanajili*,Masoud Latifzadeh,Mahmoud Bahmani

Chemical and Petroleum Engineering School,Shiraz University,71348-51154 Shiraz,Iran

1.Introduction

Carbon dioxide is one of the major impurity in natural gas which is also known as greenhouse gas.Furthermore,CO2causes pipe line corrosion and humanity illness so its removal from natural gas is necessary[1].There have been a variety of methods for separation of CO2from naturalgas component such as methane.The most commonly used methods are thermal distillation,adsorption by selective liquid and membrane technology.Among these methods,membrane technology is the most attractive one from novelty,energy consumption,minimumrequired space and minimum environmental pollution points of view[2,3].In the lastdecades,extensive studies have been conducted in the area of membrane separation.Among various types of membranes,polymeric ones have some advantages such as ease of processing and suitable strength for natural gas sweetening[4].The most commonly used commercial polymers for gas separation are the phenyl groups based polysulfone(PSf),polyethersulfone(PES)and aromatic polyimide(PI).PSf and PES are high performance engineering polymers which have good stability,permeability,selectivity,high critical pressure of plasticization and low cost[5-9].The main properties of membranes such as hydrophilicity,mechanical properties,pore size and membrane morphology could be changed by applying polymer mixture[10].Polyurethanes have many applications because of their good mechanical properties and chemical resistance,such as tensile strength,abrasion,oil resistance and long fatigue life.Furthermore,PU-based blend membranes improve gas selectivity,mechanical and thermal properties of the membranes[11-14].The capability of polyurethane(PU)and polyurethane-poly(methyl methacrylate)(PMMA)blend membranes were examined by Windm?ller et al.The effects ofblend composition,temperature and pressure were investigated on the permeability,diffusivity and solubility of CO2,H2,O2,CH4,and N2.They showed that the permeability of all gases decreases by approximately 55%with the addition of 30 wt%of PMMA[15].In 2012,Nam et al.used charged membranes prepared via blending SPAES and PES to investigate the effect of the blended polymers on water permeability as well as salt rejection[16].A successful study on fabrication of high performance blend membranes suitable for separation and purification of hydrogen was conducted by Chung et al.The effect of key parameters such as composition on miscibility and microstructure,gas permeability and selectivity of blend membranes has been investigated[17].Lin et al.desulphurized gasoline using polyethyleneglycol(PEG)/polyurethane(PU)blend membranes.They revealed that PU modifies PEG membrane and improves its performance for the separation process[18].

Recently,different methods were used to optimize and model the membrane process.However,there are many responsive variables in the field ofgas separation and membrane technologies such as pressure,applied polymer type and composition.One of the simplest and also concise methods for modeling these processes is response surface methodology(RSM).RSM is the regression method exploring the relationships between several explanatory variables and one or more response variables which was introduced by Box and Wilson in 1995.This method is important to understand the interaction effects between factors and to reduce the total number of experimental runs.The essential of RSM is for replacing a complex model by an approximate one based on results calculated at various points in the design space.RSM can thus be used to diminish the cost of function evaluation in structural optimization.The optimization is then performed at a lower cost over such response surfaces.Therefore,RSM is especially fit for long time computation consuming problems[19-21].There have been a few studies using RSM for optimizing membrane fabrication in the literature.Among them,the RSM was applied by Ahmad et al.to optimize membrane performance with thermal mechanical stretching.They investigated stretching elongation,the stretching rate and the stretching temperature as variables for their optimization purpose[22].The preparation of defect-free asymmetric membranes through applying membrane fabrication variables such as polymer concentration,solvent ratio,forced-convective evaporation time and casting shear rate were studied by Ismail and Lai using RSM.The significant and interaction effects of the mentioned variables were checked on membrane structure and performance by using factorial design and response surface methodology in order to optimize membrane formation process[23].Razali et al.used the central composite design(CCD)of the response surface method to optimize blended PES/PANI membranes.The PES concentration,PANI concentration and evaporation time during the casting process were considered as operating variables,while pure water permeability,salt rejection and contact angle values were considered as the responses[24].Another research in the optimization of the blend membranes by RSM was done by Xiangli et al.They optimized polydimethylsiloxane/ceramic composite pervaporation membranes by considering polymer concentration,crosslinking agent concentration and dip-coating time as the operating variables[25].

In this study,polymeric membranes based on polyether sulfonepolyurethane blends were made by immersion precipitation method,furthermore permeance of carbon dioxide and methane gases were studied.The properties of polyethersulfone membranes in the presence of two types of polyetherurethane and polyesterurethane were investigated by scanning electron microscopy(SEM),Fourier transform infrared spectroscopy(FTIR),thermal gravimetric analysis(TGA),and differential scanning calorimetry(DSC).The effect of key parameters such as feed pressure and different types of PU(ESPU&ETPU)on performance of blend polymeric membrane was investigated.A great effort is also made to model the key parameter of this study.In this regards,response surface methodology(RSM)is applied to optimize parameters such as pressure,PU type and composition.

2.Materials and Methods

2.1.Experimental

2.1.1.Materials

Polyethersulfone(PES Ultrason E6020P with MW=58000 g·mol-1)was purchased from BASF Company(Germany),polyesterurethane and polyethereurethane provided by CoimS.p.A Co.(LPR9025),N-methylpyrrolidone(NMP)was procured from Merck to use as solvent for PES.CO2gas(purity 99.9%,MD 3.3×10-10m)was purchased from Aboughadare Co.(Shiraz,Iran)and CH4gas(purity 99.995%,MD 3.8×10-10m)was bought from Air Products Co.

2.1.2.Preparation of membranes

Various polymer blend solutions were prepared with PES/ESPU and PES/ETPU by the mass ratios of 100/0,98.5/1.5,96.5/3.5,95/5,93/7 and concentration of 30 wt%in NMP.After stirring the mixture overnight at 75°C and ultrasonic mixing for about 5 min,the polymer solution was then casted onto a glass surface with a calculated amount of blend solution to give the desired thickness.After casting,the film was keptin distilled water bath for 24 h.In order to dry prepared membranes,they were kept in about 24 h at ambient temperature and then put them at 80°C for 24 h in vacuum oven to remove the residual solvent.

2.1.3.Membrane characterization

2.1.3.1.Differential scanningcalorimetry(DSC).Thermal properties of the film were measured by differential scanning calorimetry(DSC)using a Mettler-Toledo DSC 822e at a heating rate of 10 °C·min-1and the temperature range of 30-700°C.

2.1.3.2.Contact angle test(CA).To determine the hydrophilic characterization of the prepared membranes their contact angles were measured using a DSA100 made by KRUSS Co.of Germany.

2.1.3.3.Scanning electron microscopy(SEM).The morphology of the membranes and their surfaces were scanned using scanning electron microscope(SEM).The samples were fractured in liquid nitrogen,and then were coated with gold and tested by a Philips XL30(Philips,Netherlands)scanning electron microscope.

2.1.3.4.Fourier transform infrared spectroscopy(FTIR).FTIR spectrometer was employed for monitoring and determining the final chemical structure of the polymer membranes and also studying the consistency ofthe blend polymer.Infrared spectra for samples were collected in a Fourier transform infrared spectrophotometer(Bruker,model Equinox 55).

2.1.3.5.Porosity.The porosity of the membrane was measured by liquid displacement[26].Waterwas chosen as the displacing liquid.The membrane was immersed in a known volume(V1)of water in a graduated cylinder.The total volume of water and also water impregnated membrane was recorded as V2.The impregnated membrane was then removed from the cylinder and the residual volume was recorded as V3.The porosity of the scaffold was calculated by Eq.(1):

2.1.4.Gas permeation measurement

The permeation of pure gas through PES,PES-ESPU and PES-ETPU membranes was conducted using a constant pressure(p=0.6,0.8&1.0 MPa)method at constant temperature of 29°C applying the experimental set up shown in Fig.1.The area of the used membrane was 12.5 cm2and also the permeate side was kept at ambient pressure.The gas permeability was evaluated as following:

where P is the permeability expressed in Barrer(1 Barrer=10-9mol·m-2·s-1·Pa-1),Q is the flow rate of the permeate gas passing through the membrane(cm3·s-1),l is the membrane thickness(cm),Δp(Pa)is the trans membrane pressure difference(p1and p2are the absolute pressures of the feed side and permeate side,respectively)and A is the effective membrane area(cm2)[27].

A solution-diffusion model was applied to analyze the permeation of the gas through used membranes.The ideal gas selectivity is another significant parameter which can be calculated from pure gas permeation experiments as follows:[28,29].

where PAand PBare the gas permeabilities ofAand B gases,respectively.

Fig.1.Schematic diagram of experimental setup for gas permeation test.

2.1.5.Response surface methodology(RSM)

RSM was applied to evaluate the effects of pressure(A),composition(%)(B)and PU type(C)on the permeation and selectivity of PES,PES-ESPUand PES-ETPU membranes.The coded and uncoded independent variables used in the RSM design and their respective levels were listed in Table 1.This method provides a possible way to obtain a suitable polynomial which is applicable for experimental design.The second order models are so flexible and can cover a wide range of experimental data.Therefore,they usually can be considered as an approximation for surface response.In RSM,a model was defined for each dependentvariable which is determined the significant effects and interaction factors for each variable individually.In this study,a second-order polynomial equation was used to express the permeability and selectivity as function of the independent variables as follows:

Table 1 Codes,ranges and levels of independent variables ofpressure(A),composition(B)and PU type(C)in RSM design

Fig.2.FTIR spectra of the membranes prepared from PES,PES-ETPU and PES-ESPU.

where Y represents the response variable,β0is a constant,βi,βiiand βijare the linear,quadratic and interactive coefficients,respectively.The model was built based on the variables with correlation coefficients of 0.97 and 0.95 for permeability and selectivity respectively.

3.Results and Discussion

3.1.Membrane characterization

3.1.1.Fourier transform infrared spectroscopy(FTIR)

Fig.3.DSC thermograms of the membranes prepared from PES,PES-ETPU and PES-ESPU.

Fig.4.Thermal analysis of PES,PES-ETPU and PES-ESPU blend membranes.

Chemical characterization and functional groups of PES and PES-PU blend membranes were determined using FTIR.FTIR spectra of PES,PES-ESPU and PES-ETPU membranes with 7 wt%PU are presented in Fig.2.The PES chain includes an ether bond,a benzene ring and a sulfone(SO2)structure.The symmetrical and asymmetrical stretching vibrations of the O=S=O group for pure PES and PES blend membranes observed near 1151 and 1321 cm-1,respectively.The C--O--C stretching peaks were located at 1324 cm-1and 1239 cm-1.Three peaks between 1400 and 1600 cm-1were assigned to aromatic skeletal vibration[4,30].As seen in Fig.2,an additional peak appeared at 1728 cm-1in PES-ESPU sample which is related to the(C=O)group of polyesterurethane[31].The band observed at 1114 cm-1in PESETPUs pectrum was attributed to the antisymmetric stretching vibrations of ether(C--O--C)group in polyetherurethane[32].Compared to FTIR spectrum of pure PES,the spectra of blend membranes indicated a number of PES bands(i.e.,sulfone one)shifted to higher wavenumbers which might be due to intermolecular interactions between PES and PU[4].

3.1.2.Differential scanning calorimetry(DSC)

The miscibility of the PESand PUs wasalso studied by a DSC analysis.The DSC ther mograms for pure PES,PES-ESPU(7 wt%ESPU)and PESETPU(7 wt%ETPU)are indicated in Fig.3.The existence of the single Tgfor the PES-PUs reveals that the dispersed phase(polyurethane)is completely dissolved in the continues phase.The transition glass temperature for two different PUs is much less than PES which are 245 °C for pure PES,-39.94 °C and-39.99 °C for ESPU and ETPU respectively[33].Therefore,the PES-PU blends containing 7 wt%PU show lower Tgwith respect to pure PES.The transition glass temperature for PES-ESPU and PES-ETPU blends is 238 °C and 229 °C,respectively.The lower Tgdemonstrated the fact that less thermal energy is needed to overcome chain-chain interactions of the polymer,because the chain mobility of the PES increases in the presence of PU[34].

Thermal stability of pure PES and PES-ESPU and PES-ETPU blends containing 7 wt%PU were evaluated using thermal gravimetric analysis(Fig.4).Results showed that thermal decomposition of PES started at 500 °C and continued up to 600 °C.For PES-ESPU and PES-ESPU TGA curves,there were two decomposition stages in the range of 240-450 °C and 500-600 °C due to the degradation of PU and PES,respectively.As shown in these curves two different slopes of mass reduction were observed in the first region.These two different slopes reveal that there are two different mass reductions in PU.The first one is related to the breakage in urethane bonds,while the second step is related to the thermal decomposition of polyol[35].As it is clearfrom TGA curves the thermalstability ofPES decreases in the presence of PU,butit is high enough for gas separation purpose by membrane.

3.1.3.Contact angle

The results of water drop contact angle test for pure PES,PES-ESPU(7%ESPU)and PES-ETPU(7%ETPU)dense sheets are shown in Fig.5.As it can be seen,the presence of PU water contact angle decreases and an enhancement in hydrophilic membrane characterization due to hydrophilic nature of PU chains in comparison with PES is observed[36,37].The water contact angle and standard deviation for pure PES and blend membranes are mentioned in Table 2.

Table 2 Contact angles for pure PES and PES/ESPU(7%ESPU)and PES/ETPU(7%ETPU)

3.1.4.Scanning electron microscopy(SEM)

Fig.5.Contact angle results of(a)pure PES,(b)PES-ESPU(7%ESPU),(c)PES-ETPU(7%ETPU).

Fig.6.Cross section SEM of pure PES.

The morphology of the membrane is one of the significant properties which strongly influences on the gas transport[38].Figs.6-8 show the SEM images from top surface and cross-section of PES,PES-ESPU and PES-ETPU membranes.Cross-section images show that the prepared membranes have asymmetric structure with dense skin layer at the top of the membrane and a finger-like structure at porous sub-layer.As it is clear from the mentioned figures,adding different mass percentages of ESPU and ETPU to PES altered the membrane structure completely.It is believed that the final morphology of membranes is governed by a trade-off between thermodynamic enhancement and kinetic hindrance in the presence of a specific polymer additive[39].It is accepted as a general rule that thermodynamic enhancement leads to the formation of more porous structure,whereas kinetic hindrance delays the demixing and subsequently favors the formation of denser structures.The use of PU in the blend composition increases the viscosity of the casting solution[37]and decreases the exchange rate during the precipitation processes;hence,it suppressed the formation of macrovoids and allowed a dense skin layer to form.The addition of PU increases the hydrophilic ity of the casting solution which results in a higher inversion rate.According to Figs.7&8,it seems that at a concentration of 1.5 wt%PU,the viscosity effect was the dominant factor,because a relatively denser sponge sub-layer could be observed in comparison with the pure PES(see Table 3),while in the presence of higher concentration of 1.5 wt%,the thermodynamic instability effect was the dominant factor,which is a favored formation of membranes with larger finger-like pores.Consequently,asymmetric membranes with a thin skin layer supported by more finger type substructure were formed[38,40].When demixing is fast,the solidification occurs before the polymer molecules are fully relaxed,and the skin of top layer fractures.Therefore,the non-solvent intrudes rapidly into the cast film through these defects[41].It should be noted that,the PUs did not produce a defect free(selective)membrane at concentration higher than 1.5 wt%,so we had to set the highest concentration of the PUs at 1.5 wt%.

Fig.7.Cross section SEM of PES-ESPU with concentration of(a)1.5 wt%,(b)3.5 wt%,(c)5 wt%and(d)7 wt%.

Fig.8.Cross section SEM of PES-ETPU with concentration of(a)1.5 wt%,(b)3.5 wt%,(c)5 wt%and(d)7 wt%.

Table 3 The values of thickness of the dense top layer

Comparing the morphology of PES-ESPU and PES-ETPU samples depicted that in the presence of ETPU,membranes with larger macrovoids than those of ESPU were fabricated which can be explained by this fact that ETPU has more hydrophilicty in comparison with ESPU.Based on CA results,ETPU makes the membrane more hydrophilic than ESPU.This causes higher flux of water towards the casting solution of PES-ETPU in comparison with PES-ESPU;hence,membranes with larger finger-like pores are formed in the presence of ETPU.The thickness of created top film layer for various PUs contents is listed in Table 3.Results revealed that by increasing PU/PES ratio in the blend membranes,the thickness of top dense layer decreased.As mentioned above,polymer blend solutions showed a relatively higher viscosity than PES casting solution and their viscosity raises with PU content;however,huge thermodynamic enhancement by adding PU and increasing its percentage resulted in forming a thinner top layer in blends membranes with respect to PES.

3.2.Effect of pressure on gas permeability for PES,PES-ESPU and PES-ETPU membranes

The permeability pro files of PES,PES-ESPU and PES-ETPU membranes for CO2and CH4gases versus pressure are sketched in Figs.9 and 10.As it is obvious,the permeability of CO2is higher than that of CH4for all used membranes and pressures due to the kinetic diameter and condensability of gases,respectively[42].Due to the lower kinetic diameter and higher condensability of CO2,in comparison with CH4,CO2exhibits the higher permeability within the membranes.As seen in these figures,the permeability of gases increases by pressure increment because more gas solubility occurs at high pressures.Based on the figures,increasing the extent of PUs in the membranes causes reduction of top layer thickness and development of more finger pores in the membrane structure.Therefore the permeability of both gases increases at all pressures.Moreover,by having a conscious look at the mentioned figures,the permeability of PES-ETPU membrane is higher than that of PES-ESPU due to its larger pores and less thickness of top layer which are seen in SEM images.

Fig.9.Effect of pressure and ESPU composition on CO2 and CH4 permeance.

Fig.10.Effect of pressure and ETPU composition on CO2 and CH4 permeance.

3.3.Effect of pressure and composition on ideal selectivity of CO2/CH4

The effect of pressure and various compositions of PES/PU on ideal selectivity of CO2/CH4is sketched in Figs.11 and 12.According to pro files,by increasing pressure and composition,ideal selectivity of CO2/CH4is decreased.In order to show the reasons,we should consider each effect separately.For the pressure effect,it should be noted that pressure increment results in membrane softening and also reduction of free volume for molecular diffusion[39,40].According to higher chemical affinity of CO2for polar polymers e.g.PES than methane,more reduction in free volume by compaction may occur in the presence of CO2molecules which leads to lower selectivity of CO2over CH4with pressure in all membranes.The change in selectivity by increasing PU/PES ratio in the blends could be justified by SEM images.According to SEM images,increasing PU extent from 3 wt%up to 7 wt%in PES/PU membranes,the thickness of selective layer reduces and the size and number of finger-like voids increase.Consequently,more methane permeability is observed with pressure which leads to reduction in ideal selectivity of CO2/CH4[27].In the case of PES/ESPU membrane with the blend composition of 98.5/1.5,high feed pressure had no significant effect on CO2/CH4selectivity due to the reduction of porosity in the presence of ESPU.In order to compare two used PUs in terms of their effects on the selectivity,for PES-ETPU membranes,more reduction in selectivity is observed which can be explained by PU effect on the membrane morphology.As mentioned before,in the case of ETPU,thermodynamic parameter for the membrane precipitation increased more than that of ESPU and this enhancement is expected to lead higher gas permeability and lower CO2/CH4selectivity for PES-ETPU than that of PES/ESPU.Table 4 shows the permeation results at various feed pressures and PUs compositions.

Fig.11.Effect of pressure and ESPU composition on CO2 and CH4 selectivity.

Fig.12.Effect of pressure and ETPU composition on CO2 and CH4 selectivity.

3.4.Fitting the model

As mentioned earlier,RSM was used to optimize permeance and ideal selectivity of PES-ETPU.

PES-ESPU membranes and the experimental results were presented in Table 5.

Experimental permeance and ideal selectivity yields of used membranes were analyzed to get a regression model.The predicted values of permeance and ideal selectivity yields were calculated using the regression model and compared with the experimental values.Theestimated equations for predicting permeance and ideal selectivity yields are obtained as following:

Table 4 The results of permeation and selectivity at various pressures and compositions.

Table 5 Experimental and predicted data for the yield of permeation and selectivity from experimental design

where A and B are the pressure and PU composition respectively.

In order to verify model,correlation coefficient(R2)and standard deviation of the presented model were calculated.These values for Eqs.(5)and(6)are 0.97 and 0.045,respectively and also 0.95 and 0.32 for Eqs.(7)and(8).Based on these values,there is a good agreement between the experimental data and predicted values by model.

The effect of the variables as linear,quadratic,or interaction coefficients on the response was tested for significance by analysis of variance(ANOVA).The obtained results from variance analysis also proof model validation.Analysis of variance for per meance and selectivity results are tabulated in Tables 6 and 7.As shown in these tables,it can be found that the model F-value of 117.28 for per meance and 70.15 for selectivity implied that the model was significant.Values of Prob＞F less than 0.05 indicated that the model terms were significant,whereas thevalues greater than 0.1000 are not significant.It can be seen that for the permeance,F-values of the pressure(A),composition(B)and type of polymer(C)were 194.08,338.76 and 32.58,respectively.It can be deducted thatthe CO2permeance is more sensitive to the blend composition than the pressure and type of polymer.Moreover,for CO2/CH4selectivity,F-values of A,B and C were 15.31,170.2 and 130.07,respectively and the blend composition(B)had more effect than two parameters for this response.

Table 6 ANOVA for the regression model and respective model terms for permeance

Table 7 ANOVA for the regression model and respective model terms for CO2/CH4 selectivity

3.4.1.Response surface analysis

Response surfaces can be illustrated with three-dimensional plots by presenting the responses in function of two factors and keeping the otherconstant.Itis visualized by the yields of permeance and selectivity as a function of pressure and ether and ester compositions in PES-ETPU and PES-ESPU membranes in Fig.13.As it can be seen in these figures,increasing pressure and PU compositions lead to permeance increment and selectivity reduction.It should also be noted that the presented model has a good agreement with experimental data.

4.Conclusions

In the present study,the transport performances of carbon dioxide and methane gases were studied in PES,PES-ETPU and PES-ESPU membranes separately with different ESPU and ETPU compositions.The presence of two types of PU in pure PES membrane causes changes in physical and chemical characterization of membrane as following results:

·Reduction in the thickness of the upperlayer and enhancement on the amount of holes and their size(based on SEM).

·Increases the hydrophilicity of the solution which leads to more non-solvent(water)penetrate into the casting film(based on CA).

·Two polymers are well miscible with each other and there is an intermolecular interaction between PES and PU molecules(based on FTIR).

·More permeation of CO2and CH4and reduction in ideal selectivity of CO2/CH4.

·Higher permeability and less selectivity of PES-ETPU membrane in contract with PES-ESPU one.

Ultimately,a second order model based on RSM was suggested to predict and optimize the experimental data at wide range.The model was built based on the variables with correlation coefficients of 0.97 and 0.95 for permeability and selectivity respectively.

Fig.13.Effect of pressure and composition on permeance and selectivity(a)and(c)for PES-ESPU,(b)and(d)for PES-ETPU.(1 bar=0.1 MPa).

[1]E.Fouad,F.Ahmad,S.Sawalha,Removal of carbon dioxide from flue gases using polyethersul fone(PES)and polyimide(PI)blended membranes,Int.J.Chem.Environ.Eng.5(2014)89-93.

[2]Y.Xiao,B.T.Low,S.S.Hosseini,T.S.Chung,D.R.Paul,The strategies of molecular architecture and modification of polyimide-based membranes for CO2removal from natural gas—A review,Prog.Polym.Sci.34(2009)561-580.

[3]S.Sridhar,T.M.Aminabhavi,M.Ramakrishna,Separation of binary mixtures of carbon dioxide and methane through sulfonated polycarbonate membranes,J.Appl.Polym.Sci.105(2007)1749-1756.

[4]J.Han,W.Lee,J.M.Choi,R.Patel,B.R.Min,Characterization of polyetheretherketone with card(PEEK-WC)/polyimide blend membranes prepared by a dry/wet phase inversion:Precipitation kinetics,morphology and gas separation,J.Membr.Sci.351(2010)141-148.

[5]G.C.Kapantaidakis,S.P.Kaldis,X.S.Dabou,G.P.Sakellaropoulos,Gas permeation through PSF-PI miscible blend membranes,J.Membr.Sci.110(1996)239.

[6]G.C.Kapantaidakis,S.P.Kaldis,G.P.Sakellaropoulos,E.Chira,B.Loppinet,G.Floudas,Interrelation between state and gas permeation in polysulfone/polyimide blend membranes,J.Polym.Sci.B Polym.Phys.37(1999)2950.

[7]X.Ding,Y.Cao,H.Zhao,L.Wang,Q.Yuan,Fabrication of high performance Matrimid/polysulfone dual-layer hollow fiber membranes for O2/N2separation,J.Membr.Sci.323(2008)352.

[8]D.Li,T.S.Chung,R.Wang,Morphological aspects and structure control of dual-layer asymmetric hollow fiber membranes formed by a simultaneous co-extrusion approach,J.Membr.Sci.243(2004)155.

[9]K.Liang,J.Grebowicz,E.Valles,F.E.Karasz,W.J.MacKnight,Thermal and rheological properties of miscible polyethersulfone/polyimide blends,J.Polym.Sci.B Polym.Phys.30(1992)465.

[10]A.Bos,I.Pünt,H.Strathmann,M.Wessling,Suppression of gas separation membrane plasticization by homogeneous polymer blending,AIChE J.47(2001)1088-1093.

[11]M.Pegoraro,F.Severini,R.Gallo,L.Zanderighi,Gas transport properties of polycarbonate-polyurethane membranes,J.Appl.Polym.Sci.57(1995)421-429.

[12]K.H.Hsieh,C.C.Tsai,S.M.Tseng,Vapor and gas permeability of polyurethane membranes.Part II:effect of functional group,J.Membr.Sci.49(1990)341-350.

[13]S.-H.Chen,K.-C.Yu,S.-L.Houng,J.-Y.Lai,Gas transport properties of HTPB based polyurethane/Cosalen membrane,J.Membr.Sci.173(2000)99-106.

[14]Sh.Hassanajili,M.A.Khademi,P.Keshavarz,In fluence of various types of silica nanoparticles on permeation properties of polyurethane/silica mixed matrix membranes,J.Membr.Sci.310(453)(2014)369-383.

[15]J.A.de Sales,P.S.O.Patrício,J.C.Machado,G.G.Silva,D.Windm?ller,Systematic investigation of the effects of temperature and pressure on gas transport through polyurethane/poly(methylmethacrylate)phase-separated blends,J.Membr.Sci.310(2008)129-140.

[16]H.Y.Hwang,D.J.Kim,W.J.Yim,S.Y.Nam,PES/SPAES blend membranes for nano filtration:The effects ofsulfonic acid groups and thermaltreatment,Desalination 289(2012)72-80.

[17]S.S.Hosseini,M.M.Teoh,T.Sh.Chung,Hydrogen separation and purification in membranes of miscible polymer blends with interpenetration networks,Polymer 49(2008)1594e1603.

[18]L.Lin,Y.Kong,K.Xie,F.Lua,R.Liu,L.Guo,S.Shao,J.Yang,D.Shi,Y.Zhang,Polyethylene glycol/polyurethane blend membranes for gasoline desulphurization by pervaporation technique,Sep.Purif.Technol.61(2008)293-300.

[19]W.Hu,L.Enying,L.G.Yao,Optimization of drawbead design in sheet metal forming based on intelligent sampling by using response surface methodology,J.Mater.Process.Technol.206(2008)45-55.

[20]Q.Tang,Y.B.Lau,Sh.Hu,W.Yan,Y.Yang,T.Chen,Response surface methodology using Gaussian processes:towards optimizing the trans-stilbene epoxidation over CO2+-NaX catalysts,Chem.Eng.J.156(2010)423-431.

[21]W.Shijin,Y.Xiang,H.Zhihang,Z.H.Lili,C.H.Jianmeng,Optimizing aerobic biodegradation of dichloromethane using response surface methodology,J.Environ.Sci.21(2009)1276-1283.

[22]A.L.Ahmad,S.C.Low,S.S.R.Abd,A.Ismail,Optimization of membrane performance by thermal-mechanical stretching process using responses surface methodology(RSM),Sep.Purif.Technol.66(2009)177-186.

[23]A.F.Ismail,P.Y.Lai,Development of defect-free asymmetric polysulfone membranes for gas separation using response surface methodology,Sep.Purif.Technol.40(2004)191-207.

[24]N.F.Razali,A.W.Mohammad,N.Hilal,Ch.P.Leo,J.Alam,Optimisation of polyethersulfone/polyaniline blended membranes using response surface methodology approach,Desalination 311(2013)182-191.

[25]F.Xiangli,W.Wei,Y.Chen,W.Jin,N.Xu,Optimization of preparation conditions for polydimethylsiloxane(PDMS)/ceramic composite pervaporation membranes using response surface methodology,J.Membr.Sci.311(2008)23-33.

[26]Z.L.Mou,L.J.Zhao,Q.A.Zhang,J.Zhang,Z.Q.Zhanga,Preparation ofporous PLGA/HA/collagen scaffolds with supercritical CO2and application in osteoblast cell culture,J.Supercrit.Fluids 58(2011)398-406.

[27]H.Adib,Sh.Hassanajili,D.Mowla,F.Esmaeilzadeh,Fabrication of integrally skinned asymmetric membranes based on nanocomposite polyethersulfone by supercritical CO2for gas separation,J.Supercrit.Fluids 97(2015)6-15.

[28]L.S.Teo,C.Y.Chen,J.F.Kuo,The gas transport properties of amine-containing polyurethane and poly(urethane-urea)membranes,J.Membr.Sci.141(1998)91.

[29]M.Sadeghi,M.A.Semsarzadeh,H.Moadel,Enhancement of gas separation properties of polybenzimidazole membrane by incorporation of silica nanoparticles,J.Membr.Sci.331(2009)21.

[30]B.Deng,X.Yang,L.Xie,J.Li,Z.Hou,S.Yao,G.Liang,K.Sheng,Q.Huang,Micro filtration membranes with pH dependent property prepared from poly(methacrylic acid)grafted polyethersulfone powder,J.Membr.Sci.330(2009)363-368.

[31]V.Melnig,M.O.Apostu,V.Tura,C.Ciobanu,Optimization of polyurethane membranes morphology and structure studies,J.Membr.Sci.267(2005)58-67.

[32]L.Bistri?i?,G.Baranovic,M.Leskovac,E.Govor?in Bajsic,Hydrogen bonding and mechanical properties of thin films of polyether-based polyurethane-silica nanocomposites,Eur.Polym.J.46(2010)1975.

[33]W.J.Lau,A.F.Ismail,Effect of SPEEK content on the morphological and electrical properties of PES/SPEEK blend nano filtration membranes,Desalination 249(2009)996-1005.

[34]A.Rahimpour,S.S.Madaeni,S.Mehdipour-Ataei,Synthesis of a novel poly(amideimide)(PAI)and preparation and characterization of PAI blended polyethersulfone(PES)membranes,J.Membr.Sci.311(2008)349-359.

[35]M.Sadeghi,M.M.Talakesh,B.Ghalei,M.R.Sha fiei,Preparation,characterization and gas permeation properties of a polycaprolactone based polyurethane-silica nanocomposite membrane,J.Membr.Sci.427(2013)21.

[36]K.R.Lee,M.Y.Teng,T.N.Hsu,J.Y.Lai,A study on pervaporation of aqueous ethanol solution by modified polyurethane membrane,J.Membr.Sci.162(1999)173-180.

[37]Sh.Saedi,S.S.Madaeni,K.Hassanzadeh,A.Arabi Shamsabadi,S.Laki,The effect of polyurethane on the structure and performance of PES membrane for separation of carbon dioxide from methane,J.Ind.Eng.Chem.44(2013)2944-2950.

[38]A.Rahimpour,S.S.Madaeni,Y.Mansourpanah,Fabrication of polyethersulfone(PES)membranes with nano-porous surface using potassium perchlorate(KClO4)as an additive in the casting solution,Desalination 258(2010)79-86.

[39]A.Bansal,H.Yang,C.Li,B.C.Benicewicz,S.K.Kumar,L.S.Schadler,Controlling the thermomechanical properties of polymer nanocomposites by tailoring the polymer-particle interface,J.Polym.Sci.B Polym.Phys.44(2006)2944.

[40]J.Ahn,W.J.Chung,I.Pinnau,M.D.Guiver,Polysulfone/silica nanoparticle mixedmatrix membranes for gas separation,J.Membr.Sci.314(2008)123.

[41]M.Sadrzadeh,S.Bhattacharjee,Rational design of phase inversion membranes by tailoring thermodynamics and kinetics of casting solution using polymer additives,J.Membr.Sci.441(2013)31-44.

[42]M.Sadeghi,G.Khanbabaei,A.H.Saeedi Dehghani,M.Sadeghi,M.A.Aravand,M.Akbarzade,S.Khatti,Gas permeation properties of ethylenevinyla cetate-silica nanocomposite membranes,J.Membr.Sci.322(2008)423.