Improved thiophene solution selectivity by Cu2+, Pb2+ and Mn2+ ions in pervaporative poly[bis(p-methyl phenyl) phosphazene]desulfurization membrane

Improved thiophene solution selectivity by Cu2þ , Pb2þ and Mn2þ ions in pervaporative poly[bis(p-methyl phenyl) phosphazene] desulfurization membrane Zhengjin Yang a, Wei Zhang b, Tao Wang a, Jiding Li a,n a The State Key Laboratory of Chemical Engineering, Department of Chemical Engineering, Tsinghua University, Beijing 100084, PR China b Center for Membrane Technology, College of Environmental and Energy Engineering, Beijing University of Technology, Beijing 100124, PR China a r t i c l e i n f o Article history: Received 1 November 2013 Received in revised form 12 December 2013 Accepted 14 December 2013 Available online 22 December 2013 Keywords: Facilitated transport Pervaporation Desulfurization Poly[bis(p-methyl phenyl) phosphazene] membrane a b s t r a c t Due to interfacial defects and agglomeration, role of metal ions in mixed matrix membrane during pervaporation is not fairly demonstrated and not fully taken advantage of. Through surface modification of pristine PMePP membrane and the followed ion exchange treatments, PMePP-Mmembranes containing Cu2þ , Mn2þ , or Pb2þ ions were obtained. X-ray photoelectron spectroscopy (XPS) analysis confirms the successful exchange of Cu2þ , Mn2þ , or Pb2þ ions onto the membrane surface. Pervaporation results show that the removal of thiophene from model gasoline by pristine PMePP membrane or modified PMePP membranes is a solution-diffusion controlled process while it shifts to a solution controlled process for PMePP-Mmembranes. It provides important evidence for the facilitated thiophene mass transport via metal ions. PMePP-Mmembranes show superior selectivity in a Robeson type plot and their performance locates above the upper bound of literature reported thiophene permeability and selectivity. In the meantime, it is proved that post-treatment and control of oxidation temperature can compensate the decrease of pervaporative flux caused by the change in surface hydrophobicity and the introduction of electric charge. & 2013 Elsevier B.V. All rights reserved. 1. Introduction Pervaporation, an eco-friendly, energy efficient and easily operated membrane process, serves as a promising alternative to energy intensive distillation process in separating close boiling mixtures, azeotropic mixtures as well as substances that cannot withstand harsh operating conditions such as high temperature or high pressure [1,2]. The key to success in pervaporation lies in the development of advanced membranes with both high selectivity and high permeability. To date, two major types of membranes are developed, the polymeric ones and the inorganic ones [2,3]. Recently, the incorporation of inorganic fillers into polymeric matrix, i.e. the mixed matrix membranes (MMMs) deserves special attention. The synergistic effect between the matrix and the fillers is expected to exceed the separation obtained by the use of one single component [4]. They are now widely investigated in dehydration of alcohols [5–10] and the separation of isomers [11]. This method is even more important in fluid catalytic cracking (FCC) gasoline deep desulfurization. By incorporating Cu2þY zeolite [12], Ni2þY zeolite [13], AgþY zeolite [14] and Ag2O particles [15], the obtained MMMs for FCC gasoline desulfurization are increased in permeation flux without much sacrifice in selectivity or gains an increase in both flux and enrichment factor elsewhere. However, several scientific problems have not been clarified yet, such as the match of mass transfer properties, the particle size and size distribution, agglomeration and the interfacial contact. For metallic ions incorporated MMMs, the role played by metallic ions, such as Agþ , Cu2þ , or Pb2þ is rarely demonstrated. To our knowledge, the π-complexation between thiophene molecules and silver ions was once proposed [15,16]. This hypothesis is possible since the sulfur in thiophene can donate lone electron pair and silver ions possess empty d orbit as receptor. Therefore, an increase in permeation flux and selectivity can be probably attributed to the facilitated mass transportation by silver ions. However, this hypothesis could not be verified unless the effects of other factors, such as interfacial defects, pores inside zeolites and agglomeration are excluded. Pervaporative poly[bis(p-methyl phenyl) phosphazene] (PMePP) membrane was previously synthesized [17], functionalized and investigated in FCC model gasoline desulfurization. After surface modification, PMePP is increased in stability during the pervaporation process and is winged for grafting and ion exchange treatment. In this work, Cu2þ , Mn2þ and Pb2þ ions were exchanged onto the surface of the functionalized PMePP membrane via ion exchange treatment to investigate the role played by metallic ions in FCC gasoline desulfurization. In this specific situation, the effect of interfacial defects and agglomeration was excluded. Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/memsci Journal of Membrane Science 0376-7388/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2013.12.036 n Corresponding author. Tel./fax: þ86 10 62782432. E-mail address: [email protected] (J. Li). Journal of Membrane Science 454 (2014) 463–469 www.sciencedirect.com/science/journal/03767388 www.elsevier.com/locate/memsci http://dx.doi.org/10.1016/j.memsci.2013.12.036 http://dx.doi.org/10.1016/j.memsci.2013.12.036 http://dx.doi.org/10.1016/j.memsci.2013.12.036 http://crossmark.crossref.org/dialog/?doi=10.1016/j.memsci.2013.12.036&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1016/j.memsci.2013.12.036&domain=pdf http://crossmark.crossref.org/dialog/?doi=10.1016/j.memsci.2013.12.036&domain=pdf mailto:[email protected] http://dx.doi.org/10.1016/j.memsci.2013.12.036 2. Experimental 2.1. Materials Hexachlorocyclotriphosphazene (HCCP) was purchased from XinYi Chemical. Co., Ltd. (Jiangsu, China). It was recrystallized twice in heptane and sublimed at 50 1C before use. 4-Methyl phenol (499%) and benzophenone ketyl (499%) were purchased from Alfa Aesar (Johnson Matthey, UK). Tetrahydrofuran and toluene were purchased from FuChen Chemicals (Tianjin, China) and were purified by distilla- tion over sodium benzophenone ketyl. Sodium bisulfite (NaHSO3), copper nitrate (Cu(NO3)2), lead nitrate (Pb(NO3)2) and manganese nitrate aqueous solution (Mn(NO3)2, 50%) were obtained from Sino- pharm Chemical Reagent Co., Ltd. (Beijing, China). Potassium perman- ganate (KMnO4) was purchased from Beijing Chemical Works (Beijing, China). These reagents are analytically pure and used as received. 2.2. Pristine PMePP membrane Poly[bis(p-methyl phenyl) phosphazene] (PMePP) was synthe- sized and characterized by substituting chlorine atoms in poly (dichlorophosphazene) using sodium 4-methyl phenoxide as shown in Fig. 1 [18]. PMePP composite membranes were fabricated by the solution casting method. A homogeneous solution of PMePP (1 g) in dry THF (30 ml) was cast onto the porous PVDF (PVDF-1015, Mw�2.38�105 g/mol) support. After evaporating THF at 25 1C overnight, the as prepared membrane was heated at 80 1C for 4 h to further remove the residual solvent. The obtained membrane was stored in dry and clean atmosphere for further characterization. The thickness of the dense PMePP layer was around 15 μm. 2.3. Surface modification of PMePP membrane The as prepared PMePP composite membrane was immersed in aqueous KMnO4 solutions (KMnO4 40 mM, NaOH, 20 mM) at varied temperatures (30 1C, 50 1C and 80 1C) and time. During this process, methyl groups on membrane surface were partially oxidized into carboxylic groups. Because of the basicity of the environment, carboxylic groups were neutralized once generated. The obtained membrane was thus designated as PMePP-Na. It was then removed from the reaction solution to an aqueous NaHSO3 (0.4 mol/L) solution for 5 h to remove the MnO2. Fig. 1. Schematic of PMePP synthesis and surface functionalization treatment. Fig. 2. Schematic for ion exchange treatment on modified PMePP membrane surface. Z. Yang et al. / Journal of Membrane Science 454 (2014) 463–469464 2.4. Ion exchange treatments of modified PMePP membrane Three pieces of surface modified PMePP membranes were immersed in Cu(NO3)2(0.4M), Pb(NO3)2(0.4M) and Mn(NO3)2(50%) aqueous solutions for 48 h under stirring, respectively (Fig. 2, desig- nated as PMePP-Cu, PMePP-Pb and PMePP-Mn, respectively). They were then rinsed with deionized water for 48 h to remove the absorbed metallic salts and dried in a 50 1C oven for 5 h. 2.5. Characterization 2.5.1. X-ray photoelectron spectroscopy (XPS) analysis Surface chemical composition of PMePP membranes after ion exchange treatments was analyzed by PHI Quantera SXM XPS instrument (ULVAC-PHI, USA) using an Al Kα as the radiation source. The takeoff angle of the photoelectron was set at 451. Survey spectra were collected over the range of 0–1200 eV with a resolution of 0.5 eV. 2.5.2. Pervaporation The afore-mentioned pristine PMePP membrane and the mem- branes that were modified and ion exchanged were investigated in a self-designed apparatus for the removal of thiophene from n-heptane (representative compound of FCC gasoline). The feed mixture was continuously circulated in the upstream by a pump. The effective area of stainless steel membrane cell was about 21.67 cm2. When a steady flow was achieved, the permeated mixture was collected in a nitrogen cold trap and sulfur content was analyzed. The feed sulfur content is around 200 mg/L and the pressure of the downstream was kept at less than 300 Pa with a vacuum pump. The effect of feed temperature on permeation and selectivity was investigated. 3. Results and discussion 3.1. XPS characterization In this study, copper ion (Cu2þ) was chosen because of the report of Yang et al. [19] on the formation of π-complexation between thiophene and copper ions (Cuþ) without eliminating the effect of inherent pores and interfacial contacts and due to the similarity between Cu2þ and Cuþ . Mn2þ and Pb2þ were selected based on affinity difference. X ray photospectroscopy was used to confirm the successful and full exchange of Naþ with Cu2þ , Mn2þor Pb2þ . Fingerprint peaks for Cu2þ , Mn2þor Pb2þwere recorded in XPS survey spectrum (Fig. 3a) confirming the successful exchange of Cu2þ , Mn2þand Pb2þonto the corresponding membranes. Also, no peaks were found in the high resolution sodium XPS survey spectrum (Fig. 3b) indicating the fully substitution of Naþ . 3.2. Improved thiophene selectivity at elevated operating temperature Pervaporation performance of pristine PMePP membrane, sur- face modified PMePP membrane (PMePP-Na) and three other metal ion exchanged membranes, (i.e. PMePP-Cu, PMePP-Pb and PMePP-Mn) for the removal of thiophene from n-heptane at varying operating temperatures are presented in Figs. 4 and 5. It can be observed that the enrichment factor of pristine PMePP membrane exhibits first increasing then decreasing trend when the operating temperature was increased. This trend which is commonly attributed to the competition between solution selec- tivity and diffusion selectivity is similar to other reports [12,13]. When the operating temperature was lower than 65 1C, the solution selectivity is dominant. Consequently, the selectivity of thiophene increase when the operating temperature was elevated because of the increase in thiophene sorption. At higher operating temperature (465 1C), diffusion selectivity dominates this pro- cess. The faster increase of n-heptane mass transport exceeds that of thiophene leading to the decrease in enrichment factor. However, enrichment factor of PMePP-Na membrane is slightly decreased before 65 1C and remains stable at temperatures above 75 1C. Compared with pristine PMePP membrane, enrichment Fig. 3. XPS survey (a) and high resolution sodium spectra (b) on ion exchanged PMePP membranes. Fig. 4. Effect of operating temperature on pervaporation flux of (a) PMePP, (b) PMePP-Na, (c) PMePP-Mn, (d) PMePP-Cu, and (e) PMePP-Pb. The feed mixture is composed of thiophene and n-heptane with a sulfur content of 200 ppm. Z. Yang et al. / Journal of Membrane Science 454 (2014) 463–469 465 factor of PMePP-Na membrane was decreased which can be attributed to the more hydrophilic membrane surface due to the ionic bonding by –COONa groups (Fig. 5A). In the meantime, the pervaporation flux of both pristine PMePP and PMePP-Na are increased with an increase in operating temperature because of the increase in mass transfer driving force. However, the introduc- tion of electric charge leads to the increase in mass transfer resistance which in turn causes the slight decrease in pervapora- tion flux of PMePP-Na membrane. Fortunately, higher pervapora- tive flux of PMePP-Na membrane can be obtained by increasing operating temperatures due to the improvement in selectivity stability. As can be observed in Figs. 4 and 5 (curves c, d and e), permeation flux and selectivity increase simultaneously when Naþ is exchanged with Cu2þ , Mn2þor Pb2þ . Compared with PMePP-Na or pristine PMePP, the enrichment factors of PMePP- Cu, PMePP-Mn and PMePP-Pb are increased when the operating temperature is elevated from 45 1C to 85 1C. The result implies that after ion exchange treatment, the selectivity of as prepared PMePP-M membrane is dominated solely by thiophene solution selectivity in all the operating temperatures investigated. Thio- phene solution selectivity is improved because of the introduction of metal ions which form a π-complexation with thiophene resulting in the increase of thiophene sorption. It should be noted that when the operating temperature is higher than 75 1C, the enrichment factor of PMePP-M membrane is higher than that of both pristine PMePP membrane and surface modified PMePP membrane. 3.3. Facilitated transport mechanism As previously mentioned, the increase in selectivity and per- meation flux is realized simultaneously with increasing operating temperature after ion exchange treatment. The increase in thio- phene selectivity can be explained by the facilitated mass trans- port via metal ions as schemed in Fig. 6. Simulation results suggest that metal ions can form π-complexation with thiophene mole- cules [19,20]. Accordingly, compared with PMePP-Na membrane or pristine PMePP membrane, thiophene is preferably absorbed onto the surface of PMePP-M(Cu, Mn or Pb) membranes. Since π-complexation is similar but stronger than the van der Waals0 force, desorption of thiophene molecules from binding spots is energy required. If the energy required and the energy input are referred as Ed and Ei respectively, Ei is increased at elevated operating temperatures leading to break of binding between Cu2þ , or Pb2þ ions and thiophene. Therefore, more free thiophene molecules are available at the membrane surface. With more thiophene absorbed at the membrane surface and it is easier for them to desorb when Ei is increased, the mass transport of thiophene molecule is thus facilitated. That explains why the increase in operating temperature leads to an increase in mem- brane selectivity. Moreover, the enrichment factors of PMePP-Pb membrane and PMePP-Cu are 6.7 and 6.1, respectively, when the operating temperature is elevated to 85 1C. Accordingly, we expect that enrichment factor of both PMePP-Pb membrane and PMePP-Cu membrane would continue to increase when operating tempera- ture is further increased. Because of the greater affinity between Cu2þ and thiophene molecules, higher operating temperature might be required. The hypothesis on the formation of π-complexation between metal ions and thiophene and that the de-binding process is energy required is further supported by PMePP-Mn membrane. The binding of thiophene with metal ions is of different strength. To assess this binding strength, solubility constant (Ksp) of metal sulfides can be referred. Although solubility constant of metal sulfides is a tool to calculate their solubility, it also reflects the binding tendency between sulfur atoms and metal ions. Moreover, due to their special nature, the way metal ions bonded with sulfur ions shows somehow similarity to the π-complexation binding between sulfur in thiophene and metal ions. Thus, it is reasonable to utilize solubility constants to assess the binding of thiophene with metal ions based on the above-mentioned facts. Ksp of CuS, PbS, and MnS are 6.3�10�36, 1.0�10�28 and 2.5�10�10 respec- tively. The smaller Ksp is, the closer is in binding between sulfur atoms and metal ions. Thus, Mn2þ forms the weakest binding with thiophene. This binding tendency can also be assessed by the electro-negativity of metal ions. Mn2þ exhibits the smallest electronegativity (1.55) and thus it is less attractive to lone pair electrons. Compared with PMePP-Mn membrane, higher thio- phene concentration was formed on the surface of PMePP-Pb membrane and PMePP-Cu membrane due to stronger binding. Nevertheless, higher thiophene concentration forms a blockage to thiophene permeability at low operating temperature since the break of this strong binding required more energy. That explains why PMePP-Mn membrane exhibits both higher flux and higher enrichment factor when operating temperature is below 75 1C. However, Ei would be high enough to break the strong binding between Pb2þ ions and thiophene when operating temperature was increased. In this situation, high thiophene concentration at PMePP-Pb membrane surface leads to an increase in mass transfer driving force and facilitates the transportation of thiophene. That is the reason why PMePP-Pb membrane exhibits higher Fig. 5. Effect of operating temperature on enrichment factors of (a) PMePP, (b) PMePP-Na, (c) PMePP-Mn, (d) PMePP-Cu, and (e) PMePP-Pb. Z. Yang et al. / Journal of Membrane Science 454 (2014) 463–469466 enrichment factor at 85 1C. Because of the strongest binding between Cu2þ ions and thiophene molecules, higher operating temperature is required to overcome this blockage. 3.4. Robeson type plot Robeson [21] suggested the plot of permeability versus selec- tivity in gas separation and proved there existed the upper bound. It is now recognized as trade-off effect. The same method was introduced in pervaporation based on the proposal of Wijmans et al. [22,23] on introducing driving force normalized permeability to characterize membrane pervaporation performance. The driving force normalized permeability is calculated as in the following equation, Pi ¼ Ji � l pi0�pil ð1Þ where Ji is the pervaporation flux of component i, l is the membrane thickness, while pi0 and pil are the vapor pressure of component i at the upstream side and the downstream side respectively. The desulfurization performance of PMePP-M (Cu, Mn or Pb) in this work and some other desulfurization membranes in literatures [24–35] were recalculated according to Eq. (1). The results are presented in Fig. 7 and an upper bound was drawn for comparison. We found that comparing with classic polymeric membranes, thiophene permeability and selectivity of MMMs (mixed matrix membranes) and PMePP-M (Cu, Mn or Pb) are above the upper bound. MMMs show superior permeability by sacrificing selectivity due to the existence of interfacial defects or agglomera- tion and PMePP-M (Cu, Mn, or Pb) exhibits superior selectivity due to the facilitated mass transportation by metal ions. The above results also suggest that optimal desulfurization performance of MMMs can be accessed by the incorporation of metal ions and the elimination of interfacial defects and agglomeration. Moreover, trade-off between selectivity and flux can be exceeded when suitable transition metal ion is exchanged onto the membrane surface. Comparing the enrichment factor and permeation flux of PMePP-Cu with PMePP-Mn membrane (Figs. 4 and 5), we found that PMePP-Mn membrane exhibits both high flux and high selectivity since the strong combination of Cu2þ ions with thiophene causes great resistance both in mass transportation and the desorption of thiophene. 3.5. Performance enhancement via controlling modification conditions Compared with pristine PMePP membrane, surface modified PMePP (PMePP-Na) membrane and PMePP-M (Cu, Mn, or Pb) are both decreased in permeation flux. It can be attributed to the introduction of free electric charge and the change in surface hydrophobicity. Consequently, optimized performance shall be obtained via controlling modification conditions and by post- treatment. To testify this hypothesis, effect of acidic treatment Fig. 6. Schematic on the formation of π-complexation between transition metal ions and thiophene. Fig. 7. Desulfurization performance comparison of PMePP-Cu, PMePP-Mn and PMePP-Pb with reported polymers and mixed matrix membranes. Z. Yang et al. / Journal of Membrane Science 454 (2014) 463–469 467 and oxidation temperature on performance of surface modified PMePP membrane (PMePP-Na) was investigated. 3.5.1. Effect of acidic treatment on PMePP-Na membrane performance As previously mentioned, the decrease in permeation flux of PMePP-Na membrane is believed to be partially caused by the introduction of ionic nature and could be remedied by neutralizing surface electric charge. To demonstrate this, PMePP-Na mem- branes are immersed in HCl aqueous solutions with pH values of 1 or 4 and pure water (referred as pH¼7). During the process, the surface carboxylate anions are gradually protonated by Hþ . As can be observed in Fig. 8, the flux of PMePP-Na membrane that was immersed in HCl solution with lower pH value is higher, i.e. PMePP-Na membrane treated in HCl (pH¼1) solution shows the highest pervaporation flux. This confirms that the decrease in pervaporation flux is attributed to the introduction of electric charge. Also, permeation flux increases while enrichment factor decreases when the operating temperature is elevated. This follows the same trend with previous mentioned for PMePP-Na membrane. 3.5.2. Effect of oxidation temperature on PMePP-Na membrane performance Fig. 9 shows the effect of oxidation temperature on PMePP-Na membrane performance. As observed in Fig. 9a, PMePP-Na obtained by oxidizing pristine PMePP in basic KMnO4 solutions at 50 1C for 5 h exhibits the highest enrichment factor for the removal of thiophene from n-heptane and those oxidized at 30 1C or 80 1C is lower. Nevertheless, PMePP-Na obtained by oxidizing PMePP at 30 1C or 50 1C shows higher pervaporation flux than that obtained at 80 1C (Fig. 9b). The difference in pervaporation performance of PMePP-Na membranes can be explained by the balance between hydrophobicity and hydrophilicity of the mem- brane surface. Higher oxidation temperature means more methyl groups being oxidized into carboxylate anions which are believed to be more hydrophilic. In the meantime, thiophene is of stronger polarity than n-heptane. Thus, the increase in hydrophilicity of PMePP-Na membrane might lead to better affinity between thiophene and PMePP-Na. However, the mass transfer resistance would be increased when the binding between thiophene and PMePP-Na membrane becomes stronger. Consequently, there lies the balance between hydrophobicity and hydrophilicity. In the current study, the optimal oxidation temperature is 50 1C con- cerning both permeability and selectivity. Fig. 8. Pervaporative flux (a) and enrichment factor (b) of PMePP-Na membranes after immersed in HCl aqueous solutions (pH¼1 or pH¼4) and pure water (pH¼7) for 48 h. PMePP-Na was obtained by oxidizing PMePP in basic KMnO4 solutions (KMnO4 40 mM, NaOH, 20 mM) at 80 1C for 5 h. Fig. 9. Enrichment factor (a) and pervaporative flux (b) of PMePP-Na membranes at various operating temperatures. PMePP-Na was obtained by oxidizing PMePP in basic KMnO4 solutions (KMnO4 40 mM, NaOH, 20 mM) at 30 1C, 50 1C and 80 1C for 5 h. Z. Yang et al. / Journal of Membrane Science 454 (2014) 463–469468 4. Conclusion Through surface modification of PMePP, membranes bearing carboxylate anions are obtained. These carboxylate anions provide important binding spot for metal ions in ion exchange treatment. When Cu2þ , Mn2þ or Pb2þ ions were exchanged onto the membrane surface, the simultaneous increase in both permeabil- ity and selectivity was observed at elevated operating temperature due to the improved thiophene solution selectivity. Compared with PMePP-Pb or PMePP-Cu, PMePP-Mn shows both high enrich- ment factor and high flux in all operating temperature range investigated. It confirms the facilitated transportation of thiophene molecules by metal ions. Also, the introduction of ionic nature causes the decrease in permeation flux of PMePP-Na membrane. It is proved that the decrease can be remedied by acidic treatments. In summary, ion exchange treatment at modified PMePP mem- brane surface results in simultaneous fulfillment of high selectivity and high flux, provides supports for facilitated mass transfer by metal ions in MMMs and eliminates trade-off between selectivity and permeability when suitable metal ion is selected. The evi- dence provided here for the facilitated mass transfer via metal ions would help to design better MMMs. Acknowledgments This job is financially supported in part from the National High Technology Research and Development Program of China (No. 2012AA03A607), the National Natural Science Foundation of China (21176135, 21206001), the Beijing Natural Science Foundation (2132010). References [1] X.S. Feng, R.Y.M. Huang, Liquid separation by membrane pervaporation: a review, Ind. Eng. Chem. Res. 36 (1997) 1048–1066. [2] P. Shao, R.Y.M. Huang, Polymeric membrane pervaporation, J. Membr. Sci. 287 (2007) 162–179. [3] T.C. Bowen, R.D. Noble, J.L. Falconer, Fundamentals and applications of pervaporation through zeolite membranes, J. Membr. Sci. 245 (2004) 1–33. [4] R.D. Noble, Perspectives on mixed matrix membranes, J. Membr. Sci. 378 (2011) 393–397. [5] M. Sairam, M. Patil, R. Veerapur, S. Patil, T. Aminabhavi, Novel dense poly(vinyl alcohol)-TiO2 mixed matrix membranes for pervaporation separation of water–isopropanol mixtures at 30 1C, J. Membr. Sci. 281 (2006) 95–102. [6] R. Veerapur, K. Gudasi, T. Aminabhavi, Sodium alginate–magnesium alumi- num silicate mixed matrix membranes for pervaporation separation of water– isopropanol mixtures, Sep. Purif. Technol. 59 (2008) 221–230. [7] L.Y. Jiang, T.S. Chung, Homogeneous polyimide/cyclodextrin composite mem- branes for pervaporation dehydration of isopropanol, J. Membr. Sci. 346 (2010) 45–58. [8] N.L. Le, Y. Wang, T.-S. Chung, Pebax/POSS mixed matrix membranes for ethanol recovery from aqueous solutions via pervaporation, J. Membr. Sci. 379 (2011) 174–183. [9] S.J. Lue, C.-F. Chien, K.P.O. Mahesh, Pervaporative concentration of ethanol– water mixtures using heterogeneous polydimethylsiloxane (PDMS) mixed matrix membranes, J. Membr. Sci. 384 (2011) 17–26. [10] G.M. Shi, T. Yang, T.S. Chung, Polybenzimidazole (PBI)/zeolitic imidazolate frameworks (ZIF-8) mixed matrix membranes for pervaporation dehydration of alcohols, J. Membr. Sci. 415–416 (2012) 577–586. [11] Y. Wang, T.S. Chung, H. Wang, S.H. Goh, Butanol isomer separation using polyamide-imide/CD mixed matrix membranes via pervaporation, Chem. Eng. Sci. 64 (2009) 5198–5209. [12] L.G. Lin, Y.Z. Zhang, H. Li, Pervaporation and sorption behavior of zeolite-filled polyethylene glycol hybrid membranes for the removal of thiophene species, J. Colloid Interface Sci. 350 (2010) 355–360. [13] B. Li, D. Xu, Z.Y. Jiang, X.F. Zhang, W.P. Liu, X. Dong, Pervaporation performance of PDMS–Ni2þY zeolite hybrid membranes in the desulfurization of gasoline, J. Membr. Sci. 322 (2008) 293–301. [14] R.B. Qi, Y.J. Wang, J. Chen, J.D. Li, S.L. Zhu, Removing thiophenes from n-octane using PDMS–AgY zeolite mixed matrix membranes, J. Membr. Sci. 295 (2007) 114–120. [15] R. Qi, Y. Wang, J. Chen, J. Li, S. Zhu, Pervaporative desulfurization of model gasoline with Ag2O-filled PDMS membranes, Sep. Purif. Technol. 57 (2007) 170–175. [16] A. Takahashi, F.H. Yang, R.T. Yang, New sorbents for desulfurization by π- complexation: thiophene/benzene adsorption, Ind. Eng. Chem. Res. 41 (2002) 2487–2496. [17] Z. Yang, T. Wang, X. Zhan, J. Li, J. Chen, Poly[bis(p-methyl phenyl) phospha- zene] pervaporative membranes for separating organosulfur compounds from n-heptane and its surface functionalization, Ind. Eng. Chem. Res. 52 (2013) 13801–13809. [18] H.R. Allcock, R.J. Fitzpatrick, L. Salvati, Oxidation of poly[bis(4-methylphenoxy) phosphazene] surfaces and chemistry of the surface carboxylic acid groups, Chem. Mater. 4 (1992) 769–775. [19] R.T. Yang, A. Takahashi, F.H. Yang, New sorbents for desulfurization of liquid fuels by π-complexation, Ind. Eng. Chem. Res. 40 (2001) 6236–6239. [20] J.O. Konecny, C.M. Douglas, Polymerization of dichlorophosphinic nitrides, J. Polym. Sci., Part A: Polym. Chem. 36 (1959) 195–203. [21] L.M. Robeson, Correlation of separation factor versus permeability for poly- meric membranes, J. Membr. Sci. 62 (1991) 165–185. [22] J.G. Wijmans, Process performance¼membrane propertiesþoperating condi- tions, J. Membr. Sci. 220 (2003) 1–3. [23] R.W. Baker, J.G. Wijmans, Y. Huang, Permeability, permeance and selectivity: a preferred way of reporting pervaporation performance data, J. Membr. Sci. 348 (2010) 346–352. [24] L.S. White, R.F. Wormsbecher, M. Lesemann, Membrane separation for sulfur reduction (US 7041212 B2), W R Grace & Co.-Conn., Columbia, MD (US) United States, 2006. [25] L.G. Lin, Y. Kong, G. Wang, H.M. Qu, J.R. Yang, D.Q. Shi, Selection and crosslinking modification of membrane material for FCC gasoline desulfuriza- tion, J. Membr. Sci. 285 (2006) 144–151. [26] L.H. Wang, Z.P. Zhao, J.D. Li, C.X. Chen, Synthesis and characterization of fluorinated polyimides for pervaperation of n-heptane/thiophene mixtures, Eur. Polym. J. 42 (2006) 1266–1272. [27] R.B. Qi, Y.J. Wang, J.D. Li, C.W. Zhao, S.L. Zhu, Pervaporation separation of alkane/thiophene mixtures with PDMS membrane, J. Membr. Sci. 280 (2006) 545–552. [28] R. Cao, X. Zhang, H. Wu, J. Wang, X. Liu, Z. Jiang, Enhanced pervaporative desulfurization by polydimethylsiloxane membranes embedded with silver/ silica core–shell microspheres, J. Hazard. Mater. 187 (2011) 324–332. [29] Q.G. Zhang, B.C. Fan, Q.L. Liu, A.M. Zhu, A novel poly(dimethyl siloxane)/poly (oligosilsesquioxanes) composite membrane for pervaporation desulfuriza- tion, J. Membr. Sci. 366 (2011) 335–341. [30] R. Xu, G.P. Liu, X.L. Dong, W.Q. Jin, Pervaporation separation of n-octane/ thiophene mixtures using polydimethylsiloxane/ceramic composite mem- branes, Desalination 258 (2010) 106–111. [31] B. Li, W. Liu, Z. Jiang, X. Dong, B. Wang, Y. Zhong, Ultrathin and stable active layer of dense composite membrane enabled by poly(dopamine), Langmuir 25 (2009) 7368–7374. [32] F.W. Lu, Y. Kong, H.L. Lv, J. Ding, J.R. Yang, The pervaporation performance of polyimide-block-polyethylene glycol membranes for gasoline desulphuriza- tion: effect of PEG groups, Adv. Mater. Res. 150–151 (2010) 317–320. [33] Y. Kong, L. Lin, Y. Zhang, F. Lu, K. Xie, R. Liu, L. Guo, S. Shao, J. Yang, D. Shi, Studies on polyethylene glycol/polyethersulfone composite membranes for FCC gasoline desulphurization by pervaporation, Eur. Polym. J. 44 (2008) 3335–3343. [34] C. Zhao, J. Li, J. Chen, R. Qi, Z. Luan, Separation of sulfur/gasoline mixture with polydimethylsiloxane/polyetherimide composite membranes by pervapora- tion, Chin. J. Chem. Eng. 17 (2009) 707–710. [35] B. Li, S. Yu, Z. Jiang, W. Liu, R. Cao, H. Wu, Efficient desulfurization by polymer– inorganic nanocomposite membranes fabricated in reverse microemulsion, J. Hazard. Mater. 211–212 (2012) 296–303. Z. Yang et al. / Journal of Membrane Science 454 (2014) 463–469 469 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref1 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref1 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref2 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref2 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref3 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref3 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref4 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref4 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref5 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref5 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref5 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref5 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref5 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref6 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref6 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref6 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref7 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref7 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref7 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref8 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref8 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref8 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref9 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref9 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref9 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref10 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref10 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref10 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref11 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref11 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref11 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref12 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref12 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref12 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref13 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref13 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref13 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref13 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref13 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref13 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref13 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref14 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref14 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref14 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref15 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref15 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref15 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref15 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref16 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref16 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref16 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref17 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref17 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref17 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref17 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref18 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref18 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref18 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref19 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref19 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref20 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref20 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref21 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref21 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref22 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref23 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref23 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref23 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref24 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref24 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref24 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref25 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref25 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref25 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref26 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref26 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref26 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref27 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref27 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref27 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref28 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref28 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref28 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref29 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref29 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref29 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref30 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref30 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref30 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref31 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref31 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref31 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref32 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref32 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref32 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref32 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref33 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref33 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref33 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref34 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref34 http://refhub.elsevier.com/S0376-7388(13)00989-7/sbref34 Improved thiophene solution selectivity by Cu2+, Pb2+ and Mn2+ ions in pervaporative poly[bis(p-methyl phenyl)... Introduction Experimental Materials Pristine PMePP membrane Surface modification of PMePP membrane Ion exchange treatments of modified PMePP membrane Characterization X-ray photoelectron spectroscopy (XPS) analysis Pervaporation Results and discussion XPS characterization Improved thiophene selectivity at elevated operating temperature Facilitated transport mechanism Robeson type plot Performance enhancement via controlling modification conditions Effect of acidic treatment on PMePP-Na membrane performance Effect of oxidation temperature on PMePP-Na membrane performance Conclusion Acknowledgments References