A polyethersulfone–bisphenol sulfuric acid hollow fiber ultrafiltration membrane fabricated by a reverse thermally induced phase separation process

A novel antifouling polyethersulfone (PES) hollow fiber membrane was modified by the addition of bisphenol sulfuric acid (BPA-PS) using a reverse thermally induced phase separation (RTIPS) process. BPA-PS was synthesized by click chemistry and was blended to improve the hydrophilicity of PES hollow fiber membranes. The performance of PES/BPA-PS hollow fiber membranes, prepared with different contents of BPA-PS and at different temperatures of the coagulation water bath, was characterized by scanning electron microscopy (SEM), pure water flux (Jw), BSA rejection rate (R), atomic force microscopy (AFM), X-ray photoelectron spectroscopy (XPS), Fourier transform infrared spectroscopy (FTIR) and water contact angle measurements. SEM morphologies revealed that a finger-like cross-section emerged in the hollow fiber membrane by a non-solvent induced phase separation (NIPS) mechanism while a sponge-like cross-section appeared in the hollow fiber membrane via the RTIPS method. Both FTIR and XPS analysis indicated that the sulfate group in BPA-PS was successfully blended with the PES membranes. The results from AFM and water contact angle measurements showed that the surface roughness increased and the hydrophilicity of the PES/BPA-PS hollow fiber membrane was improved with the addition of BPA-PS. The results demonstrated that the PES/BPA-PS membrane with 1 wt% BPA-PS via RTIPS exhibited optimal properties.


Introduction
Owing to its advantages of moderate operation conditions, less energy consumption and environmental friendly properties, membrane separation techniques [1][2][3][4] have become the most popular approach in water purication. Among a variety of membrane preparation methods, phase separation methods, which include the non-solvent induced phase separation method (NIPS) [5][6][7][8] and thermally induced phase separation (TIPS) method, [9][10][11][12] are widely used for membrane preparation with outstanding merits. During the NIPS process, the fast transfer of solvent and non-solvent leads to the phase separation, which is benecial for the formation of a dense skin top surface, nger-like structure and poor mechanical properties. [13][14][15] In the TIPS process, a homogeneous doping solution is prepared at high temperature and spun at room temperature. 16,17 Nevertheless, the application of the TIPS procedure is limited by its requirement for high temperature and fewer options of appropriate diluents.
As a novel method to control phase separation, reverse thermally induced phase separation (RTIPS), which is the lower critical solution temperature (LCST)-TIPS process, 18 combines advantages of both low temperature in dope solutions preparation (similar to NIPS) and the fast rate of heat transfer (similar to TIPS). Compared to NIPS method, the structure of the membrane formed from RTIPS transforms the nger-like structure into sponge-like structure. In the meantime, the dense skin layer turns into the porous layer. The sponge-like structure accelerates the rate of BSA rejection and provides excellent mechanical properties, while porous surface contributes to high pure water ux. Contrast with TIPS, dopes are obtained at room temperature and spun at a temperature higher than cloud point, which could reduce the energy consumption and makes it more conducive to industrialization. In a word, RTIPS process is a cost-effective process and generates less waste. However, due to the membrane is more vulnerable to pollutants, the popularization and development of the membrane technology is severely hampered. In order to improve the antifouling property of the membrane, the modication is necessary. 19 Zhao et al. 20 added HBPE into PSF membrane via RTIPS method, then both the antifouling property and water permeate rate were improved to a great extent. Liu et al. 21 prepared PES/TiO 2 hybrid membrane through sol-gel process by the assistant of RTIPS method. Thus, the antifouling property of the membrane was enhanced with the addition of TiO 2 .
PES hollow ber membranes need to be fabricated to achieve excellent properties. Among the modied methods, blending approach is a simple and useful route to prepared PES membrane. Daraei et al. 22 mixed various types of polymers to modify multi-walled carbon nanotubes to improve the antifouling capability of PES membrane. Li et al. 23 added SPES into PES hollow ber membrane to improve the anti-pollution properties. However, the process and mechanism are more complex to form a membrane in these approaches. Click chemistry offers a terric way for membrane modication. [24][25][26][27][28] The prominent merits of click chemistry are its capacity to supply superior site selectivity, which can provide almost quantitative conversion under mild condition and with almost no side reaction or by-product. [29][30][31] Ge et al. 32 constructed an ionic highway in the anion-exchange membrane by click chemistry and synthesized a novel AEMs material. Zheng et al. 33 developed an antifouling silicon surface with improving the hemocompatibility by click chemistry, and found that the antifouling property enhanced in an articial kidney equipment. Besides, the modication of polyethersulfone membrane with BPA-PS by RTIPS method has not been reported in the literature. Thus, a novel approach was proposed to introduce the click sulfuric acid functional group into polyethersulfone membrane. It is very interesting to point out that the click sulfuric acid functional group in bisphenol A poly sulfuric acid (BPA-PS) 34 may facilitate the antifouling performance in the polyethersulfone membrane. As a new antifouling material, BPA-PS (synthesized by click reaction) not only displays excellent antifouling property against protein fouling but also provides an effective approach to modify the polyethersulfone membrane.
To investigate the polyethersulfone membrane modied by BPA-PS further, the effects of coagulation water bath temperature and the quantity of BPA-PS on membrane formation mechanism, membrane structure, membrane ltration performance as well as fouling resistance were investigated in this work.

Materials
Polyethersulfone (PES, M w ¼ 45 000) was purchased from BASF Co. Ltd. (German). PES was used for membrane preparation aer dried about 24 h at 333 K. N,N-Dimethyl acetamide (DMAC), diethylene glycol (DEG) were provided by Shanghai Chemical Reagent Co. Ltd. Bisphenol sulfuric acid (BPA-PS, M w ¼ 97 000) was provided by Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences. Bovine serum albumin (BSA, M w ¼ 67 000) was provided by Shanghai Lianguan Biochemical Engineering Co. Ltd. All of the pure water in this work was self-made.

Characterization of BPA-PS
The chemical structure of BPA-PS was characterized by FTIR (Nicolet 6700, Thermo Electron Scientic Instruments Corp.).

Preparation of dope solutions
The solvent (DMAc) and a certain amount of BPA-PS were premixed rstly in a conical ask about several hours, then measured amount of PES was dissolved in the conical ask at room temperature (about 293 K), nally, non-solvent (DEG) was added in the solution. The dope solutions degassed about 12 h at room temperature under atmospheric pressure for spinning. The composition of the dopes and the coagulation water bath temperatures were listed in Table 1 and 2, respectively.

Preparation of hollow ber membrane
The preparation process of the PES hybrid hollow ber membrane by wet-spinning method was illustrated in Fig. 1. Both the internal and external coagulation water bath were deionized water, the temperatures of the coagulation water bath were listed in Table 2. The dope solution ow rate, bore uid rate and spinning ow rate were constant. These prepared membranes were immersed in deionized water about 3 d which were refreshed every day to remove the remnant solvent in the hollow ber membranes, then the hollow ber membranes were immersed in glycerol aqueous solution 3 d and were dried in air at room temperature for testing.

Measurement of light transmittance
The experiment of light transmittance was operated via a selfmade device. 35 A collimated laser was directed into a glass sheet which was immersed in water bath rapidly at xed temperature. The message of light intensity was captured via a light detector and was recorded in computer. The precipitation rate of dopes was characterized by the curve of light transmittance to immersion time.

Cloud point measurement
The cloud point measurement was performed in order to conrm the phase separation temperature (i.e., LCST, which was dened as cloud point) of dopes. 18 The cloud points of dopes were measured via a self-made device. Briey, the dope solution was inserted between two cover slips which were heated on a hot stage (KEL-XMT-3100, Shanghai Weitu Optics and Electron Technology Co. Ltd.) from 293 K to 373 K at 1 K min À1 . A collimated laser was directed on coverslips, then the light intensity, which was captured by a light detector, was recorded on the computer. The intensity of transmitted light decreased rapidly with the phase separation occurred. The change of the signal was considered as the indication of cloud point.

SEM measurement
Morphologies (top surface and cross section) of hollow ber membranes were visualized via a scanning electron microscope (SEM; Nova NanoSEM, USA). The cross-sections of the membranes were broken in liquid nitrogen. Each sample was coated with gold in vacuum environment.

Permeation measurement
The pure water ux (J w ) and BSA rejection rate (R) of hollow ber membranes were measured by a self-made device. 36 All tests were executed at room temperature along with a successive pressure (0.1 MPa). The J w , R of the hollow ber membranes, which were evaluated precisely using deionized water and BSA aqueous solution (300 mg L À1 ), were calculated based on the following eqn (1) and (2), respectively.
where J w is the pure water ux (L m À2 h À1 ), A is the effective membrane area (m 2 ), Q is the permeate volume (L), t is the permeate time (h), R, C p and C F were the BSA rejection rate (%), permeate and feed solutions (wt%), respectively.

Porosity and size measurement
The membrane porosity 3 (%) was evaluated by gravimetric method using the following equation: where 3, W 1 , W 2 , r 1 and r 2 are the porosity of the membrane, wet membranes weight, dry membranes weight, water density and the polymer (PES) density, respectively. Mean pore radius r m (mm) of the membrane was dened according to the formula of Guerout-Elford-Ferry: 36 where h is the viscosity of water (8.9 Â 10 À4 Pa s À1 ), h is the thickness of the membrane (mm), DP is the operate pressure (0.1 MPa). Maximum pore size r max (mm) of the membrane could be obtained through bubble point procedure referring to the Laplace's equation: 37 where s, q, P are the water surface tension (22.8 Â 10 À3 N m À1 ), the contact angle ( ) of the membrane and the minimum bubble point pressure (MPa), respectively.

Water contact angle measurement
The contact angle (q) was acquired by the device (JC2000A, Shanghai Zhongcheng Digital Equipment Co. Ltd., China) at room temperature. Briey, the camera capturing image obtained when droplet fell on the surface of the membrane as well as sample contact angle was recorded in computer. Every sample was measured four times and then averaged.

AFM measurement
Atomic force microscopy (Veeco, NanoscopeIIIa Multimode AFM) with 5 mm Â 5 mm scanning size was employed to examine the roughness of the membrane. The AFM gure was managed by Gwyddion soware, moreover, the roughness of the membrane was analysed by NanoScope analysis soware.

XPS measurement
X-ray photoelectron spectroscopy analysis (XPS; VG Microlab II, UK) was used for determining chemical composition of the membrane.

Antifouling measurement
A dead end ltration experiment was performed to examine the long-term antifouling property of the membrane further. 38 Briey, membrane pure water ux (J w1 , L m À2 h À1 ) was obtained at 0.1 MPa. Aerwards, BSA solution (300 mg L À1 ) was permeated through the tested membrane to achieve the BSA ux (J p1 , L m À2 h À1 ) followed by washing the tested membrane by chemical process. The experiment repeated twice. The FRR (%) of the membrane, which obtained according to the permeation uxes, was dened by the following formula (6). The pure water ux and BSA ux with time also obtained to illustrate the antipollution properties of the membranes further. 3 Results and discussion

Characterization of BPA-PS
The synthetic pathway and the FTIR spectrum of BPA-PS were displayed in Fig. 2 and 3, respectively. In Fig. 2, the click reaction occurred between the -SO 3 F and -OTBS at 120 C about 1.5 h, then the -SO 4generated. As shown in Fig. 3, the absorption peak at 1150.1 cm À1 represented the symmetrical stretching vibrations of the S]O in the sulfuric acid group. In addition, the absorption peak at 878.5 cm À1 represented the stretching vibrations of the C-O-S in the sulfuric acid group. These results indicated that BPA-PS molecules had a high hydrophilicity group.

Cloud point
For the LCST dope solution which composed of PES/BPA-PS/ DMAc/DEG, phase separation occurred when the temperature of the dope solution increased to the cloud point.

Cloud point
For the LCST dope solution which composed of PES/BPA-PS/ DMAc/DEG, phase separation occurred when the temperature of the dope solution increased to the cloud point. The dope solution was thermodynamically stable at room temperature, while the interaction between the mixed polymer and solvent became increasingly weak with the increased coagulation bath temperature. Thus, the phases separated when the temperature reached cloud point. The inuence of the BPA-PS from 0 wt% to 2 wt% on the cloud point was not obvious, and the cloud point was about 313 K.  Fig. 4 indicated that the viscosity increased and exhibited a shear-thinning phenomenon in the presence of BPA-PS. This phenomenon became more pronounced when the mass ratio of BPA-PS increased from 0 wt% to 2 wt%. Moreover, Fig. 4 illustrated that the transmittance curve descended speedily at the initial stage and was accelerated with the addition of BPA-PS. In general, high viscosity 39 would decrease the precipitated speed, however, the rate of phase separation was accelerated by the increased viscosity. This phenomenon could be explained through that the membranes formation mechanism was dominated via RTIPS process when the coagulation bath temperature was higher than the cloud point, and the speed of heat transfer was much higher than the mass transfer under the RTIPS mechanism. For this reason, the transmittance curve descended speedy at the initial stage and was accelerated with the increased BPA-PS.

XPS analytical result of PES/BPA-PS hollow ber membrane
XPS results of MBPA PS-1-50 and MBPA PS-3-50 were illustrated in Fig. 5 and Table 3. There were three broaden peaks at 172 eV, 296 eV and 536 eV in the blank PES membrane (Fig. 5a) which represented S2p, C1s and O1s regions, respectively. O1 score-level spectrum of the membrane included two peaks at 531.8 eV and 533.3 eV, which corresponding represented O]S and O-C. It was worth noting that a new peak appeared at 532.3 eV, which represented that O-S appeared in the PES/BPA-PS membrane (Fig. 5d). The appearance of the new peak showed that the PES/ BPA-PS membranes had already blended with BPA-PS.

Membrane structure of PES/BPA-PS hollow ber membrane
SEM images of the membranes were exhibited in Fig. 6 and 7, respectively. The working mechanism of the dopes underwent NIPS process when the coagulation bath temperature was lower than the cloud point while RTIPS took an effective effect when coagulation bath temperature was higher than the cloud point. As illustrated in Fig. 6, the nger-like structure could be found in the MBPA PS-3-20 (by NIPS) while the sponge-like structure was shown in MBPA PS-3-50 (by RTIPS). Owing to the common Fig. 2 The synthetic route of BPA-PS by click chemistry. Fig. 3 The FTIR spectra of BPA-PS. On the other hand, as it widely known that the speed of heat transfer is much higher than the rate of mass transfer, then the heat transfer is the major driving force in membrane formation procedure. Thus, when the coagulation bath temperature is higher than the cloud point (RTIPS process), the nger-like structure turns into sponge-like structure in MBPA PS-3-50. As shown in Fig. 7, the cross-sections of the membranes varied with the addition of BPA-PS when coagulation water bath temperature was xed at 323 K. MBPA PS-1-50 showed the macroporous structure (MBPA PS-1-50-a) and dense crosssection (MBPA PS-1-50 b) under SEM images which indicated the rapid exchange of solvent and non-solvent during the membrane formation procedure. To investigate the inuence of BPA-PS on the formation of various morphologies, SEM photographs, as shown in Fig. 7, turned dense cross-section and macroporous structure into sponge-like structure with the addition of BPA-PS. This could be explained by the increased viscosity of BPA-PS, which delayed the speed of phase separation and then sponge-like structure appeared in the membranes. Stratication in the middle of the membrane was ascribed to that the pure water transferred from interior and exterior to middle layer of the membrane simultaneously with fast transfer speed, 40,41 and then led to pure water accumulated in the middle layer of the membrane. Finally, macroporous structure appeared in the cross-section of the membrane.    The porosity and pore size of the membranes were listed in Table 4, which varied with the addition of BPA-PS and membrane formation mechanism (NIPS or RTIPS). The porosity of the membranes MBPA PS-1-50, MBPA-PS-2-50, MBPA PS-3-50, MBPA PS-4-50 and MBPA PS-5-50, which prepared by RTIPS method, exhibited an increase rst and then decreased trend with the addition of BPA-PS. Moreover, the porosity of the membranes with BPA-PS was higher than that of the blank PES membrane obviously. With regard to the membrane MBPA PS-3-20 and MBPA PS-3-50, the porosity increased when the formation mechanism of the membrane changed from NIPS to RTIPS. The porosity in Table 4 was uniform with the pure water ux and the BSA rejection rate in Fig. 10 and 11. As for the MBPA PS-3 series, the maximum pore size and mean pore size of the MBPA PS-3-50 (produced from RTIPS) were higher than the MBPA PS-3-20 (produced from NIPS), and this result matched with the SEM photographs in Fig. 6.    Table 5 exhibited the three-dimensional AFM images and the roughness of the membranes. In the scan area 5 mm Â 5 mm, the R a of MBPA PS-1, MBPA PS-2, MBPA PS-3, MBPA PS-4 and MBPA PS-5 were 10.5 nm, 13.5 nm, 17.3 nm, 22.7 nm and 21.5 nm, respectively. The AFM image of the blank membrane, as illustrated in Fig. 8, was comparatively smooth. While the AFM images of the membrane became coarser with the increased BPA-PS. The anti-fouling property improved with the roughness because the pollutant accumulated in "Valley" in the membrane surface. BPA-PS, which increased the hydrophilicity of the membrane, also enhanced the adsorbance of contaminants on membrane than the blank PES membrane, and then improved the antifouling property of the membrane.
The water contact angle was metered by trailing the top surface of the membrane. As illustrated in Fig. 9, the static water contact angles of MBPA PS-1-50, MBPA PS-2-50, MBPA PS-3-50, MBPA PS-4-50 and MBPA PS-5-50 were 92.6 , 74.9 , 68.6 , 64.1 and 61.6 , respectively, which exhibited a decreasing trend with the increased BPA-PS. The presence of the hydroxyl group in BPA-PS and the transformation of BPA-PS to the top surface could explain the decreasing trend. As can be known that lower water contact angle is beneted for excellent anti-fouling property. 42 Therefore, the addition of BPA-PS improved the hydrophilicity of the membrane.   3.9 Permeation performance of PES/BPA-PS hollow ber membrane The effect of BPA-PS and different temperatures of coagulation water bath on pure water ux and BSA rejection rate were displayed in Fig. 10 and 11, respectively. The pure water ux in Fig. 10, which the coagulation water bath temperature was set at 323 K, increased rst and then decreased with BPA-PS. The increasing trend of pure water ux was attributed to the hydrophilicity group in BPA-PS while the decreasing trend was due to the increased viscosity by BPA-PS. These data were consistent with the mean pore size (r m ) listed in Table 4 and water contact angle in Fig. 9. The optimum pure water ux (739 L m À2 h À1 ) and BSA rejection (79.2%) were obtained from the membrane with 1.0 wt% BPA-PS. T20, T50, as shown in Fig. 11, were the coagulation water bath temperature. The phase separation process underwent NIPS method in T20 while the phase separation followed by RTIPS process in T50. It was evident that the pure water ux and BSA rejection rate of MBPA PS-1-50 which prepared by RTIPS method were higher than that of MBPA PS-1-20 by NIPS process. This phenomenon unanimous with that the dense skin top surface and the nger-like cross-section were obtained via NIPS while the porous top surface and the sponge-like cross-section were acquired by RTIPS. Porous top surface and sponge-like structure were conducive to high pure water ux and BSA rejection rate, respectively. This phenomenon was a convincing evidence for that RTIPS mechanism had the potential for high pure water ux and BSA rejection rate.

Antifouling property of PES/BPA-PS hollow ber membrane
The FRR, pure water ux and BSA ux with time of the membranes were shown in Fig. 12 and 13, respectively. For the membranes prepared by RTIPS process, the FRR of the membranes with BPA-PS (MBPA PS-3-50, MBPA PS-5-50) was higher than the pure PES membrane (MBPA PS-1-50). This phenomenon was coincident with the test results of water contact angle and AFM. As for the changes of pure water ux and BSA ux with time, it decreased in the initial stage then Fig. 9 The static water contact angles of PES/BPA-PS hollow fiber membrane. Coagulation temperature was 333 K.      Fig. 12 and 13, displayed a faster descend trend than the MBPA PS-3-50 and MBPA PS-5-50. These changes illustrated the better antifouling performance of the membranes with BPA-PS.

Comparison with other membranes
As listed in Table 6, the properties of the membranes in this paper were compared with other membranes in the literature. Both the high water ux and BSA rejection could be acquired via RTIPS ( Fig. 10 and 11). Furthermore, the anti-fouling properties were improved by the addition of BPA-PS.

Conclusion
This paper presented that a novel polyethersulfone (PES) hollow ber membrane was modied by the addition of bisphenol sulfuric acid (BPA-PS) which was fabricated via reverse thermally induced phase separation (RTIPS) process. Bisphenol sulfuric acid (BPA-PS) was synthesized by click chemistry and was blended to improve the hydrophilicity of PES hollow ber membranes. The inuences of different ratios of BPA-PS and cloud point on the PES membrane properties (including penetration performance, membrane morphology and anti-fouling property etc.) were investigated. The cloud point, which was used to explore the temperature of the phase separation, changed a little bit with the addition of BPA-PS. Nevertheless, the water contact angle displayed the decreasing trend from 92.6 to 61.6 with the addition of BPA-PS. The ux of pure water (J w ) and BSA rejection rate (R) of the membranes increased rst and then decreased with the increased BPA-PS. And the optimal pure water ux (739 L m À2 h À1 ) and BSA rejection rate (79.2%) were obtained from the membrane with 1 wt% BPA-PS.
With regarded to the inuence of coagulation water bath temperature on membrane performance, the mechanism of membrane formation transformed NIPS to RTIPS when the temperature of coagulation water bath increased to the cloud point. In the meantime, the nger-like cross sections changed into sponge-like structures. As for the membranes prepared by RTIPS, the pure water ux and BSA rejection rate were both higher than the membranes achieved from NIPS. This phenomenon conrmed that the membrane produced by RTIPS had better performance than that via NIPS process.
Generally speaking, the PES/BPA-PS membranes, which prepared by reverse thermally induced phase separation (RTIPS), displayed excellent properties on pure water ux, BSA rejection rate and anti-pollution performance etc. This research offered a novel approach to combine reverse thermally induced phase separation (RTIPS) and click chemistry on the PES/BPA-PS hollow ber membrane preparation.

Conflicts of interest
There are no conicts to declare.