Polysulfone hemodiafiltration membranes with enhanced anti-fouling and hemocompatibility modified by poly(vinyl pyrrolidone) via in situ cross-linked polymerization

Materials Science and Engineering C 74 (2017) 159–166

Contents lists available at ScienceDirect

Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec

Polysulfone hemodiafiltration membranes with enhanced anti-fouling and hemocompatibility modified by poly(vinyl pyrrolidone) via in situ cross-linked polymerization Lijing Zhu ⁎, Haiming Song, Jiarong Wang, Lixin Xue ⁎ Polymer and Composite Division, Key Laboratory of Marine Materials and Related Technologies, Zhejiang Key Laboratory of Marine Materials and Protective Technologies, Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, PR China

a r t i c l e

i n f o

Article history: Received 20 September 2016 Received in revised form 15 December 2016 Accepted 6 February 2017 Available online 07 February 2017 Keywords: Polysulfone hemodiafiltration membranes In situ cross-linked polymerization Hemocompatibility Anti-fouling Poly(vinyl pyrrolidone)

a b s t r a c t Poly(vinyl pyrrolidone) (PVP) and its copolymers have been widely employed for the modification of hemodiafiltration membranes due to their excellent hydrophilicity, antifouling and hemocompatibility. However, challenges still remain to simplify the modification procedure and to improve the utilization efficiency. In this paper, antifouling and hemocompatibility polysulfone (PSf) hemodiafiltration membranes were fabricated via in situ cross-linked polymerization of vinyl pyrrolidone (VP) and vinyltriethoxysilane (VTEOS) in PSf solutions and non-solvent induced phase separation (NIPS) technique. The prepared membranes were characterized by attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR), X-ray photoelectron spectroscopy (XPS) and scanning electron microscopy (SEM), which suggested that VP and VTEOS have been cross-linked copolymerized in PSf membranes. The modified PSf membranes with high polymer content showed improved hydrophilicity, ultrafiltration and protein antifouling ability. In addition, the modified PSf membranes showed lower protein adsorption, inhibited platelet adhesion and deformation, prolonged the activated partial thromboplastin time (APTT), prothrombin time (PT), and decreased the content of fibrinogen (FIB) transferring to fibrin, indicating enhanced hemocompatibility. In a word, the present work provides a simple and effective one-step modification method to construct PSf membranes with improved hydrophilicity, antifouling and hemocompatibility. © 2017 Elsevier B.V. All rights reserved.

1. Introduction In recent years, polysulfone (PSf) membrane has been widely applied in biotechnological applications, such as hemodiafiltration and protein separation, due to its excellent mechanical strength, chemical resistance, chemical and thermal durability [1–3]. However, the antifouling and hemocompatibility of the hydrophobic PSf membrane are not ideal [4–6]. Membrane fouling is often initiated by the accumulation/aggregation of suspended solids, colloidal particles and biomolecules on the surfaces and pores of the separation membranes, resulting in an increase in resistance to permeate flow [7,8]. In addition, the adsorption and deposition of proteins and platelets often cause thrombus formation and blood coagulation [9]. Therefore, it is necessary to improve the antifouling and hemocompatibility of PSf membrane applied in hemodiafiltration. Hydrophilic modification is an effective and conventional method to enhance antifouling and hemocompatibility of the polymeric membranes. A water layer formed on the hydrophilic surface can inhibit ⁎ Corresponding authors. E-mail addresses: [email protected] (L. Zhu), [email protected] (L. Xue).

http://dx.doi.org/10.1016/j.msec.2017.02.019 0928-4931/© 2017 Elsevier B.V. All rights reserved.

the adsorption and adhesion of pollutants, proteins and platelets via repulsive hydration force, thus reducing fouling and improving hemocompatibility [10–13]. Extensive hydrophilic polymers and anticoagulants, such as poly(ethylene glycol) (PEG), poly(vinyl pyrrolidone) (PVP) and polyzwitterions, have been implemented to improve the fouling and hemocompatibility of the hydrophobic polymeric membrane [4,14–20]. PVP, a non-ionic water soluble polymer, has received much attention due to its good chemical stability, physiological inertness and biocompatibility [21,22]. It was initially used as a blood plasma substitute. Up to date, PVP has been employed as a hydrophilic agent to modify polymeric membranes. Blending, surface coating and grafting methods have been reported to modify polymeric membranes [23–26]. Compared to other approaches, blending is an easier and more convenient method to improve the membrane surface and internal pore walls [27]. However, the elution of PVP from the blend membranes is almost unavoidable, leading to deterioration in long-term durability [28,29]. In order to avoid the elution of PVP, PVP-based amphiphilic copolymers, such as poly(styrene-acrylic acid-vinylpyrrolidone) (P(St-AA-VP)) [30], PVP-bpoly(methyl methacrylate)-b-PVP (PVP-b-PMMA-b-PVP) [22,31–34], PVP-b-poly(acrylate-g-poly(methyl methacrylate))-b-PVP (PVP-b-

160

L. Zhu et al. / Materials Science and Engineering C 74 (2017) 159–166

P(AE-g-PMMA)-b-PVP) [28], poly(styrene-acrylic acid)-b-PVP-b-poly(styrene-acrylic acid) (P(St-AA)-b-PVP-b-P(St-AA)) [35], were synthesized and developed as the blend additives. In the amphiphilic copolymers, the hydrophilic PVP chains migrate to the membrane surface during the phase separation process to prevent the adsorption and adhesion of proteins and platelets, meanwhile, the hydrophobic chains mingle in the membrane bulk materials. As a result, the blend membranes with amphiphilic copolymers as additive exhibited enhanced antifouling and good hemocompatibility properties and longterm durability. However, the synthesis and purification of the amphiphilic copolymers often require multi-step procedures, increasing membrane fabrication costs and limiting their practical application. Therefore, it should be to simplify the modification procedure and to improve the utilization efficiency. The in situ polymerization is an easy and effective method for the modification of polymeric membranes. Zhang et al. synthesized poly(polyethylene glycol monomethyl ether methyl methacrylatemethyl methacrylate) (P(PEGMA-MMA)) and poly(polyethylene glycol monomethyl ether methyl methacrylate-polytetrahydrofuran dimetha crylate ester) (P(PEGMA-PTMGDA)) in the poly(vinylidene fluoride) (PVDF) solutions via free radical polymerization to modify PVDF membranes with superior mechanical behaviors, enhanced antifouling properties, narrowly distributed pore size and molecular weight cut off [36,37]. However, the stability of the modified membranes might be insufficient, because crosslinking agent was not added in the polymerization. As a result, the in situ cross-linked polymerization was developed and has been used to modify polyethersulfone (PES) membranes [38–40]. Compared with the conventional blending, the in situ cross-linked polymerization approach improves the modification efficiency, avoids post-synthetic treatments and reduces the overall costs. In this study, we aim to enhance the anti-fouling and hemocompatibility properties of PSf membranes through in situ crosslinked polymerization. Generally, VP and vinyltriethoxysilane (VTEOS) were polymerized and cross-linked in PSf solutions. After being vacuum degassed, the obtained solution was directly used to fabricate membranes via non-solvent induced phase separation (NIPS) technique. The morphology, surface chemistry and hydrophilicity of the prepared membranes were investigated in detail. The ultrafiltration experiments for pure water and bovine serum albumin (BSA) solution were applied to characterize the fouling resistance property. In addition, BSA adsorption, platelet adhesion and morphology, activated partial thromboplastin time (APTT), prothrombin time (PT) and the content of fibrinogen (FIB) transferring to fibrin were measured to detect the hemocompatibility of the obtained membranes.

2. Experimental 2.1. Materials and reagents Polysulfone (PSf, S6010) was bought from BASF SE. Vinyltriethoxysilane (VTEOS), N-vinyl pyrrolidone (VP), azobisisobutyronitrile (AIBN), bovine serum albumin (BSA) and lysozyme (Lyz) were purchased from Aladdin, China. VTEOS and VP were purified with basic alumina and activated carbon before use, respectively. AIBN was recrystallized from ethanol. Platelet-rich plasma (PRP) and platelet-poor plasma (PPP) were supplied by the Blood Center of Ningbo, China. Activated partial thromboplastin time (APTT), Prothrombin time (PT) and The content of fibrinogen (FIB) transferring to fibrin reagent kits were purchased from Shanghai Sun Biotech Co., Ltd., China. All other regents, such as N,N′-dimethylacetamide (DMAc) and ethanol, were brought from Sinopharm Chemical Reagent Co., Ltd., China.

2.2. Membrane preparation The PSf membranes were prepared via the in situ cross-linked polymerization and the classical non-solvent induced phase separation (NIPS) technique. In a typical procedure, PSf (16 g), VP, VTEOS and AIBN (2 mol% with respect to VP and VTEOS) were dissolved in DMAc under mechanical stirring and N2 protection to get a transparent reaction solution. The total weigh of the reaction solution was 99 g, and the malar ration of VP to VTEOS was locked at 2.5. Then the polymerization was carried out at 80 °C for 12 h with stirring and N2 atmosphere, and stopped by quenching in ice water. Subsequently, acetic acid solution (1 mL) was added into the obtained solution under stirring for 2 h to improve the hydrolysis of the ethoxysilyl groups (Si\\O\\C2H5) and subsequent condensation of the silanol groups (Si\\OH). After being vacuum degassed, the casting solution was spread onto a glass plate and immersed into a water bath at 30 °C for liquid-liquid phase separation. Then the solid membrane was immersed in distilled water at 80 °C for 24 h to thoroughly remove unreacted monomers and solvent, and further improve the cross-linking polymerization. The fabricated PSf microporous membranes with a thickness of 68 ± 5 μm were dried via freeze drying and designated as M0, M1.5, M3, M4.5, M6 and M7, respectively. The numbers denoted the corresponding weight percentages of monomers of VP and VTEOS in the casting solution. The in situ cross-linked polymerization of VP and VTEOS in PSf solution was presented in Scheme 1.

Scheme 1. In situ cross-linked polymerization of VP and VTEOS in PSf solution.

L. Zhu et al. / Materials Science and Engineering C 74 (2017) 159–166

161

Fig. 1. (A) XPS wide-scans and (B) ATR-FTIR spectra of the PSf membranes.

2.3. Structures of the membranes Membrane morphologies were obtained with a field-emission scanning electron microscope (FESEM, Hitachi S-4800, Japan). Before the observation, the samples were sputtered with platinum. The composition of the membrane surfaces was characterized by X ray photoelectron spectrometer (XPS, Shimadzu Axis Utltradld,

Japan) with a take-off angle of 45°and attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR, Thermo-Nicolet 6700,US) between 4000 and 650 cm− 1. Zeta potential (ζ) of the membranes was measured by a SurPASS (Anton Paar GmbH, Austria) electrokinetic analyzer (1 mmol/L KCl solution, 25 °C). Pore size of the obtained PSf membranes was detected by a liquidliquid porometer (LLP-1200A, Porous Materials Inc. US). Porosity of

Fig. 2. SEM images of the PSf membranes (M0, M1.5, M3 and M6) (a: cross-session; b: enlarged cross-session; c: top surface).

162

L. Zhu et al. / Materials Science and Engineering C 74 (2017) 159–166

the obtained membranes was characterized by the liquid displacement method [41]. Briefly, a dry membrane sample was immersed into 1-butanol for 30 min and weighted after whipping off excess surface liquid. Then porosity of the sample was calculated using Eq. (1).

porosity ¼


mwet −mdry =ρ
 100% mwet −mdry =ρ þ mdry =ρPSf

ð1Þ

where mdry (g) and mwet (g) are the weight of a membrane sample before and after immersing into 1-butanol. ρ (g/cm 3) and ρPSf (g/cm3) are the density of 1-butanol and PSf, respectively.

Fig. 3. Pore size and porosity of the modified PSf membranes.

2.4. Hydrophilicity measurements measurements. The hydrophilicity/hydrophobicity of the membranes was characterized by a contact angle measurement (OCA20, Dataphysics, Germany) under the sessile mode. In a typical procedure, a deionized water droplet (1 μL) was dropped onto a membrane sample (1.0 × 5.0 cm2) attached on a glass slide. The water contact angle varying with time was auto calculated and recorded by the software of the equipment. At least five measurement results were averaged to obtain a reliable value. 2.5. Filtration and protein antifouling performances assessments Protein antifouling ability of the membranes was assessed by the ultrafiltration process of pure water and protein (BSA or Lyz) solution through the samples using a stirred filtration cell (Millipore Corporation, XFUF04701, US) at 37 °C. In the typical procedure, a membrane plate was first compacted by pure water at a pressure of 0.1 MPa. The liquid was collected, and the flux was calculated consecutively until the value was stable, and the stable water flux was denoted as Jw. Then the feed solution was switched to BSA solution (1 mg/mL in PBS buffer solution, pH 7.4) or Lyz solution (0.3 mg/mL in PBS buffer solution, pH 7.4), the protein solution ultrafiltration experiment was started using the same process, and the stable protein solution flux was named as Jprotein. At the same time, the protein rejection (Rprotein, %) was calculated using Eq. (2). After protein solution filtration, the membrane was taken out and washed with PBS buffer solution (pH 7.4) for 24 h, and the pure water flux was measured again and denoted as J wr . The flux (J, L m − 2 h− 1 ) was calculated according to Eq. (3). All given data were averaged from three different measurements.

Rprotein ¼

J¼

V St

  C permeate  100% 1− C feed

ð2Þ

FRRprotein ¼

J wr  100% Jw

ð4Þ

2.6. Protein adsorption Typically, a membrane sample was first wetted in physiological saline, then immersed into BSA solution (10 mL, 1 mg/mL in PBS buffer solution, pH 7.4) or Lyz solution (0.3 mg/mL in PBS buffer solution, pH 7.4) at 37 °C for 6 h under stirring to reach the adsorption equilibrium. The sample was carefully rinsed with PBS buffer solution (pH 7.4) to remove unstable protein. The amount of adsorbed protein (Aprotein, μg/cm2) on membrane sample was calculated by the following Eq. (5). Aprotein ¼

C before V before −C after V after S

ð5Þ

where Cbefore (mg/mL) and Cafter (mg/mL) represent the protein concentration before and after treating with membrane sample, Vbefore (mL) and Vafter (mL) represent the volume of protein solution before and after treating with membrane sample, S (12.5 cm2) is the membrane sample area. The given value was averaged from three different measurements. 2.7. Platelet adhesion Typically, a round membrane sample (d = 1 cm) was immersed in physiological saline 37 °C for 2 h. Then the physiological saline was switched to 100 μL platelet-rich plasma (PRP) at 37 °C. After 1 h, the PRP was removed, and the membrane was washed with physiological saline and immersed into 5 wt% glutaraldehyde solution for at 4 °C for

ð3Þ

where C permeate (mg/mL) and C feed (mg/mL) represent the protein concentrations in the permeate and the feed liquid, respectively. They were determined with a UV–vis spectrophotometer (Lambda 950, Perkin-Elmer, America) at wavelength of 280 nm and calculated according to a standard curve. V (L) is the volume of the collected liquid, S (18.1 cm2) is the sample area and t (h) is the filtration time, respectively. In order to evaluate the antifouling ability of the membranes, flux recovery ratio (FRRprotein, %) was defined and calculated by the following Eq. (4). All given data were averaged from three different

Fig. 4. The typical curves of water contact angle decaying with drop age for the PSf membranes.

L. Zhu et al. / Materials Science and Engineering C 74 (2017) 159–166

163

Fig. 5. (A) Flux recovery ratio (FRRprotein) and (B) adsorbed protein (Aprotein) of the PSf membranes.

12 h. Finally, the sample was washed with ultrapure water and dehydrated using a serial of graded ethanol/water solutions (10, 30, 50, 70, 90 and 100 wt%, respectively). The dried membranes were observed with SEM (Hitachi S-4800, Japan) after sputtering a platinum layer. 2.8. Activated partial thromboplastin time (APTT) A membrane sample (1.0 × 1.0 cm2) was equilibrated in physiological saline at 37 °C for 1 h. Then the sample was taken out and put into a test cup. Subsequently, fresh PPP (100 μL) was added, followed by the addition of APTT agent (100 μL, 37 °C). After incubating at 37 °C for 5 min, CaCl2 solution (100 μL, 25 mM, 37 °C) was dropped to the test cup. The APTT was recorded with a semi-automated blood coagulation analyzer PUN-2048A (Perlong, China). Each APTT value was averaged from three measurements. 2.9. Prothrombin time (PT) A membrane (1.0 × 1.0 cm2) was first equilibrated in physiological saline, and then put into a test cup. Fresh PPP (100 μL) was added and incubated at 37 °C for 3 min, followed by the addition of PT agent (200 μL, 37 °C). The PT was determined by PUN-2048A coagulation analyzer (Perlong, China). Each PT value was averaged from three measurements. 2.10. The content of fibrinogen (FIB) transferring to fibrin A membrane (1.0 × 1.0 cm2) was equilibrated in physiological saline and put into a test cup. 200 μL fresh PPP diluted with buffer solution by volume ratio of 1:10 was dropped and incubated at 37 °C for 3 min, followed by the addition of thrombin solution (100 μL). The sample

content of FIB was measured by PUN-2048A coagulation analyzer (Perlong, China). Each value was averaged from three measurements.

3. Results and discussion 3.1. Membrane chemistry and morphology In order to detect the surface chemical components of the membranes, XPS was employed. The XPS wide-scans of M0, M3 and M6 are shown in Fig. 1A. Compared with M0, the new peaks of N 1s and Si 2p3 can be observed in the spectra of M3 and M6. In addition, with the increase of the VP and VTEOS additive, the peak intensity of N 1s and Si 2p3 increased. ATR-FTIR was also applied to characterize chemical structures of the PSf membranes, and typical spectra for M0, M3 and M6 are shown in Fig. 1B. Compared with M0, a new peak appearing at about 1660 cm−1 for M3 and M6 is attributed to C_O stretching (amide I band) in pyrrolidone ring of PVP. The new peaks at 3100–3700 and 948 cm− 1 are the absorption of Si\\OH stretching [42]. Both ATR-FTIR and XPS results confirmed that VP and VTEOS have been cross-linked copolymerized in PSf membranes, and with the increase of monomers additive in PSf solution, the amount of P(VP-VTEOS) anchored on the membrane surfaces increased. The SEM images of the top surfaces and cross-sections for M0, M1.5, M3 and M6 are shown in Fig. 2. The cross-sessions morphology of the modified membranes changed obviously. M0 shows a dense skin layer and a porous sub-layer with a finger-like structure, then the porous structure. For the modified PSf membranes, diameter of the finger-like structure increased with the increase of monomers amount, and the skin layer of the modified membranes decreased from 1.4 ± 0.34 (M1.5) to 0.6 ± 0.27 μm (M6). Although, it is difficult to found the change of the membranes surfaces from the SEM images, the pore size and porosity increased slightly with the increase of the copolymers additive as shown in Fig. 3. The results indicated that the cross-sessions

Fig. 6. (A) Pure water flux (Jw) and (B) protein rejection (Rprotein) of the obtained PSf membranes.

164

L. Zhu et al. / Materials Science and Engineering C 74 (2017) 159–166

Fig. 7. SEM images of the adherent platelets on the PSf membrane surfaces before and after modification. Typical morphology of adherent platelets: R-round, D-dendritic, SD-spreading dendritic, S-spreading and FS-fully spreading.

morphology and pore size of the membranes were significantly affected by the copolymers content. The copolymers improved liquid-liquid phase separation and restrained sponge pores formation [43]. 3.2. Membrane hydrophilicity The water contact angles of the PSf membranes are shown in Fig. 4. It can be seen that the initial contact angles of M0 is as high as 75.8°. For the modified PSf membranes, the initial contact angle remarkably decreases from 74.0 (M1.5) to 43.5° (M7). Furthermore, the curves of water contact angle for the modified membranes attenuate faster with drop age than that of M0. The contact angle of M7 decreases by about 12° within 120 s, as contrast, the contact angle of the original PSf membrane changed only about 3.8°. These results suggested that the hydrophilicity of PSf membranes is improved considerably by introduction of P(VP-VTEOS). It is mainly attributed to the PVP chains and the silanol groups. 3.3. Filtration and anti-fouling performances BSA and Lyz were used as the model foulants to investigate the antifouling ability of the prepared PSf membranes. Flux recovery ratio (FRRprotein), one of the most important parameters to discuss membrane fouling, was calculated by the change of the pure water flux caused by the treatment of protein solution in a dynamic ultrafiltration process. As shown in Fig. 5A, FRRprotein increases with the increase of additive concentration, indicating improved antifouling ability in separation and purification processes. It should be pointed out that FRRprotein of M0 was not recorded, because the pure water fluxes of M0 before and after BSA treatment were 0 L m−2 h−1. In order to further investigate PSf membranes fouling, static adsorption of protein on the obtained

PSf membrane surfaces was also investigated. As shown in Fig. 5B, the amount of adsorbed protein (Aprotein) on the modified PSf membranes is much lower than that of M0. Both FRRprotein and Aprotein indicated that the modified PSf membranes with high polymer content had better antifouling performance. It is mainly attributed to the enhanced hydrophilicity of the membranes. When the modified PSf membranes contacted with proteins solution, water molecules immediately permeate the hydrophilic membrane surfaces, proteins retain on the water layer, forming proteins/water/solid composite interface. As a result, the interaction of proteins with the membrane was inhibited and membrane antifouling ability was improved [11]. In addition, electrostatic interaction plays an important role in the membrane fouling. When the pH is 7.4 that simulate the vivo environment, BSA (PI = 4.8) is negatively charged, while Lyz (PI = 10.7) is positively charged [44]. PSf membrane surfaces maintain an overall negative charge (ζ = − 55 mV, Fig. S1), which is beneficial to restrain the adsorption of negative BSA proteins due to the electrostatic repulsion [49]. As a result, the amount of adsorbed BSA on a PSf membrane is lower than that of Lyz as shown in Fig. 5B. In conclusion, membrane fouling can be significantly inhibited by improved hydrophilicity and electrostatic repulsion [45, 46]. In dynamic ultrafiltration process, pure water flux (Jw) and protein rejection (Rprotein) were calculated and shown in Fig. 6. Both Jw and Rprotein of M0 were not recorded due to its compact structure. For the modified PSf membranes, Jw exhibits an upward trend with an increase of the additive addition. It is mainly attributed to the increase of pore size (as shown in Fig. 3), which increased water permeation. Furthermore, the improved hydrophilicity decreased the water permeation resistance. As a result, pure water flux was affected by pore size and hydrophilicity of the membranes.Rprotein changes inversely with Jw. In addition, RBSA of a membrane sample is higher than that of RLyz as

Fig. 8. (A) Activated partial thromboplastin time (APTT), (B) prothrombin time (PT) and (C) the content of fibrinogen (FIB) transferring to fibrin of the PSf membranes.

L. Zhu et al. / Materials Science and Engineering C 74 (2017) 159–166

165

Table 1 Comparison of the properties of M6 with some recently reported different membranes. Membrane

PSf/P(VP-VTEOS) Pure PLA PVA/PAN PES/PVP-f-MWCNT a

Anti-fouling

Hemocompatibility

Permeability-selectivity

ABSA (μg/cm2)

APTT (s)

RLyz (%)

RBSA (%)

Jwa (Lm−2 h−1)

Ref.

10.5 ± 0.4 ~105 28.9 ± 0.3 ~11.5

55.6 ± 1.3 ~45 / /

15.2 ± 0.9 / 54.2 ~77.5

78.0 ± 6.3 67 95 ~90

90.5 ± 16.9 316 290 65.5

M6 [53] [54] [55]

The transmembrane pressure is 0.1 MPa.

shown in Fig. 6B. The trend can be mainly attributed to size exclusion and Donnan effect [47]. In order to remove middle molecules and prevent loss of albumin, M6 was preferred. 3.4. Hemocompatibility It is very important to improve the hemocompatibility of the membranes used in the field of blood-containing applications such as hemodialysis, hemodiafiltration, plasma collection et al. [47,48]. Once blood is contacted with the foreign membrane, the blood proteins and platelets are apt to adsorb and deposit on its surfaces, leading to coagulation pathways and thrombus formation and seriously complications [6,49]. The adsorbed BSA amount on the modified PSf membranes was decreased than that of neat PSf membrane as shown in Fig. 5B and did not discuss again. Here, platelet adhesion and morphology, activated partial thromboplastin time (APTT), prothrombin time (PT) and the content of fibrinogen (FIB) transferring to fibrin of M0 and M6 were investigated. The typical SEM images of the platelets adhered on the surfaces of M0 and M6 are shown in Fig. 7. It can be found that the amount of adhered platelets on M0 surface is much more than that of M6. In addition, these platelets had an irregular shape, such as D-dendritic (early pseudopodia), SD-spreading dendritic (intermediate pseudopodia), Sspreading (hyaloplasm spread between pseudopodia) and FS-fully spreading (hyaloplasm extensively spread) [50], indicating the deformation and activation of platelets. For M6, few platelets are found on the membrane surfaces, and the platelets had a round morphology without pseudopodium and deformation. This should be attributed to the highly hydrophilicity of the modified PSf membrane, which prevented the adhesion and activation of platelets. In a coagulant system, there are three pathways: the intrinsic pathways, the extrinsic pathway, and the common pathway. APTT is a global screening procedure to measure both the intrinsic and the common coagulation pathway, which reflects the level of prothrombin, fibrinogen and blood coagulation factor V, X in plasma in endogenous pathway of coagulation [38,51]. The results of APTT for the PSf membranes before and after modification are shown in Fig. 8A. Compared to M0, the APTT of M6 increased by approximately 10 s. PT is always to measure the extrinsic coagulation pathway, which will suppress the plasma coagulation pathways including factors I, II, V, X, XII [52]. The results of PT for the PSf membranes before and after modification are shown in Fig. 8B. It can be found that the PT of M6 is higher than that of M0 and PPP. FIB is often used to characterize the content of the fibrinogen transferring to fibrin. Fig. 8C shows the FIB values of M0 and M6. The FIB of M6 was about 86.8 mg/dL, which is lower than that M0 and PPP. The phenomena indicated that the modified PSf membrane exhibited enhanced hemocompatibility, which might be attributed to the reaction or combination between the hydrophilic surfaces and the coagulation factors in plasma. 3.5. Comparisons of different hemodiafiltration membranes A detailed comparison of permeability-selectivity, anti-fouling and hemocompatibility among M6 and some different hemodiafiltration membranes are shown in Table 1. M6 has lower ABSA and prolonger

APTT, which indicated that the antifouling and hemocompatibility of the modified PSf membranes were better than that of the other hemodialysis membranes. In addition, M6 is able to remove middlemolecules (lower RLyz) and prevent loss of albumin (higher RBSA), regardless of the lower water flux. Thus, the modified PSf membranes in this work had great potential to be used in hemodiafitration field. And we will further improve the performances of PSf hemodialysis membranes in our future reports. 4. Conclusions PSf membranes with improved antifouling and hemocompatibility were successfully fabricated via a simple process combining in situ cross-linked polymerization and the traditional NIPS technique. Water contact angle results suggested enhanced hydrophilicity of the modified PSf membranes, compared to neat PSf membrane. Dynamic protein filtration experiments confirmed the improved antifouling ability of the prepared membranes. In addition, the results of protein adsorption, platelet adhesion and clotting time indicated that the modified PSf membrane exhibited enhanced hemocompatibility, which might be attributed to the reaction or combination between the hydrophilic surfaces and the coagulation factors in plasma. The modified PSf membranes with enhanced hydrophilicity, antifouling and hemocompatibility properties can be applied in hemodiafitration field. Moreover, the in situ cross-linked polymerization method can be extended to other polymeric materials. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.msec.2017.02.019. Acknowledgements This work is financed by the National Natural Science Foundation of China (No. 51603214), the Ningbo Science and Technology Bureau (No. 2014B81004 and 2015B11041), the National Natural Science Foundation of Ningbo (No. 2015A610029) and the National Natural Science Foundation of Zhejiang Province (No. LQ16E030004). References [1] H.J. Song, C.K. Kim, Fabrication and properties of ultrafiltration membranes composed of polysulfone and poly(1-vinylpyrrolidone) grafted silica nanoparticles, J. Membr. Sci. 444 (2013) 318–326. [2] H.J. Song, Y.J. Jo, S.Y. Kim, J. Lee, C.K. Kim, Characteristics of ultrafiltration membranes fabricated from polysulfone and polymer-grafted silica nanoparticles: effects of the particle size and grafted polymer on the membrane performance, J. Membr. Sci. 466 (2014) 173–182. [3] A.J. Kajekar, B.M. Dodamani, A.M. Isloor, Z.A. Karim, N.B. Cheer, A.F. Ismail, S.J. Shilton, Preparation and characterization of novel PSf/PVP/PANI-nanofiber nanocomposite hollow fiber ultrafiltration membranes and their possible applications for hazardous dye rejection, Desalination 365 (2015) 117–125. [4] Y.F. Zhao, L.P. Zhu, Z. Yi, B.K. Zhu, Y.Y. Xu, Improving the hydrophilicity and foulingresistance of polysulfone ultrafiltration membranes via surface zwitterionicalization mediated by polysulfone-based triblock copolymer additive, J. Membr. Sci. 440 (2013) 40–47. [5] A. Filimon, R.M. Albu, I. Stoica, E. Avram, Blends based on ionic polysulfones with improved conformational and microstructural characteristics: perspectives for biomedical applications, Compos. Part B 93 (2016) 1–11. [6] J. Zhao, X. Zhao, Z. Jiang, Z. Li, X. Fan, J. Zhu, H. Wu, Y. Su, D. Yang, F. Pan, J. Shi, Biomimetic and bioinspired membranes: preparation and application, Prog. Polym. Sci. 39 (2014) 1668–1720.

166

L. Zhu et al. / Materials Science and Engineering C 74 (2017) 159–166

[7] S. Robinson, S.Z. Abdullah, P. Bérubé, P. Le Clech, Ageing of membranes for water treatment: linking changes to performance, J. Membr. Sci. 503 (2016) 177–187. [8] J. Ayyavoo, T.P.N. Nguyen, B.M. Jun, I.C. Kim, Y.N. Kwon, Protection of polymeric membranes with antifouling surfacing via surface modifications, Colloids Surf. A Physicochem. Eng. Asp. 506 (2016) 190–201. [9] C. Zhao, J. Xue, F. Ran, S. Sun, Modification of polyethersulfone membranes – a review of methods, Prog. Mater. Sci. 58 (2013) 76–150. [10] R. Zhou, P.F. Ren, H.C. Yang, Z.K. Xu, Fabrication of antifouling membrane surface by poly(sulfobetaine methacrylate)/polydopamine co-deposition, J. Membr. Sci. 466 (2014) 18–25. [11] G.D. Kang, Y.M. Cao, Application and modification of poly(vinylidene fluoride) (PVDF) membranes – a review, J. Membr. Sci. 463 (2014) 145–165. [12] Y. Xie, S.S. Li, X. Jiang, T. Xiang, R. Wang, C.S. Zhao, Zwitterionic glycosyl modified polyethersulfone membranes with enhanced anti-fouling property and blood compatibility, J. Colloid Interface Sci. 443 (2015) 36–44. [13] L. Zhu, F. Liu, X. Yu, L. Xue, Poly(lactic acid) hemodialysis membranes with poly(lactic acid)-block-poly(2-hydroxyethyl methacrylate) copolymer as additive: preparation, characterization, and performance, ACS Appl. Mater. Interfaces 7 (2015) 17748–17755. [14] S. Zheng, Q. Yang, B. Mi, Novel antifouling surface with improved hemocompatibility by immobilization of polyzwitterions onto silicon via click chemistry, Appl. Surf. Sci. 363 (2016) 619–626. [15] Y.B. Wang, K.H. Shi, H.L. Jiang, Y.K. Gong, Significantly reduced adsorption and activation of blood components in a membrane oxygenator system coated with crosslinkable zwitterionic copolymer, Acta Biomater. 40 (2016) 153–161. [16] X. Huang, W. Wang, Z. Zheng, W. Fan, C. Mao, J. Shi, L. Li, Surface monofunctionalized polymethyl pentene hollow fiber membranes by plasma treatment and hemocompatibility modification for membrane oxygenators, Appl. Surf. Sci. 362 (2016) 355–363. [17] P. Wang, J. Meng, M. Xu, T. Yuan, N. Yang, T. Sun, Y. Zhang, X. Feng, B. Cheng, A simple but efficient zwitterionization method towards cellulose membrane with superior antifouling property and biocompatibility, J. Membr. Sci. 492 (2015) 547–558. [18] A. Venault, M.R.B. Ballad, Y.H. Liu, P. Aimar, Y. Chang, Hemocompatibility of PVDF/ PS-b-PEGMA membranes prepared by LIPS process, J. Membr. Sci. 477 (2015) 101–114. [19] A. Venault, J.R. Wu, Y. Chang, P. Aimar, Fabricating hemocompatible bi-continuous PEGylated PVDF membranes via vapor-induced phase inversion, J. Membr. Sci. 470 (2014) 18–29. [20] G. Yilmaz, H. Toiserkani, D.O. Demirkol, S. Sakarya, S. Timur, L. Torun, Y. Yagci, Polysulfone based amphiphilic graft copolymers by click chemistry as bioinert membranes, Mater. Sci. Eng. C 31 (2011) 1091–1097. [21] J. Jiang, L. Zhu, L. Zhu, H. Zhang, B. Zhu, Y. Xu, Antifouling and antimicrobial polymer membranes based on bioinspired polydopamine and strong hydrogen-bonded poly(N-vinyl pyrrolidone), ACS Appl. Mater. Interfaces 5 (2013) 12895–12904. [22] F. Ran, S. Nie, W. Zhao, J. Li, B. Su, S. Sun, C. Zhao, Biocompatibility of modified polyethersulfone membranes by blending an amphiphilic triblock co-polymer of poly(vinyl pyrrolidone)-b-poly(methyl methacrylate)-b-poly(vinyl pyrrolidone), Acta Biomater. 7 (2011) 3370–3381. [23] D. He, H. Susanto, M. Ulbricht, Photo-irradiation for preparation, modification and stimulation of polymeric membranes, Prog. Polym. Sci. 34 (2009) 62–98. [24] C. Wang, R. Feng, F. Yang, Enhancing the hydrophilic and antifouling properties of polypropylene nonwoven fabric membranes by the grafting of poly(N-vinyl-2-pyrrolidone) via the ATRP method, J. Colloid Interface Sci. 357 (2011) 273–279. [25] N. Gomathi, R. Rajasekar, R.R. Babu, D. Mishra, S. Neogi, Development of bio/blood compatible polypropylene through low pressure nitrogen plasma surface modification, Mater. Sci. Eng. C 32 (2012) 1767–1778. [26] L.J. Zhu, L.P. Zhu, Y.F. Zhao, B.K. Zhu, Y.Y. Xu, Anti-fouling and anti-bacterial polyethersulfone membranes quaternized from the additive of poly(2-dimethylamino ethyl methacrylate) grafted SiO2nanoparticles, J. Mater. Chem. A 2 (2014) 15566. [27] Z. Zhou, S. Rajabzadeh, A.R. Shaikh, Y. Kakihana, W. Ma, H. Matsuyama, Effect of surface properties on antifouling performance of poly(vinyl chloride-co-poly(ethylene glycol)methyl ether methacrylate)/PVC blend membrane, J. Membr. Sci. 514 (2016) 537–546. [28] H. Song, F. Ran, H. Fan, X. Niu, L. Kang, C. Zhao, Hemocompatibility and ultrafiltration performance of surface-functionalized polyethersulfone membrane by blending comb-like amphiphilic block copolymer, J. Membr. Sci. 471 (2014) 319–327. [29] L.J. Zhu, L.P. Zhu, J.H. Jiang, Z. Yi, Y.F. Zhao, B.K. Zhu, Y.Y. Xu, Hydrophilic and antifouling polyethersulfone ultrafiltration membranes with poly(2-hydroxyethyl methacrylate) grafted silica nanoparticles as additive, J. Membr. Sci. 451 (2014) 157–168. [30] C. Cheng, L. Ma, D. Wu, J. Ren, W. Zhao, J. Xue, S. Sun, C. Zhao, Remarkable pHsensitivity and anti-fouling property of terpolymer blended polyethersulfone hollow fiber membranes, J. Membr. Sci. 378 (2011) 369–381. [31] S.Q. Nie, F. Ran, C. He, P.F. Zhao, X.H. Wei, J. Li, C.S. Zhao, Synthesis of amphiphilic triblock copolymer poly(vinylpyrrolidone)-b-poly(methyl methacrylate)-b-

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44] [45]

[46] [47]

[48]

[49]

[50] [51]

[52]

[53]

[54]

[55]

poly(vinylpyrrolidone) for the modification of polyethersulfone membrane, Chin. Chem. Lett. 22 (2011) 370–373. F. Ran, J. Li, Y. Lu, L. Wang, S. Nie, H. Song, L. Zhao, S. Sun, C. Zhao, A simple method to prepare modified polyethersulfone membrane with improved hydrophilic surface by one-pot: the effect of hydrophobic segment length and molecular weight of copolymers, Mater. Sci. Eng. C 37 (2014) 68–75. F. Ran, H. Song, X. Niu, A. Yang, S. Nie, L. Wang, J. Li, S. Sun, C. Zhao, Bionic design for surface optimization combining hydrophilic and negative charged biological macromolecules, Int. J. Biol. Macromol. 67 (2014) 260–269. A. Bera, C.U. Kumar, P. Parui, S.K. Jewrajka, Stimuli responsive and low fouling ultrafiltration membranes from blends of polyvinylidene fluoride and designed library of amphiphilic poly(methyl methacrylate) containing copolymers, J. Membr. Sci. 481 (2015) 137–147. S. Nie, J. Xue, Y. Lu, Y. Liu, D. Wang, S. Sun, F. Ran, C. Zhao, Improved blood compatibility of polyethersulfone membrane with a hydrophilic and anionic surface, Colloids Surf., B 100 (2012) 116–125. P.Y. Zhang, Z.L. Xu, H. Yang, Y.M. Wei, W.Z. Wu, Fabrication and characterization of PVDF membranes via an in situ free radical polymerization method, Chem. Eng. Sci. 97 (2013) 296–308. P.Y. Zhang, Z.L. Xu, X.H. Ma, H. Yang, W.Z. Wu, Y.M. Wei, Y.D. Liu, Fabrication and characterization of PVDF hollow fiber membranes employing in-situ self-assembly modulation concept, J. Membr. Sci. 486 (2015) 119–131. T. Xiang, C.D. Luo, R. Wang, Z.Y. Han, S.D. Sun, C.S. Zhao, Ionic-strength-sensitive polyethersulfone membrane with improved anti-fouling property modified by zwitterionic polymer via in situ cross-linked polymerization, J. Membr. Sci. 476 (2015) 234–242. H. Qin, C. Sun, C. He, D. Wang, C. Cheng, S. Nie, S. Sun, C. Zhao, High efficient protocol for the modification of polyethersulfone membranes with anticoagulant and antifouling properties via in situ cross-linked copolymerization, J. Membr. Sci. 468 (2014) 172–183. Z. Han, C. Cheng, L. Zhang, C. Luo, C. Nie, J. Deng, T. Xiang, C. Zhao, Toward robust pHresponsive and anti-fouling composite membranes via one-pot in-situ cross-linked copolymerization, Desalination 349 (2014) 80–93. Z. Yao, S. Du, Y. Zhang, B. Zhu, L. Zhu, A.E. John, Positively charged membrane for removing low concentration Cr(VI) in ultrafiltration process, J. Water Process. Eng. 8 (2015) 99–107. P. Hajji, L. David, J.F. Gerard, J.P. Pascault, G. Vigier, Synthesis, structure, and morphology of polymer–silica hybrid nanocomposites based on hydroxyethyl methacrylate, J. Polym. Sci. B Polym. Phys. 37 (1999) 3172–3187. S.H. Zhi, J. Xu, R. Deng, L.S. Wan, Z.K. Xu, Poly(vinylidene fluoride) ultrafiltration membranes containing hybrid silica nanoparticles: preparation, characterization and performance, Polymer 55 (2014) 1333–1340. A. Gao, F. Liu, H. Shi, L. Xue, Controllable transition from finger-like pores to interconnected pores of PLLA membranes, J. Membr. Sci. 478 (2015) 96–104. M.L. Marré Tirado, M. Bass, M. Piatkovsky, M. Ulbricht, M. Herzberg, V. Freger, Assessing biofouling resistance of a polyamide reverse osmosis membrane surface-modified with a zwitterionic polymer, J. Membr. Sci. 520 (2016) 490–498. X. Shi, G. Tal, N.P. Hankins, V. Gitis, Fouling and cleaning of ultrafiltration membranes: a review, J. Water Process. Eng. 1 (2014) 121–138. L.J. Zhu, F. Liu, X.M. Yu, A.L. Gao, L.X. Xue, Surface zwitterionization of hemocompatible poly(lactic acid) membranes for hemodiafiltration, J. Membr. Sci. 475 (2015) 469–479. F. Tong, X. Chen, L. Chen, P. Zhu, J. Luan, C. Mao, J. Bao, J. Shen, Preparation, blood compatibility and anticoagulant effect of heparin-loaded polyurethane microspheres, J. Mater. Chem. B 1 (2013) 447–453. M. Irfan, A. Idris, Overview of PES biocompatible/hemodialysis membranes: PESblood interactions and modification techniques, Mater. Sci. Eng. C 56 (2015) 574–592. S.L. Goodman, Sheep, pig, and human platelet-material interactions with model cardiovascular biomaterials, J. Biomed. Mater. Res. 45 (1999) 240–250. L. Li, C. Cheng, T. Xiang, M. Tang, W. Zhao, S. Sun, C. Zhao, Modification of polyethersulfone hemodialysis membrane by blending citric acid grafted polyurethane and its anticoagulant activity, J. Membr. Sci. 405-406 (2012) 261–274. J. Li, F. Liu, X. Yu, Z. Wu, Y. Wang, Z. Xiong, J. He, APTES assisted surface heparinization of polylactide porous membranes for improved hemocompatibility, RSC Adv. 6 (2016) 42684–42692. X. Yu, Z. Xiong, J. Li, Z. Wu, Y. Wang, F. Liu, Surface PEGylation on PLA membranes via micro-swelling and crosslinking for improved biocompatibility/ hemocompatibility, RSC Adv. 5 (2015) 107949–107956. X. Yu, L. Shen, Y. Zhu, X. Li, Y. Yang, X. Wang, M. Zhu, B.S. Hsiao, High performance thin-film nanofibrous composite hemodialysis membranes with efficient middlemolecule uremic toxin removal, J. Membr. Sci. 523 (2017) 173–184. M. Irfan, A. Idris, N.M. Yusof, N.F.M. Khairuddin, H. Akhmal, Surface modification and performance enhancement of nano-hybrid f-MWCNT/PVP90/PES hemodialysis membranes, J. Membr. Sci. 467 (2014) 73–84.