Polyethersulfone-polyvinylpyrrolidone composite membranes:Effects of polyvinylpyrrolidone content and polydopamine coating on membrane morphology,structure and performances

Yanna Wu ,Jianxian Zeng,**,Yajie Zeng ,Hu Zhou ,Guoqing Liu ,Jian Jian ,Jie Ding

1 School of Chemistry and Chemical Engineering,Hunan University of Science and Technology,Hunan Provincial Engineering Research Center for Functional Membranes,Xiangtan 411201,China

2 School of Chemistry and Chemical Engineering,Hunan Normal University,Changsha 410081,China

Keywords:Polyethersulfone membrane Polydopamine coating Hydrophilicity Permeability Antifouling

ABSTRACT Hydrophilic modification is a promising method to inhibit fouling formation on ultrafiltration membrane.In this work,different mass concentrations (1%–16%) of hydrophilic polyvinylpyrrolidone were incorporated into polyethersulfone(PES)membranes fabricated by none-solvent induced phase separation.Then,polydopamine (PDA) coating on the surface of prepared membrane was carried out at pH 8.5.The morphology and structure,surface hydrophilicity,permeation flux,BSA rejection,antifouling and stability performances of PES and PDA/PES modified membranes were investigated in detail.The results indicated that PDA was successfully attached onto the membranes.Membrane hydrophilicity was evaluated by water contact angle measurement.The contact angles of modified membranes reduced remarkably,suggesting that the membrane hydrophilicities were significantly increased.The results of filtration tests,which were done by dead-end filtration of bovine serum albumin solution,showed that the properties of permeability and fouling resistance were obviously improved by PDA modification.When polyvinylpyrrolidone mass content reached 10%,flux recovery ratio of modified membrane was up to 91.23%,and its BSA rejection were over 70%.The results of stability tests showed that the modified membranes had good mechanical stability and chemical stability.This facile fabrication procedure and outstanding performances suggested that the modified membranes had a potential in treating fouling.

1.Introduction

Ultrafiltration (UF) membranes for wastewater treatment have become a relatively mature technology and attracted increasing attention.Polyethersulfone (PES) membrane is often used as one of the materials of ultrafiltration membranes.It has excellent properties such as mechanical stability,chemical stability,heat resistance,corrosion resistance and has been widely applied in many fields such as membrane distillation,gas separation,and pollutant removal,especially water treatment[1–3].The PES polymer is soluble in N-methyl-2-pyrrolidone (NMP),N,N-dimethylacetamide(DMAc),dimethylformamide (DMF) and other some common solvents[4].The PES membranes are often prepared via conventional none-solvent induced phase separation (NIPS) or thermal induced phase separation (TIPS) methods [5,6].However,the inherent hydrophobic nature of PES membranes leads to low water permeability and serious membrane fouling by adsorbing various contaminants [7].Therefore,in order to achieve excellent performance of the membrane,modification is necessary.

Until now,a series of methods have been actualized to suppress the membrane fouling.The blending of hydrophilic polymers is considered as one of the methods to improve the antifouling performance of the membrane. Since polyvinylpyrrolidone (PVP) shows excellent biocompatibility,chemical stability,low toxicity and it can be considered as a pore-former,blending PVP into PES is a good choice to improve the hydrophilicity and antifouling of the membrane.Basri et al.[8] discussed the effect of molecular weight of PVP on the membrane properties,such as contact angle and pore size.It was found that the hydrophilicity increased and pure water flux decreased when higher molecular weight PVP was added to the casting solution.Al Malek et al.[9] reported that the addition of PVP to the casting solution strongly enhanced the permeability of membrane.Gebru et al.[10] proved that the membranes prepared by using PVP additive showed greater macro-voids and finger-like structure compared with the membranes by using poly(ethylene glycol) additive.On the other hand,some researchers studied the variation of thermodynamic and rheologic properties in polysulfone casting solution by adding PVP [11,12].The results showed that when low levels of PVP were added (less than 5.0%,mass),the changes in thermodynamic properties controlled the de-mixing process,resulting in enhanced phase separation.With further increment of PVP content,the de-mixing process was delayed due to the high viscosity of the casting solution [13].Because the reported studies do not provide enough information on the bovine serum albumin pollution mechanism indicators,such as the flux recovery ratio(FRR),reversible fouling (Rr),irreversible fouling (Rir),it is still necessary to further research optimal conditions for the preparation of PES membrane by controlling the concentration of hydrophilic pore formers.In addition,PVP can change the hydrophilicity and permeability of the membrane by controlling the membrane pore structure.If the surface modification,which is done by coating a hydrophilic layer (such as a polydopamine(PDA) layer) onto membrane surface,is combined with the addition of hydrophilic PVP into the membrane,i.e.,a dual hydrophilic modification,the antifouling performance of the membrane would be greatly improved.

PDA has the advantages of strong adhesion and simple preparation process,and is widely used in membrane synthesis and modification [14].For example,the poly(sulfobetaine methacrylate),a commonly used zwitterionic polymer,was successfully grafted from the entrapped PDA in membranes through atom transfer radical polymerization [12].Li et al.[15] prepared super-hydrophilic,excellent antifouling abilities PVDF/PDA hybrid membranes through the generation in situ of bio-inspired polydopamine(PDA) microspheres on PVDF membranes.The results demonstrated that the PVDF/PDA hybrid membrane penetrated more than five times as much water as the pristine PVDF membrane,indicating that there was an accelerated self-driving force in the membrane during the filtration process in the absence of energy input.Although PDA plays an effective role in blending modification,it is easy to be wrapped during membrane formation process.This makes the excellent performance of PDA not fully reflected.However,the membrane modification of surface-coated with PDA is facile and gets ideal performances.Due to the inherent catechol group of dopamine,dopamine can be tightly attached to the membrane surface through complex physical and chemical interactions(hydrogen bonding,chelation,π-π interaction and covalent bonding) [16].Some researchers hypothesized that dopamine molecules had multiple reaction sites,and different cross-linking reactions occurred between the molecules [17].The selfpolymerization of dopamine in alkaline solutions was due to auto-oxidation.It was proposed that the auto-oxidation of dopamine produced free radical O2-.Increasing the pH and dopamine concentration of the solution within a certain range increased the generation rate of free radical O2-[18,19].PDA plays an intermediary role or is considered a surface-modifying chemical in most modification processes.However,to the best of our knowledge,there is no report on the PDA modification of membranes with different structures.

This current study aims to develop a simple and universal strategy for improving the antifouling performance of membranes with hydrophilic additive PVP of different concentrations by PDA coating.The morphology,porosity,surface roughness,chemical structure,water flux and hydrophilic-hydrophobic property were investigated in depth to probe into the effect of PDA coating on the different structural membranes.Stability and antifouling tests were also conducted to evaluate the membrane performance.

2.Experimental

2.1.Materials

Polyethersulfone (PES Ultrason E6020P) was dried overnight in vacuum oven prior to use.Polyvinylpyrollidone(PVP,Mw,40 K),Nmethyl-pyrolidone (NMP),Dopamine hydrochloride (DA,≥98%)and Bovine serum albumin (BSA) were purchased from Aladdin Industrial Corporation (Shanghai,China).Deionized (DI) water(Resistivity,18.3 MΩ.cm) was generated by our laboratory ultrapure water system.All other chemicals and reagents were of analytical grade and were used without further purification unless otherwise indicated.

2.2.Membrane preparation

The PES membranes were prepared via NIPS,and the schematic illustration is shown in Fig.1.The ratios of PES and PVP used for membrane preparation are shown in Table 1.The ratio of PES was fixed at 16% (mass) of the total casting solution and PVP as a pore-former additive was controlled from 0 to 16% (mass) of the total casting solution.The prepared mixture was stirred for about 24 h at ambient temperature to ensure homogeneous mixing and then left overnight or more to allow complete liberation of bubbles.The solutions were cast onto glass plates using casting knife with an air–gap set at 250 μm and then immediately immersed into the coagulation bath (25 °C) of deionized water.After peeling off(happen quickly),the polymeric film was immersed in new distilled water bath for more than 24 h to complete the phase inversion.All prepared membranes were kept in deionized water before use.

Table 1Compositions of membrane casting solutions

2.3.Polydopamine coating

An aqueous dopamine solution was prepared via dopamine hydrochloride (2 mg.ml-1) dissolving in Tris buffer solution (pH 8.5,10 mmol.L-1).Since polydopamine (PDA) has a tendency to form free PDA particles at high dopamine concentration or temperature.The prepared membranes(PES-0 to PES-16)were immersed into the dopamine solution and shocked for 6 h at room temperature to avoid the formation of large PDA particles.After desired time of polymerization,dopamine molecules could selfpolymerize on the membrane surface to obtain modifying PDA/PES membranes.The PDA/PES membranes were rinsed with DI water several times and stored in DI water for 24 h to remove the excrescent dark brown precipitates.

2.4.Characterization

All the membrane samples were thoroughly rinsed with DI water,followed by drying in freeze dryer at -35 °C for 8 h prior to characterization measurements.The types of functional groups on the unmodified (PES-0 to PES-16) and PDA-modified (PES-0-PDA to PES-16-PDA) membranes were characterized by Fourier transform infrared (FTIR) spectroscopy (Nicolet 6700,Thermo Fisher Scientific).The morphologies of all membranes were examined by the field mission scanning electron microscopy (SEM)(JEOL,JSM-6380LV).The three-dimensional surface morphology and roughness of the membranes were investigated by employing an atomic force microscopy (AFM,Bruker Dimension Icon PT instrument).The static and dynamic contact angles measurements of the membranes were conducted using a contact angle goniometer(JC2000D1,Shanghai Zhongchen Digital Technic Apparatus Co.Ltd.,China) at room temperature.To minimize experimental error,the measurement was repeated five times to obtain the average.

Fig.1.Schematic illustration for non-solvent induced phase separation (a) and polydopamine coating (b).

The porosity (ε) of the membranes was measured by the gravimetric method:First step,the prepared membranes were immersed in distilled water for 24 h at 25 °C and the wet weight of membrane (2 cm × 2 cm) was measured by electronic balance after carefully wiping the surface with a clean tissue,Ww.Second step,the membrane was dried in an oven at 60 °C for 24 h and weighed again,Wd.The overall porosity (ε) of each membrane was calculated by Eq.(1):

where Wwand Wdare the mass of the wet and dry membranes,A(cm2) is the area of the membranes,L (cm) denotes the thickness of the membrane and ρw(g.cm-3) is the density of pure water.For each membrane,the final result was averaged from 5 samples so as to minimize the experimental error.

2.5.Filtration and antifouling performances assessment

A dead-end filtration apparatus (MSC300,Shanghai Mosu Science Equipment Co.Ltd.,China) with a tested membrane area of 38.5 cm2was used to characterize the filtration performance and protein solution fouling dynamics of the membranes.Permeation and rejection properties of the unmodified and the modified membranes were assessed after the membranes were pre-compacted by deionized water for 0.5 h at 150 kPa.The pure water flux (Jw1) was measured at 100 kPa.Then,the feed solution was switched to BSA solution (1 mg.ml-1in PBS,pH=7.4) and the flux (Jp) was measured at the same condition.Afterward,the membranes were taken out and rinsed with deionized water thoroughly,and were shaken in deionized water for 1 h to remove the reversibly adsorbed proteins.The cell was washed with deionized water and the water flux (Jw2) was measured again as the value after the first cycle of fouling.Two sequential cycles of dead-end filtration were performed for the membranes in the same way.All fluxes were calculated by Eq.(2):

where V is the volume of permeated feed solution(L),A is the membrane area (m2),and ΔT is the permeation time (h).BSA rejection ratio (R) is calculated by Eq.(3):

where Cpand Cf(mg.ml-1) are BSA concentrations of the permeate and feed solution measured with a UV–vis spectrophotometer(UV-2600,Shimadzu,Japan).

The flux recovery ratio (FRR),the total protein fouling (Rt),the reversible fouling ratio (Rr) and the irreversible fouling ratio (Rir)were calculated by the following Eqs.(4)–(7):

2.6.Structural stability

The coating PDA membranes were inverted in dead-end filtration apparatus with the support layer pressurized at 0.25 MPa for 15 min to study the physical stability.Then,the membrane sample was turned over again and the pure water flux was measured under the corresponding pressure.All membranes were immersed in sodium hydroxide solution with pH 13,hydrochloric acid solution with pH 2 and sodium hypochlorite solution with active chloride concentration 400–499 mg.L-1to evaluate the chemical stability according to the deterioration extent of performance[20].The samples were taken out with an interval of 3 days and then were measured water flux and BSA rejection.Each chemical stability test lasted for 15 days and was conducted at room temperature.

3.Results and Discussion

3.1.Analysis of membrane surface chemical structure

The surface chemical structure of the membranes was analyzed by FTIR spectroscopy.FTIR spectra of unmodified(PES-0 to PES-16)and PDA-modified (PES-0-PDA to PES-16-PDA) membranes with varying amounts of PVP (0–16%) are shown in Fig.2.As expected,PES-0 shows typical spectra of pristine PES membrane,aromatic bands at 1578 and 1485 cm-1from the benzene ring and C=C bond stretch and aromatic ether band around 1240 cm-1[21].Compared with PES-0,all PVP-containing PES membranes (PES-1 to PES-16)appear new absorption peaks in the infrared spectra at 1677 cm-1[22].This new band is attributed to the carbonyl existed in PVP and its peak is getting stronger as the content of PVP gradually increases.The interaction between PES and PVP may be from reciprocity of pyrrolidone groups of PVP and the sulfone groups of PES or effect of side cyclic groups of PVP and aromatic ring of PES[22,23].There are four probabilities for the interaction of PES with PVP additive in Fig.3:(i) sulfonic (O=S=O) group of PES with the C-N bond of PVP,(ii) sulfonic (O=S=O) group of PES with the ethylene unit of PVP,(iii) benzene ring of PES with carbonyl(C-O) group of PVP,and (iv) the double bond benzene ring at PES with N atom resulted from the ring-opening of pyrrole ring of PVP[24].After surface modifying with PDA,several new absorption signals(1610 and 3400 cm-1)emerge in PES-0-PDA to PES-16-PDA compared with PES-0 to PES-16,respectively.Absorption band at 1610 cm-1is attributed to the aromatic rings stretching vibrations and N-H bending vibrations.In addition,the intensive peak at 3400 cm-1is attributed to O-H/N-H stretching vibrations[15,25,26].These results indicate that PES membranes have been successfully coated by PDA coating.

Fig.2.FTIR spectra for various membranes.From bottom to top:unmodified membranes (PES-0 to PES-16) and PDA-modified membranes (PES-0-PDA to PES-16-PDA).

3.2.Morphology and structure of membranes

The cross-sectional and surface morphology and structure of all membranes were observed by SEM,as shown in Figs.4 and 5.Fig.4 shows the cross-sectional SEM images of PES and PDA coated PES membranes with different PVP contents.The cross-sectional SEM images demonstrate an asymmetric structure with a dense skin top-layer,a porous sub-layer and fully developed finger-like cavities as well as macro void structure in the substructure for all membranes (PES-0 to PES-16 and PES-0-PDA to PES-16-PDA).However,there are some different structures between them.As shown in Fig.4,PES-0 shows thick skin layer and low porous bottom layer with short finger-like structure [21].By addition of the low amount of PVP (1%–10%,mass) as a hydrophilic additive to the casting solution,the length of finger-like pore is extended,the thickness of the skin layer is reduced and the porosity is increased.Compared to PES-10,PES-16 possesses more denser finger-like cavities and less porous support layers.The finger-like structures of the membranes are formed because of a fast solvent-nonsolvent exchange during the NIPS process.The diffusion rate of solvent affects the growth of the finger-like pores[27].These changes in the membrane morphology can be explained by thermodynamics and kinetics of the casting solution.On the one hand,the addition of PVP leads to thermodynamic instability of the casting solution,which accelerates the demixing and is beneficial to the phase separation process.In addition,the interaction between PVP and PES may also cause an increment of large pores in the membrane surface [28].On the other hand,with increase of PVP more than 10%,the viscosity of dope solution enhances,slowing down the diffusion exchange rate of solvent and non-solvent during phase inversion.In such a case,de-mixing is delayed and the finger-like layer of the membrane becomes denser [29,30].

PES membranes (PES-0 to PES-16) were immersed in the prepared DA solutions.Then,the DA solutions were kept shaking for 6 h when exposed to air.The self-polymerization of DA molecules on membrane surface and pore walls was easily occurred in the presence of dissolved oxygen.From the schematic illustration of the PDA formation mechanism (see Fig.6),the quinone groups of oxidized dopamine and the primary amino groups of dopamine could form PDA coating through Schiff base formation or by Michael type addition [31].The profiles of the coated layers are shown in cross-section SEM images of PDA-modified membranes from Fig.4.A clear boundary line between the coating layer and substrate membrane can be observed,and the thin film of PDA fully covers on the surface of PES membrane.

Fig.5 exhibits the surface images of PES (PES-0 to PES-16) and PDA/PES (PES-0-PDA to PES-16-PDA) membranes.It can be observed that lower PVP concentrations (0–10%,mass) result in clear pores and narrow pore size distribution,while for higher PVP concentrations(16%,mass),the pores are observed to be smaller in size and disappear on the surface of PES membranes.As for the PDA coated membranes,it can be observed that PDA nanoparticles are evenly distributed on the membrane surface.This indicates that the PDA successfully covers the membrane surface.

Fig.3.Four possibilities for the interaction of PES with PVP additive.

Fig.4.The cross-sectional SEM images for unmodified membranes (PES-0 to PES-16) and PDA-modified membranes (PES-0-PDA to PES-16-PDA).

Porosity depends on many initial parameters such as the viscosity of the dope solution,the tensile stress on the polymer solution governed by the take-up speed,the exchange rate between solvent and non-solvent and the residence time of the dope solution in air governed by the air gap and the take-up speed [32–34].These parameters affect the morphology,pore distribution and porosity of the membrane.From Table 2,the porosities of PES-0,PES-1,PES-2,PES-5 and PES-10 are 60.7%,71.6%,78.4%,85.2% and 89.8%,respectively,indicating that the membrane porosity increases with increasing PVP content up to 10%(mass).It has been reported that the hydrophilicity plays a significant role in altering the membrane porosity [34].During the phase inversion process,the hydrophilic nature of PVP could increase the solvent and non-solvent de-mixing process and facilitate the formation of porous structure.This results the increase of porosity when the PVP content increases.In addition,the polymer casting solution with a very low viscosity facilitates the fast diffusion of the solvent and the growth of the large finger-like pores with the formation of top layer.Therefore,the thickness of the top layer decreases significantly from(140.2±0.8)μm(PES-0)to(92.5±0.7)μm(PES-10).However,at the high PVP content(16%,mass),the porosity slightly decreases to 77.3%.This phenomenon could be related to the viscosity of the PVP dope solution.At high dope viscosity,the interaction between non-solvent and polymer is strengthened,resulting in delaying the exchange rate between solvent and water during the phase inversion process.This accelerates the aggregation of polymer molecules on the surface to form a closer bond and hinder the formation of pores [35].Additionally,the thickness of the top layer increases from (92.5 ± 0.7) μm (PES-10) to (118.6 ± 0.2) μm(PES-16).The increase of the top layer thickness may be due to the increase of dope solution viscosity at high PVP content.Sofiah et al.reported that the increment in dope solution viscosity increased the density and thickness of the membrane top layer due to the delay in the de-mixing of solvent and non-solvent[36].The porosity and thickness of PDA-modified membranes are also presented in Table 2.Compared to PES membranes (PES-0 to PES-16),the corresponding PDA-modified membranes (PES-0-PDA to PES-16-PDA) have lower porosity and larger thickness.The decrease of porosity is due to the pore blocking by PDA particles.

Table 2Thickness and porosity of unmodified membranes (PES-0 to PES-16) and PDAmodified membranes (PES-0-PDA to PES-16-PDA)

Fig.5.SEM images of the surface morphologies for unmodified membranes (PES-0 to PES-16) and PDA-modified membranes (PES-0-PDA to PES-16-PDA).

Fig.6.The schematic illustration of PDA formation mechanism.

Membrane surface roughness plays a significant role in the membrane fouling and hydrophilicity[37].AFM images of unmodified membranes(PES-0 to PES-16)and PDA-modified membranes(PES-0-PDA to PES-16-PDA) are shown in Fig.7.In these images,the brightest areas represent the highest point of the membrane surface and the dark regions indicate the valleys or membrane pores [38].Table 3 summarizes the surface roughness parameters in terms of mean roughness (Ra),root mean square roughness(Rq) for all prepared membranes.

Table 3Root mean square roughness (Rq) and mean roughness (Ra) of all membranes

As displayed in Table 3,PES-0 has a relatively low Ra,while the Raof PES-1 to PES-10 gradually increases with increasing the PVP content in membranes.This is due to the pore formation process during solidification.The presence of PVP makes it easier for water to penetrate the casting solution,which will leave traces of pores on the membrane,resulting in a porous membrane with increased surface roughness [37,39].Further,PVP is used as a hydrophilic additive,and its hydrophilicity and the number of active groups are suspected to be the reasons for membrane surface roughness[38].Nonetheless,the surface roughness of PES-16 does not follow the trend.The Raof PES-16 is 9.45 nm.The sharp reduction in the roughness is due to the collapse of the nodules.The nodular structure increases and reaches a stage where it collapses,resulting in the smoother membrane surface [40].After surface modifying of PDA,it can be seen that PDA-modified membranes (PES-0-PDA to PES-16-PDA)have larger Ravalues compared with PES membranes(PES-0 to PES-16).This phenomenon is attributed to the coverage of PDA on membrane surface.One can see from the SEM images of the surface morphologies of PDA-modified membranes in Fig.5 that,PDA particles are clearly distributed on the membrane surface after the coating,resulting in the increase of the membrane surface roughness.Membranes with higher surface roughness (i.e.the amount of PDA is more on membrane surface) have higher hydrophilicity and membranes with the smoother surfaces showing lower hydrophilicity [41,42].The maximum roughness obtained in the experiments is 31.32 nm of Rafor PES-10-PDA.This is in favor of the enhancing of hydrophilicity and wettability of PES-10-PDA to a certain extent.

Fig.7.AFM images of the surfaces for unmodified membranes (PES-0 to PES-16) and PDA-modified membranes (PES-0-PDA to PES-16-PDA).

3.3.Surface wettability

The wettability represents the ability of the membrane surface to be wetted by water.It is related to the surface morphology of the membranes and the surface energy of the membrane materials.The water contact angle reflects the wettability [43–45].The wetting surfaces have been proved to be fouling resistant,because the hydrophilic groups can generate strong interaction with water molecules through hydrogen bond or electrostatic attraction.The surface hydration can lead to the formation of tightly bounded water layer on the surface,which is mainly responsible for large repulsive force to repel the adsorption of foulants[37,46].The static contact angles of unmodified (PES-0 to PES-16) and PDAmodified (PES-0-PDA to PES-16-PDA) membranes were measured,and the results are shown in Fig.8(a).It is found that the PES-0 has a contact angle of 72.03°.According to the analysis of PES-1,PES-2,PES-5 and PES-10,the static contact angles of the membranes drop significantly from 69.58°to 60.81°,indicating that the surface wettability of the membranes increases gradually with the increase of the PVP content.The increase of the hydrophilicity of the membrane can be attributed to the polar groups in PVP molecules,such as strong hydrophilic lactam groups[13].However,compared with PES-10,PES-16 has a larger static contact angle.This phenomenon can be explained as follows:One can see from the cross-sectional and surface SEM images of PES-16 that,additional PVP up to 16%(mass)results in the formation of a denser skin layer and less porous support layer,thus PES-16 possesses a lower porosity compared to PES-10.In case of porous membrane,the contact angle depends on the porosity and surface energy of membrane [39].The lower the porosity is,the higher the contact angle will be[47].Similar results were also obtained by Huang et al.in the preparation of PVDF membranes [48].

Fig.8.(a) Static contact angles and (b) dynamic contact angles for unmodified membranes (PES-0 to PES-16) and PDA-modified membranes (PES-0-PDA to PES-16-PDA)(Temperature,25 °C;each of the experiments was repeated for five times to reduce errors).

After surface modifying with PDA,the static contact angles of PDA modified membranes are smaller than those of unmodified membranes,which indicate that the modified membranes become more hydrophilic when membrane surfaces are coated with PDA.This effect is probably due to the existence of the hydrophilic groups in the PDA molecules,such as catechol and amine groups,which could improve membrane hydrophilicity [32].

The dynamic contact angles depend on the contact time between the membrane and water,and the angles decrease with increasing the time.The contact angle of membrane with better wettability decreases more rapidly [49].The dynamic contact angles of all membranes were measured to confirm the influence of PVP on the PES membranes and evaluate the wettability effect of PDA coating,the results are shown in Fig.8(b).The dynamic initial contact angles of all membranes are consistent with the static contact angles.After the PVP is added,it is easily observed that the decreasing rate of contact angle increases with increasing PVP content in cast solutions.Among PES membranes,PES-10 shows the higher decay rate of contact angle.As for PDA modified membranes,they have the faster decay rate of contact angle compared to PES membranes.Especially,PES-10-PDA exhibits the significant decay rate of contact angle.When the water droplet is dribbled on membrane surface,it spreads instantly due to the strong affinity between PDA and water molecules.This indicates that the PDA coating improves significantly the wettability of PES membranes.

3.4.Permeation flux and BSA rejection

To investigate the effects of membrane structure and hydrophilicity on membrane permeation,the filtration performances of unmodified (PES-0 to PES-16) and PDA-modified (PES-0-PDA to PES-16-PDA) membranes were firstly analyzed by measuring pure water flux and BSA rejection rate [46].As shown in Fig.9(a),due to the dense porous layer structure and inherent hydrophobicity,PES-0 shows less pure water flux under experimental conditions.After adding PVP,the water flux of membrane increases with the increase of PVP content in the range of PES-1 to PES-10,and the values of water fluxes for PES-1,PES-2,PES-5 and PES-10 are 423.5,794.4,835.9 and 935.5 L.m-2.h-1,respectively.However,the water flux of PES-16 decreases obviously compared to that of PES-10,but it is much higher than that of PES-0.The water flux of membrane mainly depends on the mass transfer resistance across the membrane surface[50].Thus,the membrane surface structures including pore size,porosity,the thickness of top layer,and membrane hydrophilicity have great effects on the membrane permeability [37,50,51].The increase of water flux of PES-1 to PES-10 is due to the enhancement of pore size and surface hydrophilicity of membranes,which facilitates the water molecules to pass through the membranes.As for PES-16,since the PVP concentration in the dope solution is up to 16%(mass),the viscosity of the dope solution increases,which leads to kinetic hindrance to phase separation and thus to cause thermodynamic instability[46].The dense skin layer and less porous support layer are formed in PES-16 as shown in SEM images,resulting in the increase of water transfer resistance.The final result is a reduction of water flux of PES-16.

The PDA coated membranes(PES-0-PDA to PES-16-PDA)show a lower water flux compared with unmodified membranes(PES-0 to PES-16).This phenomenon is explained as follows:The membrane hydrophilicity and structure have a great influence on water flux of the membrane[50,52].On the one hand,PDA coating can improve membrane surface hydrophilicity,resulting in the trend of the increase of membrane flux.On the other hand,PDA accumulates in the pore walls and causes the decrease of membrane pore size,which limits water transport.In this case,the latter factor may play a major role in determining membrane flux,resulting in the decrease of water flux.Among all modified membranes,PES-10-PDA has the highest water flux.This is due to the fact that PES-10-PDA has the porous structure and the highest hydrophilicity.

The BSA rejection results of all membranes are depicted in Fig.9(b).Although the water fluxes of PES-0 and PES-0-PDA are only about 21.4 and 11.3 L.m-2.h-1,respectively,they have the higher BSA rejection.This is because the smaller membrane pores of PES-0 and PES-0-PDA can block the passage of water molecules,and also hamper the BSA molecules.As the PVP is introduced,the BSA rejection of membranes decreases gradually from PES-1 to PES-10.The decrease in rejection is attributed to the fact that the increased membrane surface pores for PES-1 to PES-10 provide BSA molecules more opportunity to penetrate the membranes[10].However,a slight increase in BSA rejection is observed for PES-16.This is due to the dense skin layer and less porous support layer of PES-16.Further,compared with unmodified membranes(PES-0 to PES-16),the PDA modified membranes(PES-0-PDA to PES-16-PDA)have higher BSA rejections.These results can be explained as follows:on the one hand,PDA coating reduces membrane pore size.On the other hand,the roughness of the membrane surface has also certain impact on BSA rejection,leading to the higher protein rejection for the modified membranes [15,53].There is no significant difference between the rejection of PES-5-PDA and PES-10-PDA,which is probably related to the change of thickness of the membrane surface.The PES-10-PDA shows the superior permeability without sacrificing the rejection compared to other membranes.

Fig.9.(a) Water fluxes of all membranes before the filtration of BSA solution,(b) the BSA rejection rate of all membranes during the filtration of BSA solution (BSA concentration:1 mg.ml-1 in PBS,pH=7.4;Operating pressure:0.1 MPa;Temperature:25 °C;each of the experiments was repeated for three times to reduce errors).

3.5.Antifouling performance

Anti-fouling performance of membranes during actual operation is crucially important for the durability and efficiency of membranes.Each membrane was compacted with deionized water at 0.2 MPa for 30 min,and then the transmembrane pressure was kept at 0.1 MPa during the filtration process.Fig.10(a) shows the time-dependent fluxes of unmodified (PES-0 to PES-16) and PDAmodified (PES-0-PDA to PES-16-PDA) membranes.Two-cycle filtration experiments by using BSA solution as the protein fouling source were conducted to evaluate the anti-fouling performance of all membranes,where 0–30 min,60–90 min and 120–150 min represent pure water tests to obtain the pure water flux,then 30–60 min and 90–120 min represent the tests of BSA solution to obtain the BSA flux.PES-0 and PES-0-PDA show very little water flux and BSA flux due to the lack of pores on the membrane surface.In the first cycle,the unmodified membranes (PES-1 and PES-10)have higher pure water flux (Jw1) compared with the PDAmodified membranes (PES-1-PDA and PES-10-PDA).It can be explained that pore sizes of the modified membranes decrease owing to PDA coating,which plays a major role in determining pure flux.Dang et al.[37] studied the relationship between pore size and water flux,and obtained similar results.When pure water is replaced by BSA solution,the flux of each membrane declines dramatically.It is due to the adsorption or deposition of protein molecules on the surface of membranes [54].In the second cycle,after membrane cleaning,pure water fluxes(Jw2)of all membranes could not recover completely because the BSA molecules may be entrapped in the membrane pores.

The flux recovery rate (FRR) provides the information on the recycling properties of the membranes,the higher FRR value is,the better antifouling property is [55].The FRR values of prepared membranes are depicted in Fig.10(b).For the PES-0,due to the poor anti-fouling property,its FRR is 54.31% and 45.01% after the first and the second cleaning cycles,respectively.The FRRs of membranes of mixed PVP (PES-1 to PES-16) are higher than that of the base membrane (PES-0).This is because the addition of PVP increases membrane surface hydrophilicity,and effectively reduces the interaction of protein molecule with the membrane surface[56].In addition,it is obvious that the FRRs of the modified membranes (PES-0-PDA to PES-16-PDA) are higher than those of the unmodified membranes (PES-0 to PES-16) under the same operating conditions.The good anti-fouling properties of the modified membranes are mainly due to the fact that the PDA can provide more hydrophilic groups,and the surface of the membranes is more inclined to bind water molecules,and forms a hydrophilic layer to resist the invasion of BSA.On the other hand,the existing of large numbers of PDA also plays a physical role in the filtration process.PDA layer can be regarded as a barrier to decrease the probability of BSA entering membrane pores,resulting in the enhancement of anti-fouling property [53,57].Thus,it is easier to remove the protein fouling from the membrane surface and pores and obtain a high FRR for the modified membranes.PES-10-PDA obtains the highest FRR(91.23%),which indicates that it possesses the best antifouling property.

To further discuss the anti-fouling properties of membranes,three fouling parameters of the reversible fouling ratio (Rr),the irreversible fouling ratio (Rir),and the total protein fouling (Rt)are used to assess the anti-fouling properties of all membranes.Rtis attributed to the effect of total protein fouling,which is due to the adsorption and deposition of protein on the membrane[55].Rtconsists of Rrand Rir.Rrindicates that the adsorption of protein on the membrane could be reversed or removed by simple hydraulic cleaning [57].On the contrary,Rirrefers to the stable adsorption of protein molecules on the membrane surface,or the entrapment of protein molecules in pores,and it is more difficult to be removed [56].Lower Rtand higher Rr/Rtratio can be interpreted as an improvement of the antifouling property of the membrane [58].In Fig.10(c),the modified membranes show lower Rtand higher Rr/Rtratio compared with the unmodified membranes,which suggests that the PDA coating improves the membrane antifouling tendency.Riris dominant for the PES-0-PDA,PES-1-PDA and PES-2-PDA,while Rris major for the PES-5-PDA,PES-10-PDA and PES-16-PDA.Essentially,the smooth surface shows less fouling than the rough surface due to less spatial interaction between the foulant and membrane surface[59].However,the major factor that plays a significant role in this phenomenon is the hydrophilicity provided by PDA.The hydrophilicity plays an important role in increasing the antifouling property of the membrane.The fouling tendency is reduced by the repulsive interaction between protein molecule and hydrophilic membrane surface.As a result,the excellent anti-fouling property can be obtained by membrane surface modification with PDA.

Fig.10.(a)Time-dependent fluxes of all membranes under two cyclic filtration tests using BSA as a model protein,(b)the flux recovery ratio(FRR)values of the membranes during the two filtration cycles,(c) the reversible fouling ratio (Rr),the irreversible fouling ratio (Rir),the total protein fouling (Rt) and Rr/Rt ratio of membranes (BSA concentration:1 mg.ml-1 in PBS,pH=7.4;Operating pressure:0.1 MPa;Temperature:25 °C;each of the experiments was repeated for three times to reduce errors).

3.6.Stability

The PDA coating can be applied easily,but after a long time use of the membrane,the coated layer will also be easily removed from the membrane surface [60].Therefore,stability tests were performed to evaluate the practical applicability of the PDA coating.

To investigate the mechanical stability of the PDA coated layer,the back-flush (hydraulic washing in reverse direction) was performed at pressure 0.25 MPa for 15 min.As shown in Table 4,after back-flush,there is no significant change in water flux for PDA modified membranes.According to Chen’s report,the peeling strength between the PDA layer and substrate membrane is about(25.3 ± 3.0) N.m-1[61].It is stable enough to resist the hydraulic force in back-flush tests.

Table 4Specific flux values of all membranes before and after back-flush

Fig.11.(A) represents water flux and BSA rejection of the unmodified membranes after soaking in hydrochloric acid (a and b),sodium hydroxide (c and d),and sodium hypochlorite(e and f),respectively;(B)represents water flux and BSA rejection of the modified membranes after soaking in hydrochloric acid(g and h),sodium hydroxide(i and j),and sodium hypochlorite(k and l),respectively.(Back-flush pressure,0.25 MPa;temperature,25°C;each of the experiments was repeated for three times to reduce errors).

Fig.11.(continued)

The chemical stability of the PDA coated layer was investigated by soaking the membrane samples in chemical solutions,including hydrochloric acid (pH 2),sodium hydroxide (pH 13) and sodium hypochlorite solution (active chloride concentration 400–499 mg.L-1).Each solution stability test lasted for 15 days and the samples were measured water flux and BSA rejection with an interval of 3 days.Fig.11(A) and (B) show the effect of soaking time on water fluxes and BSA rejections of unmodified membranes (PES-1 and PES-16) and PDA-modified membranes (PES-1-PDA and PES-16-PDA),respectively.The water fluxes and BSA rejections of unmodified membranes almost keep unchanged for hydrochloric acid and sodium hydroxide solutions in Fig.11(a)–(d),respectively.However,the water fluxes of modified membranes increase with the soaking time for sodium hydroxide solution in Fig.11(i),and their BSA rejections decrease slightly in Fig.11(j).This may be due to loose the structure of PDA layer corroded by sodium hydroxide solution,which facilitates to the water flowing through the membrane [20].The water fluxes and BSA rejections of modified membranes almost keep unchanged for hydrochloric acid in Fig.11(g)and(h).Furthermore,after soaking in sodium hypochlorite solution,the unmodified membranes have an increase in water flux in Fig.11(e).Interestingly,the water fluxes of PES-2,PES-5 and PES-10 are found to be almost constant and about 850 L.m-2.h-1for 15 days whatever the PVP amount in the casting solution.Meanwhile,the BSA rejections of PES-2,PES-5 and PES-10 also show the constant values after 9 days of soaking in sodium hypochlorite solution,as shown in Fig.11(f).These results indicate that PES/PVP membranes are impacted in sodium hypochlorite solution to appear macro voids in the membrane structure.The increase of macro voids may be due to the oxidation of the PVP or the acceleration of the PES degradation owing to PVP oxidation by-products [12].Similar results were also reported by Kourde-Hanafi et al.in the influence of PVP content on degradation of PES/PVP membranes[62].The modified membranes after immersing in sodium hypochlorite solution have a significant increase in water flux,as shown in Fig.11(k).It may be explained that,the PDA reacts with sodium hypochlorite,resulting in producing some complex mixtures or occurring some interactions between PDA coating layer and the original membrane,which damages the membrane structure and leads to an increase in water flux [20].Compared with PES-2-PDA,PES-5-PDA and PES-16-PDA membranes,PES-10-PDA has the highest BSA rejection after 9 days of soaking in sodium hypochlorite solution in Fig.11(l).The results indicate that PES-10-PDA has better chemical stability than other membranes after they are immersed in the studied chemical solutions.

4.Conclusions

In this study,the effects of PVP content and polydopamine coating on membrane morphology,structure and performances were investigated in detail.The main conclusions are as follows:(1)The higher PVP content in the PES membrane facilitated the formation of more pores and thinner top skin layer with the slightly larger pore size.The PDA coating reduces the pore sizes of PES membranes according to SEM analysis.The roughness of the membrane surface significantly increased after the PDA coating,and it was particularly noticeable for PES-10-PDA.(2) By increasing the PVP content,the hydrophilicity of the prepared membranes was significantly enhanced.The membrane with 10%(mass)PVP exhibited the superior flux recovery and antifouling performances compared with other PES membranes.Pure water and BSA solution fluxes of PDA modified membranes decreased due to the adhesion of PDA on pore walls and thus the decrease of membrane pore size.The flux recovery of PDA modified membranes was much higher than that of PES membranes,owing to the coating of hydrophilic PDA.The Rtand Rirof PDA modified membranes were lower than those of PES membranes,implying that PDA modified membranes possessed the excellent antifouling performance.(3) The stability tests show that the coated PDA layer has a reliable mechanical stability by hydraulic back-flushing.The coated PDA layer is relatively stable in acid solution compared to alkali solution.After soaking in sodium hypochlorite solution,water fluxes of modified membranes significantly increased and BSA rejections decreased.This study concludes that the modified membranes demonstrate the good potential for application in treating fouling.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Nos.51573041,21776067),and the Demonstration Base Project of University-Enterprise Cooperation of Hunan Province (No.145812).