Chen Zhao,Yahan Ye,Xianfu Chen,Xiaowei Da,Minghui Qiu,,Yiqun Fan

1 State Key Laboratory of Materials-Oriented Chemical Engineering,College of Chemical Engineering,Nanjing Tech University,Nanjing 211800,China

2 Nanjing Membrane Material Industry Technology Institute Co,Ltd,Nanjing 211800,China

Keywords:Membranes Ultrafiltration Separation Modification Surface charge Dye wastewater treatment

ABSTRACT Tight ceramic ultrafiltration membranes have been proven to exhibit good rejection performance for reactive dye wastewater at high temperatures because of their high thermal and chemical resistance.However,the application of ceramic membranes for the treatment of cationic dye wastewater is challenging because of their surface charge.In this study,a ceramic membrane is modified by grafting aminosilane (KH-551) to enhance the positive charge of the membrane surface.The rejection performance of the charged modified ceramic membrane toward the methylene blue solution is significantly improved.The modification substance is bonded to the ceramic membrane surface via covalent bonding,which imparts good thermal stability.The modified ceramic membrane exhibits stable separation performance toward the methylene blue solution.Overall,this study provides valuable guidance for the adjustment of the ceramic membrane surface charge for treating industrial cationic dye wastewater.

1.Introduction

The treatment of textile and dyeing wastewater containing toxic dyes by using membrane processes has garnered considerable research attention for several years [1,2].Compared with nanofiltration (NF) membranes,tight ultrafiltration (UF) membranes are advantageous for the treatment of dye wastewater,because of their high dye retention and permeate flux.Furthermore,tight UF membranes enable the separation of salt electrolytes (NaCl or Na2SO4) from dyeing wastewater [3].A concentrated dye solution with trace salt amounts can be directly treated by biodegradation [4,5].

Wastewater from textile and dyeing industries typically has a temperature of 60–90°C[6].This seems unnecessary,if a filtration system can work directly at high temperatures,high-temperature feed will enhance the permeance of the membrane and improve the separation efficiency [7].Tight ceramic UF membranes are promising for the treatment of dye wastewater because of high permeance,thermal stability and corrosion resistance[8,9].In previous studies,tight ceramic UF membranes have been reported to exhibit excellent rejection performance toward reactive dyes[10,11].However,treatment of cationic dye wastewater by using tight ceramic UF membranes is still challenging because of their weak surface charge.Cationic dyes are especially suitable for the dyeing of acrylic fibres,and the dyeing temperature typically needs to be ＞60 °C for a stronger dyeing effect.Therefore,it is necessary to develop a positively charged tight ceramic UF membrane for the treatment of cationic dye wastewater.

Because cationic dyes typically have smaller molecular weights than other dyes and disperse more evenly in a solution[12,13],it is indispensable to exploit the electrostatic repulsion between the ceramic membrane surface and dye molecules when using such membranes for cationic dye wastewater treatment.Wei et al.[14]prepared positively charged NF membranes using polyethyleneimine and trimesoyl chloride as reactive monomers.They found that the resulting membranes exhibited higher separation performance toward cationic dyes than conventional hollow-fibre UF membranes.Zhang et al.[15] used dopamine and polyethyleneimine to regulate the surface charge of polyethersulfone (PES)membranes.The resulting membrane surface was highly positive,and the separation performance toward cationic dyes (methylene blue,retention 96.5%) was significantly improved.However,the surface charge of ceramic membranes completely relies on the protonation or deprotonation of the hydroxyl groups on the membrane surface,which is very weak under neutral pH conditions[16,17].Different methods (e.g.,doping and chemical grafting)have been adopted to tune the surface charge of ceramic membranes.Among them,chemical grafting using organosilanes is the most effective[18].Specific functional groups can be anchored on the surface of ceramic membranes through covalent bonds(Si-O-M) to adjust the surface properties of the membranes[19,20].Moreover,the modification effect exhibits good longterm stability because of the strong covalent bonding force between organosilanes and ceramic membranes.

To strengthen the positive charge on the membrane surface,organosilanes with amino groups were grafted onto the ceramic membrane surface.Lee et al.[21] grafted a monolayer of aminosilanes onto the surface of alumina membranes,which resulted in a strong positive charge on the modified membranes.They found that the modified membranes exhibited an obvious antifouling performance toward positively charged pollutants.Bartels et al.[22]grafted different aminosilanes on the surface of ceramic capillary membranes.After the modification,the zeta potential of the modified membranes changed from negative to positive,leading to a significantly increased virus retention capacity.In previous reports,aminosilanes were typically used to adjust the surface charge of ceramic microfiltration membranes to improve the antifouling and adsorption properties;however,negligible attention has been paid to tuning the surface properties of tight ceramic UF membranes.Moreover,the modification of tight ceramic UF membranes by grafting aminosilanes has a completely different effect on the separation performance,surface pore size,and total porosity from that of ceramic microfiltration membranes [23,24].The enhanced surface charge can even help tight ceramic UF membranes to reject charged substances (e.g.,dye molecules and ions)that are smaller than the membrane pore size.

In this study,positively charged composite ceramic UF membranes were designed by grafting an aminosilane onto tight ceramic UF membranes.Amino functional groups were attached to the surface of ceramic membranes by covalent bonds,which compensated for the surface charge weakness of ceramic membranes.The influence of the grafting layer on the pore size was studied via the retention of polyethylene glycol(PEG)with different molecular weights.Four different dyes(reactive blue 19,methylene blue,rhodamine B,and basic green 4) were used to evaluate the dye rejection performance of the modified ceramic membranes.This work will provide valuable guidance for the fabrication of positively charged ceramic membranes that can be applied for the treatment of cationic dye wastewater.

2.Experimental

2.1.Materials

3-Aminopropytrimethoxysilane(KH-551,≥97%(volume),Aladdin,China)was used for the preparation of the functionalised ceramic membranes.Ethanol (≥99.5% (volume)) was obtained from Wuxi City Yasheng Chemical Co.,Ltd.(China)for the grafting modification and cleaning of the modified membrane.Inorganicsalts were obtained from Aladdin without further treatment.The dyes were purchased from Macklin (China).Polyethylene glycol was purchased from Sigma (USA) to characterise the molecular weight cut-off(MWCO) of the ceramic membranes.Deionized water with a conductivity of 2.8 μs·cm-1was made in the laboratory.

Asymmetric tubular ceramic membranes with a zirconia selective layer(outer diameter of 12 mm,inner diameter of 8 mm,mean pore size of 2 nm,Jiangsu Jiuwu High-tech Co.Ltd.,China) were used as the support for the grafting modification.Asymmetric tubular ceramic membranes with a TiO2–ZrO2selective layer were made in the laboratory.

2.2.Modification of ceramic membranes

The schematic of the preparation of the modified ceramic membranes is shows in Fig.1.Before the modification,the ceramic UF membrane was fully rinsed with pure water and dried in an oven at 120 °C for 12 h.3-Aminopropytrimethoxysilane was dissolved in absolute ethanol to prepare the modified solution,and the solution was stirred at 25°C for 2 h.The membrane was immersed in a modification solution at 35 °C for a specific time.The modified membrane was then rinsed with pure water to remove unreacted modified substances and fully dried at 120 °C.

2.3.Membrane characterization

Field emission scanning electron microscopy (FE-SEM,S-4800,Hitachi,Japan) was conducted to observe the microstructure of the ceramic UF membrane surface.Attenuated total reflection flourier transformed infrared spectroscopy (ATR-FTIR,Spectrum 100,Perkin Elmer,USA) was performed to analyse the functional groups on the membranes.X-ray photoelectron spectroscopy(XPS,ESCALAB250Xi,Thermo,USA) was carried out to analyse the chemical composition of the membrane surface.The surface zeta potential of the ceramic membrane was determined using ZetaPALS (Brookhaven,USA).The membrane materials were dispersed in aqueous solutions with disparate pH values and were used as test samples.

Fig.1.Preparation mechanism of the modified ceramic UF membrane.

The thermal stability of the modification substance was analyzed on the basis of the thermal gravity behavior (STA 449F3 instrument,NETZSCH,Germany)of KH-551 in air.The KH-551 liquid was dissolved in ethanol and dried at 110°C to obtain KH-551 crystals.Then,the KH-551 crystals were heated to 1000°C at a rate of 10°C·min-1under a 100 ml·min-1air flow.The weight loss and endothermic curves of the samples against temperature were determined for further analysis.

To characterize the stability of the bonding between KH-551 and the ceramic membrane at high temperatures,the modified ceramic membrane was washed with hot water (60 °C) for 30 h.The heat-treated modified membrane was then used for the analysis of the surface element composition by XPS.

2.4.Membrane separation performance

A self-made cross-flow filtration device was used to perform filtration experiments on the ceramic UF membranes.A tubular ceramic membrane with a length of 11 cm and an effective area of 2010 mm2was loaded into the membrane module.Before testing,each membrane was stabilised in pure water for 1 h at 0.5 MPa.Salt solutions and mixed polyethylene glycol solutions were used as the feed to assess the ceramic membrane rejection performance of charged and neutral substances,respectively.The concentrations of these feed solutions were 1 g·L-1for salt and 3 g·L-1for PEGs,respectively.The molecular weights of PEGs (600,1500,4000,and 10,000 Da) were widely distributed,which were convenient for the characterization and analysis of the MWCO.To stabilise the feed concentration,the permeation solution was poured back into the feed.Each membrane was tested for 1 h at a pressure of 0.5 MPa and a temperature of 25 °C.To determine the error range,at least three parallel experiments were conducted.

Permeance,P (L··m-2·h-1·MPa-1),was calculated using Eq.(1):

where V is the permeate volume(L),A is the effective filtration area(m2),t is the operation time (h),and p is the operating pressure(MPa).

Rejection,R (%),was calculated using Eq.(2):

cp(g·L-1)and cf(g·L-1)in Eq.(2)are the solute concentrations of the permeate solution and feed.

Gel permeation chromatography (GPC,1515,Waters,USA) was conducted to determine the content of polyethylene glycol in the solution.An electrical conductivity meter(DDS-307A,INESA Scientific Instrument Co.,Ltd,China)was used to measure the electrical conductivity of the solution to determine the salt concentration.

2.5.Static adsorption of dye solutions

A static adsorption experiment was carried out to verify the electrostatic repulsion between the modified ceramic membrane surface and the positively charged dye.A ZrO2membrane and the modified membrane were cut to a length 11 cm.The washed and dried membranes were then immersed in a methylene blue solution (1 g·L-1).The sample was removed from the immersion solution every 24 h to determine the change in the dye concentration.A UV/Vis spectrophotometer (AQ 7000,Thermo Scientific,USA) was used to analyze the concentration of methylene blue in the sample solution.The dye adsorption rate,Ra(%),was calculated using Eq.(3):

ci(g·L-1) and co(g·L-1) in Eq.(3) represent the concentrations of methylene blue in the immersion and original solutions,respectively.

2.6.Separation performance to dye solutions

Four types of dye solutions(0.1 g·L-1)were utilized to evaluate the single dye rejection performance of the ceramic membrane at 0.2 MPa and 25 °C.To investigate the effect of the membrane surface charge on dye separation,one negatively charged dye solution(reactive blue 19) and three positively charged dye solutions(methylene blue,rhodamine B,and basic green 4) were used as the feed.Before recording the results,the filtration process was stabilized for 1 h.To ensure the reproducibility of the results,the rejection rates and fluxes of at least three parallel samples were recorded.To prove that electrostatic repulsion played a decisive role in the cationic dye rejection performance of the modified membrane,a ZrO2ceramic membrane (about 1000 Da) [25] without any modification was used as a contrast membrane to separate a methylene blue solution (0.1 g·L-1).

To evaluate the stability of the modified membranes,the filtration of the methylene blue solution was conducted for 100 h at 25 °C.The rejection rate and flux were recorded every 10 h.After each cycle,the modified membrane was washed with pure water.

3.Results and Discussion

3.1.Effect of amino functionalisation on zeta potential

Electrostatic repulsion plays a crucial role in the filtration process,especially when the particle of the substances in the feed is smaller than the pore size of the ceramic membranes.Ceramic UF membranes with ZrO2as the separation layer are negatively charged at neutral pH.However,the charge of the ceramic membrane is very weak at neutral pH because it completely relies on the protonation and deprotonation of -OH.As illustrated in Fig.2,the zeta potential of the ZrO2UF membrane is -5.36 mV at pH=7.When the ZrO2membrane was modified by KH-551,the zeta potential of the modified membrane increased to 15.6 mV at pH=7 because of the introduction of the amino groups in KH-551.At the same time,the isoelectric point of the ZrO2membrane increased from 6.9 to 7.4 after the modification.

Fig.2.Zeta potential of the ZrO2 membrane and the modified ZrO2 membrane(KH-551 concentration=30 mmol·L-1,modification time=4 h) at different pH values.

3.2.Effect of amino functionalisation on membrane material chemical composition

The functional group analysis of the ceramic membrane material was performed through FTIR and XPS.The results are shown in Fig.3.After the modification with KH-551,the ceramic membrane showed additional peaks:2919,2854,1180,and 1100 cm-1,as displayed in Fig.3(a).The peaks at 2919 and 2854 cm-1occur which are attributed to the stretching vibration of-CH2-bonds.The peak at 1180 cm-1arises because of the outside swing of the-CH2-bonds.The additional peak at 1100 cm-1is ascribed to the C-N bond in KH-551 of the modified membrane.The position of the additional peak in the modified membrane corresponded to the characteristic peak of the KH-551 FTIR spectra.Fig.3(b) shows the differences between the element peaks of the original and modified membranes.The original membrane had significant peaks of O 1s(530.26 eV)and Zr 3d(183.83 eV)because of the presence of the ZrO2separation layer.The C 1s peak at 284.8 eV was attributed to the presence of contaminants on the membrane surface.After the grafting modification,the C 1s (284.8 eV),N 1s(399.48 eV),and Si 2p (100.78 eV) peaks were clearly enhanced because of the presence of these three elements in KH-551(NH2(-CH2)3Si(OCH3)3).As illustrated in Table 1,the carbon content of the ceramic membrane increased from 19.8% to 26.3% after the modification.The nitrogen and silicon contents increased by 81.4% and 64%,respectively.The nitrogen and silicon contents of the modified membrane changed depending on the KH-551 concentration.This relationship is illustrated in Fig.3(c).The nitrogen and silicon contents increased as the concentration of the modification solution increased up to 30 mmol·L-1.However,these contents tended to be stable when the concentration was increased above 30 mmol·L-1.This is because the active hydroxyl sites on the ceramic membrane surface for grafting aminosilane are limited.In conclusion,aminosilane was successfully grafted onto the membrane surface.Furthermore,the distribution of the modification substance was uniform,which can be verified by the elemental distribution on the modified membrane surface in Fig.3(d).

Table 1 XPS composition of the original membrane and the modified membrane

Fig.3.The characterization of membrane material chemical composition.(a) FTIR spectra and (b) XPS spectra of the original membrane,the modified membrane (KH-551 concentration=30 mmol·L-1,modification time=4 h) and the KH-551.(c) The atomic composition changes of the modified membranes vs.KH-551 concentration(modification time=4 h).(d) Elemental distribution of the modified membranes (KH-551 concentration=30 mmol·L-1,modification time=4 h).

3.3.Thermal stability of grafting modification

The thermal stability of the grafting modification is determined by the stability of the modification substance and the strength of the bonding between the modification substance and the membrane surface at high temperatures.Because the weight of the modification substance grafted on the ceramic membrane accounts for only a small part of the weight of the whole membrane,there is little difference between the TG curves of the membranes before and after modification as shown in Fig.S1 (see Supplementary Material).The modification reagent (KH-551 crystals) was used to analyse the thermal stability of the modification indirectly;the results are displayed in Fig.4(a).At 100 °C,there was almost no loss in the weight of the crystal,which proves that the modification substance can withstand this temperature.From 100 to 375 °C,there was a weak endothermic peak in the DSC curve,and the sample weight decreased slightly,corresponding to the loss of crystal water.At 375°C,weight loss increased significantly;a strong endothermic peak was observed at 425°C,suggesting that the thermal decomposition of the modification substance begins at 375 °C.The hot water washing experiment (60 °C,30 h) was conducted to characterize the bonding strength of the modification substance and the ceramic membrane surface at high temperatures.The elemental composition of the membrane surface before and after the heat treatment is shown in Fig.4(b)and Table 2.The XPS spectra of the heat-treated modified membrane hardly changed compared to those of the non-treated membrane.However,the nitrogen and silicon contents on the membrane surface decreased slightly (by 0.7% and 0.5%,respectively).This result could be explained by the fact that only a small amount of unreacted modification substance is adsorbed on the surface of the ceramic membrane and,hence,is easily washed off.Only a small part of the modification substance was lost,which had no significant influence on the effect of the grafting modification.Therefore,the modified ceramic membrane can be used as a filtration system with an operating temperature of ＜100 °C.

Table 2 XPS composition of the modified membrane before and after the heat treatment

3.4.Effect of amino functionalisation on surface morphology

Chemical grafting with KH-551 can form a single-molecule modification layer on the membrane surface,as shown in Fig.5(a).The modification involved two steps.First,the water adsorbed on the membrane surface hydrolysed the aminosilane to the corresponding silanol.Silanol then dehydrated and condensed with the-OH of the ceramic membrane surface and the adjacent silanol to form a crosslinked network,which rarely affects the micromorphology of the ceramic membranes.As shown in Fig.5(b)–(f),no obvious morphological differences can be observed on the ceramic membrane surface after the grafting modification even when the concentration of the modification solution was increased to 70 mmol·L-1.

3.5.Measurement of modified membrane separation properties

Pure water,a series of PEGs (600,1500,4000,and 10,000 Da),and four salt solutions (MgCl2,MgSO2,Na2SO4,and NaCl) were used as feed to assess the separation performance of the modified membranes.As illustrated in Fig.6(a),the water permeance of the original ZrO2membrane was approximately 300 L·m-2·h-1·MPa-1.After the grafting modification,water permeance decreased by approximately 60%(120 L·m-2·h-1·MPa-1)owing to the reduction of the total porosity and the membrane pore size.The thickness of the modification layer(aminosilane)at a low modification concentration and a low modification temperature is approximately 0.2 nm [26,27],which is too small for the layer to be observed through FESEM.However,such a thin modification layer can still affect the pore size of tight ceramic UF membranes.The results are shown in Fig.6(b).The MWCO of the original membrane was 2300 Da.The corresponding membrane pore size was 2.4 nm.After the modification,MWCO decreased to 1750 Da.At the same time,the corresponding membrane pore size reduces to 2.1 nm,resulting from the presence of the thin single-molecule modification layer.It should be noted that the decrease in the ceramic membrane pore size is caused by the plugging of the aminosilane single-molecule layer in the membrane pore,which can significantly reduce the total porosity of the separation layer.Therefore,the water flux through the modified ceramic membrane inevitably decreased owing to the enhanced mass transfer resistance.Furthermore,a ZrO2ceramic membrane with a smaller pore size was used as a contrast membrane to prove that electrostatic repulsion played a decisive role in the cationic dye rejection performance of the modified membrane.As shown in Fig.6,the water permeance of the contrast membrane was approximately 350 L·m-2·h-1·MPa-1and the MWCO was 1244 Da.

Fig.4.The thermal resistance of the modified membrane.(a) TG/DSC curves of the KH-551 crystal.(b) XPS spectra of the modified membrane (KH-551 concentration=30 mmol·L-1,modification time=4 h) before and after the heat treatment.

Fig.5.The characterization of membrane surface morphology.(a) The Schematic diagram of grafting modification.(b) Surface FESEM images of the original ceramic membranes and the membranes modified with different concentrations of the KH-551 solution ((c) 10 mmol·L-1,(d) 30 mmol·L-1,(e) 50 mmol·L-1 and (f) 70 mmol·L-1).(Modification time=4 h).

The modification time and concentration are important factors determining the grafting density of KH-551 on ceramic membranes.The retention of MgCl2solution (1 g·L-1) was used to analyze the relationship between the separation performance of the modified membranes and the modification conditions.As shown in Fig.7(a),with an increase in the modification time,the rejection of MgCl2increased gradually,while the salt solution permeance gradually decreased.At a modification time of 4 h,the MgCl2retention reached 85.3%,and the permeance decreased by approximately 51.4%compared with those of original membranes.The separation performance of the modified membrane tended to be stable even with the increase in the modification time.This result can be attributed to the availability of a limited number of hydroxyl sites on the ceramic membrane surface for the grafting modification.Fig.7(b)illustrates the relationship between the salt retention performance of the modified membrane and the modification solution concentration.When the modification concentration reached 30 mmol,the MgCl2retention increased to 80.1%,while the permeance decreased to 111 L·m-2·h-1·MPa-1.With further increase in the modification concentration,the salt retention performance rarely changed because of the saturation of the single-molecule modification layer.This result is consistent with the change in the element content with the modification concentration shown in Fig.3(c).In conclusion,4 h and 30 mmol·L-1were considered as the optimal conditions for the modification of the ceramic membranes for further performance characterization.For ceramic NF membranes,this improvement in the MgCl2rejection performance is excellent.In previous reports,the rejection performance of ceramic NF membranes toward divalent cations (MgCl2or CaCl2,neutral pH,and approximately 1 g·L-1) did not exceed 40% at neutral pH [17,28,29].

Solutions of four salts were used to characterize the salt retention of the original ceramic membrane and modified membrane.As shown in Fig.8,the salt solution permeance decreased by approximately 50%after the grafting modification because of the decrease in the membrane pore size.The MgCl2,MgSO4,and NaCl retention rates of the modified membranes were higher than those of original membranes,while the Na2SO4retention rate decreased significantly.The rejection of MgCl2was higher than 80%,while the rejection of MgSO4and Na2SO4was only 8%–20%.This can be explained by the conversion of the surface charge of the ceramic membrane at a neutral pH from negative to positive after the modification,which is more beneficial for the retention of divalent cations than divalent anions.The salt retention rates of the modified membranes followed the order:MgCl2＞NaCl ＞MgSO4＞Na2SO4.Because the radius of hydrated ions is smaller than that of the modified membrane pore,electrostatic repulsion plays a dominant role in the retention process of the salt solutions [30,31].In addition,the adsorption of sulfate anions on the membrane surface can significantly reduce its positive charge [32].Therefore,the retention rates of MgCl2and NaCl were higher than those of MgSO4and Na2SO4.Furthermore,the modified ceramic membrane showed stronger electrostatic repulsion to Mg2+than Na+because of the higher valence state of Mg2+.Therefore,salt rejection occurs in the following order:MgCl2＞NaCl and MgSO4＞Na2SO4.

3.6.Anti-adsorption performance of modified membranes toward methylene blue

To directly verify the electrostatic repulsion of the modified ceramic membrane surface toward cationic dyes,a static adsorption experiment using methylene blue(0.1 g·L-1)as the adsorption solution was carried out for 7 days.Photos of the ceramic membrane after dye adsorption experiment and the UV–Vis spectrum of the soaking solution are provided in the Fig.S2.A comparison of the dye adsorption rates for the ceramic membranes before and after the modification is illustrated in Fig.9.After soaking in the dye solution for 7 days,approximately 80% of the dye molecules were adsorbed on the original membrane.However,the adsorption rate for the modified membrane toward methylene blue was only 25%after 7 days.The adsorption capacity of the original membrane toward methylene blue was more than three times that of the modified membrane,because of the stronger positive charge of the modified membrane.In conclusion,the modified membrane surface exhibited stronger electrostatic repulsion toward methylene blue molecules in the dye solution than the original membrane.

3.7.Membrane separation properties toward dyes

Four dye solutions,namely,zwitterionic (rhodamine B (RB)),cationic(methylene blue(MB)and basic green 4(BG4)),and anionic(reactive blue 19(RB19)),were used to assess the dye rejection performance of the tight ceramic UF membranes before and after the grafting modification;the results are displayed in Fig.10.The properties of the four dyes are listed in Table 3.The tight ceramic UF membrane (MWCO=2.3 kDa) exhibited good separation performance for reactive blue 19 due to the electrostatic repulsion and the more serious aggregation of reactive blue 19 in water than that of cationic dyes.The retention and permeance were 94.5%and 313 L·m-2·h-1·MPa-1respectively.However,the separation performance toward the cationic dye solutions (methylene blue and basic green 4) and the zwitterionic dye solution (rhodamine B)was unsatisfactory because the original membrane surface had a weak negative charge and the molecular size of the cationic dyes was smaller than that of the reactive dyes.This result is consistent with that reported in the literature[10].The retention of the modified membrane to the reactive blue 19 solution(98.8%)was higher than that of the original membrane because a small amount of negatively charged reactive dye molecules passing through the membrane pores was adsorbed by the positively charged membrane.At the same time,the separation performance of the modified membrane toward the other three dye solutions was improved by varying degrees,according to the charge strength of the dye molecules.The rejection toward methylene blue increased from 18.5% to 78.5%,and that toward basic green was enhanced to 72.4% after the grafting modification.However,the retention of the modified membrane toward rhodamine B was unsatisfactory and increased from 16.9% to 27.9%.This result can be explained by the different charge strengths of the dye molecules.The cationic dyes (methylene blue and basic green 4)have a strong positive charge because of the presence of a quaternary ammonium cation ([NR4]+) [13].Although rhodamine B also has a quaternary ammonium cation,the carboxylate group in the molecular structure neutralizes its positive charge,causing it to be slightly negatively charged [33].The electrostatic repulsion of the modified ceramic membrane toward the rhodamine B molecule was rarely improved.The molecular size of rhodamine B is 1.8 nm[34],which is smaller than the pore size of the modified membranes.Therefore,rhodamine B molecules can easily pass through the modified ceramic membrane pores.The zeta potentials of the four dyes were characterized by the dynamic light scattering technique;the detailed data are shown in Fig.S3.Due to the limit of pore size and surface charge,the rejection performance of ceramic membranes without any modification toward cationic dye rejection did not exceed 30% in previous reports[10,11].For ceramic membranes,the modification can improve the cationic dye rejection performance significantly.Although the rejection of modified ceramic membrane to cationic dyes can only reach about 80%,this study provides valuable guidance for the adjustment of the ceramic membrane surface charge for treating industrial cationic dye wastewater.If the modification method can be applied to ceramic membranes with smaller pore size in future research,the cationic dye rejection performance of the modified membrane is expected to be more significantly improved.

Table 3 Characterization of dyes in the study

Fig.7.The separation performance of the modified membrane to MgCl2 solutions (1 g·L-1,0.5 MPa,pH 6.8 and 25 °C) with (a) different modification time (KH-551 concentration=30 mmol·L-1) and (b) concentrations of the KH-551 solution (modification time=4 h).(P:permeance,R:rejection).

Fig.8.The salt(1 g·L-1,0.5 MPa,pH 6.8 and 25°C)separation performance.(a)The salt permeance and(b)rejection of the original membrane and the modified membrane(KH-551 concentration=30 mmol·L-1,modification time=4 h).

Fig.9.Time-dependent adsorption rate of the original membrane and the modified membrane (KH-551 concentration=30 mmol·L-1,modification time=4 h) to methylene blue.

Fig.10.The dye (0.1 g·L-1,0.2 MPa,pH 6.8 and 25 °C) separation performance.(a) The dye permeance and (b) rejection of original ZrO2 membranes and modified ZrO2 membranes (KH-551 concentration=30 mmol·L-1,modification time=4 h).(Rhodamine B (RB),methylene blue (MB),basic green 4 (BG4)),and reactive blue 19 (RB19)).

A methylene blue solution was used to compare the cationic dye rejection performance of the modified membrane with that of the contrast membrane and the result was illustrated in Fig.11.The contrast ZrO2membrane had a smaller MWCO(1244 Da) than the original ZrO2membrane (2300 Da) and the modified ZrO2membrane(1750 Da).The methylene blue rejection of the contrast membrane was still unsatisfactory,at approximately 15%.The modified membrane with a larger pore size exhibited higher cationic dye rejection of 78.5%.Electrostatic repulsion played an important role in the filtration process with ceramic membranes fora methylene blue solution.Although the grafting modification caused the reduction of the pore size of the ceramic membrane,the enhanced positive charge of the surface played a decisive role in the cationic dye rejection performance of the modified membrane.

Fig.11.The separation performance of original membrane,modified membrane(KH-551 concentration=30 mmol·L-1,modification time=4 h) and contrast membrane to methylene blue solution (0.1 g·L-1,0.2 MPa,pH 6.8 and 25 °C).

3.8.Stability of modified ceramic UF membranes

The stability of the modified membranes is vital,as it determines whether the modified membrane can meet the requirement of continuous operation.In this work,the filtration (0.2 MPa,25 °C,and 100 h) of a methylene blue solution (0.1 g·L-1) as the feed was performed to assess the stability of the grafting modification,the results are illustrated in Fig.12.After filtration for 20 h,the dye solution permeance of the modified membrane decreased from 130 to 30 L·m-2·h-1·MPa-1,resulting from the increase in membrane fouling with the increasing operating time.It should be noted that the filtration experiment is discontinuous.The drying and curing of dye contamination on the ceramic membrane surface in air can aggravate the membrane contamination in the next filtration cycle.Furthermore,the contaminants on the modified membrane surface can shield its positive charge,leading to a slight decrease in dye retention.The flux at a time of 80 h is higher than that at 70 h and 90 h,which may be caused by a slight change in the feed temperature.Because of the small membrane effective area and the low stable dye permeance,the slight permeance increase in a filtration cycle could lead to obvious abnormal permeance results,while not affect the display of experimental regular.In the filtration process,the retention of the methylene blue solution by the modified membrane decreased in the beginning but then tended to be stable(approximately 56%).In summary,the rejection of the modified ceramic membrane can to be stable from 40 h in the long-term filtration process.However,membrane fouling significantly degrades the separation properties of modified membranes to methylene blue due to the shielding effect of pollutant.

Fig.12.Separation performance stability of the modified ceramic membrane (KH-551 concentration=30 mmol·L-1,modification time=4 h) to methylene blue solution (0.1 g·L-1,0.2 MPa,pH 6.8 and 25 °C) throughout 100 h.

4.Conclusions

A positively charged ceramic membrane was successfully fabricated by grafting KH-551 onto a tight ZrO2UF membrane.The grafting modification slightly reduced the pore size of the ceramic membrane (by approximately 0.3 nm).The introduction of amino functional groups increased the isoelectric point of the ceramic membrane from pH 6.9 to pH 7.4.The enhanced positive charge of the membrane surface improved the rejection performance of the modified ceramic membrane toward divalent cations and cationic dyes.The rejection of the modified membrane to MgCl2increased from 27.8%to 83.3%,and the rejection to methylene blue increased from 18.5%to 78.5%compared with those of the original membrane.Moreover,the modified membrane showed excellent anti-adsorption performance to methylene blue.The static adsorption rate of methylene blue decreased from 80%to 25%.The membrane fouling in the filtration process can affect the rejection performance of the charged modified ceramic membranes toward methylene blue.However,the modified ceramic membrane exhibited stable separation performance toward the methylene blue solution in the filtration experiment for 100 h.

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 study was financially supported by the Project for Natural Science Research of Jiangsu Higher Education Institutions(20KJA530001),the National Natural Science Foundation of China(22078147,21808107),the Natural Science Foundation of Jiangsu Province (BK20180163) and the Research Project of National Synthetic Biotechnology Innovation Centre (TSBICIP-KJGG-002-16).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.11.007.