Anionic channel membrane encircled by SO3H-polyamide 6 particles for removal of anionic dyes

Anionic channel membrane encircled by SO3H-polyamide 6 particles for removal of anionic dyes

Author’s Accepted Manuscript Anionic Channel Membrane Encircled by SO3Hpolyamide 6 Particles for Removal of Anionic Dyes Shun Ren, Dongqing Liu, Yingb...

NAN Sizes 0 Downloads 0 Views

Author’s Accepted Manuscript Anionic Channel Membrane Encircled by SO3Hpolyamide 6 Particles for Removal of Anionic Dyes Shun Ren, Dongqing Liu, Yingbo Chen, Shulin An, Yiping Zhao, Yufeng Zhang www.elsevier.com/locate/memsci

PII: DOI: Reference:

S0376-7388(18)31612-0 https://doi.org/10.1016/j.memsci.2018.10.025 MEMSCI16541

To appear in: Journal of Membrane Science Received date: 12 June 2018 Revised date: 17 September 2018 Accepted date: 8 October 2018 Cite this article as: Shun Ren, Dongqing Liu, Yingbo Chen, Shulin An, Yiping Zhao and Yufeng Zhang, Anionic Channel Membrane Encircled by SO 3Hpolyamide 6 Particles for Removal of Anionic Dyes, Journal of Membrane Science, https://doi.org/10.1016/j.memsci.2018.10.025 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Anionic Channel Membrane Encircled by SO3H-polyamide 6 Particles for Removal of Anionic Dyes Shun Rena, Dongqing Liua,*, Yingbo Chena,b, Shulin Ana, Yiping Zhaoa, Yufeng Zhanga,c a State Key Laboratory of Separation Membranes and Membrane Processes, School of Material Science and Engineering, Tianjin Polytechnic University, 300387 Tianjin, China b Yiwu Huading Nylon Co.Ltd., 322000 Jinhua, China c School of Environmental and Municipal Engineering, Tianjin Chengjian University, 300384 Tianjin, China *Corresponding author. Email: [email protected] Abstract Membranes of anion channels in matrix were constructed through amphiphilic self-assembly of polyelectrolyte, SO3H group containing polyamide 6 (SPA6). SPA6 molecules self-aggregated into 255 nm micelles with negative charges at the surface in casting solution. Micelles switched into nanofibers, which formed pompoms of divergent arrangement of nanofibers from the center after immersed in coagulation bath. They adhered and piled up to construct the membrane. Charged groups were remained at the surface of pompoms during solidification. The gaps between spheres provided channels full of anions, which supplied great repulsion to anions. It endowed membranes the potential to remove anion dyes from aqueous solution. The rejection was up to 99.73% for 500 mg/L Congo red and 85.68% for 100 mg/L Acid blue 93 and 63.06% for 100 mg/L Methyl orange after first filtration. Total removal percentage was 95.48% for Acid blue 93 and 79.51% for Methyl orange after second filtration. Self-aggregation of dye molecules also helped their high removal efficiency. Flux of film 30M of 100 mg/L CR solution dropped to 67.34% and rejection rate was higher than 95% in 1500 min. Flux could be brought back to more than 90% of the original values after washed by deionized water, suggesting that 30M had good anti-fouling capacity. SPA6 membranes displayed a potential for effective treatment of textile wastewater. Keywords: Anionic charged membrane; SO3H containing polyamide 6; anionic dyes removal; anti-fouling.

1. Introduction Water pollution of organic compounds is attracting more and more concerns since it consistently grows stemming from industrial, agricultural and urban human activities. Dyes are a kind of persistent organic water pollutants coming from textile industry, which could take a serious environmental health problem due to their toxicity and potential hazardous effects (carcinogenicity, mutagenicity and bactericidality) on living organisms, including human beings [1-4]. Their removal is a challenging work from wastewater treatment. Conventional techniques such as adsorption [5], coagulation [6], biological degradation [7] and oxidation [8] are 1

widely applied, but they all have own limitation [9]. Membrane technique, especially nanofiltration (NF), has been increasingly investigated for colors removal because of its energy efficiency, manufacturing scalability and compact design [10-14]. However, NF was easily polluted by organic compounds in long time operation. A lot of efforts were contributed to improve its anti-fouling ability, such as modification of hydrophilicity [15,16], self-cleaning [17,18], photocatalytic [19] and photodegradation [20,21]. Charged groups introduction in membrane structure had been mentioned in several outstanding works, recently. It provided some inspiration in the development of membranes of low operation pressure and high rejection. He [22] group prepared 3-aminopropyltriethoxysilane (APTES) grafted halloysite nanotubes (A-HNTs) at first and then blended it into PVDF casting solution to fabricate positively charged membrane. A-HNTs contained hydrophilic groups, hydroxyl and amino, which helped to accelerate mass transformation between solvent and nonsolvent. It increased the thickness of top surface layer, which led to form a loose NF membrane. Meanwhile, the electrostatic interaction between membrane surface and dye molecules played an important role on dye rejection. The removal rate of Direct Red 28 was up to 94.9%. The blend membrane also exhibited excellent rejection stability and reuse performances in several hours. Zhang and Liu [23] had also reported that charged groups could adjust the thickness of surface layer in similar work. They had prepared blend membrane of polyethersulfone (PES) and a negatively charged particle, SiO2-poly (sodium-4-styrene sulfonate) (PSS). The membrane was confirmed a loose/polyethersulfone (PES) NF membrane with high permeate flux and rejection for anionic dyes, as well as low rejection for inorganic salts [24]. It was believed that charged groups could adjust the process of membrane formation and help to obtain charged surface, which was contributed to high rejection for charges [25]. Chung [26,27] added hydrophilic sulfonated polyphenylenesulfone (sPPSU) into Torlon solution to fabricate Torlon&sPPSU hollow fiber membrane through nonsolvent induced phase separation (NIPS) method. sPPSU could tailor membrane formation process, hence optimize membrane pore size and permeability. As a result, negative charge properties were provided on the membrane surface. A dense and about 150 nm nodule-like sublayer was observed underneath the inner surface, which was also mentioned in other charged groups containing membranes above. The loose membrane showed high pure water flux of 80 L·m−2·h−1·bar−1 with small pore size of 1.0 nm and low dyes rejection. After modified by hyperbranched polyethylenimine (PEI) at surface, the membrane displayed high permeation fluxes of 7.0−71.2 L·m−2·h−1 and rejections of 95.5−99.9% to various dyes under 1 bar. SO3H-polyamide 6 is an important class of chemical fiber materials. Sulfo group endowed PA6 dyeability by cationic dyes through combination of SO3H group with cations in dye molecules [28]. Therefore, several commercial SO3H-polyamide 6 were called cationic dyeable PA6 (CD-PA6) in chemical fiber field. A popular CD-PA6 was fabricated by melting polymerization of ε-caprolactam, 5-sulfobenzene-1,3-dicarboxylic acid and a diamine monomer [29]. CD-PA6 fiber was often used to manufacture bright clothing fabric. Carpets made from it were 2

fouling resistant, especially for acidic fruit juices. SO3H groups prevented the combination of acid dye in juice and fiber. It provides inspiration to develop fouling resistant membrane material using SO3H-polyamide 6. Charged particles could regulate membrane formation process through intermolecular interaction in casting solution. In this work, a negative charge containing polyamide 6, SO3H- polyamide 6 (SPA6) was chosen to prepare membranes. SPA6 was synthesized from ε-caprolactam and 5-sulphosalicylic acid (5SA). The behavior of such a polyelectrolyte in formic acid solution and its influence on film formation process were investigated. SPA6 exhibited actions of amphiphilic self-assembly, which adjusted the structure of the membranes and their performance of water treatment. Self-aggregation tailored membrane formation process and thus optimized membrane pore structure, in addition to introduce negative charge properties on the membrane surface and inside the pores. The membrane showed good rejection to anion dyes, such as Congo red. The work provided a new method to fabricate membrane with ion channels. 2. Experimental 2.1 Materials ε-Caprolactam (CPL), 5-sulphosalicylic acid (5SA), formic acid, manganese chloride tetrahydrate (MnCl2·4H2O), Congo red (CR), Acid blue 93 (AB93) and Methyl orange (MO) were purchased from Tianjin Guangfu Technology Development Co., Ltd. 6-Aminocaproic acid, bovine serum albumin (BSA) and hypophosphorous acid (HPPA) were supplied by Aladdin Industrial Corporation. Synthesis of SO3H containing polyamide 6 (SPA6) was described in detail in previous work [30]. All materials were used as received and without further purification. 2.2 Instrument The X-ray photoelectron spectroscopy (XPS) analyses were performed with a Thermofisher K-alpha spectrometer using a focused monochromatized Al Kα radiation. The surface area and mean pore size of membrane was tested by Brunauer-Emmett-Teller (BET) and Barrett-Joyner-Halenda (BJH) method. A field emission Scanning Electronic Microscopy (SEM) (S-4800, Hitachi, Japan) was used to collect surface and cross-section morphology of membranes. Water contact angle (WCA) measurements were performed by a contact angle meter (Drop Shape Analysis 100, Kruss BmbH Co., Germany). Atomic force microscopy (AFM, Agilent-S5500) was used to test roughness of the membranes. The tensile strength of membranes was investigated at room temperature using a tensile tester (JBDL-200N, China) at a drafting speed of 1 mm/min. Dynamic Light Scattering (DLS) was used to measure particle size and distribution by a ZS-90 laser light scattering spectrometer (Malvern Instruments Ltd), with a digital correlator at a scattering angle of 90°. Transmission Electron Microscope (TEM) images were acquired through a Hitachi H7650 transmission electron microscope with an accelerating voltage of 100 kV. The membrane surface charge properties were characterized by streaming zeta potential 3

measurements tested by a SurPASS electrokinetic analyzer (Anton Paar, Austria). UV-Vis spectrum of aqueous solution was measured by ultraviolet spectrophotometerTU-1901 (Beijing Purkinje General Instrument Co., Ltd.). 2.3 Preparation of SPA6 membranes A three-factor, four-level orthogonal experiment was designed and listed in Table 1. Certain amount (28 wt%, 29 wt%, 30 wt%, 31 wt%) of SPA6 resins were dissolved in 88 wt% formic acid. The solution was poured on a glass plate and scraped into film by a casting stick of certain edge thickness (100 μm, 150 μm, 200 μm, 250 μm). The membranes were stayed in air at 25℃ and 50% humidity for seconds (90s, 100s, 110s, 120s) and then immersed into coagulation bath (25℃) until peeling off from glass plate. Membranes were rinsed by deionized (DI) water and then immersed in DI water for use. Table 1. Orthogonal experiments Solid content

Time

Thickness

(wt%)

(s)

(μm)

1

28

90

100

2

28

100

150

3

28

110

200

4

28

120

250

5

29

90

200

6

29

100

250

7

29

110

100

8

29

120

150

9

30

90

150

10

30

100

100

11

30

110

250

12

30

120

200

13

31

90

250

14

31

100

200

15

31

110

150

16

31

120

100

2.4 Pure water flux Pure water flux of membrane was measured by a cross-flow module at 25℃and 0.1MPa. 0.5 h Prepress was performed using DI water at 0.1 MPa before testing. Water flux (F) was calculated by equation (1) [31]: 𝐽 𝐹= (1) 𝐴𝑇 Where J isthe volume of permeated water (L); A is effective area of membrane 2 (m , 7.07 cm2); T is the time used by permeating a fixed volume water (h). 2.5 Dye rejection Three anion dyes, CR, AB93 and MO were chosen to evaluate dye rejection of 4

membranes. The concentrations of dyes solution were determined by UV-vis. Rejection was calculated by equation (2) [22]: 𝑅(%) = (1 −

𝐶𝑝 ) × 100 (2) 𝐶𝑓

Where Cp and Cf are concentrations (mg/L) of dyes solution in the permeate and feed side, respectively. 2.6 Long-term performance test Long-term test of membrane was conducted to investigate flux stability using 100 mg/L CR aqueous solution. It was evaluated under a total recirculation mode. The flux and rejection were recorded every 60 min for 1500 min. 2.7 Anti-fouling performance measurement The anti-pollution capacity of membrane was performed with 100 mg/L CR aqueous solution. Permeate flux was recorded every 60 min. After filtration for 240 minutes, the membrane was washed by DI water for 60 min. The whole tests included three filtration and washing cycles. Flux recovery ratio (FRR) was calculated by equation (3) [32]: 𝐽𝑤.2 𝐹𝑅𝑅(%) = ( ) × 100 (3) 𝐽𝑤.1 Where Jw.1 and Jw.2 is the initial flux of first and second cycles, respectively. 3. Results and discussion 3.1 Membrane characterization Nanofiber mats of SPA6 of various amounts of 5SA were prepared in previous work [30]. It was surprisingly found that the nanofibers could neither be coloured by cationic dyes nor anionic dyes, even they were boiled at 95℃ in 1g/L methylene blue (BB9) solution or AB93 for 24 h. It attracted us to develop SPA6 a membrane material for dye waste water treatment. Table 2. Membranes properties of different 5SA content S/N

WCA

BSA-Water flux

The roughness of the

(wt/wt)

(º)

recovery rate (%)

membrane (nm)

0%

0

89

11

75.9

1.25%

0.062

74

48

61.2

2.5%

0.113

66

62

44.5

A series of films were prepared which contained different amount of 5SA units in polyamide 6. The ratio of elements S/N increased as 5SA in SPA6 and WCA decreased, meanwhile (Table 2). The increase of BSA-water flux recovery rate indicated that SO3H improved anti-fouling capacity of membrane. The average 5

roughness of surface decreased from 75.9 nm to 44.5 nm as 5SA increasing, which indicated that SO3H smoothed the surface through adjusting the phase separation process. However, the membranes were too brittle to treat water when 5SA more than 2.5 wt%. In this work, 2.5% 5SA-contained polyamide 6 was used to prepare a series of membranes for treatment of anionic dye aqueous solution. The properties of membranes and their separation mechanism were investigated. Here, SPA6 referred to 2.5% 5SA-contained polyamide 6. Table 3. Recipes of coagulation bath DI water(ml)

Ethanol(ml)

Formic acid(ml)

1

100

0

0

2

80

20

0

3

77

23

0

4

76

24

0

5

75

25

0

6

74

26

0

7

73

27

0

8

70

30

0

9

74

26

5

10

74

26

10

11

74

26

15

12

74

26

20

Fig. 1. Photos and surface SEM images of membranes from different coagulation bath (DI water/ethanol/formic, v:v:v).

It was found that recipe of coagulation bath was a crucial factor to prepare a smooth and usable film. Experiments were listed in Table 3. Fig. 1 displayed the influence on membranes and their micro-morphology at casting solution concentration of 20 wt% . The ratio of DI water/ethanol/formic acid was determined 6

as 74:26:10 (v:v:v) for coagulation bath. Ethanol adjusted the polarity of liquid system. Formic acid slowed down the dramatic solvent exchange between casting solution and nonsolvent. They both smoothed phase conversion and helped to obtain crease free films. Table 4. WCA and elements content of membranes surface WCA

XPS Data

(°)

S2p

S/N

22 %

61.5

0.27

0.0236

26 %

50.1

0.31

0.0270

30 %

40.6

0.34

0.0294

20 % casting solution membranes were too fragile to keep intact after test. Higher concentrated casting solution were investigated. As SPA6 concentration increasing, the ratio of S/N increased (Table 4), which illustrated that surface S element content increased as casting solution concentration. The more concentrated the casting solution were, the more SO3H groups were on the surface. WCA decreased as the increase of the concentration. Membrane surface became more hydrophilic as concentration increasing. In order to obtain strong enough film for water treatment, concentration should higher than 27 %. 3.2 Evaluation It was noticed that durability of films was also related to staying time and thickness. The new scraped film needed to stay a period time in air. It helped solvent adequately volatilizing out to achieve sturdy films. Such a surface enhanced density and strength of films. An orthogonal experiment of 3 factors 4 levels was designed to optimize membrane preparing process including solid content of casting solution, staying time in air and membrane thickness as factors (Table 1). Table 5. Pure water flux of membranes in orthogonal experiments

200μm 250μm

28%

29%

30%

31%

16.98

135.88

14.44

115.50

(95.23)*

(70.57)

(99.21)

(86.43)

16.14

88.32

13.59

129.09

(95.38)

(72.23)

(98.27)

(84.61)

*Data in brackets were rejection rate (%) of 100 mg/L CR solution. The unit of flux was L·m-2·h-1.The operation pressure was 0.1 MPa.

According to the result of orthogonal experiments, 16 films were prepared and only 8 films were strong enough for application test. Table 5 concluded that suitable thickness guaranteed membranes strength to take evaluation tests. Membranes thinner than 200 μm were too fragile to evaluate. 8 Testable membranes were all 200 μm or 250 μm. 7

Table 6. Pure water flux of membranes in orthogonal experiments 28% 90s

100s

110s

120s

29%

30%

31%

135.88

129.09

(70.57)*

(84.61)

88.32

115.50

(72.23)

(86.43)

16.98

13.59

(95.23)

(98.27)

16.14

14.44

(95.38)

(99.21)

*Data in brackets were rejection rate (%) of 100 mg/L CR solution. The unit of flux was L·m-2·h-1.The operation pressure was 0.1 MPa.

Table 6 showed that staying time had great impact on pure water flux. As time elongation, flux declined in the whole. There was a sudden drop between 100s and 110s. The amount of solvent volatilization increased as staying time elongating, which helped to form a compact surface. At the same staying time, the larger the SPA6 concentration, the smaller the flux in the whole. Data of yellow lattice in Table 5 and green one in Table 6 were slightly contrary to the general tendency. It demonstrated that the synergism of staying time and thickness of membrane were both important at higher casting solution concentration (30 % and 31 %).

Fig. 2. Surface and cross-sectional SEM images of different staying time: 90s (a, e), 100s (b, f), 110s (c, g) and 120s (d, h).

SEM images revealed that there were many gap cracks and crevices on film of 90s and 100s (Fig. 2), which provided higher flux. Film of 120s had smaller and narrower crevice. Adequate staying time ensured the formation of compact skin layer, which provided enough intensity. Longer staying time led to more volatization of solvent. This condensed membrane structure and enhanced its strength, but decreased water flux. Tensile strength of 90-120s membranes were 1.60 MPa, 1.75 MPa, 2.20 MPa and 2.95 MPa, respectively. From the images of cross-section, films displayed a 8

visage of particles piling-up. Particles adhered to each other and at last joined into string or merged into a piece. 120s film looked the most compact and the least pores, which also gave the smallest water flux.

6μm

Fig. 3. Cross-section SEM images and S elemental mapping by EDS of different solid concentration (22 % (a, d, g), 26 % (b, e, h), 30 % (c, f. i) ) membranes.

Basing on orthogonal experiments, SPA6 concentration in casting solution was found more important and should be discussed carefully. So, 200 μm membranes of various casting solution concentration were prepared under a fixed air staying time of 120s. Membranes clearly displayed micro-morphology of accumulation of different sizes particles (Fig. 3). 22 % membrane was composed of 10 μm particles, whose surface had holes and depressions. Particle size of membrane gradually diminished from 10 μm to 1 μm as concentration increasing. A skin layer appeared finally. When solid content up to 30 %, a 6 μm skin layer was obviously observed. It was attribute to evaporation of solvent of top surface. The layer was compact and uniform, which improved strength of granular-stacked structure. SPA6 is polyelectrolyte and could amphiphilicly self-assembly in solution. It could be deduced that SPA6 molecules aggregated into micelles with SO3H groups at surface. Although micelles merged during the process of phase conversion, the polarity difference between casting solution and coagulation bath kept SO3H remaining at particle surface. High concentration solution provided more SO3H groups to diminish surface tension, which helped to decrease the size of aggregates. The more SPA6 in casting solution, the smaller the particle size and the more compact 9

the film. SEM-EDS also confirmed this by denser S density in cross-section of 30% film. SPA6 solution were investigated by DLS and TEM later to verify the inference (Fig. 4).

a

b

c

d

2 microns

2 microns

e

f

2 microns

2 microns

Fig. 4. Average diameter of particles of 0.5 wt% SPA6 solution measured by DLS (a); Zeta potential of membranes prepared by different concentration of casting solution (b); TEM of 0.5 wt% formic acid solution of SPA6 (c); TEM of 10 wt% SPA6 solution treated by a drop of coagulation bath liquid immediately (d), after air dried 30s (e) and 60s (f).

A dilute SPA6 formic acid solution was measured by DLS. There were 255 nm micelles in 0.5 wt% solution (Fig. 4a), which could be also observed by TEM (Fig. 4c). It was very difficult to catch separated particles in higher concentration samples. Fig. 4d showed an image when a drop of coagulation bath just dropped on 10 wt% SPA6 solution. SPA6 aggregates appeared like pompoms of 10 μm or larger. Pompoms were composed of countless nanofibers protruding out of a center. Conversion of solvent polarity transformed micelles into nanofibers during blend process of two liquid phase. There were still some black particles in the picture, which were unconverted or unfully transformed micelles. They had already become larger and the edge began to blur. Under such a low concentration, pompoms were 10

separative and fully developed. In denser solution, micelles were very close to each other. Nanofibers intertwined and overlapped, making the boundary no longer clear. It illustrated the formation of particle-piled up morphology of membranes in SEM. Fig. 4e showed an image that 10 wt% SPA6 solution was treated by coagulation bath after staying in air for 30s. Nanofibers merged and shrank towards the central. The number and the size of black flacks increased. As solvent volatilizing, micelles at the top layer clustered and adhered to each other, and then dried. They could not successfully switch into pompoms while treated by coagulation bath any more. When the air dry time was 60s, there was almost no pompoms observed (Fig. 4f), only large flakes left. Hence, the density of surface layer was higher than the bulk matrix. It demonstrated the formation of the compact skin layer. It could be deduced that the gaps and crevices were encircled by particles in film bulk, which were channels full of SO3H groups. The surface of the membranes were no doubt negative charged, either. This was confirmed by zeta potential curves of membrane surface (Fig. 4b). As solution concentration increasing, pH of isoelectric points of films decreased. As a result, SPA6 membranes possessed the potential to reject negative charged particles. SO3H SO3H SO3H SO3H SO3H SO3H

a SO3H

SO3H

I

Self-assembly

SO3H SO3H SO3H SO3H

SPA 6

SO3H

SO3H

SO3H

= II

SO3H

SO3H SO3H SO3H SOSO 3H H 3 H SO3H SOSO 3H3 SOSO 3H H 3

b Coagulation Liquid

Coagulation Liquid

= III

Transition State

c

Skin Layer

Immersed in Coagulation Bath

IV

Immersed in

Stay in Air for 120s

Concentrated Solution

Coagulation Bath

V

Scheme 1 Self-assembly of SPA6 (a); Aggregates transformation induced by solvent polarity (b); Formation of skin layer (c).

Scheme 1 demonstrated the process of membrane formation. SPA6 self-associated in solution. Micelles of SPA6 (Scheme1a I) was crew-cut style (II) with SO3H groups at the outer surface. Such kind of aggregates had potential to switch into various shapes when induced by polarity change. It was believed that SO3H groups should take the end position of polyamide chains because 5SA amount played a great role on the molecular weight of SPA6 according to previous work. As concentration increasing, micelles became crowded and tended to merge into large particles (Transition state). Once accumulated micelles going into coagulation bath, they deformed and merged to form nanofibers. These nanofibers aggregated into dendrimer-like pompoms (Scheme1b III). In concentrated solution, pompoms interwined and adhered to each other, finally solidified by nonsolvent. The particle 11

piling-up membranes formed (Scheme1c IV). Solution concentration and air-dry process greatly impacted on the formation of skin layer. If solution concentration was dense enough, micelles would concentrate and merge, and even solidify during air drying. When immersing in coagulation bath, pompoms would not be successfully developed at the surface area. Therefore, the compact skin was obtained (Scheme1c V), which played a very crucial role on rejection.

charged particle

Charged Solution

V

- - - - - - - -- - - - -- - - --- - - - -- - - - - --- - - -- -- -- - --- -- ---- - --- - - -- - - - - - -- -- - - - - - -- - - - - - - - -- -- - -

SPA6 membrane Scheme 2 The separation mechanism

Scheme 2 was the separation mechanism of SPA6 membrane for anion aqueous solution. When liquid permeated the membrane, it would go through a channel full of negative charges. Particles with the same electricity were repulsed and left at the feed side. Only water and neutral small particles could pass through. Rejection of membranes in orthogonal experiments were tested using 100 mg/L CR solution under 0.1 MPa by a cross-flow module. From Table 5 and Table 6, it was concluded that the trend of rejection ratio were contrary to pure water flux at 0.1 MPa. Compact film provided high rejection but low flux. Especially, the membrane (30 % casting solution, 120s, 200 μm) showed the highest rejection (99.21 %) and almost the lowest flux (14.44 L·m-2·h-1). This membrane was named 30M, which was fully evaluated about its anionic dyes rejection. Fig. 5a displayed that flux declined from 13.87 L·m-2·h-1 to 6.51 L·m-2·h-1as CR concentration increasing. The rejection was even up to 99.73 % for 500 mg/L solution. The effluent were colorless. At higher concentration than 500 mg/L, the flux would drop down sharply and the rejection ratio also declined. The effluent of 1000 mg/L was pink. 30M was also used to treat the other two anionic dyes, AB93 and MO. Performance was the same as CR’s , but rejection ratios were lower than CR’s. For AB93 (Fig. 5b), rejection increased from 85.68 % to 91.38 % as dye concentration rising from 100 mg/L to 500 mg/L, accompany with slight declining of flux. For MO (Fig. 5c), rejection ratio greatly increased from 63.06 % to 82.66 % as from 100 mg/L to 500 mg/L , while the flux slightly declined. After second treatment for the effluents of AB93 and MO (Fig. 5d), the total rejection ratio could get to 95.48 % and 79.51 %, respectively. From above, it was also known that higher dye concentration could obtain better removal rate in a suitable range. The membrane is high effective for CR 12

remove. However, the efficiency of the first round treatment for the other two dyes was not so high as CR’s. The colors were lighter but not disappeared. Repeated treatment was necessary to eliminate the color exhaustively.

a

b

c

d

Fig. 5. Flux and rejection ratio of 30M for dyes solution: CR (a), AB93 (b), MO (c) and rejection of twice filtration for 100 mg/L AB93 and 100 mg/L MO (d). The operation pressure was 0.1 MPa.

30M were determined as ultrafiltration (UF) membrane by BET. The average pore size was 7.807 nm and specific surface area was 10.062 m2·g-1. The data indicated that the films should have no ability to intercept dye molecules, which was normally contributed to the function of NF. In order to explain the dye removal efficiency of 30M, dye solutions were also investigated. 100 mg/L CR and MO solutions both showed Tyndall effect, but AB93 did not. DLS data confirmed that there were particles in dye solutions (Table 7). The particle sizes were all up to hundreds of nanometre and larger than average pore diameter of 30M. The aggregates still had negative charges around their peripheral because of uncomplexing SO3H. Size screening effect and electrostatic interaction were both important for dye removal. However, dye particles were associated by weak forces and would scattered under a strong enough external force, the operation pressure. Driven forces for self-assembly of dye molecules mainly included three factors known from their structure. The one was π-π stacking of aromatic parts, the other was electrostatic attraction between SO3H and amino groups. The third was hydrogen bonds. The later two were stronger. As far as the molecular structure concerned, these three dyes had different amino groups. They were primary, secondary and tertiary amines for CR, AB93 and MO, respectively. Besides electrostatic attraction, there were two hydrogen bonds between primary amino group and SO3H among CR 13

molecules, one for secondary amine AB93 and none for tertiary amines MO. This distinguished the complexing strength of SO3H and amino groups between dye molecules. The order of strength was CR>AB93>MO, corresponding with the number of hydrogen bonds. The attractive force between CR molecules was the strongest. Although CR had neither the most SO3H groups nor the largest aggregates. The complexed particles in feeding side were suffered from shear stress under operating pressure. They would deformed and be destroyed to various degree. Hence, some single molecules overcame weak complexing force and left aggregates, finally passing through the film. CR aggregates was the strongest and the highest rejection, accordingly. Beside driving force, combination stress was also increase as solution concentration. As a result, rejection slightly increased as dye concentration in a limited range. Table 7. Information of different dyes Particle size in pH of 100mg/L Dyes

Molecular structure

M.W.

dye solution dye solution nm

H2N O S O O

N N

Congo red

O

O S

(CR)

H H N

Na+

O

N N

S

N N

O

O

Na+

O

696.67

466.7

6.82

799.80

582.0

7.12

327.33

415.4

6.95

N H H N N

O O S Na+ O

N H H O O S O

Na+

HN

N O S O O

Acid blue 93

Na+ N H

O S OH

O O S O Na+

O

HN

(AB 93)

N O S O O Na+ NH

O S OH

O

Na+

Methyl

O O S O

N N N

orange

O O S Na+ O N N

(MO) N

As dye concentration increased to 1000 mg/L or more, flux and rejection ratio both decreased. There were four main forces acting on aggregates of dye molecule in 14

the solution. They were electrostatic repulsion (r), osmotic pressure (Π), operation pressure (P) and attraction between molecules in their aggregates (AA). The driving force to permeate film (Δ) could be described by equation (4): ∆= r − (Π + 𝑃) + 𝐴𝐴 (4) Where r is the function of the number of charges and AA is attraction between molecules in their aggregates. For a certain membrane material, r was a fixed value. Π was the function of solution concentration. AA also had relation with solution concentration to some degree and had a maximum. There should have an upper limit for concentration, at which Δ was zero. It was also a critical point that aggregates began to be destroyed. This needed discussion in detail later. 3.3 Stability in long-term operation and anti-fouling performances Long-term tests were conducted to investigate performance stability of 30M for dye removal using a 100 mg/L CR solution as model feed at 0.1 MPa. The filtration was operated under a total recirculation mode to examine long-term membrane performance. Fig. 6a displayed the evolution of permeation flux as a function of testing duration. 30M showed a decline of permeation flux over time. The flux dropped from 13.87 L·m-2·h-1 to 9.34 L·m-2·h-1 after a continuous operation of 1500 min. The final rejection was still 95.31%. The flux decline rate was 32.66%, indicating good performance stability and low fouling propensity. It demonstrated that the structure of 30M, piled by 1 μm particles, provided enough strength to retain performance stability in dye removal under 0.1MPa.

a

b

Fig. 6. Permeation flux and rejection of 30M as a function of CR concentration (a) and flux recovery of 30M in the removal of CR (b) The testing pressure was 0.1 MPa.

Sequential steps were operated including fouling by new prepared 100 mg/L CR solution and washing by DI water to investigate reuse and anti-pollution capacity of 30M. As shown in Fig. 6b, flux of 30M reduced rapidly along with filtration of CR solution at every cycle. Besides, flux had good recovery after being washed by DI water for 60 min. Table 8 presented the FRR in three cycles. FRR of 30M could still reach 91.84% after three cycles of fouling and cleaning. The anti-fouling capacity was closely related to improvement of hydrophilicity through introduction of SO3H and also repulsion of negative charged groups. The results indicated that 30M exhibited a 15

good anti-pollution property and had potential in practical application. Table 8. FRR data of 30M in three cycles Membrane

1st cycle FRR (%)

2nd cycle FRR (%)

3rd cycle FRR (%)

30M

95.92

93.88

91.84

Fig. 7. Different salts rejection of 30M in 120 min at 0.1MPa.

Fig. 7 gave the rejection data of 30M to four 1g/L salt solutions, NaCl, Na2SO4, MgCl2 and MgSO4. 30M showed higher rejection for divalent anionic species, MgSO4 and Na2SO4, than monovalent NaCl and MgCl2 at the initial stage. It reflected that repulsion between charges of membrane and free anions in solution. The rejection for cations was irregular. As duration elongating, rejection declined to about 4.0 % for all salts. The rejection was contributed to ion exchange effect in a short period time. As leaving of protons of SO3H, it gradually decreased to a fixed value. In summary, SPA6 possesses the capacity of amphiphilic self-assembly and hence building up a membrane of uniformed distribution of SO3H groups. SO3H groups enveloped charged channels for molecules passing through, which provided enough repulsion to negative charged particles, such as anionic dyes. There were several important factors to fulfill high dye removal under a comparably low operating pressure, including: i) charged groups combined tightly with polymer during forming process; ii) compact skin layer of the film; iii) large aggregates of dye molecules. 4. Conclusion SPA6 showed ability to self-assembly into 255 nm micelles with negative charges at the surface in solution. Micelles switched into nanofibers protruding from the centre to form pompoms after immersed in coagulation bath. They piled up and solidified to form the membrane. Charged groups were remained at the surface of pompoms during concretion. The gaps between spheres provided channels full of 16

anions, which supplied great repulsion to anions. It endowed membranes the potential to remove anion dyes from aqueous solution. When the solid content of casting solution was 30 %, air staying time for 120s, the membrane of 200 μm thickness (30M) achieved 99.73% rejection ratio for 500 mg/L CR, 85.68% for 100 mg/L AB93 and 63.06% for 100 mg/L MO in filtration. Total removal percentage were 95.48% for AB93 and 79.51% for MO after second filtration. Self-aggregation of dye molecules also helped their high removal efficiency. The membrane had high rejection of 95.31% for 100 mg/L CR after filtration 1500 min. The FRR was 91.84% after three cycles of fouling and DI water washing, suggesting that 30M had good anti-fouling capacity. SPA6 membrane displayed a potential for effective treatment of textile wastewater. Acknowledgements The authors acknowledge the financial support from the National Nature Science Foundation of China (Grant No. 51373119) and Science and Technology Plants of Tianjin (No. 16YFZCASTING SOLUTIONF00330, 15PTSYJC00250). Reference [1] F.C. Moreira, R.A.R. Boaventura, E. Brillas, V.J.P. Vilar, Electrochemical advanced oxidation processes: A review on their application to synthetic and real wastewaters, Appl. Catal. B-Environ. 202 (2017) 217-261. [2] J. Lin, W. Ye, H. Zeng, H. Yang, J. Shen, S. Darvishmanesh, P. Luis, A. Sotto, B. Vander Bruggen, Fractionation of direct dyes and salts in aqueous solution using loose nanofiltration membranes, J. Membr. Sci. 477 (2015) 183-193. [3] A.E. Ghaly, R. Ananthashankar, M. Alhattab, V.V. Ramakrishnan, Production, Characterization and Treatment of Textile Effluents: A Critical Review, J. Chem. Eng. Process Technol. 5 (2014) 182. [4] M. Peydayesh, T. Mohammadi, O. Bakhtiari, Effective treatment of dye wastewater via positively charged TETA-MWCNT/PES hybrid nanofiltration membranes, Sep. Purif. Technol. 194 (2018) 488-502. [5] V. Nair, R.Vinu, Peroxide-assisted microwave activation of pyrolysis char for adsorption of dyes from wastewater, Bioresource Technol. 216 (2016) 511-519. [6] Y.Y. Lau, Y.S. Wong, T.T. Teng, N. Morad, M. Rafatullah, S.A. Ong, Degradation of cationic and anionic dyes in coagulation–flocculation process using bi-functionalized silica hybrid with aluminum-ferric as auxiliary agent, RSC Adv.5 (2015) 34206-34215. [7] D. Daâssi, S. Rodríguez-Couto, M. Nasri, T. Mechichi, Biodegradation of textile dyes by immobilized laccase from Coriolopsis gallicainto Ca-alginate beads, Int. Biodeter. Biodegr. 90 (2014) 71-78. [8] H. Song, J.A. You, C. Chen, H. Zhang, X.Z. Ji, C. Li, Y. Yang, N. Xu, J. Huang, Manganese functionalized mesoporous molecular sieves Ti-HMS as a Fenton-like catalyst for dyes wastewater purification by advanced oxidation processes, J. Environ. Chem. Eng. 4 (2016) 4653-4660. [9] X. Wei, S. Wang, Y. Shi, H. Xiang, J. Chen, Application of Positively Charged Composite Hollow-Fiber Nanofiltration Membranes for Dye Purification, Ind. Eng. Chem. Res. 53 (2014) 14036–14045. [10] J. Lin, C.Y. Tang, W. Ye, S.P. Sun, S.H. Hamdan, A. Volodin, C.V. Haesendonck, A. Sotto, P. Luis, B. Vander Bruggen, Unraveling flux behavior of superhydrophilic loose nanofiltration membranes during textile wastewater treatment, J. Membr. Sci. 493 (2015) 690-702. [11] C.Z. Liang, S.P. Sun, B.W. Zhao, T.S. Chung, Integration of Nanofiltration Hollow Fiber Membranes with 17

Coagulation–Flocculation to Treat Colored Wastewater from a Dyestuff Manufacturer: A Pilot-Scale Study, Ind. Eng. Chem. Res. 54 (2015) 11159-11166. [12] S. Zhao, P. Song, Z. Wang Z, H. Zhu, The PEGylation of plant polyphenols/polypeptide-mediated loose nanofiltration membrane for textile wastewater treatment and antibacterial application, J. Taiwan Inst. Chem. E. 82 (2018) 42-55. [13] Y.C. Xu, Z.X. Wang, X.Q. Cheng, Y.C. Xiao, L. Shao, Positively charged nanofiltration membranes via economically mussel-substance-simulated co-deposition for textile wastewater treatment, Chem. Eng. J. 303 (2016) 555-564. [14] M. Liu, Q. Chen, K. Lu, W. Huang, Z. Lv, C. Zhou, S. Yu, C. Gao, High efficient removal of dyes from aqueous solution through nanofiltration using diethanolamine-modified polyamide thin-film composite membrane, Sep. Purif. Technol. 173 (2017) 135-143. [15] J.B. Jin, D.Q. Liu, D.D. Zhang, Y.H. Yin, X.Y. Zhao, Y.F. Zhang, Preparation of thin-film composite nanofiltration membranes with improved antifouling property and flux using 2,2′-oxybis-ethylamine, Desalination355 (2015) 141-146. [16] Y. Tang, B. Tang, P. Wu, Preparation of a Positively Charged Nanofiltration Membrane Based on Hydrophilic–hydrophobic Transformation of Poly(ionic liquid), J. Mater. Chem. A 3 (2015) 12367-12376. [17] X. Zhao, W. Chen, Y. Su, W. Zhu, J. Peng, Z. Jiang, L. Kong, Y. Li, J. Liu, Hierarchically engineered membrane surfaces with superior antifouling and self-cleaning properties, J. Membr. Sci. 441 (2013) 93–101. [18] F. Chen, X. Shi, X. Chen, W. Chen, An iron (II) phthalocyanine/poly(vinylidene fluoride) composite membrane with antifouling property and catalytic self-cleaning function for high-efficiency oil/water separation, J. Membr. Sci. 552 (2018) 295-304. [19] Y. Lv, C. Zhang, A. He, S.J. Yang, G.P. Wu, S.B. Darling, Z.K. Xu, Photocatalytic Nanofiltration Membranes with Self-Cleaning Property for Wastewater Treatment, Adv. Funct. Mater. 27 (2017) 1700251. [20] J.H. Jhaveri, Z.V.P. Murthy, A comprehensive review on anti-fouling nanocomposite membranes for pressure driven membrane separation processes. Desalination 379 (2016) 137-154. [21] A.W. Mohammad, Y.H. Teow, W.L. Ang, Y.T. Chung, D.L. Oatley-Radcliffe, N. Hilal, Nanofiltration membranes review: Recent advances and future prospects, Desalination 356 (2015) 226-254. [22] G. Zeng, Y. He, Y. Zhan, L. Zhang, Y. Pan, C. Zhang, Z. Yu, Novel polyvinylidene fluoride nanofiltration membrane blended with functionalized halloysite nanotubes for dye and heavy metal ions removal, J. Hazard. Mater. 317 (2016) 60-72. [23] L. Yu, Y. Zhang, H. Zhang, J. Liu. Development of a molecular separation membrane for efficient separation of low-molecular-weight organics and salts, Desalination 359 (2015) 176-185. [24] L. Xing , N. Guo, Y. Zhang, H. Zhang, J. Liu, A negatively charged loose nanofiltration membrane by blending with poly (sodium 4-styrene sulfonate) grafted SiO2 via SI-ATRP for dye purification, Sep. Purif. Technol. 146 (2015) 50-59. [25] S. Zhao, Z. Wang, A loose nano-filtration membrane prepared by coating HPAN UF membrane with modified PEI for dye reuse and desalination, J. Membr. Sci. 524 (2017) 214-224. [26] G. Han, Y. Feng, T.S. Chung, M. Weber, C. Maletzko, Phase Inversion Directly Induced Tight Ultrafiltration (UF) Hollow Fiber Membranes for Effective Removal of Textile Dyes, Environ. Sci. Technol. 51 (2017) 14254-14261. [27] G. Han, T.S. Chung, M. Weber, C. Maletzko, Low-Pressure Nanofiltration Hollow Fiber Membranes for Effective Fractionation of Dyes and Inorganic Salts in Textile Wastewater, Environ. Sci. Technol. 52 (2018) 3676-3684. [28] L.K. Xiong, C.Y.Yin, Preparation and characterization of cationic dyeable polycaprolactam modified with 18

5-sulfoisophthalic acid sodium and poly(ethylene glycol), Text. Res. J. 87 (2016) 433-443. [29] R. A. Lofquist, Cationic dyeable polyamide of improved physical properties. United States Patent No. 3828009 (1974). [30] D.D. Zhang, D.Q. Liu, Q.J. Ren, Y.B. Chen, C.Y. Yin, Fabrication of polyamide-6 fiber of high SO3H content surface through electrospinning, Polymer 98 (2016) 11-19. [31] S.S. Madaeni, S. Zinadini, V. Vatanpour. Preparation of superhydrophobic nanofiltration membrane by embedding multiwalled carbon nanotube and polydimethylsiloxane in pores of microfiltration membrane, Sep. Purif. Technol. 111 (2013) 98-107. [32] S. Zinadini, A.A. Zinatizadeh, M. Rahimi, V. Vatanpour, H. Zangeneh, M. Beygzadeh, Novel high flux antifouling nanofiltration membranes for dye removal containing carboxymethyl chitosan coated Fe3O4 nanoparticles, Desalination 349 (2014) 145-154.

Highlights:



Membrane rejection for 500 mg/L Congo red solution was 99.71% at 0.1 MPa.



The rejection for 100 mg/L Congo red solution declined to 95.31% after 1500 min.



Electrostatic repulsion and pore size screen determined the removal of anionic dyes.

19