Journal Pre-proof

Loose nanofiltration-based electrodialysis for highly efficient textile wastewater

treatment

Wenyuan Ye, Riri Liu, Xiangyu Chen, Qin Chen, Jiuyang Lin, Xiaocheng Lin, Bart Van
der Bruggen, Shuaifei Zhao
PII: S0376-7388(20)30760-2
DOI: https://doi.org/10.1016/j.memsci.2020.118182
Reference: MEMSCI 118182

To appear in: Journal of Membrane Science

Received Date: 13 February 2020

Revised Date: 18 March 2020
Accepted Date: 19 April 2020

Please cite this article as: W. Ye, R. Liu, X. Chen, Q. Chen, J. Lin, X. Lin, B. Van der Bruggen, S. Zhao,
Loose nanofiltration-based electrodialysis for highly efficient textile wastewater treatment, Journal of
Membrane Science (2020), doi: https://doi.org/10.1016/j.memsci.2020.118182.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. 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.

© 2020 Published by Elsevier B.V.

CRediT author statement

Wenyuan Ye: Conceptualization, Methodology, Writing - Original draft, Funding

acquisition.
Riri Liu: Methodology, Investigation.
Xiangyu Chen: Methodology, Investigation.
Chen Qin: Methodology, Investigation.
Jiuyang Lin: Conceptualization, Methodology, Resources, Writing - Review &
Editing, Supervision, Funding acquisition.
Xiaocheng Lin: Conceptualization, Methodology, Resources, Supervision, Funding
acquisition.
Bart Van der Bruggen: Supervision, Writing - review & editing.
Shuaifei Zhao: Supervision, Writing - review & editing.
Table of Content
 1 Loose nanofiltration-based electrodialysis for highly efficient

2 textile wastewater treatment

3 Wenyuan Yea, Riri Liub, Xiangyu Chenb, Qin Chenb, Jiuyang Linb,*, Xiaocheng Linc,*,

4 Bart Van der Bruggend, Shuaifei Zhaoe

5 *Corresponding authors: E-mail: [email protected] (J. Lin); [email protected] (X.

6 Lin).

7
a
Fujian Provincial Key Laboratory of Soil Environmental Health and Regulation,

8 College of Resources and Environment, Fujian Agriculture and Forestry University,

9 Fuzhou 350002, China

10
b
Fujian Provincial Engineering Research Center of Rural Waste Recycling

11 Technology, School of Environment and Resources, Fuzhou University, Fuzhou

12 350116, China

13
c
College of Chemical Engineering, Fuzhou University, 350116, Fujian, China

14
d
Department of Chemical Engineering, Process Engineering for Sustainable Systems

15 (ProcESS), KU Leuven, Celestijnenlaan 200F, B-3001 Leuven, Belgium

16
e
Deakin University, Geelong, Institute for Frontier Materials, VIC 3216, Australia

17

18

19

20

1
21 Abstract: Effective fractionation of dyes and salts (NaCl) potentially paves a new

22 path to sustainable treatment of textile wastewater. Herein, a loose nanofiltration (NF)

23 membrane was innovatively used as the anion conductive membrane in electrodialysis

24 (ED) for dye and salt separation. The NF membrane with a molecular weight cut-off

25 of 678 Da showed a sufficient retention for reactive dyes with small molecular

26 weights (<627 Da) predominantly based on synergistic effect of steric repulsion and

27 Donnan repulsion. Simultaneously, the NF membrane had a considerably fast

28 transport for anions (i.e., Cl-), as its loose surface structure created extra

29 nano-channels for ion transfer. The anion conductive loose NF membrane based ED

30 process yielded excellent separation performance in a one-step dye/NaCl fractionation.

31 Specifically, a 98.9% desalination efficiency was obtained by the loose NF-based ED

32 process along with 99.4% dye recovery. Additionally, an extremely low fouling

33 propensity was observed during a 5-cycle ED operation, as the negative surface

34 charge of the loose NF membrane served as an intrinsic barrier to prevent the dye

35 species from entering the membrane pores through electrostatic repulsion. This study

36 demonstrated a desirable potential of loose NF membranes as anion conductive

37 membranes in ED for resource recovery from highly-saline wastewater.

38

39 Keywords: Loose nanofiltration; Anion conductive membrane; Electrodialysis;

40 Dye/NaCl fractionation; Textile wastewater

41

42

2
43 1. Introduction

44 The textile industry has been considered one of the most polluted industrial sectors,

45 which potentially exerts a detrimental impact on the eco-system and public health [1,

46 2]. A large quantity of textile wastewater is generated in either dye synthesis or dyeing

47 processes. Generally, textile wastewater is highly loaded with inorganic salts (i.e., 6.0

48 wt.% NaCl), which is derived from the neutralization reaction for dye synthesis or

49 used as additives to enhance the uptake of dyes during dyeing [3]. In the advanced

50 conceptual framework of sustainability, textile wastewater can be regarded as a

51 potential resource [4-7]. Particularly, the dye species in textile wastewater can be

52 feasibly extracted and reused for fabric dyeing to reduce the dyeing expenditure;

53 meanwhile, salts (mainly NaCl) with elevated concentration can be a renewable

54 resource for industrial applications, e.g., in pressure retarded osmosis [8], forward

55 osmosis [9-11], reverse electrodialysis [12] or bipolar membrane electrodialysis [13].

56 Unfortunately, the traditionally available methods, involving biological degradation

57 [14], adsorption [15], flocculation/coagulation [16] and advanced oxidation [17],

58 exclusively focus on dye degradation or removal, failing to efficiently extract the salts

59 or dyes as useful resources.

60 Loose nanofiltration (NF) membranes with molecular weight cut-offs (MWCOs) of

61 500-1000 Da have attracted increasing interests in fractionation of dyes and salts

62 [18-23]. In particular, the loose surface structure of these NF membranes renders a

63 high transmission of NaCl, but sufficiently retains dye species, showing a promising

64 feasibility in fractionation of dyes and salts. For instance, a loose NF membrane with

3
65 an MWCO of 860 Da can completely reject direct dyes (i.e., >99.6%) with 96.7%

66 permeation for NaCl [24]. Ye et al. tailored the surface property of a loose NF

67 membrane through bio-inspired co-deposition of g-C3N4 nanosheets, which enabled

68 high rejections to reactive dyes (>99.3%) but a nearly free transmission for NaCl (i.e.,

69 97.1%) [25]. Zhu et al. designed a loose nanofiltration membrane functionalized with

70 copper nanoparticles via mussel-inspired deposition, allowing for a >99.0% rejection

71 to reactive dyes with outstanding NaCl permeation (i.e., 98.0%) [26]. However, the

72 cake-enhanced concentration polarization occurs inevitably for loose NF membranes,

73 which significantly reduces the permeation flux and deteriorates the fractionation

74 efficacy of dyes and salts during the diafiltration with considerable water

75 consumption [27], in view of enriching and purifying the dyes from textile

76 wastewater.

77 Apart from loose NF, electrodialysis (ED) also poses a potential in organics/salts

78 fractionation [28-30]. Especially, cations and anions in the highly-saline dye/salt

79 mixtures can be continuously transported through the ion exchange membranes under

80 an established current field, which can potentially allow for desalination of

81 highly-saline mixtures. Xue et al. investigated the feasibility of ED for desalination of

82 an acid blue 9/salt mixture [31]. Initially, ca. 85% of salt species (NaCl) was removed

83 from this dye/salt binary mixture by ED. However, a compact dye cake layer evolved

84 and deposited onto the surface of the anion exchange membrane for pore blocking,

85 hindering further transfer of NaCl and consequently deteriorating the desalination

86 efficacy. Berkessa et al. employed a bipolar membrane for electrodialysis in salt

4
 87 valorization from textile wastewater for acid/base generation [32]. Only a moderate

88 amount (i.e., ~70%) of inorganic salts in the reactive dye/salt mixtures was

89 desalinated and converted into acid and base, mainly due to the formation of a dye

90 cake layer on the anion exchange membrane. Such a cake layer promoted the dye

91 molecules to enter the pore structure and bind with the positively charged groups of

92 anion exchange membranes through electrostatic attraction, which elevated the

93 electrical resistance, raised temperature of membrane stack and tremendously

94 jeopardized the electrodialytic performance of ion exchange membranes [32, 33]. This

95 strongly urges for advanced strategies to effectively fractionate the dye and salt

96 species in sustainable treatment of textile wastewater.

97 Taking advantage of both ED and loose NF, a novel stack configuration “loose

98 NF-based ED” using loose NF membranes instead of anion exchange membranes was

99 designed to separate dyes from salts in this study. Especially, the NF membranes

100 allow for ion migration under an established current field [34-36], showing a

101 strategical promise in ion fractionation. Considering the difference in molecular sizes

102 between dyes and salts, it calls for an advanced selective membrane for efficient

103 dye/salt fractionation via ED. Herein, a loose NF membrane with an MWCO of 678

104 Da was applied as the anion conductive membrane in ED for fractionation of dyes and

105 NaCl. Initially, the filtration behavior of the loose NF membrane with NaCl and dye

106 solutions was investigated at an applied pressure of 2 bar to elucidate its selectivity

107 towards reactive dyes (MW<627 Da) and NaCl. Afterwards, the practicability of the

108 loose NF membrane as the anion conductive membrane in ED was demonstrated for

5
109 fractionation of dyes and NaCl under different operation conditions. Ultimately, the

110 fouling propensity of the loose NF membrane in the NF-based ED process was further

111 explored by comparing with a commercial anion exchange membrane to highlight its

112 superiority in fractionation of dye/NaCl mixtures.

113 2. Experimental

114 2.1 Materials

115 Three reactive dyes, namely reactive orange 1 (RO1, MW: 615.3 Da), reactive

116 orange 16 (RO16, MW: 617.5 Da) and reactive blue 19 (RB19, MW: 626.6 Da), were

117 used as the typical model dyes. Supplementary Fig. S1 displays the properties (i.e.,

118 chemical structure and size) of these tested reactive dyes. NaCl with 99.0% purity was

119 ordered from Sigma-Aldrich. Ultrapure water with electrical resistance of 18.2

120 MΩ·cm from Milli-Q system was supplied in all the experiments. All the chemicals

121 were used as received.

122 The NF membrane coupons with surface chemistry of poly(piperazine)-amide were

123 supplied by Guangdong Yinachuan Environmental Technology Co. (China) and used

124 as anion conductive membrane in the NF-based ED process. The cation (CJMC-1)

125 and anion (CJMA-1) exchange membranes were ordered from ChemJoy Polymer

126 Material Co., Ltd (China) and were used in a conventional ED process as comparison

127 with NF-based ED process. The main properties of these ion exchange membranes are

128 shown in Supplementary Table S1.

129

130 2.2 Membrane characterization

6
131 2.2.1 Surface pore size

132 The pore size of the tested NF membrane was measured by filtrating the aqueous

133 solutions of 100 mg·L-1 polyethylene glycols with various molecular weights (i.e.,

134 200, 300, 400, 600, and 1000 Da) at 4 bar. The concentration of the polyethylene

135 glycols was determined based on the total organic carbon (TOC) concentration

136 through TOC-LCpH/CPN analyzer (Shimadzu, Japan). The rejection coefficient (R)

137 of polyethylene glycols by the NF membrane was calculated via the concentration of

138 polyethylene glycols in the permeate (Cp) and feed (Cf) by Eq. 1:

 C 
139 R (% ) = 1 − p  × 100 (1)
 Cf 

140 Specifically, the correlation between the Stokes diameter (dp) and molecular weight

141 (MW) of polyethylene glycols was expressed by Eq. 2:

142 d p = 33.46 × 10-12 × MW 0.557 (2)

143 Based on the retention of polyethylene glycols, the pore size distribution of the

144 membrane was theoretically determined with a log-normal probability density

145 function between Stokes diameters of polyethylene glycols and their rejection (R) [37,

146 38]:

dR ( d p )  ( ln d − ln µ )2 
147 =
1
exp  −
p p
 (3)
d p ln σ p 2π  ( p ) 
2
dd p
 2 ln σ

148 where μp presents the mean effective pore size of the membranes, as determined at R

149 of 50%. σp denotes geometric standard deviation of pore size, defined as the ratio

150 between dp at R of 84.13% and μp. It is noted that the MWCO of the membrane is

151 determined by dp at R of 90%, which is back-calculated using Eq. 2.

7
152 2.2.2 Surface charge

153 Surface charge of the NF membrane is critical for screening the charged solutes,

154 significantly influencing the membrane performance. Generally, the surface charge of

155 the membrane was assessed in terms of zeta potential through the SurPASS

156 electrokinetic analyzer (Anton Paar GmbH, Austria) in a 10.0 mmol·L-1 NaCl aqueous

157 solution. The solution pH for surface charge measurement ranged from 2.9 to 9.3,

158 which was adjusted through titration with HCl or NaOH. Finally, the pH-dependent

159 zeta potential of the loose NF membrane was recorded and analyzed.

160 2.3 Nanofiltration performance

161 The filtration performance of the tested NF membrane was conducted with a

162 lab-made filtration system. Firstly, the NF membrane was initially compacted at 5 bar

163 through filtration of pure water to achieve a steady permeation flux, prior to the

164 filtration experiment. Afterwards, the separation of pure NaCl solutions with variable

165 salinities from 0.05 to 15.0 g·L-1 was carried out at 2 bar and 25 °C. In addition, the

166 filtration of pure reactive dye solutions (RB19, RO16, and RO1) with variable

167 concentrations (0.1, 0.5, and 1.0 g·L-1) was conducted. Finally, the separation of a

168 dye/NaCl (i.e., RO1/NaCl) binary mixture with different salinities (i.e., up to 15.0

169 g·L-1 NaCl) was further performed to evaluate the fractionation of reactive dyes and

170 NaCl by the NF membrane.

171 The rejection (R) of reactive dyes or NaCl for the tested NF membrane was

172 determined by using Eq. 4:

8
  Cp 
173 R ( %) = 1 −  ×100 (4)
 Cf 

174 where Cp and Cf are the concentrations of the dyes or NaCl in the permeate and feed,

175 respectively. Particularly, the concentration of the dyes was measured by a UV-1780

176 double beam spectrophotometer (Shimadzu); the concentration of Cl- ions was

177 determined by a Dionex ICS-2000 ion chromatography system.

178 2.4 NF-based electrodialysis for dye/NaCl fractionation

179 In this case, the commercial NF membrane was used instead of commercial anion

180 exchange membrane (CMJA-1) in ED process to investigate its electrodialytic

181 performance for dye and NaCl fractionation. A lab-made ED stack was designed for

182 this experiment, comprised of titanium-based electrodes (i.e., an anode and a cathode)

183 coated with ruthenium and a membrane stack (Fig. 1). Specifically, three pieces of

184 cation exchange membrane (CJMC-1) and two pieces of NF membrane with effective

185 area of 1.94×10-3 m2 were alternatively inserted into the membrane stack. The NaCl

186 solution which had equivalent salt concentration with the feed was applied in the

187 concentrate compartment. A 3.0% Na2SO4 aqueous solution was applied as the

188 electrolyte media for both electrode compartments.

189
9
190 Fig. 1 Configuration of NF-based ED process for dye and salt fractionation (CM:

191 cation exchange membrane; LNF: loose NF membrane)

192

193 Initially, a 300 mL pure NaCl solution with variable concentrations (i.e., 6.0, 9.0,

194 12.0 and 15.0 g·L-1) was used as feed to test the transport behavior of ions through the

195 NF membrane as a role of anion conductive membrane at a constant current of 0.5 A.

196 Subsequently, the transport of the ions through the NF membrane was studied at

197 different currents (i.e., 0.3, 0.4, 0.5, and 0.6 A) to investigate the effect of current on

198 ion transfer behavior.

199 Additionally, a 300 mL RO1/NaCl (i.e., 1.0 g·L-1 RO1 and 12.0 g·L-1 NaCl)

200 mixture was further employed as the feed in the NF-based ED process for separation

201 of RO1 dye and NaCl. The loose NF-based ED process was operated at 0.5 A and

202 25 °C. Once the conductivity of the feed in the diluate chamber was reduced to 1.0

203 mS·cm-1, the loose NF-based ED process was terminated as 1 cycle. Simultaneously,

204 the fouling propensity of the NF membrane as anion conductive membrane caused by

205 dye species was studied based on 5-cycle ED operation at the same condition. As a

206 comparison, the conventional ED process equipped with commercial anion exchange

207 membranes (CMJA-1) was performed to demonstrate the superiority of the NF

208 membrane as anion conductive membrane in fractionation of dye/NaCl mixtures. The

209 voltage and current across the ED stack were recorded by a regulated power supply

210 with a power voltage range of 60.4 V.

10
211 The desalination efficiency (η) of the loose NF membrane during ED process was

212 calculated by Eq. 5:

 C1 
213 η ( % ) = 1 −  ×100 (5)
 C0 

214 where C0 and C1 are the concentrations of NaCl in the feed before and after ED

215 process.

216 3. Results and discussion

217 3.1 Pore size and surface charge

218 Generally, the larger pore size of the NF membranes would diminish the

219 electrostatic repulsion to the ionic species to some extent, which facilitates the

220 transmission of inorganic salts (e.g., NaCl) through the membranes for fractionation

221 of organics and salts [39]. NF membranes with loose surface structures (i.e., 500-1000

222 Da MWCO) are of increasing interest for addressing this challenge. As implied in Fig.

223 2, the NF membrane in this study had an MWCO of 678 Da, demonstrating a loose

224 surface structure for favorable permeation of salts. Additionally, the surface of this

225 loose NF membrane yielded a mean effective pore size (µ p) of 0.48 nm, which can

226 potentially retain small organic molecules.

227

11
228 Fig. 2 Relationship between rejections to polyethylene glycols and the Stokes

229 diameters of the loose NF membrane and theoretical pore size distribution (inset)

230

231 The pH-dependent surface charge of the NF membrane is illustrated in Fig. 3. It

232 can be clearly observed that the isoelectric point of the tested NF membrane was at

233 pH 3.0. This allows for a highly negative charge density on the loose NF membrane

234 surface at higher pH ranges, mainly due to the dissociation of carboxylic groups

235 present on the poly(piperazine)amide-based surface with low cross-linking degree [21,

236 40]. Specifically, the zeta potential of the NF membrane was -38 mV at neutral pH,

237 indicating an enhanced electrostatic repulsion to the charged organic solutes, such as

238 dye species.

239

240 Fig. 3 pH-dependent zeta potential of the NF membrane tested.

241 3.2 Nanofiltration performance

242 To effectively fractionate the dyes and NaCl in ED process, it calls for steric

243 retention of dyes and permeation of NaCl by the membrane. To study the potential of

244 the loose NF membrane in separation of dye/NaCl mixtures, filtration of dye/salt

245 solutions was initially conducted (Fig. 4). As demonstrated in Fig. 4A, the loose NF

12
246 membrane exhibited a very low rejection to NaCl over the NaCl concentration range.

247 Generally, the increasing salt concentration would weaken the electrostatic repulsion

248 force of the NF membrane surface to the salt solutes through suppressing the Debye

249 screening length for reduced salt retention and vice versa [41, 42]. However, only a

250 moderate rejection of 25.8±1.3% for NaCl was observed when filtrating a 0.05 g·L-1

251 NaCl solution. This is mainly attributed to the considerably large pore size of the

252 loose NF membrane, diminishing the electrostatic repulsion effect to some extent and

253 resulting in fast permeation of NaCl even at low salt concentrations. As the NaCl

254 concentration was increased to 15.0 g·L-1, the loose NF membrane showed a nearly

255 complete salt transport (i.e., 2.7±0.4% rejection). This suggests that the unique

256 surface structure of the loose NF membrane can create sufficient nano-channels for

257 salt transport. Additionally, the loose NF membrane endowed a complete rejection

258 (i.e., >99.2%) to all the reactive dyes with low molecular weights (<627 Da), based on

259 the synergistic effects of electrostatic repulsion and steric exclusion (Fig. 4B).

260

13
261

262 Fig. 4 Separation performance of the loose NF membrane with various solutions: (A)

263 pure NaCl solutions with different salinity; (B) pure reactive dye (RB19, RO16 and

264 RO1) solutions; (C) RO1/NaCl binary mixture.

265

266 In dye synthesis as well as in dyeing processes, the salts (e.g., NaCl) with high

267 content would be generated as byproduct, which potentially has a negative impact on

268 filtration performance of the membrane. Therefore, the separation behavior of the

269 loose NF membrane in the RO1/NaCl binary mixtures with different salinities (up to

270 15.0 g·L-1) was investigated (Fig. 4C). Although the elevated content of NaCl in the

271 RO1/NaCl mixture would intensify the shielding effect of the membrane surface for

272 reduced dye rejection [41, 43], an insignificant decline (i.e., ~0.2%) in rejection of

273 RO1 dye was observed over the NaCl concentration range. This suggests that steric

274 exclusion mainly dominates the separation performance of the loose NF membrane

275 for RO1 dye over the Donnan repulsion. On the other hand, the presence of RO1 had

276 no obvious impact on the permeation of NaCl by the loose NF membrane. Especially,

277 an expectably high transmission of NaCl in the RO1/NaCl binary mixtures was

278 yielded, consistently comparable to that in the pure NaCl solutions. Such a separation

14
279 performance demonstrates the superior selectivity of the loose NF membrane towards

280 dye species and NaCl.

281 3.2 NF-based electrodialytic performance

282 3.2.1 Electrodialysis of pure NaCl solutions

283 The loose surface morphology of the tested NF membrane could create extra

284 nano-channels for NaCl permeation but effectively retain the reactive dyes through

285 size exclusion. Therefore, it is of great significance to investigate the applicability of

286 this loose NF membrane as a role of anion conductive membrane for salt transport in

287 the ED process. Figs. 5 and 6 show the electrodialytic performance of the loose NF

288 membrane for salt transport in ED at different conditions (i.e., salt concentrations and

289 applied currents).

290

291

15
292 Fig. 5 Electrodialytic performance of the loose NF membrane and the anion exchange

293 membrane for desalination of NaCl solutions with different salinities at a current of

294 0.5 A. (A) Evolution of conductivity in the concentrate and diluate sides during

295 NF-based ED; (B) changes of voltage, current and stack resistance in NF-based ED;

296 (C) performance comparison between NF-based ED and conventional ED; (D)

297 transport mechanism of Cl- ions in NF-based ED and conventional ED.

298

299
300 Fig. 6 Electrodialytic performance of loose NF membrane for desalination of NaCl

301 solution (12 g·L-1) at different current densities. (A) Evolution in conductivity; (B)

302 voltage, current and electrical resistance of ED stack.

303

304 Fig. 5A shows the electrodialytic performance of the loose NF membrane for

305 desalination of NaCl solutions with different salinities at a current of 0.5 A. The

306 conductivity of the brine (i.e., NaCl) solutions in the diluate compartment decayed

307 linearly with the operation time. Simultaneously, the conductivity of the solutions in

308 the concentrate compartment increased steadily, ascribed to the transfer of Na+ and Cl-

309 ions through the cation exchange membrane and the loose NF membrane, indicating

16
310 the anionic conductivities of the loose NF membrane for effective transport of Cl- ions.

311 As the NF-based ED process proceeded, more Na+ and Cl- ions in the diluate

312 compartment were transported through the cation exchange membrane and loose NF

313 membrane, respectively. Therefore, the conductivity of the brine solutions in the

314 diluate chamber was significantly reduced to ca. 0.40 mS·cm-1, demonstrating an

315 outstanding desalination efficacy of the loose NF membrane as anion conductive

316 membrane at the established current field. For instance, the final content of NaCl in

317 the diluate chamber was reduced to 117.8 mg·L-1 when using a 12.0 g·L-1 NaCl

318 solution as feed, yielding a ca. 99.0% desalination efficiency for the NF-based ED

319 process.

320 Additionally, the voltage across the NF-based ED stack increased gradually at the

321 constant current mode, mainly ascribed to the increasing electrical resistance (Fig.

322 5B). As the Na+ and Cl- ions in the feed were further exhausted at the ultimate stage,

323 the NF-based ED stack showed a remarkable boost in the overall electrical resistance,

324 leading to an increase in voltage across the membrane stack. Once the voltage across

325 the membrane stack was beyond the range of the power supply (i.e., 60.4 V), the

326 NF-based ED stack would be operated at the constant voltage mode. In this case, the

327 current across the NF-based ED stack started to decrease with the continuously

328 increasing overall electrical resistance. On the other hand, increasing initial

329 concentration of NaCl in the feed solutions had no obvious effect on the transport of

330 NaCl through the loose NF membrane. Specifically, the NF-based ED process

331 exhibited a nearly identical decay tendency in conductivity of the brine solutions in

17
332 the diluate compartment over the tested salt concentration range, indicating the same

333 salt transfer rate at the fixed current field as driving force.

334 As a comparison, the conventional ED process equipped with commercial anion

335 exchange membranes (CMJA-1) was operated (Fig. 5C). The CMJA-1 membrane

336 yielded a faster transport of Cl- ions than the loose NF membrane, as reflected from

337 the steeper decay in conductivity of the brine solution in the diluate compartment.

338 Generally, the CMJA-1 anion exchange membrane has the functional groups carrying

339 with positive charges and offers a lower area resistance (i.e., 2.52 Ω·cm2) for

340 enhanced ion transfer at a fixed direct current field (Supplementary Fig. S2).

341 However, the loose NF membrane with larger pore size displayed a negatively

342 charged surface, establishing an intrinsic barrier for transport of anions (e.g., Cl- ions)

343 via electrostatic repulsion to some extent, as demonstrated by the higher area

344 resistance (i.e., 10.47 Ω·cm2) in Supplementary Fig. S2.

345 Fig. 6 depicts the electrodialytic performance of the loose NF membrane in a 12.0

346 g·L-1 NaCl solution at different current densities (i.e., from 0.3 to 0.6 A). As shown in

347 Fig. 6A, the feed (i.e., NaCl) solution in the diluate chamber had a conductivity that is

348 decreasing with operation time. Increasing the current density across the ED stack can

349 facilitate the transport of Na+ and Cl- ions through the membranes, as demonstrated by

350 a steeper decay in the conductivity of the feed solution. Specifically, a reversely

351 proportional relationship between the current density of the ED process and the

352 operation time was found. This indicates that a shorter operation time would be

353 required at increasing current density without sacrificing the desalination efficiency

18
354 (i.e., 99.0%).

355 3.2.2 NF-based electrodialysis for dye/NaCl mixtures

356 In order to sufficiently fractionate the dyes and NaCl, the NF-based ED process

357 was studied for separation of the dye/salt mixtures at a current of 0.5 A, aiming at

358 sustainable treatment of textile wastewater (Fig. 7).

359

360

361

19
362 Fig. 7 (A) NF-based ED for fractionation of RO1/NaCl mixture; (B) conventional ED

363 for fractionation of RO1/NaCl mixture; (C) digital pictures of the loose NF membrane

364 and conventional the anion exchange membrane (CMJA-1) before and after fouling

365 caused by dye during ED process; (D) fouling mechanism of the loose NF membrane

366 and CMJA-1 caused by RO1 dye in NF-based ED and conventional ED at an

367 established current field.

368

369 As illustrated in Fig. 7A, the NF-based ED stack yielded a nearly linear decay in

370 the conductivity of the diluate stream at an established current field, indicating the

371 effective transport of Na+ and Cl- ions in RO1/NaCl mixture through the cation

372 exchange membrane and the loose NF membrane. Compared to the case of a pure

373 NaCl solution with 12.0 g·L-1, the ion transport was limited to some extent at the

374 same operation conditions when using the RO1/NaCl mixture as feed, calling for a

375 longer operation duration for desalination. This is plausibly ascribed to the formation

376 of a dye cake layer on the loose NF membrane surface to impede the fast ion transport.

377 However, the dye cake layer had no obvious impact on the efficacy of ion transport.

378 Remarkably, the content of NaCl in the RO1/NaCl mixture was reduced from 12.0 to

379 ~0.13 g·L-1 with 98.9% desalination efficiency during NF-based ED process

380 (Supplementary Fig. S3). Furthermore, the recovery of dye reached 99.4% in this

381 case, due to the nano-channels of the loose NF membrane surface, which would

382 effectively retain the dye species through size exclusion effect (Fig. 4).

20
383 In order to explore the anti-fouling property of the loose NF membrane induced by

384 dye species at the established direct current field, the NF-based ED process was

385 performed in 5 cycles. The NF-based ED stack was observed to show a nearly

386 identical evolution in the conductivity of the concentrate and diluate in all the 5 cycles,

387 demonstrating the extremely low fouling propensity of loose NF membrane for stable

388 ED operation, which can be reflected by Fig. 7C and Supplementary Fig. S2 (i.e.,

389 slight increase in area resistance of the loose NF membrane after fouling). This is

390 mainly due to the negative surface charge of the loose NF membrane which can

391 effectively repel the dye species from entering the pore structure for pore blocking

392 through electrostatic repulsion, highlighting the superiority of the loose NF membrane

393 as anion conductive membrane in fractionation of dye and salt through ED.

394 In contrast, the conventional ED stack equipped with an anion exchange membrane

395 (i.e., CMJA-1) suffered from severe fouling caused by adsorption and deposition of

396 dye molecules into the pore structure of the CMJA-1 anion exchange membrane

397 through electrostatic attraction during a 5-cycle operation (Fig. 7B, 7C and 7D),

398 which remarkably boosted the area resistance of the CMJA-1 anion exchange

399 membrane by a factor of 40 (Supplementary Fig. S2) and thus resulted in an

400 increased electrical resistance (i.e., up to 3020 Ω) of the ED stack. This would

401 significantly constraint the ion transport of the membrane stack with reduced

402 desalination efficiency (Fig. 7B and Supplementary Fig. S3). Specifically, the

403 content of Cl- ions remained above the level of ~0.27 g·L-1 in the feed for

404 conventional ED. On the other hand, the higher electrical resistance of conventional

21
405 ED stack would produce more heats and a risk of damage in the chemical structure of

406 the CMJA-1 anion exchange membrane (Fig. 7C), thus deteriorating the performance

407 and minimizing the lifespan of the commercial CMJA-1 anion exchange membrane.

408 4. Conclusion

409 In this study, a novel loose NF-based ED process was proposed to fractionate the

410 dye/salt mixture by using a loose NF membrane as anion conductive membrane. Due

411 to the intrinsic loose surface structure, the NF membrane with an MWCO of 678 Da

412 showed a high permeation for NaCl with a sufficient retention (>99.2%) for reactive

413 dyes (MW<627 Da), enabling an outstanding selectivity between dyes and NaCl.

414 Moreover, the NF membrane yielded a 99.0% desalination efficiency as anion

415 conductive membrane during ED application in NaCl solutions, since the loose

416 surface structure created sufficient nano-channels to intensify ion transport. Based on

417 this, a one-step effective fractionation of dye and NaCl from the dye/NaCl mixture

418 can be successfully obtained by the loose NF-based ED process. Specifically, a 98.9%

419 desalination efficiency was achieved along with a 99.4% dye recovery. Compared

420 with the commercial anion exchange membrane (CMJA-1), the loose NF membrane

421 used as anion conductive membrane in NF-based ED process offered a superior

422 fractionation of dye/NaCl binary mixtures with reduced fouling propensity. This is

423 due to the fact that the intrinsic negative surface charge of the loose NF membrane

424 acted as an effective barrier to repel the dye species from entering the pore structure

425 for pore blocking via electrostatic repulsion. This study provides a conceptual

426 guideline to enable electro-nanofiltration membrane in practical ED application for

22
427 effective fractionation of organics and salt species from highly-saline wastewater, in

428 view of environmental sustainability.

429 Acknowledgment

430 J. Lin, W. Ye and X. Lin would like to thank the funding support from the National

431 Natural Science Foundation of China (Grant Nos: 21706035, 21707018, and

432 21808038), the Natural Science Foundation of Fujian Province (Grant No:

433 2019Y0006), the Fujian Agriculture and Forestry University Program for

434 Distinguished Young Scholar (Grant No.: xjq201704), the Award Program for

435 Minjiang Scholar Professorship and the Valuable Instrument & Equipment Open Test

436 Fund of Fuzhou University (Grant No.: 2019T024) for this work.

437 References

438 [1] Y.C. Xu, Z.X. Wang, X.Q. Cheng, Y.C. Xiao, L. Shao. Positively charged

439 nanofiltration membranes via economically mussel-substance-simulated

440 co-deposition for textile wastewater treatment. Chem. Eng. J. 303 (2016)

441 555-564.

442 [2] K. Paździor, L. Bilińska, S. Ledakowicz. A review of the existing and emerging

443 technologies in the combination of AOPs and biological processes in industrial

444 textile wastewater treatment. Chem. Eng. J. 376 (2019) 120597.

445 [3] W. Ye, K. Ye, F. Lin, H. Liu, M. Jiang, J. Wang, R. Liu, J. Lin. Enhanced

446 fractionation of dye/salt mixtures by tight ultrafiltration membranes via fast

447 bio-inspired co-deposition for sustainable textile wastewater management. Chem.

448 Eng. J. 379 (2020) 122321.

23
449 [4] J.S. Guest, S.J. Skerlos, J.L. Barnard, M.B. Beck, G.T. Daigger, H. Hilger, S.J.

450 Jackson, K. Karvazy, L. Kelly, L. Macpherson, J.R. Mihelcic, A. Pramanik, L.

451 Raskin, M.C.M. Van Loosdrecht, D. Yeh, N.G. Love. A new planning and design

452 paradigm to achieve sustainable resource recovery from wastewater. Environ. Sci.

453 Technol. 43 (2009) 6126-6130.

454 [5] M. Kamali, D.P. Suhas, M.E. Costa, I. Capela, T.M. Aminabhavi. Sustainability

455 considerations in membrane-based technologies for industrial effluents treatment.

456 Chem. Eng. J. 368 (2019) 474-494.

457 [6] J. Lin, F. Lin, X. Chen, W. Ye, X. Li, H. Zeng, B. Van der Bruggen. Sustainable

458 management of textile wastewater: A hybrid tight

459 ultrafiltration/bipolar-membrane electrodialysis process for resource recovery and

460 zero liquid discharge. Ind. Eng. Chem. Res. 58 (2019) 11003-11012.

461 [7] Y. Xu, J. Lin, C. Gao, B. Van der Bruggen, Q. Shen, H. Shao, J. Shen. Preparation

462 of high-flux nanoporous solvent resistant polyacrylonitrile membrane with

463 potential fractionation of dyes and Na2SO4. Ind. Eng. Chem. Res. 56 (2017)

464 11967-11976.

465 [8] B.E. Logan, M. Elimelech. Membrane-based processes for sustainable power

466 generation using water. Nature 488 (2012) 313-319.

467 [9] T.Y. Cath, A.E. Childress, M. Elimelech. Forward osmosis: Principles,

468 applications, and recent developments. J. Membr. Sci. 281 (2006) 70-87.

469 [10] T. Chung, S. Zhang, K.Y. Wang, J. Su, M.M. Ling. Forward osmosis processes:

470 Yesterday, today and tomorrow. Desalination. 287 (2012) 78-81.

24
471 [11] S. Zhao, L. Zou, C.Y. Tang, D. Mulcahy. Recent developments in forward

472 osmosis: Opportunities and challenges. J. Membr. Sci. 396 (2012) 1-21.

473 [12] J.W. Post, H.V.M. Hamelers, C.J.N. Buisman. Energy recovery from controlled

474 mixing salt and fresh water with a reverse electrodialysis system. Environ. Sci.

475 Technol. 42 (2008) 5785-5790.

476 [13] J. Lin, W. Ye, J. Huang, B. Ricard, M. Baltaru, B. Greydanus, S. Balta, J. Shen,

477 M. Vlad, A. Sotto, P. Luis, B. Van der Bruggen. Toward resource recovery from

478 textile wastewater: Dye Extraction, water and base/acid regeneration using a

479 hybrid NF-BMED process. ACS Sustain. Chem. Eng. 3 (2015) 1993-2001.

480 [14] A. Yurtsever, B. Calimlioglu, E. Sahinkaya. Impact of SRT on the efficiency and

481 microbial community of sequential anaerobic and aerobic membrane bioreactors

482 for the treatment of textile industry wastewater. Chem. Eng. J. 314 (2017)

483 378-387.

484 [15] A. Hethnawi, N.N. Nassar, A.D. Manasrah, G. Vitale.

485 Polyethylenimine-functionalized pyroxene nanoparticles embedded on Diatomite

486 for adsorptive removal of dye from textile wastewater in a fixed-bed column.

487 Chem. Eng. J. 320 (2017) 389-404.

488 [16] T.A. Nguyen, R. Juang. Treatment of waters and wastewaters containing sulfur

489 dyes: A review. Chem. Eng. J. 219 (2013) 109-117.

490 [17] L. Bilińska, M. Gmurek, S. Ledakowicz. Comparison between industrial and

491 simulated textile wastewater treatment by AOPs - Biodegradability, toxicity and

492 cost assessment. Chem. Eng. J. 306 (2016) 550-559.

25
493 [18] W. Ye, J. Lin, R. Borrego, D. Chen, A. Sotto, P. Luis, M. Liu, S. Zhao, C.Y.

494 Tang, B. Van der Bruggen. Advanced desalination of dye/NaCl mixtures by a

495 loose nanofiltration membrane for digital ink-jet printing. Sep. Purif. Technol.

496 197 (2018) 27-35.

497 [19] J. Zhu, J. Wang, A.A. Uliana, M. Tian, Y. Zhang, Y. Zhang, A. Volodin, K.

498 Simoens, S. Yuan, J. Li, J. Lin, K. Bernaerts, B. Van der Bruggen.

499 Mussel-inspired architecture of high-flux loose nanofiltration membrane

500 functionalized with antibacterial reduced graphene oxide - copper

501 nanocomposites. ACS Appl. Mater. Interfaces 9 (2017) 28990-29001.

502 [20] J. Wang, L. Qin, J. Lin, J. Zhu, Y. Zhang, J. Liu, B. Van der Bruggen. Enzymatic

503 construction of antibacterial ultrathin membranes for dyes removal. Chem. Eng. J.

504 323 (2017) 56-63.

505 [21] J. Lin, C.Y. Tang, C. Huang, Y.P. Tang, W. Ye, J. Li, J. Shen, R. Van den

506 Broeck, J. Van Impe, A. Volodin, C. Van Haesendonck, A. Sotto, P. Luis, B. Van

507 der Bruggen. A comprehensive physico-chemical characterization of

508 superhydrophilic loose nanofiltration membranes. J. Membr. Sci. 501 (2016)

509 1-14.

510 [22] Y. Xu, M. Wu, S. Yu, Y. Zhao, C. Gao, J. Shen. Ultrathin and stable graphene

511 oxide film via intercalation polymerization of polydopamine for preparation of

512 digital inkjet printing dye. J. Membr. Sci. 586 (2019) 15-22.

513 [23] W. Ye, H. Liu, M. Jiang, J. Lin, K. Ye, S. Fang, Y. Xu, S. Zhao, B. Van der

514 Bruggen, Z. He. Sustainable management of landfill leachate concentrate through

26
515 recovering humic substance as liquid fertilizer by loose nanofiltration. Water Res.

516 157 (2019) 555-563.

517 [24] J. Lin, W. Ye, H. Zeng, H. Yang, J. Shen, S. Darvishmanesh, P. Luis, A. Sotto, B.

518 Van der Bruggen. Fractionation of direct dyes and salts in aqueous solution using

519 loose nanofiltration membranes. J. Membr. Sci. 477 (2015) 183-193.

520 [25] W. Ye, H. Liu, F. Lin, J. Lin, S. Zhao, S. Yang, J. Hou, S. Zhou, B. Van der

521 Bruggen. High-flux nanofiltration membranes tailored by bio-inspired

522 co-deposition of hydrophilic g-C3N4 nanosheets for enhanced selectivity towards

523 organics and salts. Environ. Sci.: Nano 6 (2019) 2958-2967.

524 [26] J. Zhu, A. Uliana, J. Wang, S. Yuan, J. Li, M. Tian, K. Simoens, A. Volodin, J.

525 Lin, K. Bernaerts, Y. Zhang, B. Van der Bruggen. Elevated salt transport of

526 antimicrobial loose nanofiltration membranes enabled by copper nanoparticles

527 via fast bioinspired deposition. J. Mater. Chem. A. 4 (2016) 13211-13222.

528 [27] J. Lin, C.Y. Tang, W. Ye, S. Sun, S.H. Hamdan, A. Volodin, C.V. Haesendonck,

529 A. Sotto, P. Luis, B. Van der Bruggen. Unraveling flux behavior of

530 superhydrophilic loose nanofiltration membranes during textile wastewater

531 treatment. J. Membr. Sci. 493 (2015) 690-702.

532 [28] W. Pronk, M. Biebow, M. Boller. Electrodialysis for recovering salts from a

533 urine solution containing micropollutants. Environ. Sci. Technol. 40 (2006)

534 2414-2420.

535 [29] Y. Zhang, L. Pinoy, B. Meesschaert, B. Van der Bruggen. Separation of small

536 organic ions from salts by ion-exchange membrane in electrodialysis. AIChE J.

27
537 57 (2011) 2070-2078.

538 [30] T. Xu, C. Huang. Electrodialysis-based separation technologies: A critical review.

539 AIChE J. 54 (2008) 3147-3159.

540 [31] C. Xue, Q. Chen, Y. Liu, Y. Yang, D. Xu, L. Xue, W. Zhang. Acid blue 9

541 desalting using electrodialysis. J. Membr. Sci. 493 (2015) 28-36.

542 [32] Y.W. Berkessa, Q. Lang, B. Yan, S. Kuang, D. Mao, L. Shu, Y. Zhang. Anion

543 exchange membrane organic fouling and mitigation in salt valorization process

544 from high salinity textile wastewater by bipolar membrane electrodialysis.

545 Desalination 465 (2019) 94-103.

546 [33] V. Lindstrand, G. Sundström, A. Jönsson. Fouling of electrodialysis membranes

547 by organic substances. Desalination 128 (2000) 91-102.

548 [34] L. Ge, B. Wu, Q. Li, Y. Wang, D. Yu, L. Wu, J. Pan, J. Miao, T. Xu.

549 Electrodialysis with nanofiltration membrane (EDNF) for high-efficiency cations

550 fractionation. J. Membr. Sci. 498 (2016) 192-200.

551 [35] L. Bazinet, M. Moalic. Coupling of porous filtration and ion-exchange

552 membranes in an electrodialysis stack and impact on cation selectivity: A novel

553 approach for sea water demineralization and the production of physiological

554 water. Desalination 277 (2011) 356-363.

555 [36] F. Sheng, L. Hou, X. Wang, M. Irfan, M.A. Shehzad, B. Wu, X. Ren, L. Ge, T.

556 Xu. Electro-nanofiltration membranes with positively charged polyamide layer

557 for cations separation. J. Membr. Sci. 594 (2020) 117453.

558 [37] W. Ye, R. Liu, F. Lin, K. Ye, J. Lin, S. Zhao, B. Van der Bruggen. Elevated

28
559 nanofiltration performance via mussel-inspired co-deposition for sustainable

560 resource extraction from landfill leachate concentrate. Chem. Eng. J. 388 (2020)

561 124200.

562 [38] M. Jiang, K. Ye, J. Lin, X. Zhang, W. Ye, S. Zhao, B. Van der Bruggen.

563 Effective dye purification using tight ceramic ultrafiltration membrane. J. Membr.

564 Sci. 566 (2018) 151-160.

565 [39] M. Jiang, K. Ye, J. Deng, J. Lin, W. Ye, S. Zhao, B. Van der Bruggen.

566 Conventional ultrafiltration as effective strategy for dye/salt fractionation in

567 textile wastewater treatment. Environ. Sci. Technol. 52 (2018) 10698-10708.

568 [40] C.Y. Tang, Y. Kwon, J.O. Leckie. Effect of membrane chemistry and coating

569 layer on physiochemical properties of thin film composite polyamide RO and NF

570 membranes: I. FTIR and XPS characterization of polyamide and coating layer

571 chemistry. Desalination 242 (2009) 149-167.

572 [41] X. Wang, C. Zhang, P. Ouyang. The possibility of separating saccharides from a

573 NaCl solution by using nanofiltration in diafiltration mode. J. Membr. Sci. 204

574 (2002) 271-281.

575 [42] Y. Zeng, L. Wang, L. Zhang, J.Q. Yu. An acid resistant nanofiltration membrane

576 prepared from a precursor of poly(s-triazine-amine) by interfacial polymerization.

577 J. Membr. Sci. 546 (2018) 225-233.

578 [43] J. Qin, X. Dai, Y. Zhou, L. Zhang, H. Chen, C. Gao. Desalting and recovering

579 naphthalenesulfonic acid from wastewater with concentrated bivalent salt by

580 nanofiltration process. J. Membr. Sci. 468 (2014) 242-249.

29
Highlights:

1. Loose nanofiltration (NF)-based electrodialysis (ED) was used for dye/salt

separation

2. Loose NF membrane shows an excellent ion transport as anion conductive

membrane

3. Loose NF-based ED yields 98.9% desalination efficiency and 99.4% dye recovery

4. Low fouling propensity is observed for loose NF membrane as anion conductive

membrane
Declaration of interests

■ 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.

☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests: