Experiments and Modeling of Boric Acid Permeation through Double

Subscriber access provided by UNIV OF CAMBRIDGE

Article

Experiments and modeling of boric acid permeation through double-skinned forward osmosis membranes Lin Luo, Zhengzhong Zhou, Tai-Shung Chung, Martin Weber, Claudia Staudt, and Christian Maletzko Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b06166 • Publication Date (Web): 09 Jun 2016 Downloaded from http://pubs.acs.org on June 12, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 33

Environmental Science & Technology

1 2 3

Experiments and modeling of boric acid permeation through double-skinned forward

4

osmosis membranes

5 6

Lin Luo1,2 , Zhengzhong Zhou3, Tai-Shung Chung1,2,4*,

7

Martin Weber5, Claudia Staudt5, Christian Maletzko6

8 9 10 11 12 13 14 15 16 17 18 19

1

2

NUS Graduate School for Integrative Science and Engineering, National University of Singapore, Singapore 117456

Department of Chemical and Biomolecular Engineering, National University of Singapore, Singapore 117585 3

4

Water Desalination & Reuse (WDR) Center, King Abdullah University of Science and Technology, Thuwal 23955–6900, Saudi Arabia

20 21 22 23 24 25 26 27

School of Chemistry & Chemical Engineering, Jiangsu University, 301 Xuefu Road, Zhenjiang, Jiangsu Province, P. R. China 212013

5

Advanced Materials & Systems Research, BASF SE, GM-B001, 67056 Ludwigshafen, Germany 6

Performance Materials, BASF SE, G-PM/PU-F206, 67056 Ludwigshafen, Germany

*Corresponding author

28 29

Tel: +65-65166645; fax: +65-67791936; Email: [email protected]

30

ACS Paragon Plus Environment


31

Abstract

32

Boron removal is one of the great challenges in modern wastewater treatment, owing to the

33

unique small size and fast diffusion rate of neutral boric acid molecules. As forward osmosis

34

(FO) membranes with a single selective layer are insufficient to reject boron, double-skinned

35

FO membranes with boron rejection up to 83.9% were specially designed for boron

36

permeation studies. The superior boron rejection properties of double-skinned FO membranes

37

were demonstrated by theoretical calculations, and verified by experiments. The double-

38

skinned FO membrane was fabricated using a sulfonated polyphenylenesulfone (sPPSU)

39

polymer as the hydrophilic substrate and polyamide as the selective layer material via

40

interfacial polymerization on top and bottom surfaces. A strong agreement between

41

experimental data and modeling results validates the membrane design and confirms the

42

success of model prediction. The effects of key parameters on boron rejection, such as boron

43

permeability of both selective layers and structure parameter, were also investigated in-depth

44

with the mathematical modelling. This study may provide insights not only for boron removal

45

from wastewater, but also open up the design of next generation FO membranes to eliminate

46

low-rejection molecules in wider applications.

47 48

Keywords

49

Forward osmosis; double-skinned FO membranes; boron transport model; boron removal

50

efficiency


Page 2 of 33

Page 3 of 33

51


TOC art:

52 53



54

1. Introduction

55

Recently, boron has come to the forefront as a possible drinking water contaminant. The

56

global average concentration of boron in seawater is approximately 4.6 mg/L, while peaks of

57

about 100 mg/L boron have been reported on produced water streams.1 To meet the final user

58

specifications, boron-containing wastewater needs to be regulated, and the boron

59

concentration in drinking water is suggested to be 2.4 mg/L by the World Health

60

Organization. However, boron in the form of boric acid (B(OH)3 or H3BO3, a weak Lewis

61

acid of pKa=9.24,2 cannot be easily removed by conventional treatments.3 Alternatively,

62

membrane technology is one potential solution; however, its removal efficiency still remains

63

a challenge.4 At neutral pH, boric acid exists as an undissociated molecule with zero charge.

64

It has a low molecular weight of only 61.8 g/mol, and a tendency to hydrogen bond with the

65

membrane matrix.1 Consequently, commercial brackish water reverse osmosis (BWRO)

66

membranes were reported to be ineffective to remove neutral boron with rejection ratios less

67

than 65%, while seawater reverse osmosis (SWRO) operations can achieve a fair rejection

68

rate of 90%.1, 3, 5 In the case of treating high boron-containing wastewater or obtaining a high

69

water recovery, even the permeate of SWRO may not be able to meet the strict requirements

70

of boron content. Thus, some RO processes are modified by multi-step approaches at

71

evaluated pH or with the aid of boron selective resins. 4, 6

72 73

As an emerging membrane technology, forward osmosis (FO) has been investigated for boric

74

acid permeation at neutral conditions.7-10 FO is driven by the osmotic pressure difference

75

between the draw and feed solutions across the semi-permeable membrane.11-14 Under no or

76

low pressure operation, FO membranes have advantages of low fouling tendency and high

77

rejections to a wide range of contaminants.15-20 In terms of boron removal, both cellulose-

78

based and thin film composite (TFC) FO membranes have been studied.7-10 The boron


Page 4 of 33

Page 5 of 33


79

passage through FO membranes is significantly affected by membrane orientation, water flux,

80

boron permeability of the selective layer, structure parameter, as well as the pH value and ion

81

strength of feed solutions. However, most laboratory FO membranes can only achieve a

82

boron rejection similar to BWRO (≤ 65%) at neutral pH. Considering the insufficient boron

83

rejection of a single selective layer, the focus in this study lies on specially designed novel

84

double-skinned FO membranes to hinder the boron passage and improve the separation

85

performance.

86 87

The original concept of double selective FO membranes was invented by means of phase

88

inversion of cellulose acetate (CA) solutions cast on glass plates.21 It has been extended to

89

form double selective FO hollow fibers22 and double selective FO flat-sheet membranes

90

consisting of two TFC layers23, two different selective layer materials24 or layer-by-layer

91

polyelectrolytes25, 26. These double-skinned FO membranes have shown mitigated internal

92

concentration polarization (ICP)23, 25 and better fouling resistance24, 27 than single-selective

93

FO membranes. Since the former also possesses superior boron transport resistance to the

94

latter, this study aims to take the advantages of the dual-skin configuration to mitigate boron

95

transport for a higher boron rejection. Therefore, double-skinned TFC FO membranes were

96

fabricated and their boron flux and rejection performance were investigated. A transport

97

model describing boron passage was also derived in order to simulate the experimental

98

results and to understand the science among separation performance, transport phenomena

99

and crucial membrane parameters. This work may provide useful insights for better FO

100

membrane design to remove trace contaminants in water treatment.

101 102

2. Experimental section



103

2.1 Materials and membrane formation

104

Sulfonated polyphenylenesulfone polymer (sPPSU) with a sulfonation degree of 1.5 mol %

105

(IEC value 5.0 meq/100g polymer) was kindly provided by BASF and used as the support

106

material for the fabrication of FO membranes. It was synthesized using the directly

107

copolymerized sulfonation method developed by McGrath28. The sPPSU polymer has been

108

utilized for ultrafiltration29, nanofiltration30 and forward osmosis31.

109 110

The flat sheet membrane support was cast on a glass plate using the phase inversion method.

111

Details of polymer dope formula and casting conditions are described in supporting

112

information (SI). To fabricate the double-skinned TFC membrane, a polyamide TFC layer

113

was first synthesized on the top surface of the support by interfacial polymerization between

114

MPD and TMC monomers, then on the bottom surface following the same protocol as

115

described in the SI. Between the syntheses of two polyamide layers, the membranes were

116

kept hydrated by soaking in DI water. For comparison, the single-skinned FO membrane was

117

fabricated with a TFC layer only formed on the top surface of the support.

118 119

2.2 FO membrane performance

120

FO performance of the newly fabricated TFC FO membranes was evaluated through a lab-

121

scale cross-flow setup. Details can be found elsewhere31, 32. The flow rates at both the draw

122

and feed sides were kept at 0.2 L min-1, and all the experiments were conducted at room

123

temperature (23±0.5 ◦C). NaCl solutions of different concentrations and a boron solution of

124

100 ppm were prepared as draw and feed solutions, respectively. The boron feed solution was

125

prepared by dissolving boric acid into DI water, without adjusting the pH. At this pH (~6.6),

126

boric acid is poorly hydrated and exists in the form of undissociated molecules. The average


Page 6 of 33

Page 7 of 33


127

testing duration for each FO operation is around 1 hour. The single-skinned FO membrane

128

was tested under two modes: (1) the active layer facing draw solution (AL-DS) mode where

129

the selective layer faced the draw solution and (2) the active layer facing feed solution (AL-

130

FS) mode where the TFC selective layer faced the boron feed. While the double-skinned FO

131

membrane was tested with the top polyamide layer facing the boron feed solution, referred to

132

as the double-skinned (DS) mode.

133 134

The water permeation flux (Jw, L m−2 h−1 or LMH) was calculated as follows:

‫ܬ‬௪ = ஺

∆௏

೘ ∆௧

135

(1)

136

where ∆V is the volume of permeation water collected in the draw side over a predetermined

137

time ∆t, and Am is the effective membrane area. The back diffusion of the draw solute to the

138

feed side; namely, the salt reverse flux (JS, gm-2h-1 or gMH), can be calculated by measuring

139

the conductivity in the feed solution at the beginning and the end of each experiment:

‫ܬ‬௦ =

140 141

∆(஼೟ ௏೟ ) ஺೘ ∆௧

(2)

where Ct and Vt are the salt concentration and feed volume at the end of tests, respectively.

142 143

Boron solute flux (JB, mg s-1 m-2) and boron rejection (R) are calculated by the following

144

equations:

145

146

௃ಳ ൗ௃ ೢ

ܴ = ቆ1 − ஼

ಷ೐೐೏

‫ܬ‬஻ =

஼ವೝೌೢ ௏ವೝೌೢ ஺೘ ∆௧

ቇ × 100% = ቀ1 −

(3) ஼ವೝೌೢ ௏ವೝೌೢ ஼ಷ೐೐೏ ∆௏

ቁ × 100%

(4)

147

where CDraw and CFeed are the boron concentrations of the draw and feed solutions at the end

148

of tests, respectively, VDraw is the volume of the draw solution. In order to estimate the boron

149

rejection, the effective boron concentration diffusing across the membrane could be



Page 8 of 33

150

expressed as the ratio of JB/Jw7, which could be experimentally calculated as the change of the

151

boron amount in the draw solution side (CDrawVDraw) divided by the permeate volume ∆V.

152

The azomethine-H colorimetric method was utilized to detect the boron concentration in the

153

feed or draw solution. This detection method has been reported to be the most sensitive

154

spectrometric technique with a detection limit of 0.02 ppm and least suffering from

155

interferences.33, 34 More details could be found in the SI.

156 157

2.3 Mass transport characteristics of selective layers

158

The pure water permeability coefficient, A, NaCl permeability coefficient, Bs, and boron

159

permeability coefficient, BB, of the FO membranes were evaluated in a dead-end RO cell

160

under a transmembrane pressure of 3 bar. A was determined from the pure water permeation

161

flux, while the salt rejection (Rs) and boron rejection (RB) were evaluated respectively using

162

1000 ppm NaCl and 100 ppm boron solutions as feeds. Then A, Bs and BB were determined

163

with the aid of the following equations:

164

165

166

‫=ܣ‬஺

ܴ = ൬1 − ଵିோ ோ

=

ொ

೘ ∆௉

஼೛ ஼೑

൰ × 100% ஻

஺(∆௉ି∆గ)

(5) (6) (7)

167

where Q is the water permeation volumetric flow rate, Cp and Cf are the solute concentrations

168

in the permeate and feed solutions, individually. ∆P and ∆π are the hydraulic pressure and

169

osmotic pressure differences across the membrane, respectively. NaCl and boron

170

concentrations were detected respectively via conductivity and azomethine-H colorimetric

171

methods. For clarity, the TFC layer fabricated on the top surface of the substrate was

172

considered as the active layer1 with a boron permeability coefficient of B1, while the bottom


Page 9 of 33


173

TFC layer was referred to as the active layer2, which has a boron permeability coefficient of

174

B2.

175 176

3. Modeling

177

3.1 Internal concentration polarization (ICP)

178

In FO processes using a single-skinned FO membrane (Figure S1), salt concentration

179

polarization can happen internally inside the support layer and externally on the selective

180

layer.35 Usually, the ICP is a severer problem than the external concentration polarization as

181

the latter may be mitigated by increasing the flow rate. Equations had been developed to

182

model the water flux under ICP in single-skinned FO processes36-38:

183

In AL-DS mode:

184

(8)

185

In AL-FS mode:

186

(9)

187

where πD,b and πF,b are the osmotic pressures of the bulk draw and feed solutions, while πD,m

188

and πF,m represent the osmotic pressures on the surfaces of the selective layer facing the draw

189

and feed solutions, respectively. A and Bs refer to water permeability and salt permeability

190

coefficients, which can be determined by methods described in section 2.3. Since the osmotic

191

pressure of the bulk feed is very low, πF,m is assumed to be 0 in the AL-FS mode. Then the

192

solute diffusion coefficient within the porous layer Km,s can be calculated from the

193

experimental FO performance according to eq. 9. Thus, the membrane structural parameter S

194

is obtained by dividing the salt diffusivity Ds by Km,s:

195

‫ܬ‬௪ = ‫ܭ‬௠,௦ ‫݊ܫ‬

஺గವ,೘ ି௃ೢ ା஻ೞ ஺గಷ,್ ା஻ೞ

‫ܬ‬௪ = ‫ܭ‬௠,௦ ‫ ݊ܫ‬஺గ

஺గವ,್ ା஻ೞ

ܵ=௄ೄ ஽

೘,ೞ


ಷ,೘ ା௃ೢ ା஻ೞ

(10)


Page 10 of 33

196

Since the same membrane substrates and interfacial polymerization methods were utilized to

197

fabricate single- and double-skinned membranes in this study, it was reasonable to assume

198

that they have the same structure parameter.

199 200

3.2 Boron transport through double-skinned FO membranes

201

An analytical model for boron transport through double-skinned FO membranes is derived by

202

combining the solution-diffusion model for the selective layers39 and the diffusion-convection

203

transport in the support layer40, 41. A schematic of boron transport in a double-skinned FO

204

membrane is shown in Figure 1. During FO operations, active layer1 faces the boron feed

205

solution, while active layer2 faces the draw solution. Once boron permeates through the

206

active layer1, it would be carried by the water flux from the interface to the support layer.

207

Due to the retention of boron by the active layer2, boron concentration builds up at the

208

interface between the support and active layer2. This phenomenon can be considered as the

209

internal concentration polarization (ICP) of boron. Within the support, boron transport is

210

dominated by the convection induced by the water flow but also affected by the diffusion due

211

to the concentration gradient.40 At any distance x away from the interface between the

212

support and active layer1, the boron flux JB,dx can be expressed as:

213

‫ܬ‬஻, ௗ௫ = −‫ ܦ‬ௗ௫ + ‫ܬ‬ௐ ‫ܥ‬

214

where C represents the boron concentration at the distance x, JW is the water flux and D is the

215

effective boron diffusion coefficient within the porous layer. It is defined as D=DB·ε, where

216

DB is the boron diffusion coefficient and ε is the porosity. At steady state, the boron solute

217

flux JB can be expressed as:

218

ௗ஼

‫ܬ‬஻ = ‫ܤ‬ଵ (‫ܥ‬ி − ‫ ܥ‬ᇱ ) = ‫ܤ‬ଶ ൫‫ " ܥ‬− ‫ܥ‬஽ ൯

(11)

(12)

219

where B1 and B2 are the boron permeability of active layer1 and 2, while C’ and C” represent

220

the boron concentrations at the interfaces facing active layer1 and layer2, respectively. CF is


Page 11 of 33


221

the boron concentration in the feed, while CD is the boron concentration at active layer2

222

facing the draw solution7:

‫ܥ‬஽ =

223

௃ಳ

௃ೈ

(13)

224

In eq. 11, the boundary conditions of ICP are taken into consideration:, C(0)=C’ at x=0; and

225

C(τt)=C” at x=τt (τ is the tortuosity and t is the thickness of support layer)42. By solving these

226

equations, boron flux JB and boron rejection R could be expressed as:

227

228

‫ܬ‬஻ =

ಳభ ಻ೈ ಳమ౛౮౦൫಻ೈ /಼೘ ൯

ܴ = 1 − ஼ವ = 1 − ௃ ஼

ಷ

஻భ

ಳ ଵା భ ା

௃ಳ

ೈ ஼ಷ

= 1−

‫ܥ‬ி

(14) ஻భ

಻ೈ ಳభ ಳమ ౛౮౦൫಻ೈ /಼೘ ൯

௃ೈ ା஻భ ା

(15)

229

Here, Km is the mass transfer coefficient of boron within the support, which is equal to the

230

ratio of DB to S (the structure parameter):

231

‫ܭ‬௠ = ‫=ݏ‬

232

஽ಳ ௌ

(16)

௧∙ఛ ఌ

(17)

233

The calculation of Km is based on eq. 10 and 16. As the structural characteristics of the

234

membrane support layer t, τ and ε are invariant within experimental conditions and S is the

235

intrinsic property of the membrane support, the same S value is employed in the ICP

236

calculation of both boron and NaCl transports:

237

‫=ݏ‬

௧∙ఛ ఌ

= ௄ ೄ = ௄ಳ ஽

೘,ೞ

஽

೘

(18)

238

For comparison, the passage of boron through the single-skinned TFC membrane is also

239

simulated and summarized in the SI, the same equations have also been reported7. Figure S2

240

illustrates the boron transports across the single and double TFC membranes under different

241

modes. Table 1 lists the modeling equations to calculate the normalized boron flux and boron

242

rejection under these three modes. Experimentally, B1 and B2 were characterized according to

243

section 2.3 and Km was calculated from eq. 18.



244 245

4. Results and discussion

246

4.1 As-fabricated FO membranes

247

The morphologies of the sPPSU membrane support and the double-skinned FO membrane

248

are shown in Figure 2. The membrane substrate exhibits a denser top surface with small pores

249

but a relatively porous bottom surface with a thickness of around 50µm. Due to the delayed

250

demixing induced by the sulphonated material during phase inversion, the as-fabricated

251

membrane substrate displays a fully sponge-like structure29, 30. The membrane substrate has a

252

pure water permeability of 450.7 ± 21.9 LMH bar−1 with a mean pore size of 15.6 nm. Its

253

molecular weight cut-off is 94.5 kDa measured by solute rejection experiments29. In the case

254

of the double-skinned FO membrane, its top and bottom TFC layers exhibit similar ridge-

255

and-valley morphologies. However, they are different in thickness. The two inserted FESEM

256

images in Figure 2 enlarge their cross section morphologies. The polyamide thin film on the

257

bottom has a thickness of ~750 nm, which is significantly thicker than the top one (~350 nm).

258 259

The difference in layer thickness is probably caused by different interfacial polymerization

260

kinetics in both layers. Since the MPD aqueous solution is first introduced and absorbed by

261

the membrane support, the MPD monomer would diffuse out of the porous support into the

262

organic phase and react with TMC when contacting with the TMC solution.43-45 Thus, the

263

polyamide layer grows in protuberances and ridges instead of flat films. Once the thickness

264

of the polyamide film reaches a certain limit (ca. 20 nm), the MPD diffusion rate drops

265

quickly and the polymerization reaction is terminated46. Since the pore size and surface

266

porosity of the membrane substrate also affect MPD diffusion rate, the termination process

267

would be delayed on the bottom surface, where the incomplete coverage of the polyamide

268

films on large pores allows MPD to continuously diffuse out for reaction. As a result, layers


Page 12 of 33

Page 13 of 33


269

of protuberances and ridges form on the bottom surface, leading to a thick TFC layer. For

270

comparison, membrane morphologies of the single-skinned FO membrane are presented in

271

Figure S3.

272 273

The water permeability and salt rejection of TFC layers in the double-skinned membrane are

274

characterized separately through RO tests. As illustrated in Table 2, a pure water permeability

275

(PWP) of 1.64 LMH/bar is characterized for the TFC layer formed on the top, while the

276

bottom one exhibits a lower PWP of 1.19 LMH/bar. The different PWPs of two selective

277

layers could be explained by their different polyamide thicknesses (Figure 2). Both TFC

278

layers have good NaCl rejection rates of ~ 95% but relatively poor boron rejections of lower

279

than 45%. Also, the boron permeability coefficients of the top and bottom selective layers are

280

calculated based on eq. 7, and represented by B1 and B2, individually. Both layers exhibit

281

much higher boron permeability coefficients than NaCl, confirming that boron can pass

282

through the membrane more easily and freely. Unlike the hydrated sodium and chlorine ions

283

of NaCl in water, boric acid exists as an undissociated molecule with no charge at the pH of

284

6.6. Therefore, the primary separation mechanism for boron removal by RO tests is size

285

exclusion. Due to the small size of poorly hydrated boron molecules, size exclusion becomes

286

less efficient, resulting in the poor boron rejection and the high boron permeability coefficient.

287 288

4.2 Experimental FO performance

289

Figure 3 demonstrates the FO performance of as-fabricated FO membranes by adjusting the

290

draw solution concentrations under different operation modes. Importantly, AL-DS and AL-

291

FS modes are two operation modes of the single-skinned FO membrane that comprises a TFC

292

layer on its top surface, while the DS mode refers to the double-skinned FO membrane with



293

its top surface facing the feed solution. In our study, a 100 ppm boron solution is employed as

294

the feed solution, while the concentration of NaCl draw solutions varies from 0.5 M to 4 M.

295

All water fluxes and reverse salt fluxes rise up as a function of draw solution concentration in

296

three modes. Obviously, the AL-DS mode exhibits the highest water flux and reverse salt flux

297

under the same NaCl concentration, while the DS mode the lowest. Since the double-skinned

298

FO membrane has an additional selective layer, the water and salt fluxes are reduced by the

299

greater transport resistance and the ICP within the membrane support.

300

301

By using eq. 9 with aid of FO data, the solute diffusion coefficient within the porous layer

302

Km,s is calculated to be 5.51 µm/s, from which the structure parameter S is estimated to be

303

268.4 µm according to eq. 10 and a known DS from literatures7. Then the mass transfer

304

coefficient of boron within the support Km is determined to be 5.44 µm/s based on eq. 16 and

305

18. Both single- and double-skinned FO membranes are considered to share the same S and

306

Km parameters. The experimentally measured boron flux and boron rejection will be

307

illustrated in the following section and compared with the modeling results.

308

309

4.3 Theoretical and experimental boron flux

310

As shown in Table 1, the feed boron concentration CF directly affects the boron flux but has

311

little or negligible influence to boron rejection since the boron concentration here is highly

312

diluted.7,9,10 Figures 4 and 5 display the normalized boron flux (JB/CF) and boron rejection as

313

a function of water flux (JW) in FO operations using various draw solution concentrations,

314

respectively. The normalized boron fluxes in all operating modes increase with water flux to

315

different extent. The increment of JB/CF with JW in the AL-DS mode is the most significant,

316

followed by AL-FS and DS modes. This order could be explained by the different boron


Page 14 of 33

Page 15 of 33


317

transport behaviors in different operation modes. Under the AL-DS mode (Figure S2 A),

318

boric acid enters the porous support layer from the feed, driven by water convective flow and

319

diffusion. However, due to the retention by the active layer, boron concentration at the

320

interface between the support layer and active layer becomes higher than that in the feed,

321

which is reported as the concentrative ICP of boron 7. This elevates the boron concentration

322

gradient across the dense layer and thus enhances boron passage. As JW increases, the

323

magnitude of ICP grows exponentially, leading to a greater increment in boron flux. However,

324

in the AL-FS mode, the boron transport is mainly influenced by JW and the boron

325

permeability of the active layer. Thus, its

326

the DS mode, when boron passes through the active layer1 (Figure S2 C) and enters into the

327

porous support, concentrative ICP occurs due to the boron retention by the second active

328

layer. Thus an increasing boron concentration profile from active layer1 to active layer2

329

within the porous layer is established. Consequently, the boron concentrations at both

330

interfaces (i.e., C’ and C”) are raised by the ICP effect so that the boron concentration

331

gradient across the active layer1 is reduced, resulting in a lower boron flux, in comparison to

332

the AL-FS mode. It is noteworthy that although the boron concentration gradient across the

333

active layer2 is increased, the resulting boron flux across active layer2 is limited by the boron

334

flux across active layer1, since the two values must be equal at steady state. Hence, the boron

335

flux in the DS mode is always lower than that in the AL-FS mode.

௃ಳ,ಲಽషಷೄ ஼ಷ

value is less sensitive to JW. In contrast, in

336 ௃ಳ

஼ಷ

337

Interestingly, Figure 4 shows that the

values of both DS and AL-FS modes are

338

approaching the same plateau as water flux increases further. This phenomenon can be

339

explained from their modeling equations of

340

‫ܭ‬௠,஻ ൯ increases exponentially with JW, so that a large JW could bring the two terms

௃ಳ,ಲಽషಷೄ ஼ಷ

and

௃ಳ,ವೄ ஼ಷ

in Table 1. The term exp൫‫ܬ‬௪ /



341

஻భ

஻మ ୣ୶୮൫௃ೢ /௄೘,ಳ ൯

and

஻భ

௃ೈ

Page 16 of 33

to approach zero. Consequently, the denominators in these two

342

equations approach 1. Thus, at high JW values, the normalized boron fluxes of AL-FS and DS

343

modes approach the same limiting value of B1, which is 6.844 LMH or 1.901 µm/s. On the

344

other hand, the modeling equation in the AL-DS mode could be re-written as

345

஻భ ಳ భ ା భ ౛౮౦(಻ೢ /಼೘,ಳ ) ಻ೈ

346

exponentially so that the denominator is dominated by the term

347

As shown in the figure, the

348

fluxes. Physically, as Jw increases, the water convective flow increases, directly bringing

349

more boron into the FO membrane. This phenomenon not only enhances the concentrative

350

ICP of boron but also increases the boron concentration gradient across the active layer. As a

351

result, it increases boron passage. On the other hand, the AL-FS mode does not express such

352

concentrative ICP phenomenon. While in the DS mode, concentrative ICP of boron also

353

happens within the porous layer as boron flux increases by the convective flow. However, the

354

active layer 1 prevents boron from entering and results in lower ICP.

௃ಳ,ಲಽషವೄ ஼ಷ

=

. If JW increases to a certain high value, the term exp൫‫ܬ‬௪ /‫ܭ‬௠,஻ ൯ increases ௃ಳ,ಲಽషವೄ ஼ಷ

஻భ

௃ೈ

, and then

௃ಳ,ಲಽషವೄ ஼ಷ

≈ ‫ܬ‬ௐ .

value increases linearly with water flux at high water

355 356

Figure 5 presents experimental and modeled boron rejections as a function of water flux

357

under various modes. The boron rejections in AL-FS and DS modes increase with an increase

358

in JW and finally reach almost the same plateau. When Jw is small, both active layers play

359

their roles in rejecting boron, with little influence from concentrative ICP of boron within the

360

support. However, once Jw becomes larger, the boron concentration gradient within the

361

support increases due to the larger convective flow and ICP of boron, so the rejection

362

contributed by the active layer 2 is weakened. Thus, the active layer 1 determines the trend of

363

overall boron rejection. On the contrary, the boron rejection in the AL-DS mode initially


Page 17 of 33


364

increases with JW, but decreases with a further increment in JW, due to the significant boron

365

ICP at high water fluxes. Experimental data match well with the model results. The highest

366

boron rejection is 83.9% which is achieved under the DS mode at a water flux of 21.9 LMH.

367

To our best knowledge, this is the highest rejection ever achieved by experimental FO

368

membranes in neutral boron separation as summarized in Table S1 with literature data7-10.

369

The as-prepared double-skinned TFC FO membrane demonstrate a higher boron rejection

370

than commercial BWRO membranes (rejection ratios ≤ 65%), but slightly lower than SWRO

371

membranes (rejection ratios ~ 90%) at neutral pH values1.

372 373

Among the modeling curves in Figure 5, the DS mode seems to always exhibit the highest

374

boron rejection under the same JW, while the AL-DS mode has the lowest rejection. This

375

trend could be further corroborated by comparing their modeling equations. By comparing

376

RAL-FS and RAL-DS, RDS and RAL-FS according to Table 1, we can express their differences as

377

follows:

ܴ஺௅ିிௌ − ܴ‫ܮܣ‬−‫= ܵܦ‬

378

ܴ஽ௌ − ܴ‫ܮܣ‬−‫= ܵܨ‬

379

380

ଵ ೈ /௄೘,ಳ ൯

For JW>0, the term ୣ୶୮൫௃

‫ܤ‬1 ௃ೈ ൥ଵି

൥‫ܤ‬1 +

1

expቀ‫ ܹܬ‬/‫ܭ‬

‫ܹܬ‬ expቀ‫ ܹܬ‬/‫ܭ‬

݉,‫ܤ‬

ቁ

݉,‫ܤ‬

ቁ

൩

(19)

൩(஻భ ା௃ೈ )

஻భమ ௃ೈ

಻ ஻మ ୣ୶୮൬ ೢ ൰(஻భ ା௃ೈ )൦௃ೈ ା஻భ ା ಼೘,ಳ

಻ೈ ಳభ

(20)

൪

಻ ಳమ౛౮౦ቆ ೢ ቇ ಼೘,ಳ

is smaller than 1, thus both eq. 19 and 20 are positive, i.e.,

381

RAL-FS-RAL-DS>0 and RDS-RAL-FS>0. The magnitudes of RAL-FS, RAL-DS and RDS is therefore

382

following the order:

383

ܴ஽ௌ = 1 −

஻భ

಻ೈ ಳభ ಳమ ౛౮౦൫಻ೢ /಼೘,ಳ ൯

௃ೈ ା஻భ ା

> ܴ஺௅ିிௌ = 1 −

஻భ

஻భ ା௃ೈ

> ܴ஺௅ି஽ௌ = 1 −


஻భ

಻ ஻భ ା೐ೣ೛(಻ ೈ /಼ ೢ

೘,ಳ )

(21)


384

Consistent with our experimental data, the boron rejection rate in the DS mode is always the

385

highest, while the one in the AL-DS mode is the lowest under the same JW. Overall, the

386

double-skinned FO membrane exhibits better boron separation performance; namely, a lower

387

boron flux and a higher boron rejection, compared with the single-skinned one. However,

388

deviations still exist between the theoretical modelling and experimental data in Figures 4 &5.

389

This is perhaps due to the uncertainty during the membrane casting, the fabrication of TFC

390

layers which were handled manually. Overall, the close agreement between experimental and

391

modeling results implies that the newly derived model equations could help predict the boron

392

transport behaviors through double-skinned FO membranes. In the following section, the

393

effects of key parameters on the separation performance of double-skinned FO membranes

394

are modelled in order to theoretically simulate and design a better double-skinned FO

395

membrane.

396 397

4.4 Modeling predictions

398

According to eq. 15, the boron rejection of the double-skinned FO membrane is affected by

399

JW and its intrinsic properties such as boron permeability of each active layer and structure

400

parameter. Theoretically, more ideal double-skinned FO membranes should have active

401

layers with lower boron permeability (i.e., smaller B1 and B2 values) and a thinner substrate

402

with a higher porosity and lower tortuosity (i.e., smaller S parameter). Figure 6A~C presents

403

the simulated boron rejection as a function of JW by changing B1, B2 and S parameters

404

separately. In each figure, only one parameter is reduced to half or by 10 times while the

405

other two parameters are invariant. As the boron permeability coefficient decreases from B1

406

to 0.1B1 in Figure 6A, or from B2 to 0.1B2 in Figure 6B, both boron rejection curves shift up

407

and the rejections increase with increasing water flux. An ideal rejection of almost 100%

408

could be achieved in the case of 0.1B1 in the high water flux region. The influence of S


Page 18 of 33

Page 19 of 33


409

parameter on boron rejection is presented in Figure 6C. The ICP effects are reduced by

410

lowering the S parameter, leading to a higher boron rejection. The improvement becomes

411

more pronounced at higher water fluxes. Among the three key parameters, the influence of

412

changing boron permeability of active layers appears to be more significant than changing the

413

S parameter within the experimental range of this study.

414 415

Figure 6D compares the boron rejection as a function of water flux between double-skinned

416

FO membranes and single-skinned ones under the AL-FS mode. The black solid line labeled

417

with “B1, B2 and S” represents the curve simulated with parameters obtained from

418

experiments, while the other colored dotted lines are generated with one or more parameters

419

decreased by half. As observed, the rejection difference between DS and AL-FS modes (i.e.,

420

RDS-RAL-FS) is strongly dependent on water flux. The highest rejection difference always

421

appears within the low Jw range, while the value of RDS-RAL-FS decreases with a further

422

increase in Jw. As water flux increases, the rejection rates of both membranes also increase

423

and both approach to 1, inevitably leading to a smaller difference. Moreover, the ICP of

424

boron within the support under the DS mode becomes severer with high water fluxes, which

425

also contributes to the diminished difference in rejection rate. On the other hand, the highest

426

rejection difference are affected by the modeling variables B1, B2 and S. As B1 is replaced by

427

0.5B1, the value of RDS-RAL-FS is decreased to only 12%, indicating the increased similarity of

428

DS and AL-FS operations. Due to the improved boron retention by active layer1, most of the

429

boron will be rejected there, making the second active layer in the double-skinned FO

430

membrane less important. As a result, the double-skinned FO membrane behaves more like a

431

single-skinned membrane in the AL-FS mode. When further decreasing B2 to 0.5B2, an

432

enlarged boron rejection difference of 20.6% can be achieved.

433

parameter of the additional active layer in the DS mode, a reduced B2 particularly increases


Since B2 is a special


434

the rejection ability of the double-skinned FO membrane. Moreover, the RDS-RAL-FS value is

435

further improved to 21.3% when the S parameter is also reduced, corresponding to a

436

reduction in ICP effects. Future works should aim to enhance double-skinned FO membranes

437

with a higher boron rejection by improving the boron retention ability of active layers and

438

reducing the S parameter of the membrane support.

439 440

In summary, double-skinned FO membranes are proved to be superior in boron rejection.

441

However, if operated under the same clean draw /feed solutions (such as NaCl solution/boron

442

feed in this study), double-skinned FO membranes naturally exhibit a lower water flux than

443

single-skinned ones, due to the additional skin layer. But it is worth noting that double-

444

skinned membranes are reported to have a similar or higher water flux when dealing with

445

viscous draw solutions23 or foulant-containing feed solutions25. The skin layers on both sides

446

of the membrane could help to prevent the clogging or penetration of foulants into the

447

support layer. Also, organic fouling (e.g. alginate fouling) during FO is reported to have no

448

discernible influence on boron rejection8. Thus, double-skinned FO membranes are not

449

necessarily more energy-intensive or having low separation efficiency, but could be

450

promising in providing both high boron rejection and good water flux when treating high

451

fouling tendency solutions, which requires further efforts to explore.

452 453

Nomenclature

454

A

water permeability, L m-2 h-1 bar-1 or LMH/bar

455

Am

effective membrane area, m2

456

AL-DS active layer facing draw solution

457

AL-FS active layer facing feed solution


Page 20 of 33

Page 21 of 33


458

Bs

NaCl permeability coefficient, L m-2 h-1 or LMH

459

BB

boron permeability coefficient, LMH

460

B1

boron permeability of active layer1, LMH

461

B2

boron permeability of active layer2, LMH

462

C

boron concentration at the distance x

463

C’

boron concentration at the interfaces facing active layer1

464

C”

boron concentration at the interfaces facing active layer2

465

CF

boron concentration in the feed

466

CD

boron concentration at active layer2 facing the draw solution

467

Cp

solute concentration in the permeate solution, mg/L

468

Cf

solute concentration in the feed solution, mg/L

469

Ct

NaCl concentration at the end of tests, mg/L

470

CDraw boron concentration of the draw solution at the end of tests, mg/L

471

CFeed

boron concentration of the feed solution at the end of tests, mg/L

472

D

effective boron diffusion coefficient within the porous layer, m2/s

473

DB

boron diffusion coefficient, m2/s

474

Ds

NaCl diffusion coefficient, m2/s

475

DS

double-skinned

476

ICP

internal concentration polarization

477

FO

forward osmosis

478

JB

boron flux, mg s-1m-2

479

JB,dx

boron flux at the distance x

480

JS

salt reverse flux, gm-2h-1, abbreviated as gMH

481

JW

water flux, LMH

482

Km

mass transfer coefficient of boron within the support, µm/s



483

Km,s

NaCl diffusion coefficient within the support, µm/s

484

∆P

hydraulic pressure across the membrane, bar

485

Q

water permeation volumetric flow rate, L/h

486

R

boron rejection, %

487

S

membrane structural parameter, µm

488

∆t

the time duration of one FO experiment operation, h

489

TFC

thin-film-composite

490

∆V

volume of permeation water collected in the draw side over ∆t, L

491

VDraw volume of the draw solution, L

492

Vt

feed volume at the end of tests, L

493

ε

porosity

494

π

osmotic pressure, bar

495

∆π

osmotic pressure difference across the membrane, bar

496

τ

tortuosity

497

t

thickness of support layer, m

498 499

Acknowledgments

500

This research was funded by BASF SE, Germany with a grant number of R-279-000-363-597.

501

Special thanks are given to Dr. Gang Han of NUS and Dr. Natalia Widjojo of BASF

502

(Singapore) for their help and suggestions. This research was also partially funded by the

503

Singapore National Research Foundation under its Competitive Research Program for the

504

project entitled, “Advanced FO Membranes and Membrane Systems for Wastewater

505

Treatment, Water Reuse and Seawater Desalination” (grant numbers: R-279-000-336-281

506

and R-279-000-339-281).

507


Page 22 of 33

Page 23 of 33


508

Supporting Information Available

509

Information addressing membrane formation, boron detection, modeling of single-skinned

510

FO membranes. Water and boron transport mechanisms of different operational modes in FO

511

processes (Figure S1 and Figure S2); morphologies of single-skinned FO membranes (Figure

512

S3); experimental membrane performance of boron flux and boron rejection (Figure S4);

513

predictions of boron rejection of three modes under the influence of different parameters

514

(Figure S5); comparison of FO membranes in boron rejection studies (Table S1).

515 516

References

517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547

1. Tagliabue, M.; Reverberi, A. P.; Bagatin, R., Boron removal from water: needs, challenges and perspectives. J. Clean. Prod. 2014, 77, (0), 56-64. 2. Parks, J. L.; Edwards, M., Boron in the environment. Crit. Rev. Env. Sci. Tec. 2005, 35, (2), 81-114. 3. Wolska, J.; Bryjak, M., Methods for boron removal from aqueous solutions — A review. Desalination 2013, 310, (0), 18-24. 4. Shenvi, S. S.; Isloor, A. M.; Ismail, A. F., A review on RO membrane technology: Developments and challenges. Desalination 2015, 368, 10-26. 5. Tu, K. L.; Nghiem, L. D.; Chivas, A. R., Boron removal by reverse osmosis membranes in seawater desalination applications. Sep. Purif. Technol. 2010, 75, (2), 87-101. 6. Kabay, N.; Güler, E.; Bryjak, M., Boron in seawater and methods for its separation — A review. Desalination 2010, 261, (3), 212-217. 7. Jin, X.; Tang, C. Y.; Gu, Y.; She, Q.; Qi, S., Boric acid permeation in forward osmosis membrane processes: modeling, experiments, and implications. Environ. Sci. Technol. 2011, 45, (6), 2323-2330. 8. Jin, X.; She, Q.; Ang, X.; Tang, C. Y., Removal of boron and arsenic by forward osmosis membrane: Influence of membrane orientation and organic fouling. J. Membr. Sci. 2012, 389, (0), 182-187. 9. Kim, C.; Lee, S.; Shon, H. K.; Elimelech, M.; Hong, S., Boron transport in forward osmosis: Measurements, mechanisms, and comparison with reverse osmosis. J. Membr. Sci. 2012, 419–420, (0), 42-48. 10. Fam, W.; Phuntsho, S.; Lee, J. H.; Cho, J.; Shon, H. K., Boron transport through polyamide-based thin film composite forward osmosis membranes. Desalination 2014, 340, (0), 11-17. 11. Cath, T. Y.; Childress, A. E.; Elimelech, M., Forward osmosis: Principles, applications, and recent developments. J. Membr. Sci. 2006, 281, (1-2), 70-87. 12. Zhao, S. F.; Zou, L.; Tang, C. Y. Y.; Mulcahy, D., Recent developments in forward osmosis: Opportunities and challenges. J. Membr. Sci. 2012, 396, 1-21. 13. Chung, T. S.; Li, X.; Ong, R. C.; Ge, Q.; Wang, H.; Han, G., Emerging forward osmosis (FO) technologies and challenges ahead for clean water and clean energy applications. Curr. Opin. Chem. Eng. 2012, 1, (3), 246-257.



548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595

14. Cartinella, J. L.; Cath, T. Y.; Flynn, M. T.; Miller, G. C.; Hunter, K. W.; Childress, A. E., Removal of natural steroid hormones from wastewater using membrane contactor processes. Environ. Sci. Technol. 2006, 40, (23), 7381-7386. 15. Hancock, N. T.; Xu, P.; Heil, D. M.; Bellona, C.; Cath, T. Y., Comprehensive benchand pilot-scale investigation of trace organic compounds rejection by forward osmosis. Environ. Sci. Technol. 2011, 45, (19), 8483-8490. 16. Asatekin, A.; Mayes, A. M., Oil industry wastewater treatment with fouling resistant membranes containing amphiphilic comb copolymers. Environ. Sci. Technol. 2009, 43, (12), 4487-4492. 17. Liu, Y.; Mi, B., Combined fouling of forward osmosis membranes: Synergistic foulant interaction and direct observation of fouling layer formation. J. Membr. Sci. 2012, 407–408, 136-144. 18. Cornelissen, E. R.; Harmsen, D.; de Korte, K. F.; Ruiken, C. J.; Qin, J. J.; Oo, H.; Wessels, L. P., Membrane fouling and process performance of forward osmosis membranes on activated sludge. J. Membr. Sci. 2008, 319, (1–2), 158-168. 19. Coday, B. D.; Yaffe, B. G. M.; Xu, P.; Cath, T. Y., Rejection of trace organic compounds by forward osmosis membranes: A literature review. Environ. Sci. Technol. 2014, 48, (7), 3612-3624. 20. Cui, Y.; Ge, Q.; Liu, X.-Y.; Chung, T. S., Novel forward osmosis process to effectively remove heavy metal ions. J. Membr. Sci. 2014, 467, (0), 188-194. 21. Wang, K. Y.; Ong, R. C.; Chung, T. S., Double-skinned forward osmosis membranes for reducing internal concentration polarization within the porous sublayer. Ind. Eng. Chem. Res. 2010, 49, (10), 4824-4831. 22. Su, J.; Chung, T. S.; Helmer, B. J.; de Wit, J. S., Enhanced double-skinned FO membranes with inner dense layer for wastewater treatment and macromolecule recycle using Sucrose as draw solute. J. Membr. Sci. 2012, 396, (0), 92-100. 23. Wei, R.; Zhang, S.; Cui, Y.; Ong, R. C.; Chung, T. S.; Helmer, B. J.; Wit, J. S. d., Highly permeable forward osmosis (FO) membranes for high osmotic pressure but viscous draw solutes. J. Membr. Sci. 2015, 496, 132-141. 24. Duong, P. H. H.; Chung, T. S.; Wei, S.; Irish, L., Highly permeable double-skinned forward osmosis membranes for anti-fouling in the emulsified oil–water separation Process. Environ. Sci. Technol. 2014, 48, (8), 4537-4545. 25. Qi, S.; Qiu, C. Q.; Zhao, Y.; Tang, C. Y., Double-skinned forward osmosis membranes based on layer-by-layer assembly—FO performance and fouling behavior. J. Membr. Sci. 2012, 405–406, 20-29. 26. Kang, Y.; Emdadi, L.; Lee, M. J.; Liu, D.; Mi, B., Layer-by-layer assembly of zeolite/polyelectrolyte nanocomposite membranes with high zeolite loading. Environ. Sci. Technol. Lett. 2014, 1, (12), 504-509. 27. Zhang, S.; Wang, K. Y.; Chung, T. S.; Chen, H. M.; Jean, Y. C.; Amy, G., Wellconstructed cellulose acetate membranes for forward osmosis: Minimized internal concentration polarization with an ultra-thin selective layer. J. Membr. Sci. 2010, 360, (1-2), 522-535. 28. Geise, G. M.; Lee, H. S.; Miller, D. J.; Freeman, B. D.; McGrath, J. E.; Paul, D. R., Water purification by membranes: the role of polymer science. J. Polym. Sci. Pt. B-Polym. Phys. 2010, 48, (15), 1685-1718. 29. Luo, L.; Han, G.; Chung, T. S.; Weber, M.; Staudt, C.; Maletzko, C., Oil/water separation via ultrafiltration by novel triangle-shape tri-bore hollow fiber membranes from sulfonated polyphenylenesulfone. J. Membr. Sci. 2015, 476, (0), 162-170.


Page 24 of 33

Page 25 of 33

596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638


30. Zhong, P. S.; Widjojo, N.; Chung, T. S.; Weber, M.; Maletzko, C., Positively charged nanofiltration (NF) membranes via UV grafting on sulfonated polyphenylenesulfone (sPPSU) for effective removal of textile dyes from wastewater. J. Membr. Sci. 2012, 417, 52-60. 31. Widjojo, N.; Chung, T. S.; Weber, M.; Maletzko, C.; Warzelhan, V., A sulfonated polyphenylenesulfone (sPPSU) as the supporting substrate in thin film composite (TFC) membranes with enhanced performance for forward osmosis (FO). Chem. Eng. J. 2013, 220, (0), 15-23. 32. Anastasio, D.; R. McCutcheon, J., Using forward osmosis to teach mass transfer fundamentals to undergraduate chemical engineering students. Desalination 2013, 312, (0), 10-18. 33. López, F. J.; Giménez, E.; Hernández, F., Analytical study on the determination of boron in environmental water samples. Fresen. J. Anal. Chem. 1993, 346, (10-11), 984-987. 34. Sah, R. N.; Brown, P. H., Boron Determination—A Review of Analytical Methods. Microchem. J. 1997, 56, (3), 285-304. 35. Gray, G. T.; McCutcheon, J. R.; Elimelech, M., Internal concentration polarization in forward osmosis: role of membrane orientation. Desalination 2006, 197, (1-3), 1-8. 36. McCutcheon, J. R.; Elimelech, M., Modeling water flux in forward osmosis: Implications for improved membrane design. Aiche J. 2007, 53, (7), 1736-1744. 37. Zhao, S. A. F.; Zou, L. D., Relating solution physicochemical properties to internal concentration polarization in forward osmosis. J. Membr. Sci. 2011, 379, (1-2), 459-467. 38. Luo, L.; Wang, P.; Zhang, S.; Han, G.; Chung, T. S., Novel thin-film composite tribore hollow fiber membrane fabrication for forward osmosis. J. Membr. Sci. 2014, 461, 2838. 39. Geise, G. M.; Paul, D. R.; Freeman, B. D., Fundamental water and salt transport properties of polymeric materials. Progress in Polymer Science 2014, 39, (1), 1-42. 40. Mehta, G. D.; Loeb, S., Internal polarization in the porous substructure of a semipermeable membrane under pressure-retarded osmosis. J. Membr. Sci. 1978, 4, (2), 261265. 41. Tang, C. Y. Y.; She, Q. H.; Lay, W. C. L.; Wang, R.; Field, R.; Fane, A. G., Modeling double-skinned FO membranes. Desalination 2011, 283, 178-186. 42. Lee, K. L.; Baker, R. W.; Lonsdale, H. K., Membranes for power-generation by pressure-retarded osmosis J. Membr. Sci. 1981, 8, (2), 141-171. 43. Cadotte, J. E. Interfacially synthesized reverse osmosis membrane. US patent 4277344, 1981. 44. Ghosh, A. K.; Jeong, B.-H.; Huang, X.; Hoek, E. M. V., Impacts of reaction and curing conditions on polyamide composite reverse osmosis membrane properties. J. Membr. Sci. 2008, 311, (1-2), 34-45. 45. Li, X.; Wang, K. Y.; Helmer, B.; Chung, T. S., Thin-film composite membranes and formation mechanism of thin-film layers on hydrophilic cellulose acetate propionate substrates for forward osmosis processes. Ind. Eng. Chem. Res. 2012, 51, (30), 10039-10050. 46. Yan, H.; Miao, X.; Xu, J.; Pan, G.; Zhang, Y.; Shi, Y.; Guo, M.; Liu, Y., The porous structure of the fully-aromatic polyamide film in reverse osmosis membranes. J. Membr. Sci. 2015, 475, 504-510.

639 640



Page 26 of 33

Boron flux (convection, coupled with Jw)

CF

Boron flux (diffusion) Net boron flux JB

Feed solution

C”

C’ Water flux Jw

B1

x=0 Active layer1

Draw solution B2 CD

x Support layer Active layer2

Figure 1. Boron transport in a double-skinned FO membrane


1

Page 27 of 33


Top surface

Cross section

Bottom surface

Substrate

100nm

1 μm

10 μm Top

Double-skinned

Bottom

1 μm

1 μm

10 μm

Figure 2. Morphologies of the membrane support and double-skinned FO membranes ACS Paragon Plus Environment


80

Page 28 of 33

Water flux Reverse salt flux AL-DS AL-FSO DS O

70

12

Water flux (LMH)

60 6 50 3

40

0

30 20

-3

10

-6

0

0.5M 1M 1.5M 2M

AL-DS mode

0.5M 1M 1.5M 2M

AL-FS mode

1M 2M

3M

4M

Reverse salt flux (gMH)

9

-9

DS mode

Figure 3. Experimental membrane performance of water flux and reverse salt flux in different operation modes (Draw solutions: 0.5 M ~ 4 M NaCl, feed solution: 100 ppm boron solution (pH 6.6)) ACS Paragon Plus Environment

Page 29 of 33


AL-FS AL-DS

14

DS

Modeling Experimental

AL-DS

12

JB/CF (m/s)

10 8 6 4

AL-FS 2

DS

0 0

5

10

15

20

25

30

35

40

45

50

55

60

65

Water flux (LMH)

Figure 4. Experimental and modeled normalized boron fluxes as a function of water flux under various operation modes (Draw solutions: NaCl solutions, feed solution: 100 ppm boron solution (pH 6.6)) ACS Paragon Plus Environment


Page 30 of 33

100

DS

90

AL-FS

Boron rejection (%)

80 70 60 50

AL-DS

40 30 20

AL-FS AL-DS

10

DS

Modeling Experimental

0 0

5

10

15

20

25

30

35

40

45

50

55

60

65

Water flux (LMH)

Figure 5. Experimental and modeled boron rejections as a function of water flux under various operation modes (Draw solutions: NaCl solutions, feed solution: 100 ppm boron solution (pH 6.6)) ACS Paragon Plus Environment

Page 31 of 33


B

100

100

90

90

80

80

Boron rejection (%)

Boron rejection (%)

A

70 60 50 40

B1, B2, S

30

0.5B1, B2, S

20

0.1B1, B2, S

10

70 60 50 40

B1 , B2 , S

30

B1, 0.5B2, S

20

B1, 0.1B2, S

10

0

0 0

5

10

15

20

25

30

35

40

45

50

55

60

65

0

5

10

15

20

Water flux (LMH)

C

25

30

35

40

50

55

60

65

Water flux (LMH)

D

100

0.30 0.27

B1, B2, S

0.24

0.5B1, B2, S

90

0.5B1, 0.5B2, S

80 0.21

0.5B1, 0.5B2, 0.5S

70

RDS-RAL-FS

Boron rejection (%)

45

60 50 40

B1, B2, S

30

B1, B2, 0.5s

20

B1, B2, 0.1s

0.18 0.15 0.12 0.09 0.06 0.03

10

0.00

0 0

5

10

15

20

25

30

35

40

Water flux (LMH)

45

50

55

60

65

0

5

10

15

20

25

30

35

40

45

50

Water flux (LMH)

Figure 6. Predictions of boron rejection of double-skinned FO membranes under the influence of (A) boron permeability of the active layer1 B1, (B) boron permeability of active layer2 B2, (C) ACS Paragon Plus Environment structure parameter S, while (D) is the boron rejection difference between DS and AL-FS modes


Page 32 of 33

Table 1. Modeling equations for normalized boron flux and boron rejection

Mode

Normalized boron flux

Boron rejection

AL-FS mode

𝐽𝐵,𝐴𝐿−𝐹𝑆 𝐵1 = 𝐵 𝐶𝐹 1+ 1 𝐽𝑊

𝑅𝐴𝐿−𝐹𝑆 = 1 −

AL-DS mode

𝐽𝐵,𝐴𝐿−𝐷𝑆 = 𝐶𝐹

𝑅𝐴𝐿−𝐷𝑆 = 1 −

DS mode

𝐵1 exp(𝐽𝑤 /𝐾𝑚,𝐵 ) 𝐵 exp(𝐽𝑤 /𝐾𝑚,𝐵 ) 1+ 1 𝐽𝑊

𝐽𝐵,𝐷𝑆 𝐵1 = 𝐵 𝐵1 𝐶𝐹 1+ 1 + 𝐽𝑊 𝐵2 exp 𝐽𝑤 /𝐾𝑚,𝐵

𝑅𝐷𝑆 = 1 −


𝐵1 𝐵1 + 𝐽𝑊 𝐵1

𝐵1 +

𝐽𝑊 exp(𝐽𝑤 /𝐾𝑚,𝐵 ) 𝐵1

𝐽𝑊 + 𝐵1 +

𝐽𝑊 𝐵1 𝐵2 exp 𝐽𝑤 /𝐾𝑚,𝐵

Page 33 of 33


Table 2. Intrinsic separation properties of membrane selective layers

•

Selective layer

Pure Water permeability, A (LMH/bar)

NaCl rejection

NaCl permeability Bs, (LMH)

Boron rejection

Boron permeability BB, (LMH)

TFC layer on the top surface

1.64±0.14

95.5%

0.195

41.7%

6.844 (B1)

TFC layer on the bottom surface

1.19±0.13

94.7%

0.137

43.1%

4.694 (B2)

RO tests conditions: pressure 3bar, 1000 ppm NaCl feed, or 100 ppm Boron feed


Experiments and Modeling of Boric Acid Permeation through Double

Recommend Documents