Nanometric Graphene Oxide Framework ... - ACS Publications

Page 1 of 33

Environmental Science & Technology

Nanometric Graphene Oxide Framework Membranes with Enhanced Heavy Metal Removal via Nanofiltration

Yu Zhang†, Sui Zhang‡, Tai-Shung Chung†, ‡, *

†

NUS Graduate School for Integrative Science and Engineering,

National University of Singapore, 28 Medical Drive, Singapore 117456

‡

Department of Chemical and Biomolecular Engineering,

National University of Singapore, 4 Engineering Drive 4, Singapore 117585

1

ACS Paragon Plus Environment


Page 2 of 33

1

Abstract

2

A novel dual-modification strategy, including (1) the crosslinking and construction of a GO

3

framework by ethylenediamine (EDA) and (2) the amine-enrichment modification by

4

hyperbranched polyethyleneimine (HPEI), has been proposed to design stable and highly

5

charged GO framework membranes with the GO selective layer thickness of 70 nm for

6

effective heave metal removal via nanofiltration (NF). Results from sonication experiments

7

and positron annihilation spectroscopy confirmed that EDA crosslinking not only enhanced

8

structural stability, but also enlarged the nanochannels among the laminated GO nanosheets

9

for higher water permeability. HPEI 60K was found to be the most effective post-treatment

10

agent that resulted in GO framework membranes with a higher surface charge and lower

11

transport resistance. The newly developed membrane exhibited a high pure water

12

permeability of 5.01 L m-2 h-1 bar-1 and comparably high rejections towards Mg2+, Pb2+, Ni2+,

13

Cd2+ and Zn2+. These results have demonstrated the great potential of GO framework

14

materials in wastewater treatment and may provide insights for the design and fabrication of

15

the next generation 2D-based NF membranes.

16 17

Keywords:

18

Graphene oxide framework, hyperbranched polyethyleneimine, nanofiltration, heavy metal

2


Page 3 of 33


19

1. INTRODUCTION

20

The scarcity of fresh water is a global challenge due to the rapid growth in population and

21

industrialization 1. This leads to a great demand on exploring cost-effective and highly

22

efficient water reuse and desalination technologies 2. Compared with traditional wastewater

23

treatments such as chemical, precipitation and sorption methods, nanofiltration (NF)

24

technology has several advantages, including smaller footprint, lower operation cost and

25

energy consumption, and elimination of chemical residuals 3-7. One important application of

26

NF is for heavy metal removal, which is achieved by combining the size exclusion and the

27

Donnan exclusion separation mechanisms 8-11.

28 29

The ideal NF membrane should have a narrow pore size distribution to achieve a high

30

selectivity and a thin and highly porous structure to ensure a good permeability. Generally,

31

polymeric NF membranes have acceptable separation performance and are widely used

32

owing to easy fabrication with relatively low costs. However, polymeric membranes have

33

some drawbacks including poor chemical and thermal resistance, fouling and physical aging

34

12, 13

35

weaknesses of polymeric membranes and achieve high separation performance

36

Nevertheless, inorganic NF membranes are expensive and only used for special applications.

. Inorganic porous membranes, such as Al2O3, can address the aforementioned 14

.

37 38

Recently, carbon-based materials, e.g. graphene and its derivative graphene oxide (GO), have

39

shown promising to be membrane materials for their easy accessibility, high chemical and

40

mechanical stabilities 15-18. GO is only one atom thick with a lateral length of several hundred 3



41

nm (Figure S-1)

19

42

self-assemblable

20-22

43

achievements have been made for various applications, including gas separation

44

pervaporation dehydration 25-28, and ultra/nanofiltration 29-32.

Page 4 of 33

. This high aspect ratio makes GO nanosheets highly stackable and . By controlling the d-spacing of GO membranes, remarkable 23, 24

,

45 46

Nair et al. found that a sub-micrometer-thick GO membrane can completely block the

47

passage of liquids and gases under a dry state while facilitating the permeation of water vapor

48

33

49

by the empty inter-space between the nonoxidized regions of GO sheets. However, hydration

50

would increase the d-spacing of the GO membrane when it was immersed in water and allow

51

a faster permeation of molecules with sizes smaller than 0.45 nm while blocking ions or

52

molecules larger than that 34. Inspired by Nair’s findings, Huang et al. reported a modified

53

nanostrand-channeled GO membrane using copper hydroxide nanostrands as a sacrificial

54

template to enhance the membrane permeability

55

membranes of tens-of-nanometer thick for dye retention with similar rejection performance

56

but much higher water flux than commercial NF membranes 30. These results suggested that

57

laminar GO membranes are a promising candidate to develop NF membranes with an

58

ultrahigh permeability.

. They attributed this fast transport of water vapor to the low-friction nanocapillaries formed

32

. Han et al. also prepared ultrathin GO

59 60

However, since the integrity of pure GO films is solely maintained by hydrogen bond

61

interactions between the oxygen-containing groups of GO nanosheets, pure GO films suffer 4


Page 5 of 33


62

from low structural and physicochemical stability. When the membrane is immersed in an

63

aqueous solution, the hydration effect will destroy the hydrogen bond and enlarge the

64

d-spacing among GO nanosheets, leading to a significant expansion of the nanochannels in

65

the membrane. Therefore, there is an urgent need to improve the GO membrane design and

66

self-assembly technique to enhance its stability under various conditions 18, 35. So far studies

67

on the separation of small ionic species using GO membranes are limited. It was reported that

68

GO based membranes had lower rejections towards ions than common NF membranes 29, 30.

69

Several approaches have been proposed to address these issues, including layer-by-layer

70

(LbL) dip-coating 36 and construction of GO frameworks 27. However, the LbL membranes

71

assembled via electrostatic interaction tend to swell up in ionic solutions, while fully

72

crosslinked GO framework membranes suffer from a low liquid water permeability.

73 74

In this study, a novel dual-modification strategy, including (1) the crosslinking and

75

construction of a GO framework by ethylenediamine (EDA) and (2) the amine-enrichment

76

modification by hyperbranched polyethyleneimine (HPEI), is designed to fabricate stable and

77

highly charged NF membranes. The first step includes the mixing of EDA with the GO

78

solution and then deposits it on a polycarbonate substrate via a pressure-assisted assembly

79

technique, while the second step modifies and stabilizes the as-prepared membrane by HPEI.

80

Different chemical, physicochemical and morphological characterizations are conducted to

81

investigate the stability and surface properties of the pristine and modified membranes. The

82

newly developed GO composite membrane shows good stability, high water permeability and 5



83

rejections towards a variety of cationic heavy metal ions, including Pb2+, Ni2+, Cd2+ and Zn2+.

84

This study may provide useful insights on the design and fabrication of new generation NF

85

membranes.

Page 6 of 33

86 87

2. EXPERIMENTAL

88

2.1. GO framework membrane preparation

89

Prior to the GO assembly process, the substrates were modified with polydopamine to

90

enhance the adhesion between the substrate and the GO layer 37. Polydopamine coating has

91

been widely employed in membrane fabrication processes to increase substrate hydrophilicity

92

as well as to provide feasible modification sites 38, 39. Firstly, the coating solution (2 mg/mL)

93

was prepared by dissolving 0.16 g dopamine hydrochloride into an 80 mL 0.01 M tris buffer

94

at pH 8.5. The substrates were then immersed in the solution under ambient temperature for 3

95

hours. During this period, dopamine will undergo self-polymerization and form an adhesive

96

layer on the substrate to provide GO nanosheets anchor sites (Figure S3). After that, the

97

substrates were rinsed in DI water to remove the residues.

98 99

In order to prepare the GO framework, a certain amount of GO solution was fully dissolved in

100

a 20 mL 1 wt% EDA water solution. The blend was then filtrated through the modified

101

polycarbonate membrane in a dead-end filtration cell under 1 bar. After the filtration, the

102

membrane was rinsed in DI water to remove the residual EDA. GO framework membranes

103

fabricated under this condition were referred to as GO&EDA. For comparison, a control

104

membrane was prepared following the same procedure but without EDA in water. This 6


Page 7 of 33

105


membrane was referred to as pristine GO.

106 107

The surface charge properties of the as-fabricated GO&EDA membrane were further

108

modified by immersing it into an aqueous solution containing either 1 wt% amine HPEI 60K,

109

PEI 2K or dendrimer G(2,0) for a certain period of time and rinsed by DI water. The resultant

110

membrane is denoted as GO&EDA_HPEI 60K if it is modified by HPEI 60K.

111 112

2.2. Pure water permeability and salt rejection of the GO membranes

113

Pure water permeability (PWP, L m-2 h-1 bar-1, abbreviated as LMH bar-1) and salt rejection

114

(R, %) of these composite GO membranes were tested at 1 bar by a dead-end permeation

115

cell at room temperature (Figure S4). The effective membrane area was 3.14 cm2. PWP of

116

each membrane was measured with DI water and the salt rejection was determined using a

117

1000 ppm ion solution (i.e., NaCl, MgCl2). The membrane was firstly conditioned under the

118

NF mode for 2 hours before collecting permeate samples. During the test, the feed solution

119

was stirred at 500 rpm. PWP and R are calculated as follows:

120

PWP =

121

= 1 − × 100%

(1) (2)

122

where Q is the water flow rate (L/h) at the permeate side of DI water. A is the effective

123

membrane area (m2) and ∆P is the trans-membrane pressure (bar). Cp and Cf are the

124

concentrations of the ion solution in the permeate and feed, respectively, which are

125

determined by a conductivity meter (Metrohm AG). 7



Page 8 of 33

126 127

2.3. Membrane characterizations

128

X-ray photoelectron spectroscopy was employed to determine the surface chemistry of the

129

membranes using a monochromatized Al Kα X-ray source (1486.6 eV photons) at a

130

constant dwell time of 100 ms and a pass energy of 40 eV. The chemical structure of the GO

131

framework layer was analyzed by Fourier transform infrared spectroscopy (FTIR) (Bio-Rad

132

FTS-3500) over the range of 400-4000 cm-1. The polydopamine coating was not applied on

133

the polycarbonate membrane surface for the specific sample where the GO layer was peeled

134

off from the substrate, A pellet was prepared by grinding a 0.5 mg GO layer sample with

135

99.5 mg KBr powder and compressed under 8 bar, which was then characterized by the

136

direct transmittance mode. The total number of scans for each sample was 16.

137 138

The morphologies of the GO framework membranes were observed by field emission

139

scanning electron microscopy (FESEM JEOL JSM-6700LV). Prior to FESEM observation,

140

the membrane was freeze-dried before fractured in liquid nitrogen. The membrane samples

141

were then fixed on stubs and coated with a thin layer of platinum under a vacuum condition

142

by a Jeol JFC-1100e Ion Sputtering device. The GO nanosheets and membrane thicknesses

143

were examined by an Atomic force microscope (AFM) (Agilent Technologies, USA) under

144

the tapping mode. A membrane stability test was carried out via sonicating the membrane

145

sample under the degas mode by Elmasonic (S 30 H) for 10 minutes.

146 8


Page 9 of 33


147

The surface charge properties of the pristine GO, GO&EDA and GO&EDA_HPEI 60K

148

framework membranes were analyzed by a SurPASS electrokinetic analyzer (Anton Paar

149

GmbH, Austria). A 450 mL 0.01 M NaCl solution was used to measure the ζ-potential of the

150

membranes under neutral pH. In order to determine the ζ-potential as a function of pH from

151

2.5 to 10, the 0.01 M NaCl solution was firstly auto-titrated with 0.1 M HCl to pH 2.5

152

followed by auto-titration with 0.1 M NaOH to pH 10.

153 154

The microstructural evolution of GO framework membranes was investigated by Doppler

155

broadening energy spectroscopy (DBES). This measurement was conducted using positron

156

annihilation spectroscopy (PAS) in our laboratory. 22Na was used as the source of positrons

157

and experiments were performed at a counting rate of 3000-4000 counts s-1. For each

158

spectrum, a total of one millions counts was taken. The detailed testing procedures have

159

been elaborated elsewhere

160

used to determine the microstructural changes along the membrane depth profile. The mean

161

depth of the materials associated with the incident positron energy can be calculated by the

162

following equation:

163

40

. S- parameter derived from the annihilation spectra, can be

Z ( ) = × .

(3)

164

Where Z is the mean depth (nm), is the material density (g cm-3) and E+ is the incident

165

positron energy. The average density employed in the fitting was calculated via the weight

166

and volume of the membrane as 0.95 g cm-3.

9



Page 10 of 33

167

3. RESULTS AND DISCUSSION

168

3.1. Chemical

169

membranes

170

Figure 1 shows the FTIR spectra of the pristine GO, GO&EDA and GO&EDA_HPEI 60K

171

membranes under the transmission mode. The pristine GO (Figure 1a) has characteristic

172

peaks of –OH stretching at 3400 cm-1, -C=O stretching at 1732 cm-1 and -OH stretching at

173

1414 cm-1 in -COOH, C-O-C stretching at 1224 cm-1, which well corresponds to its

174

structure as illustrated in Figure S-2a in the Supporting Information 17. After crosslinked by

175

EDA, new peaks appear at about 1357 cm-1 and 3250 cm-1 (Figure 1b), which are

176

representatives of the –CN in secondary amine and the -NH bond of primary amines,

177

respectively 41. In addition, compared with the spectrum of pristine GO, the epoxy content

178

almost disappears and the intensity of carboxyl group decreases significantly. This indicates

179

that crosslinking reactions and electrostatic interaction have taken place between the amine

180

group of EDA and the epoxy or carboxylic acid group of GO. For GO&EDA_HPEI 60K, a

181

new peak can be observed at 1570 cm-1, which is attributed to the abundant secondary

182

amine on the HPEI chains. In addition, part of the free amine groups of HPEI molecules

183

may also react with the residual epoxy group in GO to contribute to the peak intensity.

and

morphological

characterizations

of

go

framework

184 185

The reactions are further confirmed by XPS as shown in Table S-1 in the Supporting

186

Information. The pristine GO membrane contains an O/C mass ratio of 0.4 and no nitrogen

187

content could be detected. An increment in N1s content and a decline in O1s content for 10


Page 11 of 33


188

GO&EDA and GO&EDA_HPEI 60K are observed Correspondingly, O-C=O, C=O, C-O

189

and C-C signals are detected on the surface of the pristine GO membrane in Figure S-5a.

190

For the GO&EDA framework membrane, the peak intensities of O-C=O, C=O and C-O

191

decline while a new peak of C-N appears due to the introduction of EDA into the GO layers

192

(Figure S-5b). Figure S-6a shows the peaks of primary, secondary and quaternary amine

193

groups on the surface of GO&EDA, indicating the presence of free amine, formation of

194

covalent bonding and electrostatic interaction between GO and EDA. The peak intensity of

195

C-N in GO&EDA_HPEI 60K increases substantially owing to the C-N bond of HPEI 60K

196

on the membrane surface (Figure S-5c). In the meantime, the peak intensities of C-O and

197

C=O further decrease while the O-C=O peak disappears, which may be due to the formation

198

of the covalent bond and ionic interaction between oxygen-containing groups of GO with

199

amine groups.

200 201

Figure 2b and 2c elucidates the chemical structures on the surfaces of GO&EDA and

202

GO&EDA_HPEI 60K membranes. It is worth noting that EDA may either perpendicularly

203

crosslink two parallel GO nanosheets or two adjacent nanosheets. Moreover, there might be

204

only one amine group of EDA reacting with GO nanosheets, which explains the presence of

205

amine group in the FTIR spectrum of the GO&EDA composite membrane. Therefore, the

206

dual modification strategy; namely, the first step by EDA to crosslink the GO sheets and the

207

second step by HPEI to impart the surface with abundant positive charge, works well on the

208

GO membranes. 11



Page 12 of 33

209 210

The structural integrity of the GO layer over the polycarbonate support is greatly enhanced

211

by the dual modifications as evidenced in Figure S-7 in the Supporting Information, which

212

shows the digital images of pristine GO and GO&EDA_HPEI 60K after a vigorous

213

sonication process. Originally, both samples have the same brown appearance indicative of a

214

thin GO layer. After the sonication process. ¾ of the pristine GO membrane becomes

215

translucent, implying most GO nanosheets have been removed by the sonication, whilst

216

GO&EDA_HPEI 60K maintains its appearance. This phenomenon suggests the structural

217

stability of GO&EDA_HPEI 60K has been improved as compared with pristine GO.

218 219

In order to optimize the preparation conditions, Figure S-8 shows the PWP and MgCl2

220

rejection of GO framework membranes as functions of GO loading and immersion duration

221

in the HPEI 60K solution. The GO framework membrane comprising 0.09375 mg GO

222

(specific deposition: 0.083 g/m2) and being modified by 1 wt% HPEI 60K aqueous solution

223

for 20 min has the best balanced separation performance; namely, a MgCl2 rejection of 96.3%

224

± 1.3% and a PWP of 5.01 ± 0.24 LMH bar -1. It is therefore chosen as a representative of

225

the GO&EDA_HPEI 60K for the subsequent studies.

226 227

Figure 3a-d displays the surface and cross section morphologies of the substrate and the

228

composite GO&EDA_HPEI 60K membrane. The polycarbonate substrate has a smooth

229

surface with uniform pores of 0.2 µm in diameter. After the deposition of GO nanosheets, a 12


Page 13 of 33


230

well packed laminated film is formed on the substrate surface (Figure 3b). The wrinkles on

231

the top surface of the GO membrane are formed from the edges and folding of GO

232

nanosheets. The thickness of the GO layer was measured by FESEM and AFM. Figure 3d

233

shows its FESEM cross section image on the substrate with a thickness of less than 100 nm,

234

while Figure 3e displays its height profile along the line across the membrane surface drawn

235

by AFM with an average thickness of 69 nm. Since the specific GO amount for this NF

236

membrane is only 0.083 g/m2, the newly developed NF membrane could be cost effective

237

and prospective for practical wastewater treatment.

238 239

3.2. Effects of EDA crosslinking and HPEI modification on rejection

240

performance in NF processes

241

Figure 4a compares the PWP values of the pristine GO, GO&EDA and GO&EDA_HPEI

242

60K membranes. GO&EDA has the highest PWP of 9.79 ± 2.20 LMH bar-1, which is

243

significantly higher than that of the pristine GO membrane. One possible explanation is that

244

the orientation of GO nanosheets in the pristine GO membrane is highly ordered and

245

densely packed

246

For perpendicularly crosslinked GO nanosheets, the expansion of d-spacing under the wet

247

condition is retarded, which would reduce water permeability. In contrast, if EDA

248

crosslinks two adjacent nanosheets, it would enlarge water transport channels and increase

249

PWP. Similarly, the incorporation of free amine and un-ordered EDA molecules among GO

250

nanosheets may introduce defects and lower their packing density, thus increasing PWP.

42

. For GO&EDA, EDA may crosslink GO nanosheets in different ways.

13



Page 14 of 33

251

These competing mechanisms result in an enhanced pure water permeability for the

252

GO&EDA membrane. The S- parameter values measured by DBES spectroscopy confirm

253

our hypotheses. Figure 5 shows the S- parameter as a function of the incident positron

254

energy for these three membranes. A smaller S- parameter indicates either a smaller free

255

volume size or a lower free volume content of the membrane

256

(i.e., corresponding to the thin GO layer), the S- parameter of the pristine GO membrane is

257

much lower than the other two. It is hence reasonable to conclude that the pristine GO has

258

the densest and most ordered packing, while GO&EDA and GO&EDA_HPEI 60K have a

259

comparable packing density. Although the packing densities of GO nanosheets in

260

GO&EDA_HPEI 60K and GO&EDA are similar, HPEI 60K crosslinks the top layer of

261

GO&EDA_HPEI 60K and thus its pure water permeability is reduced to 5.01 ± 0.24 LMH

262

bar -1.

25

. At low incident energy

263 264

The ζ-potential of the pristine GO, GO&EDA and GO&EDA_HPEI 60K as a function of

265

pH is shown in Figure 6. In a wide pH range from 3 to 11, the pristine GO membrane is

266

negatively charged, which is mainly due to the deprotonation of the carboxyl group at the

267

edges of GO nanosheets. The GO&EDA is slightly positive charge below pH 4.5 but

268

becomes neutral at pH 4.5 and then negative charge at higher pH values owing to the

269

complicated functional groups on the GO&EDA membrane surface such as hydroxyl,

270

unreacted carboxyl acid, unreacted amine and amide groups. For GO&EDA_HPEI 60K,

271

since additional amine groups are introduced on the GO framework surface, the isoelectric 14


Page 15 of 33


272

point shifts to pH 10.5 due to the protonation of the free amine groups. Moreover, its

273

ζ-potential is around 100 mV at neutral pH, which is higher than most reported data for

274

positively charged membranes 43, 44.

275 276

Figure 4 also displays the rejections of these membranes against MgCl2, MgSO4, NaCl and

277

Na2SO4 solutions. The pristine GO has the highest rejection towards Na2SO4, which is

278

consistent with the literature data 30, then the rejection follows an order of R (Na2SO4) > R

279

(MgSO4) > R (NaCl) > R (MgCl2). However, the rejection order is reversed for

280

GO&EDA_HPEI 60K. This phenomenon can be explained by the Donnan exclusion effect.

281

GO is negatively charged in a wide pH range. Therefore, it tends to extrude co-ions, such as

282

SO42- and Cl-1. In order to maintain the electroneutrality of the solutions at each side of the

283

GO composite membrane, the counter ions Na+ and Mg2+ have to be rejected as well.

284

According to Donnan exclusion theory, the rejection rate is related to the valences of the ion

285

species, following the order of Zco-ions/Zcount-ions (Z refers to the valence). On the contrary, the

286

GO&EDA_HPEI 60K membrane is positively charged after being crosslinked by

287

amine-rich HPEI molecules. In this case, the co-ions are Mg2+ and Na+. The membrane

288

thereby shows a higher rejection against divalent cations (Mg2+). Since the GO&EDA has a

289

larger free volume and close-to-neutral charge at neutral pH, the rejections towards different

290

salts are lower compared with the other two membranes.

291

15



Page 16 of 33

292

3.3. Effects of different amine-enrichment modifications

293

Besides HPEI 60K, PEI 2K and dendrimer G(2,0) were employed to investigate the effects

294

of various amine modifications. PEI 2K has a smaller molecular weight and dendrimer G(2,

295

0) is a micelle shaped molecule with abundant amine groups. The post-treatment conditions

296

for PEI 2K and dendrimer G(2,0) have been optimized and Table 1 tabulates the best NF

297

performance for each membrane.

298 299

The PEI 2K post-treatment is unfavorable for both permeability and rejections because the

300

resultant membrane has a lower rejection of 89% towards MgCl2 and a significantly

301

reduced PWP of 0.39 LMH bar -1. This is due to the fact that the molecular size of PEI 2K is

302

smaller. It can easily block surface pores of the GO layer and react with GO nanosheets 45,

303

resulting in a denser surface with a reduced water permeability. For the membrane

304

post-treated by dendrimer G(2,0), the rejection remains high while the PWP drops 4-fold.

305

This arises from the fact that dendrimer G (2.0) may form a denser crosslinking layer than

306

that from HPEI 60K because HPEI 60K has a longer molecular chain and higher chain

307

flexibility. As a consequence, the membrane post-treated by the former has a comparable

308

rejection but a lower PWP. Clearly, HPEI 60K is the most efficient agent to modify the GO

309

framework membrane with higher surface charge but lower transport resistance.

310 311

3.4. Heavy metal rejection of the HPEI 60K modified GO framework membrane

312

The newly developed GO&EDA_HPEI 60K membrane was further tested by 1000 ppm 16


Page 17 of 33


313

Pb(NO3)2, NiCl2, CdCl2 and ZnCl2 solutions, respectively. High rejections to these divalent

314

cationic heavy metal ions have been achieved as summarized in Table 2, especially for

315

Pb(NO3)2 and NiCl2. Since the newly developed GO framework membrane is highly

316

positively charged at neutral pH, it can effectively repel the cationic heavy metal ions. Table

317

2 shows a benchmarking of the current GO framework membrane with some other NF

318

membranes

319

this work.

8-10, 46, 47

. A much higher PWP with comparable rejections has been achieved in

320 321

ASSOCIATED CONTENT

322

Supporting Information

323

Materials, AFM images of GO nanosheets, chemical structures of the different compounds

324

used in this work, FESEM images of substrate membrane before and after polydopamine

325

modification, experimental apparatus, XPS C 1s and N 1s narrow scan spectra of different

326

membranes, digital photo images of GO framework membranes, effect of GO loading and

327

post-treatment duration on NF performance and XPS characterizations results. This material

328

is available free of charge via the Internet at http:// pubs. acs. org.

329 330

AUTHOR INFORMATION

331

Corresponding author

332

* Tel: +65-65166645. Fax: +65-67791936. Email: [email protected].

333

Notes 17



334

Page 18 of 33

The authors declare no completing financial interest.

335 336

ACKNOWLEDGEMENT

337

The authors would like to acknowledge the financial support from the Singapore National

338

Research Foundation (NRF) Competitive Research Program for the project entitled

339

“Advanced FO membranes and Membrane systems for Wastewater Treatment, Water Reuse

340

and Seawater Desalination” (Grants R-279-000-336-281 and R-278-000-339-281). Special

341

thanks are due to Dr. Y. Tang, Dr. K.S. Liao, Ms. J. Gao and Mr. S. Japip for their suggestions

342

on the experimental work and paper writing.

343 344

ABBREVIATIONS

345

AFM, atomic force microscopy; DBES , doppler broadening energy spectroscopy; EDA,

346

ethylenediamine; FESEM, field emission scanning electron microscopy; FTIR, Fourier

347

transform

348

polyethyleneimine; LbL, layer-by-layer, NF, nanofiltration; PAS, positron annihilation

349

spectroscopy; PWP, pure water permeability; XPS, X-ray photoelectron spectroscopy.

infrared

spectroscopy;

GO,

graphene

oxide;

HPEI,

hyperbranched

350 351

REFERENCES

352

1.

353

P.; Glidden, S.; Bunn, S. E.; Sullivan, C. A.; Liermann, C. R.; Davies, P. M., Global threats

354

to human water security and river biodiversity. Nature 2010, 467, 555-61.

Vorosmarty, C. J.; McIntyre, P. B.; Gessner, M. O.; Dudgeon, D.; Prusevich, A.; Green,

18


Page 19 of 33


355

2.

Escobar, I. C.; Van der Bruggen, B., Modern applications in membrane science and

356

technology. Oxford University Press: Oxford, 2011.

357

3.

358

solutes during NF/RO treatment--a literature review. Water Res. 2004, 38, 2795-809.

359

4.

360

membranes for water environmental remediation. J. Membr. Sci. 2015, 476, 95-104.

361

5.

362

Osmosis and Nanofiltration Membranes. Environ. Sci. Technol. 2003, 37, 4435-4441.

363

6.

364

fiber membranes through a single-step direct spinning technique. Environ. Sci. Technol.

365

2014, 48, 13933-40.

366

7.

367

der Bruggen, B., Fractionation of direct dyes and salts in aqueous solution using loose

368

nanofiltration membranes. J. Membr. Sci. 2015, 477, 183-193.

369

8.

370

Approach To Fabricate a Chitosan Rejecting Layer over Poly(ether sulfone) Support for

371

Heavy Metal Removal. Ind. Eng. Chem. Res. 2015, 54, 472-479.

372

9.

373

nanofiltration (NF) hollow fiber membranes for high efficient heavy metal removal. Water

374

Res. 2014, 63, 252-61.

375

10. Thong, Z.; Han, G.; Cui, Y.; Gao, J.; Chung, T. S.; Chan, S. Y.; Wei, S., Novel

Bellona, C.; Drewes, J. E.; Xu, P.; Amy, G., Factors affecting the rejection of organic

Cheng, X. Q.; Shao, L.; Lau, C. h., High flux polyethylene glycol based nanofiltration

Escobar, I. C.; Peng, W., Rejection Efficiency of Water Quality Parameters by Reverse

Ong, Y. K.; Chung, T. S., Mitigating the hydraulic compression of nanofiltration hollow

Lin, J.; Ye, W.; Zeng, H.; Yang, H.; Shen, J.; Darvishmanesh, S.; Luis, P.; Sotto, A.; Van

Zhang, S.; Peh, M. H.; Thong, Z.; Chung, T.-S., Thin Film Interfacial Cross-Linking

Gao, J.; Sun, S. P.; Zhu, W. P.; Chung, T. S., Chelating polymer modified P84

19



Page 20 of 33

376

nanofiltration membranes consisting of a sulfonated pentablock copolymer rejection layer

377

for heavy metal removal. Environ. Sci. Technol. 2014, 48, 13880-7.

378

11. Fu, F.; Wang, Q., Removal of heavy metal ions from wastewaters: a review. J. Environ.

379

Manage. 2011, 92, 407-18.

380

12. Van der Bruggen, B.; Mänttäri, M.; Nyström, M., Drawbacks of applying nanofiltration

381

and how to avoid them: A review. Sep. Purif. Technol. 2008, 63, 251-263.

382

13. Xu, P.; Drewes, J. E.; Kim, T.-U.; Bellona, C.; Amy, G., Effect of membrane fouling on

383

transport of organic contaminants in NF/RO membrane applications. J. Membr. Sci. 2006,

384

279, 165-175.

385

14. Van Gestel, T.; Van decasteele, C.; Buekenhoudt, A.; Dotremont, C.; Luyten, J.; Leysen,

386

R.; Van der Bruggen, B.; Maes, G., Salt retention in nanofiltration with multilayer ceramic

387

TiO2 membranes. J. Membr. Sci. 2002, 209, 379-389.

388

15. Geim, A. K.; Novoselov, K. S., The rise of graphene. Nature materials 2007, 6,

389

183-191.

390

16. Geim, A. K., Graphene: status and prospects. Science 2009, 324, 1530-4.

391

17. Dreyer, D. R.; Park, S.; Bielawski, C. W.; Ruoff, R. S., The chemistry of graphene

392

oxide. Chem. Soc. Rev. 2010, 39, 228-40.

393

18. Huang, H.; Ying, Y.; Peng, X., Graphene oxide nanosheet: an emerging star material for

394

novel separation membranes. J. Mater. Chem. A 2014, 2, 13772.

395

19. Meyer, J. C.; Geim, A. K.; Katsnelson, M. I.; Novoselov, K. S.; Booth, T. J.; Roth, S.,

396

The structure of suspended graphene sheets. Nature 2007, 446, 60-3. 20


Page 21 of 33


397

20. Chen, C.; Yang, Q.-H.; Yang, Y.; Lv, W.; Wen, Y.; Hou, P.-X.; Wang, M.; Cheng, H.-M.,

398

Self-Assembled Free-Standing Graphite Oxide Membrane. Adv. Mater. 2009, 21,

399

3007-3011.

400

21. Dikin, D. A.; Stankovich, S.; Zimney, E. J.; Piner, R. D.; Dommett, G. H.; Evmenenko,

401

G.; Nguyen, S. T.; Ruoff, R. S., Preparation and characterization of graphene oxide paper.

402

Nature 2007, 448, 457-60.

403

22. Smith, Z. P.; Freeman, B. D., Graphene oxide: a new platform for high-performance

404

gas- and liquid-separation membranes. Angew. Chem. Int. Ed. Engl. 2014, 53, 10286-8.

405

23. Kim, H. W.; Yoon, H. W.; Yoon, S. M.; Yoo, B. M.; Ahn, B. K.; Cho, Y. H.; Shin, H. J.;

406

Yang, H.; Paik, U.; Kwon, S.; Choi, J. Y.; Park, H. B., Selective gas transport through

407

few-layered graphene and graphene oxide membranes. Science 2013, 342, 91-5.

408

24. Li, H.; Song, Z.; Zhang, X.; Huang, Y.; Li, S.; Mao, Y.; Ploehn, H. J.; Bao, Y.; Yu, M.,

409

Ultrathin, molecular-sieving graphene oxide membranes for selective hydrogen separation.

410

Science 2013, 342, 95-8.

411

25. Tang, Y. P.; Paul, D. R.; Chung, T. S., Free-standing graphene oxide thin films

412

assembled by a pressurized ultrafiltration method for dehydration of ethanol. J. Membr. Sci.

413

2014, 458, 199-208.

414

26. Hung, W.-S.; An, Q.-F.; De Guzman, M.; Lin, H.-Y.; Huang, S.-H.; Liu, W.-R.; Hu,

415

C.-C.; Lee, K.-R.; Lai, J.-Y., Pressure-assisted self-assembly technique for fabricating

416

composite membranes consisting of highly ordered selective laminate layers of amphiphilic

417

graphene oxide. Carbon 2014, 68, 670-677. 21



Page 22 of 33

418

27. Hung, W.-S.; Tsou, C.-H.; De Guzman, M.; An, Q.-F.; Liu, Y.-L.; Zhang, Y.-M.; Hu,

419

C.-C.; Lee, K.-R.; Lai, J.-Y., Cross-Linking with Diamine Monomers To Prepare Composite

420

Graphene Oxide-Framework Membranes with Varyingd-Spacing. Chem. Mater. 2014, 26,

421

2983-2990.

422

28. Huang, K.; Liu, G.; Lou, Y.; Dong, Z.; Shen, J.; Jin, W., A graphene oxide membrane

423

with highly selective molecular separation of aqueous organic solution. Angew. Chem. Int.

424

Ed. Engl. 2014, 53, 6929-32.

425

29. Hu, M.; Mi, B., Enabling graphene oxide nanosheets as water separation membranes.

426

Environ. Sci. Technol. 2013, 47, 3715-23.

427

30. Han, Y.; Xu, Z.; Gao, C., Ultrathin Graphene Nanofiltration Membrane for Water

428

Purification. Adv. Funct. Mater. 2013, 23, 3693-3700.

429

31. Xu, C.; Cui, A.; Xu, Y.; Fu, X., Graphene oxide–TiO2 composite filtration membranes

430

and their potential application for water purification. Carbon 2013, 62, 465-471.

431

32. Huang, H.; Song, Z.; Wei, N.; Shi, L.; Mao, Y.; Ying, Y.; Sun, L.; Xu, Z.; Peng, X.,

432

Ultrafast viscous water flow through nanostrand-channelled graphene oxide membranes.

433

Nat. Commun. 2013, 4, 2979.

434

33. Nair, R. R.; Wu, H. A.; Jayaram, P. N.; Grigorieva, I. V.; Geim, A. K., Unimpeded

435

permeation of water through helium-leak-tight graphene-based membranes. Science 2012,

436

335, 442-4.

437

34. Joshi, R. K.; Carbone, P.; Wang, F. C.; Kravets, V. G.; Su, Y.; Grigorieva, I. V.; Wu, H.

438

A.; Geim, A. K.; Nair, R. R., Precise and ultrafast molecular sieving through graphene oxide 22


Page 23 of 33


439

membranes. Science 2014, 343, 752-4.

440

35. Mi, B., Materials science. Graphene oxide membranes for ionic and molecular sieving.

441

Science 2014, 343, 740-2.

442

36. Hu, M.; Mi, B., Layer-by-layer assembly of graphene oxide membranes via

443

electrostatic interaction. J. Membr. Sci. 2014, 469, 80-87.

444

37. Dreyer, D. R.; Miller, D. J.; Freeman, B. D.; Paul, D. R.; Bielawski, C. W., Perspectives

445

on poly(dopamine). Chem. Sci. 2013, 4, 3796.

446

38. Wang, Z.; Jiang, X.; Cheng, X.; Lau, C. H.; Shao, L., Mussel-Inspired Hybrid Coatings

447

that Transform Membrane Hydrophobicity into High Hydrophilicity and Underwater

448

Superoleophobicity for Oil-in-Water Emulsion Separation. ACS Appl. Mater. Interfaces

449

2015, 7, 9534-45.

450

39. Li,

451

polyglycerol-grafted poly(ether sulfone) hollow fiber membranes for osmotic power

452

generation. Environ. Sci. Technol. 2014, 48, 9898-907.

453

40. Zhang, S.; Wang, K. Y.; Chung, T.-S.; Chen, H.; Jean, Y. C.; Amy, G., Well-constructed

454

cellulose acetate membranes for forward osmosis: Minimized internal concentration

455

polarization with an ultra-thin selective layer. J. Membr. Sci. 2010, 360, 522-535.

456

41. Socrates, G., Infrared and Raman Characteristic Group Frequencies: Tables and

457

Charts. Wiley: 2004.

458

42. Tsou, C.-H.; An, Q.-F.; Lo, S.-C.; De Guzman, M.; Hung, W.-S.; Hu, C.-C.; Lee, K.-R.;

459

Lai, J.-Y., Effect of microstructure of graphene oxide fabricated through different

X.;

Cai,

T.;

Chung,

T.

S.,

Anti-fouling

behavior

of

hyperbranched

23



Page 24 of 33

460

self-assembly techniques on 1-butanol dehydration. J. Membr. Sci. 2015, 477, 93-100.

461

43. Gao, J.; Sun, S.-P.; Zhu, W.-P.; Chung, T.-S., Polyethyleneimine (PEI) cross-linked P84

462

nanofiltration (NF) hollow fiber membranes for Pb2+ removal. J. Membr. Sci. 2014, 452,

463

300-310.

464

44. Wang, H.; Zhang, Q.; Zhang, S., Positively charged nanofiltration membrane formed by

465

interfacial polymerization of 3,3′,5,5′-biphenyl tetraacyl chloride and piperazine on a

466

poly(acrylonitrile) (PAN) support. J. Membr. Sci. 2011, 378, 243-249.

467

45. Sun, S. P.; Hatton, T. A.; Chung, T. S., Hyperbranched polyethyleneimine induced

468

cross-linking of polyamide-imide nanofiltration hollow fiber membranes for effective

469

removal of ciprofloxacin. Environ. Sci. Technol. 2011, 45, 4003-9.

470

46. Zhu,

471

polybenzimidazole/polyethersulfone (PBI/PES) nanofiltration (NF) hollow fiber membranes

472

for heavy metals removal from wastewater. J. Membr. Sci. 2014, 456, 117-127.

473

47. Al-Rashdi, B. A. M.; Johnson, D. J.; Hilal, N., Removal of heavy metal ions by

474

nanofiltration. Desalination 2013, 315, 2-17.

W.-P.;

Sun,

S.-P.;

Gao,

J.;

Fu,

F.-J.;

Chung,

T.-S.,

Dual-layer

475 476 477

24


Page 25 of 33

478


TOC Art

479 480

481

25



C-O-C

-OH hydroxyl

(a)

C=O C=C

(b)

C-OH -CN- secondary amine

-NH secondary amine

(c)

pristine GO GO&EDA GO&EDA_HPEI 60K

-NH Primary amine 4500

Page 26 of 33

4000

3500

3000

2500

Wavenumber

2000

1500

1000

500

0

(cm-1)

Figure 1. FTIR spectra of the GO layers: (a) pristine GO, (b) GO&EDA and (c) GO&EDA_HPEI 60K framework membranes, measured in the transmission mode.


Page 27 of 33


O

OH HO2C

OH

O

OH

O

O

O

O

CO2H

OH

OH

OH

O

O

OH

HO2C

CO2H

CO2H

OH

OH

OH

CO

OH

O

(a)

OH

HO2C

OH

OH

OH

CO

OH OH

O

OH CO H 2 OH

HO2C

OH HO2C

OH

O

(b)

HO2C

OH

OH

O

(c)

Figure 2. Schematic diagrams of the purposed structures: (a) pristine GO, (b) GO&EDA and (c) GO&EDA_HPEI 60K frameworks. ACS Paragon Plus Environment


1 μm

(b)

1 μm

(c)

100 nm

(d)

100nm nm 100

(e)

20 nm

(a)

Page 28 of 33

0

-20 -40

Membrane average thickness: 69.41 ± 3.85 nm

-60 -80 -100 0

2

4

6

8

10

12

Figure 3. The top surface morphology of (a) the Whatman® Cyclopore® polycarbonate membrane and (b) the GO&EDA_HPEI 60K composite membrane. The cross section morphology of (c) the substrate layer and (d) the GO layer. (e) The AFM image of the GO layers on a mica film. ACS Paragon Plus Environment

14 μm

Page 29 of 33


PWP (LMH bar -1)

(a) 15

10

5

0 pristine GO

Salt Rejection (%)

(b)100

MgCl2

GO&EDA

NaCl

MgSO4

GO&EDA_HPEI 60K Na2SO4

80 60 40 20

0 pristine GO

GO&EDA

GO&EDA_HPEI 60K

Figure 4. (a) The PWP of the three GO framework membranes at a transmembrane pressure of 1 bar, and (b) the rejection performance ofACS theParagon membranes for four different salt solutions (each 1000 Plus Environment ppm).


0

0.55

Mean depth (µm) 1.68 3.21 5.08

Page 30 of 33

7.26

9.72

S- parameter

0.52 0.5 0.48 pristine GO

0.46

GO&EDA

0.44

GO&EDA_HPEI 60K

0.42 0

0

5

4.6

10 15 20 25 Incident positron energy (keV)

30

Mean depth (nm) 13.9 26.6 40.0 60.2 80.6 103.1 127.6

S- parameter

0.47 0.46 GO framework layer

0.45 0.44 0.43 0.42 0

0.25

0.5 0.75 1 1.25 1.5 1.75 Incident positron energy (keV)

2

Figure 5 . S- parameters of the three GO framework membranes against the incident positron energy. ACS Paragon Plus Environment

Page 31 of 33


150

ζ-potential (mV)

100 50 0 0

2

4

6

8

12

pH

-50

-100

10

pristine GO GO&EDA

-150

GO&EDA_HPEI 60K

Figure 6. ζ-potential as a function of pH of the three GO framework membranes. ACS Paragon Plus Environment


Page 32 of 33

Table 1. Effects of post-treatment agents on NF performance of 1000 ppm MgCl 2 .

Reaction duration*

PWP (LMH bar -1)

Rejection (%)

ζ-potential (mV)

30 s

0.39

89.0

72.6

HPEI 60K

20 min

5.01

96.3

92.6

Dendrimer G(2,0)

20 min

1.20

97.2

101.0

PEI 2K

*

The reaction duration was optimized to the highest rejection performance


Page 33 of 33


Table 2. A benchmarking of NF membranes for heavy metal removal.

Membrane Chitosan PES composite membrane

PWP (LMH bar -1)

Ion

Testing condition

Rejection (%)

Refs.

3.45

NiCl2

1000 ppm, 10 bar

96.3

6

Pb(NO3)2

93.0

Chelating polymer modified P84

˃1

Pb(NO3)2

1000 ppm, 10 bar

˃ 99

7

Kraton matrimid composite membrane

2.4

Pb(NO3)2

1000 ppm, 10 bar

99.8

8

˃ 98

CdCl2 PBI/PES dual-layer hollow fiber

0.826

Pb(NO3)2

200 ppm, 1 bar

93

42

Dow membrane NF270

13.2

Pb(NO3)2

1000 ppm, 4 bar

≈ 60

43

≈ 68

CdCl2 HPEI modified GO&EDA framework membrane

5.01

Pb(NO3)2

1000 ppm, 1 bar

95.7 ± 0.7

NiCl2

96.0 ± 3.8

ZnCl2

97.4 ± 2.0

CdCl2

90.5 ± 0.1


This work

Nanometric Graphene Oxide Framework ... - ACS Publications

Recommend Documents