Impaired Performance of Pressure-Retarded Osmosis due to

Subscriber access provided by EPFL | Scientific Information and Libraries

Article

Impaired Performance of Pressure-Retarded Osmosis due to Irreversible Biofouling Edo Bar-Zeev, François Perreault, Anthony P. Straub, and Menachem Elimelech Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.5b03523 • Publication Date (Web): 01 Oct 2015 Downloaded from http://pubs.acs.org on October 9, 2015

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 28

Environmental Science & Technology

1 2 3 4

Impaired Performance of Pressure-Retarded

5

Osmosis due to Irreversible Biofouling

6 7 8


9

Revised: September 26, 2015

10 11 12

Edo Bar-Zeev1*, François Perreault2, Anthony P. Straub3, and Menachem

13

Elimelech3

14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

1

2

Department of Environmental Hydrology & Microbiology, Zuckerberg Institute for Water Research (ZIWR), Ben-Gurion University of the Negev, Israel.

School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ, 85287. 3

Department of Chemical and Environmental Engineering, Yale University, New Haven, Connecticut 06520-8286.

*Corresponding author; Address: Ben-Gurion University, Sade Boker, 8499000, Israel email: [email protected]

1 ACS Paragon Plus Environment


30

Page 2 of 28

ABSTRACT

31

Next-generation pressure-retarded osmosis (PRO) approaches aim to harness the energy potential 32

of streams with high salinity differences, such as wastewater effluent and seawater desalination 33

plant brine. In this study, we evaluated biofouling propensity in PRO. Bench-scale experiments 34

were carried out for 24 hours using a model wastewater effluent feed solution and simulated 35

seawater desalination brine pressurized to 24 bar. For biofouling tests, wastewater effluent was 36

inoculated with Pseudomonas aeruginosa and artificial seawater desalination plant brine was 37

seeded with Pseudoalteromonas atlantica. Our results indicate that biological growth in the feed 38

wastewater stream channel severely fouled both the membrane support layer and feed spacer, 39

resulting in ~50% water flux decline. We also observed an increase in the pumping pressure 40

required to force water through the spacer-filled feed channel, with pressure drop increasing 41

from 5 to 17 bar m-1 due to spacer blockage from the developing biofilm. Neither the water flux

42

decline nor the increased pressure drop in the feed channel could be reversed using a pressure43

aided osmotic backwash. In contrast, biofouling in the seawater brine draw channel was 44

negligible. Overall, the reduced performance due to water flux decline and increased pumping 45

energy requirements from spacer blockage highlight the serious challenges of using high fouling 46

potential feed sources in PRO, such as secondary wastewater effluent. We conclude that PRO 47

power generation using wastewater effluent and seawater desalination plant brine may become 48

possible only with rigorous pretreatment or new spacer and membrane designs. 49


Page 3 of 28

50


TOC Art

51 52 53

54 55



Page 4 of 28

56

INTRODUCTION

57

Global climate change due to increasing carbon dioxide emissions has led scientists to seek

58

new, sustainable energy sources with a low carbon footprint. Capitalizing on the chemical

59

potential released when mixing two streams of different salt concentrations, researchers have

60

recognized salinity gradient energy as a promising renewable energy source

61

potential to provide up to 13% of global electricity requirements 3.

62

1,2

with the

Salinity gradient energy can be extracted by several different processes, such as pressure4–6

, and capacitive mixing

9,10

retarded osmosis (PRO)

64

PRO is well studied

65

PRO technology generates power by using an osmotic pressure difference across a

66

semipermeable membrane to produce a flux of water from a low concentration feed solution to

67

a high concentration draw solution. The expanding volume of the draw solution can be

68

restricted to create hydraulic pressure and drive a hydro-turbine.

5,11–13

, reverse electrodialysis

7,8

63

. Of these,

and considered promising in terms of both efficiency and cost

14,15

.

69

Studies of PRO technology focused primarily on harnessing the energy released from

70

mixing fresh river water and saline seawater. However, it was recently shown that

71

river/seawater pairing might not be feasible due to the relatively low theoretical extractable

72

energy and other challenges such as pumping costs, non-ideal membranes, and pretreatment

73

requirements 16–18. Instead, alternative salinity combinations, such as seawater reverse osmosis

74

(SWRO) brine mixing with wastewater effluent or hypersaline Great Salt Lake water mixing

75

with river water, were suggested to yield greater specific energies of up to 2.26 kWh per cubic

76

meter of total source water used 19.

77

Mixing SWRO brine with wastewater effluent in PRO appears especially promising due to

78

several factors. First, the osmotic pressure difference (∆π) is twice that of the river water and

79

seawater solution pairing, hence doubling the maximum obtainable specific energy to a value

80

of 0.551 kWh m-3

81

reducing adverse impacts once the brine is discharged into the aquatic environment. Third,

82

various trace contaminates that pass the secondary wastewater treatment stage will be retained

83

by the PRO membrane and thus will avoid discharge to the environment.

19

. Second, the SWRO brine is diluted during the PRO process, thus

84

One possible hurdle associated with the use of wastewater effluent as a feed source is

85

membrane fouling, which is known to be detrimental to system performance and shortens the 4 ACS Paragon Plus Environment

Page 5 of 28


20,21

86

membrane life. It was recently demonstrated that fouling with inorganic matter

87

organic matter

88

water flux in bench-scale PRO setups. However, these studies have suggested that inorganic

89

and organic fouling were reversibly attached and could be partially cleaned by osmotic

90

backwashing, resulting in permeate water flux recovery of 50% to 63% 18,23.

18,22,23

24,25

, and other organic foulants

, natural

resulted in severe decline in permeate

91

In fully operational PRO systems, feed water will always comprise a combination of

92

microbial, organic, and inorganic foulants that will cause biofouling, a phenomenon that can

93

cripple membrane performance and lifetime

94

clusters of live and dead bacterial cells encased in self-produced extracellular polymeric

95

substances (EPS), primarily composed of polysaccharides and proteins

96

biofilms are notoriously resistant to removal by chemical and physical treatments due to the

97

protection provided by the EPS matrix 32. Despite the ubiquitous nature of biofouling and the

98

detrimental effect of biological fouling on membrane performance, no study has yet explored

99

the effects of microbial fouling on PRO feasibility.

26–28

. Biofilm is typically found as multilayered 29–31

. Once established,

100

In the present study, we explore the efficiency and practicability of PRO under biofouling

101

conditions using synthetic wastewater effluent and SWRO-brine pairing. Our results indicate

102

that biofouling was irreversibly attached to the feed channel spacers and throughout the

103

membrane support structure, severely impairing membrane performance. We further provide

104

new insights on the implications of pairing wastewater effluent with SWRO-brine for PRO

105

system operation and viability.

106 107

MATERIALS AND METHODS

108

Lab-Scale PRO Setup, Membranes, and Spacers. Experiments were conducted in a

109

bench-scale PRO setup (Figure S1), with commercial flat-sheet TFC forward osmosis

110

membranes, obtained from Hydration Technology Innovations (Albany, OR). The permeate

111

flow rate, temperature, and conductivity were automatically recorded in real time. Both the

112

feed and draw channels were 10.7 cm long and 3.6 cm wide, with a smaller 9.9 cm by 3.4 cm

113

section of the membrane exposed. The feed and draw channel heights were 0.5 and 1 mm,

114

respectively, to mimic a membrane module design. Figure 1 presents the spacers and



Page 6 of 28

115

membrane orientation in the cell: the feed channel was filled with two fabric spacers (Sp1 and

116

Sp2), each 0.225 mm thick (Hornwood, Gloversville, NY), while the draw channel was filled

117

with a ~1 mm thick plastic spacer (Sp3). The draw solution flow rate was maintained at 0.4 L

118

min-1 and the feed solution flow rate was initially set to 0.04 L min-1. Both initial flow rates

119

were set to allow for suitable hydrodynamic mixing without incurring excessive frictional

120

pressure losses along the membrane channel 33.

121

Figure 1.

122

Media and Bacterial Strains. Ten liters of artificial sterile wastewater secondary effluent

123

(hereafter WW) was used as feed solution (recipe modified from Glueckstern et al.

124

supplemented with 0.01 % D-glucose as additional carbon source. Pseudomonas aeruginosa

125

(ATCC® 27853TM) monoculture was used as a model biofilm strain in all WW biofouling

126

experiments.

127

34

)

Assuming a 50% recovery 35, SWRO brine solution (10 L) was made by doubling artificial 36

128

seawater (F2) concentration

. SWRO-brine was supplemented with 0.005 % D-glucose as

129

additional carbon source. Pseudoalteromonas atlantica (ATCC® BAA-1087) monoculture was

130

used as a model biofilm strain in the SWRO-brine (draw) biofouling runs. A detailed

131

description of the WW and SWRO-brine recipes as well as growing conditions of both model

132

bacteria strains is available in the Supporting Information.

133

Biofouling Experiments. The membrane and spacers were fitted into the PRO unit after

134

an extensive cleaning procedure that included circulating 10% bleach and 95% ethanol through

135

the system for one hour. The system was then stabilized at a zero permeate water flux with DI

136

water in both the feed and draw reservoirs (2 L and 10 L volumes, respectively). After

137

increasing the hydraulic pressure to 26.2 bar, the draw reservoir was supplemented with 5 M

138

NaCl to reach a final concentration of 50 mM, and salt rejection (> 96%) was evaluated to

139

verify membrane integrity. The draw solution was then replaced with 10 L of synthetic SWRO-

140

brine (excluding D-glucose) and the feed solution was replaced with 10 L of artificial WW

141

(excluding sodium citrate and D-glucose). The osmotic pressure difference (∆π) between the

142

solutions was 59.9 bar, calculated using OLI software. Once initial permeate water flux was

143

stabilized at 12 ± 2 L m-2 h-1, a baseline was collected for 24 hours to determine the decrease in


Page 7 of 28


144

water flux due solely to dilution of the draw solution and concentration of the feed reservoir as

145

water and salt diffuse across the membrane.

146

At the conclusion of the baseline run, both the feed and draw were replaced with 10 L of

147

fresh WW or SWRO brine solutions, while maintaining the hydraulic pressure (26.2 bar). Prior

148

to addition, late exponential stage P. aeruginosa and P. atlantica (OD600 of 0.6-0.8) were

149

centrifuged for 20 minutes at 4,000 rpm and 4 °C to remove the Lysogeny broth (LB) or

150

marine broth (MB). Cells were re-suspended in 10 mL of sterile WW (for P. aeruginosa) or

151

seawater (for P. atlantica) by vortexing for 30 seconds.

152

Two feed biofouling experiments were carried for ~24 hours by inoculating the feed WW

153

reservoir with P. aeruginosa to achieve an initial bacterial concentration of ~2.4 × 106 cells

154

mL-1. WW was then supplemented with 0.01% D-glucose. Two draw solution biofouling runs

155

were carried for ~24 hours by inoculating the reservoir with P. atlantica to achieve an initial

156

bacterial concentration of ~3.7 × 106 cells mL-1. D-glucose (0.005%) was then added to the

157

SWRO brine draw solution.

158

Pressure drop across the feed channel, permeate water flux, temperature, and conductivity

159

were measured continuously throughout the experiments. Feed and draw water sub-samples

160

(50 mL) were routinely collected (every 6 to 8 hours) to monitor pH, conductivity, and

161

bacterial concentration (via plate counts and OD600 measurements). The pH and the

162

temperature of the feed (7.3 ± 0.3, 25 ± 0.5 °C) and draw (8.13 ± 0.02, 25 ± 0.5 °C) reservoirs

163

were kept constant. Bacteria and nutrient concentrations used in this study were in the range

164

reported for secondary wastewater effluents from wastewater treatment facilities 34,37

165

PA-OBW Procedure. Two pressure-aided osmotic backwash (PA-OBW) experiments

166

were performed to test biofouling reversibility and cleaning efficiency (Figure S2). PA-OBW

167

experiments were initiated after the feed channel was inoculated and fouled with P. aeruginosa

168

for ~20 hours as described earlier. After fouling, PA-OBW was carried out by switching the

169

WW feed reservoir with SWRO brine (1.2 M NaCl) and the draw reservoir to DI water while

170

maintaining the hydraulic pressure, ∆P, of 26.2 bar (Figure S2).

171

hydraulic pressure differences both driving flow from the draw to the feed reservoir, the PA-

172

OBW produced a reverse water flux of 32 ± 4 L m-2 h-1 for 60 min. PA-OBW was terminated


With the osmotic and


Page 8 of 28

173

by replacing both the feed and draw reservoirs with the initial configuration (WW and SWRO

174

brine, respectively) for 1 hour.

175

Biofilm Characterization with CLSM. Biofilm biovolume and morphology were

176

determined using confocal laser scanning electron microscopy (CLSM). At the end of the

177

experimental run, a central section (1 cm × 1 cm) was cut from the membrane coupon as well

178

as the draw or feed spacers (Sp1 and Sp3, respectively). Sub-samples were then imaged and

179

analyzed in a custom made biofilm in vivo characterization chamber according to Bar-Zeev et

180

al., 2014 38. Further details of the CLSM sample preparation, staining, and image analysis are

181

available in the Supporting Information.

182

SEM and TEM Analysis. Scanning electron microscopy (SEM) and transmission

183

electron microscopy (TEM) were used to further understand the extent of biofilm formation.

184

Membrane and spacer samples were cut and fixed in a 1.5 mL eppendorf with a Karnovsky

185

fixative solution (2% paraformaldehyde, 2.5% glutaraldehyde in 0.2 M Sorenson’s buffer, pH

186

7.2) overnight at 4 °C. After fixation, samples were stained with osmium tetroxide and

187

sequentially dehydrated. Samples were then freeze fractured in liquid nitrogen, and imaged

188

with an Hitachi SU-70 SEM (Hitachi High Technologies America, Inc. Clarksburg, MD).

189

For TEM imaging, dried samples were infiltrated with epoxy resin (Embed 812, Electron

190

Microscopy Sciences, Hatfield, PA) at room temperature. Embedded samples were cut into

191

0.07 µm thick slices and imaged with an FEI Tecnai Osiris microscope, operating at an

192

acceleration voltage of 200 kV. Energy dispersive X-ray spectroscopy elemental mapping was

193

performed in scanning TEM mode (STEM-EDX) with 10 cycles compiled per image.

194

Bacterial Cell Counts. Membrane or spacers (Sp1, Sp2, and Sp3) sub-sections were re-

195

suspended in 1.5 mL eppendorf tubes with 1 mL sterile WW for the feed and F2 for the draw

196

samples

197

(AQUASONIC, PA, USA) for 5 minutes to remove cells from the surface, plated on LB or MB

198

agar plates, and incubated for 24 hours at 37 °C (feed samples) and 26 °C (draw samples).

199

Bacteria abundance was determined by counting colony forming units (CFU) and normalized

200

by the sample area.

to

maintain

experimental

conditions.

Samples

were

then

bath-sonicated

201

Protein Assay. Membrane or spacer sub-sections were re-suspended in 1.5 mL

202

eppendorf tubes with 1 mL 1X Lauber buffer (50 mM HEPES (pH 7.3), 100 mM NaCl, 10% 8 ACS Paragon Plus Environment

Page 9 of 28


203

sucrose, 0.1% 3-[(3-cholamidopropyl)-dimethylammonio]-1 propanesulfonate, and 10 mM

204

dithiothreitol). Samples were then probe sonicated on ice (two 30-second cycles) using an

205

ultra-cell disruptor (MISONIX Inc. NY, USA). Cell extracts were centrifuged at 12,000 rpm

206

for 10 minutes, and supernatant was collected for protein quantification (BCA Protein Assay

207

Kit, Thermo Scientific, Rockford, IL, USA). Protein concentrations were normalized according

208

to the sample area.

209

TOC Analysis. For total organic carbon (TOC) analysis, membrane or spacer sub-

210

sections were placed in acid-cleaned glass vials and resuspended in 20 mL DI water with 40

211

µL of 1 M HCl. Samples were probe-sonicated (two 30 second cycles) using an ultra-cell

212

disruptor (MISONIX Inc. NY, USA) to remove all the cells from the surface. Total organic

213

carbon was then analyzed using a TOC analyzer (TOC-V, Shimadzu Corp., Japan) and

214

normalized according to the sample area.

215 216

RESULTS AND DISCUSSION

217

Effect of Biofouling on PRO Performance. In our bench-scale PRO system, an initial

218

permeate water flux of 13.1 ± 1.8 L m-2 h-1 was achieved by the osmotic pressure difference

219

(∆π = 59.9 bar), which was opposed by the applied hydraulic pressure (∆P = 26.2 bar) in the

220

draw channel (Figure 1). Prior to each biofouling run, the membrane active layer properties,

221

structural parameter, and draw channel mass transfer coefficient were determined. All our

222

experiments showed consistent values for these parameters, with an average water permeability

223

coefficient, A, of 3.71 ± 0.14 L m-2 h-1 bar-1; salt permeability coefficient, B, of 1.20 ± 0.12 L

224

m-2 h-1; structural parameter, S, of 627 ± 124 µm; and draw mass transfer coefficient, k, of 11.3

225

± 1.2 µm s-1. A detailed description of the methodology used to determine these values is

226

available in the Supporting Information. Using these initial operating parameters, the

227

performance of PRO under biofouling conditions was evaluated over 24 hours by

228

independently inoculating the feed or draw streams with either P. aeruginosa (WW feed) or P.

229

atlantica (SWRO-brine draw), respectively.

230

Inoculating the WW feed solution with P. aeruginosa (1.6 to 3.4 ×106 cells mL-1) induced a

231

permeate water flux decline 3 hours after the addition of bacteria, which corresponded to a 9 ACS Paragon Plus Environment


Page 10 of 28

232

normalized cumulative permeate volume of ~36 L m-2 (Figure 2A). A sharp decrease in

233

permeate water flux (~50%) occurred for 8 hours after the initial fouling event. In the final 11

234

hours of the experiment, permeate water flux remained relatively constant, and was 6 ± 0.9 L

235

m-2 h-1 at the end of the ~24 hour run (Figure 2A). Throughout the experiment, bacterial

236

concentration in the WW reservoir increased, reaching a final value of ~1.9 × 107 cells mL-1 Figure 2.

237 238

We attribute the permeate water flux decline during the biofouling experiments on the feed

239

side of the membrane to three factors: (i) biofilm development within the feed channel, across

240

the spacers (Sp1 and Sp2), and on the membrane support surface, resulting in cake-enhanced

241

concentration polarization; (ii) internal biofilm formation within the membrane support layer,

242

leading to increased internal concentration polarization, and (iii) biofilm growth at the active

243

layer−support interface, which increases the hydraulic resistance of the membrane 23.

244

In addition to the reduction in permeate water flux, biofouling in the feed spacers resulted

245

in decreased cross flow rate and increased pressure loss along the feed channel (Table S1). In

246

all experiments, the cross flow rate through the feed channel spacers (Sp1 and Sp2) was

247

initially set at 0.038 ± 0.005 L min-1 and decreased to 0.013 ± 0.004 L min-1 at the end of the

248

biofouling experiment. Normalized pressure drop along the feed channel amounted to 6.4 ± 0.8

249

bar m-1 at the start of the biofouling experiment and increasing by 84% over the course of the

250

run, resulting in a normalized feed channel pressure loss of 11.7 ± 0.2 bar m-1. This dramatic

251

increase in the feed channel pressure loss is due to the accumulation of a dense biofilm within

252

the fabric feed spacer.

253

Unlike results in the feed channel, inoculating the SWRO brine draw solution with P.

254

atlantica (1.4 to 6.0 ×106 cells mL-1) did not lead to any change of permeate water flux (Figure

255

2B). Moreover, no pressure drop was measured along the draw channel and the flow rate

256

through the draw channel spacer (Sp3) remained constant at 0.4 L min-1. We suggest that the

257

convective permeate water flux (from the support surface to the active layer side) transports

258

foulants and bacteria away from the membrane surface, thereby limiting deposition and

259

hindering the development of biofouling. Additionally, we surmise that the hypersaline (1.2 M

260

NaCl) environment restricted both bacterial growth in the SWRO-brine stream, which was

261

reduced to 3.4 ± 1.4 × 104 cells mL-1at the end of the run, and biofilm development on the 10 ACS Paragon Plus Environment

Page 11 of 28


262

active layer surface. However, it is possible that longer exposure of the membrane surface to

263

natural SWRO-brine will result in biofilm development by various halophilic microorganisms,

264

such as protobacteria and archaea 39,40.

265

Biofilm Development and Characteristics in PRO. Overall, PRO experimental

266

measurements indicated extensive biofouling of the fabric spacers and the TFC membrane

267

structure on the WW feed side. However, minimal changes were observed after fouling along

268

the spacer (Sp3) or membrane active layer on the SWRO brine draw side. We further probed

269

biofilm formation in the system using microscopic imaging and biofilm characterization

270

techniques.

271

In the fabric spacers that filled the feed channel, biofilm was found, by SEM imaging, to be

272

clogging openings between the spacer fibers (Figures 3A, B). CLSM imaging confirmed that

273

these large clusters—hundreds of micrometers in length—were composed of both live cells

274

and EPS-polysaccharides (Figure 4A).

275

Biofilm was also observed on the membrane support surface, which faces the wastewater

276

feed stream. SEM (Figures 3C, 3D) and CLSM (Figure 4B) images showed biofilm on the

277

surface of the membrane support, primarily around the openings of the exposed polyester

278

mesh. CLSM image analysis indicated that the average biofilm thickness was 60 ± 7.1 µm,

279

while biovolume was 35.7 ± 6.8 µm3 µm-2 (biovolume was calculated as the sum of live/dead

280

cells and EPS-polysaccharides). However, large biofilm structures (a few hundreds of

281

micrometers thick) were often captured above the membrane support surface (Figure 4B),

282

indicating that biofilm was interlocked with the overlying spacer (Sp1).

283

The permeating flow of water into the support layer led to biofilm formation inside open

284

voids on the surface of the membrane. Polysaccharides were found along the membrane

285

support layer surface as well as in funnel-shaped structures within openings in the support

286

layer that formed around the polyester mesh during fabrication (Figure S3). In addition, dense

287

biofilm was observed attached in voids beneath the support surface (Figure 3E).

288

Figure 3

289

Figure 4

290

Growth of cells was also observed far within the fingerlike polysulfone support layer

291

(Figure 3F). TEM images of the support layer cross-section, complemented with EDX 11 ACS Paragon Plus Environment


292

elemental mapping, confirmed the presence of bacteria in voids throughout the support

293

structure (Figure 5). Cells were typically found as small clusters within the fingerlike

294

polysulfone matrix, as close as ~200 nm from the active layer (Figure 5). We conclude that the

295

development of biofilm within the membrane structure resulted from continuous flow of WW

296

feed through breaches in the support surface facilitating nutrient (e.g. polysaccharides)

297

transport as well as bacterial deposition and growth.

298

Figure 5

Page 12 of 28

299

Biofilm biomass developed on the membrane and spacers was also analyzed by

300

determining cell, TOC, and protein coverage as well as biofilm biovolume (Figure 6).

301

Throughout feed spacers Sp1 and Sp2, cell coverages were respectively four- and two-fold

302

higher than on the TFC membrane (Figure 6A). The extensive cell proliferation in the spacers

303

may have resulted from favorable growth conditions, such as high nutrient levels and a well-

304

oxygenated environment

305

biomass, measured as TOC and proteins, between the feed spacers and membrane structure

306

(Figures 6B and 6C). We surmise that the discrepancy between the spacers and membrane with

307

respect to cell coverage (higher coverage on the spacers) and biofilm biomass (proteins and

308

TOC similar in both spacers and membrane) is attributable to the yield in extracting the

309

biomass for analysis. Specifically, cells remained attached within the membrane support

310

structure, while TOC and proteins, for which a stronger cell-disrupting sonication treatment

311

was used, were completely removed from the membrane.

312

29,41,42

. However, no significant difference was detected in biofilm

Figure 6

313

In contrast to the feed channel, no viable cells were found attached to the draw spacer (Sp3)

314

or the surface of the membrane active layer (Figures 6A, 6D, and S4). The absence of cells on

315

the draw spacer and active layer is attributed to the outward convective permeate water flux

316

and to induced osmotic stress by the hypersaline SWRO-brine solution, hindering P. atlantica

317

growth. In addition, the adsorbed biomass (TOC and proteins) extracted from the draw spacer

318

and membrane active layer was ~80% lower than the amount found in the feed channel

319

(Figures 6B and 6C). The lower fouling propensity in the draw channel is in agreement with

320

previous studies of organic fouling propensity in PRO18.


Page 13 of 28


321

Membrane Orientation is the Culprit for Severe Biofouling in PRO. The

322

substantially greater biofilm accumulation on the feed side, as evidenced by the marked water

323

flux decline and biofilm accumulation, is a direct consequence of the membrane orientation in

324

PRO. Since the porous support layer faces the WW feed stream with high organic and bacteria

325

concentrations (Figure 1), the flow of WW in the feed channel continuously carries dissolved

326

nutrients and bacterial cells to the spacers, membrane surface, and membrane support layer.

327

We suggest that proliferation of cell clusters and EPS secretions within the membrane support

328

structure may lead to irreversible biofilm attachment. Moreover, it is highly likely that

329

continuous production of EPS will gradually accumulate at the membrane support−active layer

330

interface and further hinder permeate flux, as previously suggested for organic fouling

331

Therefore, we estimate that, in PRO systems using treated WW effluent as feed, biofouling

332

proliferation inside the support structure will cause detrimental effects to membrane

333

performance over short (days to weeks) time scales.

22

.

334

Minimizing cell attachment on the TFC membrane support surface and preventing bacteria

335

from colonizing the polysulfone structure was attempted by reversing the membrane

336

orientation, with the active layer facing the WW feed solution rather than the SWRO brine.

337

However, membrane failure was recorded at 6.9 bar, indicated by a rapid increase in reverse

338

salt flux and water flux, most probably due to the delamination of the active layer.

339

Is Biofouling Reversible? Pressure-aided osmotic backwashing (PA-OBW) was carried

340

out to physically disperse biofouling from the membrane and spacer by reversing the permeate

341

water flow direction

342

and permeate water flux was reduced by 50% (Figure 7A). PA-OBW was initiated through a

343

reversal of the osmotic driving force by exchanging the draw solution with deionized water and

344

the feed solution with model SWRO brine, while maintaining a constant ∆P of 26.2 bar (Figure

345

S2). Reverse water flux during PA-OBW (from the active layer side to the porous support side)

346

was 32 ± 4 L m-2 h-1 and remained constant for 60 min (Figure 7B).

347

18,23,43,44

. We applied PA-OBW after the PRO membrane was biofouled

Figure 7

348

The intensive PA-OBW resulted in low (12%) permeate water flux recovery (Figure 7A)

349

and no significant change in the feed channel pressure drop (Table S1). Overall, the surface of

350

the spacers (Figure S5) and membrane support were cleaner following the PA-OBW as 13 ACS Paragon Plus Environment


Page 14 of 28

351

indicated by a reduction in cell, TOC, protein coverage, and total biovolume (Table 1). In

352

addition to a reduction in cell biovolume after PA-OBW, there were 36% more dead cells than

353

live cells on the membrane surface and the feed (Sp1) spacer (Table 1, Figure S6). We attribute

354

the high dead to live cell ratio to an osmotic shock to bacteria induced by the highly saline

355

solution 45. Table 1

356 357

However, PA-OBW was only effective for the outermost layers of the biofilm, and no

358

change was found in the distribution of biofilm within the support layer after the 60-minute

359

PA-OBW (Figure S7). Biofilm was irreversibly attached within the polysulfone support matrix,

360

and sporadic bacterial cells remained in the finger-like structure (Figures S7 E and F). In

361

addition to bacterial cells, organic fouling, such as EPS, also remained trapped within the

362

support layer, even after rigorous backwash (Table 1).

363

It has been recently reported that organic fouling in PRO resulted in 30% to 46% permeate

364

flux decline, while various backwashing procedures recovered significant percentage (50-63%)

365

of the water flux, suggesting that organic fouling was in part reversible

366

biofouling can be partially removed from the feed spacer and the membrane support surface by

367

PA-OBW. However, biofilm and EPS as well as bacterial cells remain irreversibly confined

368

within the polysulfone support layer. We attribute the limited recovery of permeate water flux

369

(12%) following PA-OBW to the detrimental effects of biofilm that remained irreversibly

370

attached to the membrane support structure.

18,23

. We show that

371

Surprisingly, although feed spacers appeared cleaner following PA-OBW (Figure S5),

372

pressure drop along the feed channel was not significantly different (Table S1). We suggest

373

that the recorded pressure drop along the feed channel may be the result of biofilm residues

374

blocking micro-channels within the spacers, thus retaining the overall hydraulic resistance.

375

Feasibility of PRO using Wastewater and SWRO-Brine. Compared to river and

376

seawater pairing in PRO, harnessing the energy of mixing WW and SWRO-brine for power

377

production has been conceived as more viable due to the increased extractable energy 46–48. The

378

promise of this system configuration was further emphasized by the high reversibility of

379

organic and inorganic fouling observed in PRO, with 50% to 63% permeate flux recovery

380

following OBW 18,23.


Page 15 of 28


381

Our present work conducted with model WW, simulated SWRO-brine, and live bacteria

382

indicates that significant irreversible biofouling along the feed channel and within the TFC

383

membrane support layer causes (i) 50% permeate water flux decline and (ii) severe pressure

384

drop along the feed channel (up to 17 bar m-1). A rigorous PA-OBW cleaning procedure was

385

only able to recover 12% of the permeate water flux and could not lessen the feed channel

386

pressure drop.

387

The amount of energy extractable from a PRO module is proportional to the applied

388

hydraulic pressure, ∆P, multiplied by the water flow rate across the membrane 19. Hence, the

389

50% reduction in water flux of the fouled membrane will cause a correspondingly reduced

390

power output and require a PRO system with significantly larger membrane area 49. Both these

391

effects would significantly reduce the feasibility of the process.

392

The effect of biofouling will extend beyond reducing the water flux across the membrane

393

module; it will also necessitate a higher energy input into the system. At the end of the feed

394

fouling tests, we observed a dramatic increase in pressure loss across the feed channel,

395

reaching values up to 17 bar m-1 as compared to 6.4 bar m-1 at the beginning of the

396

run. Pressurizing the feed solution before it enters the membrane module represents a

397

substantial energy cost

398

pumped in at 17 bar, while the energy extractable from the system would come from the flow

399

rate permeating across the membrane entering at turbine at 26.2 bar. Since the flow rate

400

entering the feed side of the module must always exceed the flow rate across the membrane.

401

into the turbine, we can estimate that a fouled meter-long module would consume at least 65%

402

of the energy that is extractable from the system pumping the feed solution.

24

. For a meter-long module, the feed flow rate would need to be

403

Our results stress that using a high fouling potential solution, such as secondary WW

404

effluent, as a feed solution for PRO will cripple the performance of the process in short time

405

periods, due to irreversible biofilm formation. These results also extend to other natural waters,

406

such as river water, since dissolved organic matter and bacteria are ubiquitous in natural

407

waters. However, for other feedwaters, the time scale for biofilm development hindering PRO

408

performance will depend on the source of the water and the level of pretreatment.

409

To minimize biofouling development in the feed channel, rigorous feedwater pretreatment

410

will be necessary, requiring both more energy input into the system and enhanced chemical

411

use. This energy input for pretreatment will be significant, and a more extensive analysis is 15 ACS Paragon Plus Environment


Page 16 of 28

412

necessary to determine the net energy output of a system accounting for the cost of

413

pretreatment. Alternatively, it may also be possible to reduce biofouling using new and

414

innovative membrane and spacer designs that reduce bacteria adhesion with the membrane

415

support or allow for improved cleaning efficacy.

416 417

SUPPORTING INFORMATION AVAILABLE

418

Detailed descriptions of the membrane characterization method; media and bacteria that were

419

used; cleaning procedures of the PRO setup; detailed description of CLSM, SEM and TEM

420

preparation; PRO experimental setup (Figure S1); pressure aided-osmotic backwash procedure

421

(Figure S2); pressure drop and WW flow rate through the feed channel (Table S1); CLSM

422

orthogonal views of biofouling on the membrane support (Figure S3); CLSM 3D images of

423

biofilm on the active layer and draw spacer (Figure S4); feed channel spacers before and after

424

OBW (Figure S5); CLSM 3D images (Figure S6) and SEM micrographs (Figure S7) of biofilm

425

that remained attached following OBW. This material is available free of charge via the

426

Internet at http://pubs.acs.org.

427 428

Notes

429

The authors declare no competing financial interest.

430 431

ACKNOWLEDGMENTS

432

We acknowledge support received from the National Science Foundation under Award Number 433

CBET 1232619. We would also like to thank Dr. Ming Xie and Evyatar Shaulsky for special 434

technical assistance. A.P.S. acknowledges support from the National Science Foundation 435

Graduate Research Fellowship (DGE-1122492). F.P. acknowledges financial support from the 436

Natural Sciences and Engineering Research Council of Canada postdoctoral fellowship. E.B-Z 437

acknowledges financial support of the postdoctoral fellowship provided by the United States438

Israel Binational Agricultural Research and Development (FI-474-12) Fund. Facilities used were 439

supported by the Yale Institute of Nanoscale and Quantum Engineering (YINQE) and NSF 440

MRSEC DMR 1119826. 16 ACS Paragon Plus Environment

Page 17 of 28


441

REFERENCES

442 443

(1)

Logan, B. E.; Elimelech, M. Membrane-based processes for sustainable power generation using water. Nature 2012, 488, 313–319.

444 445

(2)

Pattle, R. E. Production of electric power by mixing fresh and salt water in the hydroelectric pile. Nature, 1954, 174, 660–660.

446 447

(3)

La Mantia, F.; Pasta, M.; Deshazer, H. D.; Logan, B. E.; Cui, Y. Water salinity difference. Nano 2011, 11, 1810–1813.

448 449

(4)

Loeb, S.; Van Hessen, F.; Shahaf, D. Production of energy from concentrated brines by pressure-retarded osmosis. J. Memb. Sci. 1976, 1, 249–269.

450 451

(5)

Helfer, F.; Lemckert, C.; Anissimov, Y. G. Osmotic power with pressure retarded osmosis: theory, performance and trends - A review. J. Memb. Sci. 2014, 453, 337–358.

452 453

(6)

Achilli, A.; Childress, A. E. Pressure retarded osmosis: From the vision of Sidney Loeb to the first prototype installation - Review. Desalination 2010, 261, 205–211.

454 455 456

(7)

Post, J. W.; Hamelers, H. V. M.; Buisman, C. J. N. Energy recovery from controlled mixing salt and fresh water with a reverse electrodialysis system. Environ. Sci. Technol. 2008, 42, 5785–5790.

457 458 459

(8)

Daniilidis, A.; Vermaas, D. a.; Herber, R.; Nijmeijer, K. Experimentally obtainable energy from mixing river water, seawater or brines with reverse electrodialysis. Renew. Energy 2014, 64, 123–131.

460 461 462

(9)

Rica, R.; Ziano, R.; Salerno, D.; Mantegazza, F.; van Roij, R.; Brogioli, D. Capacitive mixing for harvesting the free energy of solutions at different concentrations. Entropy 2013, 15, 1388–1407.

463 464 465

(10)

Hatzell, K. B.; Beidaghi, M.; Campos, J. W.; Dennison, C. R.; Kumbur, E. C.; Gogotsi, Y. A high performance pseudocapacitive suspension electrode for the electrochemical flow capacitor. Electrochim. Acta 2013, 111, 888–897.

466 467 468

(11)

Yip, N. Y.; Elimelech, M. Performance limiting effects in power generation from salinity gradients by pressure retarded osmosis. Environ. Sci. Technol. 2011, 45, 10273– 10282.

469 470

(12)

Thorsen, T.; Holt, T. The potential for power production from salinity gradients by pressure retarded osmosis. J. Memb. Sci. 2009, 335, 103–110.

471 472 473

(13)

Chou, S.; Wang, R.; Fane, A. G. Robust and High performance hollow fiber membranes for energy harvesting from salinity gradients by pressure retarded osmosis. J. Memb. Sci. 2013, 448, 44–54. 17 ACS Paragon Plus Environment


Page 18 of 28

474 475 476

(14)

Yip, N. Y.; Elimelech, M. Comparison of energy efficiency and power density in pressure retarded osmosis and reverse electrodialysis. Environ. Sci. Technol. 2014, 48, 11002–11012.

477 478

(15)

Ramon, G. Z.; Feinberg, B. J.; Hoek, E. M. V. Membrane-based production of salinitygradient power. Energy Environ. Sci. 2011, 4, 4423.

479 480 481

(16)

Straub, A. P.; Lin, S.; Elimelech, M. Module-scale analysis of pressure retarded osmosis: performance limitations and implications for full-scale operation. Environ. Sci. Technol. 2014, 48, 12435–12444.

482 483

(17)

http://www.statkraft.com/media/news/News-archive/2013/Statkraft-halts-osmoticpower-investments.

484 485 486

(18)

Kim, D. I.; Kim, J.; Shon, H. K.; Hong, S. Pressure retarded osmosis (PRO) for integrating seawater desalination and wastewater reclamation: Energy consumption and fouling. J. Memb. Sci. 2015, 483, 34–41.

487 488

(19)

Lin, S.; Straub, A.; Elimelech, M. Thermodynamic limits of extractable energy by pressure retarded osmosis. Energy Environ. Sci. 2014, 7, 2706–2714.

489 490

(20)

Zhang, M.; Hou, D.; She, Q.; Tang, C. Y. Gypsum scaling in pressure retarded osmosis: Experiments, mechanisms and implications. Water Res. 2014, 48, 387–395.

491 492 493

(21)

Chen, S. C.; Wan, C. F.; Chung, T.-S. Enhanced fouling by inorganic and organic foulants on pressure retarded osmosis (PRO) hollow fiber membranes under high pressures. J. Memb. Sci. 2015, 479, 190–203.

494 495

(22)

Thelin, W. R.; Sivertsen, E.; Holt, T.; Brekke, G. Natural organic matter fouling in pressure retarded osmosis. J. Memb. Sci. 2013, 438, 46–56.

496 497 498

(23)

Yip, N. Y.; Elimelech, M. Influence of natural organic matter fouling and osmotic backwash on pressure retarded osmosis energy production from natural salinity gradients. Environ. Sci. Technol. 2013, 47, 12607–12616.

499 500 501

(24)

She, Q.; Hou, D.; Liu, J.; Tan, K. H.; Tang, C. Y. Effect of feed spacer induced membrane deformation on the performance of pressure retarded osmosis (PRO): Implications for PRO process operation. J. Memb. Sci. 2013, 445, 170–182.

502 503 504

(25)

Li, X.; Cai, T.; Chung, T.-S. Anti-fouling behavior of hyperbranched polyglycerolgrafted poly(ether sulfone) hollow fiber membranes for osmotic power generation. Environ. Sci. Technol. 2014, 48, 9898–9907.

505 506 507

(26)

Matin, A.; Khan, Z.; Zaidi, S. M. J.; Boyce, M. C. Biofouling in reverse osmosis membranes for seawater desalination: Phenomena and prevention. Desalination 2011, 281, 1–16. 18 ACS Paragon Plus Environment

Page 19 of 28


508 509

(27)

Mansouri, J.; Harrisson, S.; Chen, V. Strategies for controlling biofouling in membrane filtration systems: challenges and opportunities. J. Mater. Chem. 2010, 20, 4567.

510 511 512

(28)

Bar-Zeev, E.; Passow, U.; Romero-Vargas Castrillón, S.; Elimelech, M. Transparent exopolymer particles: from aquatic environments and engineered systems to membrane biofouling. Environ. Sci. Technol. 2015, 49, 691–707.

513 514 515

(29)

De Beer, D.; Stoodley, P. The Prokaryotes. In The Prokaryotes; Rosenberg, E.; DeLong, E. F.; Lory, S.; Stackebrandt, E.; Thompson, F., Eds.; Springer Berlin Heidelberg: Berlin, Heidelberg, 2013; pp. 343–373.

516 517 518

(30)

Herzberg, M.; Kang, S.; Elimelech, M. Role of extracellular polymeric substances (EPS) in biofouling of reverse osmosis membranes. Environ. Sci. Technol. 2009, 43, 4393– 4398.

519 520

(31)

Herzberg, M.; Elimelech, M. Biofouling of reverse osmosis membranes: Role of biofilm-enhanced osmotic pressure. J. Memb. Sci. 2007, 295, 11–20.

521 522

(32)

Flemming, H.-C.; Wingender, J. The biofilm matrix. Nat. Rev. Microbiol. 2010, 8, 623– 633.

523 524 525

(33)

Straub, A. P.; Osuji, C. O.; Cath, T. Y.; Elimelech, M. Selectivity and mass transfer limitations in pressure-retarded osmosis at high concentrations and increased operating pressures. Environ. Sci. Technol. 2015, In Press. DOI: 10.1021/acs.est.5b01317

526 527

(34)

Glueckstern, P.; Priel, M.; Gelman, E.; Perlov, N. Wastewater desalination in Israel. Desalination 2008, 222, 151–164.

528 529

(35)

Elimelech, M.; Phillip, W. A. The future of seawater and the environment. Science. 2011, 333, 712–717.

530 531

(36)

Guillard, R. R.; Ryther, J. H. Studies of marine planktonic diatoms: Cyclotella nana hustedt, and Detonula confervacea (cleve) gran. Can. J. Microbiol. 1962, 8, 229–239.

532 533 534

(37)

Gombos, E.; Felföldi, T.; Barkács, K.; Vértes, C.; Vajna, B.; Záray, G. Ferrate treatment for inactivation of bacterial community in municipal secondary effluent. Bioresour. Technol. 2012, 107, 116–121.

535 536 537

(38)

Bar-zeev, E.; Zodrow, K. R.; Kwan, S. E.; Elimelech, M. The importance of microscopic characterization of membrane biofilms in an unconfined environment. Desalination 2014, 348, 8–15.

538 539 540

(39)

Khan, M. T.; Manes, C. D. O.; Aubry, C.; Gutierrez, L.; Croue, J. P. Kinetic study of seawater reverse osmosis membrane fouling. Environ. Sci. Technol. 2013, 47, 10884– 10894.



Page 20 of 28

541 542 543

(40)

Zodrow, K. R.; Bar-Zeev, E.; Giannetto, M. J.; Elimelech, M. Biofouling and microbial communities in membrane distillation and reverse osmosis. Environ. Sci. Technol. 2014, 48, 13155–13164.

544 545

(41)

Donlan, R. M. Biofilms: Microbial life on surfaces. Emerg. Infect. Dis. 2002, 8, 881– 890.

546 547

(42)

Stoodley, P.; Sauer, K.; Davies, D. G.; Costerton, J. W. Biofilms as complex differentiated communities. Annu. Rev. Microbiol. 2002, 56, 187–209.

548 549

(43)

Qin, J.; Liberman, B.; Kekre, K. A. Direct osmosis for reverse osmosis fouling control: principles, applications and recent developments. Open Chem. Eng. J. 2009, 3, 8–16.

550 551 552

(44)

Bar-Zeev, E.; Elimelech, M. Reverse osmosis biofilm dispersal by osmotic backflushing: cleaning via substratum perforation. Environ. Sci. Technol. Lett 2014, 2, 162–166.

553 554 555

(45)

Katebian, L.; Jiang, S. C. Marine bacterial biofilm formation and its responses to periodic hyperosmotic stress on a flat sheet membrane for seawater desalination pretreatment. J. Memb. Sci. 2013, 425-426, 182–189.

556 557 558

(46)

Achilli, A.; Prante, J. L.; Hancock, N. T.; Maxwell, E. B.; Childress, A. E. Experimental results from RO-PRO: A next generation system for low-energy desalination. Environ. Sci. Technol. 2014, 48, 6437–6443.

559 560 561

(47)

Song, X.; Liu, Z.; Sun, D. D. Energy recovery from concentrated seawater brine by thinfilm nanofiber composite pressure retarded osmosis membranes with high power density. Energy Environ. Sci. 2013, 6, 1199.

562 563 564

(48)

Prante, J. L.; Ruskowitz, J. a.; Childress, A. E.; Achilli, A. RO-PRO desalination: An integrated low-energy approach to seawater desalination. Appl. Energy 2014, 120, 104– 114.

565 566

(49)

Banchik, L. D.; Sharqawy, M. H.; Lienhard, J. H. Limits of power production due to finite membrane area in pressure retarded osmosis. J. Memb. Sci. 2014, 468, 81–89.

567 568


Page 21 of 28

569 570


TABLE 1. Biofilm characteristics and percent cleaning efficiency for each biofilm component following 60 min of pressure aided osmotic backwash.

571 Units Cell coverage

TOC

Protein

Membrane 2

“Dead cells”

EPS Polysaccharides Total biovolume

Spacer

(Sp1)

(Sp2)

Cell / µm

0.7 ±

Impaired Performance of Pressure-Retarded Osmosis due to

Recommend Documents