Activated carbon nanofiber nonwoven for removal of emulsified oil from water

Journal Pre-proof Activated carbon nanofiber nonwoven for removal of emulsified oil from water Basma I. Waisi, Jason T. Arena, Nieck E. Benes, Arian Nijmeijer, Jeffrey R. McCutcheon PII:

S1387-1811(19)30825-X

DOI:

https://doi.org/10.1016/j.micromeso.2019.109966

Reference:

MICMAT 109966

To appear in:

Microporous and Mesoporous Materials

Received Date: 23 October 2019 Revised Date:

6 December 2019

Accepted Date: 17 December 2019

Please cite this article as: B.I. Waisi, J.T. Arena, N.E. Benes, A. Nijmeijer, J.R. McCutcheon, Activated carbon nanofiber nonwoven for removal of emulsified oil from water, Microporous and Mesoporous Materials (2020), doi: https://doi.org/10.1016/j.micromeso.2019.109966. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.

Author Contributions section Basma I. Waisi: Validation, Methodology, Writing - Original Draft, Formal analysis, Investigation Jason T. Arena: Resources Nieck E. Benes: Supervision, Formal analysis, Writing - Review & Editing Arian Nijmeijer: Supervision, formal analysis, Writing - Review & Editing Jeffrey R. McCutcheon: Project administration, Supervision, Conceptualization, Writing- Reviewing and Editing,

1

Activated Carbon Nanofiber Nonwoven for Removal of Emulsified Oil from Water

2

Basma I. Waisi1,2, Jason T. Arena3, Nieck E. Benes1, Arian Nijmeijer1, and Jeffrey R. McCutcheon3*

3 4 5

1.

6

Inorganic membrane, Department of science and technology. MESA+ Institute for nanotechnology, University of Twente. Enschede, The Netherlands

7

2.

8

Department of Chemical Engineering, College of Engineering, University of Baghdad, Baghdad, Iraq

9

3.

10

Department of Chemical &Biomolecular Engineering. Center for Environmental Sciences and Engineering, University of Connecticut, Storrs, CT

11 12 13 14 15

Keywords: Activated carbon, nonwoven, emulsion, batch, flow-through, and adsorption

16 17 18

1

19

Abstract

20

Activated carbon nanofiber nonwoven (ACNFN) has the potential for being high capacity and

21

throughput sorbent for a variety of contaminants in water due to high porosity and permeability,

22

respectively. In this work, we explore the removal of emulsified oil from water using ACNFNs in both

23

a batch and flow-through mode. The ACNFNs were prepared by pyrolysis of electrospun

24

polyacrylonitrile and characterized for specific surface area, hydrophobicity, and mechanical strength.

25

The ACNFNs exhibit a removal efficiency of up to 95% in batch mode and outperform our control tests

26

with non-activated carbon nanofiber nonwoven and a commercial granular activated carbon.

27

Adsorption tests in a flow-through mode demonstrated characteristic breakthrough behavior with

28

subsequent clogging. Removal efficiency increased with the thickness of the mat and with prolonged

29

operation due to ripening.

30

2

31

1. Introduction

32

Oily wastewater is a common discharge from petroleum refineries [1,2], manufacturing plants [3],

33

mining, and metal-processing industries [4]. These waters threaten terrestrial and aquatic ecosystems

34

due to their toxicity and low rate of oil degradation [5,6]. These wastewaters contain oil in three

35

general classifications based on the size of the oil droplet (dp): free-floating oil (dp > 150 µm),

36

dispersed oil (150 µm > dp > 20 µm) and emulsions (dp < 20 µm) [7,8]. Free-floating and dispersed oil

37

can be removed with relative ease using conventional methods including gravity separation, skimming,

38

air flotation, centrifugation, and coagulation and flocculation [9–11]. These methods, however, struggle

39

to effectively remove stable emulsions [12,13].

40

An adsorption is a common approach to removing emulsions from wastewater [8,14]. A variety of

41

sorbent materials have been considered for emulsified oil removal, including biomaterials [15], coal

42

[16], barley straw [17], silica aerogel [18] bentonite [19], meshed corncobs [20], and carbon [21].

43

Commercially available activated carbon, in powder or granular form, is widely used as an adsorbent to

44

remove a range of pollutants, including oil [22]. However, for oil, such activated carbons have shown

45

slow adsorption kinetics and limited oil removal capacity [8,23]. This can be explained by the large

46

diameter of oil droplets compared to the diameter of pores of activated carbon, resulting in blocking of

47

the pores and preventing further adsorption of oil droplets [24]. Favorable adsorption behavior can be

48

achieved by optimizing the porous structure [25] and ensuring surface hydrophobicity [18], though

49

increasing pore size will reduce specific surface area.

50

Activated carbon nanofiber nonwoven (ACNFN) is a form of activated carbon that is attracting

51

increasing attention as a sorbent due to its highly accessible specific surface area combined with its

52

nonwoven form factor 3

[26]. ACNFNs can be fabricated through the pyrolysis of electrospun

53

nanofibers. The use of electrospinning offers a degree of customization as mat and fiber properties are

54

readily tailored to the end-use [27]. Polyacrylonitrile (PAN) is one of the most widely used precursor

55

materials. Also used to produce some conventional carbon fibers, PAN pyrolysis produces a high

56

carbon yield and is known for its thermal stability [28].

57

A series of thermal treatments convert a PAN nanofiber mat into ACNFN [28–30]. The general

58

sequence includes stabilization

(usually in air) [31], carbonization (inert atmosphere) [32], and

59

activation (using oxidizing agent or water) [33]. The produced activated carbon nanofiber (ACNFN)

60

has a high specific surface area and hierarchical porous structure [34,35]. The mat has a high macro-

61

porosity which allows convective flow while the small fibers are mesoporous, providing high surface

62

area. The small fibers combined with the accessibility of the fiber surface to convective flow reduce

63

diffusion path lengths and improve the rate and capacity of adsorption [28]

64

ACNFN has been considered in gas phase adsorption. The adsorption of toluene [36], gaseous

65

formaldehyde [37], other volatile organic compounds (VOCs) [29], and the treatment of toxic industrial

66

gas [38] have all been demonstrated. The rigidity and brittleness of ACNFNs [39,40] are common

67

drawbacks of the material and are the reason to be cautious when considering the material for aqueous

68

systems.

69

Recently, we reported the optimized fabrication conditions to obtain ACNFN with good

70

mechanical strength, a high specific surface area, and a high hydrophobicity in nonwoven structures.

71

This is aided by the fact that ACNFNs (and nanofiber nonwovens in general) have exceptional

72

hydraulic permeability that enable their use in applications that require integrated or flexible sorbents

73

[35]. Here, we describe the emulsified oil adsorption capacity and rate of such ACNFNs using a batch

74

test. The effects of the initial oil concentration and the agitating speed on adsorption performance are 4

75

studied in detail. A comparison is made between the performance of the ACNFNs, CNFNs and a

76

commercial granular activated carbon (GAC) adsorbent. Our belief was that CNFNs had better

77

assessable surfaces than GAC and that assessability ACNF has accessible surface areas that match PAC

78

and the permeability of GAC. Then, ACNFN mats were applied in a flow-system for oil removal from

79

water. The results showed that the oil removed by adsorption on and in the fibers itself and size

80

exclusion by the open spaces between the fibers.

81

2 Materials and Methods

82

2.1 Materials

83

Petroleum oil (18% aromatics basis, B.pt (180-220)

o

C) was purchased from Sigma-

84

Polyacrylonitrile (PAN) powder was purchased from Scientific Polymer Products Inc. (MW avg.

85

150,000). Dimethylformamide (DMF) was purchased from Acros Organics. For comparison purposes,

86

Granular Activated Carbon (GAC) was used in this study as a commercial adsorbent. GAC (Pfaltz and

87

Bauer) was ordered from Fisher Scientific.

88

2.1.1 Fabrication of Nonwovens

89

Polyacrylonitrile (PAN) and dimethylformamide (DMF) were mixed to prepare 14 wt. % PAN in

90

DMF solutions, under constant stirring at 60 °C for 2 hr. The solution was dispensed using a syringe

91

pump (KD Scientific), at a constant rate of 1cc/h from a 20 gauge blunt-end needle. The collector was a

92

grounded drum rotating at 70 rpm. The needle to collector distance was 18 cm and the applied voltage

93

difference between needle and collector was 22- 24 kV.

94

The electrospun nonwoven precursor was then stabilized in air at 280 °C for 1 h in a muffle

95

furnace (Carbolite), followed by a carbonization step in a tube furnace (Lindberg Blue M, Thermo 5

96

Scientific) at 600 °C for 2 h in an inert nitrogen atmosphere to make carbon nanofiber nonwoven

97

(CNFN). The ramp rate was 1 °C/min for stabilization and 3 °C/min for carbonization. Activated

98

carbon nanofiber nonwovens (ACNFNs) were produced by exposing the CNFN to 1 g/min of steam in

99

an inert nitrogen atmosphere at 750 °C for 1 h in the same furnace. The ramp rate for activation was 5

100

°C/min.

101

2.1.2 Preparation of the Emulsion

102

Petroleum oil was used to prepare synthetic oily wastewater. Petroleum oil was added to 500 mL

103

of distilled water to make mixtures of 100, 500, and 1000 mg/l. The mixture was emulsified in a high-

104

speed blender (Waring Two-Speed Blenders) at 24000 rpm for 8 minutes, at room temperature.

105

Emulsions were prepared with three different petroleum oil concentrations: 100, 500, and 1000 mg/L

106

Emulsion stability was measured to ensure a reasonably constant emulsion size during the

107

handling of the emulsion. Droplets' size distributions of the prepared emulsions were evaluated by laser

108

light scattering using a submicron Particle Sizer (NICOMP, Santa Barbara, California, USA). The

109

measurements were assessed at different time intervals from preparation (1, 2, and 6 hours) for all oil

110

concentrations. The total measurement time for each sample was 10 minutes.

111

Oil concentrations were measured using UV/Vis spectrophotometer (Thermo Scientific Genesys

112

10S) at a wavelength of 260 nm. The viscosity and refractive index of the samples were measured at 25

113

o

114

Abbe), respectively.

C using Brookfield viscometer (LVDV-I Prime Cone/Plate) and Refractometer (Bausch and Lomb

115

2.1.3 Nonwovens Characterization

116

Scanning electron microscopy (SEM) images of ACNFN samples were obtained before and after 6

117

tests using field emission scanning electron microscopy (FESEM, JEOL 6335F) to determine the

118

surface morphology. The average fiber diameter of fifty individual fibers imaged was measured using

119

Image J software (National Institutes of Health, USA).

120

Nitrogen adsorption/desorption isotherms of the adsorbent samples were obtained using the

121

Physisorption Analyzer (Micromeritics Instrument Corporation). The specific surface area of each

122

sample was calculated through the application of the BET model. For carbon materials, the surface

123

hydrophobicity was determined by measuring their contact angle with water, using a CAM 101 series

124

contact angle goniometer. The average of ten measurements was taken, with a water droplet volume of

125

5 ± 0.5 µL.

126

127

2.2 Emulsified Oil Removal Experiments

2.2.1 Batch Experiments

128

Batch oil adsorption tests were conducted by placing 0.1 g (a rectangular sample of dimensions 4

129

cm X 2 cm) of adsorbent material into 250 mL synthetic emulsified oil at room temperature. The

130

solutions were stirred by a magnetic stirrer at a rate of 100, 200, and 300 rpm. A cover was used over

131

the stir bar to prevent the sample from coming into direct contact with the stir bar.

132

Control experiments (without using adsorbent) were conducted to ensure that a decrease in oil

133

concentration was not caused by instability of the emulsion during the experiments or adsorption onto

134

the glass. Oil concentrations of 100, 500, and 1000 mg/L were used.

135 136

Adsorption performance of the ACNFNs was compared with equivalent masses of CNFNs (and GACs. The initial pH was 7.3 for all solutions.

7

137

2.2.1 Flow-Through Experiments

138

A schematic diagram of the experimental setup used to remove emulsified oil from water by the

139

fabricated membranes is shown in Fig. 1. The setup consisted of a separating module that would hold

140

small coupon stacks of the ACNFN, a syringe pump, mixer, pressure gauge and a UV detector with a

141

flow-through cell to detect the variation of oil concentration over time. As shown in Fig.2, the

142

separating module included two custom-designed conical pieces designed to hold the nonwoven

143

membrane layers in the system (the thickness of each layer was 200 µm). The flow-through module

144

was designed as two mirrored pieces. The pieces to assemble the module incorporated a mesh to

145

support the ACNFN and an o-ring channel to seal around the material. The two pieces of the module

146

were designed in Solid works and printed on Form 1 stereolithographic 3D printer (Formlabs,

147

Somerville, USA). Form 1 has used a laser to cure a proprietary resin consisting of a photoinitiator and

148

acrylic oligomers.

149

The nonwoven membrane mats were cut into 3.8 cm discs to fit the designed holder. The effective

150

area of the nonwoven membrane was 7 cm2. The feed solution was continuously mixed throughout the

151

experiment using an agitating tank at a stirring speed of 500 rpm to avoid the destabilization of the oil

152

droplets during the experiments. A syringe pump (KD Scientific) was used to deliver the emulsion at a

153

constant rate. The outlet solution was sent to a UV detector which is connected to a computer to

154

monitor the oil concentration.

155 156

Fig. 1-Schematic diagram of the flow-through system.

157 158 8

159

Fig. 2- Images of the separating module

160

3. Results and Discussion

161

3.1 Nonwovens Characterization

162

The surface morphology and corresponding fiber-size distributions of 14 wt. % PAN-based

163

nonwoven precursor, CNFN and ACNFN membranes are shown in Fig. 3 (a-c), respectively. The range

164

of fiber-size distributions and the average fiber size decreased during carbonization and activation. This

165

is consistent with previous work [33,39,41] and is related to weight loss during activation.

166

The precursor nanofibers exhibited a range of diameters between 430 and 700 nm (Fig. 3 a). The

167

CNFN morphology and showed a significant reduction in fiber size (Fig. 3 b). The reduction in size is

168

due to gases (e.g. H2O, N2, and HCN) evolving from the polymer during carbon cyclization [33,39].

169

Further reduction in fiber size during activation occurs as steam reacts with the carbon, releasing CO

170

and H2 and creating “defects” on the surface [34].

171

Table 1 presents the contact angle and the specific surface area of the nonwoven membranes at

172

various stages in their fabrication. The PAN-based precursor membrane has a hydrophilic surface and

173

small specific surface area. The surface of the CNFN is hydrophobic due to the increase in carbon

174

content, and the specific surface area increases slightly due to the release of gasses and the loss of non-

175

carbon elements. Steam activation induces small defects in the pore structure, which increases the

176

specific surface area. Steam activation also introduces hydroxyl groups onto the surface which

177

increases hydrophilicity [39]

178 179

Fig. 3-The SEM images and fiber size distribution of 14 wt. % PAN-based (a) Electrospun

180

Precursor, (b) CNFN, and (c) ACNFN. (n=50) 9

181

Fig. 4 depicts typical N2 adsorption/desorption isotherms for the carbon-based adsorbents. The

182

CNFN sample shows an increase in N2 adsorption as the relative pressure (P/P0) increases. This is

183

referred to as type II adsorption according to the IUPAC classification [42]. Most of the nitrogen

184

adsorption on GAC occurs at a low relative pressure (P/P0 < 0.1), and a long plateau appears at higher

185

relative pressures. This is referred to as Type I sorption and is characteristic for micro-porous materials

186

(pore size < 2 nm). The adsorption isotherm of ACNFN shows moderate N2 adsorption at P/P0 < 0.1

187

and a continuous increase of N2 adsorption along with the enlarged over the region of 0.1 < P/P0 < 0.9,

188

suggesting the presence of both microporous and mesoporous in ACNFN which was consistent with

189

previous work [39].

190

The specific surface area and the contact angle of each used adsorbent are presented in Table 1.

191

The high contact angle of the CNFN indicates a surface hydrophobicity. After steam activation, the

192

ACNFN surface exhibits a lower hydrophobicity. This is attributed to the increased oxygen content of

193

the material added during exposure to high-temperature steam [39,43,44]. As compared to the GAC the

194

hydrophobicity of both the CNFN and ACNFN is high. This is due to the differences in the surface

195

hierarchical structures [37].

196 197

Fig. 4- Nitrogen adsorption isotherms of ACNFN, GAC, and CNFN from BET analysis

198

Table 1- Specific surface areas and contact angles of the membranes tested

199 200 201

10

202

3.2 Emulsion Characterization

203

Fig. 5 presents the histograms of the oil droplets size distribution of emulsions over time. The

204

histogram divides the particle counts into discrete size ranges along the horizontal axis. The height of

205

each vertical bar corresponds to the volume percentage of oil droplets in each range. The histograms of

206

the emulsion confirm the successful preparation of an emulsion without the use of an emulsifier. The

207

average droplet size is less than 6 µm and remains below this value for more than 6 h. This indicates

208

that the oil in water mixture produced in our experiments can be classified as emulsions.

209 210 211

Fig. 5 - Evolution of the droplet size distributions and average droplet size of emulsions (500

212

mg/L oil) at 25 oC.

213

3.3 Oil Removal Dynamic Study

214

3.3.1 Effect of Initial Oil Concentration

215

Fig. 6 a shows the removal of emulsified oil by ACNFN as a function of time, at 25 oC, at

216

different initial oil concentrations. For the various initial oil concentrations, the plots show increasing

217

oil adsorption over time. For short times, the change of the normalized concentration in time is similar

218

to the three oil concentrations. This would imply that the kinetics of the process are roughly first order

219

in the oil concentration (ln(c/c0) ~ (-k.t), with k the overall rate constant). Indeed, for all three

220

emulsions the ln(c/c0) versus ~ (-k.t) (Fig. 6 b) gives a straight line with slope k= (1.5 + 1)*10-4 s-1. For

221

low concentration, this linear relation persists for up to 4 hours. For the highest initial concentration of

222

1000 mg/L, a deviation of the linear trend is observed, which is attributed to saturation of the adsorbent

223

material, and a more pronounced shift to larger droplet sizes for the more concentrated oil in water 11

224

emulsions (Fig. 7). First-order kinetics of sorption processes have been observed frequently by others,

225

see e.g., [21,45]. In our study, the reason can be external mass transport resistance, from the liquid bulk

226

towards the surface (of the individual fibers) in the ACNFN, even at the high stirring speed that was

227

applied.

228 229

Fig. 6 - The effects of the initial oil concentration on (a) oil removal percentage, and (b) the

230

kinetics of the adsorption process. All experiments were done at 25 oC, the mixing rate was

231

300 rpm, the ACNFN dose was 100 mg, and the emulsion volume was 250 mL.

232 233

Fig. 7 - Oil drop size distribution and average droplets size for different oil concentration

234

emulsions after 1 h of preparation, at 25 oC.

235

3.3.2 Effect of Stirrer Speed

236

Fig. 8 (a-c) depicts the effect of the stirring speed on the oil removal rate. Clearly, the rate of oil

237

removal increases with increasing stirrer speed. The same behavior was noticed by [20] for oil

238

adsorption using meshed corncobs. It confirms that external mass transport resistance has a pronounced

239

effect on the kinetics of the adsorption process. An increase in the stirrer speed corresponds to an

240

increase in the Reynolds number, which is well-known to result in an increase in the Sherwood number

241

and hence in the mass transfer coefficient. In the batch set-up used in this study, the stirrer speed could

242

not be increased further, implying that the kinetics of oil removal is dictated by the process conditions

243

rather than by the inherent kinetics of the sorption of oil into the nano-sized pores of the ACNFN.

244 245

(a)

(b)

246 247

(c) 12

248

Fig. 8 - The effect of the rotating speed on the adsorption process. All experiments were

249

done at 25 °C, the initial oil concentration was 500 mg/L, the ACNFN dose was 100 mg, and

250

the solution volume 250 mL.

251

3.3.3 Effect of Adsorbent Type

252

Fig. 9 represents the oil removal efficiency of an ACNFN when compared with a CNFN and

253

GAC. The adsorbents show different oil removal rates and capacities. The GAC displays a low rate of

254

oil removal, in spite of its relatively high surface area. This behavior has been associated with the

255

hydrophilicity of the pore surface of the GAC, which would favor the formation of "water clusters" that

256

block some of the pores [46]. The rate of oil uptake of the CNFN and GAC are similar for the first ~4

257

hours. After this, the rate of oil removal appears to increase for the CNFN. This is attributed to the

258

more hydrophobic surface characteristics of the CNFN, which implies a higher affinity of oil for this

259

surface and prevents the formation of water clusters. For the ACNFN, the kinetics are limited by

260

external mass transport resistances. Consequently, the oil removal rate is first order in the oil

261

concentration, and much faster than observed for the other adsorbents, With the ACNFN, ~20% of the

262

emulsified oil was removed within in 2 h.

263 264

Fig. 9 - Effect of adsorbent type on the adsorption effectivity. All experiments were done at

265

25 oC, the mixing rate was 300 rpm, initial oil concentration was 500 mg/L, ACNFN dose

266

was 100 mg, and the solution volume was 250 mL.

267

3.4 Flow Through Adsorption Performance

268

3.4.1 Effect of Membrane Material

269

Fig. 10 a shows the removal of emulsified oil as a function of time using eight layers of the PAN-

270

based precursor, CNFN and ACNFN mats in a flow-through system using an influent emulsion flow

271

rate of 400 mL/h. The corresponding pressure drop in the system is shown in Fig. 10 b. Table 2 shows 13

272

SEM images of both sides of the first layer of the membranes used for oil removal.

273

The PAN-based precursor nonwoven membrane decreased the concentration of the permeate oil to

274

half that in the influent stream, which continued to decrease during the experiment. The pressure drop

275

through the module increased over time (Fig. 10 b). This suggests that clogging is occurring and this is

276

confirmed in Table 2. This phenomenon may have been exacerbated by the coalescence of oil droplets

277

at the surface of the nonwoven as they attempted to move through the matrix.

278

The CNFN behaved differently. In the permeate, the concentration of the oil e was initially quite

279

low, then increased over 15 min, and then decreased. This characteristic breakthrough curve suggested

280

adsorption of oil and then breakthrough, but then coalescence of oil and clogging. The same behavior

281

was seen for the ACNFN, which exhibited even better oil removal and prevented breakthrough for a

282

longer period.

283

This behavior can be explained based on the removal mechanisms. The precursor simply removes

284

the oil by coalescence and surface screening. This means that as oil droplets coalesce, they grow in size

285

and the nonwoven screens them at the feed surface. As the PAN fibers clogs, the effective pore size of

286

the nonwoven decreases and the fiber becomes more size-selective. This is a type of filter ripening

287

effect and is indicated by the increased pressure drop across the fiber mat during the test (Fig. 10 b).

288

There may be some adsorption on the PAN fibers as well, though the hydrophilicity of the material

289

would make that less likely (or have less effect) than the more hydrophobic carbon materials.

290

With more hydrophobic materials, we can expect that the wicking of the oil into the fiber matrix

291

would occur. This wicking produces the breakthrough effect noted in the first hour of the test for both

292

materials. Eventually, the fiber is overwhelmed with oil and begins to clog, causing the oil removal

293

mechanism to shift from adsorption to size exclusion as the effective pore sizes drop. This behavior is 14

294

enhanced by the ACNFN mesoporosity, which allows for the adsorption of oil into the individual

295

fibers. The increased capacity for oil adsorption enables a longer breakthrough event prior to nonwoven

296

clogging and pore size shrinkage. The lower propensity to clog for both CNFN and ACNFN is also

297

demonstrated in the lower pressure drop across the mat for both materials (Fig 10 b). Table 2 would

298

also suggest that even for these materials, most of the clogging occurs in the upper layers of the

299

nonwoven.

300 301

Fig. 10 (a) Oil removal performance of the various membrane materials used in this study.

302

(b) Pressure drop in the system for the various membranes. The inlet flow oil concentration

303

was 500 mg/L, the inlet flow rate was 400 ml/h and eight membrane layers were used.

304

15

305

Table 2-SEM images of the various membrane materials after their use in the flow-through

306

system for oil removal at room temperature (25 °C). The inlet flow oil concentration was 500

307

mg/L, the inlet flow rate was 400 mL/h and eight membrane layers were used.

308 309

310

3.4.2 Effect of Membrane Thicknesses

311

The effect of the number of the nonwoven precursor, CNFN and ACNFN layers on the efficiency

312

of oil removal is shown in Fig. 11 (a-c), respectively. Using fewer layers of the precursor membrane

313

had no obvious effect on the oil concentration in the effluent. This is due to the fact that oil drops were

314

removed by size exclusion on the top of the first layer, so subsequent layers played a little roll in

315

removal. The number of nonwoven layers of the CNFN and ACNFN membranes, however, had a clear

316

effect on the amount of oil in the permeate stream.

317

responsible for oil removal, thicker nonwovens provided more material for adsorption and longer

318

residence times to adsorb. Activation of the carbon had a modest effect on effluent quality as more

319

surface area provided more adsorption area for oil removal. In this case, while breakthrough occurred

320

relatively quickly, the permeate oil concentration was substantially lower than that exiting the CNFN

321

material. We note that these experiments lasted 2 hours, over which the emulsion droplet size changed

322

modestly.

Since adsorption mechanisms were in part

323 324

Fig. 11-Effect of layer number on oil removal using a) ACNFN, b) CNFN and c) precursor

325

membranes. Inlet oil concentration was 500 mg/L, and the Inlet flow rate was 400 mL/h.

326

3.4.3 Effect of Flow Rate of the Inlet Solution

327

The effect of influent flow rate on the removal of emulsified oil droplets by eight layers of 16

328

ACNFN membrane mats at an inlet oil concentration of 500mg/L using various influent flow rates (50,

329

100 and 400 mL/h) is shown in Fig. 12. ACNFN membranes removed emulsified oil droplets by a

330

combination of two mechanisms (adsorption and size exclusion) until the adsorption sites were fully

331

saturated. Removal peaked at this point, and the oil concentration then decreased due to the size

332

exclusion of oil droplets on the top surface of the membrane. Decreasing the flow rate had a substantial

333

effect on the concentration of permeate oil and on the time required to reach the peak (Fig. 12 a). The

334

contact time between the oil droplets and the adsorption sites increased by decreasing the inlet flow

335

rate, leading to an increase in oil removal by adsorption.

336

The slope of the breakthrough curve increased with a volumetric inlet flow rate, reaching the

337

peak faster than with the other flow rates. The residence time of the emulsified oil droplets among the

338

nanofibers was too short for occupying all the adsorption sites. Oil droplets accumulated quickly on the

339

surface of the ACNFN mat, creating a distinct layer of oil on the top layer. The contact time between

340

the adsorbate and the adsorbent was thus minimized, leading to an earlier peak, increasing the flow rate

341

and reducing the time for saturation.

342

The variation of concentration of the permeate oil with the treated volume of the emulsion is

343

shown in Fig. 12 b. Treating the same volume of solution at a higher flow rate increased the oil

344

concentration in the permeate flow due to a decrease in adsorption. Decreasing the inlet flow rate,

345

however, increased the time needed for adsorption and enhanced oil removal.

346 347

Fig. 12-Effect of inlet flow rate on oil removal breakthrough at room temperature using

348

eight layers of ACNFN membrane, and Inlet oil concentration was 500 mg/L, and the Inlet

349

flow rate was 400 mL/h.

350

4. 17

Conclusions

351

In this work, it has been demonstrated that PAN based electrospun ACNFN can be used for

352

emulsified oil adsorption from oily wastewater. The ACNFN has unique surface properties and exhibits

353

a high rate of, and capacity for, oil removal. It has a high specific surface area, mesoporous structure,

354

and high hydrophobicity. In batch experiments, the initial oil removal rate is roughly first order in the

355

oil concentration, and strongly dependent on the stirrer speed. This implies that the inherent oil

356

sorption kinetics of the ACNFN are fast as compared to the external mass transport resistances. The

357

results show additional adsorption of (aggregated) oil droplets on the outside of the fibers in the

358

ACNFN, in addition to the uptake of oil in the nano-sized pores of the fibers. The results imply that the

359

ACNFN allows for much faster removal of larger amounts of oil from oil-water emulsions, as

360

compared to other (commercially available) forms of activated carbon. The authors acknowledge that

361

this study neglects to discuss the cost of the materials relative to one another. ACNFN is likely to be an

362

expensive sorbent material compared to GAC. That said, it is producible at-scale and also offers

363

unique adsorption properties that might relegate it to specialty adsorption applications where more

364

expensive materials can be justified so long as their performance is improved.

365

5. Acknowledgement

366

The authors acknowledge financial support from the Higher Committee for Education

367

Development (HCED) in Iraq (Grant Number is 1000324). The authors also acknowledge the NWRI-

368

AMTA Fellowship for Membrane Technology and National Science Foundation GK-12 Program,

369

which provided support for Jason T. Arena.

370 371

6. Reference [1]

B.I.H. Waisi, U.F. a. Karim, D.C.M. Augustijn, M.H.O. Al-Furaiji, S.J.M.H. Hulscher, A study 18

372

on the quantities and potential use of produced water in southern Iraq, Water Sci. Technol.

373

Water Supply. 15 (2015) 370. doi:10.2166/ws.2014.122.

374

[2]

M.H.O. Al-Furaiji, U.F.A. Karim, D.C.M. Augustijn, B.I.H. Waisi, S.J.M.H. Hulscher,

375

Evaluation of water demand and supply in the south of Iraq, J. Water Reuse Desalin. 6 (2016).

376

doi:10.2166/wrd.2015.043.

377

[3]

W. Lan, G. Gang, W. Jinbao, Biodegradation of oil wastewater by free and immobilized

378

Yarrowia lipolytica W29, J. Environ. Sci. 21 (2009) 237–242. doi:10.1016/S1001-

379

0742(08)62257-3.

380

[4]

V.B. Lazarevic, I.M. Krstic, M.L. Lazic, D.S. Savic, D.U. Skala, V.B. Veljkovic, Scaling up the

381

chemical treatment of spent oil-in-water emulsions from a non-ferrous metal-processing plant,

382

Hem. Ind. 67 (2013) 59–68. doi:10.2298/HEMIND120317055L.

383

[5]

P. Kundu, I.M. Mishra, Removal of emulsified oil from oily wastewater ( oil-in-water emulsion

384

) using packed bed of polymeric resin beads, Sep. Purif. Technol. 118 (2013) 519–529.

385

doi:10.1016/j.seppur.2013.07.041.

386

[6]

C.S. Kim, Y.K. Cho, B.J. Choi, K.T. Jung, S.H. You, Improving a prediction system for oil

387

spills in the Yellow Sea: Effect of tides on subtidal flow, Mar. Pollut. Bull. 68 (2013) 85–92.

388

doi:10.1016/j.marpolbul.2012.12.018.

389

[7]

M. Cheryan, N. Rajagopalan, Membrane processing of oily streams . Wastewater treatment and

390 391

waste reduction, J. Memb. Sci. 151 (1998) 13–28. [8]

D. Wang, T. Silbaugh, R. Pfeffer, Y.S. Lin, Removal of emulsified oil from water by inverse

392

fluidization

393

doi:10.1016/j.powtec.2010.05.021.

394

[9]

aerogels,

Powder

Technol.

203

(2010)

298–309.

Separation, Sep. Sci. Technol. 40 (2005) 3407–3418. doi:10.1080/01496390500423755. [10]

M.L.S. Welz, N. Baloyi, D.A. Deglon, Oil removal from industrial wastewater using flotation

397 398

hydrophobic

Y. Suzuki, T. Maruyama, Removal of Emulsified Oil from Water by Coagulation and Foam

395 396

of

in a mechanically agitated flotation cell, Water SA. 33 (2007) 453–458. [11]

F.-R. Ahmadun, A. Pendashteh, L.C. Abdullah, D.R.A. Biak, S.S. Madaeni, zurina Z. Abidin, 19

399

Journal

400

doi:10.1016/j.jhazmat.2009.05.044.

401

[12]

of

Hazardous

Materials,

J.

Hazard.

Mater.

170

(2009)

530–551.

Y. r. QIU, H. ZHONG, Q. x. ZHANG, Treatment of stable oil/water emulsion by novel felt-

402

metal supported PVA composite hydrophilic membrane using cross flow ultrafiltration, Trans.

403

Nonferrous Met. Soc. China (English Ed. 19 (2009) 773–777. doi:10.1016/S1003-

404

6326(08)60348-9.

405

[13]

X. Xiong, B. Jianguo, S.H. Omer, G. Hui, Z. Yu, W. Hong, DStudy for adsorption behaviors of

406

emulsion oil on a novel ZrO 2 / PVDF modified membrane, Desalin. Water Treat. 3994 (2015)

407

37–41. doi:10.1080/19443994.2015.1044918.

408

[14]

N.S. Majeed, H.A. Sabbar, N.O.A. Al-musawi, Preparation Activated Carbon from Scrap Tires

409

by Microwave Assisted KOH Activation for Removal Emulsified Oil, Iraqi J. Chem. Pet. Eng.

410

18 (2017) 57–69.

411

[15]

A. Srinivasan, T. Viraraghavan, Oil removal from water using biomaterials, Bioresour. Technol.

412 413

101 (2010) 6594–6600. doi:10.1016/j.biortech.2010.03.079. [16]

L.I. Xiaobing, Z. Chunjuan, L.I.U. Jiongtian, Adsorption of oil from waste water by coal :

414

characteristics and mechanism, Min. Sci. Technol. 20 (2010) 778–781. doi:10.1016/S1674-

415

5264(09)60280-5.

416

[17]

S. Ibrahim, S. Wang, H.M. Ang, Removal of emulsified oil from oily wastewater using

417

agricultural

418

doi:10.1016/j.bej.2009.11.013.

419

[18]

waste

barley

straw,

uidization

421

doi:10.1016/j.powtec.2010.05.021. [19]

of

hydrophobic

J.

49

(2010)

78–83.

aerogels,

Powder

Technol.

203

(2010)

298–309.

K. Okiel, M. El-sayed, Treatment of oil – water emulsions by adsorption onto activated carbon

423

,

424

doi:10.1016/j.ejpe.2011.06.002.

425

Eng.

D. Wang, T. Silbaugh, R. Pfeffer, Y.S. Lin, Removal of emulsi fi ed oil from water by inverse fl

420

422

Biochem.

[20]

bentonite

and

deposited

carbon,

Egypt.

J.

Pet.

20

(2011)

9–15.

B.A. Olufemi, L.A. Jimoda, N.F. Agbodike, Adsorption of Crude Oil using Meshed Corncobs, 20

426

Asian

427

https://pdfs.semanticscholar.org/16f6/b5d3ead4bad007d1f1d7ebf27757eef7759f.pdf.

428

[21]

J.

Appl.

Sci.

Eng.

3

(2014)

7–21.

E.A. Emam, Modified Activated Carbon and Bentonite Used to Adsorb Petroleum

429

Hydrocarbons Emulsified in Aqueous Solution, Am. J. Environ. Prot. 2 (2013) 161.

430

doi:10.11648/j.ajep.20130206.17.

431

[22]

D. Mysore, T. Viraraghavan, Y.C. Jin, Treatment of oily waters using vermiculite, Water Res.

432 433

39 (2005) 2643–2653. doi:10.1016/j.watres.2005.04.034. [23]

A. Li, H.-X. Sun, D.-Z. Tan, W.-J. Fan, S.-H. Wen, X.-J. Qing, et al., Superhydrophobic

434

conjugated microporous polymers for separation and, Energy Environ. Sci. 4 (2011) 2062–

435

2065. doi:10.1039/c1ee01092a.

436

[24]

G.R. Alther, Organically modified clay removes oil from water, Waste Manag. 15 (1995) 623–

437 438

628. doi:10.1016/0956-053X(96)00023-2. [25]

F.M.T. Luna, A.A. Pontes-Filho, E.D. Trindade, I.J. Silva, D.C.S. Azevedo, C.L. Cavalcante,

439

Removal of aromatic compounds from mineral naphthenic oil by adsorption, Ind. Eng. Chem.

440

Res. 47 (2008) 3207–3212. doi:10.1021/ie071476v.

441

[26]

S. Rafiei, B. Noroozi, S. Arbab, A.K. Haghi, Characteristic assessment of stabilized

442

polyacrylonitrile nanowebs for the production of activated carbon nano-sorbents, Chinese J.

443

Polym. Sci. 32 (2014) 449. doi:10.1007/s10118-014-1410-4.

444

[27]

X. Wang, J. Yu, G. Sun, B. Ding, Electrospun nanofibrous materials: a versatile medium for

445

effective

446

http://linkinghub.elsevier.com/retrieve/pii/S1369702115003855.

447

[28]

separation,

Mater.

Today.

00

(2015)

0–11.

C.K. Liu, K. Lai, W. Liu, M. Yao, R.J. Sun, Preparation of carbon nanofibers through

448 449

oil/water

electrospinning and thermal treatment, Polym. Int. 58 (2009) 1341–1349. doi:10.1002/pi.2669. [29]

Y. Bai, Z.H. Huang, M.X. Wang, F. Kang, Adsorption of benzene and ethanol on activated

450

carbon nanofibers prepared by electrospinning, Adsorption. 19 (2013) 1035–1043.

451

doi:10.1007/s10450-013-9524-5.

452

[30]

H. Liu, C.-Y. Cao, F.-F. Wei, P.-P. Huang, Y.-B. Sun, L. Jiang, et al., Flexible macroporous 21

453

carbon nanofiber film with high oil adsorption capacity, J. Mater. Chem. A. 2 (2014) 3557.

454

doi:10.1039/c3ta14468b.

455

[31]

X. Huang, Fabrication and properties of carbon fibers, Materials (Basel). 2 (2009) 2369–2403.

456 457

doi:10.3390/ma2042369. [32]

K.A. Khalil, H. Eltaleb, H.S. Abdo, S.S. Al-deyab, H. Fouad, Carbon Nanofibers Containing

458

Ag/TiO2 Composites as a Preliminary Stage for CDI Technology, J. Mater. Sci. Chem. Eng. 2

459

(2014) 31–37. doi:10.4236/msce.2014.21006.

460

[33]

H. Lee, H. Kang, K. An, H. Kim, B. Kim, Comparative studies of porous carbon nanofibers by

461 462

various activation methods, Carbon Lett. 14 (2013) 180–185. doi:10.5714/CL.2013.14.3.180. [34]

Z. Shao, M. Pang, W. Xia, M. Muhler, C. Liang, Creation of surface defects on carbon

463

nanofibers by steam treatment, J. Energy Chem. 22 (2013) 804–810. doi:10.1016/S2095-

464

4956(13)60107-2.

465

[35]

B.I. Waisi, S.S. Manickam, N.E. Benes, A. Nijmeijer, J.R. McCutcheon, Activated Carbon

466

Nanofiber Nonwovens: Improving Strength and Surface Area by Tuning the Fabrication

467

Procedure,

468

https://pubs.acs.org/doi/10.1021/acs.iecr.8b05612.

469

[36]

Ind.

Eng.

Chem.

Res.

58

(2019)

4084–4089.

G. Oh, Y. Ju, M. Kim, H. Jung, H. jin Kim, W.-J. lee, Adsorption of toluene on carbon

470

nanofibers prepared by electrospinning, Sci. Total Environ. 393 (2008) 341–347.

471

doi:10.1016/j.scitotenv.2008.01.005.

472

[37]

K.J. Lee, N. Shiratori, G.H. Lee, J. Miyawaki, I. Mochida, S. Yoon, et al., Activated carbon

473

nanofiber produced from electrospun polyacrylonitrile nanofiber as a highly efficient

474

formaldehyde

475

doi:10.1016/j.carbon.2010.07.034.

476

[38]

adsorbent,

Carbon

N.

Y.

48

(2010)

4248–4255.

P. Sullivan, J. Moate, B. Stone, J.D. Atkinson, Z. Hashisho, M.J. Rood, Physical and chemical

477

properties of PAN-derived electrospun activated carbon nanofibers and their potential for use

478

as an adsorbent for toxic industrial chemicals, Adsorption. 18 (2012) 265–274.

479

doi:10.1007/s10450-012-9399-x.

22

480

[39]

S.S. Manickam, U. Karra, L. Huang, N. Bui, B. Li, J.R. McCutcheon, Activated carbon

481

nanofiber

482

http://dx.doi.org/10.1016/j.carbon.2012.10.009.

483

[40]

anodes

for

microbial

fuel

cells,

Carbon

N.

Y.

53

(2013)

19–28.

B.I. Waisi, S.M. Al-jubouri, R. Mccutcheon, Fabrication and Characterizations of Silica

484

Nanoparticle Embedded Carbon Nanofibers, Ind. Eng. Chem. Res. 58 (2019) 4462–4467.

485

doi:10.1021/acs.iecr.8b05825.

486

[41]

H. Tavanai, R. Jalili, M. Morshed, Effects of fiber diameter and CO2 activation temperature on

487

the pore characteristics of polyacrylonitrile based activated carbon nanofibers, Surf. Interface

488

Anal. 41 (2009) 814–819. doi:10.1002/sia.3104.

489

[42]

M.E. Ramos, P.R. Bonelli, A.L. Cukierman, M.M.L. Ribeiro Carrott, P.J.M. Carrott, Influence

490

of thermal treatment conditions on porosity development and mechanical properties of

491

activated carbon cloths from a novel nanofibre-made fabric, Mater. Chem. Phys. 116 (2009)

492

310–314. doi:10.1016/j.matchemphys.2009.03.042.

493

[43]

C. Moreno-Castilla, Adsorption of organic molecules from aqueous solutions on carbon

494 495

materials, Carbon N. Y. 42 (2004) 83–94. doi:10.1016/j.carbon.2003.09.022. [44]

P. Serp, B. Machado, Carbon (Nano)Materials for Catalysis, in: Nanostructured Carbon Mater.

496 497

Catal., Royal Society of Chemistry, 2015. doi:10.1039/9781782622567-00001. [45]

H.H. Sokker, N.M. El-Sawy, M.A. Hassan, B.E. El-Anadouli, Adsorption of crude oil from

498

aqueous solution by hydrogel of chitosan based polyacrylamide prepared by radiation induced

499

graft

500

doi:10.1016/j.jhazmat.2011.03.055.

501

[46]

polymerization,

J.

Hazard.

Mater.

190

(2011)

359–365.

A.P. Terzyk, Further insights into the role of carbon surface functionalities in the mechanism of

502

phenol adsorption, J. Colloid Interface Sci. 268 (2003) 301–329. doi:10.1016/S0021-

503

9797(03)00690-8.

504

23

Fig. 1-Schematic diagram of the flow-through system.

3.8 cm 1.5 cm

Fig. 2- Images of the separating module

Fig. 3-The SEM images and fiber size distribution of 14 wt. % PAN based (a) Electrospun Precursor, (b) CNFN, and (c) ACNFN. (n=50)

Fig. 4- Nitrogen adsorption isotherms of ACNFN, GAC, and CNFN from BET analysis

Table 1- Specific surface areas and contact angles of the membranes tested

Membrane type

Specific surface area (m2/g)

Contact angle ( o)

Preursor

7±2

80 ± 5

CNFN

15 ± 5

135 ± 2

ACNFN

520 ± 20

120 ± 3

GAC

260 + 10

25 ± 3

Fig. 5 - Evolution of the droplet size distributions and average droplet size of emulsions (500 mg/L oil) at 25 oC.

Fig. 6 - The effects of the initial oil concentration on (a) oil removal percentage, and (b) the kinetics of the adsorption process. All experiments were done at 25 oC, the mixing rate was 300 rpm, the ACNFN dose was 100 mg, and the emulsion volume was 250 mL.

Fig. 7 - Oil drop size distribution and average droplets size for different oil concentration emulsions after 1 h of preparation, at 25 oC.

(a)

(b)

(c) Fig. 8 - The effect of the rotating speed on the adsorption process. All experiments were done at 25 °C, the initial oil concentration was 500 mg/L, the ACNFN dose was 100 mg, and the solution volume 250 mL.

Fig. 9 - Effect of adsorbent type on the adsorption effectivity. All experiments were done at 25 oC, mixing rate was 300 rpm, initial oil concentration was 500 mg/L, ACNFN dose was 100 mg, and solution volume was 250 mL.

Fig. 10 (a) Oil removal performance of the various membrane materials used in this study. (b) Pressure drop in the system for the various membranes. The inlet flow oil concentration was 500 mg/L, inlet flow rate was 400 ml/h and eight membrane layers were used.

Fig. 11-Effect of layer number on oil removal using a) ACNFN, b) CNFN and c) precursor membranes. Inlet oil concentration was 500 mg/L, and Inlet flow rate was 400 mL/h

Fig. 12-Effect of inlet flow rate on oil removal breakthrough at room temperature using eight layers of ACNFN membrane, and Inlet oil concentration was 500 mg/L, and Inlet flow rate was 400 mL/h.

Table 2-SEM images of the various membrane materials after their use in the flow-through system for oil removal at room temperature (25 °C). The inlet flow oil concentration was 500 mg/L, inlet flow rate was 400 mL/h and eight membrane layers were used.

Feed side

Precurso r

CNFN

ACNFN

Permeate side

Highlights: 1- PAN based electrospun ACNFN can be used for emulsified oil adsorption from oily wastewater. 2- The ACNFN has unique surface properties, high specific surface area, mesoporous structure, and high hydrophobicity. 3- This implies that the inherent oil sorption kinetics of the ACNFN are fast as compared to the external mass transport resistances. 4- The results imply that the ACNFN allows for much faster removal of larger amounts of oil from oil-water emulsions, as compared to other (commercially available) forms of activated carbon.

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: