chlorine Process

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES

Remediation and Control Technologies

Trace Organic Pollutant Removal by VUV/UV/chlorine Process: Feasibility Investigation for Drinking Water Treatment on a Minifluidic VUV/UV Photoreaction System and a Pilot Photoreactor Mengkai Li, Mengyu Hao, Laxiang Yang, Hong Yao, James R. Bolton, Ernest R. Blatchley, and Zhimin Qiang Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b00611 • Publication Date (Web): 24 May 2018 Downloaded from http://pubs.acs.org on May 25, 2018

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 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 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.

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

Page 1 of 33

Environmental Science & Technology Submission to Environ. Sci. Technol.

by Li et al.

1 2

Trace Organic Pollutant Removal by VUV/UV/chlorine Process:

3

Feasibility Investigation for Drinking Water Treatment on a Mini-fluidic

4

VUV/UV Photoreaction System and a Pilot Photoreactor

5 6

Mengkai Li,†,‡ Mengyu Hao,†,§ Laxiang Yang,† Hong Yao,∗,§ James R. Bolton,ǁ

7

Ernest R. Blatchley, III,‡ and Zhimin Qiang∗,†

8 9

†

Key Laboratory of Drinking Water Science and Technology, Research Center for

10

Eco-Environmental Sciences, University of Chinese Academy of Sciences, Chinese

11

Academy of Sciences, 18 Shuang-qing Road, Beijing 100085, China

12

‡

13

United States.

14

§

15

University, Beijing 100044, China.

16

ǁ

17

Edmonton, AB T6G 1H9, Canada.

Lyles School of Civil Engineering, Purdue University, West Lafayette, Indiana 47907,

Department of Municipal and Environmental Engineering, Beijing Jiaotong

Department of Civil and Environmental Engineering, University of Alberta,

18 19

∗

20

Phone: +86 10 62849632; e-mail: [email protected] (Z. Qiang)

21

Phone: +86 10 51682157; e-mail: [email protected] (H. Yao)

Corresponding authors.

22 23

1

ACS Paragon Plus Environment


The

Page 2 of 33 by Li et al.

vacuum-ultraviolet/ultraviolet/chlorine

(VUV/UV/chlorine)

24

ABSTRACT:

25

process, with a VUV/UV mercury lamp used as the light source, was found to be a

26

highly efficient advanced oxidation process (AOP) in a previous study. Hence, its

27

application feasibility for trace organic pollutant removal from drinking water

28

becomes attractive. In this work, a bench-scale mini-fluidic VUV/UV photoreaction

29

system was used to determine the degradation kinetics of sulfamethazine (SMN), a

30

model sulfonamide antibiotic frequently detected with trace levels in aquatic

31

environments. Results indicated that SMN (0.1 mg L−1) could be degraded rapidly by

32

VUV/UV/chlorine, and a synergism was observed between the VUV/UV and

33

UV/chlorine processes. Photon-fluence based rate constants of SMN degradation were

34

determined to be 6.76 × 103 and 8.51 × 103 m2 einstein−1 at chlorine doses of 0.05 and

35

0.5 mg L−1, respectively. The presence of natural organic matter in real waters

36

significantly inhibited SMN degradation. In addition, pilot tests were conducted to

37

explore the practical performance of the VUV/UV/chlorine process, thereby allowing

38

electrical energy per order to be calculated for cost evaluation. The effect of flow

39

pattern on photoreactor efficiency was also analyzed by computational fluid dynamics

40

simulations. Both bench- and pilot-scale tests have demonstrated that the

41

VUV/UV/chlorine process, as a new AOP, has potential applications to trace organic

42

pollutant removal in small-scale water treatment.

43

Key words: vacuum-ultraviolet (VUV); VUV/UV/chlorine; advanced oxidation

44

process; trace organic pollutant; water treatment

45

2


Page 3 of 33


by Li et al.

46

■ INTRODUCTION

47

In recent years, the frequent occurrence of trace organic pollutants in water has

48

received increasing attention because these pollutants threaten aquatic ecosystems and

49

human health.1,2 Advanced oxidation processes (AOPs), which generate abundant

50

hydroxyl radicals (HO•) of a high oxidation potential (2.80 eV),3−5 can effectively

51

degrade recalcitrant organic constituents. Since ultraviolet (UV) irradiation has been

52

widely used in water treatment, which has merits of low by-product formation, small

53

footprint and high efficiency,6,7 UV-based AOPs (or UV-AOPs) are regarded as highly

54

efficient technologies for trace organic pollutant removal in drinking water,

55

wastewater, and reclaimed water.8−10

56

Commercially

available

UV light

sources

include

low-pressure

(LP),

57

medium-pressure (MP) and vacuum-UV/UV (VUV/UV) mercury lamps as well as

58

UV light emitting diodes (UV-LEDs). The VUV/UV lamp is characterized with

59

emission wavelengths at 185 and 254 nm, and has manufacturing and operational

60

costs similar to those of a conventional LP lamp.11 Thus, a photoreactor using a

61

VUV/UV lamp as the light source has advantages over that using a conventional LP

62

lamp in disinfection and trace organic pollutant removal for two reasons: (1) VUV

63

photolysis of water can generate additional HO• (eq 1),12 and (2) VUV irradiation may

64

induce a synergistic effect on treatment efficiency in the UV-AOPs.

65

hv185

H2 O → HO • + H •

(1)

66

A previous study has revealed that the VUV/UV/chlorine process is highly

67

efficient for methylene blue (MB) degradation in water.13 The MB degradation rate in

3




68

the VUV/UV/chlorine process was higher than a sum of those in the individual

69

VUV/UV and UV/chlorine processes, indicating a synergistic effect on pollutant

70

removal. The observed synergism was mainly ascribed to: (1) the addition of chlorine

71

inhibited the recombination of HO• to produce H2O2, which would otherwise occur

72

prominently near the outer surface of the protective quartz sleeve where a highest HO•

73

concentration was expected; and (2) the simultaneously formed, longer-lived

74

secondary radicals (e.g., OCl•) could degrade effectively the target pollutant as well.

75

The new VUV/UV/chlorine process involves two commonly-used water treatment

76

processes (i.e., VUV/UV and chlorine), so it has a high potential for drinking water

77

treatment.

78

A mini-fluidic VUV/UV photoreaction system (MVPS), which enables an

79

accurate determination of VUV and UV photon-fluences, considerably facilitates the

80

kinetic and mechanistic studies on photochemical degradation of organic pollutants at

81

bench-scale.14 In addition, pilot tests are inevitable for performance and cost

82

evaluations and reactor optimal design. Considering the short transmittance of VUV

83

light in water [about 90% VUV photons were absorbed by a 5.5 mm deionized (DI)

84

water layer], the VUV/UV/chlorine process is targeted at small-scale water treatment

85

(e.g., water supplies in rural and remote areas, or point of use applications).

86

In this study, the feasibility of VUV/UV/chlorine for enhanced removal of trace

87

organic pollutants in water was investigated at both bench- and pilot-scales.

88

Sulfamethazine (SMN), a sulfonamide antibiotic frequently detected at trace levels in

89

aquatic environments,15,16 was selected as a model pollutant. Bench-scale tests were

4


Page 5 of 33


by Li et al.

90

carried out on the MVPS to explore the kinetics and mechanism of SMN degradation

91

by VUV/UV/chlorine. The influences of chlorine dose, solution pH, and water matrix

92

were also examined. Thereafter, pilot-scale tests were conducted to evaluate the

93

performance in annular reactors and energy consumption (i.e., electrical energy per

94

order, EEO) of this new AOP for trace organic pollutant removal from drinking water.

95

■ EXPERIMENTAL SECTION

96

Bench-scale Tests. The MVPS, with an 8 W cold-cathode LP lamp (arc length =

97

200 mm, Wanhua Co., Zhejiang, China) as the light source, was utilized as the

98

bench-scale photoreactor (Figure S2). Details of its construction and the monitoring

99

micro-fluorescent silica detector were provided previously.13,17 A straight synthetic

100

quartz tube [VUV/UV tube, high VUV transmittance (roughly 60% per millimeter)]

101

and a straight Ti-doped quartz tube (UV tube, opaque to VUV irradiation) were used

102

for VUV/UV and UV exposures, respectively. Water samples were collected from the

103

solution container at various experimental times, so that a range of exposure fluences

104

could be obtained. A reduction equivalent exposure time (tree, s) was defined as the

105

total experimental time (t, s) multiplied by the ratio of the exposure volume of the

106

quartz tube (πr2L, m3) to the total sample volume (V, m3).13 Hence, the photon

107

fluences (einstein m–2) for VUV (Fp,VUV), UV (Fp,UV) and total exposure (Fp,total) were

108

calculated as follows:14,18

π r 2h

109

t ree =

110

0 Fp,UV = Ep,UV tree

V

t

(2) (3)

5


Environmental Science & Technology

Page 6 of 33

Submission to Environ. Sci. Technol.

by Li et al.

111

0 Fp,VUV = Ep,VUV tree

(4)

112

Fp,total = Fp,UV + Fp,VUV

(5)

113

where r and h are the radius (1.2 mm) and length (100 mm) of the VUV/UV and UV

114

tubes, respectively; and Ep,UV and Ep,VUV are the UV photon fluence rate (einstein

115

m−2 s−1) in the UV (or VUV/UV) tube and the VUV photon fluence rate (einstein m−2

116

s−1) in the VUV/UV tube, respectively. Ep,UV and Ep,VUV were determined to be

117

3.07 × 10−4 and 0.27 × 10−4 einstein m−2 s−1, respectively, whose determined method

118

was described in a previous paper.14 Note that the term symbols follow those

119

previously recommended.13,14

0

0

0

0

120

Pilot-scale Tests. To evaluate the practical performance of the VUV/UV/chlorine

121

process for trace organic pollutant removal, pilot testes were conducted in an annular

122

stainless steel photoreactor (inner diameter = 35 mm, inner length = 950 mm), which

123

contained a 105 W VUV/UV mercury lamp (Foshan Comwin Co. China, length = 780

124

mm) centered inside a high-purity quartz sleeve (outer diameter = 23 mm), as

125

illustrated in Figure 1. DI water (30 L), spiked with trace SMN ([SMN]0 = 0.05 mg

126

L−1), was stored in a tank and pumped through the pilot photoreactor at various flow

127

rates (Qs). Computational fluid dynamics (CFD) analysis was performed to explore

128

the effect of flow pattern on the photoreactor efficiency (Text S1). In certain tests,

129

humic acid (HA, 5 mg L−1 as DOC) was added to examine the competition of natural

130

organic matter (NOM) for reactive species. Influent and effluent samples were

131

collected to determine the removal efficiency of SMN.

132

Figure 1

6


Page 7 of 33


by Li et al.

133

Chemicals and Analysis. All chemicals used without otherwise stated were of

134

reagent grade or higher. SMN was obtained from Thermo Fisher Scientific (Fair Lawn,

135

NJ, USA) and dissolved in water buffered with phosphate (5 mM, pH 5−8) or borate

136

(5 mM, pH 10). Benzoic acid (BA), HA (technical grade) and NaCl were purchased

137

from Sigma-Aldrich (St. Louis, MO, USA). Free chlorine was prepared freshly from a

138

stock NaOCl solution (10−15% by weight, Sigma-Aldrich), whose concentration was

139

measured by the N,N-diethyl-p-phenylenediamine colorimetric method on a UV-vis

140

spectrophotometer (Hach DR6000, USA).19 All solutions in the bench-scale tests were

141

prepared with ultrapure water produced from a Milli-Q system (Advantage A10,

142

Millipore, USA) unless otherwise stated, and DI water was used for all pilot tests. In

143

addition, to evaluate the performance of the VUV/UV/chlorine process for real waters,

144

a surface water from Miyun reservoir (MYR) and a granular carbon-filtered water

145

from a local drinking water treatment plant (WTP) were selected. The real waters

146

were filtered through 0.45-µm membranes and stored at 4 °C until use.

147

SMN

and

BA

were mass

analyzed

using

spectrometry

ultra-high-performance and

148

chromatography-tandem

149

chromatography (Text S1). The principle degradation products were evaluated by

150

using by using ultra-performance liquid chromatography quadrupole time-of-flight

151

mass spectrometry (UPLC/Q-TOF-MS, Alliance-2695 Waters, Text S1). To scavenge

152

reactive species and identify their contributions to trace organic pollutant removal, 10

153

mM tert-butanol (TBA, Sigma-Aldrich) or 0.5 mM nitrobenzene (NB, Sigma-Aldrich)

154

were added individually to relevant reaction solutions in advance (the preliminary

7


high-performance

liquid liquid



155

tests for the applied concentrations of TBA and NB were described in Text S1). In

156

addition, the antibacterial activity variation of SMN during VUV/UV/chlorine process

157

was discussed. The bioassay was carried out by using Escherichia coli ATCC 25922

158

(E. coli) as a challenge microorganism and the initial SMN concentration was selected

159

as 10 mg L-1. The detailed experimental process was described in a previous study.14

160

■ RESULTS AND DISCUSSION

161

SMN Degradation by VUV/UV/chlorine. Figure 2 shows comparatively the

162

degradation kinetics of SMN ([SMN]0 = 0.1 mg L−1) by UV/chlorine, VUV/UV and

163

VUV/UV/chlorine. Chlorine doses of 0.05 (Figure 2a) and 0.5 mg L−1 (Figure 2b)

164

were selected as common residual chlorine concentrations at the tap and in WTP

165

effluent, respectively. Only slight SMN degradation was observed in the UV/chlorine

166

process at [chlorine]0 = 0.05 mg L−1. The notably higher reaction rate constant

167

(k′VUV/UV = 4.00 × 103 m2 einstein−1) in the VUV/UV process could be ascribed to the

168

additional oxidation by HO•. Note that in this study, photon-based fluence, rather than

169

the commonly used energy-based fluence, should be used to properly compare the

170

photochemical kinetics involving two wavelengths (i.e., 185 and 254 nm).18

171

Figure 2

172

Previous research found a synergistic effect on MB (5 mg L-1) degradation by

173

VUV/UV/chlorine.13 For SMN with a considerably lower concentration (0.1 mg L-1)

174

in this study, a degradation rate (k′VUV/UV/Cl) of 8.51 × 103 m2 einstein−1 was achieved

175

by VUV/UV/chlorine, which is again greater than the sum of the degradation rates in

176

the UV/chlorine (k′UV/Cl) and VUV/UV (k′VUV/UV) processes (Figure 2). Moreover, the 8


Page 9 of 33


by Li et al.

177

Figure S3 shows that the chlorine decays in VUV/UV/chlorine faster than that in

178

UV/chlorine process. The results demonstrate that synergism was also evident in the

179

degradation of trace organic pollutant by VUV/UV/chlorine.

180

The effect of chlorine dose on the UV/chlorine and VUV/UV/chlorine processes

181

is manifested in Figure S4a. Both k′UV/Cl and k′VUV/UV/Cl increased with increasing

182

chlorine dose, and the latter clearly increased faster. To assess the synergistic effect of

183

VUV/UV/chlorine, as compared to the total performance of VUV/UV and

184

UV/chlorine, an enhancement factor (R) was defined as follows:

185

RCl =

′ kVUV/UV/Cl ′ ′ + kVUV/UV kUV/Cl

−1

(6)

186

As shown in Figure S4b, the R value increased from 0 to 1.42 as the chlorine dose

187

changed from 0 to 1.0 mg L−1, indicating that the VUV/UV/chlorine process should

188

be effective at common chlorine doses for drinking water disinfection. The higher the

189

chlorine dose, the stronger the enhancement.

190

Contributions of Reactive Species to SMN Degradation. HO• and reactive

191

chlorine species (RCS, e.g., Cl•, ClO•, and Cl2•−) are principal reactive species in the

192

VUV/UV/chlorine process. TBA has a high reactivity toward HO•, Cl• and ClO•,9 and

193

NB reacts rapidly with HO• but slowly with RCS.9,20 As shown in Table 1, after

194

adding 10 mM TBA and 0.5 mM NB in the UV/chlorine process, k′UV/Cl decreased

195

from 0.51 × 103 to 0.04 × 103 and 0.25 × 103 m2 einstein−1, respectively. It implies that

196

the relative contributions of HO• and principle RCS (including Cl• and ClO•) to SMN

197

degradation were about 50% and 41%, respectively, and the residual 9% could be

198

ascribed to the direct UV photolysis and the oxidation of Cl2•− that rarely react with 9




199

TBA.9 In the VUV/UV/chlorine process, k′VUV/UV/Cl decreased from 8.51 × 103 to 2.58

200

× 103 and 4.19 × 103 m2 einstein−1, respectively, after adding 10 mM TBA and 0.5 mM

201

NB. Since TBA could scavenge HO•, Cl• and ClO•, about 32% of SMN degradation

202

might be attributed to the direct VUV photolysis and the oxidation of Cl2•−. HO• and

203

principle RCS (including Cl• and ClO•) contributed 49% and 19% to the SMN

204

degradation, respectively.

205

Effects of pH, NOM, Chloride and Bicarbonate. SMN degradation by

206

VUV/UV/chlorine was examined comparatively at pH values of 5.0, 7.0, 8.0 and 10.0,

207

with a maximum k′VUV/UV/Cl observed at pH 7.0 (Figure 3). The pH effect on SMN

208

degradation was mainly attributed to the different forms of chlorine, SMN, phosphate

209

(buffer), and HO• under various pH conditions. On the one side, because of the

210

dissociation equilibrium between HOCl and OCl− (pKa = 7.5 at 25 °C), chlorine

211

photolysis at a lower pH has a higher quantum yield for reactive radicals, which

212

enhanced the SMN degradation.21,22 On the other side, the anionic (SMN−) and neutral

213

(SMN0) forms of SMN (Figure S5) have different degradation rates with HO•, which

214

caused an inhibited SMN degradation with a decreasing pH.14 In addition, because the

215

reaction rate of H2PO4- and HPO42- with HO• are 2 × 104 and 1.5 × 105 M-1 s-1,

216

respectively, increasing pH could cause a higher competition of HO• and decrease the

217

k′VUV/UV/Cl.23 As a result, the synergistic effect of VUV/UV/chlorine on SMN

218

degradation was maximized near pH 7.0.

219

Figure 3

10


Page 11 of 33


by Li et al.

220

A relatively lower k′VUV/UV/Cl was found at pH 10.0 than that at pH 5.0 (Figure 3),

221

implying that the dissociation of HOCl/OCl− had a stronger impact on this process

222

than that of SMN. In addition, the dissociation of HO• into oxygen anion radical (O•−)

223

(eq 7)24 could reduce the HO• concentration, thus decreasing the k′VUV/UV/Cl at a higher

224

pH to some extent. It was also found that MB was degraded by VUV/UV/chlorine

225

more effectively at pH 5.0 than at pH 10.0.13

226

HO• ↔ H+ + O •− , pKa = 11.9

(7)

227

Table 1 shows that 1 and 10 mM chloride additions resulted in slight reductions

228

of k′VUV/UV/Cl, while 0.1 and 1 mM bicarbonate additions obviously increased

229

k′VUV/UV/Cl. As a common ion in surface waters, chloride can scavenge reactive species

230

(e.g., HO• and Cl•) to form secondary radicals such as ClOH•− and Cl2•− (eqs 8 and 9)9

231

with relatively lower oxidation potentials. Thus, the presence of chloride in water will

232

somewhat reduce the rate of organic pollutant degradation. Yet, the bicarbonate can

233

scavenge reactive species to form CO3•− (eqs 10 and 11) that has high reaction rate

234

with the compounds containing aromatic amine groups (e.g., SMN).25 Hence the

235

bicarbonate addition could enhance the SMN degradation. In addition, HA (to

236

simulate NOM) addition induced an obvious inhibition on SMN degradation (Table 1),

237

because it competed significantly for the reactive species against SMN (kHO•,NOM = 2.5

238

× 104 L mgC−1 s−1, kCl•,NOM = 1.3 × 104 L mgC−1 s−1).9,26 In the presence of 3 mg L−1

239

HA, the SMN degradation rate dropped from 8.51 × 103 to 1.12 × 103 m2 einstein−1.

240

HO• + Cl− → ClOH •−

(8)

241

Cl• + Cl− → Cl2•−

(9)

11




242

HO• + HCO3− → H2O + CO3•−

(10)

243

Cl• + HCO3− → HCl + CO3•−

(11)

244

Effect of Water Matrix. To examine water matrix effect on SMN degradation by

245

VUV/UV/chlorine, two real waters were employed including MYR and WTP with

246

their water quality given in Table S1. For the MYR water, because of a relatively high

247

DOC content (i.e., 4.2 mg L−1) competing strongly for reactive species, a low

248

degradation rate of SMN (0.86 × 103 m2 einstein−1) was observed at 0.05 mg L−1

249

chlorine dose in the VUV/UV/chlorine process (Figure 4a). By raising the chlorine

250

dose to 0.5 mg L−1, there was only a slight increase in the SMN degradation rate (0.94

251

× 103 m2 einstein−1), indicating that chlorine had an insignificant impact on SMN

252

degradation, and hence the VUV/UV process played a dominant role.

253

Figure 4

254

The WTP water had a lower DOC content (i.e., 2.6 mg L−1) than that of the MYR

255

water, so a considerably higher degradation rate of SMN (2.84 × 103 m2 einstein−1)

256

was obtained at 0.05 mg L−1 chlorine dose in the VUV/UV/chlorine process (Figure

257

4b). Moreover, it should be pointed out that the DOC composition in the WTP water

258

was also much different from that in the MYR water. The hydrophobic and large

259

molecule fraction of NOM in the WTP water was generally removed by the

260

coagulation, sedimentation, sand filtration and granular activated carbon filtration

261

processes, leaving mostly the hydrophilic and small molecule fraction which has a

262

much less reactivity toward chlorine and reactive species.27 This is also evident when

263

comparing the NOM effects in the WTP water (Figure 4b) and in the synthetic water

12


Page 13 of 33


by Li et al.

264

(Table 1) with a similar DOC content (i.e., 2.6 vs. 3.0 mg L−1). A further increase in

265

the chlorine dose to 0.5 mg L−1 made the SMN degradation rate quickly go up to 4.89

266

× 103 m2 einstein−1. Figure 4b also shows that with a photon fluence of 0.6 × 10−3

267

einstein m−2 (i.e., a very short exposure time of 1.3 s), SMN degradation could reach

268

natural log removals of 1.9 (85%) and 3.2 (96%) at chorine doses of 0.05 and 0.5 mg

269

L−1, respectively.

270

Antibacterial Activity Assessment. The antibacterial activity removal of SMN

271

during the VUV/UV/chlorine was evaluated by the bioassay, which is calculated as

272

follows:

273

I ( %) =

Amax − A ×100 Amax − Amin

(12)

274

where I (%) represents the growth inhibition of E. coli; Amax, Amin and A are the

275

absorbance of the positive control (i.e., no growth inhibition), the negative control

276

(i.e., 100% growth inhibition) and the test sample, respectively.

277

Figure S6 shows the antibacterial activity variation of SMN solution (10 mg L−1)

278

during the VUV/UV/chlorine treatment. The results indicate that the I value decayed

279

with increasing irradiated photon fluence from 0 to 9.96 × 103 einstein m−2, indicating

280

that the VUV/UV/chlorine process could eliminate the antibacterial activity

281

effectively. The decay rate of I was lower than that of SMN, and I remained about 12%

282

even with exposure of 9.96 × 103 einstein m-2 fluence, implying that the treated

283

solution still had certain residual antibacterial activity although the parent compound

284

(i.e., SMN molecule) was removed.

13




285

Pilot Tests and EEO Calculation. Pilot tests were carried out with an annular

286

photoreactor to evaluate the practical performance and calculate EEO values of the

287

VUV/UV/chlorine process. Figure S7 shows that at [chlorine]0 = 0.5 mg L−1, the

288

removal efficiency of SMN ([SMN]0 = 0.05 mg L−1) decreased slightly with

289

decreasing Q for both VUV/UV and VUV/UV/chlorine in the absence of NOM. This

290

is reasonable because a high Q results in a short exposure time (or fluence). A higher

291

removal efficiency of SMN was achieved by VUV/UV/chlorine than by VUV/UV. At

292

Q = 0.75 m3 h−1, the removal efficiency increased with the chlorine dose increasing

293

from 0.05 to 5 mg L−1 in the VUV/UV/chlorine process (inset of Figure S7). As

294

expected, in the presence of 5 mg L−1 HA, the removal efficiency of SMN dropped

295

significantly at each Q.

296

EEO has been applied as a figure-of-merit to evaluate the energy consumption of

297

AOPs when the target pollutant concentration is low (< 100 mg L−1) so that its decay

298

follows pseudo-first-order kinetics.28 By using the pilot test data, the EEO values at

299

various Q values and chlorine doses were calculated (eq 13) and are summarized in

300

Table 2. The equivalent energy consumption for chlorine production (ECl) was

301

assumed to be 0.005 kWh mg L−1 that could be expressed in terms of EEO calculation.

302

EEO =

ECl [chlorine] P + 3600Q log([SMN]o / [SMN]) log([SMN]o / [SMN])

(13)

303

where P = electrical power (kW); t = exposure time (s); and V = volume of water

304

treated (m3). Because the EEO simultaneously takes account of Q and t, one should

305

expect the same EEO at various Q values. However, Table 2 shows that for the

306

VUV/UV process, the EEO decreased with decreasing Q, which could arise from a 14


Page 15 of 33


by Li et al.

307

special characteristic of the pilot photoreactor. Previous studies found that because of

308

a high absorption coefficient (1.8 cm−1) of the 185 nm VUV light,29 a high

309

concentration of HO• was generated and accumulated in the near-lamp region of the

310

photoreactor.13 The HO• had difficulty in diffusing to the far-lamp region so that H2O2

311

could be formed, which wasted a certain portion of the VUV irradiation.29 A higher Q

312

could promote the mass transfer of reactants (e.g., SMN and HO•) in the photoreactor,

313

so a higher SMN removal efficiency (or a lower EEO) was obtained.

314

By using CFD, water velocity (v) at each node in the pilot photoreactor was

315

simulated. This vector v could be divided into two fractions: the axial velocity va

316

(lamp axis direction) and the radial velocity vr (photoreactor radial direction). The

317

profiles of va and vr at various Q values in the central cross-section of the pilot

318

photoreactor in this study are illustrated in Figure S8. The average vr (vave,r) and the

319

relative average va (v’ave,a), which is equal to the average va in a certain water layer

320

normalized by that in the pilot photoreactor, are shown in Figure 5 for various water

321

layers. Results indicated that the v’ave,a at the near-sleeve or near-photoreactor-wall

322

water layer was lower than that in the central water layer. By contrast, the vave,r in the

323

central water layer was lower than that in the layer near reactor wall or quartz sleeve.

324

With increasing Q, both v’ave,a and vave,r at the near-sleeve water layer increased,

325

implying that a higher Q could enhance convection between the near- and far-lamp

326

regions to increase the pollutant removal.

327

Figure 5

15



Page 16 of 33


by Li et al.

328

For the VUV/UV/chlorine process, a higher Q also promoted the mass transfer of

329

HOCl/OCl−, which could further enhance the reactive species formation. The

330

enhancement of Q on the EEO was larger than that for the VUV/UV process (Table 2).

331

Meanwhile, a higher chlorine dose could induce a lower EEO value because of the

332

stronger synergism of the VUV/UV/chlorine process. As a comparison, much higher

333

EEO values for the VUV/UV/H2O2 process were observed under similar test

334

conditions (Table S2). It is also noted that the pilot tests yielded obviously higher EEO

335

values in the presence of 5 mg L−1 HA for the VUV/UV/chlorine process. However,

336

as aforementioned, the DOC in the sand-filtered water of WTPs has a much less

337

reactivity toward chlorine and reactive species than the HA spiked in this study, and

338

meanwhile, its concentration usually ranges from 0.5 to 1.5 mg L−1 in most cases.

339

Therefore,

340

VUV/UV/chlorine process is applied to real water treatment.

considerably

lower

EEO

values

can

be

expected

when

the

341

Discussion on the Combinatorial Application of MVPS and Pilot

342

Photoreactor for VUV/UV-AOP studies. This study demonstrates that the

343

VUV/UV-AOPs (e.g., VUV/UV/chlorine) have a good potential on trace organic

344

pollutant removal from water. The pilot photoreactor allows an evaluation of the

345

practical performance of VUV/UV-AOPs quickly and intuitively; but it is not suitable

346

for fundamental rate constant determination. The MVPS provides an accurate and fast

347

determination of photochemical kinetic parameters because of its unique optical

348

construction allowing accurate measurements of UV and VUV fluences. However,

349

since the UV tube (or VUV/UV tube) cannot capture all photons emitted from the

16


Page 17 of 33


by Li et al.

350

light source, the MVPS is not suitable for the energy consumption calculation.

351

Therefore, the combinatorial application of the MVPS and the pilot photoreactor

352

provides a good approach for VUV/UV-AOP studies. In addition, although with the

353

similar irradiated 254 nm/185 nm photon ratio, because of the different constructions

354

of the MVPS and industrial reactors, one can suspect that if the special effects

355

obtained in MVPS (e.g., synergistic enhancement effect of VUV/UV/chlorine) can

356

happen in practical application. Table 2 shows that the EEO values of VUV/UV are

357

1.38 – 1.96 times higher than those of VUV/UV/chlorine at chlorine dose of 0.5 mg

358

L-1. It implies that the synergistic enhancement also occurred in the pilot reactor, and

359

validates the results from MVPS.

360

Discussion on the Potential Applications of VUV/UV/chlorine. In recent years,

361

trace organic pollutant removal has become an important issue in water treatment.

362

Ozonation has been used as an oxidation process of the trace organic pollutant in

363

many large water treatment plants followed by biological active activated carbon.

364

However, for small-scale water supply systems, the ozone application is usually

365

uneconomic. The VUV/UV/chlorine process involves two existing water treatment

366

methods (i.e., VUV/UV irradiation and chlorination). Moreover, the chlorine doses of

367

0.05 and 0.5 mg L−1 match those commonly used in water treatment plants. Therefore,

368

this new oxidation/disinfection technology can be conveniently applied to new and

369

existing small-scale water treatment systems only by installing a VUV/UV

370

photoreactor.

17




371

The formation of chlorinated by-products (CBPs) in SMN degradation by

372

VUV/UV/chlorine was evaluated by using ultra-performance liquid chromatography

373

quadrupole time-of-flight mass spectrometry (UPLC/Q-TOF-MS). A high initial

374

concentration of SMN (10 mg L−1) was applied. According to the chromatogram (in

375

Figure S9) and mass spectrum of SMN degradation in VUV/UV/chlorine process with

376

various photon fluences, five by-products were summarized in Table S3 and possible

377

degradation pathway was proposed in Figure S10.

378

Because the HO• and RCS contributed to the organic pollutant degradation in

379

VUV/UV/chlorine process, the principle HO• oxidative products were found,

380

including P1 (C6H10N3O, m/z 140), P2 (C6H10N3, m/z 124), P3 (m/z 295) and P4 (m/z

381

295). These products were also found during the SMN degradation by UV/H2O2 and

382

VUV/UV processes, whose dominant oxidant was HO•.14,30 With the RCS attack, a

383

CBP (P5, C12H12N4Cl, m/z 275) was found at the retention time of 1.762 min. The

384

possible reaction involved the SO2 elimination driven by direct photolysis and the

385

substitution of Cl on the amine. Nassar et al. also found this product during the

386

chlorination of SMN, and pointed out that the toxicity of this product is not higher

387

than that of SMN.31 Therefore, because of the low concentration of SMN in drinking

388

water and low toxicity of the principle CBP, SMN removal by VUV/UV/chlorine

389

could not introduce obvious toxicity increase as compared to the existing processes

390

(e.g., VUV/UV or UV/chlorine).

391

The nitrite formation has been paid attention during the irradiation of

392

low-wavelength UV ( 99%). b

Not detected.

26


Page 27 of 33


by Li et al.

Table 2. EEO Values Determined in the Pilot Photoreactor for SMN (0.05 mg L−1) Degradation by VUV/UV and VUV/UV/chlorine EEO (kWh m−3 order−1) Q

[chlorine]0

(m3 h−1)

(mg L−1)

a

VUV/UV

VUV/UV/chlorine

VUV/UV/chlorine

DI water

DI water

with 5 mg L−1 HA

1.02

0.50

0.137 ± 0.001

0.070 ± 0.005

1.433 ± 0.128

0.75

0.50

0.162 ± 0.001

0.094 ± 0.006

1.600 ± 0.114

0.42

0.05

−a

0.204 ± 0.008

−

0.42

0.50

0.220 ± 0.002

0.159 ± 0.007

2.505 ± 0.026

0.42

5.00

−

0.150 ± 0.006

−

Not detected.

27




Figure 1. Schematic illustration of the pilot VUV/UV photoreactor.

28


Page 29 of 33


by Li et al.

(a)

ln([SMN]/[SMN]o)

0

-4

-8

k' (m2 einstein-1) UV/chlorine, k'UV/Cl = 0.08 x 103 VUV/UV, k'VUV/UV = 4.00 x 103 VUV/UV/chlorine, k'VUV/UV/Cl = 6.76 x 103

(b)

ln([SMN]/[SMN]o)

0

-5

-10

k' (m2 einstein-1) UV/chlorine, k'UV/Cl = 0.51 x 103 VUV/UV, k'VUV/UV = 4.00 x 103 VUV/UV/chlorine, k'VUV/UV/Cl = 8.51 x 103 0.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7

Photon fluence (x 10-3 einstein m-2)

Figure 2. SMN degradation by UV/chlorine, VUV/UV and VUV/UV/chlorine with chlorine doses of 0.05 (a) and 0.5 (b) mg L−1. Conditions: [SMN]0 = 0.1 mg L−1, pH = 7.0.

29



Page 30 of 33


by Li et al.

10000

2

-1

k' (m einstein )

8000

6000

4000

2000

0

4

5

6

7

8

9

10

11

pH

Figure 3. Photo-fluence based rate constant (k′) as a function of pH for SMN degradation by VUV/UV/chlorine. Conditions: [SMN]0 = 0.1 mg L−1 , [chlorine]0 = 0.5 mg L−1. Each data point (k′) represents the linear regression result of six experimental data points (R2 > 99%).

30


Page 31 of 33


by Li et al.

0.0

(a)

ln([SMN]/[SMN]o)

-0.5 -1.0 -1.5

2

-1

k' (m einstein ) 3

k'UV/Cl = 0.03 x 10 , [chlorine]0 = 0.05 mg L

-2.0

3

k'UV/Cl = 0.10 x 10 , [chlorine]0 = 0.5 mg L -2.5

-1

-1

3

k'VUV/UV/Cl = 0.86 x 10 , [chlorine]0 = 0.05 mg L 3

k'VUV/UV/Cl = 0.94 x 10 , [chlorine]0 = 0.5 mg L

-3.0

-1

-1

ln([SMN/SMN]o)

0

(b)

-2

-4 k' (m2 einstein-1)

k'UV/Cl = 0.02 x 103, [chlorine]0 = 0.05 mg L-1

-6

k'UV/Cl = 0.44 x 103, [chlorine]0 = 0.5 mg L-1 k'VUV/UV/Cl = 2.85 x 103, [chlorine]0 = 0.05 mg L-1

-8

k'VUV/UV/Cl = 4.89 x 103, [chlorine]0 = 0.5 mg L-1 0.0

0.3

0.6

0.9

1.2

1.5

1.8

-3

2.1

2.4

2.7

-2

Photon fluence (x 10 einstein m )

Figure 4. SMN degradation by UV/chlorine and VUV/UV/chlorine at chlorine doses of 0.5 and 0.05 mg L−1 in MYR (a) and WTP (b) waters. Conditions: [SMN]0 = 0.1 mg L−1, pH = 7.0.

31



Page 32 of 33


by Li et al.

3

2.4 11.5 < r < 12.0 mm 12.5 < r < 13.5 mm 14.5 < r < 15.5 mm 16.5 < r < 17.5 mm

2.0

v'ave,a

1.6

12.0 < r < 12.5 mm 13.5 < r < 14.5 mm 15.5 < r < 16.5 mm

2

1.2

1

0.8

vave, r (x 10-3 m s-1)

Water layer

0.4

0

0.0

1.02

0.75

1.02

0.42

0.75

0.42

Q (m3 h-1) Figure 5. Estimated relative average axial velocity (v’ave,a) and average radial velocity (vave,r) values at various water layers and flow rates in the central cross-section of the pilot photoreactor.

32


Page 33 of 33



by Li et al.

TOC Art

Bench-scale test

Pilot-scale test

ln([SMN/SMNo])

0

-4

-8

3

2

-1

UV/chlorine k'UV/Cl = 0.080 x 10 m einstein 3 2 -1 VUV/UV k'VUV/UV = 4.000 x 10 m einstein 3 2 -1 VUV/UV/chlorine k'VUV/UV/Cl = 6.760 x 10 m einstein

0.0

0.3

0.6

0.9

1.2

1.5 -3

1.8

2.1

2.4

2.7

-2

Photon fluence (x 10 einstein m )

33


chlorine Process

Recommend Documents