Potential formation of mutagenicity by low pressure-UV/H2O2 during the treatment of

nitrate-rich source waters

S. Semitsoglou-Tsiapoua,b,† , M. R. Templetona, N. J. D. Grahama, S. Mandalc, L. Hernández

Lealb and J. C. Kruithofb

a. Department of Civil and Environmental Engineering, Imperial College London, South

Kensington Campus, SW7 2AZ, London, UK.

b. Wetsus, European centre of excellence for sustainable water technology, P.O. Box 1113,

8900 CC, Leeuwarden, the Netherlands.

c. Duisburg-Essen University, Universitätsstraße 2, 45117, Essen, Germany.

† Corresponding author: s.semitsoglou-tsiapou12@imperial.ac.uk, Oostergoweg 9, PO Box

1113, 8900 CC, Leeuwarden

1

1

2

3

4

5

6

7

8

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

LP-UV/H2O2 treatment of NOM-containing synthetic waters led to nitrite, nitrophenol and

measurable but not mutagenic Ames responses in the presence of nitrate, where the NOM

type affected the response levels.

Water Impact Statement

Despite the increasing application of UV advanced oxidation processes in water treatment,

few studies have considered the potential formation of nitrated by-products and potential

toxicity issues. This study showed that UV advanced oxidation of waters containing nitrate

and natural organic matter can produce nitrite and nitrophenols, but measurable mutagenicity

formation in Salmonella typhimurium was not significant compared to standard thresholds.

Introduction

Advanced oxidation processes (AOPs) have been increasingly incorporated for drinking

water treatment applications, due to the rise of recalcitrant micro-pollutants in the water, and

because of the effective non-selective degradation achieved1. More specifically, UV-based

AOPs (e.g. UV/O3, UV/H2O2, UV/TiO2, UV/HOCl) have been widely applied in the past

decade due to the combined effect of UV photolysis and hydroxyl radical oxidation for the

degradation of organic micro-pollutants2 (as well as for disinfection purposes). AOPs have

been shown to produce reaction products3, which depend greatly on the water matrix, the

process and the conditions applied. This work has focused specifically on the application of

the Low Pressure (LP)-UV/H2O2 process. UV/H2O2 treatment is becoming popular since it

2

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

has been shown in numerous previous studies to successfully treat various organic

contaminants, e.g. endocrine disrupting compounds (EDCs)4, polycyclic aromatic

hydrocarbons (PAHs)5-6, pharmaceuticals7 and pesticides8-9. It is a promising technology due

to the effective, non-selective degradation of micropollutants via OH-radical-assisted

oxidation with an additional contribution of UV photolysis, while disinfection of the water

takes place at the same time. Specifically, Low Pressure (LP) lamp applications exhibit

advantages over Medium Pressure (MP) applications due to higher energy efficiency, longer

lifetime, and minimum formation of by-products of concern, such as nitrite and bromate10.

The formation of nitrogenous reaction products via the photolysis of nitrate, followed by a

complex series of reactions, has been studied previously only to a limited extent11. In contrast,

the photolysis of nitrate has been studied extensively, especially for medium pressure (MP)-

UV wavelengths (λ < 240 nm) where its absorption is the highest, leading to peroxynitrite

(OONO-) formation as the main intermediate species. Peroxynitrite decomposition leads to

the production of nitro- and nitroso- radicals, with nitrate and nitrite as end products12-13. In

the presence of an organic matrix, the incorporation of inorganic nitrogen into the organic

matrix has been demonstrated14. The main mechanisms shown to take place during the

photolysis of nitrate or nitrite in the presence of low molecular weight organic molecules,

mainly aromatic substances such as benzene and phenol15-18, are hydroxylation, nitration and

nitrosation, the latter being enhanced in the absence of oxygen19. In these studies, the

formation of a variety of nitrated compounds was observed (e.g. 2-nitrophenol, 4-nitrophenol,

4-nitropyrocatechol, 4-nitrosophenol). Studies investigating the reaction product formation in

NOM-nitrate rich water by UV/H2O2 treatment are few; Martijn et al.17-18 reported the

formation of 2- and 4-nitrophenol and 4-nitrocatechol when irradiating phenol as surrogate

for NOM in the presence of nitrate with MP-UV, whereas Kolkman et al.20 found a variety of

nitrogen containing compounds by MP-UV photolysis and MP-UV/H2O2 treatment of

3

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

synthetic (nitrate-rich, with Pony Lake NOM) and full-scale water samples, respectively,

confirming the following three reaction products: 4-nitrophenol, 4-nitrocatechol, and 2-

methoxy-4,6-dinitrophenol.

Since some nitrated compounds are known to be toxic21 and as a result may pose a health risk

in drinking water, it is critical that their presence and potential contribution to the water’s

toxicity are investigated. For that reason, various bioassays have been developed, such as the

Ames assay, which is widely used as a standard screening method for the detection of

mutagenic compounds22. The Ames assay was selected, since regulators consider that, ‘‘a

substance that is mutagenic in the Salmonella typhimurium bacterium is more likely than not

to be a carcinogen in laboratory animals, and thus, by extension, present a risk of cancer to

humans’’ (U.S. Department of Health and Human Services, 2016)23. The test is advantageous

in terms of its easy application, rapid generation of results and low cost, thus rendering it a

useful tool for assessing potential carcinogenicity. It makes use of strains of the bacterium

Salmonella typhimurium that carry gene mutations that inhibit histidine synthesis; these

mutations can be reversed in the presence of mutagenic chemical compounds. The Ames II

assay used in this study is a modification of the standard Ames test; it uses a liquid

microplate format and much smaller sample volumes are required24. There are two commonly

applied Salmonella typhimurium strains, the frameshift strain TA98 and the base-pair

detecting strain TA100. It has been shown by previous studies that the TA98 strain (and

especially the TA98(-S9) combination, where (-S9) denotes the absence of a rat liver

metabolic activation system) is most responsive in samples treated with oxidation

processes17,25-28, whereas the TA100 strain (both with and without S9 addition) has exhibited

positive responses in cases of chlorinated samples29-32. Therefore, in this study, the strain

TA98, in the absence of a rat liver metabolic activation system (S9), was selected.

4

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

The mutagenicity of MP-UV/H2O2-treated water has been investigated previously and

showed a significant Ames assay response and the important role that the presence of nitrate

played in this response17-18, 20, 27, 33 with only one study giving negative Ames assay results

where natural quartz sleeves with a cut-off of UV light< 240 nm were utilized 34. Martijn et

al.17-18 showed in a preliminary risk assessment, by converting the Ames responses of full-

scale water samples treated by MP-UV and MP-UV/H2O2 into 4-nitroquinoline oxide (4-

NQO) equivalent concentrations, that there is reason for concern and further investigation is

required. Although in theory LP-UV and LP-UV/H2O2 processes are expected to cause little,

if any, mutagenicity formation, very few studies have examined this in detail, and these have

suggested that LP-UV-based treatment did not cause any mutagenicity formation33 or a very

weak response35. In order to provide further clarification of potential effects, this paper

summarises the results of recent research which has evaluated: a) the formation of

mutagenicity, nitrite and nitrophenol in laboratory tests using synthetic water containing

NOM and nitrate, and b) the formation of mutagenicity in samples from a full-scale drinking

water plant where LP-UV/H2O2 treatment was applied.

Methods

Chemicals

Sodium nitrate, hydrogen peroxide (30%), HPLC-grade acetone, ammonia (30%), methanol

and ethyl acetate (99.9%) were supplied by VWR (the Netherlands). Acetonitrile was

supplied by Merck (the Netherlands), formic acid (99%) by Boom (the Netherlands) and

fenoprofen by Sigma-Aldrich (the Netherlands). Laboratory-grade water (LGW) was

produced by a Milli-Q Advantage A10 system (Merck Millipore, Darmstadt, Germany).

Two types of reference natural organic matter were obtained from the IHSS (International

Humic Substances Society) as dry solid extracts: Suwannee River NOM (2R101N) and Pony

5

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

Lake NOM (1R109F). Suwannee River NOM is a widely used reference aquatic NOM, and

Pony Lake NOM is a reference fulvic acid-containing NOM. The nitrogen content for

Suwannee River NOM was 1.27 % (w/w), whereas for Pony Lake NOM it was 6.51 %

(w/w); the highest nitrogen content among the IHSS NOMs available. Their difference is

reflected by the SUVA254 (SUVA254=Abs254/DOC) values obtained experimentally, 3.6 and

1.6 L mg−1 m−1 for Suwannee and Pony Lake, respectively, showing that Suwannee NOM

consists of humic, highly aromatic and hydrophobic matter with high molecular weight

organic fractions, whereas Pony Lake is mostly hydrophilic non-humic (fulvic) matter. The

A254/A203 UV absorbance ratios36 were calculated as 0.42 and 0.32, respectively, suggesting

aromatic rings highly substituted with hydroxyl, carbonyl, ester and carboxyl groups for

Suwannee River NOM, and rings predominantly substituted with aliphatic functional groups

for Pony Lake NOM.

Phenomenex Strata-X Polymeric Solid Phase Extraction (SPE) columns (200 mg Oasis HLB

3 mL glass cartridges) were supplied by Phenomenex (the Netherlands).

Stock solutions of the two types of NOM and NaNO3 were prepared in MilliQ water, and test

solutions were produced from the required amounts of the stock solutions and hydrogen

peroxide solution.

UV collimated beam experiments

UV exposure experiments were carried out with a bench scale collimated beam apparatus,

equipped with a 25-Watt low pressure (LP) mercury arc discharge lamp without a lamp

sleeve. The emission spectrum of the UV lamp, obtained from Trojan UV Technologies

(London, Ontario, Canada) consisted of a strong, almost exclusive emission at 254 nm

(Figure S.1, ESI). The experimental conditions for the LP-UV collimated beam experiments

are given elsewhere37. The exposure time for the desired UV fluences was calculated using a

fluence calculation spreadsheet based on Bolton and Linden38. According to this protocol, the

6

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

UV fluence applied, and consequently the irradiation time, were adjusted for the absorbance

of each solution at 254 nm.

Regarding the study of mutagenicity formation by LP-UV/H2O2 treatment, experiments were

performed in laboratory-grade water spiked with either Suwannee River NOM or Pony Lake

NOM (both at 4 mg/L, which corresponded to 2 and 2.1 mg C/L as DOC for Suwannee River

and Pony Lake NOM, respectively) in the presence of nitrate (50 mg/L). The UV fluences

applied were 0 (no irradiation), 1500 and 2000 mJ/cm2, corresponding to the upper-end of the

fluence range that is commonly applied (1500 mJ/cm2) and worst case (2000 mJ/cm2) UV

fluences, with a peroxide dose of 15 mg/L. All experiments were performed in duplicate and

at room temperature (23-25 ˚C).

Analytical methods

Nitrite and nitrate ions were quantified by ion chromatography (detection limits of 0.05 and

0.1 mg/L, respectively), using a Metrohm IC Compact 761 ion chromatograph (IC) equipped

with a Metrohm Metrosep A Supp 5 (150/4.0 mm) column, a Metrohm Metrosep A Supp 4/5

Guard pre-column and a conductivity detector.

The analysis for the detection of nitrophenols was performed by a method involving

compound extraction using polymeric SPE tubes (Phenomenex Strata-X Polymeric 200 mg /

3 ml SPE extraction tubes), followed by liquid chromatography tandem mass spectrometry

(LC-MS/MS). For the SPE extraction, acetone/ethyl acetate (1:1), methanol and acidified

blank sample (pH=2) were used consecutively for the activation of the SPE cartridges. The

cartridges were then loaded with 0.5 L of acidified sample. Elution was performed with 7.5

mL of acetone/ethyl acetate (1:1). A volume of 50 µL of the internal standard solution

(fenoprofen in methanol) was added to the eluates, which were further evaporated with a

nitrogen stream until the volume reached 0.2 mL. Finally, 0.8 mL of ultrapure water was

added before analysis.

7

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

For the negative ion electrospray LC-MS/MS analysis, an Agilent 6410 QQQ Mass Analyzer

with electrospray ion source was used. A Phenomenex Gemini Phenyl-hexyl column

(150mm*3mm, 3µm particle size) was employed and equipped with an appropriate guard

column for separation. The mobile phases used were A: 2.5 L Milli-Q water with 0.75 mL

formic acid (99%) and 1.75 mL ammonia (30%), and B: 2.5 L acetonitrile. The flow rate was

0.6 mL/min. As internal standard, fenoprofen was used. The compounds were measured with

specific QQQ transitions. For instrument control and data analysis Agilent MassHunter

software was used. The quantitation limit (QL) and detection limit (DL) of the method were

0.02 and 0.007 μg/L, respectively.

NOM characterisation was performed by Liquid Chromatography-Organic Carbon Detection

(LC-OCD) (Model 8, DOC-LABOR, Karlsruhe, Germany), equipped with both DOC and

DON detection, an organic carbon detector (NDIR), an organic nitrogen detector (UV 220

nm), as well as a UV detector (254 nm), all integrated within the LC-OCD system. The

column used was a Toyopearl HW-50S (30μm, 250 mL) and a phosphate buffer was used as

eluent. Data analysis was performed with DOC-LABOR software (ChromLog version).

H2O2 concentrations were measured using the triiodide method39.

All experiments were performed in duplicate and statistical t-test comparisons were

conducted with a 95% confidence interval.

Water treatment works sampling

Samples were collected from a selected drinking water treatment works (WTW) in the UK,

incorporating LP-UV/H2O2 technology and treating surface water from a lowland river

known for its poor water quality caused by the presence of industrial discharges, pesticides

and high nitrate concentrations. The WTW comprises river bank side storage where surface

water is collected from the adjacent river, followed in sequence by roughing granular

activated carbon (GAC) filtration, submerged ultrafiltration, LP-UV/H2O2 oxidation,

8

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

polishing post GAC filtration, UV disinfection and residual chlorine addition. The LP-

UV/H2O2 treatment process makes use of four UV reactors, each containing 96 UV 1kW

lamps, and a H2O2 dosing system. Two separate samplings (referred to as (a) and (b)) took

place in February and April 2016, to cover expected differences in water composition and the

maximum design UV/H2O2 conditions that were applied as part of testing the efficiency of the

system in a worst-case scenario ((UV fluences of 2000 and 1750 mJ/cm2 for sampling (a) and

(b), respectively, and a H2O2 dose of 40 mg/L in both cases). These doses, along with water

quality and operational parameters for the WTW sampling events can be found in Table 1.

Water samples were collected from the inlet (only for sampling (b)), pre-AOP, post-AOP and

post-GAC treatment steps, to assess nitrite formation, NOM fate and potential mutagenicity

in terms of Ames assay response during treatment.

Ames II Mutagenicity assay

Both the synthetic and full-scale water samples were subjected to the Ames II Mutagenicity

assay (the TA98(-S9) combination) in order to evaluate their mutagenic potential to induce

reverse mutations in Salmonella typhimurium. All water samples were extracted by SPE

according to the method described by Heringa et al.27. The extracts were 20,000x

concentrated after SPE extraction, diluted 25x in the Ames II assay (i.e. relative enrichment

factor was 800). The Ames II assay was performed according to the Xenometrix protocol40. A

blank (SPA Reine = spa bottled water) was treated identically as a method control. Solvent

control (100% DMSO), and positive controls (2-nitrofluorene + 4-nitroquinoline N-oxide)

were included. In total, 9 replicates in the control conditions (method and solvent controls)

and 3 replicates for each extract (i.e. sample) were prepared. Per replicate, 48 wells are used,

therefore the mean number of positive wells was expressed as a percentage (%) of the 48.

Both the sample extraction and the Ames assay were performed by VITO laboratories in

Belgium.

9

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

In order to interpret the findings in terms of positive or negative mutagenicity, the number-

fold inductions over the solvent control (ratio of the mean number of positive wells for the

test item divided by the mean number of positive wells for the solvent control) and the

number-fold inductions over the baseline value (ratio of the mean number of positive wells

for the test item divided by the baseline value for the solvent control) were calculated. The

baseline value represented the mean number of positive wells in solvent control conditions +

1 SD. A Student’s t-test (1-sided, unpaired) was also carried out to determine significance at

the p ≤ 0.05 level between the range of data for the solvent control and the data for the test

substances. The raw data and calculated values are given in Table S.1 (ESI).

Number-fold inductions in revertant numbers over the solvent control are considered as

positive if > 3.0, whereas number-fold inductions in revertant numbers over the baseline

value are positive if > 2.0. A sample (or compound) that shows a clear number-fold increase

> 2.0 (baseline) and significant difference (p<0.05 in the t-test), is classified as mutagenic.

When the number-fold induction is below these values the test substances are not mutagenic

towards the Salmonella typhimurium strain TA98.

Results and Discussion

Ames II assay of synthetic water samples

The Ames II assay was performed to assess the mutagenicity of the treated water, using two

different types of NOM, Pony Lake and Suwannee River NOM, in the absence and presence

of nitrate (50 mg/L) under the following conditions: UV fluences of 0 (no irradiation), 1500

and 2000 mJ/cm2 and a dose of 15 mg/L H2O2. The results for the water samples, including

the method and solvent control, are given in Figure 1 (a-b).

Comparing the combinations of 0 UV+H2O2+NO3 and 2000 UV+H2O2+NO3 the Ames II

assay results for Suwannee River NOM (Figure 1a) were not statistically different. In

10

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

contrast, for Pony Lake NOM, the 2000 UV+H2O2+NO3 combination showed a significant

increase in positive wells levels (Figure 1b). Therefore, it can be concluded that LP-UV

photolysis and/or hydroxyl radical oxidation was responsible for the increase in the number

of positive wells. A statistically significant increase was also observed for the 2000

UV+H2O2+NO3 combination compared to the 2000 UV+H2O2 (no NO3), indicating that NO3

photolysis played a major role in the increased number of positive wells. This finding agrees

with the findings from Martijn et al. (2014, 2015)17-18 treating samples containing Pony Lake

NOM with and without the presence of nitrate with either MP-UV photolysis or MP-

UV/H2O2 treatment, producing an increase in the Ames II assay response only when nitrate

was present. Overall, significant variations in the Ames II assay response were observed in

Pony Lake NOM samples in the presence of nitrate after LP-UV/H2O2 treatment.

Nevertheless, the levels produced (around 10 positive wells) with a high nitrate concentration

(50 mg/L) and high UV fluence (2000 mJ/cm2) were only half of those observed by Martijn

et al. (2014)17 (around 20 positive wells) when MP-UV/H2O2 treatment was applied with a

lower UV fluence and H2O2 dose (560 mJ/cm2 with 6 mg/L H2O2).

In order to interpret the results in this work in terms of mutagenicity with the Ames II assay,

the number-fold inductions over the solvent control and the number-fold inductions over the

baseline value (explained in the Methods Section) were calculated. Figure 2 shows that all

number-fold inductions over the solvent control were below 3 and all number-fold inductions

over the baseline value were below 2, indicating that none of the SPE extracts of the samples

could be classified as mutagenic towards the Salmonella typhimurium strain TA98. It should

be noted that negative results are not conclusive of the total absence of mutagenic potency, as

other mutagenic mechanisms cannot be excluded.

Based on the calculated SUVA values (3.6 and 1.6 L mg−1 m−1 for Suwannee River and Pony

Lake NOM, respectively), Suwannee River NOM is expected to have a more aromatic nature

11

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

compared to Pony Lake NOM, and consequently be more prone to OH-radical attack under

UV/H2O2 treatment at the electron-rich sites41. Nevertheless, this would be the case for low

molecular weight organic compounds42, whereas based on the high SUVA value it is

suggested that Suwannee River NOM contains high molecular weight organic fractions.

According to literature values, the reaction rate constant of these two NOMs with OH-

radicals are reported to be 3.3 108 M-1 s-1 for Suwannee River NOM43 and 2.03 108 M-1 s-1 for

Pony Lake NOM44 (by competition kinetics), so the difference in reactivity is small. If,

however, the degradation of Suwannee River NOM takes place to a greater extent than Pony

Lake NOM, and mutagenic products are formed as intermediates, the degradation of those

intermediates could also take place, therefore lowering the final Ames levels.

The difference between the Ames levels obtained by the two NOMs was not statistically

significant; nevertheless, the role of photolysis as well as the role of the presence of nitrate

were apparent only for Pony Lake NOM under the conditions applied, as already explained

earlier. The incorporation of the nitrate-nitrogen into the organic matrix and the formation on

nitrated/nitrosated compounds by MP-UV treatment has been demonstrated for both

Suwannee River14 and Pony Lake NOM17,20. This could be the reason why mutagenicity levels

(even though low) were measured for both NOMs. It can be suggested that this N-

incorporation is also affected by the nitrogen content of the NOM, especially when it comes

to N-nitrosation45; since this content is 5 times higher for Pony Lake NOM (1.27 % (w/w) and

6.51 % (w/w) for Suwannee River and Pony Lake NOM, respectively), nitrosation could have

been favoured, giving rise to more, potentially mutagenic, compounds. Additionally, the

more aliphatic composition of Pony Lake NOM could have contributed to greater nitrosation,

since, according to Thorn and Cox45, the greater the aliphatic carbon content, the greater the

concentration of activated methylene and methyl carbons that are available for nitrosation and

subsequent rearrangement. Although in the Thorn and Cox study45 the NOMs were treated

12

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

with nitric acid and not a nitrate salt, they state that some detected N-compounds were also

observed when sodium nitrate was used.

Nitrite and nitrophenol formation

While no nitrite was formed in dark samples (i.e. no UV irradiation), the nitrite

concentrations obtained for the irradiated samples (1500/15 and 2000/15, in mJ/cm2 / mg/L)

for both NOMs were 0.08-0.09 mg/L (Figure S.2, ESI), which were very close to the 0.1

mg/L EU regulatory limit for nitrite. However, this limit might be exceeded if the H 2O2 dose

is increased46. It should be noted that conflicting results have been reported regarding the role

of hydrogen peroxide. Thus, Lu et al.47, utilizing a LP lamp at pH = 9.5 for a concentration of

10 mg/L NO3--N (44 mg/L NO3

-), found that 0.8 mg/L NO2- was produced both in the absence

of H2O2 and with 10 mg/L H2O2, while in contrast Sharpless et al.48 reported 0.16 mg/L NO2-

formed by MP-UV photolysis of water containing 10 mg/L of nitrate with a UV fluence of

150 mJ/cm2 and 0.19 mg/L NO2- with the addition of 10 mg/L H2O2, thereby showing that

nitrite formation by MP-UV can be enhanced by H2O2 addition. The nitrite formation with

and without H2O2, under the same UV fluences as in this work, and both in the presence and

absence of NOM, can be found in our previous study46. It was observed that, a) the H2O2

concentration (up to 50 mg/L) and nitrite yield (0.05-0.13 mg/L) were directly proportional

when 50 mg/L nitrate was used in NOM-free water, and b) the presence of NOM (either

Suwannee River, Nordic Lake or Pony Lake NOM) increases the formation of nitrite for all

UV/H2O2 dose combinations and NOM concentrations, compared to UV photolysis and

NOM-free waters. Specifically for Pony Lake NOM (for comparison reasons) the nitrite

concentrations under the fluences of 1500 and 2000 mJ/cm2 without H2O2 were 0.11 and 0.13

mg NO3- /L, respectively, compared to the 0.09 mg NO3

- /L value for both 1500/15 and

2000/15 (mJ/cm2 / mg/L) combinations.

13

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

In addition, the presence of nitrophenols was investigated (only with Pony Lake NOM) since

their formation has been observed previously (as mentioned in the Introduction), due to the

reaction of phenol with UV photolysis intermediates of NO3. Martijn et al. (2014)17 used

Pony Lake NOM in the presence and absence of nitrate, hypothesizing the Ames response

was caused by the incorporation of inorganic nitrogen into the organic matrix by MP-UV and

MP-UV/H2O2 treatment, and subsequently observed the formation of nitrophenols by MP-UV

photolysis when phenol was used as a model compound for NOM. Therefore, it was

considered worthwhile to investigate any nitrophenol formation in the tests in this study

stemming from reactions in the NOM-NO3 system, induced by LP-UV photolysis (i.e. 254

nm).

The LC-MS/MS method was used to analyse for the following nitrophenols: 2-nitrophenol, 4-

nitrophenol and a combination of both. All samples yielded trace amounts of mono-

nitrophenol (0.014-0.046 μg/L), without the possibility to distinguish between 2-nitrophenol

and 4-nitrophenol, since the analytical method used only determines the total mono-

nitrophenol concentration. The experiments were performed in duplicate and the results are

shown in Figure 3. With no UV irradiation, 0.01 μg/L of nitrophenol was found, which was

very close to the detection limit (0.007 μg/L) and lower than the quantitation limit (0.02

μg/L). For the two UV/H2O2 combinations, a significant nitrophenol concentration (~0.04

μg/L) above the quantitation limit was observed.

Nitrophenol production is expected since phenolic groups are one of the aromatic structures

present either in the NOM or from the degradation products of NOM molecules by oxidation.

The results agree with Kolkman et al. (2015)20 who found a variety of nitrogen-containing

compounds, including 4-nitrophenol, when MP-UV photolysis was applied in samples with

Pony Lake NOM, as well as in full-scale water samples treated with MP-UV/H2O2. The two

phenols detected, 2- and 4-nitrophenol, were also found by Martijn et al. (2014)17 along with

14

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

4-nitrocatechol, when irradiating samples containing phenol and NO3 with a MP lamp. In our

experiments with the LP lamp, a significant concentration of nitrophenol was found for both

UV fluences (i.e. 1500 and 2000 mJ/cm2). The measured nitrophenol concentration was less

for the 2000 mJ/cm2 than for the 1500 mJ/cm2 fluence (0.037 and 0.046 μg/L, respectively),

although the difference was not statistically significant (Figure 3). This finding agrees with

Martijn et al. (2014)17 who showed an initial increase in nitrophenol concentration as a

function of the irradiation time, followed by a decrease at longer irradiation times, attributed

to oxidation of the previously produced nitrophenols.

Ames II assay of full-scale water samples

Samples collected from the UK drinking water treatment plant after various treatment steps

on two days in different months (sampling (a) was in February and sampling (b) in April)

were also analysed with the Ames II assay. On both sampling days, duplicate samples were

collected. The water quality and operational parameters, relevant for the LP-UV/H2O2

treatment of the WTW, for the two samplings are given in Table 1.

The number of positive wells from the Ames II assay for the water treatment works samples

are shown in Figure 4. Sampling (a) showed a small but significant decrease in Ames II assay

response as the water passed through the AOP step. However, in sampling (b) (when the inlet

water was also analysed) the number of positive wells significantly decreased from the inlet

to the pre-AOP step (the submerged ultrafiltration), but remained unchanged for the AOP

step and significantly increased after passing the GAC filters. Although all samples produced

low levels of positive wells compared to the positive control, the two sampling sets differed

significantly for the pre-AOP and post-AOP steps, indicating that the Ames II assay response

was indeed influenced by the different water characteristics and/or treatment conditions of the

two periods.

15

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

The increased response after the post-GAC filtration found during the second sampling could

be attributed to either: a) toxic chemical compound formation on the GAC filter via

biological activity, or b) pathogen colonization of the GAC filter, e.g. by Salmonella

bacterium, both of which would induce undesired biological activity on the GAC filters.

Regarding the first factor, examples of microbiologically assisted processes like methylation

of mercury, hydroxylamine formation, pesticide-related molecule conversion to toxic

metabolites, nitrosamine formation with nitrite as a precursor and co-metabolism of

refractory compounds, have been observed in laboratory models or natural water bodies.

Nevertheless, for drinking water treatment production, the possibility of bio-degradation by-

products has been demonstrated only after O3-BAC treatment49. Regarding the second factor,

according to Camper et al.50, pathogens such as Salmonella can colonize and persist in the

carbon bed, especially on virgin GAC filters. The presence of Salmonella could increase the

response of the Ames test, which makes use of strains of the bacterium Salmonella

typhimurium, giving a false positive. The likelihood, though, of this speculation is also very

low since the water samples would have a reduced microbial load by having been enriched by

the SPE method as well as having been subjected to high UV fluences before entering the

GAC filters. The most probable explanation is an experimental error during the second

sampling.

Figure 5 depicts the number-fold increases obtained for the WTW samples. All number-fold

increases over the solvent control were < 3.0, and all number-fold increases over the baseline

value were < 2.0, indicating that none of the SPE extracts of the samples could be identified

as mutagenic towards the Salmonella typhimurium strain TA98.

In an attempt to explain the differences in the Ames II assay responses between the two

samplings (Figure 4), all samples were analysed for nitrite concentrations, and LC-OCD

fractionation was performed to assess any changes in the NOM content due to the pre-

16

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

treatment (up to UV/H2O2 treatment), the UV/H2O2 treatment itself and post GAC filtration.

The nitrite concentrations for all samples were at the detection limit (0.05 mg/L). The results

from the LC-OCD fractionation are given in Figure 6. For both samplings, the decrease in the

humics-C and humics-N content through all the treatment stages (inlet, pre-AOP, post-AOP

and post-GAC steps) was statistically significant (except for the post-AOP/post-GAC for the

humics-N in sampling (a)). Comparison for either humics-C or humics-N content between the

two samplings showed that the differences were significant for all the treatment steps except

for the GAC filtration step, for the humics-N, suggesting that the NOM composition

variations were period-dependent.

The SUVA254 values obtained (SUVA254 > 4 L/m mg) for both samplings (Table 1) suggested

that pre-AOP the water contained humic, aromatic and hydrophobic matter with high

molecular weight DOM fractions. The SUVA values decreased post-AOP, suggesting that the

aromatic compounds present in the water pre-AOP (which comprise a major fraction of the

humic content51, the NOM fraction most susceptible to photolysis and oxidation via

UV/H2O2), underwent degradation to lower molecular weight (LMW) compounds which are

more likely to have sustained loss of aromaticity (e.g. loss of cyclicity, conjugated π-system).

This hypothesis was supported by the decrease in humic content, as well as by the increase in

LMW acids evident from the LC-OCD fractionation (see Figure 6), which were found to be

statistically significant pre- and post-AOP for both samplings. Nevertheless, these results

only suggest a link between the fate of NOM in the water and the Ames II assay responses

obtained, without proving a direct cause and effect relationship.

MP vs LP-based treatment

The findings from the work of Martijn52 were compared to ours, since the procedure followed

for the Ames II assay was identical in both studies and the concentration factor (cf) of the

samples from the SPE extraction before the Ames II assay (i.e. 20,000 cf) was also the same

17

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

for both studies. Martijn52 measured the Ames II assay response of synthetic water with Pony

Lake NOM (2.5 mg C/L) after MP-UV/H2O2 treatment (600 mJ/cm2 and 6 mg/L H2O2) and

reported a yield of 25 positive wells (Figure 7.252) in the presence of nitrate (12 mg/L); in our

case where Pony Lake NOM (2.1 mg C/L) was treated by LP-UV/H2O2 (1500 mJ/cm2 and 15

mg/L H2O2) the number of positive wells was 3 times lower (8 positive wells), even though

the initial nitrate concentration was 4 times higher (50 mg/L). Applying a maximum fluence

of 2000 mJ/cm2, 4 and 10 positive wells were obtained in the absence and presence of nitrate,

respectively. Therefore, even though the Ames II assay responses obtained in this work were

much lower than the ones obtained by Martijn52 the Ames II assay response in nitrate-rich

water was still significant.

A distinction between significant and positive results (in terms of testing by the Ames assay)

should be kept in mind here; a detectable response is not necessarily considered significant.

For example, an organic compound in levels high enough to be detected may be below the

quantitation limit, therefore it can be considered not significant, and no sound conclusions

can be drawn from this finding. In the example of the Ames responses, the levels obtained

were detectable and statistically significant from one another in many cases; nevertheless,

without reaching the mutagenic level defined by the Ames protocol, the responses are not

considered positive in that respect.

Illustrating the significant role of nitrate for both MP and LP-UV processes, the Ames II

assay response as a function of nitrite formation was shown by Martijn52 (Figure 5.3).

Plotting our nitrite data (0.08-0.09 mg NO3-/L) obtained from the synthetic Pony Lake NOM

in Figure 7, gave an Ames II assay response of 5.5 positive wells, which is within the range

of the Ames II assay response we obtained for Pony Lake NOM for the 1500 mJ/cm 2 (8 ±

3.46 positive wells) (Figure 1b). Although different treatment processes (MP-UV vs LP-

UV/H2O2) and different nitrate contents (12 mg/L vs 50 mg/L) were involved, the overlap

18

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

observed supports the earlier indication of the impact of UV photolysis of nitrate on both the

nitrite formation and the Ames II assay response (Martijn et al. 2014)17, and is a confirmation

of the clear relation between the two (nitrite and Ames II assay response).

Based on the observations of MP-UV processes, it became clear that due to the significant

Ames responses observed, a risk assessment was required. Martijn et al.53 performed a

preliminary risk assessment by converting the Ames test responses into 4-nitroquinoline

oxide (4-NQO) equivalent concentrations, in order to obtain a risk indication via the Margin

of Exposure (MOE) approach54. They found that the 4-NQO equivalent concentrations

exceeded the Estimated Daily Intake (EDI), associated with a negligible risk, indicating a

concern of the water quality, should it be distributed “as drinking water without further post

treatment”. Such an approach was not required for this work, since it became evident from

our findings that the LP-UV/H2O2 treatment gave little, if any, reason for concern from a

health-related aspect.

Conclusions

From the current work, it can be concluded that by LP-UV/H2O2 treatment of nitrate-rich

water, nitrite and mono-nitrophenol formation may be observed. The principal two mono-

nitrophenols, 2- and 4-nitrophenol, reported in previous research, were also detected in this

work. These nitrophenols are known to have higher toxicity to organisms than the parent

compound, phenol, i.e. the oral LD50 values in mice are 0.3, 1.30-2.08 and 0.38-0.47 g/kg for

phenol, 2- and 4-nitrophenol, respectively55. However, under the conditions applied in this

research study, the concentrations of these phenols were in the range of trace levels and

would not be expected to cause health-related issues. Moreover, previous studies of 2- and 4-

nitrophenol have shown that no mutagenicity was observed when different variations of the

Ames II assay were applied56.

19

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

These nitrated organics are the result of the reaction of intermediate radicals (formed by LP-

UV photolysis of nitrate) with the organic matrix of the water samples. Since the nitrate

photolysis is much more enhanced in the MP region due to its absorption spectrum,

especially at wavelengths below 240 nm, the formation of radical species as well as nitrite

(the end product of a complex series of photolysis reactions) is correspondingly enhanced,

compared to the case of LP-UV (254 nm). Therefore, as expected, the incorporation of

inorganic nitrogen into the organic matrix and consequently the reaction product formation

and mutagenicity response are lower in LP-UV applications, giving little reason for concern.

From the present work, it can be concluded that LP-UV/H2O2 treatment is not expected to

produce significant mutagenic activity as shown by the applied Ames II assay results

(although other mutagenic mechanisms cannot be excluded), even when high nitrate

concentrations are present and high UV fluences and H2O2 doses are applied. Nevertheless,

case-specific studies should be conducted since the nitrite levels produced by the LP-

UV/H2O2 treatment are not always negligible (0.08-0.09 mg NO2-/L in our case) and the

conjunctive effect of all factors contributing to health effects is usually complex.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was performed in the cooperation framework of Wetsus, European Centre of

Excellence for Sustainable Water Technology (www.wetsus.eu). Wetsus is co-funded by the

Dutch Ministry of Economic Affairs and Ministry of Infrastructure and Environment, the

Province of Fryslân and the Northern Netherlands Provinces. The authors wish to thank the

participants of the ‘Priority Compounds’ research theme for the fruitful discussions, and

20

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

especially Dr. Bram Martijn for his guidance and advice in the experimental design and

relating the findings of the two studies. The authors would also like to thank Trojan

Technologies Inc. for supplying the low-pressure UV lamp and VITO laboratories for

performing the Ames II assays. The authors also gratefully acknowledge the financial support

of Anglian Water Services Ltd.

References

1 S. Vilhunen and M. Sillanpää, Rev Environ Sci Biotechnol, 2010, 9, 323.2 W. Yang, H. Zhou and N. Cicek, Crit. Rev. Environ. Sci. Technol., 2014, 44, 1443.3 S. D. Richardson, M. J. Plewa, E. D. Wagner, R. Schoeny and D.M. DeMarini, Mutat.

Res., 2007, 636, 178.4 E.J. Rosenfeldt and K.G. Linden, Environ. Sci. Technol., 2004, 38, 5476. 5 F.J. Beltrán, G. Ovejero and J. Rivas, Ind. Eng. Chem. Res., 1996, 35, 883. 6 S. Ledakowicz, J.S. Miller and D. Olejnik, Int. J. Photoenergy, 1999, 1, 1. 7 K. Ikehata, N.J. Naghashkar and M.G. El-Din, Ozone Sci. Eng., 2006, 27, 83. 8 H.D. Burrows, L.M. Canle, J.A. Santaballa and S. Steenken, J. Photochem. Photobiol.

B Biol., 2002, 67, 71. 9 E. Kowalska, M. Janczarek, J. Hupka and M. Grynkiewicz, Water Sci. Technol., 2004,

49, 261.10 U. von Gunten and Y. Oliveras, Environ. Sci. Technol., 1998, 32, 63. 11 T. Bond, M.R. Templeton and N. Graham, J. Hazard. Mater., 2012, 235−236, 1.12 J. Mack and J.R. Bolton, J. Photochem. Photobiol., A, 1999, 128, 1.13 S. Goldstein and J. Rabani, J. Am. Chem. Soc., 2007, 129, 10597.14 K.A. Thorn and L.G. Cox, J. Environ. Qual., 2012, 41, 865-881.15 J. Dzengel, J. Theurich and D.W. Bahnemann. Formation of Nitroaromatic Compounds

in Advanced Oxidation Processes: Photolysis versus Photocatalysis. Environ. Sci. Technol., 1998, 33, 294.

16 D. Vione, V. Maurino, C. Minero, M. Lucchiari and E. Pelizzetti, Chemosphere, 2004, 56, 1049.

17 A.J. Martijn, M.G. Boersma, J.M. Vervoort, I.M.C.M. Rietjens and J.C. Kruithof, Desalin. Water Treat., 2014, 52, 6275–6281.

18 A.J. Martijn, J.C. Kruithof, R.A.M. Hughes, R.A. Mastan, A.R. Van Rompay and J.P. Malley Jr., J. - Am. Water Works Assoc., 2015, 107, 301.

19 F. Machado and P. Boule, J. Photochem. Photobiol., A, 1995, 86, 73.20 Kolkman, B.J. Martijn, D. Vughs, K.A. Baken and A.P. van Wezel, Environ. Sci.

Technol., 2015, 49, 4458.

21

495

496

497

498

499

500

501

502

503

504505506507508509510511512513514515516517518519520521522523524525526527528529530531532

21 H.S. Rosenkranz and R. Mermelstein, J. Environ. Sci. Health, Part C: Environ. Carcinog. Rev., 1985, 3, 221.

22 B.N. Ames, E.G. Gurney, J.A. Miller and H. Bartsch, Proceedings of the National Academy of Sciences of the United States of America, 1972, 69, 3128.

23 National Toxicology Program, http://ntp.niehs.nih.gov/testing/types/genetic/invitro/sa/index.html , (accessed September 2017).

24 S. Flückiger-Isler, M. Baumeister, K. Braun, V. Gervais, N. Hasler-Nguyen, R. Reimann, J. Van Gompel, H.G. Wunderlich and G. Engelhardt, Mutat. Res., Genet. Toxicol. Environ. Mutagen., 2004, 558, 181.

25 F. van Hoof, J.G. Janssens and H. van Dijck, Chemosphere, 1985, 14, 501.26 M.A. van der Gaag, J.C. Kruithof and L.M. Puijker, Sci. Total Environ., 1986, 47, 137.27 M.B. Heringa, D.J.H. Harmsen, E.F. Beerendonk, A.A. Reus, C.A.M. Krul, D.H. Metz

and G.F. IJpelaar, Water Res., 2011, 45, 366.28 A.J. Martijn and J.C. Kruithof, Ozone Sci. Eng., 2012, 34, 92. 29 C.F. van Kreijl, H.J. Kool, M. de Vries, H.J. van Kranen and E. de Greef, Sci. Total

Environ., 1980, 15, 137.30 R.F. Christman, K. Kronberg, R. Singh, L.M. Ball and J.D. Johnson. Identification of

mutagenic by-products from aquatic humic chlorination. Department of Environmental Sciences and Engineering University of North Carolina, 1991.

31 X. Lv, Y. Lu, X. Yang, X. Dong, K. Ma, S. Xiao, Y. Wang and F. Tang, Sci. Rep., 2015, 5, 1.

32 T. Manasfi, M. De Méo, B. Coulomb, C. Di Giorgio and J.-L. Boudenne, presented in part at the 6th International Conference on Swimming Pool and Spa, Amsterdam, March, 2015.

33 R.C. Hofman-Caris, D.J. Harmsen, L. Puijker, K.A. Baken, B.A. Wols, E.F. Beerendonk and L.L. Keltjens, Water Res., 2015, 74, 191.

34 E.J.M. Penders, A.J. Martijn, A. Spenkelink, G.M. Alink, I. Rietjens and W. Hoogenboezem, J. Water Supply: Res. Technol.--AQUA, 2012, 61, 435.

35 G.F. IJpelaar, A.J. van der Veer, G.J. Medema and J.C. Kruithof, J. Environ. Eng. Sci., 2005, 4, S51.

36 G.V. Korshin, C-W Li and M.M. Benjamin, Water Res., 1997, 31, 1787.37 S. Semitsoglou-Tsiapou, M.R. Templeton, N.J.D. Graham, L. Hernández Leal, B. J.

Martijn, A. Royce, J.C. Kruithof, Water Res., 2016(a), 91, 285.38 J.R. Bolton and K.G. Linden, J. Environ. Eng., 2003, 129, 209.39 N.V. Klassen, D. Marchington and H.C.E. McGowan, Anal. Chem., 1994, 66, 2921.40 S. Flückiger-Isler, M. Baumeister, K. Braun, V. Gervais, N. Hasler-Nguyen, R.

Reimann, J. Van Gompel, H.-G. Wunderlich, G. Engelhardt, Mutat. Res., 2012, 558, 181.

41 D. Minakata, K. Li, P. Westerhoff and J. Crittenden, Environ. Sci. Technol., 2009, 43, 6220.

42 G. McKay, J.L. Kleinman, K.M. Johnston, M.M. Dong. F.L. Rosario-Ortiz and S.P. Mezyk, J Soils Sediments, 2014, 14, 298.

43 Y. Ahn, D. Lee, M. Kwon, I. Choi, S-N. Nam and J-W. Kang, Chemosphere, 2017, 184, 960.

22

533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577

44 J. E. Donham, E. J. Rosenfeldt and K. R. Wigginton, Environ Sci Process Impacts, 2014, 16, 764.

45 K.A. Thorn and L.G. Cox, PLoS ONE, 2016, 11, 1/20. 46 S. Semitsoglou-Tsiapou, A. Mous, M.R. Templeton, N.J.D. Graham, L. Hernández Leal

and J.C. Kruithof, Water Res., 2016(b), 106, 312.47 N. Lu, N.Y. Gao, Y. Deng and Q.S. Li, Water Sci. Technol., 2009, 60, 1393. 48 C.M. Sharpless, M.A. Page and K.G. Linden, Water Res., 2003, 37, 4730. 49 P. Jin, X. Jin, X. Wang, Y. Feng and X.C. Wang, in Biomass Now - Cultivation and

Utilization, ed. Dr. Miodrag Darko Matovic, InTech, 2013, ch. 7, pp. 153-192. 50 A.K. Camper, M.W. Lechevallier, S.C. Broadaway and G.A. Mcfeters, Appl. Environ.

Microbiol., 1985, 50, 1378–1382.51 J. Chen, B. Gu, E.J. LeBoeuf, H. Pan and S. Dai, Chemosphere, 2002, 48, 59. 52 A.J. Martijn. Ph.D. Thesis, Wageningen University, the Netherlands, 2015.53 B.J. Martijn, A.R. Van Rompay, E.J.M. Penders, Y. Alharbi, P.K. Baggelaar, J.C.

Kruithof and I.M.C.M. Rietjens, Chemosphere, 2016, 144, 338. 54 EFSA. Guidance of the Scientific Committee on a request from EFSA on the use of the

benchmark dose approach in risk assessment, The EFSA Journal, 2009, 1150, 1.55 WHO. Concise International Chemical Assessment Document 20, Mononitrophenols,

World Health Organization, Geneva, 2000. 56 USEPA. Health Advisories for Drinking Water Contaminants. United States

Environmental Protection Agency. Office of Water Health Advisories. Lewis Publishers, USA, 1993.

Table 1. Water quality and operational parameters for the two WTW sampling events. * denotes ‘’not calculated’’, since the DOC value was below the detection limit (1 mg/L); UVT: Ultraviolet Transmission

Figure 1. Number of positive wells generated by the Ames II assay (TA98 – S9) under different experimental conditions with a) Suwannee River, and

23

Sampling (a)

Sampling (b)

Temperature (ₒC) 5.0 10pH (-) 8.2 8.5

Works inlet DOC (mg/L) 4.05 4.71Pre-AOP DOC (mg/L) 2.68 2.11Post-AOP DOC (mg/L) 1.52 <1.00Inlet SUVA (L/m mg) 4.61 3.74

Pre-AOP SUVA (L/m mg) 4.70 4.18Post-AOP SUVA (L/m mg) 2.90 *

Inlet UVT (%) 65.0 66.7Pre-AOP UVT (%) 74.8 81.7Post-AOP UVT (%) 90.4 92.7UV fluence (mJ/cm2) 2000 1750

Hydrogen peroxide dose (mg/L) 40 40

Works inlet nitrate (mg/L) 30.5 31.7Works effluent nitrate (mg/L) 30.0 30.2

578579580581582583584585586587588589590591592593594595596597598599

600

601

602

603

604605606

607

608

609

610611 612

613614615616617618619620

b) Pony Lake NOM (H2O2 concentration 15 mg/L). The numbers 0, 1500 and 2000 represent the UV fluences applied, in mJ/cm2. The positive control (2-nitrofluorene + 4-nitroquinoline N-oxide) produced 46.9 (±0.782) positive wells. The error bars represent the standard deviation values for the mean number of positive wells, where the mean number of positive wells represents the average of the positive wells for 9 replicates in the control conditions and 3 replicates for each extract.

Figure 2. Number-fold increases over a) solvent control, and b) baseline, calculated from the Ames II assay (TA98 – S9) values for the different experimental conditions with either Pony Lake or Suwannee River NOM (H2O2 concentration 15 mg/L). The numbers 0, 1500 and 2000 represent the UV fluences applied, in mJ/cm2. The error bars represent the standard deviation values for the mean number of positive wells, where the mean number of positive wells represents the average of the positive wells for 9 replicates in the control conditions and 3 replicates for each extract.

24

621622623624625626

627628

629630631632633634635636

637

Figure 3. Concentrations of mono-nitrophenol (2-nitrophenol, 4-nitrophenol or a combination of both) produced by the LP-UV/H2O2 treatment of synthetic water samples containing Pony Lake NOM and nitrate. The bar representing the value for the 0/15 UV/H2O2

combination is shaded because the nitrophenol concentration is higher than the detection limit of 0.007 μg/L but lower than the quantification limit of 0.02 μg/L. The numbers 0, 1500 and 2000 represent the UV fluences applied, in mJ/cm2, while the 15 represents the H2O2 dosage in mg/L.

Figure 4. Number of positive wells generated by the Ames II assay (TA98 – S9) for the full-scale water samples during different stages of treatment for the two sampling dates. The positive control (2-nitrofluorene + 4-nitroquinoline N-oxide) produced 46.9 (±0.782) positive wells. The error bars represent the standard deviation values for the mean number of positive wells, where the mean number of positive wells represents the average of the positive wells for 9 replicates in the control conditions and 3 replicates for each extract.

25

638639640641642643644645

646

647

648649650651652653654

655

Figure 5. Number-fold increases over a) solvent control, and b) baseline, calculated from the Ames II assay (TA98 – S9) values for the full-scale water samples during different stages of treatment for both samplings (a) and (b). The error bars represent the standard deviation values for the mean number of positive wells, where the mean number of positive wells represents the average of the positive wells for 9 replicates in the control conditions and 3 replicates for each extract.

26

656

657658659660661662663

664

Figure 6. LC-OCD fractionation of full-scale water samples: a) sampling (a), and b) sampling (b) (C: Carbon, N: Nitrogen, LMW: Low Molecular Weight).

Figure 7. Ames II response in water samples (20,000 concentration factor) as a function of the nitrite formation by MP UV treatment at WTP Andijk and in CB experiments with IHSS Pony Lake NOM (current LP-UV/H2O2 results superimposed on adapted Figure 5.3 of Martijn, 2015)52.

27

665666667

668

669

670671672673674