Accepted Manuscript

Degradation of sulfamethoxazole by UV, UV/H2O2 and UV/persulfate (PDS):

Formation of oxidation products and effect of bicarbonate

Yi Yang, Xinglin Lu, Jin Jiang, Jun Ma, Guanqi Liu, Ying Cao, Weili Liu, Juan Li,
Suyan Pang, Xiujuan Kong, Congwei Luo
PII: S0043-1354(17)30244-0
DOI: 10.1016/j.watres.2017.03.054
Reference: WR 12791

To appear in: Water Research

Received Date: 7 September 2016

Revised Date: 20 March 2017
Accepted Date: 25 March 2017

Please cite this article as: Yang, Y., Lu, X., Jiang, J., Ma, J., Liu, G., Cao, Y., Liu, W., Li, J., Pang,
S., Kong, X., Luo, C., Degradation of sulfamethoxazole by UV, UV/H2O2 and UV/persulfate (PDS):
Formation of oxidation products and effect of bicarbonate, Water Research (2017), doi: 10.1016/
j.watres.2017.03.054.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

2 Degradation of sulfamethoxazole by UV, UV/H2O2 and

3 UV/persulfate (PDS): Formation of oxidation products

and effect of bicarbonate

PT
4

RI
6 Submitted to

SC
7 Water Research

9
U
Yi Yang1, Xinglin Lu1, Jin Jiang1,*, Jun Ma1,*, Guanqi Liu1, Ying Cao1, Weili
AN
10 Liu1, Juan Li1, Suyan Pang2, Xiujuan Kong1 and Congwei Luo1
M

11
1
12 State Key Laboratory of Urban Water Resource and Environment,
D

13 Harbin Institute of Technology, Harbin, 150090, China.

TE

14
2
EP

15 Key Laboratory of Green Chemical Engineering and Technology of College of Heilongjiang

16 Province, College of Chemical and Environmental Engineering, Harbin University of Science

C

17 and Technology, Harbin, Heilongjiang 150040, China.

AC

18

19 *Corresponding authors contact details:

20 *E-mail: jiangjinhit@126.com; Tel: +86 451 86283010; Fax: +86 451 86283010.

21 *E-mail: majun@hit.edu.cn; Tel: +86 451 86283010; Fax: +86 451 86283010.

-1-
 ACCEPTED MANUSCRIPT
22 Abstract

23 The frequent detection of sulfamethoxazole (SMX) in wastewater and surface waters gives

24 rise of concerns about their ecotoxicological effects and potential risks to induce antibacterial

25 resistant genes. UV/hydrogen peroxide (UV/H2O2) and UV/persulfate (UV/PDS) advanced

26 oxidation processes have been demonstrated to be effective for elimination of SMX, but there

PT
27 is still a need for a deeper understanding of product formations. In this study, we identified

RI
28 and compared the transformation products of SMX in UV, UV/H2O2 and UV/PDS processes.

29 Because of the electrophilic nature of SO4•-, the second-order rate constant for the reaction of

SC
30 sulfate radical (SO4•-) with the anionic form of SMX was higher than that with the neutral

form, while •OH exhibited comparable reactivity to both forms. The direct photolysis of SMX

U
31 AN
32 predominately occurred through cleavage of the N-S bond, rearrangement of the isoxazole

33 ring, and hydroxylation mechanisms. Hydroxylation was the dominant pathway for the
M

34 reaction of hydroxyl radical (•OH) with SMX. SO4•- favored attack on –NH2 group of SMX

35 to generate a nitro derivative and dimeric products. The presence of bicarbonate in UV/H2O2
D

36 inhibited the formation of hydroxylated products, but promoted the formation of the nitro
TE

37 derivative and the dimeric products. In UV/PDS, bicarbonate increased the formation of the
EP

38 nitro derivative and the dimeric products, but decreased the formation of the hydroxylated

39 dimeric products. The different effect of bicarbonate on transformation products in UV/H2O2

C

40 vs. UV/PDS suggested that carbonate radical (CO3•-) oxidized SMX through the electron
AC

41 transfer mechanism similar to SO4•- but with less oxidation capacity. Additionally, SO4•- and

42 CO3•- exhibited higher reactivity to the oxazole ring than the isoxazole ring of SMX.

43 Ecotoxicity of transformation products was estimated by ECOSAR program based on the

44 quantitative structure-activity relationship analysis as well as by experiments using Vibrio

45 fischeri, and these results indicated that the oxidation of SO4•- or CO3•- with SMX generated

-2-
 ACCEPTED MANUSCRIPT
46 more toxic products than those of •OH.

47 Keywords: sulfamethoxazole; hydroxyl radical; sulfate radical; carbonate radical;

48 transformation products.

PT
RI
U SC
AN
M
D
TE
C EP
AC

-3-
 ACCEPTED MANUSCRIPT
49 1. Introduction

50 Sulfamethoxazole (SMX) is an important antibiotic and has been widely used in human

51 and veterinary medicines to treat diseases and infections. A large portion of SMX is excreted

52 unchanged into the sewage system (Lienert et al. 2007, Ternes and Joss 2006). Due to the less

53 efficient elimination of SMX in conventional wastewater treatment plants, it eventually enters

PT
54 into the aquatic environment. SMX has been regularly detected in wastewater effluent in a

RI
55 concentration range of 100−2500 ng/L, in surface water in a concentration range of 60−150

56 ng/L, and even in drinking water at a concentration of 12 ng/L (Al Aukidy et al. 2012, Kolpin

SC
57 et al. 2002, Morasch et al. 2010, Padhye et al. 2014, Ratola et al. 2012). The chronic exposure

58 of bacteria to trace levels of antibiotics in the aquatic environment raises concerns about their

U
AN
59 ecotoxicological effects and potential risks to induce antibacterial resistant genes among

60 native bacterial populations. Many studies have previously investigated the occurrence and
M

61 fate of SMX in the aquatic environment. Compared with biodegradation, photodegradation is

62 a predominant pathway for SMX elimination in surface waters (Boreen et al. 2003). However,
D

63 the efficiency of SMX degradation by natural sunlight varied extensively with solar exposure
TE

64 by seasons (Bonvin et al. 2012, Bonvin et al. 2011). The frequent occurrence of SMX in the

65 aquatic environment indicates that more efficient treatment is needed to destruct SMX in
EP

66 order to meet water demands.

C

67 Advanced oxidation processes (AOPs) are promising technologies to destruct organic

AC

68 contaminants. Hydroxyl radical-based AOPs (e.g., UV photolysis of hydrogen peroxide

69 (UV/H2O2)) have been used for the destruction of recalcitrant organic contaminants from

70 drinking water and wastewater, as hydroxyl radicals (•OH) react with many organic

71 chemicals at near diffusion-controlled rates (Buxton et al. 1988). Pharmaceuticals could be

72 effectively removed by •OH-based AOPs in various conditions (Baeza and Knappe 2011,

-4-
 ACCEPTED MANUSCRIPT
73 Keen and Linden 2013, Wols et al. 2013). Sulfate radical-based AOPs (e.g., UV photolysis of

74 peroxodisulfate (UV/PDS)) have attracted more interests, because sulfate radical (SO4•-) also

75 reacts with a wide range of organic contaminants with near diffusion-limited rate constants

76 (Neta et al. 1988). Although SO4•- is a strong oxidant with a redox potential (2.5−3.1 V)

77 (Neta et al. 1988) comparable to that of •OH (1.8−2.7 V) (Buxton et al. 1988), SO4•- is a more

PT
78 selective oxidant than •OH.

The rate constant for the reaction of SMX with SO4•- was determined to be (11.7−16.1)

RI
79

80 × 109 M-1 s-1 in the pH range from 6 to 9, which was slightly higher than that with •OH at

SC
81 (7.02−7.89) × 109 M-1 s-1 (Zhang et al. 2015). Meanwhile, direct photolysis at 254 nm was

82 identified as an effective pathway for SMX degradation with a relatively high quantum yield

83
U
(Canonica et al. 2008, Wols et al. 2013). These results suggest a potential application of UV-
AN
84 based AOPs in the removal of SMX from various waters. The apparent destruction efficiency
M

85 of SMX by UV at 254 nm, UV/H2O2 and UV/PDS in reverse osmosis (RO) brines from

86 municipal wastewater reuse facilities has been investigated in our previous study (Yang et al.
D

87 2016). The average UV fluence for 50% removal of SMX was less than 180 mJ/cm2 in two
TE

88 RO brine samples. Anions at high concentration in RO brines acted as •OH and SO4•-

89 scavengers to generate reactive species. Bicarbonate has been suggested to be an important

EP

90 scavenger of •OH, SO4•- and halogen radicals to form carbonate radical (CO3•-) (Anastasio
C

91 and Matthew 2006, Grebel et al. 2010, Yang et al. 2014). Modeling results indicated that

CO3•- concentration exceeded those of other radicals by several orders of magnitude in RO

AC

92

93 brines, especially in UV/PDS process (Yang et al. 2014). CO3•- is a more selective oxidant

94 than •OH and SO4•-, but it is sufficient to oxidize some pharmaceuticals (Canonica et al.

95 2005). The presence of bicarbonate promoted the degradation of SMX by UV/PDS, but

96 slightly inhibited that by UV/H2O2 (Yang et al. 2016). Zhang (Zhang et al. 2015) found

97 similar results in synthetic human urine samples and explained this difference by the lower
-5-
 ACCEPTED MANUSCRIPT
98 rate constants for the reaction of CO3•- with PDS than that with H2O2. The rate constant for

99 the reaction of CO3•- with SMX was determined to be (2.68 ± 0.71) × 108 M-1 s-1 (Zhang et al.

100 2015).

101 Understanding the transformation products (TPs) of SMX and characterizing the

102 reaction pathways in different oxidation processes are critical to evaluating the risk associated

PT
103 with the presence of SMX in the aquatic environment. The TPs of SMX by photolysis have

RI
104 been well investigated in solar and UV 254 nm photolysis (Bonvin et al. 2012, Trovó et al.

105 2009, Zhou and Moore 1994), and many efforts have also been put on investigating TPs

SC
106 formation of SMX in •OH- and SO4•--based processes including electrochemistry, thermo-

107 activated persulfate and peroxymonosulfate/cobalt(II) systems (Ji et al. 2015, Lin et al. 2013,

108
U
Mahdi Ahmed et al. 2012). Nevertheless, the formation mechanisms of the TPs of SMX by
AN
109 direct photolysis combined with •OH or SO4•- oxidation (i.e., UV/H2O2 and UV/PDS), as well
M

110 as the relative yields of TPs, are not well documented. Also, little work has been conducted to

111 address the TPs of SMX by CO3•- oxidation.

D

112 The objective of this study is to identify and compare the TPs of SMX in UV, UV/H2O2
TE

113 and UV/PDS processes. The contributions of •OH and SO4•- on the TPs formation were

114 investigated in UV-based AOPs. The relative yields of TPs by •OH and SO4•- were evaluated,
EP

115 revealing the different reactivity of •OH and SO4•- to the specific groups of SMX. This study

is the first time to investigate the TPs of SMX by CO3•- and assess the impact of bicarbonate
C

116
AC

117 on ecotoxicity during UV/H2O2 and UV/PDS processes.

118 2. Materials and methods

119 2.1. Materials

120 Sulfamethoxazole (SMX), potassium perdisulfate (PDS), ammonium acetate were

121 purchased from Sigma-Aldrich. Sulfamethoxazole-d4 (SMX-d4) was purchased from Santa

-6-
 ACCEPTED MANUSCRIPT
122 Cruz Biotech. Hydrogen peroxide (H2O2) solution (35% w/w), tert-butanol (t-BuOH) and

123 sodium bicarbonate were purchased from Sinopharm Chemical Reagent Co. Ltd., China.

124 HPLC grade methanol and acetic acid were received from Thermo Fisher Scientific, and

125 acetonitrile was received from Merck. All solutions were prepared in deionized (DI) water

126 (18.2 MΩ/cm) from a Milli-Q purification system (Millipore, Billerica, MA).

PT
127 2.2. Experimental procedures

RI
128 UV apparatus was applied using a semi-collimated beam system comprising four 10 W

129 low pressure mercury lamps emitting at 254 nm shining down onto a 100 mL crystallization

SC
130 dish (pathlength = 4.1 cm), as described previously (Yang et al. 2015). Samples were

131 withdrawn periodically and supplemented with 20 µL methanol per mL sample to quench any

132
U
radicals formed by thermolysis of PDS or H2O2, and kept in a 4 °C refrigerator for further
AN
133 analysis within 12 hours. The surface irradiance (8.69 × 10−7 Einsteins L−1 s−1) was
M

134 determined by iodide-iodate actinometry (Rahn et al. 2003). Experiments were conducted in

135 10 mM phosphate buffer at 20 ± 2 °C.

D

136 Some experiments were conducted with filtrated waters from two drinking water plants
TE

137 using surface water and ground water as water sources. These water samples were filtered

138 through 0.7 µm nominal pore size glass fiber filters (Whatman) and stored at 4°C. The
EP

139 surface water had a low DOC and low alkalinity (DOC 4.24 mg/L, alkalinity 0.2 mM HCO3-,
C

140 UV254 = 0.045, pH 7.2), while the ground water showed a relatively high DOC and high

alkalinity (DOC 6.94 mg/L, alkalinity 8 mM HCO3-, UV254 = 0.034, pH 8.25).

AC

141

142 2.3. Kinetics of the reactions of SMX with •OH and SO4•-.

143 Since pKa,1 (1.7) of SMX is not relevant to natural water conditions, only pKa,2 (5.7) was

144 considered herein. The pH-dependent kinetics of direct photolysis of SMX were performed in

145 the pH range of 3−8 (see Text S1 for details). To evaluate the specific rate constants for the

146 reactions of •OH and SO4•- with the neutral and anionic forms of SMX, the apparent pH-
-7-
 ACCEPTED MANUSCRIPT
147 dependent rate constants ( k app, • OH and k app,SO• - ) were determined by competition kinetics (see
4

148 Text S2 for details). k app, • OH and k app,SO• - are expressed as equation 1 and equation 2,
4

149 respectively.

150 k app, • OH = αSH k SH, • OH + αS− k S− , • OH (1)

PT
151 k app,SO• - = αSH k SH,SO• - + αS− k S− ,SO• - (2)
4 4 4

RI
152 where kSH, • OH , kS− , • OH , k SH,SO• - and k S− ,SO• - are the specific rate constants for the reactions of
4 4

153 OH and SO4•- with neutral and anionic forms, respectively; αSH and αS− are the proportions
•

SC
154 of neutral (SH) and anionic (S-) forms at a given pH and can be calculated by equation 3 and

U
155 equation 4, respectively.
AN
[H + ]
156 α HS = (3)
K a,2 + [H + ]
M

Ka 2
157 α S− = (4)
K a,2 + [H + ]
D

158 Therefore, the second-order rate constants of •OH or SO4•- with specific forms of SMX were
TE

159 calculated by nonlinear regression of pH-dependent k app, • OH and k app,SO• - data.

4
EP

160 2.4. Identification of oxidation products

161 SMX was analyzed using a Waters 1525 HPLC with a Waters 2487 dual λ detector.
C

162 Chromatographic separations were performed using a Waters symmetry C18 column (150
AC

163 mm × 4.6 mm, 5 µm). The concentrations of SMX were quantified at λ = 260 nm, with an

164 eluent consisting of 0.1% acetic acid and methanol with a ratio of 40:60 (v/v) at a flow rate of

165 1 mL/min.

166 To identify the oxidation products of SMX, a triple quadrupole TOF mass spectrometer

167 (Triple TOF 5600, AB Sciex) coupled with the Ekspert ultralLC110 was used.

-8-
 ACCEPTED MANUSCRIPT
168 Chromatographic separations were performed using a Poroshell 120 EC-C18 column (50 mm

169 × 3.0 mm, 2.7 µm). Accurate MS and MS/MS patterns of SMX and its oxidation products

170 were analyzed in a molecular ion scanning mode (m/z 50 to 1000) in both positive and

171 negative electrospray ionization (ESI) modes. Analyst (version 1.6, AB Sciex) and PeakView

172 (version 1.2.0.3, AB Sciex) were employed to identify and analyze peaks. The obtained

PT
173 fragmentation patterns of products were incorporated into a multiple reaction mode (MRM)

RI
174 method to quantify the formation of products by a triple quadrupole mass spectrometer

175 (QTrap 5500 MS, ABSciex) coupled with the Agilent 1260 HPLC. The LC program was

SC
176 identical to that described above. Peak areas of TPs were normalized using the internal

177 standard (SMX-d4). Further information on the applied setup, the optimized chromatographic

178
U
conditions and the conditions of data acquisition parameter is provided in Text S3.
AN
179 2.5. Toxicity analysis
M

180 The acute toxicity assay of samples treated by UV, UV/H2O2, and UV/PDS was carried

181 out by measuring the decrease in the bioluminescence of Vibrio fischeri (Hamamatsu
D

182 Photonics, Beijing,China). Residue H2O2 or PDS in samples were quenched by bovine
TE

183 catalase and ascorbic acid, respectively. Both bovine catalase and ascorbic acid had a

184 negligible impact on luminescence. As 500 mM bicarbonate exceeded the osmolality of

EP

185 Vibrio fischeri, only 50 mM bicarbonate was tested for the acute toxicity. The luminescence
C

186 was detected by a luminometer (Promega GloMax®-Multi Jr). The acute toxicity of SMX
AC

187 and its TPs were also estimated using the Ecological Structure Activity Relationships

188 (ECOSAR) program for fish, daphnid and green algae.

189 3. Results and discussion

190 3.1. Oxidation kinetics of the reactions of SMX with •OH and SO4•-

191 The pH dependence of SMX degradation by direct photolysis has been investigated by

-9-
 ACCEPTED MANUSCRIPT
192 Canonica et al (Canonica et al. 2008). The quantum yields of neutral and anionic forms of

193 SMX determined in this study were 0.187 ± 0.002 M Einstein-1 and 0.028 ± 0.002 M

194 Einstein-1 (Text S1), respectively, which were consistent with their results.

195 As pH varied from 3 to 8, no significant difference of k app, • OH for reaction of SMX with

•
196 OH was observed (Fig. 1 A). This result indicated that •OH exhibited similar reactivity to the

PT
197 neutral form and anionic form of SMX. Therefore, the same value of kSH, • OH and kS− , • OH was

RI
198 suggested herein and determined to be (7.63 ± 0.85) × 109 M-1 s-1. Interestingly, k app,SO•-
4

SC
199 increased gradually with increasing pH from 3 to 8 (Fig. 1 B). This phenomenon was

200 ascribed to deprotonation of the secondary amine group in SMX, and SO4•- was expected to

201
U
have high reactivity as a result of its electrophilic nature. k S− ,SO• - was measured to be (1.34 ±
AN
4

202 0.02) × 1010 M-1 s-1, which was 14-fold higher than k SH,SO• - ((9.29 ± 1.5) × 108 M-1 s-1). The
4
M

203 reaction kinetics of SO4•- with substructure model substrates of SMX demonstrated that

204 aniline was the most reactive group to SO4•- (Ji et al. 2015). Substitution by the sulfonic
D

205 group or the sulfonamide group at the para position of aniline reduced the rate constants
TE

206 significantly, while 3-amino-5-methyl-isoxazole was also reported to be less reactive than
EP

207 aniline (Ji et al. 2015). When –NH2 group of SMX was substituted by –NHCOCH3, the

208 second-order rate constants of acetyl-SMX was only one fifth of that of SMX (Zhang et al.
C

209 2015). These results indicate that the reaction of SO4•- with the SMX structure takes place
AC

210 primarily at the p-sulfonylaniline group. The pH dependence of k app,SO•- likely reflects
4

211 enhancement of the electron-donating effect of SMX’s aniline group by sulfonamide nitrogen

212 deprotonation, which may favor the electrophilic attack by SO4•-.

213 3.2. Product identification

214 Several studies have examined transformation products (TPs) arising from direct

- 10 -
 ACCEPTED MANUSCRIPT
215 photolysis, •OH- and SO4•--based oxidations of SMX. The major photoproducts were formed

216 from cleavage of the N-S bond (Boreen et al. 2004, Trovó et al. 2009). Hydroxyl derivatives

217 of SMX were produced during •OH- or SO4•--dominant process (Hu et al. 2007, Zhang et al.

218 2016). In addition, SO4•- was reported to favor oxidation of –NH2 group of SMX (Mahdi

219 Ahmed et al. 2012). However, only few studies quantified the formation products among UV,

PT
220 UV/H2O2 and UV/PDS (Zhang et al. 2016).

TPs produced by direct photolysis, •OH, SO4•- and CO3•- were identified by ESI-TOF-

RI
221

222 MS. The accurate m/z values are provided in Table S2. For structural elucidation, the

SC
223 fragmentation pathways of TPs were studied by performing product ion scans. The product

224 ion spectra and the fragmentation pathways of SMX ([m + H]+ = 254.0591) were illustrated

225
U
in Fig. S3, where m/z 156, m/z 108, m/z 99 and m/z 92 were the characteristic fragment ions
AN
226 of SMX. The chemical structures of TPs could be elucidated based on the fragment ions of
M

227 SMX. Normalized peak areas were used to compare the relative yields of TPs of SMX.

228 3.2.1. Direct photolysis

D

229 The formation of TPs by direct photolysis of SMX was conducted at pH 3, where more
TE

230 efficient degradation of SMX could result in detectable TPs. The normalized peak areas of

231 main TPs at pH 3 were shown in Fig. 2. TP 98 ([m + H]+ = 99.0554) and TP 173 ([m - H]- =
EP

232 172.0074) were confirmed as 3-amino-5-methylisoxazole and sulfanilic acid (Fig. S4 and Fig.
C

233 S5), respectively, resulting from cleavage of the N-S bond of SMX. The peak area of TP 98
AC

234 increased in the first 30 min and then reached a plateau, which was consistent with the

235 degradation trend of SMX. This result indicated that 3-amino-5-methylisoxazole was not

236 susceptible to be photodegraded.

237 TP 253 ([m + H]+ = 254.0588, Fig. S7) reported by Zhou and Moore (Zhou and Moore

238 1994) as an isomer of SMX was produced through photoisomerization of the isoxazole ring.

239 TP 253 was accumulated during the initial 20 min and then gradually decreased, suggesting
- 11 -
 ACCEPTED MANUSCRIPT
240 that the isomer of SMX underwent photodegradation. Two peaks with a nominal mass [m +

241 H]+ = 270.0542 were found (see Fig. S8 for example). Their MS2 spectra exhibited similar

242 fragment ions. The MS2 spectra of TP 269 showed characteristic fragment ions at m/z 172,

243 124 and 108, suggesting the addition of one oxygen atom to the sulfanilic group. The

244 occurrence of the m/z 99 fragment ion indicated that the isoxazole ring remained unchanged.

PT
245 Although the retention time of TP 269 at 8.5 min matched that of an authentic standard of N4-

RI
246 hydroxy-sulfamethoxazole, the relative intensity of fragment ions was different. This

247 indicated the addition of hydroxyl group of the benzene ring, which was also suggested by

SC
248 previous studies (Gao et al. 2014, Trovó et al. 2009). Due to the insufficient separation of

249 isomers during chromatography, a total peak area of TP 269 was calculated. The peak area of

250
U
TP 269 reached the maximum value at 20 min and then gradually decreased. TP 189 ([m - H]-
AN
251 = 188.0013) might result from cleavage of the N-S bond by the photodegradation of TP 269

or the hydroxylation of TP 173. The product ion scan of both TP 271 ([m + H]+ = 272.0700)
M

252

253 and TP 287 ([m + H]+ = 288.0635) showing the fragment ions at m/z 156, 108 and 92
D

254 signified an unmodified sulfanilic group (Fig. S9 and Fig. S11). The absence of fragment ion
TE

255 at m/z 99 suggested the high reactivity of the isoxazole ring during direct photolysis. TP 271

256 and TP 287 were considered to result from the addition of one and two oxygen atoms to the
EP

257 double bond in the isoxazole ring to form corresponding alcohols, respectively. Two peaks of
C

258 TP 287 were separated at distinct retention times. The peak of TP 287 at 3.6 min was
AC

259 identified to derive from TP 253, while the peak of TP 287 at 6.8 min was identified to derive

260 from SMX (see details in Section 3.2.4 below). The peak area of TP 271 decreased after 20

261 min, while those of TP 287-1 and TP 287-2 remained constant. Similar TPs of SMX by direct

262 photolysis were observed at pH 8. Note that the maximum peak value of TP 269 at pH 8 was

263 eight times higher than that at pH 3. This indicated that the photolysis efficiency of TP 269

264 was different as pH varied.

- 12 -
 ACCEPTED MANUSCRIPT
265 3.2.2. Identification of reactive species during direct photolysis

266 Besides the degradation of SMX by direct photolysis, reactive species (e.g., 1O2 and
•
267 OH ) produced during SMX photolysis could also contribute to the degradation of SMX

268 (Zhou and Moore 1997). •OH reaction can be scavenged by t-BuOH with the second-order

269 rate constant of 6.0 × 108 M-1 s-1 (Buxton et al. 1988), but 1O2 is inert to t-BuOH (Rodgers

PT
270 1983). Therefore, the oxidation pathway by •OH would be completely inhibited in a large

RI
271 excess of t-BuOH (10 mM). The degradation rate of SMX was not affected in the presence of

272 t-BuOH (Fig. 2). The generation of TP 98, TP 253, TP 271 and TP 287-1 also exhibited

SC
273 similar trends as that without t-BuOH. However, the maximum peak area of TP 269 was

274 reduced by 64%, while no TP 287-2 was observed in the presence of t-BuOH. The decrease

275
U
of these hydroxylated products suggested that •OH contributed to the formation of secondary
AN
276 products. Specifically, TP 287-2 could derive from SMX by •OH, while TP 253 might be one
M

277 precursor of TP 269.

278 NaN3 was an effective scavenger of 1O2 with the second-order rate constant of 1 × 109
D

279 M−l s−l (Gsponer et al. 1987), while it also effectively quenched •OH with the second rate
TE

280 constant of 1.2 × 1010 M−l s−l (Buxton et al. 1988). The presence of NaN3 inhibited the

281 degradation of SMX, due to an inner filter effect of NaN3 for direct photolysis (Fig. 2). The
EP

282 quantum yield of SMX was equal to that without NaN3, indicating that the degradation of

SMX by 1O2 was negligible, which was consistent with previous studies (Boreen et al. 2004,
C

283
AC

284 Zhou and Moore 1997). The peak area of TP 254 increased, as the peak area of TP 271

285 decreased. These results suggested that 1O2 might contribute to the degradation of TP 253

286 leading to the formation of TP 271. Neither TP 269 nor TP 287-2 was detected, because

287 NaN3 scavenged •OH as well.

288 3.2.3. UV/H2O2

289 pH 3. The apparent degradation rate of SMX in UV/H2O2 was comparable to that by
- 13 -
 ACCEPTED MANUSCRIPT
290 direct photolysis (Fig. 2). Because of the higher phototransformation quantum yield at pH 3,

291 the direct photolysis was a major pathway for SMX degradation. Although •OH contributed

292 to only 13% of SMX degradation, TPs were significantly affected by the involvement of •OH.

293 The major differences between UV/H2O2 and direct photolysis were the formation of

294 hydroxylated products. For UV/H2O2 process, the maximum peak area of TP 269 increased

PT
295 by around 15 times, and the subsequent degradation of TP 269 was attributed to further

oxidation by •OH. Interestingly, •OH promoted the formation of TP 287-2 (derived from

RI
296

297 SMX), while it inhibited the formation of TP 287-1 (derived from TP 253, an isomer of

SC
298 SMX). These results suggested that •OH favored the addition reaction on the isoxazole ring

299 over the oxazole ring, which was consistent with t-BuOH quenching experiment above. In

300
U
contrast to direct photolysis, the peak area of TP 98 decreased after 25 min, implying further
AN
301 oxidation of TP 98 by •OH.
M

302 pH 8. Fig. 3 depicted the degradation of SMX and the formation of TPs at pH 8. The

303 contribution of •OH to SMX degradation increased to 54% at pH 8. Maximum peak areas of
D

304 both TP 253 and TP 271 in UV/H2O2 were only half of those formed in direct photolysis.
TE

305 Meanwhile, the maximum peak area of TP 269 was seven times higher in UV/H2O2 than that

306 in direct photolysis. The formation of TP 287-1 and TP 287-2 was not observed. This result
EP

307 indicated that the further oxidation of TP 287-1 and TP 287-2 by •OH, leading to their peak
C

308 areas below the detection limit.

AC

309 3.2.4. UV/PDS

310 pH 3. The apparent degradation rate of SMX by UV/PDS was close to that by direct

311 photolysis (Fig. 2), where SO4•- contributed to only 7% of SMX degradation. However, the

312 generation of TPs in UV/PDS was significantly different from that in direct photolysis. The

313 peak of TP 253 was hardly detected in UV/PDS, while the maximum peak area of TP 271

314 was as half as that in direct photolysis. Meanwhile, SO4•- resulted in an increase in the
- 14 -
 ACCEPTED MANUSCRIPT
315 formation of TP 287-1, and the maximum peak area of TP 287-1 was three times greater than

316 that in direct photolysis, suggesting a correlation between TP 253 and TP 287-1. The peak

317 area of TP 287-2 was comparable to that in direct photolysis. Considering that the retention

318 time of SMX was longer than TP 253 by the separation of chromatography, a longer retention

319 time of TP 287 derived from SMX was expected. This evidence confirmed that TP 287-1 and

PT
320 TP 287-2 contained an oxazole ring and an isoxazole ring, respectively. Therefore, SO4•-

RI
321 showed higher reactivity to the oxazole ring than the isoxazole ring. These results also

322 indicated that SO4•- favorably destructed the TPs formed in direct photolysis (i.e., TP 253).

SC
323 Furthermore, four new TPs were produced in UV/PDS. The retention time of TP 269

324 was different from those observed in direct photolysis and UV/H2O2 (Fig. S14). Their similar

325
U
MS2 spectra indicated that •OH and SO4•- exhibited different reactivity to different positions
AN
326 on the sulfanilic group. Due to different structures of TP 269, the peak area of TP 269 in

UV/PDS was not compared with those in direct photolysis and UV/H2O2. TP 283 ([m + H]+ =
M

327

328 284.0328, Fig. S10) was identified as nitro-SMX. A loss of 46 Da corresponding to the nitro
D

329 group resulted in a m/z 238 fragment ion. The fragment ion m/z 122 resulted from cleavage
TE

330 of the N-S bond to generate nitrobenzene group. Two dimeric products were also detected

331 during the degradation of SMX in UV/PDS. TP 502 ([m + H]+ = 503.0813, Fig. S12) was
EP

332 proposed as azosulfamethoxazole, and TP 518 ([m + H]+ = 519.0755, Fig. S13) was a
C

333 hydroxyl substitution of TP 502.

pH 8. Since the rate constant for the reaction of SO4•- and SMX increased at a higher pH,
AC

334

335 SO4•- and direct photolysis contributed to 72% and 28% of SMX degradation in UV/PDS,

336 respectively. No TP 253 was observed in Fig. 4. A higher relative yield of TP 98 was found

337 in UV/PDS over both direct photolysis and UV/H2O2. The peak area of TP 284 reached the

338 maximum value as complete removal of SMX. The decrease of TP 502 peak area after 60

339 min suggested further oxidation by SO4•- to form other TPs, like TP 518.
- 15 -
 ACCEPTED MANUSCRIPT
340 3.2.5. Effect of bicarbonate

341 The effect of bicarbonate was investigated at pH 8. Relatively high concentrations of

342 bicarbonate to generate a CO3•- dominant condition were chosen as 50 and 500 mM, which

343 scavenged 74% and 97% of •OH, and 34% and 84% of SO4•-, respectively. In the presence of

344 bicarbonate (Fig. 3), the degradation of SMX was inhibited in UV/H2O2. The addition of 50

PT
345 mM bicarbonate in UV/PDS slightly enhanced the degradation of SMX, while the effect of

RI
346 500 mM bicarbonate on SMX degradation was negligible (Fig. 4). These results might be

347 explained by the lower rate constant of CO3•- with PDS than that with H2O2 (Zhang et al.

SC
348 2015).

349 The presence of bicarbonate at 50 mM in UV/ H2O2 inhibited the formation of TP 253,

350
U
TP 269 and TP 271 (Fig. 3). Neither TP 253 nor TP 269 was observed with 500 mM
AN
351 bicarbonate. Consistent with the relatively low yield of TP 271, TP 287-1 was detected in the
M

352 presence of bicarbonate. Meanwhile, the peak area of TP 98 was promoted with 500 mM

353 bicarbonate. The formations of TP 283, TP 502 and TP 518, derived from the oxidation of
D

354 aniline group, were also enhanced in the presence of bicarbonate. Since the rate constant of
TE

•
355 OH with SMX was 30 times higher than that of CO3•-, the degradation of SMX was still

356 predominant by •OH in the presence of 50 mM bicarbonate, and thus producing only small
EP

357 amounts of these TPs. As bicarbonate concentration increased to 500 mM, the relative yields
C

358 of these TPs were promoted substantially. The formation trends of these TPs were similar to

those in UV/PDS, suggesting that the reaction pathway of SMX degradation by SO4•- likely
AC

359

360 also applied to CO3•-.

361 UV/PDS produced similar TPs in the presence or absence of bicarbonate (Fig. 4). The

362 generation of CO3•- increased the relative formation yield of TP 98. The accumulation of TP

363 283 was slower in the presence of bicarbonate than that in the absence of bicarbonate in 60

364 min, and continued increasing after that. This result implied the stronger oxidation capacity
- 16 -
 ACCEPTED MANUSCRIPT
365 of SO4•- than CO3•- to further oxidize TP 284. This analysis also explained a greater peak area

366 of TP 502, but a smaller peak area of TP 518 obtained in the presence of bicarbonate.

367 3.3. Proposed transformation pathways

368 3.3.1. Direct photolysis pathway

369 The formation of the identified TPs occurred through cleavage reaction by photolysis or

PT
370 oxidation by reactive species (i.e., 1O2, •OH, SO4•- and CO3•-). For direct photolysis (Scheme

RI
371 1), TP 98 and TP 173 were formed through cleavage of the N-S bond of SMX. Previous

372 studies also found the formation of aniline from cleavage of the C-S bond by solar irradiation

SC
373 (Bonvin et al. 2012, Zhou and Moore 1994). However, no peak matched an authentic

374 standard of aniline in this study. TP 253 was generated by rearrangement of the isoxazole

375
U
ring, which can be further degraded in direct photolysis. The accumulation of TP 253 and the
AN
376 smaller peak area of TP 271 with the presence of NaN3 indicated that TP 253 was one of the

precursors for TP 271. The reaction of 1O2 involved a cycloaddition to the electron-rich
M

377

378 olefinic double bond. The exact mechanism of hydroxylated products was not yet known.
D

379 Possible mechanisms may include a rearrangement of addition product to form epoxide
TE

380 (Frimer 1979), which could be further reduced by photosensitized electron transfer reactions

381 (Zhou and Moore 1997). Additionally, the formation of TP 271 and TP 287-1 in the presence
EP

382 of 1O2 and •OH scavengers indicated that an oxidation pathway still occurred. This was likely
C

383 attributed to the generation of intermediate radicals by t-BuOH and NaN3, as indicated by

previous studies that •CH2C(CH3)2OH and N3• possibly reacted with some organic
AC

384

385 compounds (NIST, 2016).

386 3.3.2. •OH pathway

•
387 OH exhibited high reactivity with olefinic double bonds and anilines (Buxton et al.

388 1988). One of the main reaction pathways of •OH was an addition reaction to the double bond

389 located on the isoxazole (oxazole) ring (Scheme 1). The formation of TP 287-1 and TP 287-2
- 17 -
 ACCEPTED MANUSCRIPT
390 were somewhat surprising since they were saturated alcohols. Hu et al. (Hu et al. 2007)

391 studied the transformation products of SMX by TiO2 photolysis and also found the formation

392 of TP 287 (no isomers). They proposed that the addition of •OH to C=C of the isoxazole ring

393 lead to the formation of a tertiary carbon-centered radical (Hu et al. 2007). This radical

394 reacted with oxygen to give a peroxy radical, which may abstract a hydrogen from a donor

PT
395 and form the corresponding hydroperoxide (Hu et al. 2007). The homolytic cleavage of

RI
396 hydroperoxide produced a hydroxyl radical and an alkoxy radical (Larson and Weber 1994),

397 which can further abstract a hydrogen atom to form the corresponding saturated alcohol

SC
398 product (Hu et al. 2007). This mechanism provided a possible explanation for the

399 degradation of TP 98 by •OH.

400
U
In the reaction of •OH with aniline, the abstraction of hydrogen is a preferred reaction
AN
401 pathway. •OH was expected to attack aniline on various positions. Solar et al.(Solar et al.

1986) investigated the absorbance characteristics of formed aniline radicals by •OH attack.
M

402

403 They found that the main primary radical (54%) resulted from the reaction of •OH with the
D

404 ortho-position of aniline, while 36% of •OH reacted directly with –NH2 group, and the
TE

405 remaining 10% of •OH probably attacked the para-position (Solar et al. 1986). Hydrolysis of

406 aniline radicals formed hydroxylated radical products, which further reacted with O2 to
EP

407 produce hydroxylated anilines. This mechanism explained the formation of TP 269 through
C

408 the reaction on the ortho- or para-positions of aniline. TP 502 was subsequently formed
AC

409 through coupling of the N-centered radical derived from –NH2 group.

410 3.3.3. SO4•- pathway

411 One important reaction pathway of SO4•- was initiated by the electrophilic attack at the

412 olefinic double bond in the isoxazole (oxazole) ring to generate olefinic radical (Scheme 2).

413 Hydrolysis of the olefinic radical and subsequent reactions are similar to that of •OH, which

414 gave rise to alcohol products (TP 287-1 and TP 287-2). The relative yield of TP 287-1 by
- 18 -
 ACCEPTED MANUSCRIPT
415 SO4•- was much higher than that by •OH, whereas the relative yield of TP 287-2 by •OH was

416 much higher than that by SO4•-. Because SO4•- was an electrophilic and more selective

417 oxidant than •OH, electron-donating groups could enhance the reactivity towards SO4•-. In

418 particular, the nitrogen atom in the oxazole ring could enhance electron density of the olefinic

419 double bond and exhibited stronger electron-donating properties than the olefinic double

PT
420 bond in the isoxazole ring. It indicated that SO4•- preferred to react with the oxazole ring than

RI
421 with the isoxazole ring, resulting in the higher relative yield of TP 287-1. The higher relative

422 yield of TP 287-2 by •OH was ascribed to the relatively higher concentration of SMX than TP

SC
423 253.

424 Furthermore, the reaction of SO4•- with –NH2 group was proposed to produce N-

425
U
centered radical (Scheme 2). The N-centered radical readily reacted with water, and this
AN
426 reaction was followed by a fast elimination of H+ to form anilino radical. The addition of
M

427 oxygen to the formed anilino radical resulted in the formation of hydroxylamine with the

428 release of HO2•. Further oxidation of hydroxylamine led to the formation of nitroso derivative
D

429 (Mahdi Ahmed et al. 2012). Hydrolysis of the nitroso derivative resulted in the formation of
TE

430 nitro derivative (i.e., TP 284). Meanwhile, the coupling of the N-centered radical generated

431 TP 502, which was further oxidized to generate TP 518. The different behaviors of TPs
EP

432 between •OH and SO4•- depended on their reactivity to specific groups of SMX. SO4•- would
C

433 attack –NH2 group more efficiently and thus favor the formation of nitrogen oxidized
AC

434 products.

435 3.3.4. CO3•- pathway

436 CO3•- was a more selective oxidant than •OH and SO4•-(Buxton et al. 1988, Neta et al.

437 1988, Neta et al. 1977), and thus reacted with organic compounds through electron transfer as

438 SO4•-. The conversion of •OH to less reactive CO3•- was responsible for the decrease of TP

439 269 (Fig. 3). CO3•- was more selective to attack –NH2 group, increasing the formations of TP
- 19 -
 ACCEPTED MANUSCRIPT
440 283 and TP 502 in both UV/H2O2 and UV/PDS. Interestingly, the formation of TP 518 was

441 slow in UV/H2O2 at initial 7 min, and then the accumulation of TP 518 was accelerated (Fig.

442 3). This result indicated that TP 518 was the secondary product and might derive from TP

443 502 or the coupling reaction between SMX and TP 269. However, this lag phase was not

444 obviously observed in UV/PDS (Fig. 4), which was ascribed to the higher formation yield of

PT
445 SO4•- than •OH to shorten the initial lag phase. Additionally, the formation of TP 518

increased in UV/H2O2, but decreased in UV/PDS, indicating that SO4•- was a stronger oxidant

RI
446

447 than CO3•- to form the further oxidized products. CO3•- attacked the secondary amine leading

SC
448 to cleavage of the N-S bond, which might explain the high yield of TP 98.

449 3.5. Toxicological implications of transformation products

450 3.5.1. Acute toxicity to Vibrio fischeri

U
AN
451 The luminescence inhibition of Vibrio fischeri by SMX after different treatments was
M

452 shown in Fig. 5. L0 is the luminescence of each sample without treatment, while L is the

453 luminescence of the sample with corresponding treatment. The ratio of L/L0 was used as an
D

454 indicator of acute toxicity; namely, the lower value of L/L0 indicated a higher acute toxicity,
TE

455 and vice versa. The dash line represents the luminescence induced by the remaining SMX in

456 the sample.

EP

457 With the degradation of SMX in UV and UV/H2O2, the luminescence slightly increased.
C

458 Comparatively, the luminescence of samples treated by UV/PDS reduced by 65%, indicating
AC

459 a high toxicity of TPs. The higher toxicity of the samples treated by UV/PDS was also

460 observed by Vibrio qinghaiensis in freshwater (Zhang et al. 2016). The presence of 50 mM

461 bicarbonate had no significant effect on luminescence in both UV/H2O2 and UV/PDS. This

462 result was ascribed to the fact that bicarbonate at 50 mM did not provide a CO3•- dominant

463 condition to destruct SMX under the conditions investigated. Unfortunately, due to the

464 limitation of the osmolality of Vibrio fischeri, a higher concentration of bicarbonate cannot be
- 20 -
 ACCEPTED MANUSCRIPT
465 applied in this experiment. Nevertheless, the similar products generated by SO4•-and CO3•-

466 implied that these products were more toxic than those generated by •OH.

467 3.5.2. Toxicity assessment of TPs by ECOSAR

468 The antibacterial activity of SMX is originated from its competition with p-

469 aminobenzoic acid for enzyme inhibition and metabolic interference, which is essential for

PT
470 bacterial folic acid synthesis (Majewsky et al. 2014). Three types of TPs were observed in

RI
471 this study: (i) TPs retaining the sulfonamide toxicophore; (ii) breakdown TPs not exhibiting

472 the toxicophore; (iii) dimeric products. To demonstrate different responses of SMX and TPs

SC
473 on various species (i.e., fish, daphnid and green algae), QSAR analysis by ECOSAR program

474 was conducted (Table S4). According to the previous result of 48-h half effective

475
U
concentration (EC50) value for D. magna (Kim et al. 2007), the value of class aniline
AN
476 (unhindered) was selected for prediction.
M

477 Except for TP 253, the acute and chronic toxicity of four TPs retaining sulfonamide

478 group exhibited lower toxicity than those of SMX. This modeling result showed a similar
D

479 trend with the experimental result using Vibrio fischeri (Majewsky et al. 2014). Additionally,
TE

480 Lienert et al. (Lienert et al. 2007) suggested that the toxicity of parent compound and its

481 metabolite was positively related to their octanol-water partition coefficients (Kow). The
EP

482 hydroxylated products of SMX increased hydrophilicity, which reduced the toxicity of TPs.
C

483 TP 283 was expected to possess higher toxicity. Majewsky et al. (Majewsky et al. 2014)
AC

484 investigated the antibacterial activity of selected SMX TPs in the aquatic environment. Their

485 results indicated that the transformation of SMX at the para-position still remained or even

486 increased its toxicity. The increasing toxicity of TP 283 was explained by a negative

487 inductive effect of the nitro group, which might either favor release of the proton in the amide

488 group to enhance the affinity of the key enzyme for folic acid synthesis, or replace p-

489 hydroxybenzoic acid to form faulty products (Majewsky et al. 2014).

- 21 -
 ACCEPTED MANUSCRIPT
490 For breakdown products, TP 173 and TP 189 were less toxic than SMX. In contrast,

491 higher toxicity of TP 98 was predicted by ECOSAR, which was opposite to the experimental

492 data (Majewsky et al. 2014). The prediction of the isoxazole ring by the class of aniline

493 (unhindered) might lead to a large bias to the real value.

494 For dimeric products, both TP 502 and TP 518 exhibited higher toxicity towards fish.

PT
495 The mechanism of azo carcinogenicity was proposed including: (i) cleavage of the azo bond

RI
496 to generate anilines during metabolic reactions; (ii) oxidation of azo to highly reactive

497 electrophilic diazonium salts (Brown and De Vito 1993). It was likely that the oxidation of

SC
498 SO4•- or CO3•- with SMX generated more toxic products than that of •OH.

499 3.6. Oxidation in authentic water

500
U
Most of the TPs identified in the synthetic solutions could also be found in the authentic
AN
501 waters (Fig. 6). No dimeric products were observed in both water samples treated by
M

502 UV/H2O2. This result might be ascribed to the relatively low concentration of SMX in the

503 authentic waters than that in the synthetic water, resulting in their peak areas below the
D

504 detection limit. Interestingly, TP 283 was detected in both water samples treated by UV/PDS,
TE

505 while TP 502 and TP 518 were only identified in the ground water. The absorption at 254 nm

506 represented the content of aromatic moieties of DOC. The ground water contained higher
EP

507 DOC with lower UV254 in relative to those of the surface water. As selective radicals, SO4•-

and CO3•- might be more reactive to DOC in the surface water than that in the groundwater.
C

508

Amine functional group on the aromatic structure also increased its reactivity towards SO4•-
AC

509

510 and CO3•-. Therefore, the N-centered radical derived from SMX might couple with another

511 N-centered radical derived from DOC, and lead to the lower formation of TP 502 and TP 518

512 in the surface water. The impact of natural organic matter properties on TPs formation and

513 following toxicological potential need further investigation.

- 22 -
 ACCEPTED MANUSCRIPT
514 4. Conclusions.

515 SO4•- showed the higher reactivity towards the anionic form of SMX than the neutral

516 form due to the electrophilic nature of SO4•-, while •OH exhibited the comparable reactivity to

517 SMX species. The generation of the TPs in direct photolysis of SMX was derived from

518 cleavage of the N-S bond, rearrangement of the isoxazole ring, and hydroxylation of SMX.

PT
519 Although direct photolysis of SMX was the predominant pathway at pH 3, the combination

RI
520 of •OH with direct photolysis enhanced the formation of hydroxylated products, while SO4•-

521 further oxidized the TPs generated by direct photolysis. At pH 8, •OH was confirmed to

SC
522 produce hydroxylated products, while SO4•- favored attack on –NH2 group, and thus

523 generating the nitro derivative. Additionally, SO4•- promoted the formation of oxazole ring

U
oxidized product rather than that of the isoxazole ring, while •OH reacted in the opposite
AN
524

525 pathway. More dimeric products were determined by SO4•-, resulting from recombination of
M

526 the N-centered radical.

527 The high concentration of bicarbonate scavenged •OH and SO4•- to generate CO3•-, which
D

528 was a more selective oxidant than •OH and SO4•-. CO3•- exhibited higher reactivity towards –
TE

529 NH2 group, and thus increasing the formation of the TPs related to the N-centered radical, but

530 inhibiting the formation of hydroxylated products on the benzene ring. Both the acute toxicity
EP

531 experiments and ECOSAR assessments suggested the TPs generated by SO4•- or CO3•-
C

532 possess higher toxicity than those by •OH. Nevertheless, further studies are needed to
AC

533 evaluate the impact of natural organic matter properties on TPs formation and their

534 toxicological potential when UV-based oxidation processes are applied in different

535 background waters.

536 Acknowledgments

537 This research was financially supported by the National Natural Science Foundation of

- 23 -
 ACCEPTED MANUSCRIPT
538 China (No. 51578203), the Funds of State Key Laboratory of Urban Water Resource and

539 Environment (HIT, 2016DX13), and the Foundation for the Author of National Excellent

540 Doctoral Dissertation of China (201346). We are also thankful to the International

541 Postdoctoral Exchange Fellowship Program (No. 20160074) to Y.Y. and the Excellent

542 Graduate Student Scholarship from the Shanghai Tongji Gao Tingyao Environmental Science

PT
543 and Technology Development Foundation awarded to X.L.

RI
544

545 Appendix. Supplemental material

SC
546 Supplementary information for this manuscript can be downloaded.

547

U
AN
548 References

549 Al Aukidy, M., Verlicchi, P., Jelic, A., Petrovic, M. and Barcelò, D. (2012) Monitoring
550 release of pharmaceutical compounds: Occurrence and environmental risk assessment of two
M

551 WWTP effluents and their receiving bodies in the Po Valley, Italy. Science of the Total
552 Environment 438, 15-25.
553 Anastasio, C. and Matthew, B.M. (2006) A chemical probe technique for the
D

554 determination of reactive halogen species in aqueous solution: Part 2 - chloride solutions and
555 mixed bromide/chloride solutions. Atmospheric Chemistry and Physics 6, 2439-2451.
TE

556 Baeza, C. and Knappe, D.R.U. (2011) Transformation kinetics of biochemically active
557 compounds in low-pressure UV Photolysis and UV/H2O2 advanced oxidation processes.
558 Water Research 45(15), 4531-4543.
EP

559 Bonvin, F., Omlin, J., Rutler, R., Schweizer, W.B., Alaimo, P.J., Strathmann, T.J.,
560 McNeill, K. and Kohn, T. (2012) Direct Photolysis of Human Metabolites of the Antibiotic
561 Sulfamethoxazole: Evidence for Abiotic Back-Transformation. Environmental Science &
562 Technology 47(13), 6746-6755.
C

563 Bonvin, F., Rutler, R., Chèvre, N., Halder, J. and Kohn, T. (2011) Spatial and Temporal
564 Presence of a Wastewater-Derived Micropollutant Plume in Lake Geneva. Environmental
AC

565 Science & Technology 45(11), 4702-4709.

566 Boreen, A., Arnold, W. and McNeill, K. (2003) Photodegradation of pharmaceuticals in
567 the aquatic environment: A review. Aquatic Sciences 65(4), 320-341.
568 Boreen, A.L., Arnold, W.A. and McNeill, K. (2004) Photochemical Fate of Sulfa Drugs
569 in the Aquatic Environment:  Sulfa Drugs Containing Five-Membered Heterocyclic Groups.
570 Environmental Science & Technology 38(14), 3933-3940.
571 Brown, M.A. and De Vito, S.C. (1993) Predicting azo dye toxicity. Critical Reviews in
572 Environmental Science and Technology 23(3), 249-324.
573 Buxton, G.V., Greenstock, C.L., Helman, W.P. and Ross, A.B. (1988) Critical review of
574 rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals.
- 24 -
 ACCEPTED MANUSCRIPT
575 Journal of Physical and Chemical Reference Data 17(2), 513-886.
576 Canonica, S., Kohn, T., Mac, M., Real, F.J., Wirz, J. and von Gunten, U. (2005)
577 Photosensitizer method to determine rate constants for the reaction of carbonate radical with
578 organic compounds. Environmental Science & Technology 39(23), 9182-9188.
579 Canonica, S., Meunier, L. and von Gunten, U. (2008) Phototransformation of selected
580 pharmaceuticals during UV treatment of drinking water. Water Research 42(1–2), 121-128.
581 Frimer, A.A. (1979) The reaction of singlet oxygen with olefins: the question of
582 mechanism. Chemical Reviews 79(5), 359-387.
583 Gao, S., Zhao, Z., Xu, Y., Tian, J., Qi, H., Lin, W. and Cui, F. (2014) Oxidation of

PT
584 sulfamethoxazole (SMX) by chlorine, ozone and permanganate—A comparative study.
585 Journal of Hazardous materials 274, 258-269.
586 Grebel, J.E., Pignatello, J.J. and Mitch, W.A. (2010) Effect of halide ions and carbonates

RI
587 on organic contaminant degradation by hydroxyl radical-based advanced oxidation processes
588 in saline waters. Environmental Science & Technology 44(17), 6822-6828.
589 Gsponer, H.E., Previtali, C.M. and García, N.A. (1987) Kinetics of the photosensitized

SC
590 oxidation of polychlorophenols in alkaline aqueous solution. Toxicological and
591 Environmental Chemistry 16(1), 23-37.
592 Hu, L., Flanders, P.M., Miller, P.L. and Strathmann, T.J. (2007) Oxidation of
593 sulfamethoxazole and related antimicrobial agents by TiO2 photocatalysis. Water Research

U
594 41(12), 2612-2626.
595 Ji, Y., Fan, Y., Liu, K., Kong, D. and Lu, J. (2015) Thermo activated persulfate
AN
596 oxidation of antibiotic sulfamethoxazole and structurally related compounds. Water Research
597 87, 1-9.
598 Keen, O.S. and Linden, K.G. (2013) Degradation of Antibiotic Activity during
M

599 UV/H2O2 Advanced Oxidation and Photolysis in Wastewater Effluent. Environmental

600 Science & Technology 47(22), 13020-13030.
601 Kim, Y., Choi, K., Jung, J., Park, S., Kim, P.-G. and Park, J. (2007) Aquatic toxicity of
D

602 acetaminophen, carbamazepine, cimetidine, diltiazem and six major sulfonamides, and their
603 potential ecological risks in Korea. Environment International 33(3), 370-375.
604 Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B.
TE

605 and Buxton, H.T. (2002) Pharmaceuticals, Hormones, and Other Organic Wastewater
606 Contaminants in U.S. Streams, 1999−2000:  A National Reconnaissance. Environmental
607 Science & Technology 36(6), 1202-1211.
EP

608 Larson, R.A. and Weber, E.J. (1994) Reaction mechanisms in environmental organic
609 chemistry, CRC press.
610 Lienert, J., Gudel, K. and Escher, B.I. (2007) Screening method for ecotoxicological
C

611 hazard assessment of 42 pharmaceuticals considering human metabolism and excretory

612 routes. Environmental Science and Technology 41(12), 4471-4478.
AC

613 Lin, H., Niu, J., Xu, J., Li, Y. and Pan, Y. (2013) Electrochemical mineralization of
614 sulfamethoxazole by Ti/SnO2-Sb/Ce-PbO2 anode: Kinetics, reaction pathways, and energy
615 cost evolution. Electrochimica Acta 97, 167-174.
616 Mahdi Ahmed, M., Barbati, S., Doumenq, P. and Chiron, S. (2012) Sulfate radical anion
617 oxidation of diclofenac and sulfamethoxazole for water decontamination. Chemical
618 Engineering Journal 197, 440-447.
619 Majewsky, M., Wagner, D., Delay, M., Bräse, S., Yargeau, V. and Horn, H. (2014)
620 Antibacterial Activity of Sulfamethoxazole Transformation Products (TPs): General
621 Relevance for Sulfonamide TPs Modified at the para Position. Chemical Research in
622 Toxicology 27(10), 1821-1828.
623 Morasch, B., Bonvin, F., Reiser, H., Grandjean, D., De Alencastro, L.F., Perazzolo, C.,
- 25 -
 ACCEPTED MANUSCRIPT
624 Chèvre, N. and Kohn, T. (2010) Occurrence and fate of micropollutants in the Vidy Bay of
625 Lake Geneva, Switzerland. Part II: Micropollutant removal between wastewater and raw
626 drinking water. Environmental Toxicology and Chemistry 29(8), 1658-1668.
627 Neta, P., Huie, R.E. and Ross, A.B. (1988) Rate constants for reactions of inorganic
628 radicals in aqueous solution. Journal of Physical and Chemical Reference Data 17(3), 1027-
629 1284.
630 Neta, P., Madhavan, V., Zemel, H. and Fessenden, R.W. (1977) Rate constants and
631 mechanism of reaction of sulfate radical anion with aromatic compounds. Journal of the
632 American Chemical Society 99(1), 163-164.

PT
633 NIST, 2016. National Institute of Standards and Technology. NDRL/NIST Solution
634 Kinetics Database on the Web. http://kinetics.nist.gov/solution/ (Accessed May13, 2016).
635 Padhye, L.P., Yao, H., Kung'u, F.T. and Huang, C.-H. (2014) Year-long evaluation on

RI
636 the occurrence and fate of pharmaceuticals, personal care products, and endocrine disrupting
637 chemicals in an urban drinking water treatment plant. Water Research 51, 266-276.
638 Rahn, R.O., Stefan, M.I., Bolton, J.R., Goren, E., Shaw, P.S. and Lykke, K.R. (2003)

SC
639 Quantum yield of the iodide-iodate chemical actinometer: Dependence on wavelength and
640 concentration. Photochemistry and Photobiology 78(2), 146-152.
641 Ratola, N., Cincinelli, A., Alves, A. and Katsoyiannis, A. (2012) Occurrence of organic
642 microcontaminants in the wastewater treatment process. A mini review. Journal of Hazardous

U
643 materials 239–240, 1-18.
644 Rodgers, M.A.J. (1983) Solvent-induced deactivation of singlet oxygen: additivity
AN
645 relationships in nonaromatic solvents. Journal of the American Chemical Society 105(20),
646 6201-6205.
647 Solar, S., Solar, W. and Getoff, N. (1986) Resolved multisite OH-attack on aqueous
M

648 aniline studied by pulse radiolysis. International Journal of Radiation Applications and
649 Instrumentation. Part C. Radiation Physics and Chemistry 28(2), 229-234.
650 Ternes, T.A. and Joss, A. (2006) Human pharmaceuticals, hormones and fragrances: the
D

651 challenge of micropollutants in urban water, IWA Publishing, London.

652 Trovó, A.G., Nogueira, R.F.P., Agüera, A., Sirtori, C. and Fernández-Alba, A.R. (2009)
653 Photodegradation of sulfamethoxazole in various aqueous media: Persistence, toxicity and
TE

654 photoproducts assessment. Chemosphere 77(10), 1292-1298.

655 Wols, B., Hofman-Caris, C., Harmsen, D. and Beerendonk, E. (2013) Degradation of 40
656 selected pharmaceuticals by UV/H 2 O 2. Water Research 47(15), 5876-5888.
EP

657 Yang, Y., Jiang, J., Lu, X., Ma, J. and Liu, Y. (2015) Production of Sulfate Radical and
658 Hydroxyl Radical by Reaction of Ozone with Peroxymonosulfate: A Novel Advanced
659 Oxidation Process. Environmental Science & Technology 49(12), 7330-7339.
C

660 Yang, Y., Pignatello, J.J., Ma, J. and Mitch, W.A. (2014) Comparison of halide impacts
661 on the efficiency of contaminant degradation by sulfate and hydroxyl radical-based advanced
AC

662 oxidation processes (AOPs). Environmental Science & Technology 48(4), 2344-2351.
663 Yang, Y., Pignatello, J.J., Ma, J. and Mitch, W.A. (2016) Effect of matrix components
664 on UV/H2O2 and UV/S2O82− advanced oxidation processes for trace organic degradation in
665 reverse osmosis brines from municipal wastewater reuse facilities. Water Research 89, 192-
666 200.
667 Zhang, R., Sun, P., Boyer, T.H., Zhao, L. and Huang, C.-H. (2015) Degradation of
668 Pharmaceuticals and Metabolite in Synthetic Human Urine by UV, UV/H2O2, and UV/PDS.
669 Environmental Science & Technology 49(5), 3056-3066.
670 Zhang, R., Yang, Y., Huang, C.-H., Li, N., Liu, H., Zhao, L. and Sun, P. (2016)
671 UV/H2O2 and UV/PDS Treatment of Trimethoprim and Sulfamethoxazole in Synthetic
672 Human Urine: Transformation Products and Toxicity. Environmental Science & Technology
- 26 -
 ACCEPTED MANUSCRIPT
673 50(5), 2573-2583.
674 Zhou, W. and Moore, D.E. (1994) Photochemical decomposition of sulfamethoxazole.
675 International Journal of Pharmaceutics 110(1), 55-63.
676 Zhou, W. and Moore, D.E. (1997) Photosensitizing activity of the anti-bacterial drugs
677 sulfamethoxazole and trimethoprim. Journal of Photochemistry and Photobiology B: Biology
678 39(1), 63-72.
679
680

PT
RI
U SC
AN
M
D
TE
C EP
AC

- 27 -
 ACCEPTED MANUSCRIPT
681 A

1.2x1010

1.0x1010
k OH,app (M-1 s-1)

8.0x109

PT
6.0x109

4.0x109
●

RI
2.0x109

0.0

SC
3 4 5 6 7 8
pH
682

683 B

U
AN
1.4x1010
1.2x1010
M
(M-1 s-1)

1.0x1010
8.0x109
D
,app

6.0x109
●-

TE
kSO

4.0x109
2.0x109
0.0
EP

3 4 5 6 7 8
pH
684
C

685 Fig. 1. Apparent second order rate constants for reactions of SMX with •OH (A) and SO4•- (B)
AC

686 as a function of pH.

687

- 28 -
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP

688
C

689 Fig. 2. Degradation of SMX and evolution of main TPs during UV, UV/H2O2 and UV/PDS at
AC

690 pH 3. Peak areas were normalized using the peak area of the internal standard. Experimental

691 condition: [SMX]0 = 20 µM, [H2O2]0 =1 mM, and [PDS]0 =1 mM.

692

- 29 -
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP

693
C

694 Fig. 3. Effect of bicarbonate on degradation of SMX and evolution of main TPs during
AC

695 UV/H2O2 at pH 8. Peak areas were normalized using the peak area of the internal standard.

696 Experimental condition: [SMX]0 = 20 µM, and [H2O2]0 =1 mM.

697

- 30 -
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D

698

699 Fig. 4. Effect of bicarbonate on degradation of SMX and evolution of main TPs during
TE

700 UV/PDS at pH 8. Peak areas were normalized using the peak area of the internal standard.
EP

701 Experimental condition: [SMX]0 = 20 µM, and [PDS]0 =1 mM.

702
C
AC

- 31 -
 ACCEPTED MANUSCRIPT

PT
RI
SC
703

704 Scheme 1. Proposed reaction pathway for SMX degradation by UV and UV/H2O2. The solid

705
U
arrows represent major pathways, and dash arrows represent minor pathways.
AN
706
M

707

708
D
TE
C EP
AC

- 32 -
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
709
M

710 Scheme 2. Proposed reaction pathway for SMX degradation by UV/PDS in the absence and

711 presence of bicarbonate. The solid arrows represent major pathways, and dash arrows
D

712 represent minor pathways.

TE
C EP
AC

- 33 -
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M

713

714 Fig. 5. Impact on Vibrio fischeri luminescence by SMX after different treatments. The dash
D

715 line represents the luminescence induced by the remaining SMX. Errors represent the
TE

716 standard deviation (n = 4).

717
C EP
AC

- 34 -
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP

718

719 Fig. 6. SMX degradation and main TPs formation in authentic waters. SW and GW represent
AC

720 the surface water and the ground water, respectively. Peak areas were normalized using the

721 peak area of the internal standard. Experimental condition: [SMX]0 = 1 µM, [H2O2]0 or

722 [PDS]0 =1 mM.

723

724
- 35 -
 ACCEPTED MANUSCRIPT

Reaction rate constants of •OH and SO4•- with sulfamethoxazole were determined.

Hydroxylated products was predominant in UV/H2O2.

SO4•- favored attack on –NH2 group to form a nitro derivative and dimeric
products.

CO3•- enhanced oxidation of –NH2 group.

PT

SO4•- or CO3•- oxidation generated more toxic products than •OH oxidation.

RI
U SC
AN
M
D
TE
C EP
AC