Carbon, Hydrogen, and Nitrogen Isotope Fractionation Trends in N

Oct 16, 2017 - (19-21) CSIA was applied, for example, to monitor changes of 13C/12C ratios in chloroform produced upon chlorination of lake water. ...
1 downloads 13 Views 458KB Size
Subscriber access provided by Eastern Michigan University | Bruce T. Halle Library

Article

Carbon, hydrogen, and nitrogen isotope fractionation trends in N-nitrosodimethylamine reflect the formation pathway during chloramination of tertiary amines Stephanie Spahr, Urs von Gunten, and Thomas B. Hofstetter Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b03919 • Publication Date (Web): 16 Oct 2017 Downloaded from http://pubs.acs.org on October 16, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

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

Page 1 of 32

Environmental Science & Technology

Carbon, Hydrogen, and Nitrogen Isotope Fractionation Trends in N-Nitrosodimethylamine Reflect the Formation Pathway during Chloramination of Tertiary Amines Stephanie Spahr,†,‡ Urs von Gunten,†,‡, ¶ and Thomas B. Hofstetter∗,†, ¶ Eawag, Swiss Federal Institute of Aquatic Science and Technology, CH-8600 Dübendorf, Switzerland, School of Architecture, Civil and Environmental Engineering (ENAC), Ecole Polytechnique Federale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland, and Institute of Biogeochemistry and Pollutant Dynamics, ETH Zürich, CH-8092 Zürich, Switzerland E-mail: [email protected]

∗ To

whom correspondence should be addressed

† Eawag ‡ EPF ¶ ETH

Lausanne Zürich

1 ACS Paragon Plus Environment

Environmental Science & Technology

Abstract

1

2

Assessing the precursors and reactions leading to the carcinogenic N-nitrosodimethylamine

3

(NDMA) during drinking water disinfection is a major challenge. Here, we investigate whether

4

changes of

5

trends that can be used to infer NDMA formation pathways. We carried out compound-

6

specific isotope analysis (CSIA) of NDMA during chloramination of four tertiary amines that

7

produce NDMA at high yields, namely ranitidine, 5-(dimethylaminomethyl)furfuryl alcohol,

8

N,N-dimethylthiophene-2-methylamine and N,N-dimethylbenzylamine. Carbon and hydrogen

9

isotope ratios of NDMA function as fingerprints of the N(CH3 )2 moiety and exhibit only

10

minor isotope fractionation during the disinfection process. Nitrogen isotope ratios showed

11

that NH2 Cl is the source of the N atom of the nitroso group. The large enrichment of 15 N in

12

NDMA was indicative of the isotope effects pertinent to bond cleavage and bond formation

13

reactions during chloramination of tertiary amines. Correlation of δ15 N vs. δ13 C values of

14

NDMA resulted in trend lines that were not affected by the type of tertiary amine and treatment

15

conditions suggesting that the observed C and N isotope fractionation in NDMA may be

16

diagnostic for NDMA precursors and formation pathways during chloramination.

13 C/12 C, 2 H/1 H,

and

15 N/14 N

ratios of NDMA give rise to isotope fractionation

2 ACS Paragon Plus Environment

Page 2 of 32

Page 3 of 32

Environmental Science & Technology

17

Introduction

18

N-Nitrosamines form as drinking water disinfection by-products (DBPs) and are of public and reg-

19

ulatory concern due to their mutagenicity and potential carcinogenicity. 1,2 N-nitrosodimethylamine

20

(NDMA) is a frequently detected DBP in finished drinking waters and often exceeds guidance val-

21

ues of 9-100 ng/L. 2–5 NDMA is produced unintentionally with typically used disinfectants, that is

22

chlorine, chloramine, and ozone through reactions with organic compounds in raw waters including

23

natural organic matter, 6–9 anthropogenic contaminants such as pharmaceuticals and pesticides, 10–14

24

and chemicals used for water treatment (e.g., polymeric coagulants). 15,16 The different molecular

25

structures of NDMA precursors have led to the conclusion that pathways of NDMA generation

26

differ widely among the various disinfection procedures. 17,18 Nevertheless, detailed knowledge of

27

chemical reaction mechanisms remains scarce so that a systematic prediction and prevention of

28

NDMA formation during water treatment is currently hampered.

29

One promising new tool with which NDMA can be related to its precursors and formation

30

reactions is compound-specific isotope analysis (CSIA). Previous studies have demonstrated that

31

changes in the natural stable isotope composition of DBPs, that is the so called stable isotope

32

fractionation, provide evidence for reactive precursor materials and DBP formation pathways. 19–21

33

CSIA was applied, for example, to monitor changes of 13 C/12 C ratios in chloroform produced upon

34

chlorination of lake water. 19 Chloroform was enriched in 12 C due to a preferential reaction of light

35

(i.e., 12 C containing) isotopologues when produced from resorcinol-like moieties of natural organic

36

matter (NOM). Conversely, chloroform was enriched in

37

of phenolic functional groups of NOM. These contrasting isotopic preferences reflect different

38

chloroform formation pathways. Each of them exhibits different kinetic isotope effects (KIEs),

39

which reflect which chemical bonds are broken and formed. 22,23 Because KIEs are specific features

40

of a reaction mechanism, the stable isotope fractionation observed in the reactants and products

41

can serve as proxy for the transformation pathway. 24–28

13 C

when generated from chlorination

42

We have recently introduced analytical procedures for 13 C/12 C, 2 H/1 H, and 15 N/14 N isotope ratio

43

measurements of N-nitrosamines. 29 However, this method has not been applied to study disinfection 3 ACS Paragon Plus Environment

Environmental Science & Technology

Page 4 of 32

44

processes and it is currently unknown how C, H, and N isotope fractionation in NDMA can be

45

indicative of its formation pathway(s). Elucidation of reaction mechanisms with CSIA typically

46

focuses on the analysis of substrate disappearance, where substrate isotope fractionation reveals the

47

isotope effects of the first elementary reaction steps leading to irreversible bond cleavage(s). 24–27

48

Gaining such information from product isotope fractionation is less common and CSIA of reaction

49

products is applied mostly to study isotope effects of well-defined reactions leading to known

50

and measurable products (e.g., 30–34 ). None of the previous applications of CSIA reflect cases

51

similar to the one of NDMA, which forms in a sequence of (partially) unknown reactions. Each

52

of these reactions exhibits an isotope effect and thus contributes to the final isotopic composition

53

of NDMA. 11,35 However, the identification of reactive processes from the isotope fractionation in

54

reaction products is a common approach in (bio)geosciences. In this discipline, the elucidation

55

of events in the past is often made through isotopic analysis of frequently occurring molecules

56

and minerals in natural samples such as hydrocarbons, sulfur species, oxides, and carbonates. 36,37

57

The methodology is based on observations made in multiple isotope systems, that is from the

58

comparison of isotope fractionation from different elements in a molecule (e.g., 2 H/1 H vs.

59

of methane,

18 O/16 O

vs.

60

18 O/16 O

17 O/16 O

in O2 and 34 S/32 S vs.

61

understanding of isotope effects of individual reactions, evidence for reactive processes can then

62

be obtained from correlations of isotope fractionation, which, in most cases, result in indicative

63

isotope fractionation trendlines.

vs.

34 S/32 S

13 C/12 C

in sulfate) and different isotopes of the same elements (e.g., 33 S/32 S

in sulfates). 38–43 Regardless of a quantitative

64

The objective of this work was to evaluate whether multi-element isotope fractionation analysis

65

is applicable to track precursor moieties and formation pathways of NDMA through analysis

66

of C, H, or N isotope ratios of the DBP. This work builds on two of our recent studies on

67

the NDMA formation from the chloramination of tertiary amines 29,35 and information on the

68

reaction mechanism provided by others. 17,44–46 Based on the evidence of radical intermediates and

69

reactions of molecular oxygen in the NDMA formation pathway, 35 we hypothesized that tertiary

70

amines undergo a sequence of substitution, electron transfer, oxygenation, and radical coupling

4 ACS Paragon Plus Environment

Page 5 of 32

Environmental Science & Technology

H H 3C

-H+

O

N

R

H 3C

O

H 3C

N

+H+

products

R

CH3

1

2

H 2N

products

Cl

Tertiary amine precursors used in this work

H N

O N

Cl –

H N

S NO 2

Ranitidine

NH 2

H 3C

R

products

H 3C

O N

O

N 3

OH

H 2N

Cl

DFUR Cl – + NH 2

H+ S N NH H 3C

DMTA

O

N

R

products

R

products

H 3C N

4

O2 DMBA O O NH H 3C

O N

H 3C 5

O

O

O

O

HN O R

N

NH CH3

H 3C

CH3

H 3C

O N

R 6

H 2O2 H 3C N N O

2

+ 2

H 2C

O R

H 3C 7

8

Scheme 1 NDMA formation pathway during chloramination of selected tertiary amines (adapted from 35 ) including the precursor compounds examined in this study (ranitidine, DFUR, DMTA, and DMBA). At pH 8.0 deprotonation of tertiary amines (1 to 2) occurs prior to nucleophilic attack on NH2 Cl. Reactive intermediates include substituted dimethylhydrazines (3), aminyl radicals (4), Nperoxyl radicals (5), and tetroxide species (6). Decomposition of 6 leads to two equivalents of NDMA (7) and carbocations (8). Based on a molar NDMA yield < 100%, precursors or intermediates 3 - 5 also undergo reactions to unidentified products.

5 ACS Paragon Plus Environment

Environmental Science & Technology

71

reactions shown in Scheme 1. Those reactions will be the source of isotope fractionation in

72

NDMA. In our previous work on stable isotope analysis, 29 we have used the chloramination of

73

reaction of ranitidine to validate our methodology and we found that NDMA formation processes

74

can be studied at low µM concentrations if NDMA yields are high. Therefore, we focused our

75

current investigation on chloramination of tertiary amines, which can give rise to yields of NDMA

76

that exceed 60%. 11,14,29,35,47 We conducted chloramination experiments with four tertiary amines,

77

namely ranitidine, 5-(dimethylaminomethyl)furfuryl alcohol (DFUR), N,N-dimethylthiophene-2-

78

methylamine (DMTA), and N,N-dimethylbenzylamine (DMBA, see Scheme 1), and investigated

79

C, H, and N isotope fractionation in NDMA during its formation from these precursors. First,

80

we inferred the sources of C, H, and N atoms in NDMA by examining the initial and site-specific

81

isotope ratios of selected precursors and by conducting experiments with isotopically distinct

82

monochloramines. Second, we investigated to which extent the C and N isotope fractionation

83

trends in two model tertiary amines and NDMA reflect NDMA formation pathways involving

84

multiple isotope-sensitive reaction steps. Finally, we assessed whether the multi-element isotope

85

fractionation trends of NDMA are characteristic for chloramination of tertiary amines and could

86

potentially be used as proxy for the NDMA formation pathway during drinking water disinfection.

87

Experimental Section

88

Chemicals

89

A list of all chemicals including suppliers and purities is provided in the Supporting Information

90

(SI).

91

NDMA formation experiments

92

Monochloramine (NH2 Cl) stock solutions (30 mM) were prepared daily as described previously 48,49

93

by mixing hypochlorite (OCl – ) with either ammonium chloride (NH4 Cl) or ammonium sulfate

6 ACS Paragon Plus Environment

Page 6 of 32

Page 7 of 32

Environmental Science & Technology

94

((NH4 )2 SO4 ) at pH 9.5 with a molar Cl:N ratio of 1:1.05. The ammonium salts exhibited distinctly

95

different and well defined N isotope ratios corresponding to N isotope signatures, δ15 N, of −1.4h

96

and +53.7h for NH4 Cl and (NH4 )2 SO4 , respectively. 50

97

Chloramination experiments were carried out in 14 amber glass bottles containing 1 L of either

98

10 mM phosphate buffer (pH 8.0), 10 mM phosphate buffer (pH 7.0), 50 mM phosphate buffer (pH

99

8.0), or 10 mM borate buffer (pH 8.0). Each reactor was spiked with 100 µL of a methanolic stock

100

solution to obtain initial concentrations of 3 µM ranitidine or DFUR and 40 µM DMTA or DMBA.

101

The formation of NDMA was initiated through addition of NH2 Cl in 15-fold excess corresponding

102

to initial NH2 Cl concentrations of 45 µM or 600 µM, respectively. At predefined time points,

103

one 1 L reactor was sacrificed for chemical analyses. We measured the solution pH and NH2 Cl

104

concentration, and quenched the chloramine reaction by adding 0.5 g of Na2 S2 O3 to the reactor.

105

To quantify the concentrations of NDMA, ranitidine, and DFUR, 1 mL of the solution was filled

106

into an 1.5 mL amber autosampler glass vial. For analyses of DMTA and DMBA, 40 mL of the

107

solution was transferred into 50 mL amber glass flasks and the pH of the solution was adjusted to

108

11.3 through addition of 5 M NaOH. All samples were stored in the dark at 4℃ until concentration

109

analyses and further processing for stable isotope analyses. Two control experiments were set up

110

(i) to quantify losses of the organic amine precursor in the absence of NH2 Cl and (ii) to determine

111

the self-decay rate of NH2 Cl in the absence of organic amines. Unless stated otherwise, reported

112

NH2 Cl concentrations were corrected by the self-decay of NH2 Cl. The presence of methanol had

113

no effect on the consumption of NH2Cl and the formation of NDMA as shown previously. 35

114

Chemical analyses

115

The concentration of aqueous NH2 Cl stock solutions (30 mM) was quantified as described pre-

116

viously using a Varian Cary 100 Bio UV-visible spectrophotometer. 29,48,49,51 In reaction mix-

117

tures containing tertiary amines, reactive intermediates, and NDMA, NH2 Cl was quantified with

118

a colorimetric method using 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium

119

salt (ABTS). 29,52 A detailed comparison of chloramine quantification methods can be found in the 7 ACS Paragon Plus Environment

Environmental Science & Technology

120

SI section S5 of Spahr et al. 35 . Concentrations of NDMA, ranitidine, and DFUR were determined

121

by reverse phase HPLC with UV detection (Dionex UltiMate 3000). 29,35

122

Concentrations of DMTA and DMBA were measured by solid-phase microextraction (SPME)

123

coupled to a GC/MS (Thermo TRACE GC Ultra and Thermo TRACE DSQ II). 2 mL amber

124

autosampler glass vials, which contained 0.3 g NaCl, were filled with 1.3 mL of aqueous sample

125

in 10 mM phosphate buffer (pH 11.3) and shaken on a Vortex mixer. Direct immersion SPME

126

was carried out with a polydimethylsiloxane/divinylbenzene coated fibre (PDMS/DVB, 65 µm,

127

Supelco) after conditioning the fiber daily for 30 min at 250℃. Analytes were extracted for 45 min

128

at 40℃ and desorbed within 3 min at 270℃ in the split/splitless injector of the GC. 53 The GC was

129

equipped with 1 m DPTMDS (methyl/phenyl) deactivated fused-silica guard column (0.53 mm i.d.,

130

BGB) and a 30 m × 0.25 mm ZB-5ms column (0.25 µm, Zebron, Phenomenex). Helium carrier gas

131

was used at a constant pressure of 130 kPa. The temperature program was 1 min at 50℃, 10℃/min

132

to 250℃, and 5 min at 250℃. DMTA and DMBA concentrations were quantified with an external

133

calibration of 0.1 - 1.5 µM.

134

Stable isotope analyses

135

Stable C, H, and N isotope ratios of NDMA were measured using gas chromatography isotope

136

ratio mass spectrometry (GC/IRMS) coupled to solid-phase extraction (SPE) as reported in Spahr

137

et al. 29 C and N isotope analysis of DMTA and DMBA in aqueous samples was conducted with

138

SPME-GC/IRMS. The SPME procedure, GC setup, and temperature program was identical to

139

that for GC/MS analysis, but a 30 m × 0.32 mm ZB-5ms column (1 µm, Zebron, Phenomenex) was

140

used. For all C and N isotope measurements, a Ni/Ni/Pt reactor was operated at 1000℃. 29 Method

141

quantification limits (MQLs) of the SPME-GC/IRMS measurements of DMTA and DMBA were

142

determined according to the moving mean procedure of Jochmann et al. 54 and measurements were

143

made in concentration ranges of 0.3 - 0.6µM and 2.5 -16 µM for DMTA and DMBA, respectively

144

(Figure S1). The 15 N equilibrium isotope effect associated with the deprotonation of DMTA was

145

investigated by SPME-GC/IRMS at pH 8.4, 8.7, 9.4, 10.4, and 11.3 in 10 mM phosphate buffer at 8 ACS Paragon Plus Environment

Page 8 of 32

Page 9 of 32

Environmental Science & Technology

146

ionic strength of 4 M using DMTA concentrations of 66 µM, 12 µM (pH 8.7 and 9.4), 6.6 µM, and

147

5 µM, respectively. 53

148

Carbon, hydrogen, and nitrogen isotope ratios are reported as δ13 C, δ2 H, and δ15 N relative

149

to Vienna PeeDee Belemnite, Vienna standard mean ocean water, and air, respectively. 24,29 All

150

isotope signatures are reported in permil (h) as arithmetic mean of triplicate measurements (±σ).

151

To ensure the accuracy of the measured isotope ratios, we used a series of isotopic standard materials

152

purchased from Indiana University 55,56 as documented in Spahr et al. 29 as well as calibrated in-

153

house standards in standard bracketing procedures. In-house standards of ranitidine, DFUR,

154

DMTA, and NH4 Cl were obtained through C and N isotope ratio measurements with an elemental

155

analyzer IRMS (Table S1). Isotopic analysis of NH2 Cl was impeded owing to its thermal instability

156

and self-decay to ammonia. Instead, we used the δ15 N values of NH4 Cl or (NH4 )2 SO4 , from which

157

NH2 Cl was produced, as a proxy for the initial δ15 N values of NH2 Cl. This assumption was based

158

on high molar NH2 Cl yields (>94%) from the reaction of HOCl with NH4 Cl or (NH4 )2 SO4 .

159

Data evaluation

160

We conducted chloramination experiments with two model compounds, namely DMTA and DMBA,

161

to study isotope fractionation in the tertiary amines. Bulk isotope enrichment factors for carbon and

162

nitrogen, εC and εN , were derived from linear regression of δ13 C and δ15 N values versus fractional

163

amount of remaining precursor (c/c0 ) according to eq. 1 (see Figure S3). 25    δhE + 1 c ln h = εE · ln c0 δ E0 + 1 

(1)

164

where δ h E0 and δ h E are isotope ratios of an element E in the precursor at the beginning

165

and during the reaction, respectively. Apparent kinetic isotope effects, AKIEE , were calculated

166

according to eq. 2 considering the total number of atoms of an element (n), the number of atoms

167

in reactive positions (x), and the number of atoms in intramolecular competition (z). The AKIEC

168

values are reported as average secondary isotope effect for all C atoms in DMTA (n = x = 7, z = 1)

9 ACS Paragon Plus Environment

Environmental Science & Technology

Page 10 of 32

169

and DMBA (n = x = 9, z = 1). AKIEN values stand for primary isotope effects and both precursors

170

contain one N atom (n = x = z = 1). Uncertainties of εE and AKIEE values are reported as 95%

171

confidence interval.

AKIEE =

1 1 + (n/x) · z · εE

(2)

172

The observable AKIEN of DMTA or DMBA during chloramination at pH 8.0 originates from

173

the combination of a deprotonation step (eq. 3) and the subsequent reaction of the neutral tertiary

174

amine (eq. 4). As we have shown previously, 53,57 the observable AKIEN therefore consists of a

175

combination of a 15 N-equilibrium isotope effect for the quarternary amine deprotionation, EIEBH N

176

and an apparent kinetic isotope effect of the subsequent reaction, AKIEBN (eq. 5).

+ −B

k1 BG B + H+ BH+ FGGGGGG GGGGG k −1 k2 B GGGGGGA P BH AKIEN = fBH+ · EIEN

+ −B

(3) (4) · AKIEBN + (1 − fBH+ ) · AKIEBN

(5)

177

where k 1 and k−1 are rate constants of H+ exchange reaction, k 2 is the rate constant for transformation

178

of the neutral amine, and fBH+ is the fraction of the conjugate acid of the tertiary amine (see section

179

S4 for details).

180

Results and Discussion

181

Observable C, H, and N isotope fractionation trends in NDMA

182

We used 5-(dimethylaminomethyl)furfuryl alcohol (DFUR) as model compound for the chlorami-

183

nation of tertiary amines and studied the NDMA formation kinetics as well as C, H, and N isotope 10 ACS Paragon Plus Environment

Page 11 of 32

Environmental Science & Technology

184

ratios of NDMA. Figure 1a shows the formation of NDMA during the reaction of DFUR (3 µM)

185

with NH2 Cl (45 µM) in 10 mM phosphate buffer at pH 8.0. The reaction was completed within 10 h

186

with a molar NDMA yield of 65 ± 2%. 35,47,58 Consistent with our earlier observations, we found

187

a lag-phase of approximately one hour, in which only 0.2 µM DFUR was transformed to 0.1 µM

188

NDMA (dashed line in Figure 1a). 35 After 1 h, DFUR disappeared at a faster rate concomitant

189

with the formation of NDMA. This kinetic behavior implies that reactive intermediates such as the

190

N,N-dimethylhydrazine species (compound 3 in Scheme 1) and possible radical intermediates (4

191

and 5) are short-lived and transformed to NDMA and other unidentified products more rapidly than

192

the initial transformation of DFUR to compound 3. The total consumption of NH2 Cl amounted to

193

12.0 µM and thus exceeded the initial concentration of DFUR by a factor of 4.1 in agreement with

194

previous findings. 35 No lag-phase was observed for the disappearance of NH2 Cl (Figure 1a) indicat-

195

ing that side reactions, which did not lead to NDMA, likely contributed to the over-stoichiometric

196

consumption of NH2 Cl.

197

Figure 1b shows C, H, and N isotope signatures of NDMA during its formation. δ15 N values

198

of NDMA (depicted as upward triangles) increased within 10 h from –24.8h to −8.7h. This N

199

isotope fractionation is caused by primary kinetic isotope effects that occur when chemical bonds

200

to N are broken or formed in rate-determining reaction steps. 25,59 In contrast, δ13 C and δ2 H values

201

of NDMA changed only slightly from –36.8h to –34.3h and –133.5h to –110.7h, respectively.

202

This C and H isotope fractionation is small compared to reactions, in which bonds to C and H are

203

broken (e.g., 60,61 ) and likely due to secondary kinetic isotope effects of atoms that do not participate

204

in chemical reactions.

205

Isotope ratios of NDMA reveal the origin of C, H, and N atoms in NDMA

206

Origin and isotopic composition of the N,N-dimethylamine moiety of NDMA

207

We have shown previously that the N,N-dimethylamine group (N(CH3 )2 ) of the tertiary amine is

208

transferred to NDMA without being chemically altered. 29 This observation was key to elucidate the

209

mechanisms of N atom oxygenation and formation of radical intermediates in the NDMA formation 11 ACS Paragon Plus Environment

Environmental Science & Technology

(a)

Page 12 of 32

45

35 t = 1h

Concentration (µM)

40

3 2

DFUR ( NDMA NH 2 Cl (

1

control) control)

0 0

5

10 15 Time (h)

25

b15N0(NH2Cl-B) 40

20

+27.6‰

b15N0(average of DFUR and NH2Cl-B)

0 b15N0(DFUR)

+26.9‰

b13C (‰)

b15N (‰)

(b)

20

b15N0(NH2Cl-A) b13C0(DFUR)

-20

b2H (‰)

-40 -100 -120 -140 0

5

10 15 Time (h)

20

25

Figure 1 NDMA formation from the reaction of DFUR (3 µM) with NH2 Cl (45 µM) in 10 mM phosphate buffer (pH 8.0). Panel (a) shows DFUR abatement, NH2 Cl consumption, and NDMA formation over time. Symbols in panel (b) illustrate δ15 N, δ13 C, and δ2 H values of NDMA. Grey and yellow solid lines represent the initial δ13 C and δ15 N value of DFUR, respectively. The red and blue lines depict the initial δ15 N values of two different NH2 Cl batches with which separate NDMA formation experiments were conducted leading to NDMA with different δ15 N signatures (red-yellow vs. blueyellow triangles). The blue-yellow line represents the average of the initial δ15 N values of DFUR and NH2 Cl-B. Standard deviations of triplicate δ15 N, δ13 C, and δ2 H measurements were