Reactivity of Triplet Excited States of Dissolved Natural Organic Matter

Subscriber access provided by UNIV LAVAL

Article

Reactivity of triplet excited states of dissolved natural organic matter in stormflow from mixed-use watersheds Andrew J. McCabe, and William A. Arnold Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b01914 • Publication Date (Web): 26 Jul 2017 Downloaded from http://pubs.acs.org on July 31, 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 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Environmental Science & Technology

Reactivity of triplet excited states of dissolved natural organic matter in stormflow from mixed-use watersheds Andrew J. McCabe and William A. Arnold* Department of Civil, Environmental, and Geo- Engineering, University of Minnesota-Twin Cities, 500 Pillsbury Dr. SE, Minneapolis MN, 55455 *Corresponding author: [email protected]

Submitted to Environmental Science and Technology Word count: 5823 text + 2100 (6 figures) + 300 (1 table) = 8223 Abstract

15

Dissolved organic matter (DOM) quantity and composition control the rate of formation (Rf,T) of

16

triplet excited states of dissolved natural organic matter (3DOM*) and the efficiency of 3DOM*

17

formation (the apparent quantum yield, AQYT). Here, the reactivity of 3DOM* in stormflow

18

samples collected from watersheds with variable land covers is examined. Stormflow DOM

19

reflects variability in DOM quantity and composition as a function of land cover and may be

20

important in controlling the fate of co-transported pollutants. Rf,T and AQYT were measured using

21

2,4,6-trimethylphenol in stormflow samples under simulated sunlight. The DOM source and

22

composition was characterized using absorbance and fluorescence spectroscopies and high-

23

resolution mass spectrometry. Rf,T and the total rate of light absorption by the water samples (Ra)

24

increased with the dissolved organic carbon (DOC) concentration. AQYT was independent of

25

DOC concentration, but varied with DOM source: developed land cover (4 – 6%) ≈ open water >

26

vegetated land cover (3%). AQYT was positively related to an index for microbial/algal DOM

27

content and negatively related to DOM molecular weight, DOM aromaticity, and the content of

28

polyphenols. This work demonstrates that TMP is an effective probe for the determination of Rf,T and

29

AQYT in whole water samples after accounting for the inhibition of TMP photodegradation by DOM.

1 ACS Paragon Plus Environment


Page 2 of 36

30 31

Introduction

32

Triplet excited states of dissolved natural organic matter (3DOM*) are produced when

33

chromophoric dissolved natural organic matter (CDOM) absorbs sunlight. 3DOM* are highly

34

reactive species that exist in surface waters for micro-seconds, reaching steady-state

35

concentrations ([3DOM*]ss) on the order of femto- to pico-molar with excited-state energies

36

above 100 kJ mol-1 (~30–50% of 3DOM* species have energies ≥250 kJ mol-1)1 and reduction

37

potentials between 1.4 to 1.9 eV (relative to the standard hydrogen electrode).2 The progress

38

made in clarifying the nature of 3DOM* was recently reviewed by McNeill and Canonica.2 There

39

is accumulating evidence that 3DOM* are active photo-physical and chemical processes in sunlit

40

surface waters, including photosensitized production of reactive oxygen species (singlet oxygen

41

and hydroxyl radical),3 production of reactive halogen species,4 inactivation of pathogens

42

(through formation of 1O2),5 and reaction with organic micro-contaminants (pesticides,6,7

43

endocrine disruptors,8 and pharmaceuticals9,10).

44 45

Chromophores within DOM are typically categorized into two overlapping groups (1) discreet

46

chromophores (e.g., carbonyls and aromatic moieties11) and (2) charge-transfer (CT) complexes.

47

3

48

singlet states (1DOM*) that undergo forbidden electron spin flips to the lower energy excited

49

state, 3DOM*. 3DOM* decay radiatively or non-radiatively, and in oxic systems, are efficiently

50

quenched by dissolved oxygen.13 Sharpless and Blough recently reviewed the evidence

51

purporting the existence of CT complexes in DOM.12 These complexes are relatively stable

52

excited state species11 that form between closely-associated donor (phenols) and acceptor

DOM* form when discreet chromophores (e.g., aromatic ketones) absorb light forming excited


Page 3 of 36


53

(quinones) moieties within the DOM.12 Their formation may lower yields of 1DOM* and

54

3

DOM* because 1DOM*-precursors may act as acceptor moieties.14

55 56

There are several approaches to study the role of DOM source and composition in CDOM light

57

absorption and 3DOM* formation, including chemical transformations through photobleaching

58

or borohydride reduction,15 adjusting solution constituents that act as 3DOM* quenchers (ionic

59

strength16,17 or dissolved oxygen18), size fractionation,19,20 use of model 3DOM* compounds,21

60

and using whole water samples across water chemistry and DOM gradients.22–24 Size

61

fractionation of DOM shows that 3DOM* yields are inversely related to DOM molecular weight

62

while the steady-state concentration, [3DOM*]ss, does not show consistent trends with molecular

63

weight.19,20 Sewage-derived DOM may produce 3DOM* that efficiently react with trace organic

64

contaminants.9 It has been suggested, however, that while sewage organic matter may have high

65

3

66

depend on surface water hydroperiod (relative surface water residence time), suggesting that

67

DOM photobleaching and possibly inputs of algal-derived DOM result in high 3DOM* yields.24

DOM* yields, sewage DOM may also quench 3DOM* more efficiently.23 3DOM* yields also

68 69

Landscape-level characteristics, such as watershed land cover and water residence times,

70

influence DOM quantity and composition of inland waters.25–28 DOM subject to long surface

71

water residence times, for example, tends have relatively slow decomposition rates26 and

72

agricultural land use tends to increase the algal/microbial character of DOM.28 There are few

73

systematic studies, however, attempting to understand how DOM source influences the

74

photoproduction of 3DOM*. This has practical applications for studies aimed at modeling

75

regional and global surface water steady-state concentrations of 3DOM* and secondary reactive



Page 4 of 36

76

species such as singlet oxygen. The goals of this research are to identify the way in which land

77

cover, DOC concentration, and DOM composition influence the formation rate (Rf,T) and

78

apparent quantum yield (AQYT) of 3DOM* in stormflow. Stormflow has received comparatively

79

little attention in 3DOM* photochemistry literature, but it is critically important in transporting

80

DOM29,30 and trace organic contaminants31–34 to inland aquatic and marine environments.

81 82

Experimental

83

Sample Collection. Stormflow, baseflow, and snowmelt water samples were collected in

84

collaboration with six watershed/conservation districts from mixed-use watersheds in the upper

85

Mississippi River watershed near Minneapolis-St. Paul, Minnesota, U.S.A. Land cover of the

86

studied watersheds range from highly developed (>50% impervious) to highly vegetated (0%

87

impervious) to largely open water (~20%). Details about the sampling bottles, cleaning

88

procedures, and filtration are in the Supporting Information (SI Section 1).

89 90

Stormflow samples (186 total) were collected over the period of September 2014 to October

91

2015 from 31 sites, 29 baseflow samples were collected from 22 sites in the spring and/or fall of

92

2015, and 18 snowmelt samples were collected in March 2015 (Figure 1 and Table S1, SI

93

Section 2, list of site names and watershed districts). Stormflow samples were collected when

94

daily precipitation was >0.25 cm (Figure S1, SI Section 3). Samples were collected either as

95

composites (the duration of the storm event) or as grab samples during the stormflow period.

96

Baseflow samples were collected during periods of no precipitation within a 24–≥48-h period.

97

Samples were kept at ~4 °C until they were transported to the laboratory for processing. Samples

98

are labelled with the site name used by the watershed/conservation district and date of collection


Page 5 of 36


99

following the month-day-year convention (e.g., CMH07-100815 corresponds to a sample from

100

site CMH07 collected on October 8, 2015). The methods for determination of land cover for

101

each site are in the SI, Section 4.

102 103

Water chemistry and DOM optical measurements. Water chemistry measurements, including

104

pH, specific conductance, anion concentrations, and dissolved organic (DOC) and inorganic

105

carbon (DIC) concentrations are in the SI, Section 5. Absorbance spectra (λ=200–800 nm) were

106

collected with a Shimadzu UV-1601PC spectrophotometer using 1-cm quartz cuvettes. The

107

instrument was zeroed with Milli-Q water and spectra were corrected by subtracting the

108

spectrum of Milli-Q water to remove noise caused by the transition from halogen to deuterium

109

lamps. The E2/E3 ratio (abs250/abs365, an inverse proxy for molecular weight and a direct proxy

110

for the degree of photobleaching) and specific UV absorbance at 254 nm (SUVA254, a direct

111

proxy for aromaticity) were computed for each sample.15,20,35–37

112 113

Excitation-emission matrices (EEMs) were collected with a Horiba Aqualog in a 1-cm quartz

114

cell. Instrument specifications are detailed in Gilmore et al.38 Depending on CDOM content,

115

EEMs were collected with a 2 or 3 s integration time and either 1, 3, or 5-nm excitation

116

wavelength intervals. Samples were diluted with Milli-Q if the absorbance at 254 nm was >0.6.

117

EEMs were corrected using the drEEM toolbox in MATLAB (R2014A, Mathworks).39 Raman

118

scattering signals were removed by blank subtraction, inner-filter effects were corrected,40 and

119

spectra were normalized to the area of the water Raman scattering peak at an excitation of 350

120

nm. The fluorescence index (FI, a proxy for microbially- or terrestrially-derived DOM),41 the

121

humification index (HIX, a proxy for the degree to which fluorescence emissions red-shift as



Page 6 of 36

122

DOM humification occurs),42 and the β/α ratio (a proxy for recently produced/algal-derived

123

DOM) were used to assess DOM source.28,43

124 125

Calculations of E2/E3, SUVA254, FI, HIX, and β/α are described in the SI, Section 6. Optical

126

parameters and water chemistry measurements of the stormflow samples were compared

127

between watershed groups using the nonparametric Kruskal-Wallis analysis of variance with

128

post-hoc Dunn tests at a significance level of 0.05 in MATLAB (Anderson-Darling tests

129

suggested the data were not normally distributed).

130 131

FT-ICR MS analysis. A subset of samples (n=23, two baseflow and 21 stormflow samples) were

132

analyzed by FT-ICR MS to assess DOM aromaticity and molecular character.44,45 These samples

133

were selected to be representative of the observed range in DOM optical properties and the range

134

of watershed land covers. This technique assigns molecular formulas to highly resolved DOM

135

molecular compositions based on m/z (mass to charge ratios).46 In combination with principal

136

component analysis (PCA),47 FT-ICR MS data was used to assess molecular-level differences in

137

DOM collected from watersheds with different land covers.

138 139

Samples were prepared by extracting and concentrating DOM from filtered water samples by

140

solid-phase extraction using styrene divinylbenzene polymer-packed cartridges (0.5 g, Agilent

141

PPL) using a vacuum manifold.48 Samples were analyzed using a custom-built FT-ICR mass

142

spectrometer with 9.4 T magnet and negative-ion electron spray ionization at the National High

143

Magnetic Field Laboratory, Florida State University. Extraction details, instrument

144

specifications, and operating conditions are in the SI, Section 7. Ion masses were assigned a


Page 7 of 36


145

molecular formula constrained to C1-100H4-200N0-4O1-25S0-1 if the molar mass of the assigned

146

formula had a root-mean square error of ≤1 ppm relative to the theoretical molar mass (full

147

composition assignment criteria are in SI Section 7).

148 149

PCA was used to identify sample groupings based on FT-ICR MS relative abundances. This is a

150

frequently used multivariate technique applied to FT-ICR MS data (e.g., ref.47) that reduces

151

multidimensional data (samples×relative abundances of assigned compositions) to two or three

152

dimensions. PCA was performed using variance-normalized relative abundances in MATLAB.

153

Only compositions with relative abundances >0.01 and present in ≥75 % of samples were

154

included in the PCA matrix. Previously established thresholds27,45,49 based on the aromaticity

155

index (AI=[1 + C – O – S – 0.5H]/[C – O – S – N])45 and DOM bio-lability49 were used to

156

define the molecular classes that were causing separation of samples along the principal

157

component axes: aliphatic and bio-labile compounds were defined by H/C≥1.5, highly

158

unsaturated and phenolic compounds by H/C99%, Mallinckrodt), respectively.

211

For the photobleaching treatments, samples were exposed for 15, 30, 45, and 60 h in the solar

212

simulator (intensity from λ=300–800 nm set to 765 W m-2) prior to initiating the TMP

213

experiments. Concentrations of DOC and DIC, specific conductance, pH, UV-vis absorbance, 9 ACS Paragon Plus Environment


Page 10 of 36

214

and EEMs were measured following photobleaching, but no water chemistry adjustments were

215

made.

216 217

The second order rate constant, kT,TMP, was estimated by measuring kobs,TMP as a function of

218

[TMP]0 between 0–750 µM for a subset of the water samples covering the range in watershed

219

land covers and sampling seasons. The data were fit to a linearized form of eq. 1 (eq. S16c). This

220

experimental procedure has previously been used with TMP18,22 and is analogous to the way in

221

which the probe, trans,trans-2,4-hexadienoic acid is used to quantify 3DOM* formation.58

222

Additional details on the kinetic models are in the SI, Section 11.

223

Kinetic modeling was performed using Microsoft Excel 2016. Photochemical results were

224

compared between watershed groups using Kruskal-Wallis analysis of variance with post-hoc

225

Dunn tests at a significance level of 0.05 in MATLAB. Spearman rank correlation coefficients

226

(ρ) were computed between DOM optical properties, [DOC], AQYT, and Rf,T in MATLAB.

227

Spearman correlation coefficients were also computed between relative abundances of

228

compositions identified with FT-ICR MS and AQYT for compositions with relative abundances

229

>0.01 in ≥75% of the samples analyzed. This threshold was selected to avoid correlations with

230

compositions that had high occurrences of low relative abundance. The significance level was set

231

to α=0.001 to avoid type I (false positive) errors. Compositions with correlations meeting these

232

criteria were then plotted on a van Krevelen diagram. This correlational analysis technique has

233

previously been used to establish links between the relative abundances of molecular

234

compositions and DOM bio-lability and age,60 EEMs components identified by parallel factor

235

analysis,61 and hydrological and climatic variables.27 The non-parametric Spearman rank

236

correlation analysis was used because it does not assume a specific model (e.g., linear) and


Page 11 of 36


237

neither the raw data nor residuals are required to adhere to a specific distribution. A Spearman

238

correlation coefficient, ρ, greater than +0.6 or less than –0.6 is considered a strong trend.

239 240

Inhibition of TMP Photodegradation. Experiments using CBP as a 3DOM* model and TMP

241

were performed following the protocol of Canonica and Laubscher62 to assess the influence of

242

DOC inhibition on TMP photodegradation. Four experimental treatments were used to estimate

243

an inhibition factor (IF) that describes the ratio of rates of 3DOM* -induced TMP loss with and

244

without DOM present. The treatments were: (1) TMP in pH 8 10 mM borate buffer, (2)

245

TMP+CBP in pH 8 10 mM borate buffer, (3) TMP in whole water stormflow samples, and (4)

246

TMP+CBP in whole water stormflow samples. The first treatment corrects the second treatment

247

for the direct photolysis of TMP and the third treatment corrects the fourth treatment for

248

3

249

superscript corr in eq. 2–4). Appropriate screening factors were also applied (SI, Section 11).

250

The IF was computed using eq. 2–4.

DOM*-induced TMP loss with only DOM present (these corrections are indicated with the

251

, =

252

=

253

() =

(2)

!"]

(3)

, [

, [

!"] − $% ∙ [•' ] ∙ [ !]

2344 +,-.,/01

(4)

2344 +,-.

254

Where kCBP,DOMcorr (s-1) is the pseudo-first order rate constant for the reaction between TMP and

255

the triplet excited state of CBP (3CBP) with DOM present, k3CBP,TMP (M-1 s-1) is the second order

256

rate constant for the reaction between 3CBP and TMP, kred (M--1 s-1) is the second order rate

257

constant for the reaction between TMP•+ and reduced moieties of DOM. and kCBPcorr (s-1) is the

258

pseudo-first order rate constant for reaction between TMP and 3CBP without DOM present.



Page 12 of 36

259

Substituting eq. 2 and 3 into eq. 4 gives eq. 5a (eq. 5b is the linearized form analogous to the

260

model in ref.62).

261

262

() = + 8 9:

=

+35

(5a)

467 ∙[]'+35

+467 +35

∙ [ !] + 1

(5b)

263

Where kox is the pseudo-first order rate constant for the reaction between TMP•+ and O2. Eq. 5b

264

shows that 1/IF should be linearly dependent on [DOC] (a proxy for the concentration of reduced

265

moieties in DOM) with slope kred/kox. IF was computed from eight experiments using a subset of

266

five samples (giving a range of [DOC] and watershed land covers). The resulting data were fit to

267

eq. 5b. Additional details are in SI, Section 11.

268 269

Results and Discussion

270

Average water chemistry parameters (pH, concentrations of DOC and DIC, specific

271

conductance, and anion concentrations) for each site are summarized in Table S5 (SI Section 12).

272

All data from the photochemical experiments and optical properties are provided in a

273

supplemental spreadsheet and available on the Data Repository for the University of Minnesota.

274

Details about sample stability and the length of time between sample collection and water

275

chemistry or photochemistry measurements are in the SI, Section 13.

276 277

Water chemistry summary. pH was relatively consistent between stormflow, baseflow, and

278

snowmelt samples and across watershed land covers. The average pH of the stormflow samples

279

was 7.9 (pH range 7.2–8.6); samples from four high-intensity developed watersheds were

280

slightly more acidic (pH ~7.5). Specific conductance ranged between 200–750 µS cm-1 and was


Page 13 of 36


281

not significantly different between watershed land cover groups (excluding the outlier site

282

ST19). Specific conductance was higher in baseflow and snowmelt samples (86% and 67%,

283

respectively, had specific conductance measurements greater than average stormflow

284

measurements).

285 286

Conversely, DOC concentrations were lower in baseflow and snowmelt samples (69% and 89%,

287

respectively, were less than the average stormflow DOC concentrations). The observation that

288

[DOC] is higher in stormflow than baseflow is consistent with previous observations and has

289

been attributed to the activation of alternative water flow paths during rain events.29,63,64 DOC

290

concentrations in stormflow samples from vegetated watersheds were significantly higher

291

compared to developed and open water watersheds (average vegetated: 10×10-4 M vs. open water

292

and developed: 5–6×10-4 M). This is consistent with observations made comparing stormflow

293

DOC concentration in forested, agricultural, and golf course dominated watersheds (~8–30×10-4

294

M C) to watersheds with high impervious cover (

Reactivity of Triplet Excited States of Dissolved Natural Organic Matter

Recommend Documents