Temperature Dependence of the Photochemical ... - ACS Publications

Subscriber access provided by UNIV OF NEW HAMPSHIRE

Article

Temperature Dependence of the Photochemical Formation of Hydroxyl Radical from Dissolved Organic Matter Garrett McKay, and Fernando L. Rosario-Ortiz Environ. Sci. Technol., Just Accepted Manuscript • Publication Date (Web): 26 Feb 2015 Downloaded from http://pubs.acs.org on February 26, 2015

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 25

Environmental Science & Technology

1

Temperature Dependence of the Photochemical Formation of

2

Hydroxyl Radical from Dissolved Organic Matter

3

Garrett McKay and Fernando L. Rosario-Ortiz*

4

Department of Civil, Environmental and Architectural Engineering, 428 UCB, University

5

of Colorado, Boulder, CO 80309, USA

6 7 8 9 10

• Corresponding author: [email protected]; Phone: 303-492-7607; Fax: 303-492-7317

Abstract

11

In this paper, the temperature dependence of the photochemical production of the

12

hydroxyl radical (•OH) from dissolved organic matter (DOM) was investigated by

13

measuring the apparent temperature dependence of the quantum yield (Φa) for this

14

process. Temperature dependent Φa values were analyzed using the Arrhenius equation.

15

Apparent activation energies obtained for DOM isolates purchased from the International

16

Humic Substances Society ranged from 16 to 32 kJ mol-1. Addition of 40 units mL-1

17

catalase, used to hinder the hydrogen peroxide (H2O2)-dependent pathway to •OH, did

18

not impact the observed activation energy. However, an increase in activation energy was

19

observed in lower molecular weight DOM obtained by size fractionation. We also

20

measured the temperature dependence of p-benzoquionone photolysis as a model

21

compound for DOM and observed no temperature dependence (slope p = 0.41) for the

22

formation of phenol from oxidation of benzene (the •OH probe used), but a value of

23

about 10 kJ mol-1 for p-benzoquinone loss, which is consistent with formation of a water-

24

quinone exciplex. These data provide insight into DOM photochemistry as well as

1

ACS Paragon Plus Environment


25

provide parameters useful for modeling steady state •OH concentrations in natural

26

systems.

27 28

Introduction

29

Reactive oxygen species (ROS) play an important role in the photochemistry of

30

natural aquatic systems. These ROS include the hydroxyl radical (•OH), singlet oxygen

31

(1O2), superoxide (O2•-), hydrogen peroxide (H2O2) and triplet excited states. From an

32

engineering standpoint, degradation of many organic contaminants is increased by

33

ROS.1,2 Ecologically, ROS are involved in the mineralization of organic carbon,3

34

contributing to the global carbon cycle. The presence of H2O2 was first detected in

35

marine waters in 1966 and was shown later to result from photochemical reactions.4

36

Since then, other ROS such as •OH and 1O2 have been shown to be produced

37

photochemically in natural waters.3,5,6

38

Multiple chemical species undergo photochemical reactions to produce ROS, such

39

as nitrate, nitrite, and dissolved organic matter (DOM).3,5,7 In particular, DOM photolysis

40

produces •OH and 1O2, as well as H2O2, triplet states (here abbreviated as 3DOM*), and

41

O2•-. The generally accepted model is that 1DOM* is formed following absorption of a

42

photon by ground state DOM, which can subsequently undergo internal conversion,

43

vibrational relaxation, or intersystem crossing to 3DOM*. Early work by Zepp et al.

44

established an average energy of around 250 kJ mol-1 for 3DOM* molecules,8 sufficiently

45

high to form

46

disproportionation of O2•-, which is formed by reaction of dissolved oxygen with some

47

photo produced reductant.9 The •OH radical is believed to be formed by multiple

48

pathways, however, steady state concentrations are limited to around 10-18 to 10-16 M in

1

O2 from reaction with dissolved oxygen. H2O2 results from the

2


Page 2 of 25

Page 3 of 25


49

natural waters due to its high reactivity.3,10 1O2 is a weaker and more selective oxidant

50

than •OH and thus reaches steady state concentrations a few orders of magnitude higher.7

51

Additionally, steady state concentrations of •OH and 1O2 are generally higher for effluent

52

organic matter compared to DOM isolates.11,12 Reported apparent quantum yields (Φa) for

53

•OH radical are on the order of 10-4 to 10-5, again being generally higher for wastewater

54

samples.11

55

While it is clear that DOM is a photochemical source of •OH, the underlying

56

mechanisms are not completely understood. Some of the formation pathways are

57

summarized in Scheme 1. There likely exist two pathways to •OH: an H2O2-dependent

58

and H2O2-independent pathway.5,7,11 The H2O2-dependent pathway involves the photo

59

Fenton reaction13,14 and possibly direct photolysis of photochemically produced H2O2.15

60

The importance of this pathway has been demonstrated by using catalase to decompose

61

H2O2, which decreases •OH production by up to 50 % for some samples.16,17 Super-

62

oxidized iron species (e.g. ferryl, FeO2+), which are operational at neutral pH,18 may also

63

account for some •OH-like activity by reacting with various organic compounds.19 The

64

H2O2-independent pathway involves species that lead to free •OH as well as so-called

65

low energy hydroxylating species.17,20,21 H-atom abstraction from water by some DOM

66

excited state (3DOM*, 1DOM*, or charge transfer species-DOM•+/•-) is one mechanism

67

suggested to produce free •OH, but this has not been confirmed.7,22 Generation of low

68

energy hydroxylators has been argued largely on the quenching of •OH-like activity with

69

addition of methane. For example, Page et al. (2011) showed that some DOM isolates

70

exhibited less than calculated decreases in photochemical •OH production with addition

71

of known concentrations of methane.17 Though uncertain, these species could be modeled

3



72

by the quinone-water exciplex formed during irradiation of 2-methyl-p-benzoquinone.

73

For example, Pochon et al. (2002)20 and Gan et al. (2008)21 demonstrated no methyl

74

radical production upon irradiation of methyl-p-benzoquinone containing methane as an

75

•OH probe.

76

This paper presents data on the apparent temperature dependence of Φa for bulk

77

DOM isolates, with and without catalase present, as well as apparent molecular weight

78

fractions. The temperature dependence of the photolysis of p-benzoquinone was also

79

investigated as a model compound for comparison. These data provide insight into DOM

80

characterization and photochemistry by comparing the temperature dependence of

81

different DOM samples to a model compound. In addition, the measured activation

82

energies can be used for kinetic modeling of steady state •OH concentrations in natural

83

and engineered systems exhibiting temperature fluctuations or large deviations from

84

room temperature.

85

Materials and Methods

86

DOM isolates including Suwannee River Fulvic Acid I (SRFA, 1S101F),

87

Suwannee River Humic Acid II (SRHA, 2S101H), Suwannee River NOM (SRNOM,

88

2R101N) Pony Lake Fulvic Acid (PLFA, 1R109F), and Elliot Soil Humic Acid (ESHA,

89

1S102H) were purchased from the International Humic Substance Society. All other

90

chemicals were obtained from commercial sources. The solution conditions, irradiation

91

conditions, and analytical methods used in this study are presented in detail elsewhere.11

92

Briefly, solutions used in irradiation experiments were buffered at pH 7.2 in equimolar

93

KH2PO4/K2HPO4 (total [phosphate] = 10 mM). Samples were irradiated in clear, glass

94

vials free of headspace. Irradiations were ran in duplicate for 2.5 to 8 hours (depending

4


Page 4 of 25

Page 5 of 25


95

on sample to achieve measurable •OH concentrations) with an Oriel 1429A Solar

96

Simulator equipped with a 1000 W Xenon lamp (see Figure S1 for lamp spectrum). Over

97

the irradiation period, samples were collected at 5 time intervals. Temperature control

98

within ± 2 °C was accomplished by laying the vials flat in a water-jacketed petri dish.

99

Benzene was used as an •OH probe by following the formation of phenol through the

100

Dorfman reaction.23,24 Though not described in detail here, this mechanism involves

101

reaction of the cyclohexadienyl radical with dissolved oxygen, which facilitates the

102

formation of phenol. Based on the experimental conditions used in this study (gas tight

103

vials, ~ µM phenol formation) no oxygen depletion due either to temperature changes or

104

phenol formation during the course of experiments was expected. A benzene

105

concentration of 3 mM was used to quantitatively scavenge •OH radicals. Phenol

106

concentrations were measured by HPLC utilizing a gradient mobile phase of pH 2.8, 10

107

mM H3PO4 and methanol with a flow rate of 1.0 mL min-1. The retention time of phenol

108

under these conditions was around 4.7 minutes. Using the same HPLC method, p-

109

benzoquinone eluted at 2.2 minutes (column dead time ~ 1.6 minutes). The column used

110

was a Zorbax Eclipse XBD (4.6 x 150 mm) C18 column. Bovine liver catalase was used

111

in some experiments to quench H2O2.

112

Apparent molecular weight fractions were collected using < 5 kDa and < 1 kDa

113

ultrafiltration membranes (Millipore, USA). Membranes were soaked three times in fresh

114

1 L batches of Milli-Q water before use. To remove any remaining organic material, 200

115

mL of Milli-Q water was passed through the membrane before fractionating DOM

116

samples. Approximately 200 mL of DOM solution was added to the ultrafiltration cell.

117

The first 100 mL of permeate was collected.

5



118

Page 6 of 25

The apparent quantum yield of •OH from DOM photolysis is given by

R•OH Φa = kDOM−a [DOM]

119

(1)

120

where R•OH is the rate of •OH radical formation in M s-1 and kDOM-a is the specific rate of

121

light absorption by DOM in s-1. The specific rate of light absorption was calculated by the

122

following equation 400nm

123

kDOM−a = ∑ λ =290nm

− ε DOM ( λ )[DOM]z

E p0(λ )ε DOM ( λ )(1−10

)

ε DOM (λ )[DOM]z

(2)

εDOM(λ) is

124

where ‫ܧ‬௣଴ (λ) is the photon irradiance of the solar simulator in Einstein-1 cm-2,

125

the molar absorptivity of DOM in MC-1 cm-1, [DOM] is the DOM concentration in MC,

126

and z is the cell depth in cm. Absolute irradiance was measured with a spectroradiometer.

127

Absorbance was measured with a Cary 100 UV/Vis spectrophotometer. Molar

128

absorptivity data for p-benzoquinone were taken from the National Institute of Standards

129

and Technology Chemistry WebBook.25

130

The rate of phenol formation in this system is given by

131

d[C6H5OH] = kC6H6-OH[ OH][C6H5 ]= Yphenol R OH dt

(3)

132

where kC6H6-OH is the second-order rate constant between benzene and •OH radical and

133

Yphenol is the yield of phenol from this reaction. Making the steady state assumption for

134

•OH radical gives

135

d[ OH] = kDOM−a Φ a [DOM]-∑ ks [S][ OH] ≈ 0 dt

6


(4)

Page 7 of 25

136


[ OH]ss =

kDOM−aΦ a [DOM]

∑ k [S]

(5)

s

137

where kS is the second order rate constant between •OH and scavenger S. Since the

138

experimental conditions are such that Σks[S][•OH] ≈ kC6H6-OH[C6H6][•OH], that is,

139

benzene is the dominant •OH scavenger equation (3) becomes

140 141

142

d[C6H5OH] = kDOM−a Φ a[DOM] dt

(6)

and solving for Φa gives

Yphenol Φ a = R OH kDOM−a[DOM]

(7)

143

Clearly, Φa depends on the yield of phenol from hydroxylation of benzene. We discuss

144

below the temperature dependence of Yphenol.

145 146 147

Activation energies were determined by plotting the measured temperature dependent Φa according to the Arrhenius equation:

E 1 ln(Φ a ) = − a + ln(A) RT

(8)

148

Quantum yields were measured at 4 to 5 temperatures over a range of around 10 to 40 °C.

149

Temperature dependent quantum yields have been analyzed in this way before26-28 and

150

the present data generally fit the Arrhenius model, with the average relative error (based

151

on a linear fit of the Arrhenius plots) in activation energies being 15 %.

152

Results and Discussion

153

Temperature Dependence of Phenol Yield from Benzene

154

In order to use equation (7) at different temperatures, the temperature dependence of

155

Yphenol needed to be assessed. Work by Chu and Anastasio (2003) has demonstrated that 7



Page 8 of 25

156

the •OH probe benzoic acid is not sensitive to temperature,27 suggesting that this may be

157

the case for benzene.

158

Measuring Yphenol involves knowing the actual rate of phenol produced (Rphenol)

159

and the theoretical or calculated amount of •OH resulting from photolysis of H2O2

160

(R•OH,calc) – i.e. Yphenol = Rphenol / R•OH,calc. R•OH,calc depends on the rate constant k for the

161

photochemical transformation, which is calculated in general by

162

R•OH ,calc = k [H 2 O 2 ]

163

k = 2.303

(9)

400nm

∫ (Φλε λ I λl)dλ

λ =290nm

(10)

164

where Φλ, ελ, and Iλ represent the quantum yield of •OH from H2O2, molar absorptivity,

165

and irradiance at a given wavelength and l is the pathlength. Note that the (small)

166

temperature dependence of Φλ and ελ measured by Chu and Anastasio was used in

167

equation (9).29 Solutions of 90 to 100 µM H2O2 were spiked with 3 mM benzene and

168

irradiated for periods of 100 to 220 min. Rphenol was found from linear fits of the resulting

169

data (not shown). Experiments were conducted at temperatures ranging from 11.6 to 38.4

170

°C and resulted in Yphenol values of 0.63 ± 0.07 (see Figure S2) with no apparent

171

temperature dependence (slope p = 0.14). In addition, an activation energy of 16.6 ± 5.0

172

kJ mol-1 for the production of •OH from nitrate photolysis was measured using benzene

173

as the •OH probe. This value is in good agreement with that measured by Chu and

174

Anastasio (2003)27 of 19 ± 4 kJ mol-1 and Zellner et al. (1990)26 of 16 ± 4 kJ mol-1.

175

The Yphenol determined here is within the range of values in the literature (0.43 to

176

0.95 depending on pH)23,30-32 and agrees well with the recently determined value of 0.69

177

measured by Sun et al. (2014).33 In previous publications by our group11,22,34 and others,35

8


Page 9 of 25


178

a yield of 0.85 was used. In this work, the determined yield was 26 % lower. This

179

highlights the importance authors reporting •OH formation rates should report the value

180

of Yphenol to allow for comparison to other studies.

181

One limitation to this study is that irradiations were performed using a solar

182

simulator (opposed to a monochromatic source) as the quantum yield for •OH radical are

183

strongly wavelength-dependent.5,36 Subsequently, the quantum yields reported here

184

should be interpreted as an average over the spectral output of the solar simulator (λ >

185

290 nm). Nevertheless, this paper provides the first temperature dependent measurements

186

of Φa for •OH; also, previous studies have used polychromatic irradiation to investigate

187

DOM photochemistry.

188

Temperature Dependence for Bulk DOM Isolates

189

DOM isolates were chosen to give a range of bulk chemical functionalities, type (e.g.

190

humic acid vs. fulvic acid), and source (e.g. Pony Lake vs. Suwannee River). A typical

191

data set is shown in Figure 1 for ESHA. Apparent quantum yields were plotted according

192

to the Arrhenius equation (eq. 8) in order to obtain the apparent activation energy

193

(hereafter called activation energy). The term apparent activation energy is used to

194

acknowledge the complexity of the system in that DOM is a heterogeneous, complex

195

material and that the activation energies are obtained using a polychromatic source.

196

Activation energies for all samples are summarized in Table 1. There is surprisingly little

197

variation in the values for bulk DOM isolates, besides SRFA (34.3 ± 7.2 kJ mol-1), which

198

suggests a consistent mechanism for •OH production. In addition, data are plotted

199

according to the Eyring equation (see Table S1) and some discussion is provided in the

200

Supplemental Information.

9



Page 10 of 25

201

Values for bulk samples are in the range of reported activation energies for other

202

photochemical processes (see Table 2), though there is some variation depending on the

203

reaction. For example, aqueous photolysis of H2O2 and NO3- results in primary

204

photochemical processes characterized by different photofragments as shown in

205

equations 11-13.

206

hν •OH H2O2 →

(11)

207

[NO3–]* → NO2- + O(3P)

(12)

208

2O [NO3–]* → NO2• + O•- H  → NO2• + •OH + OH–

(13)

209

The lower activation energies of ≈ 16 to 18 kJ mol-1 for ESHA, SRHA, and SRNOM are

210

closer to values characterizing diffusion-controlled reactions (≈15 kJ mol-1),37 while

211

those for PLFA and SRFA, greater than 20 kJ mol-1, are slightly higher. An explanation

212

for the higher activation energy for nitrate photolysis (relative to H2O2) is that the

213

primary photoproduct of H2O2 photolysis (•OH radical) diffuses more easily from the

214

solvent cage than that of nitrate photolysis (O•-).26,29,38,39 Therefore the activation energies

215

for Φa from DOM could reflect either secondary thermal chemistry or diffusion of

216

different photoproducts from the solvent cage.

217

Considering the activation energies for photochemical reactions of organic

218

molecules, values for H-atom abstraction by ketones can vary from 9.2 to 29.3 kJ mol-1

219

depending on the reaction conditions and organic substrate.40 These Norrish Type

220

reactions involve H-atom abstraction from the triplet state, subsequently producing a

221

carbon-centered diradical. It is interesting that the activation energies in Table 1 match

222

these values. However, the above activation energies are for production of a carbon-

10


Page 11 of 25


223

centered radical in organic solvents, while H-atom abstraction from water by 3DOM*

224

would lead to •OH in aqueous solution.

225

Catalase Experiments

226

The activation energies shown in Table 1 are for DOM samples not containing catalase,

227

which is used to quench photochemically produced H2O2 in order to minimize the

228

contribution of direct photolysis and photo-Fenton reactions to •OH production.16,17

229

Activation energies for the photochemical production of H2O2 from DOM have recently

230

been reported by Kieber et al. (2014) for seawater samples, ranging from about 8 to 53 kJ

231

mol-1 depending on the irradiation wavelength.28 Their analysis revealed that when taken

232

as an average across all samples, activation energies ranged from 15 to 30 kJ mol-1 over

233

290 to 400 nm.

234

To examine the H2O2-independent pathway, Φa values were measured for ESHA

235

solutions containing increasing concentrations of catalase. A similar experiment

236

including the same DOM sample is described by Page et al. for the terephthalate probe.17

237

These data for ESHA are shown in Figure S3 and demonstrate an almost identical

238

quenching of •OH production for the same sample (ESHA) determined by Page et al.

239

(2011) - about 50 %, at similar concentrations of catalase (≈ 40 units mL-1).17 Using a

240

catalase concentration of 40 unit mL-1, an activation energy of 16.6 ± 3.4 kJ mol-1 was

241

measured for ESHA, which matches well with the non-catalase value of 17.6 ± 1.3. This

242

result indicates that the H2O2-independent pathway to •OH (H-atom abstraction by

243

3

244

for ESHA. The addition of catalase also limits photo-Fenton reactions from occurring.

245

Again the similarity in activation energies with addition of catalase, which essentially

DOM* and/or low-energy hydroxylator) has an activation energy of about 17 kJ mol-1

11



246

shuts down any Fenton chemistry, suggests that the H2O2-dependent pathway has a

247

similar or lower activation energy than the H2O2-independent pathway.

248

Temperature dependence of p-Benzoquinone photolysis

249

To gain further insight into the H2O2-independent pathway, the temperature-dependent

250

photolysis of p-benzoquinone was investigated using benzene as an •OH probe. Others

251

have investigated this system at room temperature using dimethylsulfoxide (DMSO) to

252

trap •OH, leading to the mechanism proposed in Scheme 2.20,21

253

In our experiments, solutions of 92.5 µM p-benzoquinone produced about 500 nM

254

phenol over 12 minute irradiations (Figure S4). The fact that phenol from irradiation of p-

255

benzoquinone shows that benzene may be susceptible to hydroxylation by species other

256

than free •OH, e.g. low energy hydroxylators. Though there is some debate as to whether

257

irradiation of quinones produces free •OH, there is evidence that hydroxylation occurs

258

through a quinone-water exciplex.20,21 As quinone moieties are believed to be present in

259

DOM, this may explain the presence of low energy hydroxylators in irradiated DOM

260

observed previously.17,20

261

As seen in Figure 2a, the Φa of phenol from p-benzoquinone irradiation was

262

essentially temperature independent with an activation energy of ≈ 0 kJ mol-1 (slope p =

263

0.41). The loss of p-benzoquinone, however, shows more complex behavior (Figure 2b).

264

At temperatures lower than ≈ 23 °C, p-benzoquinone loss is temperature dependent (9.7

265

kJ mol-1), while above 23 °C, there is almost no temperature dependence. These data are

266

consistent with formation of a water-quinone exciplex that can hydroxylate benzene or

267

proceed by other reactions to form hydroxy-p-benzoquinone or hydroquinone. The loss of

268

p-benzoquinone, and production of hydroxy-p-benzoquinone and hydroquinone was

12


Page 12 of 25

Page 13 of 25


269

confirmed in separate experiments by measuring the spectral changes in irradiated

270

solutions containing 75 µM p-benzoquinone with and without 3 mM benzene (with

271

benzene experiment shown in Figure S5). The presence of 3 mM benzene did not

272

eliminate the production of hydroxy-p-benzoquinone, suggesting that the intermediate

273

species formed is not quantitatively quenched by benzene at a concentration of 3 mM.

274

Conversely, others have reported quenching of this intermediate by DMSO at

275

concentrations of 100 mM.21

276

The reactions shown in Scheme 2 are likely affected by temperature in two ways.

277

First, forward reactions to form phenol, hydroxy-p-benzoquinone, and the hydroquinone

278

are expected to increase with increasing temperature. Additionally, an increase in

279

temperature likely encourages dissociation of the exciplex to re-form ground state p-

280

benzoquinone. Based on these data, it seems that an increase in temperature favors

281

production of hydroxy-p-benzoquinone and the hydroquinone up to a certain point

282

(around 23 °C), while at higher temperatures, dissociation of the exciplex dominates.

283

This explanation also accounts for the net zero temperature dependence of Φa for phenol.

284

Negative and near zero activation energies have been taken as evidence for exciplex

285

formation in other chemical systems,41-43 albeit better characterized systems than DOM.

286

Taken as a whole, these experiments demonstrate that the activation energies observed

287

for •OH production from DOM cannot be attributed solely to a simple quinone model,

288

even though the experimental approach depended on the measurement of the formation of

289

phenol from the hydroxylation of benzene.

290 291

13



Page 14 of 25

292

Effect of Molecular Size

293

Additional observation of Table 1 reveals differences in activation energies between

294

humic (average of 17.1 ± 0.9 kJ mol-1) and fulvic (average of 27.9 ± 9.0 kJ mol-1) acids.

295

It is difficult to specify what could account for this phenomenon because there are many

296

differences between these samples. For example, although SRFA and PLFA are both

297

fulvic acids, they differ widely in geographic source (Antarctica vs. Georgia, USA) and

298

type (microbial vs. terrestrial influence). However, fulvic acids in general have lower

299

average molecular weights.44

300

To further discern the effect of molecular size on the temperature dependence of

301

Φa, SRNOM was size-fractionated with

Temperature Dependence of the Photochemical ... - ACS Publications

Recommend Documents