Temperature dependence of 185 nm photochemical water treatment – The photolysis of water

Accepted Manuscript Title: Temperature dependence of 185 nm photochemical water treatment – The photolysis of water Author: Laith Furatian Madjid Mohseni PII: DOI: Reference:

S1010-6030(17)31258-3 https://doi.org/doi:10.1016/j.jphotochem.2017.12.030 JPC 11068

To appear in:

Journal of Photochemistry and Photobiology A: Chemistry

Received date: Revised date: Accepted date:

29-8-2017 9-12-2017 20-12-2017

Please cite this article as: Laith Furatian, Madjid Mohseni, Temperature dependence ¨ ¨/>nm photochemical water treatment ndash The photolysis of of 185 (2017), https://doi.org/10.1016/j.jphotochem.2017.12.030 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.

3

Laith Furatian, Madjid Mohseni

5

Department of Chemical and Biological Engineering, University of British Columbia, Vancouver BC, Canada

cr

4

ip t

2

Temperature Dependence of 185 nm Photochemical Water Treatment - The Photolysis of Water

1

Abstract

7

The photochemical treatment of water using 185 nm radiation forms the basis of an

8

advanced oxidation process (AOP) that does not require chemical addition. The

9

185 nm photolysis of water generates the hydroxyl radical ( ·OH) able to degrade

10

trace organic contaminants. However, the strong absorbance of water (αH2 O ) at

11

185 nm imposes severe geometrical constraints on reactor design. Investigation of

12

the reported temperature dependence of αH2 O on treatment efficiency was conducted

13

between 5 and 35 ◦C, in model solutions using carbamazepine as a radical probe.

14

Comparison was made with the temperature dependence of the 254 nm photolysis of

15

hydrogen peroxide under similar conditions. It was found that the 185 nm AOP is

16

less sensitive to temperature under the conditions tested, suggesting the absorbance

17

of water has a negligible effect in this range. This is postulated to be due to the

18

fundamental structure of liquid water, whereby 185 nm photons are absorbed by a

19

small population of interstitial H2 O monomers, with no hydrogen bonds to impede

20

the escape of ·OH from the solvent cage.

21

Keywords: advanced oxidation, ultraviolet, water photolysis, hydroxyl radical,

22

monomer

Ac ce p

te

d

M

an

us

6

Email address: [email protected] (Madjid Mohseni) Preprint submitted to Journal of Photochemistry and Photobiology A

December 4, 2017

Page 1 of 22

1. Introduction

24

The conventional low-pressure mercury-vapour discharge lamp generates high inten-

25

sity UV radiation primarily at 254 and 185 nm (1). The irradiation of water at 254 nm

26

is particularly effective at the inactivation of chlorine resistant protozoan pathogens

27

(2; 3), and is now widely employed in drinking water treatment at municipal scale

28

(4; 5). In addition to disinfection, 254 nm photolysis of H2 O2 generates the highly re-

29

active ·OH, thus constituting an advanced oxidation process (AOP) (6). AOPs may

30

be applied to augment drinking water treatment when sources are impaired by trace

31

organic chemical contaminants refractory to conventional treatment. The 254 nm

32

photolysis of H2 O2 is recognized as one of the most practical AOPs for drinking wa-

33

ter treatment, offering a compact footprint, the absence of a waste stream, avoidance

34

of bromate formation, and simultaneous disinfection. However, the on-site storage of

35

concentrated H2 O2 , its addition upstream of UV reactors, and its subsequent removal

36

prior to distribution introduce increased process complexity and cost. The 185 nm

37

emission is of approximately five to ten times lower intensity than that of 254 nm (1),

38

and is unused by 254 nm-H2 O2 based AOPs. The potential to exploit 185 nm radia-

39

tion for photochemical purification of water has been recognized for some time (7; 8).

cr

us

an

M

d

te

Ac ce p

40

ip t

23

41

The 185 nm wavelength occurs at the start of the broad absorption band of the first

42

excited state of water (9), with the decadic absorption coefficient for pure water

43

(aH2 O ) of 1.8 cm−1 (10), and resulting in the photolysis of water to generate ·OH

44

with a quantum yield (ΦH2 O ) of 0.3 (11):

185 nm

H2 O −−−−→ H· + ·OH

ΦH2 O = 0.3

(1)

2

Page 2 of 22

In the presence of dissolved oxygen, H· is scavenged at diffusion limited rates (12),

46

to produce the relatively nonreactive acid-base pair HO2· /O2· – (13):

HO2· O2·− + H+

k = 2 × 1010 M−1 s−1

(2)

pKa = 4.8

cr

H· + O2 −→ HO2·

ip t

45

(3)

The relatively longer lifetime of ·OH results in oxidative conditions and thus an AOP

48

that does not require chemical addition.

us

47

an

49

The high absorbance of water at 185 nm imposes substantial geometrical constraints

51

on reactor design, with approximately 99% of photons absorbed within a path length

52

of 1 cm in natural waters. As the lifetime of photogenerated radicals is on the or-

53

der of microseconds (14), substantial mixing is required to ensure treatment of any

54

non-irradiated reactor volume. A temperature coefficient for aH2 O of 0.05 cm−1 ◦C−1

55

has been reported between 20 and 50 ◦C (10; 15). The behaviour at lower tempera-

56

tures is not known. The impact of changes in aH2 O due to temperature depends on

57

reactor design. For reactors that minimize the non-irradiatd volume by using short

58

optical path lengths (≤ 1 cm), operation at lower temperatures may result in a 5 to

59

10% of 185 nm radiation reaching the reactor wall, with a corresponding decrease in

60

absorption and treatment efficiency.

d

te

Ac ce p

61

M

50

62

Seasonal fluctuations in surface water temperature may span a range from 0 to 20 ◦C

63

or more at latitudes far from the equator. Treatment processes in such locations

64

must ensure adequate performance regardless of temperature. While UV disinfec-

65

tion is relatively insensitive to water temperature, the influence of temperature on 3

Page 3 of 22

66

UV based AOPs is not well documented.

67

The temperature dependence of the quantum yield ΦH2 O at 185 nm is not known,

69

but may be assumed to follow an Arrhenius type relation with an activation energy

70

dependent on competing rates of radical recombination kr and escape ke from the

71

solvent cage:

us

cr

ip t

68

185 nm

ke · · − * → H· + ·OH H2 O − ) −− −− − − [H , OH]aq −

(4)

an

kr

The effective activation energy EΦ will depend on kr and ke , which were not di-

73

rectly accessible experimentally in the current study. The situation is analogous for

74

254 nm photolysis of H2 O2 , though more information is available on the tempera-

75

ture dependence of the quantum yield ΦH2 O2 , which is approximately unity at 25 ◦C

76

(16; 17):

254 nm

te

d

M

72

ke · · − * → HO· + ·OH H2 O2 − ) −− −− − − [HO , OH]aq −

Ac ce p

(5)

kr

77

Based on the limited data reported in the literature for 254 nm photolysis of H2 O2 ,

78

an activation energy EΦ for ΦH2 O2 at 254 nm is estimated to be in the range of 11 to

79

13 kJ mol−1 (16; 18; 19).

80

81

The temperature dependence of the 185 nm based AOP is a critical factor for con-

82

sideration in efficient 185 nm reactor design and is reported here for the first time.

83

A comparison with the temperature dependence of the 254 nm-H2 O2 AOP and fun-

4

Page 4 of 22

damental insight into the structure of liquid water are also discussed.

85

2. Theory

86

The ·OH driven treatment of UV based AOPs involves composite chemical reactions.

87

Component steps include photolytic generation of ·OH, reaction with target contam-

88

inants, and competition reactions with major solutes that act as radical scavengers.

89

These component steps themselves are composed of multiple elementary reactions.

90

Despite this complexity, the net reaction rate of many composite reactions may be

91

represented by a single Arrhenius expression, involving an overall experimentally

92

observed activation energy (20):

M

an

us

cr

ip t

84

k 0 = Atot e−Etot /RT

(6)

where k 0 is an overall reaction rate constant, Atot is the pre-exponential factor, Etot is

94

the activation energy, R is the universal gas constant 8.314 46 J K−1 mol−1 , and T is

95

the absolute temperature. Component steps may have individual activation energies.

96

Whether ·OH is generated by 185 nm photolysis of H2 O, or 254 nm photolysis of

97

H2 O2 , the experimentally observed k 0 for the ·OH degradation of a trace organic

98

contaminant C, in a solution containing a much larger concentration of the scavenger

99

S, and under steady-state conditions, may be expressed by:

Ac ce p

te

d

93

k0 =

100

kOH,C ΦΘ kOH,S

(7)

where kOH,C and kOH,S are the relevant second order ·OH rate constants, Φ is the 5

Page 5 of 22

corresponding quantum yield, and Θ contains all other terms assumed to be tem-

102

perature independent. In order for k 0 to remain constant throughout the irradiation,

103

the condition kOH,C [C]  kOH,S [S] must be satisfied. Equation 7 may be expanded

104

in terms of the three components as:

(10)

cr

Ac ce p

te

d

with Etot obtained from the experimental data and:

ln(k 0 ) = ln(Atot Θ) − Etot /RT 107

(9)

M

Etot = EC + EΦ − ES 106

(8)

from which the Etot may be related to the component activation energies by:

an

105

AC e−EC /RT AΦ e−EΦ /RT Θ AS e−ES /RT

us

k0 =

ip t

101

108

If Θ possesses a significant temperature dependence, then a plot of ln(k 0 ) vs. 1/T

109

will reveal curvature. In the case of 185 nm radiation, the term Θ contains aH2 O .

110

The absence of such curvature indicates that aH2 O temperature sensitivity is negligi-

111

ble under the conditions tested. Furthermore, it allows the estimation of one of the

112

component activation energies if the others are known. For diffusion limited reac-

113

tions, such as those involving ·OH, E is often in the range of 10 to 20 kJ mol−1 (12).

114

Examples of E for ·OH reactions include 21 kJ mol−1 for HCO3 – and 10 kJ mol−1 for

115

tert-butanol (12). It is important to note that an activation energy is not necessarily 6

Page 6 of 22

associated with an elementary reaction nor a transient intermediate. Nevertheless,

117

the concept may be useful in gaining mechanistic insights.

118

3. Methods and Materials

119

3.1. Overall Approach

120

Experiments consisted of a series of batch irradiations of assembled solutions of

121

known composition spiked with a probe compound and held in a temperature con-

122

trolled vessel. Under steady-state conditions, degradation of the probe compound C

123

follows pseudo-first order kinetics with an experimentally observed rate constant k 0 :

an

us

cr

ip t

116

M

ln([Ct ]/[Co ]) = −k 0 t

(11)

with changes in k 0 correlated to changes in solution temperature using Equations 7

125

and 8. All solutions contained approximately 0.25 µM of carbamazepine (CBZ) as the

126

probe, and 8 mg L−1 as C (0.2 mM) of tert-butanol as scavenger. For 254 nm-H2 O2

127

studies, 3.5 mg L−1 (0.1 mM) of H2 O2 was added prior to irradiation. All exposures

128

were performed in triplicate. Bovine catalase was used to quench H2 O2 following

129

irradiation. Dark controls for 254 nm-H2 O2 tests were also included.

te

Ac ce p

130

d

124

131

3.2. Materials

132

Solutions were made using ultrapure water produced by an Elga Purewater Option-Q

133

system (Elga Labwater, UK) and analytical grade reagents (Sigma Aldrich, USA).

134

The 185 nm optical path in the collimation tube was purged with ultrapure nitrogen

7

Page 7 of 22

135

(Praxair, Canada).

136

3.3. Apparatus

138

UV irradiations using conventional low pressure mercury lamps (Light Sources Inc,

139

USA) were conducted using two types of quasi-collimated beam apparatus (21). For

140

254 nm-H2 O2 experiments, a 42 W lamp was used with a quartz envelope doped

141

with titanium to block 185 nm. For experiments using both 185 and 254 nm, a sec-

142

ond apparatus was equipped with a 10 W lamp not doped with titanium and thus

143

transparent to both wavelengths. The optical path of the 185 nm enabled beam was

144

purged with ulatrapure nitrogen gas to prevent ozone formation. Samples exposed

145

to 254 nm alone were placed in open vessels with small stir bars and a liquid depth

146

of approximately 2.0 cm. Samples exposed to both 185 and 254 nm were sealed in

147

cylindrical fused silica cells of 1.0 cm path length (Starna, UK) with miniature teflon

148

coated stir bars. The distance from the lamps to liquid surface were made at least

149

five times the aperture to ensure quasi-collimation and held fixed for all experiments.

cr

us

an

M

d

te

150

ip t

137

The 185 nm irradiations used a custom made Peltier temperature controlled cell

152

holder (Quantum Northwest, USA). The 254 nm-H2 O2 tests used a custom water

153

jacketed borosilicate beaker (Cansci, Canada) with temperature controlled by a re-

154

circulating chiller (Thermo Fisher Scientific, USA). In both cases, temperature con-

155

trol within ± 0.1 ◦C was verified via fine gage thermocouple (Omega, Canada).

156

Ac ce p

151

157

3.4. Probe Compound Selection and Characterization

158

Since typically molar absorption coefficients () are less than 105 M−1 cm−1 and quan-

159

tum yields (Φ) are less than unity (22), direct photolysis at 185 nm is generally neg8

Page 8 of 22

ligible for contaminants or probes in aqueous systems at concentrations of 1 µg L−1

161

or less. Thus, the two main degradation processes for trace contaminants include

162

direct photolysis at 254 nm and oxidation by ·OH generated by 185 nm photolysis of

163

water. In order to study the effects of 185 nm alone, a simple method involves se-

164

lecting a probe compound for which the rate constant for direct photolysis at 254 nm

165

is negligible. In solutions of low absorbance (Aλ < 0.02), the rate of direct pho-

166

tolysis at 254 nm (kd0 ) is proportional to the product of molar absorption coefficient

167

(254 ) and photolysis quantum yield (Φ254 ) (23). Carbamazepine (CBZ) was found

168

to satisfy this condition well, with experimentally determined values 254 and Φ254

169

listed in Table 1. While para-chlorobenzoic acid (pCBA) has often been used as

170

a probe compound in AOP studies, the use of CBZ eliminates the need to correct

171

for direct photolysis at 254 nm under typical fluence values (doses), and its superior

172

chromatography by HPLC provide wider dynamic range and a lower limit of quanti-

173

tation. The second-order ·OH rate constant for CBZ was determined experimentally

174

by competitive kinetics with pCBA as the reference (24).

d

M

an

us

cr

ip t

160

te

175

Ac ce p

Table 1: Photochemical reaction parameters for probe compounds at 254 nm

Compound 254 (M−1 cm−1 ) CBZ pCBA

a

Φ254

kOH,C (M−1 s−1 )

6759 ± 190 0.00067 ± 0.00002 6.8 ± 0.6 × 109 3410 ± 75 0.011 ± 0.003 5.0 × 109 a

Reference (24)

176

3.5. Analysis

177

CBZ and pCBA were quantified by HPLC using a Dionex UltiMate 3000 System

178

(Thermo Fisher Scientific, USA). A 100 µL injection volume was delivered to a mobile 9

Page 9 of 22

phase composed of 30% acetonitrile and 70% water acidified to pH 2.5 with 10 mM

180

phosphoric acid, in an isocratic flow of 1.0 mL min−1 though a Nova-Pak C18 column

181

(Waters Corp., USA) maintained at 35 ◦C. UV detection was performed using 211

182

and 239 nm. The t-butanol concentration was verified by a GE Sievers M9 TOC An-

183

alyzer (GE Analytical Instruments, USA), using the UV-persulfate method. Absence

184

of anionic impurities was verified by HPLC using a Dionex Ion Chromatography sys-

185

tem. Hydrogen peroxide was measured using the I3 – method (25). Fluence rates at

186

254 nm, used for determinations of Φ254 , were measured using KI−KIO3 actinometry

187

(26; 27). Absorbance measurements for 254 determinations, H2 O2 quantitation, and

188

KI−KIO3 actinometry, were made using a Shimadzu UVmini-1240 Spectrophotome-

189

ter (Shimadzu, Japan).

190

4. Results

191

All decay curves exhibit first-order kinetics. Note that all experimental data and

192

calculations presented here are available in the associated dissertation (28).

193

4.1. 254 nm - H2 O2 Regime

194

The results for 254 nm - H2 O2 tests are displayed in Figure 1. The reaction rate

195

observed increases with temperature from 5 to 35 ◦C, as expected.

Ac ce p

te

d

M

an

us

cr

ip t

179

10

Page 10 of 22

3.4. Results 0

ip t

5 C

cr

0.4

20 C

us

0.6

0.8

an

ln([CBZ]t /[CBZ]o )

0.2

35 C

1

10

20

30

M

0

40

50

60

Time (min)

te

d

3.1: Temperature dependence 254 nm - H2 O [HL2−1 O2. ] [tBuOH] = 2 2 regime. Figure 1:Figure Temperature dependence in 254 nm - Hin ] = 3.5 mg = 2 O2 regime. [H 2O −1 3.5 mg L 1 . [tBuOH] = 8 mg L 1 as C. [CBZ] ' 0.25 µM. 8 mg L as C. [CBZ]o ' 0.25 µM. o

The values of k 0 are calculated from linear regression of the triplicate measurements

197

for each irradiation time used, with the standard error of the slope used to express

198

uncertainty σk0 . The uncertainty of ln(k 0 ) is calculated from the approximation

199

σln(k0 ) ≈ σk0 /k 0 (29). The calculated values are displayed in Table 2.

Ac ce p

196

Table 2: Effect of temperature on removal rate of CBZ in 254 nm-H2 O2 regime

T (◦C)

k 0 × 103 (min−1 ) ln(k 0 )

5

20

35

7.3 ± 0.3 11.4 ± 0.2 16.1 ± 0.5 −4.92 ± 0.04 −4.47 ± 0.02 −4.13 ± 0.03

[H2 O2 ] = 3.5 mg L−1 , [tBuOH] = 8 mg L−1 as C, [CBZ]o ' 0.25 µM

60

11

Page 11 of 22

200

The Arrhenius plot for the 254 nm-H2 O2 regime is displayed in Figure 2 using the

201

data tabulated in Table 2 and the linearized Equation 6. 3.4. Results

ip t

4

cr

4.2

us

ln(k0 )

4.4

an

4.6

4.8

5 3.1

3.3

3.4

3.5

M

3.2

1/T ⇥

103

K

3.6

3.7

1

te

d

Figure 2:Figure Arrhenius for 254plots nm-Hfor regime. ] = 3.5 mg [tBuOH] 3.2: plots Arrhenius nm-H[H [H2LO−12]. = 3.5 mg L= 81 .mg L−1 as 2 O2254 2O 2O 2 2regime. 1 C. [CBZ][tBuOH] = 8 mg L as C. [CBZ]o ' 0.25 µM. o ' 0.25 µM.

Temperature E↵ects in the 185 nm Regime 4.2. 1853.4.2 nm Regime

203

The results for 185 nm tests are displayed in Figure 3.3. As with the preThe results for 185 nm tests are displayed in Figure 3. As with the previous case,

Ac ce p

202

vious case, the reaction rate is observed to increase with temperature from

204

205

the reaction is observed to increase with temperature from 5 to 35 ◦C, though 5 C torate 35 C. to a lesser extent. As with the 254 nm-H2O2 regime, the observed pseudo-first order rate constants k 0 are calculated from linear regression of the triplicate measurements for each irradiation time used, with the standard error of the slope used to express uncertainty

k0 .

from the approximation in Table 3.2.

As before, the uncertainty of ln(k 0 ) is calculated

ln(k0 )

⇡

k0 /k

0.

The calculated values are displayed

The Arrhenius plot for the 185 nm regime is displayed in Figure 3.4 using the data tabulated in Table 3.2.

12 61

Page 12 of 22

3.4. Results 0

ip t

5 C

cr

2

us

20 C

3

35 C

an

ln([CBZ]t /[CBZ]o )

1

0

2

4

M

4

6

8

10

12

Time (min)

te

d

3.3: Temperature dependence in 185[tBuOH] nm regime. Figure 3:Figure Temperature dependence in 185 nm regime. = 8 mg[tBuOH] L−1 as C.=[CBZ]o ' 1 as C. [CBZ] ' 0.25 µM. 8 mg L o 0.25 µM. The slopes of both Arrhenius plots, as calculated by linear regression, allow

206

The values of k 0 and associated uncertainty are calculated in the same fashion as

207

before and displayed in Table 3. and are 254 nm-H 2O2 regimes. The values are displayed in Table 3.3.

Ac ce p

the determination of the experimental activation energy for both the 185 nm

Table 3: Effect of temperature on removal rate of CBZ in 185 nm regime

T (◦C)

k 0 × 102 (min−1 ) ln(k 0 )

5

20

35

24.7 ± 0.2 30.5 ± 0.5 35.6 ± 0.5 −1.40 ± 0.01 −1.19 ± 0.02 −1.03 ± 0.01

[tBuOH] = 8 mg L−1 as C, [CBZ]o ' 0.25 µM

208

The Arrhenius plot for the 185 nm regime is displayed in Figure 4 using the data

209

tabulated in Table 3.

62

13

Page 13 of 22

Table 3.3: Experimental (overall) activation energies for carbamazepine degradation in the presence of tert-butanol in 254 nm-H2O2 and 185 nm regimes

Ea (kJ mol

1)

254 nm-H2O2

185 nm

18.7 ± 0.9

8.6 ± 0.5

NB: These values pertain to the composite reactions and not elementary steps.

ip t cr

1.2

us

ln(k0 )

1

1.4

3.2

3.3

3.4

3.5

an

3.1

1/T ⇥

103

K

3.6

3.7

1

1

M

3.4: Arrhenius plotsin for regime.[tBuOH] [tBuOH] 8 mg L as as Figure 4: Figure Temperature dependence 185185 nmnm regime. == 8 mg L−1 C. C. [CBZ]o ' 0.25 µM. [CBZ]o ' 0.25 µM.

The slopes of both Arrhenius plots, as calculated by linear regression, 63 allow the

211

determination of the experimental activation energy Etot for both the 185 nm and

212

254 nm-H2 O2 regimes and are displayed in Table 4.

te

d

210

Ac ce p

Table 4: Experimental (overall) activation energies for carbamazepine degradation in the presence of tert-butanol in 254 nm-H2 O2 and 185 nm regimes

Ea (kJ mol−1 )

254 nm-H2 O2

185 nm

18.7 ± 0.9

8.6 ± 0.5

213

5. Discusion

214

The results indicate that the 185 nm AOP is less temperature sensitive than the

215

254 nm-H2 O2 AOP under the conditions tested. This result is understood to apply

216

to water matrices for which H2 O is the major absorber of 185 nm photons, where the 14

Page 14 of 22

scavenging term Σki [Si ] has a magnitude greater than 105 s−1 and where the activa-

218

tion energy of the target contaminant is less than the effective activation energy for

219

the scavenging term (i.e. EC < ES ). If the last condition is reversed (i.e. EC > ES ),

220

the observed rate k 0 may follow an inverse relationship with temperature.

ip t

217

221

In the present case, the scavenger is the pure substance tert-butanol, with an ·OH

223

activation energy ES reported as 10 ± 3 kJ mol−1 (30). Using an average of the

224

reported values of the activation energy of H2 O2 photolysis at 254 nm, EΦ = 12 ±

us

cr

222

1 kJ mol−1 , allows the estimation of the activation energy of ·OH with CBZ, EC ,

226

using equation 9. In this manner, the value EC = 17 ± 5 kJ mol−1 is obtained.

227

Application of equation 9 for the 185 nm regime allows for the estimation of EΦ , the

228

activation energy of the 185 nm photolysis of water itself. A value of EΦ ≈ 0 kJ mol−1

229

is obtained. The fundamental activation energies deduced from experimental data

230

are listed in Table 5.

M

an

225

te

d

Table 5: Summary of fundamental activation energies estimated from this work ·

Ac ce p

Ea (kJ mol−1 )

185 nm

OH + CBZ

H2 O −−−−→ H· + ·OH

17 ± 5

≈0

231

Explanation for the EΦ ≈ 0 kJ mol−1 value of H2 O photolysis at 185 nm relates to

232

the presence of H2 O monomers in the structure of the liquid state. While the pre-

233

cise structure of liquid H2 O remains in dispute (31), the conventional view based

234

on evidence from X-ray and neutron diffraction depicts virtually all molecules of

235

H2 O as dynamically hydrogen bonded to an average of four neighbours in an ice-like

236

tetrahedral motif with distorted bond angles (32; 33). The existence of non-hydrogen

237

bonded interstitial H2 O molecules (monomers) is supported by evidence from Raman

238

and infrared spectroscopy, though the proportion of such molecules is interpreted to 15

Page 15 of 22

239

be small (< 1%).

240

Evidence from far-UV absorption also supports the existence of interstitial H2 O

242

monomers. Extensive measurements of 185 nm absorption of ultrapure water in

243

both the vapour and liquid state have been reported (34), confirming observations

244

by others that the molar absorption coefficient of H2 O vapour is three orders of mag-

245

nitude greater than that of the liquid (10; 15; 35). Values of v = 22.1 M−1 cm−1

246

and ` = 0.0274 M−1 cm−1 at 23.5 ◦C were reported for the vapour and liquid states

247

respectively, with measurements of liquid absorption made at increasing temperature

248

(34). The ratio ` /v remained approximately 0.0012 between 23 and 27 ◦C, rising

249

sharply above 30 ◦C to 0.0090 at 91.8 ◦C. Such observations may be explained by

250

the existence of gas-like monomers in the liquid state, representing a fraction of all

251

molecules on the order of 10−3 in the vicinity of 20 ◦C, and with monomers responsible

252

for virtually all 185 nm photon absorption. The temperature dependence reported

253

by Stevenson is interpreted as an increase in monomer population with temperature

254

that nevertheless remains a minority even near the boiling point. This potentially

255

oversimplified depiction is consistent with the considerable degree of hydrogen bond-

256

ing remaining at the boiling point and the relatively high critical temperature of

257

water (36).

cr

us

an

M

d

te

Ac ce p

258

ip t

241

259

Upon 185 nm excitation of an H2 O monomer, only a relatively weak van der Waals

260

force must be overcome in order for the photo-products to escape the solvent cage,

262

since no hydrogen bonds are involved. Though the excited H2 O molecule may lose √ energy to the solvent by collision with a rate constant kr proportional to T , the

263

excess energy of the excited molecules itself is likely sufficient to overcome a van der

264

Waals energy of ∼ 5 kJ mol−1 . The activation energy for the photolysis of an H2 O

261

16

Page 16 of 22

265

molecule, Eφ ≈ 0, applies to the excited-state molecule that has absorbed a 185 nm photon. Such photons possess an energy of 647 kJ mol−1 while the bond-dissociation

267

energy between HO and H is approximately 494 kJ mol−1 (37). The excess energy, is

268

more than sufficient to overcome a postulated van der Waals force.

269

6. Conclusion

270

The temperature studies conducted indicate that, under the conditions tested, the

271

185 nm-AOP is relatively insensitive to temperature, potentially simplifying reactor

272

design. An activation energy for the ·OH reaction with carbamazepine has been

273

estimated as 17 ± 5 kJ mol−1 , which should be compared to the effective activation

274

energy of the background organic matter in kinetic studies when temperature is

275

varied. The activation energy for the 185 nm photolysis of H2 O has been estimated

276

to be approximately 0 kJ mol−1 and supports the view that 185 nm photon absorption

277

occurs in interstitial non-hydrogen bonded H2 O monomers present as an approximate

278

10−3 fraction of all molecules. Additional temperature dependence studies should be

279

investigated in water matrices for which H2 O is not the major absorber of 185 nm

280

photons.

281

Acknowledgements

282

The Natural Sciences and Engineering Council of Canada (NSERC) and the RES'EAU-

283

WaterNET Strategic Network are acknowledged for financial support of this work.

285

286

287

cr

us

an

M

d

te

Ac ce p

284

ip t

266

Refrences

[1] R. Phillips, Sources and Applications of Ultraviolet Radiation, Academic Press, 1983. 17

Page 17 of 22

[2] J. R. Bolton, B. Dussert, Z. Bukhari, T. Hargy, J. L. Clancy, Inactivation of

289

Cryptosporidium parvum by medium-pressure ultraviolet light in finished drink-

290

ing water, in: Proc. AWWA 1998 Annual Conference, Dallas, TX, Vol. A, 1998,

291

pp. 389–403.

ip t

288

[3] S. A. Craik, G. R. Finch, J. R. Bolton, M. Belosevic, Inactivation of Giardia

293

muris cysts using medium-pressure ultraviolet radiation in filtered drinking wa-

294

ter, Water Research 34 (18) (2000) 4325–4332.

us

cr

292

[4] A. D. Dotson, C. E. Rodriguez, K. G. Linden, UV disinfection implementation

296

status in US water treatment plants, Journal - American Water Works Associ-

297

ation 104 (5) (2012) 77–78.

an

295

[5] J. C. Kruithof, R. Van der Leer, W. Hijnen, Practical experiences with UV disin-

299

fection in the Netherlands., Journal of Water Supply: Research and Technology

300

- Aqua 41 (2) (1992) 88–94.

d

M

298

[6] W. H. Glaze, J. Kang, D. H. Chapin, The chemistry of water treatment processes

302

involving ozone, hydrogen peroxide and ultraviolet radiation, Ozone: Science &

303

Engineering 9 (4) (1987) 335–352.

305

306

307

308

Ac ce p

304

te

301

[7] O. Legrini, E. Oliveros, A. Braun, Photochemical processes for water treatment, Chemical Reviews 93 (2) (1993) 671–698.

[8] K. Zoschke, H. B¨ornick, E. Worch, Vacuum-uv radiation at 185 nm in water treatment–a review, Water Research 52 (2014) 131–145.

[9] H. Okabe, Photochemistry of Small Molecules, Wiley New York, 1978.

18

Page 18 of 22

[10] J. L. Weeks, G. Meaburn, S. Gordon, Absorption coefficients of liquid water

310

and aqueous solutions in the far ultraviolet, Radiation Research 19 (3) (1963)

311

559–567.

313

[11] N. Getoff, G. Schenck, Primary products of liquid water photolysis at 1236, 1470 and 1849 ˚ a, Photochemistry and Photobiology 8 (3) (1968) 167–178.

cr

312

ip t

309

[12] G. V. Buxton, C. L. Greenstock, W. P. Helman, A. B. Ross, Critical review of

315

rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl

316

radicals ( ·OH/ ·O – ) in aqueous solution, Journal of Physical and Chemical Ref-

317

erence Data 17 (2) (1988) 513–886.

an

us

314

[13] B. H. Bielski, D. E. Cabelli, R. L. Arudi, A. B. Ross, Reactivity of HO2· /O2· –

319

radicals in aqueous solution, Journal of Physical and Chemical Reference Data

320

14 (4) (1985) 1041–1100.

323

324

325

326

d

Water Science and Technology 35 (4) (1997) 1–8.

te

322

[14] J. Hoign´e, Inter-calibration of OH radical sources and water quality parameters,

[15] J. Barrett, A. L. Mansell, Ultraviolet absorption spectra of the molecules H2 O,

Ac ce p

321

M

318

HDO and D2 O, Nature 187 (1960) 138.

[16] J. H. Baxendale, J. A. Wilson, The photolysis of hydrogen peroxide at high light intensities, Transactions of the Faraday Society 53 (1957) 344–356.

327

[17] S. Goldstein, D. Aschengrau, Y. Diamant, J. Rabani, Photolysis of aqueous

328

h2o2: quantum yield and applications for polychromatic uv actinometry in pho-

329

toreactors, Environmental science & technology 41 (21) (2007) 7486–7490.

19

Page 19 of 22

[18] D. H. Volman, J. C. Chen, The photochemical decomposition of hydrogen per-

331

oxide in aqueous solutions of allyl alcohol at 2537 ˚ a. 1, Journal of the American

332

Chemical Society 81 (16) (1959) 4141–4144.

ip t

330

[19] J. P. Hunt, H. Taube, The photochemical decomposition of hydrogen perox-

334

ide. Quantum yields, tracer and fractionation effects, Journal of the American

335

Chemical Society 74 (23) (1952) 5999–6002.

cr

333

[20] K. J. Laidler, Chemical Kinetics, 3rd Edition, Harper & Row, New York, 1987.

337

[21] J. R. Bolton, K. G. Linden, Standardization of methods for fluence (uv dose)

338

determination in bench-scale uv experiments, Journal of Environmental Engi-

339

neering 129 (3) (2003) 209–215.

an

M

341

[22] R. P. Wayne, Principles and Applications of Photochemistry, Oxford University Press, 1988.

d

340

us

336

[23] J. R. Bolton, M. I. Stefan, Fundamental photochemical approach to the concepts

343

of fluence (uv dose) and electrical energy efficiency in photochemical degradation

344

reactions, Research on Chemical Intermediates 28 (7) (2002) 857–870.

Ac ce p

te

342

345

[24] P. Neta, L. M. Dorfman, Pulse radiolysis studies. XIII. Rate constants for the

346

reaction of hydroxyl radicals with aromatic compounds in aqueous solutions,

347

348

in: E. J. Hart (Ed.), Radiation Chemistry: Volume I. Aqueous Media, Biology, Dosimetry, ACS Publications, 1968, Ch. 15.

349

[25] N. V. Klassen, D. Marchington, H. C. McGowan, H2 O2 determination by the I3 –

350

method and by KMnO4 titration, Analytical Chemistry 66 (18) (1994) 2921–

351

2925.

20

Page 20 of 22

[26] R. O. Rahn, Potassium iodide as a chemical actinometer for 254 nm radiation:

353

use of lodate as an electron scavenger, Photochemistry and Photobiology 66 (4)

354

(1997) 450–455.

ip t

352

[27] J. R. Bolton, M. I. Stefan, P.-S. Shaw, K. R. Lykke, Determination of the

356

quantum yields of the potassium ferrioxalate and potassium iodide–iodate acti-

357

nometers and a method for the calibration of radiometer detectors, Journal of

358

Photochemistry and Photobiology A: Chemistry 222 (1) (2011) 166–169.

361

362

us

dissertation, University of British Columbia, Vancouver, Canada (2017).

an

360

[28] L. Furatian, The Use of 185 nm Radiation for Drinking Water Treatment, PhD

[29] D. C. Harris, Quantitative Chemical Analysis, 8th Edition, W. H. Freeman, 2010.

M

359

cr

355

[30] B. Ervens, S. Gligorovski, H. Herrmann, Temperature-dependent rate constants

364

for hydroxyl radical reactions with organic compounds in aqueous solutions,

365

Physical Chemistry Chemical Physics 5 (9) (2003) 1811–1824.

te

d

363

[31] P. Ball, Water: An enduring mystery, Nature 452 (7185) (2008) 291–292.

367

[32] F. Franks, Water: A Matrix of Life, Royal Society of Chemistry, 2000.

368

[33] H. S. Frank, Structural models, in: F. Franks (Ed.), Water a Comprehensive

369

Treatise: Volume 1: The Physics and Physical Chemistry of Water, Plenum

370

371

372

Ac ce p

366

Press, 1972, Ch. 14.

[34] D. P. Stevenson, On the monomer concentration in liquid water, The Journal of Physical Chemistry 69 (7) (1965) 2145–2152.

21

Page 21 of 22

373

374

[35] K. Watanabe, M. Zelikoff, Absorption coefficients of water vapor in the vacuum ultraviolet, Journal of the Optical Society of America 43 (9) (1953) 753–754. [36] T. W. Marin, I. Janik, D. M. Bartels, D. M. Chipman, Vacuum ultraviolet

376

spectroscopy of the lowest-lying electronic state in subcritical and supercritical

377

water, Nature Communications 8.

cr

d

M

an

us

lication 1 (1).

te

379

[37] B. Darwent, Bond dissociation energies in simple molecules, NIST Special Pub-

Ac ce p

378

ip t

375

22

Page 22 of 22