Reaction of bromine and chlorine with phenolic compounds and natural organic matter extracts – Electrophilic aromatic substitution and oxidation

Water Research 85 (2015) 476e486

Contents lists available at ScienceDirect

Water Research journal homepage: www.elsevier.com/locate/watres

Reaction of bromine and chlorine with phenolic compounds and natural organic matter extracts e Electrophilic aromatic substitution and oxidation Justine Criquet a, b, c, Eva M. Rodriguez a, d, e, Sebastien Allard b, Sven Wellauer a, Elisabeth Salhi a, Cynthia A. Joll b, Urs von Gunten a, e, f, * a

Eawag, Swiss Federal Institute of Aquatic Science and Technology, CH-8600 Dübendorf, Switzerland Curtin Water Quality Research Centre, Curtin University, GPO Box U1987, Perth, WA 6845, Australia Universit
e Lille 1 Sciences and Technologies, LASIR, UMR CNRS 8516, 59655 Villeneuve d'Ascq, France d Departamento de Ingeniería Química y Química Física, Universidad de Extremadura, Avda. Elvas s/n, 06006 Badajoz, Spain e School of Architecture, Civil and Environmental Engineering (ENAC), Ecole Polytechnique F
ed
erale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland f Institute of Biogeochemistry and Pollutant Dynamics, ETH Zürich, CH-8092 Zürich, Switzerland b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 April 2015 Received in revised form 27 August 2015 Accepted 27 August 2015 Available online 2 September 2015

Phenolic compounds are known structural moieties of natural organic matter (NOM), and their reactivity is a key parameter for understanding the reactivity of NOM and the disinfection by-product formation during oxidative water treatment. In this study, species-specific and/or apparent second order rate constants and mechanisms for the reactions of bromine and chlorine have been determined for various phenolic compounds (phenol, resorcinol, catechol, hydroquinone, phloroglucinol, bisphenol A, phydroxybenzoic acid, gallic acid, hesperetin and tannic acid) and flavone. The reactivity of bromine with phenolic compounds is very high, with apparent second order rate constants at pH 7 in the range of 104 to 107 M
1 s
1. The highest value was recorded for the reaction between HOBr and the fully deprotonated resorcinol (k ¼ 2.1  109 M
1 s
1). The reactivity of phenolic compounds is enhanced by the activating character of the phenolic substituents, e.g. further hydroxyl groups. With the data set from this study, the ratio between the species-specific rate constants for the reactions of chlorine versus bromine with phenolic compounds was confirmed to be about 3000. Phenolic compounds react with bromine or chlorine either by oxidation (electron transfer, ET) or electrophilic aromatic substitution (EAS) processes. The dominant process mainly depends on the relative position of the hydroxyl substituents and the possibility of quinone formation. While phenol, phydroxybenzoic acid and bisphenol A undergo EAS, hydroquinone, catechol, gallic acid and tannic acid, with hydroxyl substituents in ortho or para positions, react with bromine by ET leading to quantitative formation of the corresponding quinones. Some compounds (e.g. phloroglucinol) show both partial oxidation and partial electrophilic aromatic substitution and the ratio observed for the pathways depends on the pH. For the reaction of six NOM extracts with bromine, electrophilic aromatic substitution accounted for only 20% of the reaction, and for one NOM extract (Pony Lake fulvic acid) it accounted for <10%. This shows that for natural organic matter samples, oxidation (ET) is far more important than bromine incorporation (EAS). © 2015 Elsevier Ltd. All rights reserved.

Keywords: Bromine Chlorine Hydroxy aromatic compounds Natural organic matter Electrophilic aromatic substitution Electron transfer

1. Introduction

* Corresponding author. Eawag, Swiss Federal Institute of Aquatic Science and Technology, CH-8600 Dübendorf, Switzerland. E-mail address: [email protected] (U. von Gunten). http://dx.doi.org/10.1016/j.watres.2015.08.051 0043-1354/© 2015 Elsevier Ltd. All rights reserved.

Chemical disinfection is widely used for the provision of safe drinking water. However, disinfection by-products (DBPs) are formed during disinfection processes from reactions between the applied chemical oxidants (e.g. chlorine) and naturally occurring organic and inorganic precursors, such as natural organic matter

J. Criquet et al. / Water Research 85 (2015) 476e486

(NOM) and bromide (Sedlak and von Gunten, 2011; Heeb et al., 2014). Some DBPs pose potential health risks and are subject to regulation (Richardson et al., 2007; Hrudey, 2009; Hrudey and Charrois, 2012; Sedlak and von Gunten, 2011; Narotsky et al., 2013; Yang et al., 2014). Chlorination of bromide-containing water usually induces the formation of a mix of chlorinated and brominated DBPs (Richardson et al., 2003; Chowdhury et al., 2010; Hua and Reckhow, 2012; Pan and Zhang, 2013; Roccaro et al., 2014). During chlorination, bromide is oxidized to hypobromous acid/hypobromite (kHOCl=Br
¼ 1550 M
1 s
1; kClO
=Br
¼ 9  10
4 M
1 s
1; Kumar and Margerum, 1987), which are the main bromine species formed in fresh waters (Heeb et al., 2014). Phenolic groups are known constituents of NOM and phenolic compounds (the term phenols is also used in the text to represent polyhydroxy aromatic compounds) have often been used as model compounds to represent substructures within NOM (Bond et al., 2009). Hypohalous acids (chlorine, bromine, iodine) can react with phenols through electrophilic aromatic substitution (EAS), as well as by oxidation, i.e., electron transfer (ET) reactions (Rook, 1976, 1977; Gallard and von Gunten, 2002). HOBr is generally more reactive than HOCl. For phenolic compounds, QSAR modelling showed a ratio of the second order rate constants for the reactions of HOBr or HOCl with phenolic compounds, kHOBr/kHOCl of z3  103, demonstrating the high propensity of phenolic compounds to bromination as compared to chlorination if bromide is present in the water (Heeb et al., 2014). The rate for the reactions of chlorine and bromine with phenolic compounds is pH-dependent due to the differences in reactivity of the hypohalous acid and the hypohalite anion and the higher reactivity of the phenolate compared to the phenol (Rebenne et al., 1996; Gallard and von Gunten, 2002; Deborde and von Gunten, 2008; Heeb et al., 2014). This can be explained by the higher electrophilicity of HOX compared to XO
and the increased electron density of the phenolate compared to the phenol (Heeb et al., 2014). Based on this reactivity pattern, apparent second order rate constants are pH dependent and show a maximum at the mean of the pKa values of the two involved acid-base couples (HOX/OX
and phenol/phenolate) (Heeb et al., 2014; Criquet et al., 2012). Phenol groups are strongly activating and direct EAS reaction in ortho- or para-positions. The products of EAS reactions of phenols with HOX are mono-, di- and tri-halophenols. These reactions have been studied in some detail because of the off-flavour characteristics of the products and their role for taste and odour of drinking waters (Gallard et al., 2003; Acero et al., 2005; Heitz et al., 2003). For some phenols, particularly those with two hydroxy groups in meta position to each other, the reaction pathway includes ringopening of the aromatic ring and production of regulated DBPs such as the trihalomethanes (THMs) (Boyce and Hornig, 1983). The formation of THMs from a variety of phenolic compounds has been extensively studied (Ichihashi et al., 1999; Chang et al., 2006; Arnold et al., 2008). However, depending on the substitution patterns on the aromatic ring, the yield of THM formation can be vastly variable. Among phenols, resorcinol (1,3-dihydroxybenzene) showed the highest yield of THM formation (Norwood et al., 1980; Bichsel and von Gunten, 2000). In contrast, Bond et al. (2009) demonstrated that despite a high chlorine demand, tannic acid (a polymer of gallic acid (3,4,5-trihydroxybenzoic acid); for structure, see Supporting Information (SI), Fig. S1) showed a low DBP substitution efficiency (0.5% of Cl-incorporation compared to ca. 45% for resorcinol). It was proposed that chlorine was consumed in oxidation of tannic acid and/or in production of other DBPs, which weren't measured (Bond et al., 2009). Oxidation of phenolic compounds can potentially lead to quinone type structures (Wenk et al., 2013). These electron transfer (ET) reactions depend strongly on the electronic properties of the substituents and electron-withdrawing

477

substituents can inhibit this reaction completely. Moreover, halobenzoquinones have been identified as chlorine/bromine disinfection by-products in drinking water (Zhao et al., 2012; Wang et al., 2014). Haloquinones were previously predicted as potential DBPs (Bull et al., 2006) and they may be relevant to the known elevated risk of human bladder cancer due to consumption of chlorinated drinking water (Mills et al., 1998). In the current study, apparent and species-specific second order rate constants for the reactions of bromine and chlorine with a variety of phenolic compounds were determined. Furthermore, the extent of the reaction of bromine with phenolic compounds and NOM extracts by electrophilic aromatic substitution and/or oxidation processes was investigated. 2. Materials and methods 2.1. Chemicals All reagents used were of the highest available purity. Bromine solutions were produced from a chlorine stock solution and bromide as described previously (Criquet et al., 2012). Briefly, an aliquot of a chlorine stock solution (1.2 M) is added to a solution of bromide with a slight excess of bromide compared to chlorine concentration (5%). Experiments were performed in ultrapure water (Barnstead Nanopure (Skan); TOC <0.2 mg L
1; 18.2 mU cm). The pH was controlled by 5 mM phosphate, borate, or carbonate buffers. Poorly soluble compounds were solubilized in methanol; the effect of methanol (usually <0.5% v/v, except for the determination of the reactivity of chlorine with flavone, where MeOH was present at <5% v/v) in our experimental systems was tested and found not to interfere with the experiments related to kinetics or elucidation of reaction pathways. The structures of the phenolic compounds used for the kinetic studies are presented in Table 1. The NOM extracts were purchased from the International Humic Substances Society (IHSS); the percentage of aromatic carbon has been reported previously (Thorn et al., 1989). Suwannee River DOM (SRDOM; catalogue number: 2R101N; aromatic carbon: 23%), Suwannee River humic acid (SRHA; catalogue number: 2S101H; aromatic carbon: 31%), Suwannee River fulvic acid (SRFA; catalogue number: 2S101F; aromatic carbon: 22%), Nordic reservoir NOM (NR; 1R108N; aromatic carbon: 19%), Pony Lake fulvic acid (PL; 1R109F; aromatic carbon: 12%), Elliot soil humic acid (ESHA; catalogue number: 1S102H; aromatic carbon: 50%) and Leonardite humic acid (LEHA; catalogue number: 1S104H; aromatic carbon: 58%) were used as received. We chose to use commercially available DOM extracts because they have been used as references in oxidation studies, and their chemical properties are known. Furthermore, these extracts represent a wide variety of organic materials from allochthonous (SR and NR) and autochthonous (PL) aquatic systems to extracts from agricultural soil (ES) and lignite coal mining (LE). 2.2. Analytical methods Bromide analyses were performed by ion chromatography (Dionex ICS 3000) on an AS9-HC column with a quantification limit of 10 mg L
1 and a standard deviation of ±10% (Salhi and von Gunten, 1999). The oxidant concentration was measured using the colorimetric method based on diethyl-p-phenylene diamine (DPD) by an absorbance measurement of the sample at 515 nm (Rodier et al., 2009). A linear calibration was performed with oxidant concentrations in the range of 0.1e5 mM with a 5 cm cell for better accuracy at low concentrations. When needed, the gallic acid concentration was analyzed by high performance liquid chromatography (HPLC) (Agilent 1100) with a 15 cm long, 0.4 cm i.d.

478

J. Criquet et al. / Water Research 85 (2015) 476e486

Table 1 The selected phenolic compounds and flavone with the corresponding pKa values and method(s) used for the determination of the second order rate constant for their reaction with bromine and/or chlorine. Compounds

Structure

pKa

Method(s) for rate constant determination

Phenol

10.0

Quench-Flow

Resorcinol

9.4 11.2

Quench-Flow and competition kinetics

Catechol

9.3 11.6

Stopped-Flow

Hydroquinone

9.9 11.4

Stopped-Flow

Phloroglucinol

8.0 9.2 14

Quench-Flow and competition kinetics

Bisphenol A

9.6 10.2

Quench-Flow

p-Hydroxybenzoic acid

4.5 9.3

Quench-Flow (batch experiments for chlorination)

Gallic acid

4.3 8.7 11.5

Stopped-Flow (batch experiments for chlorination)

Quench-Flow (batch experiments for chlorination)

Flavone

9.7

Hesperetin

Tannic acid

Fig. S1 (SI)

Competition kinetics (Quench-Flow for chlorination)

Stopped-Flow

J. Criquet et al. / Water Research 85 (2015) 476e486

Kromasil C18 column with 20/80 v/v acetonitrile/water (with 0.1% phosphoric acid) as mobile phase (flow rate 1 mL min
1) with detection at 254 nm (limit of quantification (LOQ) 0.08 mM; %RSD 1.7). 2.3. Determination of apparent second order rate constants All kinetic experiments were performed at 24 ± 1  C. For slow kinetics (kapp  103 M
1 s
1; e.g., chlorination of p-hydroxybenzoic acid, gallic acid and flavone), experiments were performed in a 500 mL batch reactor equipped with a dispenser. Small volumes of sodium hypochlorite stock solutions were added at t ¼ 0 to a buffered solution containing the target compound, and aliquots of the reaction solution were withdrawn after certain reaction times and quenched in a DPD/phosphate buffer mixture at pH 6.5 to determine the residual oxidant concentrations. Concentrations of sodium hypochlorite stock solutions were chosen to obtain sufficient sensitivity for the analysis of chlorine and a manageable reaction time. For gallic acid, aliquots were withdrawn at different times, quenched with sodium thiosulfate and the gallic acid concentration was analyzed by HPLC. For fast kinetics (104 < kapp < 8  106 M
1 s
1), a quench-flow system (Biologic SFM-400/Q) was used which allows fast mixing of the reactants via drive syringes. The speed of the driving, as well as the size of the mixer and the delay line (190 mL), determine the reaction time. The reaction mixtures (phenolic compound þ oxidant) were aged for 20 ms to 3 s and the residual HOX was quenched by DPD, which was introduced in excess in a second mixing-T. The reaction mixture (5 mL) was collected and the concentration of the residual oxidant was determined by the DPD method described above. These measurements were repeated at least twice for each reaction time. For very fast kinetics (kapp > 8  106 M
1 s
1; e.g., bromination of resorcinol), a competition kinetics method was selected to determine the apparent second order rate constants. For this method, phenol was selected as the competitor. The apparent second order rate constants were determined relative to the reactivity of the oxidant towards the competitor as described elsewhere ~ oz and von Sonntag, 2000). (Mun Intermediates formed from chlorination and bromination of polyhydroxyphenols with OH groups arranged in o- or p-positions (catechol, hydroquinone, gallic acid and tannic acid) are reactive semiquinones/quinones capable of interacting with the competitor or with DPD, as well as to disproportionate and/or polymerize (Alegría et al., 1996; Uchimiya and Stone, 2009). Therefore, competition kinetics was not suitable for these polyhydroxyphenols. In these cases, kinetic studies were performed by stoppedeflow with a UVeVis detector (HI-TECH KinetAsyst SF61DX2, TgK Scientific) working in excess of polyphenol under pseudo-first order conditions (polyphenol/HOX initial molar ratio, 10:1) and measuring spectrophotometrically the formation rate of the corresponding quinonic intermediate (at a wavelength of 390 nm for catechol, 246 nm for hydroquinone and 500 nm for gallic and tannic acids). Since o- and p-hydroxyphenols are easily oxidised by dissolved oxygen at pH  8, apparent second order rate constants for the reaction of these compounds with HOBr were only determined at pH 7. 2.4. Determination of the species-specific rate constants and modelling the apparent second order rate constants as a function of the pH To determine the species-specific rate constants from the pHdependence of the apparent second order rate constants for the target compound-HOBr reactions, the reaction systems were fitted

479

by a kinetic model considering the speciation of the oxidant and the target compound (Gallard et al., 2003). Each experiment was carried out in the pH-range 7 to 11.5, and a set of 11 apparent second order rate constants was used to determine the species-specific rate constants. The second order rate constants were determined based on Eqs. (1)e(8) HOX þ PhOH / products

k1

(1)

HOX þ PhO
/ products

k2

(2)

HOX þ PhðO
Þ2 /products

k3

(3)

XO
þ PhO
/ products

k4

(4)

XO
þ PhðO
Þ2 /products

k5

(5)

XO
þ Hþ %HOX

KaHOX

(6)

.
PhOH PhðOHÞ2 %PhO
PhðOHÞO
þ Hþ Ka1

(7)

PhðOHÞO
%PhðO
Þ2 þ Hþ

(8)

Ka2

with pKaHOBr ¼ 8.8 and pKaHOCl ¼ 7.5; pKa values of phenolic compounds are given in Table 1. The following phenolic functional groups were considered: PhOH represents the non-dissociated compounds; PhO
represents the compounds after dissociation of one Hþ; and Ph(O
)2 represents doubly deprotonated compounds with two or more hydroxyl groups on the aromatic ring. For monohydroxy aromatic compounds, the rate law considering the set of equations (Eqs. (1)e(8)) is given by Eq. (9):

i h d½HOXtot ¼ k1 ½HOX½PhOH þ k2 ½HOX PhO

dt h ih i þ k4 XO
PhO


(9)

In the case of polyhydroxy aromatic compounds, the doubly deprotonated species also have to be taken into account (Eq. (10)):

  d½HOX tot ¼ k1 ½HOX  PhðOHÞ2 þ k2 ½HOX ½PhðOHÞO

dt     þ k3 ½HOX  PhðO
Þ2 þ k5 ½ XO
 PhðO
Þ2

(10)

with [HOX]tot ¼ [HOX] þ [XO
] The kinetics of the two reactions between the bromine species and the phenolic compounds (HOBr þ PhO
and BrO
þ PhOH) cannot be distinguished kinetically since they have the same pHdependence. Rather than considering the two reactions in parallel, it is usually assumed that only one of the two pathways is important (Criquet et al., 2012; von Gunten and Oliveras, 1997). From a chemical point of view, HOBr is a stronger electrophile and PhO
is a stronger nucleophile. Therefore, the reaction of HOBr with PhO
(Eq. (2)) is considered to be the dominant pathway for the reaction of bromine with phenolic compounds. The same argument can also be applied for the reactions between bromine and dihydroxybenzene (HOBr þ Ph(O
)2 and BrO
þ Ph(OH)O
) for which only the reaction between HOBr and Ph(O
)2 is considered. The same procedure was applied for the determination of the species-specific rate constants for chlorine. Also, in this case, the reactivity of ClO
was neglected. Details on the determination of species-specific second order rate constants from experimental apparent rate constants are given in the SI (Text S1).

480

J. Criquet et al. / Water Research 85 (2015) 476e486

2.5. Pathways for the reaction of bromine with phenolic compounds and NOM extracts (EAS vs ET) To characterize the pathways of the reactions between bromine and the selected phenolic compounds or flavone, a 50-fold excess of the phenolic compound was added to a buffered solution containing 3 mM HOBr. An aliquot of target compound stock solution was added resulting in a final concentration of 150 mM of the target compound. In the case of phenolic compounds with low solubility, concentrations of 1 and 50 mM were used, respectively, for bromine and the phenolic compounds. For experiments with NOM extracts, concentrations of 3.6 mg C L
1 and 1 mM of bromine were used. This concentration of the NOM extract was chosen to achieve a similar carbon concentration as in the experiments with the target compounds (50 mM corresponding to 3.6 mg C L
1). Even if a part of the carbon atoms in NOM are not included in aromatic structures, the large excess (50:1) guarantees a stoichiometric excess of aromatic structures compared to the oxidant. To test the extent of the ET reaction (reduction of HOBr to Br
), the bromide concentration was analyzed by ion chromatography after completion of the reaction (no oxidant residual). In these conditions the reaction is very fast, and the bromide concentration has been determined at least 24 h after mixing. As a blank measurement, an equivalent bromine solution was quenched by an excess of sodium sulfite to determine the sum of the initial and the reducible bromide concentration in the bromine solutions ([HOBr]0 þ [Br
]0). With this approach, the molar excess of bromide (5%) relative to chlorine used for the preparation of the HOBr stock solution was taken into account in the calculations. The experiments were performed at pH 7.0 and 8.0 in a phosphate buffer (5 mM) and the fraction of bromine which reacted by EAS was determined by ([HOBr]0 þ [Br
]0
[Br
]f)/ [HOBr]0, where [Br
]f is the concentration of bromide in solution after complete reaction of HOBr with the target compound or the NOM isolate. 3. Results and discussion 3.1. Determination of species-specific rate constants To determine the species-specific rate constants of the reaction of bromine and chlorine with phenolic compounds or flavone, experiments were performed at various pH values. The determined apparent second order rate constants for the reactions between bromine and selected phenolic compounds are plotted as a function of the pH in Fig. 1. The species-specific rate constants were determined by a non-linear least-squares regression as a function of the pH, according to Section 2.4 and Text S1 (SI) considering the pKa values shown in Table 1 (raw data are presented in Fig. S2 and Text S2, SI). The pKa value associated with the carboxylic acid functional group of p-hydroxybenzoic acid was not considered, as the compound is mainly dissociated in the investigated pH-range. The third pKa of phloroglucinol (pKa3 ¼ 14) was also neglected, as it is significantly above the pH-range of this study. The modeled pHdependent apparent second order rate constants for selected phenolic compounds are shown in Fig. 1 (lines) and are in good agreement with the measured data (symbols). In Fig. 1, also the speciation of HOBr/BrO
and of the various phenolic compounds are shown. For phenol and p-hydroxybenzoic acid, the apparent second order rate constants show a maximum at the mean of the two pKa values of the involved species (1/2(pKaHOBr þ pKa1); pKaHOBr ¼ 8.8; pKa1 ¼ 10.0 and 9.3 for phenol and p-hydroxybenzoic acid, respectively). For phenol, the speciation of phenol and bromine as a function of pH is included in Fig. 1, confirming that the main reaction occurs between HOBr and the phenolate (Criquet et al., 2012;

von Gunten and Oliveras, 1997). This is further underlined by the relatively high value of the species-specific rate constant k2 (Table 2). In the case of phenolic compounds with two or more acid-base equilibria (i.e. two or more hydroxyl substituents, e.g., resorcinol), the pH-dependence is weighted by mainly k2 and k3, and no rules could be derived to predict the pH of the maximum apparent second order rate constant. Table 2 summarizes the species-specific second order rate constants for the reaction of bromine with the selected phenolic compounds or flavone and/or their apparent rate constants at pH 7 determined in this study. Few rate constants for bromination of phenolic compounds were published previously; our results are in agreement with studies performed with the same kind of apparatus (Echigo, 2002; Echigo and Minear, 2006) where species-specific rate constants k2 (HOBr/PhO
) of 4.1  107 and 1.4  107 M
1s
1 were found, respectively for phenol and p-hydroxybenzoic acid. The previously published data for phenol (Gallard et al., 2003; Guo and Lin, 2009) were overestimated and are not considered here (for details see Heeb et al., 2014). The reactivity of bromine with phenolic compounds or flavone is very high, with apparent second-order rate constants at pH 7 in the range 104 to 107 M
1 s
1 (Table 2; raw data presented in SI, Fig. S2, Texts S2 and S3). The reactions between HOBr and the deprotonated forms of the phenolic compounds, i.e., the phenolate forms for phenol and p-hydroxybenzoate (k2) and the diphenolate forms for resorcinol, phloroglucinol and bisphenol A (k3), are dominant for the apparent rate constants at pH 7. The corresponding species-specific rate constants for the phenolates can be up to 108 e 109 M
1 s
1. Resorcinol, phloroglucinol and bisphenol A have a higher reactivity than phenol. For resorcinol and phloroglucinol, compared to phenol, the extra hydroxyl group(s) (either protonated or deprotonated) on the aromatic ring strongly activate(s) the aromatic system and enhance(s) the relative reactivity with bromine. For bisphenol A, the reactivity is enhanced due to the presence of two phenolic rings. The effect of each substituent also depends on their relative position on the aromatic ring. A Quantitative Structure-Activity Relationship (QSAR) between the species-specific rate constants and the aromatic ring substitution patterns can account for these effects (for details see Hansch et al., 1991; Heeb et al., 2014). A QSAR-type correlation determined in a previous study (Heeb et al., 2014) for the reaction of HOBr with substituted phenols was expanded with data from the current investigation to increase the range of the previous correlation by including negative Hammett constants, corresponding to activated phenolates. The updated QSAR-type correlation in P this study (log (k) ¼ 7.8 e 3.2 s with R2 ¼ 0.8) is in good agreement P with the previous one (log (k) ¼ 7.8 e 3.5 s) (Heeb et al., 2014) (Fig. S3, SI). Apparent and, in some cases, species-specific second order rate constants for the reaction of chlorine with the selected phenolic compounds and flavone from this study and from previous studies are shown in Table 3 (raw data are given in Figs. S4 and S5 and Text S4). The second order rate constants for chlorine are all smaller than for the corresponding reactions with bromine. Since the electronegativity of Cl (3.16) is higher than for Br (2.96), the partial positive charge (dþ) on the bromine atom in HOBr (or BrO
) is higher than the one for chlorine in HOCl (or ClO
). Therefore the electrophilic aromatic substitution reaction of bromine with phenol is faster than the corresponding reaction with chlorine. A good correlation between the species-specific second order rate constants of chlorine and bromine was obtained when combining the previously reported data with the rate constants determined in this study (Fig. S6, SI). The ratio of the second order rate constants for the reactions with HOBr or HOCl, kHOBr/kHOCl was approximately 3000, which is in good agreement with our previous study (Heeb et al., 2014). However, the ratio between the reactivity of bromine

J. Criquet et al. / Water Research 85 (2015) 476e486

481

Fig. 1. pH dependence of the apparent second order rate constants for the reactions of various phenolic compounds with bromine and pH-dependent speciation of active species. Symbols: experimental data; bold lines: least squares fit according to Section 2.4.; fine lines: speciation of bromine and the phenolic compounds.

and chlorine with resorcinol is smaller, and the corresponding data point is below the linear correlation line (i.e., about a factor of 250 and 20, respectively, for Ph(OH)O
and Ph(O
)2). This could be due to the very high species-specific rate constants observed for the reaction of bromine with resorcinol (Table 2), which are close to the diffusion limitation. Therefore, no increase of the rate constant is possible for bromine, while for chlorine the rate constant can still increase. These results confirm the existing correlation between the reactivity of the two oxidants towards phenolic compounds for a larger range of k-values, but also show the limitations for highly reactive compounds such as resorcinol. This type of correlation is only valid for species-specific rate constants and cannot be applied

to apparent rate constants. Due to the influence of pKa values, the apparent rate constants of the reaction of bromine with the selected compounds are always higher than the corresponding rate constants of chlorine but the ratios vary between 10 and 9  103 and no simple correlation can be made. For the reasons explained in the materials and methods section, stopped-flow was used for the determination of the reactivity of bromine and chlorine with selected hydroxyphenols (hydroquinone, catechol, gallic and tannic acids). The reactivity was investigated by monitoring the evolution of the absorbance at 500 nm (raw data are presented in Texts S3 and S4, SI). However, in contrast to the previous compounds, it was not possible to determine the

482

J. Criquet et al. / Water Research 85 (2015) 476e486

Table 2 Species-specific and apparent second order rate constants for reaction of bromine with phenolic compounds and flavone at pH 7. Compound

Species-specific second order rate constants (M
1 s
1) k1

Bisphenol A Catechol Flavone Gallic acid Hesperetin Hydroquinone Phenol Phloroglucinol p-Hydroxybenzoic acid Resorcinol Tannic acid

HOBr/PhOH

k2

HOBr=PhO

k3




(2.1 ± 1.2)  105

(8.0 ± 0.2)  107

<1 (1.3 ± 0.3)  107 (9.5 ± 3.7)  104 (7.9 ± 1.4)  106

(6.6 (2.9 (1.1 (3.5

± ± ± ±

1.5) 0.6) 0.1) 0.2)

   

107 107 107 108

k4




HOBr=PhðO Þ2




BrO =PhO




(2.4 ± 0.1)  108

(2.5 ± 0.2)  108 (2.1 ± 0.8)  109

k5







BrO =PhðO Þ2

(4.8 ± 1.6)  105

(3.5 ± 6.5)  104 (7.7 ± 4.3)  104

(2.1 ± 0.8)  107 (1.5 ± 2.8)  106

Apparent second order rate constant at pH 7 kapp (M
1 s
1) (4.0 (2.7 (1.4 (1.5 (8.2 (6.4 (6.5 (1.4 (1.5 (9.0 (1.7

± ± ± ± ± ± ± ± ± ± ±

1.2) 0.1) 0.1) 0.1) 1.1) 0.1) 1.5) 0.3) 0.4) 1.4) 0.1)

          

105 105 104 104 106 104 104 107 105 106 105

Table 3 Species-specific and apparent second order rate constants for reaction of chlorine with phenolic compounds and flavone at pH 7. Compound

Species-specific second order rate constants (M
1 s
1) k1

Bisphenol A Catechol Flavone Gallic acid Hesperetin Hydroquinone Phenol Phloroglucinol p-Hydroxybenzoic acid Resorcinol Tannic acid

HOCl/PhOH

k2

HOCl=PhO




1.8

3.1  104

0.36 ± 0.28

2.2 ± 0.1  104

(3.7 ± 2.1) <330

(3.4 ± 0.1)  103 (1.4 ± 0.3)  106

k3

HOCl=PhðO
Þ2

6.6  104

(1.2 ± 0.1)  108

species-specific rate constants. It has to be noted that the presence of borate buffer at 8 < pH < 10 stabilizes o-hydroxyphenols (among other diols) in solution through the formation of stable complexes (Babcock and Pizer, 1980; Power and Woods, 1997). The formation of borate complexes has been corroborated spectrophotometrically for catechol, gallic and tannic acids and is shown in Fig. S8. The presence of borate leads to a shift of lmax to higher wavelengths and the absorption between 220 nm and 240 nm increases. The electrophilicity of the borate-polyhydroxy aromatic compound complexes (e.g., catechol) seems to be lower than expected for the free molecules. This is illustrated in Fig. S9 where the influence of borate on the catechol bromination rate is shown. In the pH range 8e10 where typically the maximum kapp would be expected, the presence of borate clearly leads to a much lower reactivity. In conclusion, for kinetic studies involving these types of compounds, the use of borate buffer should be carefully evaluated. Therefore, for catechol and gallic and tannic acids, only apparent second order rate constants at pH 7, determined in the presence of phosphate buffer, are provided in Tables 2 and 3. In the particular case of gallic acid, the chlorination showed two consecutive kinetic phases (2 mse100 ms; 100 mse1000 ms) for which the obtained apparent second order rate constants differ by about one order of magnitude (Fig. 2a and b). To further elucidate this, the kapp of the HOCl-gallic acid reaction at pH 7 was also determined under pseudo first-order conditions working in excess of the oxidant. From the results shown in Fig. 2c, a kapp value of 1600 M
1 s
1 was determined for the chlorination of gallic acid at pH 7 (raw data are presented in Fig. S5, SI). This value is clearly lower than kapp obtained by stopped-flow for the first phase (Fig. 2a, 1.17  104 M
1 s
1) and within a factor of two for the second phase (Fig. 2b, 815 M
1 s
1). Such a difference between the two obtained values is acceptable for a determination of rate constants with two different methods. The two observed phases

Apparent second order rate constant at pH 7 kapp (M
1 s
1)

References

62 217.5 ± 1.3 13.1 ± 1.0 1600 ± 48 (2.6 ± 0.1)  105 21.6 ± 0.3 18 ± 1 (1.2 ± 0.1)  106 15.9 ± 2.0 z4000 (1.9 ± 0.1)  104

(Gallard et al., 2004) this study this study this study this study this study (Gallard and von Gunten, 2002) this study this study (Rebenne et al., 1996) this study

probably correspond to the two consecutive electron transfer reactions from Clþ to Cl
and the formation of the intermediate semiquinone radical. In the case of gallic acid this radical could be stabilized by delocalization due to the three adjacent hydroxyl groups. In the corresponding experiments for bromine, no biphasic behaviour was observed (Fig. S7, SI). It can be hypothesized that in the case of chlorination, the lower overall reactivity allowed to resolve the two steps of oxidation; these two oxidation steps were not distinguishable by the time resolution of the stopped-flow instrumentation in the case of bromine. A determination of species-specific rate constants for the reaction of bromine or chlorine with natural organic matter is not possible because NOM has a continuum of pKa values. Westerhoff et al. (2004) studied the reactivity of chlorine and bromine with two different natural organic matter extracts and found a ratio of approximately 1e2 orders of magnitude between the two oxidants at pH 5. For the organic matter surrogate compound, tannic acid, a ratio 10 has been found between the apparent rate constants of chlorine and bromine. This is in the range of the previously reported ratio for NOM (Westerhoff et al., 2004). 3.2. Reactions of bromine with phenolic compounds and NOM: electrophilic aromatic substitution (EAS) versus oxidation (electron transfer (ET)) Bromination of phenol via EAS has been described previously (Acero et al., 2005). Alternatively, an electron transfer reaction can occur, whereby bromine is reduced to bromide and e.g., dihydroxybenzene compounds are oxidized to the corresponding quinones. The release of bromide can be used to characterize the type of reaction involved for a particular phenolic compound or NOM isolate. To examine the propensity of different phenolic compounds to EAS versus ET, bromination experiments were

J. Criquet et al. / Water Research 85 (2015) 476e486

Fig. 2. Kinetics of the gallic acid e HOCl reaction at pH 7.0. (a) Stopped-flow data at 500 nm in the time range 2 mse100 ms; (b) stopped-flow data at 500 nm in the time range 100 mse1000 ms; (c) measurement of the gallic acid decrease by HPLC under conditions of excess HOCl. Experimental conditions for stopped-flow experiments: (a), (b), [gallic acid] ¼ 2 mM; [HOCl] ¼ 0.2 mM; Black lines: Fitting curves (pseudo-first order kinetics); experimental conditions for direct measurement: (c) - [Gallic acid]0 1 uM and [HOCl]0 ¼ 50 uM; , [Gallic acid]0 1 uM and [HOCl]0 ¼ 25 uM; lines: Fitting curves (pseudo-first order kinetics).

performed in excess of phenolic compounds, flavone or NOM to avoid bromine reactions with the reaction products. In some cases, these products can be as reactive or even more reactive than the parent phenolic compounds due to the lower pKa of the halogenated phenols (Acero et al., 2005). The determination of the % bromine incorporation into the phenolic compound or NOM isolate has been discussed in the Materials and Methods section and the results of these experiments are shown for the selected phenolic compounds and NOM isolates in Fig. 3.

483

Varying propensities to EAS versus ET are shown across the range of phenolic compounds, flavone and NOM extracts. Phenol, phydroxybenzoic acid and bisphenol A undergo exclusively and flavone 90% EAS at pH 7 and 8. In contrast, for hydroquinone, catechol, gallic acid and tannic acid, a complete reduction of bromine to bromide is observed, indicating that these phenolic compounds are mainly oxidised. For phloroglucinol, the main pathway was also ET, with a contribution of about 75%. For all the other tested compounds, the fraction of released bromide was low, which indicates that the reaction of bromine occurs mainly by EAS. However, oxidation also occurred to some extent. Fig. 3 shows that the ratio between ET and EAS depends on the pH mainly for three compounds, i.e., hesperetin, 3,5-dihydroxybenzoic acid and resorcinol. The oxidation pathway was more important at pH 7. The common property of these three compounds is a meta-dihydroxy structure. From the species-specific rate constants of the reaction of bromine with resorcinol (Table 2), it can be calculated that the reaction of HOBr with Ph(OH)2 contributes 85% to the overall reaction at pH 7 (mainly Eqs. (1) and (2) at pH 7). The contribution of this reaction drops to 35% at pH 8. Since the relative proportion of EAS observed in Fig. 3 was higher at pH 8 than at pH 7, it can be hypothesized that HOBr reacts with the non-dissociated di-phenols partially by an ET, whereas HOBr reacts with the dissociated form mostly by EAS. This hypothesis is currently being further tested in our laboratory. The selected NOM extracts generally showed bromine incorporation (EAS) of about 20%, with the exception of Pony Lake fulvic acid (PLFA) for which EAS was <10%. Even though NOM is responsible for the formation of halogenated organic compounds as disinfection by-products (Rook, 1977), it is evident that the reaction of bromine with NOM moieties is dominated by electron transfer. These results allow to explore which fraction of the HOBr reacts with EAS or electron transfer, respectively, as the reactive DOM moieties are in excess. This gives an estimate on the ratio of the corresponding organic moieties reacting with either mechanism. Based on the results of the selected phenolic compounds, which are
, 2003), strucknown to be moieties of NOM (Leenheer and Croue tures such as hydroquinone, catechol and gallic acid could be responsible for the bromine consumption via electron transfer. These results are consistent with a previous study (Wenk et al., 2013), in which humic substances containing phenolic and hydroquinone moieties showed that their electron donating capacities were reduced by reaction with chlorine. This decrease was explained by the formation of quinone-type moieties from hydroquinone and/or catechol moieties (Wenk et al., 2013). Echigo and Minear (2006) stopped their reaction after 1 s and about 20% of HOBr consumption and they found about 80% of bromine incorporation into 3 NOM extracts. This could be an indication that the initial reactions proceed via EAS followed by ET processes. This hypothesis is in agreement with the difference of apparent second order rate constants at pH 7, which are one to two orders of magnitude higher for resorcinol than for the moieties reacting via ET (e.g. catechol, hydroquinone or gallic acid). Surprisingly, the bromine incorporation for NOM extracts showed almost no difference for the different types of extracts (Fig. 3), except for the PLFA sample. This difference could be attributed to the microbial source of PLFA, which is reported to have a proteinaceous character in high abundance and to be composed of more N- and S-containing moieties (D'Andrilli et al., 2013). These moieties are particularly reactive with bromine through an oxidation-reduction process (Heeb et al., 2014) and could act as competitors with the phenolic moieties towards reaction with bromine. This is further supported by the particularly low aromatic carbon content of PLFA (i.e., 12% compared to values between 20 and 60% for the other NOM extracts (Thorn et al., 1989) (see Section

484

J. Criquet et al. / Water Research 85 (2015) 476e486

Fig. 3. Electrophilic aromatic substitution (% Br-incorporation) for the reaction of bromine with phenolic compounds, flavone or natural organic matter extracts at pH 7 or 8. Bromide concentrations were determined after completion of the reaction. Experimental conditions: phenolic compounds and flavone 150 or 50 mM and HOBr 3 or 1 mM, respectively; NOM concentration equivalent to 3.6 mg C L
1 in presence of 1 mM HOBr (see Section 2.5 for details) (ND: No Data available; NI: No Br-Incorporation observed).

2.1)). Although a high disparity is noticed between the aromatic carbon content of the different NOM extracts, the % Brincorporation cannot be linked to this parameter. This parameter, as well as the SUVA, is related to the aromaticity and does not allow a distinction between the aromatic structures reacting via EAS or ET reactions. Nevertheless, based on our results, it can be hypothesized, that the ratio between NOM moieties reacting by EAS or ET remains fairly constant across various sources of NOM. Furthermore, these types of experiments could contribute to a better understanding of the relative reactivity of NOM and the subsequent formation of DBPs. 3.3. Oxidation reactions with HOBr/HOCl: formation of quinones The very low bromine incorporation observed in the reaction of bromine with o- and p-hydroxyphenols (hydroquinone, catechol and gallic and tannic acids), presented in Fig. 3, suggests an electron transfer process as the main reaction mechanism for these phenolic compounds. The UVeVis spectra of catechol and hydroquinone before and after reaction with bromine are shown in Fig. 4. Based on the spectra of o-quinone (Albarr
an et al., 2010) and p-benzoquinone matching well with the corresponding post-reaction spectra of catechol and hydroquinone, respectively (Fig. 4), it can be concluded that the major products of reaction of bromine with o- and p-hydroxyphenols are the corresponding quinones when a 1:1 molar ratio of oxidant and target compound are used. The reaction mechanisms for EAS and ET for polyhydroxy aromatic compounds are illustrated in Fig. 5. Once formed, quinones are unstable in solution due to dimerization/polymerization reactions and/or their reduction. The latter is more pronounced for p-quinone (e.g., 1 mol of p-benzoquinone oxidizes the DPD reagent giving a DPDþ signal equivalent to ~0.5 mol of bromine or chlorine) than for o-quinone. The same quinonic intermediates were spectrophotometrically identified during chlorination of catechol and hydroquinone (Figs. S10 and S11). However, the stability of o-quinone in chlorination processes is much smaller than in the presence of bromine. As shown

Fig. 4. Reaction of bromine with (a) catechol and (b) hydroquinone: Spectra before (solid lines) and after (dashed lines) the reaction. Experimental conditions: [hydroxyphenol] ¼ 33 mM; [HOBr] 33 mM; pH 7 phosphate buffer 0.01 M. Dotted lines:
n et al., 2010) and (b) p-benzo(a) spectra of o-quinone 33 mM (according to Albarra quinone 33 mM (experimentally determined). In (b), the dashed and dotted lines overlap.

Fig. 5. Mechanisms for reactions of bromine with polyhydroxyaromatic compounds.

J. Criquet et al. / Water Research 85 (2015) 476e486

in Fig. S10, for an initial molar ratio [HOCl]/[Catechol] ¼ 0.25, the UVeVis absorbance spectrum recorded after 5 min contact time almost matches the expected spectrum (sum of the spectra of the expected amount of catechol and o-quinone in solution according to a molar ratio 1:1 for the reaction of chlorine with catechol leading to o-quinone). For higher [HOCl]/[catechol] ratios, according to Fig. S10, the o-quinone concentration is lower than expected and the absorbance in the UV-C range increases. These results suggest that, under these conditions, the o-quinone formed competes with catechol for chlorine and/or chlorine accelerates the dimerization/polymerization rate of o-quinone. In the case of hydroquinone, the stoichiometry of the reaction between chlorine and hydroquinone was found to be 2:1 instead of 1:1. As shown in Fig. S11, a ratio of [HOCl]/[Hydroquinone] ¼ 1 leads to only 50% yield of p-benzoquinone (similar results have been obtained working with other initial ratios). The oxidation of gallic and tannic acids with chlorine and bromine led, in a first step, to the formation of pinkered intermediates (lmax 500 nm) that further evolved to yellow compounds (Fig. S12, change of the UVeVis spectrum during chlorination of tannic acid at pH 7 is shown), which is in agreement with a previous study (Hartzfeld et al., 2002). Similar to the pathways of catechol and hydroquinone, semiquinones/quinones are likely formed during the oxidation of gallic and tannic acids (Bors et al., 2000), which then undergo further transformation/polymerization reactions. 4. Conclusions A selection of 10 phenolic structures and flavone were tested for their reactivity with bromine and chlorine. It was confirmed that bromine is more reactive towards these moieties than chlorine (apparent rate constants at pH 7 from 104 to 107 and 101 to 106 for bromine and chlorine, respectively) and it was found that the ratio of the species-specific rate constants (bromine/chlorine) is about 3000. This allows an estimation of species-specific second order rate constants for the reaction of bromine with phenolic compounds based on the corresponding chlorine kinetics or vice versa, except for highly reactive compounds such as resorcinol. The higher reactivity of bromine species compared to chlorine induces an enhancement of the overall reactivity during chlorination in presence of bromide. The influence of bromide depends on pH, concentrations and also the mechanism involved. The reaction of phenolic compounds with bromine (and chlorine) can involve electrophilic aromatic substitution, incorporating halogen, or electron transfer (oxidation), potentially forming quinonic structures. The latter is favored for dihydroxybenzenes with hydroxy groups in ortho or para substitution patterns. In natural waters, a mix of the two types of reactions can be expected. The bromide concentration is a key factor for (i) ET processes where it induces an overall higher oxidation rate during chlorination due to a catalytic effect and for (ii) EAS for which the incorporation of bromine into the structure may lead to the formation of more harmful by-products. These results have many implications in terms of formation of disinfection by-products as phenolic compounds have been found to be precursors of DBP formation. Seven out of 8 NOM extracts tested exhibited approximately the same percentage of bromine incorporation (EAS, 20%), showing that the major fraction of the oxidant was consumed by ET. Among the reducing moieties present in NOM, the hydroquinone, catechol and gallic acid type structures could potentiality be precursors of emerging DBPs (e.g., haloquinones), which need to be further studied. The simple method presented here based on the bromide analysis after bromination as an estimate for the extent of ET vs EAS could provide information on the potential formation of DBPs from natural organic matter.

485

Acknowledgments The authors would like to acknowledge funding and support from the Australian Research Council (ARC LP100100285), Water Corporation (Western Australia), Curtin University, Eawag, EPFL and Water Research Australia. Dr Eva M. Rodríguez thanks the
n, Cultura y Deporte for the support Spanish Ministerio de Educacio
Castillejo” program (CAS12-00146). provided through the “Jose Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.watres.2015.08.051. References Acero, J.L., Piriou, P., von Gunten, U., 2005. Kinetics and mechanism of formation of bromophenols during drinking water chlorination: assessment of taste and odor development. Water Res. 39, 2979e2993.
n, G., Boggess, W., Rassolov, V., Schuler, R.H., 2010. Absorption spectrum, Albarra mass spectrometric properties, and electronic structure of 1,2-benzoquinone. J. Phys. Chem. A 114 (28), 7470e7478. Alegría, A.E., Lopez, M., Guevara, N., 1996. Thermodynamics of semiquinone disproportionation in aqueous buffer. J. Chem. Soc. Faraday Trans. 92 (24), 4965e4968. Arnold, W.A., Bolotin, J., von Gunten, U., Hofstetter, T.B., 2008. Evaluation of functional groups responsible for chloroform formation during water chlorination using compound specific isotope analysis. Environ. Sci. Technol. 42 (21), 7778e7785. Babcock, L., Pizer, R., 1980. Dynamics of boron acid complexation reactions. Formation of 1:1 boron acid-ligand complexes. Inorg. Chem. 19 (1), 56e61. Bichsel, Y., von Gunten, U., 2000. Formation of iodo-trihalomethanes during disinfection and oxidation of iodide containing waters. Environ. Sci. Technol. 34, 2784e2791. Bond, T., Henriet, O., Goslan, E.H., Parsons, S.A., Jefferson, B., 2009. Disinfection byproduct formation and fractionation behavior of natural organic matter surrogates. Environ. Sci. Technol. 43, 5982e5989. Bors, W., Michel, C., Stettmaier, K., 2000. Electron paramagnetic resonance studies of radical species of proanthocyanidins and gallate esters. Arch. Biochem. Biophys. 374 (2), 347e355. Boyce, S.D., Hornig, J.F., 1983. Reaction pathways of trihalomethane formation from the halogenation of dihydroxyaromatic model compounds for humic acid. Environ. Sci. Technol. 17, 202e211. Bull, R.J., Reckhow, D.A., Rotello, V., Bull, O.M., Kim, J., 2006. Use of Toxicological and Chemical Models to Prioritize DBP Research; Denver. Chang, E.E., Chiang, P.C., Chao, S.H., Lin, Y.L., 2006. Relashionship between chlorine consuption and chlorination by-products formation for model compounds. Chemosphere 64, 1196e1203. Chowdhury, S., Champagne, P., McLellan, P.J., 2010. Investigating effects of bromide ions on trihalomethanes and developing model for predicting bromodichloromethane in drinking water. Water Res. 44, 2349e2359. Criquet, J., Allard, S., Salhi, E., Joll, C., Heitz, A., von Gunten, U., 2012. Iodate and iodotrihalomethane formation during chlorination of iodide-containing waters: role of bromide. Environ. Sci. Technol. 46, 7350e7357. D'Andrilli, J., Foreman, C.M., Marshall, A.G., McKnight, D.M., 2013. Characterization of IHSS Pony Lake fulvic acid dissolved organic matter by electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry and fluorescence spectroscopy. Org. Geochem. 65 (0), 19e28. Deborde, M., von Gunten, U., 2008. Reactions of chlorine with inorganic and organic compounds during water treatment e kinetics and mechanisms: a critical review. Water Res. 42, 13e51. Echigo, S., Minear, R.A., 2006. Kinetics of the reaction of hypobromous acid and organic matters in water treatment processes. Water Sci. Technol. 53 (11), 235e243. Echigo, S., 2002. Kinetics and Speciation of Brominated Disinfection By-products during Ozonation (PhD Thesis). University of Illinois, Urbana-Champaign, Illinois.
, J.P., 2004. Chlorination of bisphenol A: kinetics and Gallard, H., Leclercq, A., Croue by-products formation. Chemosphere 56, 465e473.
, J.P., Legube, B., 2003. Rate constants of reactions of Gallard, H., Pellizzari, F., Croue bromine with phenols in aqueous solution. Water Res. 37, 2883e2892. Gallard, H., von Gunten, U., 2002. Chlorination of phenols: kinetics and formation of chloroform. Environ. Sci. Technol. 36, 884e890. Guo, S.C., Lin, F., 2009. The bromination kinetics of phenolic compounds in aqueous solution. J. Hazard. Mater. 170, 645e651. Hansch, C., Leo, A., Taft, R.W., 1991. A survey of hammett substituent constants and resonance and field parameters. Chem. Rev. 91 (2), 165e195. Hartzfeld, P.W., Forkner, R., Hunter, M.D., Hagerman, A.E., 2002. Determination of hydrolyzable tannins (gallotannins and ellagitannins) after reaction with potassium iodate. J. Agric. Food Chem. 50 (7), 1785e1790.

486

J. Criquet et al. / Water Research 85 (2015) 476e486

Heeb, M.B., Criquet, J., Zimmermann-Steffens, S.G., von Gunten, U., 2014. Oxidative treatment of bromide-containing waters: formation of bromine and its reactions with inorganic and organic compounds d A critical review. Water Res. 48 (0), 15e42. Heitz, A., Blythe, J., Allpike, B., Joll, C.A., Kagi, R.I., 2003. Plastic tastes in drinking water: factors affecting the chemistry of bromophenol formation. Water Sci. Technol. Water Supply 2, 179e184. Hrudey, S.E., 2009. Chlorination disinfection by-products, public health risk tradeoffs and me. Water Res. 43 (8), 2057e2092. Hrudey, S.E., Charrois, J.W.A., 2012. Disinfection By-products and Human Health. IWA Publishing, London. Hua, G., Reckhow, D.A., 2012. Evaluation of bromine substitution factors of DBPs during chlorination and chloramination. Water Res. 46 (13), 4208e4216. Ichihashi, K., Teranishi, K., Ichimura, A., 1999. Brominated-trihalomethane formation from phenolic derivatives as a model of humic materials by the reaction with hypochlorite and hypobromite ions. Chem. Lett. 9, 957e958. Kumar, K., Margerum, D.W., 1987. Kinetics and mechanism of general-acid-assisted oxidation of bromide by hypochlorite and hypochlorous acid. Inorg. Chem. 26, 2706e2711.
, J.P., 2003. Characterizing aquatic dissolved organic matter. Leenheer, J.A., Croue Environ. Sci. Technol. 37 (1), 18Ae26A. Mills, C.J., Bull, R.J., Cantor, K.P., Reif, J., Hrudey, S.E., Huston, P., 1998. Health risks of drinking water chlorination by-products: report of an expert working group. Chronic Dis. Can. 19, 91e102. ~ oz, F., von Sonntag, C., 2000. Determination of fast ozone reactions in aqueous Mun solution by competition kinetics. J. Chem. Soc. Perkin Trans. 2, 661e664. Narotsky, M.G., Klinefelter, G.R., Goldman, J.M., Best, D.S., McDonald, A., Strader, L.F., Suarez, J.D., Murr, A.S., Thillainadarajah, I., Hunter, E.S., Richardson, S.D., Speth, T.F., Miltner, R.J., Pressman, J.G., Teuschler, L.K., Rice, G.E., Moser, V.C., Luebke, R.W., Simmons, J.E., 2013. Comprehensive assessment of a chlorinated drinking water concentrate in a rat multigenerational reproductive toxicity study. Environ. Sci. Technol. 47 (18), 10653e10659. Norwood, D.L., Johnson, J.D., Christman, R.F., Hass, J.R., Bobenrieth, M.J., 1980. Reaction of chlorine with selected aromatic models of aquatic humic material. Environ. Sci. Technol. 14 (2), 187e190. Pan, Y., Zhang, X., 2013. Four groups of new aromatic halogenated disinfection byproducts: effect of bromide concentration on their formation and speciation in chlorinated drinking water. Environ. Sci. Technol. 47 (3), 1265e1273. Power, P.P., Woods, W.G., 1997. The chemistry of boron and its speciation in plants. Plant Soil 193 (1e2), 1e13. Rebenne, L.M., Gonzalez, A.C., Olson, T.M., 1996. Aqueous chlorination kinetics and mechanism of substituted dihydroxybenzenes. Environ. Sci. Technol. 30, 2235e2242. Richardson, S.D., Thruston Jr., A.D., Rav-Acha, C., Groisman, L., Popilevsky, I., Juraev, O., Glezer, V., McKague, A.B., Plewa, M.J., Wagner, E.D., 2003.

Tribromopyrrole, brominated acids, and other disinfection byproducts produced by disinfection of drinking water rich in bromide. Environ. Sci. Technol. 37 (17), 3782e3793. Richardson, S.D., Plewa, M.J., Wagner, E.D., Schoeny, R., DeMarini, D.M., 2007. Occurrence, genotoxicity, and carcinogenicity of regulated and emerging disinfection by-products in drinking water: a review and roadmap for research. Mutat. Res. Rev. Mutat. 636 (1e3), 178e242. Roccaro, P., Korshin, G.V., Cook, D., Chow, C.W.K., Drikas, M., 2014. Effects of pH on the speciation coefficients in models of bromide influence on the formation of trihalomethanes and haloacetic acids. Water Res. 62, 117e126. Rodier, J., Legube, B., Merlet, N., 2009. L'analyse de l'eau, 9 ed. Dunod, p. 1600. Rook, J.J., 1977. Chlorination reactions of fulvic acids in natural waters. Environ. Sci. Technol. 11 (5), 478e482. Rook, J.J., 1976. Haloforms in drinking water. J. Am. Water Works Assoc. 68 (3), 168e172. Salhi, E., von Gunten, U., 1999. Simultaneous determination of bromide, bromate and nitrite in low mg l
1 levels by ion chromatography without sample pretreatment. Water Res. 33, 3239e3244. Sedlak, D.L., von Gunten, U., 2011. The chlorine dilemma. Science 331 (6013), 42e43. Thorn, K.A., Folan, D.W., MacCarthy, P., 1989. Characterization of the International Humic Substances Society Standard and Reference Fulvic and Humic Acids by Solution State Carbon-13 (13C) and Hydrogen-1 (1H) Nuclear Magnetic Resonance Spectrometry.; Denver, CO, p. 93. Uchimiya, M., Stone, A.T., 2009. Reversible redox chemistry of quinones: impact on biogeochemical cycles. Chemosphere 77 (4), 451e458. von Gunten, U., Oliveras, Y., 1997. Kinetic of the reaction between hydrogen peroxide and hypobromous acid: implication on water treatment and natural systems. Water Res. 31, 900e906. Wang, W., Qian, Y., Li, J., Moe, B., Huang, R., Zhang, H., Hrudey, S.E., Li, X.F., 2014. Analytical and toxicity characterization of halo-hydroxyl-benzoquinones as stable halobenzoquinone disinfection byproducts in treated water. Anal. Chem. 86 (10), 4982e4988. Wenk, J., Aeschbacher, M., Salhi, E., Canonica, S., von Gunten, U., Sander, M., 2013. Chemical oxidation of dissolved organic matter by chlorine dioxide, chlorine, and ozone: effects on its optical and antioxidant properties. Environ. Sci. Technol. 47 (19), 11147e11156. Westerhoff, P., Chao, P., Mash, H., 2004. Reactivity of natural organic matter with aqueous chlorine and bromine. Water Res. 38 (6), 1502e1513. ~ as, B.J., Plewa, M.J., Yang, Y., Komaki, Y., Kimura, S.Y., Hu, H.Y., Wagner, E.D., Marin 2014. Toxic impact of bromide and iodide on drinking water disinfected with chlorine or chloramines. Environ. Sci. Technol. 48 (20), 12362e12369. Zhao, Y., Anichina, J., Lu, X., Bull, R.J., Krasner, S.W., Hrudey, S.E., Li, X.F., 2012. Occurrence and formation of chloro- and bromo-benzoquinones during drinking water disinfection. Water Res. 46 (14), 4351e4360.