Article pubs.acs.org/est
Halogenation of Bisphenol-A, Triclosan, and Phenols in Chlorinated Waters Containing Iodide Peter J. Vikesland,*,† E. Matthew Fiss,† Krista R. Wigginton,†,∥ Kristopher McNeill,‡,⊥ and William A. Arnold§ †
Via Department of Civil and Environmental Engineering, Virginia Tech, 415 Durham Hall, Blacksburg, Virginia 24061, United States Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, United States § Department of Civil Engineering, University of Minnesota, 500 Pillsbury Drive SE, Minneapolis, Minnesota 55455, United States ‡
S Supporting Information *
ABSTRACT: Free chlorine reacts with the iodide (I−) present in disinfected waters to produce free iodine. Past research has indicated that this nonchlorinated oxidant exhibits greater reactivity and potentially produces more toxic byproducts than free chlorine alone. In this study, we examined the reactivity of the phenolic compounds 2,4-dichlorophenol, triclosan, and bisphenol A in chlorinated waters containing I−. The free iodine mediated reactions were probed as a function of the initial I− concentration and the solution pH. Over the pH range of 5.5 to 10 for an initial I− concentration of 10 μM, the observed transformation kinetics of 2,4-dichlorophenol were generally 2−15× faster than reactions with free chlorine alone, while for triclosan and bisphenol A the free iodine mediated transformations were ≈3−20× and ≈230−660× faster, respectively. A comprehensive reaction model that simultaneously accounts for free chlorine and free iodine reactivity in these systems was developed to determine second-order rate constants for the chlorinated and iodinated oxidants. For all test compounds, iodinated daughter products are rapidly produced when free chlorine reacts in the presence of I−.
■
k1
INTRODUCTION Phenols are a ubiquitous class of environmental contaminant commonly detected in a range of environmental matrices.1,2 Because of the electron donating capacity of the hydroxyl group, phenols are especially susceptible to electrophilic substitution reactions with electrophiles (e.g., Cl+, Br+, I+).3−5 In water, these reactions result in ring substitution followed by cleavage3,4,6 and the formation of regulated and unregulated byproducts.6−10 Free chlorine (a term inclusive of HOCl, OCl−, Cl2, Cl2O, and H2OCl+)11 is often added to drinking water and to wastewater treatment plant effluents for disinfecting purposes. One issue that has been generally overlooked in the majority of the kinetic studies examining phenol reactivity in free chlorine solutions is the potential complicating effects of common electrolytes, such as I− and Br−, on these reactions. In natural waters, I− is found at concentrations ranging from 0.5 to 100 μg/L (0.004−0.8 μM) with a mean concentration of 20 μg/L (0.2 μM).12 In the presence of free chlorine, I− is oxidized to hypoiodous acid (HOI)4,13−16 by a complex series of diffusion-limited reactions involving the intermediates HOClI− and ICl.15−18 The overall irreversible reaction for I− oxidation is given by eq 1: © 2013 American Chemical Society
HOCl + I− → HOI + Cl−
k1 = 4.3 × 108 M−1s−1 −
−
(1) +
Once formed, HOI equilibrates to form I2, OI , I3 , H2OI , and HOI2− as described by eqs 2−6.16,19−21 I 2 + H 2O ⇄ HOI + I− + H+
K 2 = 5.44 × 10−13 M2 (2)
−
HOI ⇄ OI + H
I 2 + I− ⇄ I3−
+
−11
K3 = 2.3 × 10
M
K4 = 715 M−1
H 2OI+ ⇄ HOI + H+
HOI 2− ⇄ HOI + I−
K5 = 2 M
K 6 = 3.1 × 10−3 M
(3) (4) (5) (6)
Prior studies have shown that iodinated oxidants react with organic material to produce iodinated4,22 byproducts. At present, relatively little is known about the kinetics of Special Issue: Rene Schwarzenbach Tribute Received: Revised: Accepted: Published: 6764
December 3, 2012 March 1, 2013 March 7, 2013 March 7, 2013 dx.doi.org/10.1021/es304927j | Environ. Sci. Technol. 2013, 47, 6764−6772
Environmental Science & Technology
Article
iodination in chlorinated waters that also contain I−. Although past work has examined the effects of I− on electrophilic substitution in the presence of free chlorine, these studies focused on product distributions23−25 rather than on the determination of reaction rates. Published kinetic studies with I− and free chlorine focused on the disproportionation of HOI to produce iodate or the reactivity of solutions of I− in the presence of HOI and free chlorine.4,13,26−29 Iodinated disinfection byproducts (DBPs) have been found to be up to 54× more toxic than their chlorinated analogues22 thus supporting further study of iodination reactions. Additionally, the presence of I− during organic matter oxidation has been shown to produce, on a mass basis, more THMs and HAAs than are produced by free chlorine alone.14 Unfortunately, free iodine cannot be differentiated from free chlorine (eqs 7−105,11) by DPD-based methods (Supporting Information (SI) Figure S1) and thus the presence of iodinated species during water disinfection is often overlooked. HOCl + H+ ⇄ Cl+ + H 2O
K 7 = 130 M−1
HOCl + Cl− + H+ ⇄ Cl 2 + H 2O
The objective of this study was to evaluate how I− affects triclosan and BPA reactivity and daughter product formation in laboratory-prepared chlorinated waters. 2,4-Dichlorophenol (2,4-DCP) was also studied for comparative purposes. A comprehensive kinetic model was developed to describe the reaction kinetics in these solutions.
■
MATERIALS AND METHODS Deionized water (>18 MΩ-cm) was produced using an Aries water purification system. pH measurements were obtained using a Fisher Scientific model 60 pH meter and a ThermoOrion Ross PerpHect combination electrode. Laboratory glassware was cleaned by sequentially soaking in 10% nitric acid and then in concentrated chlorine (≈ 5.25%). BPA, phenol, 4-chlorophenol, 4-bromophenol, 4-iodophenol, 2,4DCP, and 2,4,6-trichlorophenol (2,4,6-TCP) were purchased from Alfa Aesar and used without further purification. Triclosan and the DPD-FAS reagents were purchased from Aldrich and were used without further purification. Solution Preparation. Stock phenolic solutions were prepared by dissolving 3−85 micromoles of test phenol in 50 mL of reagent grade methanol. Stocks of free chlorine were prepared using a commercial solution of sodium hypochlorite (purified grade 4−6% NaOCl; Fisher Scientific). Experimental solutions were prepared using reagent grade water containing 2 mM sodium bicarbonate buffer adjusted to the desired pH. Aliquots of pretitrated free chlorine stock were spiked into experimental solutions to obtain a desired concentration. After allowing 5 min for the solutions to equilibrate, each was adjusted to the desired pH using sodium hydroxide and sulfuric acid. Sulfuric acid was used to minimize chloride addition because chloride ions shift the solution equilibrium from HOCl toward Cl2/H2OCl+ at low pH.6 All experiments were performed at 25 °C. Total Oxidant Consumption. Short-duration oxidant consumption experiments were performed using a quench flow system (QFS) based on the design of Buffle et al.43 (see the SI for details). In a QFS experiment the free chlorine and I− solutions were contained in two separate syringes. These two solutions were then mixed for a period of 20 s prior to mixing with a phenol in a reaction flow loop. Using this system and a constant flow-rate, we achieved highly accurate, short reaction times (ranging from 0.25 to 10 s) by simply changing the volume of each reaction flow loop. Total oxidant concentrations (defined as the sum of the Cl2, H2OCl+, HOCl, OCl−, Cl2O, I2, H2OI+, HOI, HOI2−, and OI− concentrations for experiments in the presence of I−) were determined using a modified DPD-FAS colorimetric method44 developed for use with the QFS. Our definition for [Total Oxidant] is operational because the DPD-FAS method measures the combined concentrations of the oxidants and cannot differentiate between free chlorine and free iodine. With the QFS system, it was possible to have stable reactant solutions that were mixed during the course of an experiment for a defined period of time. In addition, because the syringes were headspace free it was possible to use a bicarbonate buffer under low pH conditions where outgassing of CO2 would normally be a challenge. To promote pseudo-first-order conditions, oxidant demand experiments were performed with phenol concentrations in excess relative to the oxidant concentration. Following a given reaction period (fixed by the flow loop volume), the final solution was mixed with DPD reagent and phosphate buffer to quench the reaction. The quenched solution then flowed
(7)
K8 = 1960 M−2 (8)
HOCl ⇄ OCl− + H+
K 9 = 2.9 × 10−8 M
2 HOCl ⇄ Cl 2O + H 2O
K10 = 8.7 × 10−3
(9) (10)
Triclosan (5-chloro-2-(2,4-dichlorophenoxy)phenol) is an antimicrobial agent found in many common household hygienic products at concentrations up to 0.3%. Although its effectiveness in the home is a topic of debate,30 triclosan is added to these products because it exhibits antibacterial properties. Incomplete removal during wastewater treatment and the land application of triclosan-laden biosolids have resulted in widespread environmental distribution of triclosan.31−37 Similar to triclosan, bisphenol A (BPA, 2,2-bis-(4hydroxy-phenyl)propane) is also a commonly detected phenolic contaminant of surface waters. BPA has been used in the manufacture of a range of consumer products, as a monomer in the synthesis of polycarbonates, and as a polymerization inhibitor in PVC pipes.38 As of 2007, the estimated U.S. production of BPA was ≈1 billion kg,39 while approximately 0.45 million kg of triclosan are produced annually. Both triclosan and BPA are commonly detected in wastewater effluents and in surface waters used for drinking water supply. In the latter matrix, concentrations as high as 58 μg/L for triclosan and 41 μg/L for BPA have been reported.40 Triclosan and BPA present in wastewater effluents and in drinking water supplies can come into contact with disinfectants during treatment, there is evidence of BPA leaching from plastics and epoxy resins used within water distribution systems,41 and many triclosan- and BPA-containing products currently on the market are designed to be used during activities such as showering and dish washing. Each of these activities results in contact with disinfectants. Studies have shown that both triclosan and BPA react quickly with free chlorine to produce chlorinated daughter products.1,6,7,9,10,42 Similarly, studies have shown that phenolic compounds are susceptible to electrophilic substitution by free iodine. No work to date, however, has examined the kinetics of triclosan or BPA halogenation in free chlorine solutions that also contain I−. 6765
dx.doi.org/10.1021/es304927j | Environ. Sci. Technol. 2013, 47, 6764−6772
Environmental Science & Technology
Article
through an additional flow loop with a fixed retention time of 1 min to allow color development. The absorbance at 515 nm was determined using a Cary5000 UV−vis-NIR spectrophotometer and a 1-cm path length flowcell. The absorbance of each solution was measured for each reaction time (fixed by the flow loop volume) until the absorbance signal stabilized for a minimum of 30 s (Figure S2). The total oxidant concentration was determined from the average absorbance over the 30-s collection period using a total oxidant calibration curve. All experiments were performed in at least duplicate to ensure data reproducibility. Results obtained from replicate experiments showed 2,4-DCP. This result can be explained by differences in the phenol ring substituents of each compound. 2,4-DCP contains two deactivating chloro substituents, in ortho and para position, that act to stabilize the phenolate form (PhO−) of 2,4-DCP and slow the phenolate’s reaction toward electrophiles.53 Triclosan, in contrast, has one meta-substituted chlorine substituent and an ortho-substituted aryloxy (−O-C6H3Cl2) group while BPA is activated by paraalkyl substitution. Both triclosan and BPA are more electron-
d[free chlorine] = −k Cl 2 ,PhOH[Cl 2][PhOH] dt − kHOCl,PhOH[HOCl][PhOH] − kHOCl,PhO−[HOCl][PhO−]
(15)
For a diphenol such as BPA the model takes a slightly different form: 6767
dx.doi.org/10.1021/es304927j | Environ. Sci. Technol. 2013, 47, 6764−6772
Environmental Science & Technology
Article
Table 1. Model-Determined Rate Coefficients for Free Chlorine and Free Iodine Reactivity with the Listed Phenolic Compounds phenol
pKaa
2,4-DCP
7.85
triclosan
7.9
kCl2PhOH (M−1 s−1)
kHOCl,PhOH (M−1 s−1) 23 (±13) 373
30 (±17)b
kHOCl,PhOH2 (M−1 s−1)
kCl2PhOH (M−1 s−1) BPA
9.59 and 11.3
47 (±25)
kHOCl,PhO− (M−1 s−1)
kHOCl,PhOH− (M−1 s−1)
kI2,PhOH (M−1 s−1)
kHOCl,PhO2− (M−1 s−1)
660 (±130) 5.4 × 103c
240 (±73) 4.0 (±0.7) × 103 kI2,PhOH− (M−1 s−1)
kHOCl,PhOH− (M−1 s−1)
kHOCl,PhO2− (M−1 s−1)
kI2, PhOH2 (M−1 s−1)
3.2 (±0.6) × 104
5.5 (±0.7) × 104
4.0 (±1.8) × 102
kI2,PhO− (M−1 s−1) 1.0 (±0.26) × 104 1.1 (±0.1) × 105 kHOI,PhOH2 (M−1 s−1)
3.1 (±0.4) × 107
2.4 (±0.2) × 105
kHOI,PhO− (M−1 s−1) 650 (±150) 1.1 (±0.6) × 103 kHOI,PhO2− (M−1 s−1) 1.4 (±0.5) × 107
a
pKa values for all phenols were reported in reference 5 and the references therein. bError limits are 95% confidence intervals reported by Micromath Scientist. cExperimentally determined in Rule et al.6
constitutes less than 10% of the total iodinated oxidant for pH values below 9.75. Importantly, under these conditions, free chlorine accounts for up to 50% of the total oxidant concentration when the pH is ≤8. For this reason, the free iodine mediated reaction kinetics could only be evaluated while simultaneously considering the free chlorine reactions. In our model, the rate constants for the free-chlorine-mediated reactions were fixed at the values determined in the absence of I−. In natural waters there will almost always be an excess of free chlorine and thus reactions involving free iodine are always intimately linked to those involving free chlorine. Similar to what was done for the solutions containing only free chlorine, we developed an expression for free iodine loss that initially incorporated reactions between each of the known iodinated reactants (I2, HOI, H2OI+, OI−, HOI2−, I3−) with both the phenolate and the neutral forms of the phenols. Using experimental data collected with 10 μM of I− present and the rate constants for the free chlorine mediated reactions determined in the absence of I−, this general expression was systematically evaluated to determine the rate-determining reactions over the pH range between pH 5 and 10. The resulting model was then used in a predictive mode to model the reaction rates when only 4 μM of I− was present. Based upon this modeling exercise, the change in the free iodine concentration for the monophenols can be described as follows:
d[free chlorine] = −kHOCl,PhOH 2[HOCl][PhOH 2] dt − kHOCl,PhOH−[HOCl][PhOH−] − kHOCl,PhO2−[HOCl][PhO2 −]
(16)
Model determined rate constants for each tested phenol (2,4DCP, triclosan, and BPA) are tabulated in Table 1. Considering our definition of kobs we can write for the monophenols: kobs,chlorine = (k Cl 2,PhOHfCl fPhOH + KHOCl,PhOHfHOCl fPhOH 2
+ kHOCl,PhO−fHOCl fPhO− )[Total Phenol]
(17)
For BPA: kobs,chlorine = (KHOCl,PhOH2fHOCl fPhOH + kHOCl,PhOH− f HOCl 2
fPhOH− + kHOCl,PhO2−fHOCl fPhO2− ) [Total Phenol]
(18)
with f PhOH, f PhO−, f PhOH2, f PhOH−, and f PhO2− representing the fractions of phenol in a given form and f Cl2 and f HOCl representing the fractions of free chlorine as HOCl and Cl2, respectively. As shown in Figures 1 and 2, kobs values calculated using our rate constants and eqs 17 and 18 generally predict free chlorine loss in the presence of the three different phenols. The model predictions are most accurate in the circumneutral pH range, but tend to overpredict free chlorine loss at low pH, while slightly underpredicting at high pH. This exercise shows that for our reaction conditions that neither H2OCl+ nor Cl2O has a measurable effect on the observed reaction rate. This result contradicts previous models developed for these systems,3,55,56 but corroborates experimental evidence showing that Cl2 is an active reactant at low pH.58 It is possible that if the pH range were extended below 5.0 that the relative importance of these additional species could become apparent. Although significant amounts of OCl− are present in basic solution the model consistently overestimates the kinetics of oxidant loss if reactions involving OCl− are incorporated. This observation is in agreement with previous studies examining free chlorine reactivity with phenols.3 Free Iodine−Phenol Reactivity. When I− is present, the solution chemistry complexity increases considerably. For a solution initially containing 10 μM free chlorine and 10 μM of I−, I2 is the most prevalent iodinated species below pH 7.25, while HOI predominates at higher pH (Figure S3). OI−
d[free iodine] = k1[HOCl][I−] − k I2,PhOH[I 2][PhOH] dt − (k I2,PhO−[I 2] + kHOI,PhO−[HOI])[PhO−] (19)
And for BPA: d[free iodine] = k1[HOCl][I−] − k I2,PhOH2[I 2][PhOH 2] dt − k I2,PhOH−[I 2][PhOH−] − kHOI,PhOH2[HOI][PhOH 2] − kHOI,PhO2−[HOI][PhO2−]
(20)
−
where k1 is the rate constant for I oxidation by HOCl (eq 1) and the remaining rate constants are tabulated in Table 1 for the different phenols. In our model we assume that I+ reacts with a phenol to produce iodinated products. Electrophilic substitution reactions involving free iodine species containing multiple iodine atoms (i.e., I2, HOI2−, and I3−), however, also produce I− that can subsequently result in the regeneration of HOI by eq 1. 6768
dx.doi.org/10.1021/es304927j | Environ. Sci. Technol. 2013, 47, 6764−6772
Environmental Science & Technology
Article
Figure 3. Summary of phenolic products produced when free chlorine reacts with 2,4-DCP, triclosan, or BPA in the presence of iodide under conditions with an excess of free chlorine relative to the initial iodide concentration. The arrows indicate a substitution reaction involving free chlorine or free iodine. The variables X, Y, Z, and Z′ define a substitution reaction product that contains chlorine or iodine. All products included in this schematic were detected via GC-MS.
Accordingly, our kinetic model also accounts for the fate of I−. For the monophenols:
where f Cl/TOTOX and f I/TOTOX are the fractions of total oxidant present either as free chlorine or free iodine, respectively. As shown in Figure S4, under our experimental conditions it was necessary to simultaneously consider the reactivity of both free chlorine and free iodine to accurately predict the overall reaction rate. At pH 7 for a 2,4-DCP solution with an initial free chlorine concentration of 10 μM and [I−]0 = 10 μM, if only free iodine was considered the overall reaction rate was underpredicted by 11%. Conversely, if only free chlorine was considered, the reaction rate was underpredicted by 89%. Such an outcome can be expected under reaction conditions that more closely resemble those employed in water treatment where the free I− concentration will generally be much lower than the free chlorine concentration. Under such conditions, only a model simultaneously accounting for free chlorine and free iodine reactivity can provide accurate predictions. For a majority of the test conditions, the comprehensive model defined by eqs 15−25 predicts the kinetic results (Figure 1) and the calculated kobs values reflect the pH trends in the experimental data (Figure 2). The general predictive capacity of the model over the range of conditions supports the validity of the fitted rate constants and illustrates the ability of the model to capture the speciation of free iodine in these solutions. In only one case, with BPA and 4 μM I−, was the model prediction less than satisfactory. Under those conditions the model can predict the magnitude of kobs but does not fully reflect how changes in pH alter the reaction kinetics. To illustrate the predictive capacity of this model, additional experiments were performed under excess oxidant conditions to allow prediction of phenol loss. Simply by using the initial phenol, free chlorine, and I− concentrations as inputs and
d[I−] = −k1[HOCl][I−] + k I2,PhOH[I 2][PhOH] dt − k I2,PhO−[I 2][PhO−]
(21)
and for BPA: d[I−] = −k1[HOCl][I−] + k I2,PhOH2[I 2][PhOH 2] dt − k I2,PhOH−[I 2][PhOH−]
(22)
Following the arguments described above, we can define the first-order rate constants for the free iodine mediated reactions of the monophenols as follows: k iodine,obs = ((k I2,PhOHfPhOH + k I2,PhO−(1 − fPhO− ))fI
2
+ kHOI,PhO−fHOI (1 − fPhOH ))[Total phenol] (23)
and for BPA: k iodine,obs = ((k I2,PhOH2fPhOH + k I2,PhOH−(1 − fPhOH− ))fI 2
2
+ (kHOI,PhOH2fPhOH + kHOI,PhO2−fPhO2− )fHOI ) 2
[Total phenol]
(24)
For conditions where both free chlorine and free iodine are present the overall first-order rate constant is: kobs,overall = kchlorine,obsfCl/TOTOX + k iodine,obsfI/TOTOX
(25) 6769
dx.doi.org/10.1021/es304927j | Environ. Sci. Technol. 2013, 47, 6764−6772
Environmental Science & Technology
Article
Triclosan. Iodinated intermediates and products form when triclosan and free chlorine react in the presence of I− (Figure 3). These species form in parallel to the chlorinated intermediates and products described elsewhere.6,9,10 In the presence of I −, monoiodinated (5-chloro-4-iodo-2-(2,4dichlorophenoxy)phenol; 5-chloro-6-iodo-2-(2,4dichlorophenoxy)phenol), diiodinated (5-chloro-4,6-diiodo-5− 2-(2,4-dichlorophenoxy)phenol), and monochlorinated/di-iodinated (either 4,5-dichloro-6-iodo-2-(2,4-dichlorophenoxy)phenol or 5,6-dichloro-4-iodo-2-(2,4-dichlorophenoxy)phenol; these two possible congeners cannot be distinguished by GCMS) forms of triclosan were detected. 2,4-DCP and 2,4,6-TCP were two triclosan daughter products detected in the presence of I− (Figures S6−S7). This result is consistent with our previous hypotheses that 2,4DCP is produced solely via scission of the ether bond of triclosan and not by substitution of its phenoxy ring, and that 2,4,6-TCP is produced by subsequent chlorination of 2,4-DCP. Once produced, 2,4-DCP reacts with free iodine to produce 2,4-dichloro-6-iodophenol. No iodinated THMs were detected at pH 7. BPA. When BPA and free chlorine react in the presence of I−, iodinated intermediates and products are detected (Figure 3). These species form in parallel to the chlorinated intermediates and products described elsewhere60 and in the SI. Their formation and reactivity are functions of the halide concentration (Figure S9). In the presence of I−, a monoiodinated (2iodo-BPA) and a monoiodinated/monochlorinated product (either 2-chloro-2′-iodo-BPA or 2-chloro-6-iodo-BPA; again these products could not be distinguished by GC-MS) of BPA were detected. Detection of these iodinated products suggests stepwise addition of iodine and chlorine to one of the phenol rings of BPA. Environmental Implications. This study has demonstrated that I− rapidly reacts with free chlorine to produce free iodine. This highly reactive species has the capacity to react with organic material to produce a variety of iodinated products. Unfortunately, because free iodine cannot be differentiated from free chlorine using DPD-based tests its presence in chlorinated waters is often undetected unless the reaction products are quantified and iodinated products are observed. The mechanistic model developed herein helps clarify the importance of free iodine in these systems. To illustrate the predictive value of the model, a simulation was run for both triclosan and BPA under conditions typical of natural waters. Assuming 1 μM of I− is present in pH 7 water that also contains 28.2 μM (= 2 mg/L) free chlorine and 0.0003 μM (= 100 ng/L) of triclosan, the half-life of triclosan decreases from 45 s when I− is absent to 40 s when I− is present. This change in kinetics is insignificant, however, we have shown herein that iodinated daughter products rapidly form instead of chlorinated products when I− is present and thus, although the apparent reaction kinetics have not changed, the product speciation has. When a similar simulation is performed for 0.0004 μM (= 100 ng/L) of BPA using an initial I− concentration of 0.2 μM (the mean concentration found in natural waters), a very different result occurs: the half-life of BPA in the absence of I− was 250 s, but when I− was present, the half-life was only 13 s. Such a change in reaction kinetics would be readily apparent and thus the change in product speciation would be anticipated. The reason for the difference in the relative effects of I− is shown in Figure 2addition of I− to the highly reactive BPA system has a much larger effect on the reaction kinetics than for the
assuming a 1:1 stoichiometry for the rate limiting substitution reaction,6 the model was able to accurately predict phenol loss both in the presence and absence of iodine (Figures S5 and S6). Furthermore, as discussed in the SI, our comprehensive model was readily extended to characterize the reactivity of phenol, 4-bromophenol, 4-chlorophenol, and 4-iodophenol. For the sake of clarity we emphasize that we did not isolate any of the individual reactants invoked in the models described by eqs 15−25. As such, our model does not consider some potential interhalogen intermediates (e.g., HOClI− and ICl) that could be important reactants. Given the current analytical challenges associated with differentiation of free iodine and free chlorine, our model represents the current state of understanding of these complicated reactive systems. As new data is acquired, we expect that this model will be refined and updated. Reaction Products. It is well established that when phenols react with halogenating agents, a number of halogenated and nonhalogenated byproducts are produced.3−5 In this effort we examined the formation of chlorinated and iodinated byproducts under excess total oxidant conditions (i.e., [Total oxidant] > [Total phenol]) in the presence and absence of 10 μM I−. We note that this concentration of I− was used to facilitate daughter product detection. Under natural conditions the relative concentrations of iodinated and chlorinated products would be expected to differ from those determined here. Details about product identification and quantification may be found in the SI; herein we focus upon the variety of products formed and overall product formation mechanisms. 2,4-DCP. Past studies have established that 2,4-DCP undergoes electrophilic substitution in the presence of free chlorine and free iodine.59 In the absence of I− the primary product is 2,4,6-TCP, while in its presence both 2,4,6-TCP and 2,4-dichloro-6-iodophenol are produced (Figure 3). The measured concentration of 2,4-dichloro-6-iodophenol after a 5-min contact period was a function of the initial I− concentration with lower levels produced for an initial I− concentration of 4 μM relative to 10 μM. In kinetic experiments at pH 7, 2,4,6-TCP was found to build up over time in the absence of I− (Figure S5). 2,4,6-TCP reacts with free chlorine, however, it does so at a rate considerably slower than the rate at which 2,4-DCP reacts, thus resulting in buildup of 2,4,6-TCP over short periods of time.6 Neither tetrachlorophenol nor pentachlorophenol were detected thus indicating that 2,4,6-TCP undergoes ring cleavage, as previously shown,6 to produce chloroform and other THMs. In the presence of 10 μM of I−,
