Role of halogen(I) cation-transfer mechanisms in water chlorination in the presence of bromide ion

Abstract

Bromide ion is rapidly converted to HOBr via BrCl by reaction with HOCl. The subsequent slow reactions of (HOCl, OCl⁻)/(HOBr, OBr⁻) mixtures are monitored directly by multiwavelength UV–vis absorbance methods and simultaneously by ion chromatographic measurement of ClO₂⁻, ClO₃⁻, and BrO₃⁻ (p[H⁺] 5.6–7.6). A first-order loss of HOCl is observed which is catalyzed by trace concentrations of Br⁻ and BrCl. Chlorite ion forms first and is subsequently oxidized to ClO₃⁻. The loss of HOBr is slower and is second-order in HOBr, so that BrO₃⁻ formation takes longer than ClO₃⁻ formation. Under the conditions of this work, the relative yield of BrO₃⁻ increases with increase in pH. The decomposition of HOCl by bromide proceeds primarily by a series of halogen(I) cation-transfer reactions with subsequent halide release. The presence of HOCl increases the BrO₃⁻ yield three-fold from HOBr decay alone.

---

Introduction

Chlorine (typically 0.2–40 mg L⁻¹)¹ is one of the most common disinfectants used by water utilities.² Aqueous chlorine hydrolyzes rapidly to HOCl and Cl⁻ at pH 7 with a half-life of 0.1 s.³ The decay rate of HOCl, the main disinfectant in the absence of ammonia, is very slow,⁴ and sufficient amounts remain for effective disinfection. However, the decay of HOCl is accelerated by the presence of bromide ion. In natural waters, the bromide concentration varies from 0.01–3.0 mg L⁻¹ in freshwater to 67 mg L⁻¹ in seawater.⁵ HOCl and Cl₂ react quickly with Br⁻ to form BrCl, which rapidly hydrolyzes to HOBr and Cl⁻.^6,7 Even when HOCl is present in excess of total bromine, (where [Br]_T = [HOBr] + [OBr⁻] + [BrCl] + [Br⁻]), an additional acceleration of the decay rate of HOCl is observed. The acceleration has been attributed to HOBr catalysis in previous work.^8–10 Although the rates of decomposition of HOCl and HOBr are accelerated when both hypohalous acids are present, the rate dependence and reaction mechanism are not known. The present work shows that the HOCl decomposition is catalyzed by bromine species other than HOBr and proceeds by a series of halogen(I) cation-transfer reactions with subsequent halide ion release. The products of the decomposition of hypohalous acids [eqn. (1)] are chlorate ion (ClO₃⁻) and bromate ion (BrO₃⁻).BrO₃⁻ is a known carcinogen.¹¹ For this reason, EPA has promulgated a 10 µg L⁻¹ maximum contaminant level (MCL) for BrO₃⁻ in drinking water.¹² The reaction mechanism for the decomposition of HOCl in the presence of Br⁻ is important in order to understand the fate of disinfection byproducts in the environment. The kinetic information should be useful in regard to disinfection efficacy and factors that influence BrO₃⁻ formation.

Experimental

Reagents and materials

All solutions were prepared using distilled, deionized (18 MΩ cm) water. NaOCl was synthesized by bubbling chlorine gas (Matheson, UHP) through continuously stirred 0.30 M sodium hydroxide solution held at 4 °C. NaOCl solution (pH > 12) was standardized spectrophotometrically using the molar absorptivity of OCl⁻ as 362 M⁻¹ cm⁻¹ at 292 nm.¹³

Chromatographic standards were made from sodium salts (analytical reagent grade, dried at 120 °C for 2 h), preserved with ethylenediamine,¹⁴ and used within four weeks of preparation. NaClO₂ was purified and ClO₂⁻ standards were standardized spectrophotometrically using the molar absorptivity of ClO₂⁻ as 154 M⁻¹ cm⁻¹ at 260 nm.¹³

Methods and instrumentation

A ROSS^® combination electrode was calibrated by titration of standardized perchloric acid and sodium hydroxide in media with a total ionic strength (μ_T) = 1.0 M (Na₂SO₄), to permit p[H⁺] to be calculated from the measured pH.

UV–vis spectra and kinetic scans were obtained on a PerkinElmer Lambda-9 UV–vis–NIR spectrophotometer. Ion chromatographic (IC) data were collected with a Dionex DX-500 HPLC. The separation method is an adaptation of EPA Method 300.1.¹⁴ Samples were injected via an autosampler (AS40) through a 25 µL injection loop onto quaternary amine anion exchange guard (AG9HC) and separation (AS9HC) columns. The analytes were eluted with 9.0 mM Na₂CO₃ at a 1.0 mL min⁻¹ flow rate. Suppressed conductivity detection (ED40), with an ASRS-Ultra suppressor in the gas-assisted self-regeneration mode and a current of 100 mA, was used to detect the analytes.

To initiate the reaction, sodium bromide, in phosphate buffer ([PO₄]_T = [H₂PO₄⁻] + [HPO₄²⁻]), was added to a sodium hypochlorite solution. The reaction mixture was protected from light, stirred, and thermostatted at 25.0(2) °C or allowed to react at room temperature (22 °C) where indicated. For analysis, aliquots were removed from the reaction mixture and quenched at timed intervals. To determine [HOCl]_T and [HOBr]_T (where [HOX]_T = [HOX] + [OX⁻], X = Cl, Br), a NaOH quench was applied to give pH > 12 so that all HOBr and HOCl were converted to OBr⁻ and OCl⁻, respectively. These samples were analyzed by UV–vis absorbance spectrophotometry with multiwavelength analysis. To calculate [HOCl]_T and [HOBr]_T, eqn. (2) was applied, where A_j = absorbance at wavelength j, ε(Cl)_j and ε(Br)_j are the molar absorptivities of OCl⁻ and OBr⁻, respectively, at wavelength j, and j ranges from 280 to 340 nm (1 nm intervals).

To determine the products by IC, aliquots of the [HOCl]_T/[Br]_T reaction mixture were quenched by addition to sulfite in basic solution (10.5 < pH < 12). Sulfite reacts rapidly with OCl⁻ and OBr^−15,16 and prevents column oxidation by these hypohalites. The concentrations of ClO₃⁻ and BrO₃⁻ were determined directly. ClO₂⁻ reacts slowly with sulfite,¹⁷ but the peak areas are accurate if the sample is injected immediately after the quench. The total chloride is the sum of Cl⁻ formed in the [HOCl]_T/[Br]_T reaction, the Cl⁻ present in the OCl⁻ stock solution, and the Cl⁻ released from the reaction of [HOCl]_T with SO₃²⁻. The total bromide is the sum of the Br⁻ formed in the [HOCl]_T/[Br]_T reaction, the Br⁻ released from the reaction of BrO₂⁻ with SO₃²⁻, and the Br⁻ released from the reaction of [HOBr]_T with SO₃²⁻.

Results and discussion

Kinetic species distribution

The kinetic species distribution for the decomposition of [HOCl]_T in the presence of [Br]_T is shown in Fig. 1. The [HOCl]_T decay and ClO₃⁻ formation both fit a first-order rate expression [eqn. (3)], where n is a stoichiometric factor.The following integrated rate expressions [eqns. (4) and (5)] may be derived from the differential equation, where t = time, i = initial, and f = final.The observed rate constants (k_obs^HOCl) are calculated from the integrated rate expressions, where k_obs^HOCl = (3.63 ± 0.04) × 10⁻⁵ s⁻¹ for [HOCl]_T loss and k^obsHOCl = (4.06 ± 0.08) × 10⁻⁵ s⁻¹ for ClO₃⁻ formation at 25.0 °C and μ_T = 1.0 M (Na₂SO₄). The precision given in the rate constants reflect the error in the curve fit. The agreement of the rate constants is about 10%, which is good considering two different methods were used to measure the kinetics.

Fig. 1 a,b Kinetic species distribution for [HOCl]_T/[Br]_T reaction. Reaction conditions: 5.03 mM [HOCl]_T, 0.508 mM [Br]_T, 0.36 M [PO₄]_T, p[H⁺] 7.00, μ_T = 1.0 M (Na₂SO₄), 22 °C. [HOCl]_T, (□); [ClO₃⁻] (▲); [BrO₃⁻] (○); [HOBr]_T (△). [HOCl]_T and [ClO₃⁻] fit a first-order decay and formation, respectively. [HOBr]_T and [BrO₃⁻] fit a second-order decay and formation, respectively. Fig. 1b is a detailed view of changes in [HOBr]_T and [BrO₃⁻] with time.

In contrast, the [HOBr]_T decay and bromate formation fit second-order kinetics [eqn. (6)].

The observed rate constants (k_obs^HOBr) are measured from the integrated rate expressions [eqns. (7) and (8)] and for the conditions given in Fig. 1, k_obs^HOBr = (1.77 ± 0.04) × 10⁻² M⁻¹ s⁻¹ for [HOBr]_T loss and k_obs^HOBr = (1.7 ± 0.1) × 10⁻² M⁻¹ s⁻¹ for BrO₃⁻ formation.The ratio of [HOBr]_T reacted to [BrO₃⁻] formed is one while [HOCl]_T is present.

The kinetic traces for total chloride, total bromide, and chlorite are measured but not shown in Fig. 1 for clarity. The initial total chloride measured by IC agrees with the expected value. Also, the total bromide measured as a function of time agrees with [HOBr]_T within the error of the analysis, which indicates that very little Br⁻ or BrO₂⁻ builds up when HOCl is present. Chlorite ion is detected at the micromolar level near the beginning of the reaction and then decays.

Observed [HOCl]_T decay rate constant dependence on [Br]_T

The rate expression for the decay of [HOCl]_T is first order in [HOCl]_T and is not directly dependent on [HOBr]_T, which also decays as shown in Fig. 1. Nevertheless, the observed first-order rate constants, k_obs^HOCl, do depend on [Br]_T. Although [HOBr]_T approximately equals [Br]_T, minor species, such as Br⁻ and BrCl, are believed to be important in the catalysis. The observed first-order rate constants for [HOCl]_T decay were measured as a function of [Br]_T (0.36 M [PO₄]_T, p[H⁺] 7.00, [HOCl]_T/[Br]_T ≥ 4.0). The k_obs^HOCl values increase as [Br]_T increases as seen in Fig. 2. The data are fit to a quadratic equation: k_obs^HOCl = y_o + a[Br]_T + b[Br]_T². The results of the fit are a = (6.5 ± 0.3) × 10⁻² M⁻¹ s⁻¹ and b = 24 ± 1 M⁻² s⁻¹. The y-intercept, y_o, is the [HOCl]_T decay rate constant in the absence of [Br]_T and is negligible with respect to k_obs^HOCl. The [HOCl]_T decay rate constant in the absence of [Br]_T has a third-order dependence on [HOCl]_T and can be calculated [eqns. (9)–(11), K_a^HOCl = 10^−7.5 M].⁴

Fig. 2 Dependence of observed first-order rate constant for [HOCl]_T decay on [Br]_T, where [Br]_T = [HOBr] + [OBr⁻] + [BrCl] + [Br⁻]. Reaction conditions: 0.36 M [PO₄]_T, p[H⁺] 7.00, μ_T = 1.0 M (Na₂SO₄), 22 °C, [HOCl]_T/[Br]_T > 4.0. Best fit line to quadratic equation: k_obs^HOCl = y_o + a[Br]_T + b[Br]_T², where a = (6.5 ± 0.3) × 10⁻² M⁻¹ s⁻¹ and b = 24 ± 1 M⁻² s⁻¹. y_o is calculated from eqn. (11) and equals 2.7 × 10⁻⁷ s⁻¹ at [HOCl]_T,i = 7.2 mM. Error bars represent error in k_obs^HOCl.

[Cl]_T dependence of observed second-order rate constant for [HOBr]_T decay and BrO₃⁻ yield

The observed rate of decay of HOBr in Fig. 1 is second order in [HOBr]_T and is not affected by the large changes in HOCl concentration during its decay. Nevertheless, the observed values for these second-order rate constants (k_obs^HOBr, M⁻¹ s⁻¹) do depend on the initial concentrations of total chlorine ([Cl]_T = [HOCl] + [OCl⁻] + [BrCl] + [Cl⁻], approximately 50% [HOCl]_T, 50% [Cl⁻]) as shown in Fig. 3 (0.36 M [PO₄]_T, p[H⁺] 7.04, [HOBr]_T,i = 0.504 mM). The observed rate constant for [HOBr]_T loss has a squared dependence in [Cl]_T and the data are fit to the relationship (k_obs^HOBr = k_HOBr + b[HOCl]_T²) where k_HOBr = (1.2 ± 0.3) × 10⁻² M⁻¹ s⁻¹ and b = (1.06 ± 0.09) × 10² M⁻³ s⁻¹. The curve-fitted y-intercept value (k_HOBr) is in agreement with the expected second-order rate constant for HOBr disproportionation (k_disp = 0.015 M⁻¹ s⁻¹) in the absence of HOCl [eqns. (12) and (13), where k_1a = 2 × 10⁻³ M⁻¹ s⁻¹, k_B^P = 0.05 M⁻² s⁻¹, k_B^OH = 15 M⁻² s⁻¹, k_1b = 6 × 10⁻⁷ M⁻¹ s⁻¹, K_a^HOBr = 10^−8.59 M, and n  = 1].¹⁸The BrO₃⁻ yield in Fig. 3 is the percentage of initial [HOBr]_T converted to BrO₃⁻ after 68 h total reaction time. The BrO₃⁻ yield increases with increasing initial [Cl]_T.

Fig. 3 Dependence of observed second-order rate constant for [HOBr]_T loss (○) and %BrO₃⁻ yield (×) on initial [Cl]_T, where [Cl]_T = [HOCl] + [OCl⁻] + [BrCl] + [Cl⁻]. k_disp at zero [Cl]_T,i (△) calculated by eqn. (13). %BrO₃⁻ yield at zero [Cl]_T,i (□) calculated by eqn. (1). Best fit line to [HOBr]_T rate constant is curve fit to the following equation: k_obs^HOBr = k_HOBr + b[Cl]_T², where k_HOBr = (1.2 ± 0.3) × 10⁻² M⁻¹ s⁻¹ and b = (1.06 ± 0.09) × 10² M⁻³ s⁻¹. %BrO₃⁻ = [BrO₃⁻]/[HOBr]_T,i after 68 h total reaction time. Reaction conditions: 0.504 mM [HOBr]_T, 0.36 M [PO₄]_T, p[H⁺] 7.04(2), μ_T = 1.0 M (Na₂SO₄), 25.0(2) °C. Dashed line shows trend in data. Error bars represent error in k_obs^HOBr.

pH dependence of [HOCl]_T/[Br]_T reaction

The [HOCl]_T and [HOBr]_T decay rate constants were measured as a function of p[H⁺]. These results are shown in Fig. 4. The [HOCl]_T decay rate constant is at a maximum near p[H⁺] 6.4. The [HOBr]_T decay rate constant values maximize near p[H⁺] 7.5, which is consistent with eqn. (13).

Fig. 4 p[H⁺] dependence of observed rate constant (k_obs) for [HOCl]_T (○) and [HOBr]_T (●). Error bars represent error in linear-least squares curve fit to kinetic traces. The lines show data trends. Reaction conditions: 7.3 mM [HOCl]_T, 0.72 mM [Br]_T, 0.36 M [PO₄]_T, μ_T = 1.0 M (Na₂SO₄), 25.0(2) °C.

The chlorate/bromate yields from the decomposition of HOCl in the presence of [Br]_T were measured after a reaction period of 1 week. The results are shown in Table 1. The bromate yields increase with increasing pH, and chlorate yields decrease with increasing pH. The percent total halate ion yield is statistically constant (32.0 ± 0.3)%. This is consistent with the stoichiometry for hypohalous acid decomposition [eqn. (1)].

Table 1 Bromate and chlorate yield dependence on p[H⁺]^a

p[H^+]	%BrO3^−^b	%ClO3^−^b	%Total halate ion^b
a Reaction conditions: 5.14 mM [HOCl]T and 0.518 mM [Br]T in 0.36 M [PO4]T, 22 °C, μT = 1.0 M (Na2SO4). Yields were measured after 7 days total reaction time. b %XO3^− = [XO3^−] yield/([HOCl]T,i + [HOBr]T,i) × 100 (where X = Cl or Br). %Total halate ion = %BrO3^− + %ClO3^−.
5.57	3.92	27.9	31.8
5.88	4.36	27.8	32.2
6.23	4.81	27.5	32.3
6.59	5.37	26.9	32.3
6.94	5.93	26.2	32.1
7.29	6.44	25.2	31.6
7.60	7.23	24.7	31.9

---

Proposed mechanism and validation

The proposed mechanism is given in Table 2. The decay of [HOCl]_T is first order while the decay in [HOBr]_T is second order. The first-order fit to [HOCl]_T is valid for at least three half-lives (87.5% loss). During this time, the [HOBr]_T concentration decreases by 36%. If HOBr was catalyzing the reaction, its concentration would have to be constant to give a good first-order curve fit for the loss of HOCl. [HOBr]_T is not constant, so this implies that no direct reaction between HOCl and HOBr occurs.

Table 2 Proposed mechanism for HOCl decomposition in the presence of [Br]_T

Reaction	Rate constant^a	Ref.
a 25.0 °C, X = Cl or Br, μT = 0.5–1.0 M.
	k 14 = 1.55 × 10^3 M^−1 s^−1	6,7
	k −14 = 5.2 × 10^7 M^−1 s^−1
	k 15 = 3.0 × 10^6 s^−1	7
	k −15 = 2.3 × 10^10 M^−2 s^−1
	k 16 > 10^6 M^−1 s^−1	This work
	k 18 = 2 × 10^−3 M^−1 s^−1	18
	k 19 > 10^3 M^−1 s^−1	This work
	k 22 > 10^6 M^−1 s^−1	This work
	k 23 > 10^3 M^−1 s^−1	This work
	k 24 > 10^3 M^−1 s^−1	This work
	k 26 ≈ 97 M^−1 s^−1	13
	k 30 = 17 M^−1 s^−1	9

---

Because HOBr does not directly react with HOCl and the HOCl decay rate constant depends on [Br]_T, some other bromine species must be causing the HOCl decay. Upon reaction initiation, HOCl reacts with Br⁻ to give BrCl [eqn. (14)].⁶

Then, BrCl rapidly hydrolyzes to form HOBr [eqn. (15)].⁷ Trace amounts of Br⁻ and BrCl exist (approximately 10⁻⁹ M under equilibrium conditions).BrCl is a very reactive species and we propose that BrCl rapidly reacts with OCl⁻ to form BrOCl [eqn. (16)]. BrOCl is a known species in the gas phase.¹⁹Eqn. (17) describes the electronic rearrangement.The bromine atom of BrCl, which is slightly positive, reacts with the nucleophilic oxygen on OCl⁻. The mechanism is also favored because it is easier to expand the number of electrons around bromine than chlorine. The overall reaction in eqn. (17) can be described as a halogen cation-transfer (Br⁺ transfer to OCl⁻) with the loss of Cl⁻.

If HOBr does not react with HOCl, then HOBr must decay by some other means. HOBr disproportionation is a second-order process and produces BrO₂⁻ and Br⁻ [eqn. (18)].¹⁸

A second-order rate constant for HOBr disproportionation is calculated as shown in eqns. (12) and (13). When HOCl is present, n = 1, because HOCl rapidly converts the Br⁻ produced from disproportionation back to HOBr. The observed rate constant found in Fig. 1, (1.77 ± 0.04) × 10⁻² M⁻¹ s⁻¹, is similar to the rate constant calculated from eqn. (13) (k_disp = 1.5 × 10⁻² M⁻¹ s⁻¹) and is in agreement when the low precision in the rate parameters and the differences in ionic strength are considered. This indicates that HOBr decay in the presence of HOCl is mainly due to disproportionation of HOBr.

Chlorite ion is detected in the [HOCl]_T/[Br]_T reaction mixtures at micromolar levels. Although BrOCl reactions have not been studied, BrOCl is expected to produce chlorite because the reaction of a similar species, Cl₂O, with OCl⁻ is known [eqn. (10)]. For these reasons, a BrOCl/OCl⁻ reaction [eqn. (19)] is proposed.

This reaction proceeds by a series of steps that are probably similar to the Cl₂O/OCl⁻ reaction.⁴ The electronic rearrangement is shown by eqns. (20) and (21). The bromine atom of BrOCl reacts with the oxygen atom of OCl⁻.HOCl is regenerated and the unstable intermediate, HOClOBr⁻, dissociates to form ClO₂⁻, H⁺, and Br⁻.

The oxygen atom of OBr⁻ should also be a good nucleophile, and for this reason a BrCl/OBr⁻ reaction [eqn. (22)] and a BrOCl/OBr⁻ reaction (23) could occur.

BrOBr could hydrolyze (with OX⁻ assistance, X = Cl or Br) to form Br⁻ and BrO₂⁻ [eqn. (24)].In either case, Br⁻ rapidly forms so that HOBr is regenerated by the HOCl/Br⁻ reaction.

The kinetics for the HOCl/Br⁻ reaction have been studied by Kumar and Margerum.⁶ The following rate expression for HOCl loss may be applied for the conditions in the present work [eqn. (25)], where k′ = 1.55 × 10³ M⁻¹ s⁻¹, k′_H = 1.32 × 10⁶ M⁻² s⁻¹, k′_P = 502 M⁻² s⁻¹.

The bromide concentration at equilibrium conditions may be calculated from eqns. (14) and (15). This gives [Br⁻]_eq = 4.7 × 10⁻⁹ M. If we apply this value to eqn. (25), the expected first-order rate constant for the HOCl/Br⁻ reaction is 5.6 × 10⁻⁶ s⁻¹. The rate constant measured (Fig. 1) is (3.63 ± 0.04) × 10⁻⁵ s⁻¹, a factor of 6.5 larger. This suggests that Br⁻ is not in equilibrium but that its concentration is at a higher, steady-state level.

ClO₂⁻ and BrO₂⁻ are rapidly oxidized by either HOBr or HOCl to form the final products, ClO₃⁻ and BrO₃⁻. The ClO₂⁻ and BrO₂⁻ concentrations are small in comparison to [HOCl]_T and [HOBr]_T so that pseudo-first-order conditions prevail. Eqns. (26) and (27) represent the reaction and observed rate constant for the HOBr/ClO₂⁻ reaction,¹³ where k₁ = 97 M⁻¹ s⁻¹, k₂^H/k₋₁ = 3.1 × 10⁵ M⁻¹, k₂^H₂PO₄/k₋₁ = 8.3 M⁻¹, K_a^HOBr = 10^−8.59 M. At p[H⁺] = 7.0 (0.36 M [PO₄]_T), the observed rate constant for the HOBr/ClO₂⁻ reaction is k_obs = 39.4[HOBr]_T.

Eqns. (28)–(29) represent the reaction and observed rate constant for the HOCl/ClO₂⁻ reaction,²⁰ where k₁ = 1.6 M⁻¹ s⁻¹, k₂^H/k₋₁ = 1.6 × 10⁴ M⁻¹, k₂^H₂PO₄/k₋₁ = 8.5 M⁻¹, K_a^HOCl = 10^−7.5 M. At p[H⁺] = 7.0 (0.36 M [PO₄]_T), the observed rate constant for the HOCl/ClO₂⁻ reaction is k_obs = 0.146[HOCl]_T. As a result the HOBr/ClO₂⁻ reaction dominates chlorate formation when [HOCl]_T/[HOBr]_T < 54. In Fig. 1, this condition is satisfied for about 4 half-lives of [HOCl]_T decay.

Less kinetic information is known about reactions with bromite, but approximate rate constants are known. The rate constant for the HOCl/BrO₂⁻ reaction⁹ is 17 M⁻¹ s⁻¹ [eqn. (30)] and the rate constant for the HOBr/BrO₂⁻ reaction²¹ is 0.018 M⁻¹ s⁻¹ [eqn. (31)]. The result is that the HOCl/BrO₂⁻ reaction dominates bromate formation when [HOCl]_T/[HOBr]_T > 0.0053.

The final products are chlorate and bromate. By the mechanism given in Table 2, the chlorate formation should follow the [HOCl]_T decay. The rate constants for chlorate formation and [HOCl]_T decay are relatively close. The rate constants for bromate formation and [HOBr]_T decay are in better agreement, which indicates that BrO₃⁻ formation follows directly from [HOBr]_T decay. However, as [Cl]_T increases, more BrCl is available to react with OBr⁻ and/or form a higher steady-state amount of BrOCl. Then, the [HOBr]_T decay rate constant increases and more BrO₃⁻ forms as [Cl]_T increases (as observed in Fig. 3). For this reason, the BrCl/OBr⁻ reaction and the BrOCl/OBr⁻ reaction become more important at higher [Cl]_T, and they cause the [Cl]_T² dependence of k_obs^HOBr.

The pH dependence of the observed rate constants for HOCl and HOBr decay show that the reaction rates are somewhat sensitive to pH. The pH dependence of HOBr decay rate constant follows the relationship given in eqn. (13). The HOCl decay rate constant should not decrease with decreasing pH according to eqn. (25). The observed pH dependence suggests that the OCl⁻/BrCl reaction [eqn. (16)] affects the observed HOCl decay rate at lower pH. At lower pH, less OCl⁻ is available for reaction, which decreases the observed rate constant.

The pH dependence of the bromate/chlorate yields indicates bromate formation is favored at a high pH and chlorate formation at low pH. This can be explained in terms of the given mechanism. The disproportionation of HOBr [eqns. (12) and (13)] is general-base assisted and the HPO₄²⁻ term (k_B^P) dominates the rate. As pH decreases (until p[H⁺] 6.4), the HOCl/Br⁻ reaction rate constant increases while the HOBr self-disproportionation reaction rate constant decreases. Also, the [BrO₃⁻]/[HOBr]_T stoichiometry decreases from 1.0 (when HOCl is present) to 0.33 (when HOCl is absent). If the length of time for HOCl availability is shorter, then there is less time for bromate to be produced under high reaction stoichiometry conditions. For this reason, the bromate yield is lower. However, the overall halate yield as a function of total hypohalous acid remains constant with pH, so that the relative chlorate yield increases as pH increases.

The [Br]_T dependence of rate constant for [HOCl]_T decay shows a squared term (Fig. 2). As [Br]_T increases, the formation of Br₂ is favored [eqns. (32) and (33)].⁷

As [Br]_T increases, the [BrCl]/[Br]_T ratio increases with a squared dependence on [Br]_T. This suggests that steady-state BrCl contributes to the rate-determining steps.

HOCl decay in the presence of environmentally relevant bromide ion and buffer levels

The bromide and chlorine concentrations in most of this work are higher than typical freshwater levels, because higher levels are needed for detailed kinetic analysis. However, the decay of [HOCl]_T in low-level Br⁻ concentrations was studied by following the decay of the [HOCl]_T absorbance at 290 nm. At this wavelength, the [HOBr]_T absorbance is negligible, so the [HOCl]_T loss may be monitored directly. Fig. 5 shows the decay of HOCl in the presence 9.24 µM [Br]_T and in zero [Br]_T. The decay of HOCl with no Br⁻ is expected to be third order, according to Gordon's mechanism [eqns. (9)–(11)].⁴

Fig. 5 Decomposition of 11.2 mM [HOCl]_T in the presence of 9.24 µM [Br]_T (○) and 0 µM [Br]_T (△). Reaction conditions: 0.36 M [PO₄]_T, p[H⁺] 7.02, μ_T = 1.0 M (Na₂SO₄), 25.0(2) °C.

When the decomposition of [HOCl]_T is measured in the presence of environmentally relevant [Br]_T, the presence of [Br]_T accelerates [HOCl]_T decay (Fig. 5). We propose that very small amounts of BrCl catalyze the decomposition. However, a change in rate-controlling paths has occurred. Since the [HOCl]_T/[HOBr]_T ratio is greater than 54, HOCl, and not HOBr, reacts with ClO₂⁻ to form ClO₃⁻.

Table 3 includes the ClO₃⁻ and BrO₃⁻ yields as a function of time for the decay of HOCl. The presence of 14.2 µM [Br]_T results in a three-fold increase in the ClO₃⁻ yield in comparison to no added [Br]_T. Also, the BrO₃⁻ yield is far above the MCL within 10 h. The AWWA Disinfection Committee Survey reported that 90% of water utilities employ a contact time (the time between the point of chlorine application and the first consumer) of 10 h.¹

Table 3 Chlorate and bromate production from HOCl decomposition^a

Time/min	0 µM [Br]T	14.2 µM [Br]T
	[ClO3^−]/µM	[ClO3^−]/µM	[BrO3^−]/µM
a Reaction conditions: 5.07 mM [HOCl]T, 0.36 M [PO4]T, p[H^+] 6.98, μT = 1.0 M (Na2SO4), 25.0(2) °C.
22	7.1	9.1	< 0.5
602	24.5	85.0	0.77
1005	31.1	132	1.2
1387	41.7	164	2.3
1787	57.8	200	2.8
2607	75.4	260	3.8

---

In this present work, pH control was achieved by the use of 0.36 M phosphate buffer. However, this level is much higher than expected for natural waters. Instead, carbonate buffers mediate pH in natural waters and the HCO₃⁻ concentration is typically about 1 mM.²² We can use this value to predict the pH effect on HOBr decomposition. By applying the Brønsted–Petersen relationship,²³ the rate constant for HCO₃⁻ assistance of HOBr disproportionation can be calculated (k_B^HC = 0.030 M⁻² s⁻¹). Thus, the rate constant for HOBr disproportionation can be altered to account for carbonate buffers as in eqn. (34) [k_B^C = 0.32 M⁻² s⁻¹, other rate parameters are defined by eqns. (12) and (13)].¹⁸

With the adjustment for buffer type, the pH dependence of the observed rate constant for HOBr decay changes. HOBr decay is faster at low pH because the k_1a term dominates the rate. Therefore, BrO₃⁻ production is slowest at high pH. Under typical drinking water conditions (pH 6–8.5, [HCO₃⁻] = 1 mM), the presence of carbonate buffer contributes very little to the rate of HOBr decay. For example, at pH 7.5, the HOBr decay rate constant in the presence of 0 mM and 1 mM HCO₃⁻ is 1.7 × 10⁻³ M⁻¹ s⁻¹. However, as the pH and the carbonate buffer concentration increase, the effect of CO₃²⁻ catalysis will be observed. CO₃²⁻ is a stronger Brønsted base than HCO₃⁻, and this increases the rate. As a result, at pH 9, the rate constant for HOBr decay doubles when the carbonate buffer concentration is increased from 1 mM to 36 mM.

Conclusions

The reaction mechanism for the decomposition of [HOCl]_T in the presence of moderate levels (0.1 mM–1.0 mM) of [Br]_T is given. In contrast to previous work, no direct reaction is observed between the molecular species HOCl and HOBr. We propose that decomposition of HOCl in the presence of Br⁻ proceeds by a series of halogen(I) cation-transfer reactions with halide ion release. Br⁻ and BrCl occur in trace amounts and catalyze the HOCl decomposition. For example, in the presence of 9.2 µM Br⁻, the decay rate constant of HOCl (k_obs^HOCl = 1.9 × 10⁻⁶ s⁻¹ at 11 mM [HOCl]_T, p[H⁺] 7.0, 0.36 M [PO₄]_T, μ_T = 1.0 M, 25.0 °C) is 25% greater than in the absence of Br⁻. HOBr decay is attributed to disproportionation except at high total chlorine concentrations. ClO₃⁻ and BrO₃⁻ are formed from the rapid reactions of HOBr with ClO₂⁻ and HOCl with BrO₂⁻, respectively. For this reason, chlorate and bromate formation rates are directly related to the decay rates of HOCl and HOBr. The stoichiometry of the oxidation of HOBr to BrO₃⁻ is 1∶1 when HOCl is present because HOCl rapidly converts BrO₂⁻ to BrO₃⁻ and HOBr is regenerated by HOCl/Br⁻ reaction [eqn. (35)].As a consequence, the BrO₃⁻ yield from HOBr decay is higher in the presence of HOCl than in the absence of HOCl.

A pH dependence in the BrO₃⁻ and ClO₃⁻ yields was observed; for high [PO₄]_T, BrO₃⁻ is favored at higher pH and ClO₃⁻ is favored at lower pH. However, when carbonate buffer instead of phosphate buffer is present, the HOBr decay rate increases with decreasing pH. Carbonate buffer is more typically found in natural waters and increasing the pH of water treatment can help to minimize BrO₃⁻ formation.

Past work has focused on bromate formation in drinking water by ozonation of bromide-containing source water.²⁴ This present work, however, shows that bromate formation in chlorination is possible and reaction pathways exist for bromate formation above the MCL. Fortunately, bromate formation is slower in hypochlorous acid than in ozone. Bromate formation in chlorinated systems can be limited by increasing pH, decreasing the concentration and contact time of [Cl]_T, and decreasing buffer concentration. Therefore in the absence of other reactants, such as amines, organic matter, or light, the decomposition of HOCl in the presence of Br⁻ may be an important pathway for BrO₃⁻ formation in drinking water.

Acknowledgements

This work was supported by NSF grant CHE-9818214 and by the Purdue Research Foundation. Preliminary studies were conducted by M. J. Tranovich (Senior Thesis, Purdue University, 1993).

References

1. G. C. White, Handbook of Chlorination and Alternative Disinfectants, Van Nostrand Reinhold, New York, 3rd edn., 1992, pp. 377–379.
2. J. L. Oxenford, in Disinfection By-Products in Water Treatment, ed. R. A. Minear and G. L. Amy, Lewis Publishers, Boca Raton, FL, 1996, pp. 3–16.
3. T. X. Wang and D. W. Margerum, Inorg. Chem., 1994, 33, 1050.
4. L. C. Adam, I. Fábián, K. Suzuki and G. Gordon, Inorg. Chem., 1992, 31, 3534.
5. (a) P. Westerhoff, M. Siddiqui, J. Debroux, W. Zhai, K. Ozekin and G. Amy, in Critical Issues in Water and Wastewater Treatment, ed. J. N. Ryan and M. Edwards, ASCE, 1994, pp. 670–677; (b) S. W. Krasner, M. J. McGuire, J. G. Jacangelo, N. L. Patania, K. M. Reagan and E. M. Aieta, J. – Am. Water Works Assoc., 1989, 81, 41; (c) CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press, Boca Raton, FL, 1995.
6. K. Kumar and D. W. Margerum, Inorg. Chem., 1987, 26, 2706.
7. Q. Liu and D. W. Margerum, Environ. Sci. Technol., 2001, 35, 1127.
8. M. Lewin and M. J. Avrahami, J. Am. Chem. Soc., 1955, 77, 4491.
9. T. F. Perrone, Ph.D. Thesis, Purdue University, 1999.
10. C. S. Furman, Ph.D. Thesis, Purdue University, 1997.
11. Y. Kurokawa, A. Maekawa, M. Takahashi and Y. Hayashi, Environ. Health Perspect., 1990, 87, 309.
12. Fed. Reg., 1998, 63(241), 69390.
13. C. S. Furman and D. W. Margerum, Inorg. Chem., 1998, 37, 4321.
14. EPA method 300.1, US EPA, Cincinnati, OH, 1997.
15. K. D. Fogelman, D. M. Walker and D. W. Margerum, Inorg. Chem., 1989, 28, 986.
16. R. C. Troy and D. W. Margerum, Inorg. Chem., 1991, 30, 3538.
17. L. C. Adam and G. Gordon, Anal. Chem., 1995, 67, 535.
18. R. C. Beckwith and D. W. Margerum, Inorg. Chem., 1997, 36, 3754.
19. J. B. Burkholder, G. Knight and J. J. Orlando, J. Photochem. Photobiol. A, 2000, 134, 133.
20. Z. Jia, D. W. Margerum and J. S. Francisco, Inorg. Chem., 2000, 39, 2614.
21. P. Engel, A. Oplatka and B. Perlmutter-Hayman, J. Am. Chem. Soc., 1954, 76, 2010.
22. S. E. Manahan, Environmental Chemistry, Lewis Publishers, Boca Raton, FL, 7th edn., 2000, pp. 75–77.
23. J. H. Espenson, Chemical Kinetics and Reaction Mechanisms, McGraw-Hill, New York, 2nd edn., 1995, pp. 233–236.
24. M. S. Siddiqui, G. L. Amy and R. G. Rice, J. – Am. Water Works Assoc., 1995, 87, 58.

---