Article pubs.acs.org/est
Perchlorate Production by Photodecomposition of Aqueous Chlorine Solutions Balaji Rao,† Nubia Estrada,‡ Shelly McGee,§ Jerry Mangold,∥ Baohua Gu,† and W. Andrew Jackson*,‡ †
Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, United States Depratment of Civil and Environmental Engineering, Texas Tech University, Lubbock, Texas 79409-1023, United States § Department of Mathematics, Our Lady of the Lake University, San Antonio, Texas 78207, United States ∥ Department of Civil, Architectural and Environmental Engineering, University of Texas at Austin, Austin, Texas, United States ‡
S Supporting Information *
ABSTRACT: Aqueous chlorine solutions (defined as chlorine solutions (Cl2,T) containing solely or a combination of molecular chlorine (Cl2), hypochlorous acid (HOCl), and hypochlorite (OCl−)) are known to produce toxic inorganic disinfection byproduct (e.g., chlorate and chlorite) through photoactivated transformations. Recent reports of perchlorate (ClO4−) productiona wellknown thyroid hormone disruptor from stored bleach solutions indicates the presence of unexplored transformation pathway(s). The evaluation of this potential ClO4− source is important given the widespread use of aqueous chlorine as a disinfectant. In this study, we perform detailed rate analysis of ClO4− generation from aqueous chlorine under varying environmental conditions including ultraviolet (UV) light sources, intensity, solution pH, and Cl2,T concentrations. Our results show that ClO4− is produced upon UV exposure of aqueous chlorine solutions with yields ranging from 0.09 × 10−3 to 9.2 × 10−3% for all experimental conditions. The amount of ClO4− produced depends on the starting concentrations of Cl2,T and ClO3−, UV source wavelength, and solution pH, but it is independent of light intensity. We hypothesize a mechanistic pathway derived from known reactions of Cl2,T photodecomposition that involves the reaction of Cl radicals with ClO3− to produce ClO4− with calculated rate coefficient (kClO4−) of (4−40) × 105 M−1 s−1 and (3−250) × 105 M−1 s−1 for UV-B/C and UV-A, respectively. The measured ClO4− concentrations for both UV-B and UV-C experiments agreed well with our model (R2 = 0.88−0.99), except under UV-A light exposure (R2 = 0.52−0.93), suggesting the possible involvement of additional pathways at higher wavelengths. Based on our results, phototransformation of aqueous chlorine solutions at concentrations relevant to drinking water treatment would result in ClO4− concentrations (∼0.1 μg L−1) much below the proposed drinking water limits. The importance of the hypothesized mechanism is discussed in relation to natural ClO4− formation by atmospheric transformations.
■
INTRODUCTION Perchlorate, a thyroid hormone disruptor, is an emerging contaminant of growing concern with a U.S. federal health advisory level of 15 μg L−1 1 and a state mandated maximum contaminant limit (MCL) of 6 μg L−1 and 2 μg L−1 for California and Massachusetts, respectively.2,3 Additionally, the Environmental Protection Agency (EPA) is currently in the process of setting a federally mandated MCL for ClO4−. The occurrence of ClO4− in the environment can be attributed to both natural and anthropogenic sources.4−7 Man-made sources of ClO4− are primarily manufactured by electrochemical oxidation of chlorate (ClO3−),7 whereas the mechanisms involved in natural ClO4− formation are not well understood. Chlorine and oxygen isotopic ratios of natural ClO4− show patterns that strongly indicate a stratospheric source and suggest both an ozone and photo-oxidation mediated formation process.5 Recent studies have also demonstrated the production of ClO4− from nonelectrochemical reactions both to assess their role in indirect anthropogenic production and to determine plausible pathways for natural ClO4− formation.8−11 © 2012 American Chemical Society
Examples of experimentally studied pathways include the reaction of ozone (O3) with inorganic chlorine species such as chloride (Cl−), hypochlorite (OCl−), hypochlorous acid (HOCl), chlorite (ClO2−), and chlorine dioxide (ClO2), and ultraviolet light (UV) mediated oxidation of ClO2− and OCl−.8−11 An underexplored source of ClO4− is aqueous chlorine solutions (defined as aqueous solutions containing solely or a combination of molecular chlorine (Cl2), HOCl and OCl−) that are commonly used as a disinfectant for applications including water treatment systems, swimming pools, drip irrigation lines, household products, and numerous other applications.12 Aqueous chlorine solutions decompose during prolonged storage, resulting in a reduction of disinfection potential and formation of disinfection byproducts (DBPs).13,14 Received: Revised: Accepted: Published: 11635
April 17, 2012 August 26, 2012 September 10, 2012 September 10, 2012 dx.doi.org/10.1021/es3015277 | Environ. Sci. Technol. 2012, 46, 11635−11643
Environmental Science & Technology
Article
Table 1. Comparison of the Rate Coefficient of Cl2,T Decay (kCl2,T) and ClO4− Production (kClO4−) at Different Experimental Conditionsa evaluated param. effect of UV-source
effect of starting Cl2,T concn. ([Cl2,T]o)
effect of pH/Cl2,T speciation effect of UV intensity
lamp type
initial Cl2,T ([Cl2,T]o) (mM)
UV photon intensity (μEinstein L−1 s−1)
initial pH
final pH
Cl2,T transformation 1st order rate coefficient, kCl2,T (× 10−3 s−1)
overall rate coefficient of ClO4− production, kClO4 (× 105 M−1 s−1)
UV-C UV-B UV-A UV-B UV-B UV-A UV-A UV-A UV-B UV-B UV-B UV-B
16.0 17.1 18.0 1.6 8.2 2.3 282 1433 16.2 17.0 17.2 15.4
26 34 50 34 34 50 50 50 34 34 9 17
10.6 11.0 10.9 8.7 10.0 8.8 11.5 11.4 2.7 6.6 10.7 10.1
7.5 10.4 10.1 7.3 9.3 8.8 11.3 8.3 2.2 1.7 10.8 10.4
4.8 8.3 0.4 8.3 8.3 0.4 0.4 0.4 2.4 2.4 2.1 4.2
17 8 18 4 6 3 100 250 40 22 8 9
The details on the three types of UV sources used and the methodology employed to calculate the first-order rate coefficient of Cl2,T (kCl2,T) from eq 1 are available in the Supporting Information. The second-order rate co-efficient of ClO4− (kClO4) is obtained by regressed fit to the experiment data calculated from eq 21 for 1 minute time steps. Hydrochloric acid (HCl) was used to lower the pH for experiments evaluating the speciation of Cl2,T, and no external buffers were added to maintain a constant pH during photolysis. a
This decay is substantially enhanced under exposure to light resulting in accelerated formation of chloride (Cl−), chlorite (ClO2−), chlorate (ClO3−), and oxygen (O2).13−16 The mechanisms of these photodecomposition reactions upon exposure to ultraviolet (UV) light sources are well-known.16 However, recent reports of perchlorate (ClO4−) production as a minor product (∼3 orders of magnitude lower than ClO3− and Cl− yield during aging of commercial bleach solutions), indicates the presence of additional transformation pathway(s) of aqueous chlorine solutions that have so far been unexplored.9,17 Production of ClO4− from UV photolysis (at 254 nm) of alkaline aqueous chlorine (as OCl−) was previously demonstrated at relatively high concentrations (∼ 200 mM of OCl−), while experiments performed at lower concentrations were inconclusive, likely due to the presence of background ClO4−.9 The authors in that study proposed a general mechanism involving the production of chlorite (ClO2−) and chlorine dioxide (ClO2) from OCl−, followed by their subsequent photo-oxidation to chlorine trioxide (ClO3) radicals, which, upon hydrolysis, generates ClO4−.9 However, no information was provided on the rate of ClO4− formation and that of the major products (Cl− and ClO3−) during the UV photolysis of chlorine species (Cl2/HOCl/OCl−). Further the impact of variables, in particular, the effect of concentration, chemical speciation of aqueous chlorine solutions, and different UVsources relevant to outdoor radiation and disinfection were not examined. Therefore, the primary objective was to (1) determine the photoformation yields of ClO4− from Cl2,T under varying environmental conditions such as Cl2,T concentration, UV source wavelengths, UV intensities, and solution pH; (2) develop a kinetic model of ClO4− and ClO3− production; and (3) propose an overall mechanistic pathway for ClO 4− formation from phototransformation of aqueous chlorine. We used stock aqueous chlorine solutions (obtained from calcium hypochlorite (Ca(OCl)2) salt) containing negligible amounts of starting ClO4− for our experiments, which allowed the evaluation of Cl2,T phototransformation to ClO4− up to 2 orders of magnitudes lower than the previous study.9 The model and results are used to predict the potential ClO4−
impact in the environment from widespread use of chlorine (Cl2,T) as a disinfectant.
■
MATERIALS AND METHODS Experimental Set-up. Details of chemicals, reagents, and the analysis are provided in the Supporting Information. Aqueous chlorine solutions were prepared from calcium hypochlorite salt (available chlorine ∼ 65%; Sigma-Aldrich) and were filtered through a 0.2 μm filter (PVDF Membrane, Millipore) before use in photodegradation experiments. Aqueous chlorine solutions thus obtained were irradiated without any chemical modification for all experiments except the low pH experiments (Table 1). Hydrochloric acid (HCl) was used to lower the pH for experiments evaluating the speciation of Cl2,T, and no external buffers were added to maintain a constant pH during photolysis. Irradiation experiments were conducted using a Rayonet photochemical chamber reactor (Model RPR-200; Southern New England Ultraviolet Co., Branford, CT) equipped with a merry-go-round (MGR) sample holder (Figure S1, Supporting Information). Inside the reactor was a cylindrical chamber (20.4 cm in diameter; 40.5 cm in depth) that can hold up to 16 UV lamps. Three types of UV lamps were usedUV-C, UV-B, and UV-Aoperating at light frequency regimes corresponding to peak emission wavelengths at 254, 300, and 350 nm, respectively (Table 1 and S2, Supporting Information). The procedure used to determine the photon-intensity and the path-length of the quartz vial used in our experiments is provided in the Supporting Information text. Cylindrical quartz tubes (Rayonet, Model RQV-5; 1.27 cm inside diameter, 10.5 cm length, and 1 mm thickness) were used for all irradiation experiments. To ensure equal irradiation of all samples regardless of lamp position or intensity fluctuations, the quartz tubes were held in the reactor using a two-story MGR, which rotated at 5 rpm. Each tube was filled with a known concentration of aqueous Cl2,T (15 mL/tube) and capped with a silicone stopper. The minimum concentration of Cl2,T in solution used for our experiments was dependent on the analytical ability to accurately quantify the ClO4− produced by light exposure. Time zero controls were obtained by sacrificing samples just before UV exposure. Dark control experiments were conducted 11636
dx.doi.org/10.1021/es3015277 | Environ. Sci. Technol. 2012, 46, 11635−11643
Environmental Science & Technology
Article
Figure 1. Percent final yield of ClO4− (red solids and left y-axis) and ClO3− (black solids and right y-axis) obtained from (a) UV-A, and (b−d) UV-B irradiation of Cl2,T solution at different experiment conditions (see Table 1 for details). The error bars on the vertical bars represent the calculated standard deviations for the duplicate experiments.
than a 15% change in the Cl2,T, Cl−, ClO2−, ClO3−, and ClO4− concentrations compared to time zero controls. Due to the much longer experimental run time (∼5000 min) for the highest concentration UV-A experiment, the formation of ClO4− in the dark controls was not negligible (30% increase from time zero controls). Thermal decomposition of aqueous chlorine solutions are known to produce ClO4− at very low rates in the dark 17 and may account for the observed ClO4− in our dark controls (more details in the Supporting Information). The measured amount in the dark controls was still a fraction (∼10%) of the final ClO4− concentration produced by UV-A light exposure. Influence of Cl2,T Concentration on Yields. Detailed evaluation of ClO4− production at different Cl2,T concentrations was carried out for UV-B and UV-A lamps. The solution was not pH adjusted with starting values ranging from 9 to 11 and most of the Cl2,T in OCl− form (Table 1). Increasing the initial Cl2,T concentration from 2.3 mM to 1433 mM (UV-A) and 1.6 mM to 17 mM OCl− (UV-B) resulted in an enhancement of ClO4− yield, from 0.09 × 10−3% to 9.2 × 10−3% and from 0.3 × 10−3% to 0.6 × 10−3%, respectively (Figure 1). The ClO4− yield exhibited a log−linear increase, with initial Cl2,T concentration under both UV-A (slope = 0.49, y-intercept = −6.0, R2 = 0.95) and UV-B (slope = 0.43, y-intercept = −5.8, R2 = 0.99) radiation (Figure 1). The increase in ClO4− yield with respect to the starting Cl2,T concentration ([Cl2,T]o) implies that the production rate is not a simple first-order equation with respect to the Cl2,T concentration. The ClO3− yield ranged from 11 to 30% and did not show a consistent dependency with starting Cl2,T concentration (Figure 1). Influence of UV Wavelength and Intensity on Yields. Irradiation of ∼18 mM Cl2,T (Cl2,T as OCl−, starting pH = 10) using UV-C lamps produced the highest ClO4− yield ((1.2 ± 0.03) × 10−3 mol mol−1 %) compared to UV-A and UV-B at similar conditions ((0.59 ± 0.04) × 10−3 (UV-B) and (0.46 ± 0.04) × 10−3% (UV-A)). These yields are lower compared to a
by covering the sample cuvettes with aluminum foil while placed inside the reactor. Although not controlled, the reaction temperature reached steady state at 30 (±1) o C in the laboratory fume hood after an initial increase from ambient. It should be noted that the UV lamps were switched on at least an hour prior to the start of experiments, ruling out the possible effects of changes in UV intensity during the kinetic runs. The kinetic measurement data points for Cl2,T, Cl−, ClO2−, ClO3−, and ClO4− were obtained by sacrificing samples at different time intervals. All samples and controls were duplicated for each experiment. A significant loss of total chlorine species (∼25 %) was noticed in the UV-B and UV-C experiments. However, there was complete closure of the chlorine mass balance for the UV-A experiments in which the residual Cl2,T was quenched with hydrogen peroxide before the IC analysis (see SI for details). The apparent loss of chlorine species in the earlier experiments may be attributed to the interference of the residual Cl2 in the measurement of Cl−.18
■
RESULTS AND DISCUSSION Overall Product Yields. To evaluate the influence of common reaction variables on the production of ClO4− from phototransformation of aqueous chlorine solutions, we performed different experiments by changing the starting pH (thereby the speciation of Cl2,T), starting Cl2,T concentration ([Cl2,T]o), UV source wavelength (for UV-A, UV-B, and UVC), and UV-B source intensity (Table 1). Photolysis of Cl2,T produced ClO4− as a minor product ((0.09 to 9.2) × 10−3%) for all conditions tested. Major reaction products identified were Cl− (50−91%) and ClO3− (11−30%) (Figure 1). Chlorite was detected (limit of detection (LOD) = 15 μM) only for UVA irradiation at starting Cl2,T ≥ 18 mM and behaved as an intermediate species (Figure 3a and Figure S3, Supporting Information). Analysis of the dark control samples at the end of all experiments except the UV-A experiment involving the highest Cl2,T concentration ([Cl2,T]o = 1433 mM) exhibited less 11637
dx.doi.org/10.1021/es3015277 | Environ. Sci. Technol. 2012, 46, 11635−11643
Environmental Science & Technology
Article
Figure 2. Kinetics of ClO4− (solid circles) production from light exposure of Cl2,T at different (a) UV wavelengths, and for UV-B irradiation (b) light intensities, (c) Cl2,T concentrations, and (d) starting pH (see Table 1 for details on experimental parameters). The solid circles and error bars represent the average measured concentration and its standard deviation (n = 2) for duplicate experiments. The lines represent the modeled curves and their regressed fit to the data calculated from eq 21 for 1 min time steps (kCl2,t and kClO4 are values are provided in Table 1). Data for the UV-A experiments is provided in the Supporting Information (Figure S4).
previously reported yield of 2.4 × 10−3 mol mol−1% for the exposure of 194 mM OCl− solution to UV-C radiation.9 The greater reported yield is likely due to the use of a higher starting concentration of OCl− in the previous study, a trend that is clearly witnessed in our UV-A and UV-B experiments. The change in ClO4− yield with wavelength for Cl2,T photolysis is notably different than change in yields with respect to wavelength for ClO2− photolysis, where higher yields were obtained at longer wavelengths (0.6 × 10−2% and 3.0 × 10−2% for UV-C and UV-A, respectively).8 The yield of ClO3− was less variable with similar values for all sources of UV radiation (average yield ∼17−19%) (Figure 1). Decreasing the UV-B intensity by a factor of 4 (total UV irradiance from 60 to 15 W −2 ) did not considerably change the yield of ClO4− (ranged from 0.6 × 10−3 to 0.7 × 10−3%) or ClO3− (ranged from 18 to 20%) (Figure 1). Influence of Starting pH on Yields. The impact of pH was evaluated under UV-B radiation (photon intensity = 34 μEinstein L−1 s−1) and a starting Cl2,T ∼ 17 mM. Switching the predominant Cl2,T species from OCl− to HOCl/Cl2 by lowering the pH to 0.9, Figure 2). 11641
dx.doi.org/10.1021/es3015277 | Environ. Sci. Technol. 2012, 46, 11635−11643
Environmental Science & Technology information on all reaction parameters (including rate constants) involving these species and is beyond the scope of this paper. Implications. The rapid production of ClO4− from the photo-decay of aqueous chlorine solutions suggests a potential route for ClO4− exposure.22,28 For comparison, the calculated yields of ClO4− reported in a study involving the enhanced thermal decomposition (temperature = 50 °C) of utility bleach solutions (([Cl2,T]o = 1200 − 2200 mM and [ClO3−]o = 53− 275 mM) ranges between (10 − 20) × 10−4 (mol/mol).22 This is at least an order of magnitude greater than the highest aqueous chlorine concentration measured in our photolysis study (Figure 1) and represents a critical ClO4− formation process in stored bleach solutions. However, the importance of photolytic processes as a major source of ClO4− depends on [Cl2,T]o, UV light source characteristic, pH and background Cl species (Cl−, ClO3−, and ClO2−). Phototransformation of Cl2,T solutions at concentrations relevant to drinking water treatment ([Cl2,T]o < 0.1 mM and pH ∼ 7.0) would result in maximum ClO4− concentrations of ∼0.1 μg L−1 and therefore would not constitute a significant source on its own. For chlorinated swimming pools (constant Cl2,T ∼ 0.04 mM) exposed to a 12 h sunlight period for only 3 days, (assuming the predominant UV source in sunlight is in the UV-A regime 19) our model predicts enhancement of ClO4− production with concentrations reaching the most stringent regulatory levels of 2 μg L−1 in 3 days. Mitigation strategies, in such cases, to minimize ClO4− production should include avoiding exposure of chlorinated solutions to UV light source, and inhibiting the production of ClO3− and ClO2−. Further studies are required to determine the applicability of our kinetic analysis at very low concentrations of Cl2,T (
