Article pubs.acs.org/est
Peracetic Acid Oxidation of Saline Waters in the Absence and Presence of H2O2: Secondary Oxidant and Disinfection Byproduct Formation Amisha D. Shah,†,‡ Zheng-Qian Liu,†,§ Elisabeth Salhi,† Thomas Höfer,∥ and Urs von Gunten*,†,⊥ Eawag, Swiss Federal Institute of Aquatic Science and Technology, Ü berlandstrasse 133, CH-8600 Dübendorf, Switzerland School of Environmental Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074, China ∥ Federal Institute for Risk Assessment (BfR), Max-Dohrn-Strasse 8-10, D-10589 Berlin, Germany ⊥ School of Architecture, Civil and Environmental Engineering (ENAC), Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland † §
S Supporting Information *
ABSTRACT: Peracetic acid (PAA) is a disinfectant considered for use in ballast water treatment, but its chemical behavior in such systems (i.e., saline waters) is largely unknown. In this study, the reactivity of PAA with halide ions (chloride and bromide) to form secondary oxidants (HOCl, HOBr) was investigated. For the PAA−chloride and PAA− bromide reactions, second-order rate constants of (1.47 ± 0.58) × 10−5 and 0.24 ± 0.02 M−1 s−1 were determined for the formation of HOCl or HOBr, respectively. Hydrogen peroxide (H2O2), which is always present in PAA solutions, reduced HOCl or HOBr to chloride or bromide, respectively. As a consequence, in PAA-treated solutions with [H2O2] > [PAA], the HOBr (HOCl) steady-state concentrations were low with a limited formation of brominated (chlorinated) disinfection byproducts (DBPs). HOI (formed from the PAA−iodide reaction) affected this process because it can react with H2O2 back to iodide. H2O2 is thus consumed in a catalytic cycle and leads to less efficient HOBr scavenging at even low iodide concentrations (0.6 mM) at a constant bromide concentration does not significantly affect the PAA residual concentration. Alternatively, iodide does affect the PAA residual concentration because it consumes PAAH rapidly, with a second-order rate constant that is 3 orders of magnitude higher than that of bromide, to form HOI, which is then converted back to iodide through the reaction with H2O2 (HO2−), generating an iodidecatalyzed process. Scheme 1 illustrates the overall reactions in the PAA−Br−−I−−H2O2 system and how these two catalytic cycles are connected.
Figure 2. Effect of H2O2 dose on HOBr formation after 110 min under synthetic brackish water conditions. Conditions: [PAA]0 = 2.0 mM, [I−]0 = 0.26 μM, [Br−]0 = 0.42 mM, [Cl−]0 = 0.27 M, [carbonate buffer] = 50 mM, [2,6-diclorophenol]0 =3 mM (only for [H2O2] = 0 mM), [H2O2] = 0−5.6 mM, pH 7.0 ± 0.1. Symbols represent experimental data, and lines represent model calculations obtained at 110 min for individual H2O2 doses. For a description of the modeling, see Text S8 (SI) and main text.
Scheme 1. Reactions Occurring in the PAA−Br−−I−−H2O2 System in the Absence and Presence of DOM (or 2,6-DCP), where the Bromide- and Iodide-Catalyzed Reactions Are Connected by H2O2a−c
included in the model. The results indicated that the model overpredicted HOBr formation compared to the experimental results by approximately 0.1−0.2 mM over the H2O2 range of 0−1.2 mM with a factor of 2.6 at 0.6 mM H2O2 (Figure 2, dotdashed line). These findings are consistent with the results found in Figure 1 for a H2O2 concentration of 0.65 mM, where the model results (dashed line) overpredicted the experimental data (solid circles) by a factor of 2. This is likely due to additional HOBr sinks not included in the model, such as bromate, which could lead to lower experimental HOBr values (Figure 2). A good agreement between the measured and modeled data was found for H2O2 concentrations of ≥1.5 mM, where model simulations matched the experimental results and no HOBr was formed in either case (Figure 2). The PAA residual concentration was also found to depend on the H2O2 dose. After 110 min, the residual PAA concentration was about 1 mM for a dose of 0−0.79 mM H2O2 (Figure 2, open squares). However, it decreased with increasing H2O2 concentration, reaching 0.2 mM PAA for a H2O2 dose of 5.6 mM (Figure 2). To better understand this decrease, additional model simulations were performed in which the PAA residual concentration after 110 min was determined using the same modeling conditions as described for HOBr formation (Kintecus, Text S8, SI). With this model, three different simulations were performed in which either (i) PAA and bromide were present (Figure 2, solid line, PAA−Br model); (ii) PAA and iodide were present (Figure 2, dotted line, PAA−I model); or (iii) PAA, bromide, and iodide were present (Figure 2, dashed line, PAA−Br−I model). The PAA− Br model results indicate that 0.42 mM bromide leads to a relatively small PAA decrease from 1.6 to 1.4 mM when the H2O2 concentration is increased from 0 to 5.6 mM. However, PAA with 0.26 μM iodide (PAA−I model) leads to a larger PAA decrease from 2.0 to 1.25 mM, suggesting that iodide can play a dominant role in PAA degradation, even at low concentrations. The additive effect of these two models is then represented in the PAA−Br−I model curve (dashed line), which well matches the experimental PAA values (Figure 2, open squares) for H2O2 in the range of 0.6−2 mM. This differs
a
Bold arrows in the iodide cycle indicate that these reactions are faster. Reaction of HOBr (HOI) with DOM can lead to Br (I) DBPs. c Reaction for HOBr quenching with 2,6-DCP leading to 4-Br-2,6-DCP is also shown. b
To further investigate the effect of iodide, experiments were conducted with varying iodide concentrations (0−10.9 μM) and constant levels of chloride (0.27 M) and bromide (0.46 mM) in the presence of PAA (2.15 mM) and H2O2 (0.65 mM). The range of iodide concentrations was chosen to include an average seawater iodide concentration (0.5 μM) as well as very high iodide concentrations (≤10.9 μM) to evaluate the effect of HOI on the quenching of H2O2. A complete quenching of H2O2 would result in the HOBr FP shown in Figure 3 (solid line). Figure 3 shows the evolutions of HOBr, H2O2 and PAA at pH 7. The results indicate that an increase in iodide from 0 to 0.5 μM increased the HOBr formation rate and also the HOBr concentration at 120 min (from 0.14 to 0.21 mM). At the same time, the H2O2 and PAA depletion rates increased (Figure 3b,c). This effect was further enhanced by increasing the iodide concentration to 10.9 μM. Under these conditions, the H2O2 concentration decreased to [PAA]. The kinetics and mechanisms of the reactions involved in bromate formation in PAA systems are currently being investigated in our laboratory. THMs and HAAs Formation during PAA Treatment of Brackish Waters. To further elucidate other potential sinks for
Figure 3. Effect of the iodide concentration on the evolution of (a) HOBr, (b) H2O2, and (c) PAA. Experimental conditions: [PAA]0 = 2.15 mM, [H2O2]0 = 0.65 mM, [Cl−]0 = 0.270 M, [Br−]0 = 0.46 mM, [I−] = 0−10.9 μM, [carbonate buffer] = 50 mM, pH 7. Experimental data from the HOBr formation potential (FP) experiment from Figure 1 for the HOBr and PAA residual concentrations were added for comparison but are represented as solid lines.
H2O2 loss and iodide is re-formed from this reaction (Scheme 1). These iodide-catalyzed reactions affect the PAA−Br−−H2O2 system significantly by consuming both PAA and H2O2. Because H2O2 is completely depleted but there is still residual PAA in the presence of high iodide concentrations, the HOBr formation approaches the HOBr FP curve. It was expected that the two data sets should match if plotted against the PAA exposure to compensate for differing PAA stability in the two experiments. A plot of the HOBr formation versus the PAA exposure showed a good match for the initial phase (≤85 mM· min) but deviated at higher PAA exposures (≤190 mM·min; Figure S7, SI). This is likely to be a direct consequence of the depletion of H2O2 at this point, which enables PAA to oxidize HOBr further to bromate (see below). This leads to lower HOBr formation for high iodide values as compared to the FP values. Overall, these results indicate that the iodide 1703
DOI: 10.1021/es503920n Environ. Sci. Technol. 2015, 49, 1698−1705
Article
Environmental Science & Technology
Industries AG and jointly by Nippon Yuka Kogyo Co. Ltd. and Katayama Chemical, Inc. (SKY-System) apply a maximum dose of 150 mg/L (2 mM) PERACLEAN OCEAN PAA in which H2O2 is present in excess (as discussed above). However, these systems differ in that the Evonik system adds excess catalase simultaneously with PAA to quench H2O2 and prevent its environmental discharge6 whereas the SKY-System adds sodium sulfite (Na2SO3) to quench H2O2, although this reaction is rather slow at circumneutral pH,33 but only immediately prior to treated water discharge.34 Differences were observed in the types and levels of THMs and HAAs formed during pilot-scale and full-scale testing of these two treatment systems. In both cases, brackish water was treated with 150 mg/L PERACLEAN OCEAN PAA, and THMs and HAAs formation was monitored.6,34 After a holding time of 5 days for the Evonik system, mainly brominated THMs and HAAs were formed (8.5 μg/L CHClBr2, 180 μg/L CHBr3, 23 μg/L MBAA, and 100 μg/L DBAA).6 These results are expected, given that catalase is added simultaneously with PAA and H2O2 and, thus, the H2O2 is immediately quenched and can no longer react with the resulting HOCl (eq 4) and HOBr (eq 5) that are formed when chloride and bromide, respectively, react with PAA (Table 1). This subsequently leaves HOCl and HOBr to react with DOM to generate mostly brominated THMs and HAAs. Alternatively, the same set of THMs and HAAs were monitored in the SKY-System for brackish water treated after 5 days in which no significant formation (
