Oxidation of Manganese (II) during Chlorination: Role of Bromide

Jul 16, 2013 - Curtin Water Quality Research Centre, Department of Chemistry, Curtin University, GPO Box U1987, Perth WA 6845, Australia. ‡. WA Orga...
0 downloads 0 Views 901KB Size
Download PDF
Article pubs.acs.org/est

Oxidation of Manganese(II) during Chlorination: Role of Bromide S. Allard,† L. Fouche,† J. Dick,‡ A. Heitz,† and U. von Gunten*,§,∥ †

Curtin Water Quality Research Centre, Department of Chemistry, Curtin University, GPO Box U1987, Perth WA 6845, Australia WA Organic and Isotope Geochemistry Centre, Department of Chemistry, Curtin University, GPO Box U1987, Perth WA 6845, Australia § Eawag, Swiss Federal Institute of Aquatic Science and Technology, ETH Zürich, Zürich, Switzerland ∥ School of Architecture, Civil and Environmental Engineering (ENAC), Ecole Polytechnique Fédérale Lausanne (EPFL), Switzerland ‡

S Supporting Information *

ABSTRACT: The oxidation of dissolved manganese(II) (Mn(II)) during chlorination is a relatively slow process which may lead to residual Mn(II) in treated drinking waters. Chemical Mn(II) oxidation is autocatalytic and consists of a homogeneous and a heterogeneous process; the oxidation of Mn(II) is mainly driven by the latter process. This study demonstrates that Mn(II) oxidation during chlorination is enhanced in bromide-containing waters by the formation of reactive bromine species (e.g., HOBr, BrCl, Br2O) from the oxidation of bromide by chlorine. During oxidation of Mn(II) by chlorine in bromide-containing waters, bromide is recycled and acts as a catalyst. For a chlorine dose of 1 mg/L and a bromide level as low as 10 μg/L, the oxidation of Mn(II) by reactive bromine species becomes the main pathway. It was demonstrated that the kinetics of the reaction are dominated by the adsorbed Mn(OH)2 species for both chlorine and bromine at circumneutral pH. Reactive bromine species such as Br2O and BrCl significantly influence the rate of manganese oxidation and may even outweigh the reactivity of HOBr. Reaction orders in [HOBr]tot were found to be 1.33 (±0.15) at pH 7.8 and increased to 1.97 (±0.17) at pH 8.2 consistent with an important contribution of Br2O which is second order in [HOBr]tot. These findings highlight the need to take bromide, and the subsequent reactive bromine species formed upon chlorination, into account to assess Mn(II) removal during water treatment with chlorine.

■

INTRODUCTION Manganese (Mn) concentrations can vary between sub mg/L to several mg/L in water resources under anoxic conditions.1−4 While Mn is an essential trace element in the human diet, in excess it is potentially toxic and may impair children’s intelligence.5 The World Health Organization postulates that a concentration Mn(II) (s)

(1)

(fast equilibrium) K (2)

>Mn(II) (s) + HOCl + H 2O → 2MnO2 (s) + Cl− + 3H+

(moderate)

(3)

>Mn(II) refers to adsorbed Mn(II) species, i.e., MnO2·Mn(II) (s). Dissolved manganese is initially slowly oxidized to MnO2. Thereafter, an autocatalytic process takes place. Mn(II) is quickly adsorbed onto the newly formed MnO2 and is oxidized at its surface. The homogeneous oxidation of Mn(II) with HOCl is the rate limiting step with a second order rate constant ([Mn(II)], [HOCl]) of approximately 6.4 × 10−4 M−1 s−1 at pH 8, whereas the heterogeneous oxidation is the major mechanism for dissolved manganese removal with a third order rate constant ([Mn(II)], [MnO2], [HOCl]) of 4 × 106 M−2 s−1 at pH 8.16 Hao et al.16 demonstrated that the efficiency of this autocatalytic process is strongly influenced by the composition of the raw water. The Mn(II) speciation is also an important parameter, as the various hydroxo species of Mn(II) (Mn(OH)+, Mn(OH)2), similarly to Fe(II) (Fe(OH)+, Fe(OH)2), have been shown to be orders of magnitude more reactive with oxygen than their respective free aqueous ions (Mn2+, Fe2+).19−23 Therefore, the kinetics of these reactions are highly pH dependent. It has been demonstrated that bromine is more reactive than chlorine toward organic compounds such as phenols,24,25 steroid estrogens,26 or more generally amines or phenolic moieties within dissolved organic matter (DOM).27−29 Some inorganic species such as ammonia and hydrogen peroxide are also oxidized more quickly by bromine30,31 than by chlorine.32,33 Concentrations of bromide (Br−) in natural waters usually range from 10 μg/L to several mg/L.34−36 During chlorination, bromide readily reacts with hypochlorous acid (HOCl) yielding bromine (HOBr + BrO−).37 HOCl + Br − → HOBr + Cl−

k = 1550 M−1 s−1

Figure 1. Speciation of chlorine, bromine, and dissolved manganese. pKaHOCl/OCl‑ = 7.5,37 pKaHOBr/OBr‑ = 8.8,39 pKaMn2+/Mn(OH)+ = 10.6,23 pKaMn(OH)+/Mn(OH)2 = 11.6,23 (I.S. = 0 M).

constants for the heterogeneous system were measured, and the relative contribution of chlorine and bromine to the heterogeneous Mn(II) oxidation was determined. The effect of the pHdependent speciation of Mn(II) and the oxidant on the Mn(II) oxidation rates was studied.

■

EXPERIMENTAL SECTION Reagents and Analytical Methods. Mn(ClO4)2, NaOH, HClO4, NaClO4, KBr, and Na2B4O7 were all reagent grade chemicals. All solutions were prepared using ultrapure water from an IBIS Technology Ion Exchange System followed by Elga Purelab Ultra System. Standard stock solutions of chlorine were prepared from a reagent grade sodium hypochlorite solution (10% active chlorine) and calibrated weekly by iodometric titration.40 Working solutions were prepared daily and calibrated with the DPD method.40 Bromine solutions were prepared by adding an excess of bromide (5−10%) to a chlorine solution at pH 6. The solution was stirred for 30 min to allow complete formation of HOBr. Mn(II) was analyzed spectrophotometrically using the formaldoxime method41 with a SHIMADZU UV Pharmaspec 1700 spectrophotometer. Colloidal MnO2 was analyzed at 350 nm18 using the same spectrophotometer. Experimental Procedures. All experiments were performed at room temperature (25 ± 1 °C). Two different setups were used depending on the kinetics of the experiments (Scheme S1 in the SI). For slow reactions (t1/2 ≥ 2 min for both homogeneous and heterogeneous processes), kinetic experiments were carried out in a 1 L beaker mounted on a balance. First the chlorine or bromine solution was added, and the desired pH was established by automatic addition of NaOH or HClO4 using an automated pH titrator (Metrohm 857 Titrando). The solution was vigorously mixed with a rod stirrer. The solutions were purged with nitrogen for 1 h prior to the start of all experiments to avoid oxidation with oxygen especially for experiments conducted in alkaline media,10 and the pH titrator was used to keep the pH constant. During all experiments, N2 was constantly bubbled through the solutions. During the oxidative process the pH was controlled by addition of NaOH with concentrations ranging from 0.01 to 0.1 M to ensure a minimal variation of the initial volume. The pH variation was always lower than 0.05 pH units. It was also verified that the concentration of the titrant solution did not change the kinetics of the reaction by producing localized

(4)

38

Melichova et al. showed that bromine oxidizes dissolved Mn(II) in an alkaline medium. An autocatalytic process analogous to that in the chlorine-Mn(II) system is expected with hypobromous acid (HOBr). A simplified pH-dependent speciation diagram of both oxidants and Mn(II) is presented in Figure 1. As illustrated for the oxidation of Mn(OH)+ by HOCl and HOBr in the enlarged portion of Figure 1, the region of coexistence between Mn(OH)+ and HOBr is larger than that for Mn(OH)+ and HOCl. Together with the higher reactivity of HOBr, it is expected that the kinetics of manganese oxidation in chlorinated drinking water can be affected by the presence of bromide. The objective of this study was to assess the role of bromide during chlorination of Mn(II)-containing waters. The mechanism of the bromine-Mn(II) reaction in both the homogeneous and heterogeneous system was investigated. Second order rate 8717

dx.doi.org/10.1021/es401304r | Environ. Sci. Technol. 2013, 47, 8716−8723

Environmental Science & Technology

Article

Figure 2. Mn(II) oxidation by (a) HOCl and (b) HOBr. Experimental conditions: pH 8.0, [Mn(II)] = 180 μM (10 mg/L), I.S. = 1.0 × 10−2 M, with (a) [HOCl]tot = 1.8 mM (95 mg/L) and (b) [HOBr]tot = 1.8 mM (175 mg/L).

[HOBr] < 1 × 10−4 M). The analytical methods described previously are not sensitive enough to quantify such low concentrations of Mn(II), and a 10 times excess of oxidant is needed to work under pseudo first order conditions. However, in terms of kinetics, molar ratios are more relevant than actual concentrations. Therefore, we expect that the outcomes will translate well to the lower concentration ranges encountered for typical drinking water conditions.

high pH values, which would increase the reactivity of Mn(II) (formation of hydroxo species) and decrease the reactivity of HOBr or HOCl. Aliquots were withdrawn periodically and filtered using 0.2 μm filters (hydrophilic polyethersulfone, Bulk Acrodisc, PALL Life Sciences). The filtrate was then quenched with an excess of sodium sulfite (at least 10 times the initial oxidant concentration) to remove the excess oxidant before analysis of dissolved manganese. It was verified that no oxidized manganese (colloids) passed through the filters. Samples were withdrawn and filtered after complete consumption of the oxidant. An aliquot of the filtrate was poured into a solution of ascorbic acid to reduce the colloidal MnO2. Mn(II) was analyzed in both quenched and unquenched samples, and no detectable manganese was found in the filtrate. It was not practical to manually withdraw, filter, and quench enough of the sample to accurately derive rate constants at alkaline pH using the slow kinetic setup, because the heterogeneous process was too fast for our experimental conditions. Therefore, for fast reactions (t1/2 ≤ 2 min for both homogeneous and heterogeneous processes), the formation of colloidal manganese oxide was followed using the time drive of the spectrophotometer, recording the absorbance at 350 nm.18 The oxidant and the Mn(II) solutions were prepared in two separate beakers containing the desired amount of borate buffer. As previously, N2 was bubbled through the solutions, and NaOH or HClO4 were used to adjust the pH. An equal volume of the oxidant and the Mn(II) solutions was introduced in the cuvette under vigorous mixing of the reactant before analysis by UV spectrophotometry. The use of buffer could not be avoided with this setup. Preliminary experiments were carried out with different buffer concentrations at several pHs and showed that for the fast kinetic experiments, the buffer concentration did not significantly affect the kinetics of Mn(II) oxidation (Figure S1). The experiments were carried out using 1 mM borate buffer to minimize pH deviation. This setup was not appropriate for slow reactions since colloidal MnO2 polymerized after a few minutes leading to precipitation in the cuvette. The ionic strength (I.S.) was maintained at 1.0 × 10−2 M with NaClO4 for all experiments. Particulate manganese was not characterized in this study since different manganese oxide phases are expected to be formed depending on the oxidation process. However, the primary products with relatively low crystallinity should have comparable effects on Mn(II) oxidation. The concentration of reactants used in this study (1 × 10−4 M < Mn(II) < 1.8 × 10−4 M and 1 × 10−3 M < HOCl/HOBr < 3.6 × 10−3 M) are higher than those encountered in real drinking waters (in most cases [Mn(II)] < 5 × 10−5 M and [HOCl] or

■

RESULTS AND DISCUSSION

Oxidation of Manganese(II): Autocatalysis. Mn(II) undergoes both homogeneous (lag period) and heterogeneous (acceleratory period) oxidation processes.10 It has been demonstrated that manganese oxidation by oxygen is an autocatalytic process (analogous to eqs 1−3).10 This catalysis phenomenon is induced by the adsorption of Mn(II) onto reactive sites of the freshly formed Mn(III,IV) oxides. Under pseudo first order conditions (excess of oxidant), the corresponding rate law for the oxidation of Mn(II) by oxygen was formulated as10 −

d[Mn(II)] = ko[Mn(II)] + kobs[Mn(II)][MnO2 ] dt

(5)

k0 and kobs are the homogeneous and heterogeneous rate constants for Mn(II) oxidation, respectively. Autocatalytic patterns for the oxidation of Mn(II) by HOCl and HOBr are shown in Figures 2a and b, respectively. For the chosen experimental conditions, a first step consisting of the slow oxidation of Mn(II) and the formation of Mn(III,IV) oxides (homogeneous oxidation) occurred in the first 20−25 min of the experiment for HOCl and in ∼1 min for HOBr in 10-fold excess of oxidant concentrations (1.8 mM). This was followed by a heterogeneous oxidation consisting of the formation of a surface complex and a faster oxidation of the adsorbed Mn(II). For example, under the experimental conditions of Figure 2a and b, ∼90% of the initial Mn(II) was oxidized in 15 min with HOCl, and in only ∼1 min with HOBr during the heterogeneous process while almost no change in the Mn(II) concentration was observed during the homogeneous process. This clearly shows that the oxidation of dissolved Mn(II) is governed by a heterogeneous oxidation process. This reaction is rather complex, but an approximation by an autocatalytic rate law seems reasonable to describe this process. After integration, eq 5 leads to eq 6 (see appendix in ref 42) 8718

dx.doi.org/10.1021/es401304r | Environ. Sci. Technol. 2013, 47, 8716−8723

Environmental Science & Technology log

Article

⎤ ⎡ [Mn(II)]0 kobs[Mn(II)]0 + log⎢ − 1⎥ ⎦ ⎣ [Mn(II)] k0

=

kobs[Mn(II)]0 t 2.3

oxidant concentrations, Mn(II) was completely oxidized in 1.5 h by chlorine and in only 2 min by bromine. The shapes of the curves in Figure 4a show autocatalytic behavior similar to the experiments where chlorine or bromine were used separately. Nevertheless, the chlorination process in the presence of bromide is somewhat more complicated. First, bromide is oxidized by HOCl to HOBr (reaction 4). Then, a two step process is expected when the bromide concentration is lower than the initial Mn(II) concentration. As shown in Figure 4a, there is a substantial and consistent increase in the rate of Mn(II) oxidation with increasing bromide concentration. Even for the highest Br−/HOCl ratio, the bromide concentration is present understoichiometrically relative to the Mn(II). For our experimental conditions, the fast oxidation of adsorbed Mn(II) by HOBr (reduction to bromide) competes with the slower oxidation of adsorbed Mn(II) by HOCl. The reaction should follow the HOCl kinetics, once bromide is used up if it would not be a catalytic process. Therefore, we can conclude that the released bromide is reoxidized by chlorine to form HOBr (eq 4). As a consequence, bromide is recycled and hence acts as a catalyst. Figure 4b shows the second order rate constant kobs determined by eq 6 as a function of the [Br−]/[HOCl] ratio. kobs increases linearly with increasing [Br−]/[HOCl] ratios or is proportional to the Br− concentration since HOCl was kept constant. If bromide was not recycled the heterogeneous oxidation would mainly be governed by the chlorine oxidation and the kobs value would only partially depend on the bromide concentration. This finding gives further evidence to the assumption of a Br−-catalyzed process. In previous studies, a catalytic effect of bromide has been observed for ammonia oxidation, N-nitrosodimethylamine formation from dimethylsulfamide during ozonation of bromide-containing waters and iodate formation during chlorination of iodide- and bromidecontaining waters.43−45 Figure 5 shows the contribution of chlorine and bromine to the rate of >Mn(II) oxidation for varying [Br−]/[HOCl] ratios (see Text S1 for explanation). Based on these results the oxidation of >Mn(II) by HOBr becomes the predominant pathway for [Br−]/[HOCl] ratios >0.009. For a typical chlorine concentration of 1 mgCl2/L in drinking water treatment, this corresponds to a bromide concentration of only ∼10 μg/L, which is at the lower end of Br− concentrations in natural waters (see above). These results highlight the need to take Br− into account to assess the rate of heterogeneous Mn(II) oxidation during chlorination of almost any natural water.

(6)

When the autocatalytic process predominates, a plot of log (([Mn(II)]0/[Mn(II)])−1) versus time gives a straight line as shown in Figure 3. An empirical second order rate constant kobs

Figure 3. Example of linearization of the kinetics of Mn(II) oxidation by chlorine based on eq 6 (raw data shown in Figure 2a). The black circles correspond to the phase of the heterogeneous oxidation (1500−2500 s), which was used to determine kobs.

for the heterogeneous part of the reaction can be computed from the slope of the linear portion of the curve. k0 could also be theoretically computed by the y-intercept but was not used in this study due to the large uncertainty of the results (high standard deviation). In water treatment, dissolved manganese is usually removed by oxidation and subsequent filtration. In this case manganese oxide is already present in the filter media, and heterogeneous oxidation would be the major mechanism for the entire duration of the oxidation. Catalytic Effect of Bromide and Mechanistic Consideration. To investigate the role of bromide during chlorination of Mn(II)-containing waters, the [Br−]/[HOCl] ratio was varied, ranging from 0 to 1. It was demonstrated that for the same experimental conditions ([Mn(II)] = 180 μM (10 mg/L), [HOX]/[Mn(II)] ratio =10 (HOX being HOCl or HOBr), pH 7.5, I.S. = 1.0 × 10−2 M) the presence of bromide significantly enhances the rate of Mn(II) oxidation (Figure 4a). For the same

Figure 4. Influence of the molar bromide/chlorine ratio on the manganese oxidation rate. (a) Mn(II) decrease as a function of the [Br−]/[HOCl] ratio, (b) second order rate constant kobs (eq 6) as a function of [Br−]/[HOCl] ratio. Experimental conditions: [HOX]tot = 1.8 mM, [Mn(II)] = 180 μM (10 mg/L), pH 7.5, I.S. = 1.0 × 10−2 M, 0 ≤ [Br−] ≤ 1.8 mM. 8719

dx.doi.org/10.1021/es401304r | Environ. Sci. Technol. 2013, 47, 8716−8723

Environmental Science & Technology

Article

Figure 5. Relative contribution of chlorine and bromine to the heterogeneous manganese oxidation rate as a function of the molar [Br−]/[HOCl] ratio at pH 7.5.

Kinetics of Heterogeneous Manganese(II) Oxidation by Chlorine and Bromine: Species-Specific Rate Constants. The influence of pH on Mn(II) oxidation kinetics was studied between pH 6 and 9.2 for chlorination and between pH 7.2 and 9.2 for bromination (pH-range relevant for drinking water). Experiments were carried out using both methods (i.e, the slow and fast kinetic setup), and similar results were obtained (Figure S2). However, only experiments performed with the fast kinetic setup were used for fitting. In the slow kinetic experiment, the range of kobs values as a function of pH was relatively small which resulted in a high uncertainty. As shown in Figure 6a, b, and c (data points), for the same oxidant/Mn(II) ratio for both chlorine and bromine, the rate of Mn(II) oxidation increases with increasing pH up to a maximum at pH ≈ 8.6 and then decreases for higher pH values. We postulate, that similar to the oxidation by oxygen,22,23 the pH dependency of kobs can be explained by the speciation of the adsorbed Mn(II) and the speciation of chlorine and bromine. Similar pH-dependencies have been observed for the reaction of inorganic and organic compounds with chlorine and bromine.46,47 The different adsorbed Mn(II) species, i.e. >Mn2+, >Mn(OH)+ and >Mn(OH)2, were considered to quantify the oxidation kinetics. In addition, even if the main chlorine or bromine species were HOCl/ClO− and HOBr/BrO−, respectively, other reactive haline species were considered because they may be important for the overall reactivity due to their high electrophilicity even if they are present at low concentrations. Recently, it was demonstrated that under certain water treatment conditions, Cl2, Cl2O and BrCl, Br2, Br2O, and BrOCl can influence the rate of the oxidation of certain organic compounds.48−50 Therefore, to better characterize the pHdependence of the reaction, both the speciation of adsorbed Mn(II) and of the oxidants has to be considered. The reactivities of BrO− and ClO− were assumed to be negligible because of the higher electrophilicity of HOX than XO−.51 As shown in Figure 2a and b when the autocatalytic process takes place, the heterogeneous oxidation clearly outweighs the homogeneous process. Therefore, during the heterogeneous process, the homogeneous oxidation could be neglected in the kinetic rate law, and, introducing the equilibrium constant K from eq 2 into eq 5 leads to eq 7.

Figure 6. Example of pH-dependent contribution of species specific rate constants to the apparent second order rate constants for >Mn(II) oxidation and ΔpKa = 1 during (a) chlorination: fitting with HOCl, Cl2, and Cl2O, k(HOCl/>Mn(OH)2) = kHOCl/>Mn(OH)2 α>Mn(OH)2[HOCl]; (b) bromination: fitting with HOBr only, k(HOBr/>Mn(OH)+) = k HO Br />M n( OH )+ α > M n ( O H) + [HOBr], k(HOBr/>Mn(OH) 2 ) = kHOBr/>Mn(OH)2 α>Mn(OH)2[HOBr]; and (c) bromination: fitting with HOBr, Br2, Br2O, and BrCl, k(BrCl/>Mn(OH)+) = kBrCl/>Mn(OH)+ α > M n ( O H ) + [BrC l], k( B r 2 O / > M n ( O H ) 2 ) = k B r 2 O / > M n ( O H ) 2 α>Mn(OH)2[Br2O], (k(HOBr) and k(Br2) are close to zero and not visible in the graph for all the adsorbed manganese species). Circles: experimental data; lines: model fits for species specific rate constants. kmodel HOCl, kmodel HOBr only, kmodel Br species correspond to the sum of each species specific rate constants. Experimental conditions: [Mn(II)] = 100 μM, [HOX]tot = 1.0 mM, I.S. = 1.0 × 10−2 M, 1.0 × 10−3 M borate buffer.

−

d[Mn(II)] = kobs[Mn(II)][MnO2 ] dt 1 = kobs[>Mn(II)]tot K 1 = [>Mn(II)]tot ∑ αik i K

(7)

αi represents the fraction of hydrolyzed >Mn(II) species: α>Mn 2+ = [>Mn 2 +]/[>Mn(II)]tot , α>Mn(OH)+ = [>Mn(OH)+ ]/[>Mn(II)]tot , α>Mn(OH)2 = [>Mn(OH)2 ]/[>Mn(II)]tot

with 8720

dx.doi.org/10.1021/es401304r | Environ. Sci. Technol. 2013, 47, 8716−8723

Environmental Science & Technology

Article

This diagram shows that the hydroxo species are shifted to lower pH lowering the fraction of the >Mn2+ species in the pH range studied. Initial model calculations showed that >Mn2+ was nonreactive. The contribution of >Mn2+ to the global rate of >Mn(II) oxidation did not significantly affect the model accuracy and was not further considered in the calculations. The best estimate for the chlorination experiments showed that HOCl was the sole oxidative reactive species and reacted with >Mn(OH)2 (Figure 6a). Similar results were found for the range of ΔpKa values studied. HOCl is more than 4 orders of magnitude higher in concentration than the other chlorine species (Figure S3) and >Mn(OH)2 is a stronger nucleophile than either >Mn2+ or >Mn(OH)+ and thus more reactive toward the electrophile chlorine. As a consequence of the dominant contribution of >Mn(OH)2 and HOCl to the overall oxidation rate, the rate constant for the monohydroxo species as well as for the chlorine species Cl2, and Cl2O could not be determined accurately. For the bromination experiments, several alternatives were considered. Results from calculations with HOBr as the sole oxidant are shown in Figure 6b (Bromine-1). The oxidation rate is dominated by the hydroxo species >Mn(OH)+ and >Mn(OH)2. Both >Mn(OH)+ and >Mn(OH)2 contribute significantly to the overall oxidation rates kmodel HOBr alone (Figure 6b). However, the statistical outputs, R2 value of 0.975 and the wide 95% confidence interval, indicate that the species-specific rate constants are not accurately determined. Inclusion of Br2, BrCl, and Br2O in the model (see eq 9) provides improved fits to the data for the pH range studied (Figure 6c, Bromine-2). The oxidation rate was dominated by the hydroxo species >Mn(OH)+ and >Mn(OH)2, and, interestingly, the main reactive species in this case seem to be Br2O and BrCl. Outputs from the fitting exercise for 0.5 < ΔpKa < 2 (Figure S6) showed that the couple HOBr/>Mn(OH)2 significantly contributed to the overall rate only for ΔpKa = 0.5. For all other trials, the contribution of HOBr to the overall rate of >Mn(II) oxidation was found to be negligible even though Br2O is present at much lower concentration (Figure S4). Therefore, results allow the hypothesis that the two couples BrCl/>Mn(OH)+ and Br2O/ >Mn(OH)2 are the main reactive species involved in this oxidation process. The reaction between BrCl and >Mn(OH)+ was found to govern the reactivity for pH Mn(OH)2 gives the main contribution for pH >7.5. In the pH range studied, the formation of Br2O is favored over BrCl and BrOCl because the experiments were carried out in excess of bromide. The much better fit obtained by inclusion of BrCl and Br2O is clearly apparent in Figure 6c (kmodel Br species) and is confirmed by the statistical parameters with an R2 value of 0.996 and a much narrower 95% confidence interval. These results emphasize the role of the latter oxidants and suggest that HOBr is of minor importance under the conditions studied. To further support our hypothesis, additional experiments were carried out with various bromine concentrations under pseudo first order conditions ([HOBr]tot/[Mn(II)]>10), to determine the reaction order in [HOBr]tot according to eq 10.

[>Mn(II)]tot = [>Mn 2 +] + [>Mn(OH)+ ] + [>Mn(OH)2 ]

ki represents species specific rate constants for the >Mn(II) speciation: k >Mn 2+ ,

k >Mn(OH)+ ,

k >Mn(OH)2

Each adsorbed Mn(II) species can potentially react with each halogen species; therefore, the following reactivity models were used to determine species specific rate constants: kobsHOCl =

kobsHOBr =

∑ αik i/Cl [Cl 2] + ∑ αik i/Cl O[Cl 2O] + ∑ αik i/HOCl[HOCl] 2

2

(8)

∑ αik i/Br [Br2] + ∑ αik i/Br O[Br2O] + ∑ αik i/BrCl[BrCl] + ∑ αik i/HOBr[HOBr] 2

2

(9)

BrOCl was not included in the model because the bromination experiments were carried out with an excess of bromide. However, it was shown to be an important species under typical water chlorination conditions i.e. [HOCl]tot ≫ [Br−].50 The chlorine and bromine speciation diagrams were constructed with ChemEQL V3.052 (see Table S1 for equations and Figures S3 and S4). Because of a lack of literature data for pKa values for the >Mn(II) species, it was not possible to derive a speciation diagram using ChemEQL V3.0 and relative >Mn(II) concentrations had to be determined by fitting. Similar to e.g. substituted phenols with electron-withdrawing substituents, their pKa values are expected to be lower than their dissolved analogues due to an electron-withdrawing effect of the bulk MnO2. A program was developed in the R software environment53 to retrieve values of the species specific rate constants (ki/ox in eqs 8 and 9) using a linear least-squares regression. Both sets of kobs data and oxidant speciation data were used as input (i.e., n = 27 for the chlorination experiments and n = 33 for the bromination experiments) (see Text S2 in SI for more information). The models were constrained so that all coefficients (rate constants) would be zero or positive (Penalized Constrained Least Squares: “pcls” function in R package “mgcv”). To reduce the degrees of freedom of the modeling, the difference between pK a >Mn(OH)+/>Mn(OH)2 (pKa2) and pKa >Mn2+/>Mn(OH)+ (pKa1) was fixed (ΔpKa = pKa2 − pKa1 = constant), so that only pKa1 was used as a variable. The program adjusts the pKa1 and pKa2, for a ΔpKa. For trial values of ΔpKa for >Mn(II), the Cl and Br models were individually fitted, then the combined R2 for both models was calculated. The pKa values of the >Mn(II) species were optimized by iterative adjustment, to maximize the combined R2 for both of the chlorine and bromine models. Examples of fitting are provided in Text S2 for 0.5 < ΔpKa < 2, and reasonable statistic outputs (0.9831 < combined R2 < 0.9927) were obtained for the range of ΔpKa’s studied. As expected, the pKa values of the adsorbed species obtained from fitting are lower than their dissolved analogues for all ΔpKa’s tested. Since the combined R2 decreases for increasing ΔpKa, no trial above ΔpKa = 2 was carried out. As an example, for ΔpKa = 1, the best fits were obtained for pKa values of 8.2 and 9.2 for the couples >Mn2+/ >Mn(OH)+ and >Mn(OH)+ />Mn(OH)2, respectively. Based on these pKa values a revised pH-dependent speciation diagram for the adsorbed Mn(II) species can be obtained (Figure S5).

kobsHOBr = kapp([HOBr]tot )n

(10)

n was calculated via linear regression of log kobsHOBr vs log [HOBr]tot (Figure S7). Unlike previous research, which reported a first order reaction rate with respect to the bromine concentration,38 n was found to be equal to 1.33 (±0.15) at pH 7.8 and increased to 1.97 (±0.17) at pH 8.2. These results 8721

dx.doi.org/10.1021/es401304r | Environ. Sci. Technol. 2013, 47, 8716−8723

Environmental Science & Technology

Article

suggest that first order and second order reactions occur simultaneously. First order oxidative species comprise HOBr, BrCl, or Br2; however, only Br2O can contribute to a second order dependence, because it is formed from 2 HOBr (Table S1). The close to second order dependence obtained for pH 8.2 supports the hypothesis that Br2O governs the reaction as shown in Figure 6c. The decrease of n with decreasing pH also underlines the trend observed in Figure 6c with a decreasing contribution of Br2O and increasing contribution of BrCl (first order dependence) to the overall rate of reaction. A comparison with previously published models is difficult since the modeling approaches are different. A mechanistic model (determination of species specific rate constants) was developed for data interpretation in the present study, while empirical fitting procedures were used in previous studies.16,38 Even though more data are needed to further evaluate speciesspecific rate constants, this study shows that HOBr is by far not the sole reactive bromine species contributing to >Mn(II) oxidation. Implications for Mn(II) Oxidation during Chlorination of Natural Waters. This study demonstrates that the oxidation of Mn(II) during chlorination is significantly enhanced in bromide-containing waters. Scheme 1 summarizes the possible

determination. Speciation diagrams of chlorine, bromine, and adsorbed manganese. Example of R-program, input and output. Example of pH-dependent contribution of species specific rate constants for 0.5 < ΔpKa < 2. Determination of the reaction order in [HOBr]tot. Influence of DOM on Mn(II) adsorption onto MnO2. This material is available free of charge via the Internet at http://pubs.acs.org.

■

Corresponding Author

*Phone: +41 58 765 5270. Fax: +41 58 765 5210. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Andrew Chan for his assistance in the laboratory and the Water Corporation of Western Australia for funding. We gratefully acknowledge J. Sivey for his assistance and his insightful comments on this work. We also acknowledge Saskia Zimmermann, Caroline Gachet Aquillon, and Michèlle Heeb for their contributions.

■

Scheme 1. Manganese(II) Oxidation during Chlorination of Bromide-Containing Water: Bromide-Catalysisa

a

REFERENCES

(1) Barceloux, D. G. Manganese. J. Toxicol., Clin. Toxicol. 1999, 37 (2), 293−307. (2) Knocke, W. R.; Van Benschoten, J. E.; Kearney, M. J.; Soborski, A. W.; Reckhow, D. A. Kinetics of manganese and iron oxidation by potassium permanganate and chlorine dioxide. J. - Am. Water Works Assoc. 1991, 83 (6), 80−87. (3) Knocke, W. R.; Occiano, S. C.; Hungate, R. Removal of soluble manganese by oxide-coated filter media - sorption rate and removal mechanism issues. J. - Am. Water Works Assoc. 1991, 83 (8), 64−69. (4) Hallberg, R. O.; Martinell, R. Vyredox − in situ purification of ground water. Ground Water 1976, 14, 88−93. (5) Bouchard, M. F.; Sauve, S.; Barbeau, B.; Legrand, M.; Brodeur, M.; Bouffard, T.; Limoges, E.; Bellinger, D. C.; Mergler, D. Intellectual impairment in school-age children exposed to manganese from drinking water. Environ. Health Perspect. 2011, 119 (1), 138−143. (6) WHO, Manganese in drinking-water, Background document for development of WHO Guidelines for Drinking-water Quality. Available at http://www.who.int/water_sanitation_health/dwq/chemicals/ manganese.pdf (accessed July 2, 2013). (7) Kohl, P. M.; Medlar, S. J. Occurrence of Manganese in Drinking Water and Manganese Control; Denver, AwwaRF and IWA: 2007. (8) Anonymous, Committee report - research needs for the treatment of iron and manganese. J. - Am. Water Works Assoc. 1987, 79, (9), 119122. (9) Sly, L. I.; Hodgkinson, M. C.; Arunpairojana, V. Deposition of manganese in a drinking water distribution system. Appl. Environ. Microb. 1990, 56 (3), 628−639. (10) Morgan, J. J. Chemistry of aqueous manganese II and IV; Harvard, Cambridge, 1964. (11) Kessick, M. A.; Morgan, J. J. Mechanism of autoxidation of manganese in aqueous solution. Environ. Sci. Technol. 1975, 9 (2), 157− 159. (12) Morgan, J. J.; Stumm, W. Colloid-chemical properties of manganese dioxide. J. Colloid Sci. 1964, 19 (4), 347−359. (13) Pankow, J. F.; Morgan, J. J. Kinetics for the aquatic environment. Environ. Sci. Technol. 1981, 15 (11), 1306−1313. (14) Carlson, K. H.; Knocke, W. R. Modeling manganese oxidation with KMnO4 for drinking water treatment. J. Environ. Eng. 1999, 125 (10), 892−896. (15) Van Benschoten, J. E.; Lin, W.; Knocke, W. R. Kinetic modeling of manganese(II) oxidation by chlorine dioxide and potassium permanganate. Environ. Sci. Technol. 1992, 26 (7), 1327−1333.

The possible interferences by reactions with DOM are also shown.

reactions involved in this Br-catalyzed process. Due to the regeneration of bromide and subsequently HOBr and/or other bromine-containing oxidants only a few μg/L Br− are necessary to greatly enhance Mn(II) oxidation during chlorination. The catalytic process induced by the regeneration of bromine may be terminated by the incorporation of bromide into dissolved organic matter (DOM). However, it was demonstrated that a large fraction of HOBr is reduced and leads to bromide when reacting with DOM.54,55 It was estimated that the reaction of HOBr with DOM yields about 25% of bromo-organic compounds and about 75% bromide.54 Furthermore, DOM can be adsorbed onto manganese oxide and compete with dissolved manganese for adsorption sites.56,57 Adsorption experiments were carried out with some DOM extracted from a surface water (Figure S8). These experiments show that DOM may reduce the adsorption capacity of Mn(II) on MnO2. This can partially inhibit the catalytic effect of MnO2 during heterogeneous oxidation.

■

AUTHOR INFORMATION

ASSOCIATED CONTENT

S Supporting Information *

Scheme of experimental setups. Influence of buffer on the Mn(II) adsorption/oxidation process. Calculation of the chlorine and bromine contribution to the rate of adsorbed Mn(II) oxidation. Comparison of experimental setup for kobs 8722

dx.doi.org/10.1021/es401304r | Environ. Sci. Technol. 2013, 47, 8716−8723

Environmental Science & Technology

Article

(16) Hao, O. J.; Davis, A. P.; Chang, P. H. Kinetics of manganese(II) oxidation with chlorine. J. Environ. Eng. 1991, 117 (3), 359−375. (17) Jacobsen, F.; Holcman, J.; Sehested, K. Oxidation of manganese(II) by ozone and reduction of manganese(III) by hydrogen peroxide in acidic solution. Int. J. Chem. Kinet. 1998, 30 (3), 207−214. (18) Reisz, E.; Leitzke, A.; Jarocki, A.; Irmscher, R.; Von Sonntag, C. Permanganate formation in the reactions of ozone with Mn(II): A mechanistic study. J. Water Supply: Res. Technol.–AQUA 2008, 57 (6), 451−454. (19) Millero, F. J.; Sotolongo, S.; Izaguirre, M. The oxidation kinetics of Fe(II) in seawater. Geochim. Cosmochim. Acta 1987, 51 (4), 793−801. (20) Millero, F. J. The effect of ionic interactions on the oxidation of metals in natural waters. Geochim. Cosmochim. Acta 1985, 49 (2), 547− 553. (21) King, W. D. Role of carbonate speciation on the oxidation rate of Fe(II) in aquatic systems. Environ. Sci. Technol. 1998, 32 (19), 2997− 3003. (22) Rosso, K. M.; Morgan, J. J. Outer-sphere electron transfer kinetics of metal ion oxidation by molecular oxygen. Geochim. Cosmochim. Acta 2002, 66 (24), 4223−4233. (23) Morgan, J. J. Kinetics of reaction between O2 and Mn(II) species in aqueous solutions. Geochim. Cosmochim. Acta 2005, 69 (1), 35−48. (24) Acero, J. L.; Piriou, P.; Von Gunten, U. Kinetics and mechanisms of formation of bromophenols during drinking water chlorination: Assessment of taste and odor development. Water Res. 2005, 39 (13), 2979−2993. (25) Guo, G.; Lin, F. The bromination kinetics of phenolic compounds in aqueous solution. J. Hazard. Mater. 2009, 170 (2−3), 645−651. (26) Lee, Y.; Von Gunten, U. Transformation of 17α-ethinylestradiol during water chlorination: Effects of bromide on kinetics, products, and transformation pathways. Environ. Sci. Technol. 2009, 43 (2), 480−487. (27) Westerhoff, P.; Chao, P.; Mash, H. Reactivity of natural organic matter with aqueous chlorine and bromine. Water Res. 2004, 38 (6), 1502−1513. (28) Pattison, D. I.; Davies, M. J. Kinetic analysis of the reactions of hypobromous acid with protein components: implications for cellular damage and use of 3-bromotyrosine as a marker of oxidative stress. Biochemistry 2004, 43 (16), 4799−4809. (29) Echigo, S.; Minear, R. A. Kinetics of the reaction of hypobromous acid and organic matters in water treatment processes. Water Sci. Technol. 2006, 53, 235−243. (30) Wajon, J. E.; Morris, J. C. Rates of formation of N-bromo amines in aqueous solution. Inorg. Chem. 1982, 21 (12), 4258−4263. (31) von Gunten, U.; Oliveras, Y. Kinetics of the reaction between hydrogen peroxide and hypobromous acid: Implication on water treatment and natural systems. Water Res. 1997, 31 (4), 900−906. (32) Weil, I.; Morris, J. C. Kinetic studies on the chloramines. I. The rates of formation of monochloramine, N-chlormethylamine and Nchlordimethylamine. J. Am. Chem. Soc. 1949, 71 (5), 1664−1671. (33) Held, A. M.; Halko, D. J.; Hurst, J. K. Mechanisms of chlorine oxidation of hydrogen peroxide. J. Am. Chem. Soc. 1978, 100 (18), 5732−5740. (34) von Gunten, U.; Hoigne, J. Bromate formation during ozonation of bromide-containing waters: Interaction of ozone and hydroxyl radical reactions. Environ. Sci. Technol. 1994, 28 (7), 1234−1242. (35) D’Alessandro, W.; Bellomo, S.; Parello, F.; Brusca, L.; Longo, M. Survey on fluoride, bromide and chloride contents in public drinking water supplies in Sicily (Italy). Environ. Monit. Assess. 2008, 145 (1−3), 303−313. (36) Magazinovic, R. S.; Nicholson, B. C.; Mulcahy, D. E.; Davey, D. E. Bromide levels in natural waters: Its relationship to levels of both chloride and total dissolved solids and the implications for water treatment. Chemosphere 2004, 57 (4), 329−335. (37) Kumar, K.; Margerum, D. W. Kinetics and mechanism of generalacid-assisted oxidation of bromide by hypochlorite and hypochlorous acid. Inorg. Chem. 1987, 26 (16), 2706−2711. (38) Melichova, Z.; Treindl, L.; Valent, I. Kinetics and mechanism of the autocatalytic oxidation of Mn(II) by bromine. React. Kinet. Catal. Lett. 2001, 74 (1), 79−86.

(39) Haag, W. R.; Hoigne, J. Ozonation of bromide-containing waters: Kinetics of formation of hypobromous acid and bromate. Environ. Sci. Technol. 1983, 17 (5), 261−267. (40) APHA, A. WPaCF Standard Methods for the Examination of Water and Wastewater, 20th ed.; WA, 1998. (41) Chiswell, B.; O’Halloran, K. R. Comparison of three colorimetric methods for the determination of manganese in freshwaters. Talanta 1991, 38 (6), 641−647. (42) Wilson, D. E. Surface and complexation effects on the rate of Mn(II) oxidation in natural waters. Geochim. Cosmochim. Acta 1980, 44 (9), 1311−1317. (43) Haag, W. R.; Hoigne, J. Kinetics and products of the reactions of ozone with various forms of chlorine and bromine in water. Ozone: Sci. Eng. 1984, 6 (2), 103−114. (44) von Gunten, U.; Salhi, E.; Schmidt, C. K.; Arnold, W. A. Kinetics and mechanisms of N -nitrosodimethylamine formation upon ozonation of N,N-dimethylsulfamide-containing waters: Bromide catalysis. Environ. Sci. Technol. 2010, 44 (15), 5762−5768. (45) Criquet, J.; Allard, S.; Salhi, E.; Joll, C. A.; Heitz, A.; von Gunten, U. Iodate and iodo-trihalomethane formation during chlorination of iodide-containing waters: Role of bromide. Environ. Sci. Technol. 2012, 46 (13), 7350−7357. (46) Deborde, M.; von Gunten, U. Reactions of chlorine with inorganic and organic compounds during water treatment-Kinetics and mechanisms: A critical review. Water Res. 2008, 42 (1−2), 13−51. (47) Heeb, M.; Criquet, J.; Zimmermann, S.; von Gunten, U. Bromine production during oxidative water treatment of bromide containing waters and its reactions with inorganic and organic compounds: A critical review. Submitted to Water Res. (48) Sivey, J. D.; Roberts, A. L. Assessing the reactivity of free chlorine constituents Cl2, Cl2O, and HOCl toward aromatic ethers. Environ. Sci. Technol. 2012, 46 (4), 2141−2147. (49) Sivey, J. D.; McCullough, C. E.; Roberts, A. L. Chlorine monoxide (Cl2O) and molecular chlorine (Cl2) as active chlorinating agents in reaction of dimethenamid with aqueous free chlorine. Environ. Sci. Technol. 2010, 44 (9), 3357−3362. (50) Sivey, J. D.; Arey, J. S.; Tentscher, P. R.; Roberts, A. L. Reactivity of BrCl, Br2, BrOCl, Br2O, and HOBr toward dimethenamid in solutions of bromide + aqueous free chlorine. Environ. Sci. Technol. 2013, 47 (3), 1330−1338. (51) Schwarzenbach, R. P.; Gschwend, P. M.; Imbonden, D. M., Environmental Organic Chemistry, 2nd ed.; Wiley: New York, 2003. (52) Muller, B. ChemEQL V3.0, a program to calculate chemical speciation equilibria, Swiss Fed. Inst. of Aquat. Sci. and Technol., Kastanienbaum, Switzerland, 2004. Available at http://www.eawag.ch/ research_e/surf/Researchgroups/sensors_and_analytic/chemeql.html (accessed July 2, 2013). (53) R core team, R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, Austria, 2012. Available at http://www.R-project.org (accessed July 2, 2013). (54) Westerhoff, P.; Song, R.; Amy, G.; Minear, R. Numerical kinetic models for bromide oxidation to bromine and bromate. Water Res. 1998, 32 (5), 1687−1699. (55) von Gunten, U. Ozonation of drinking water: Part II. Disinfection and by-product formation in presence of bromide, iodide or chlorine. Water Res. 2003, 37 (7), 1469−1487. (56) Tipping, E.; Heaton, M. J. The adsorption of aquatic humic substances by two oxides of manganese. Geochim. Cosmochim. Acta 1983, 47 (8), 1393. (57) Bernard, S.; Chazal, P.; Mazet, M. Removal of organic compounds by adsorption on pyrolusite (δ-MnO2). Water Res. 1997, 31 (5), 1216.

8723

dx.doi.org/10.1021/es401304r | Environ. Sci. Technol. 2013, 47, 8716−8723