Environ. Sci. Technol. 2001, 35, 2525-2531
Bromate Minimization during Ozonation: Mechanistic Considerations ULRICH PINKERNELL† AND URS VON GUNTEN* Swiss Federal Institute for Environmental Science and Technology (EAWAG), U ¨ berlandstrasse 133, CH-8600 Du ¨ bendorf, Switzerland
Bromate formation during ozonation of bromide-containing natural waters is somewhat inversely connected to the ozone characteristics: an initial fast increase followed by a slower formation rate. During the initial phase mostly OH radical reactions contribute to bromate formation, whereas in the secondary phase both ozone and OH radicals are important. To minimize bromate formation several control options are presented: ammonia addition, pH depression, OH radical scavenging, and scavenging or reduction of hypobromous acid (HOBr) by organic compounds. Only the two first options are applicable in drinking water treatment. By both methods a similar effect of a bromate reduction of approximately 50% can be achieved. However, bromate formation during the initial phase of the ozonation cannot be influenced by either method. Ammonia (NH3) efficiently scavenges HOBr to NH2Br. However, this reaction is reversible which leads to higher required NH3 concentrations than expected. The rate constant kNH2Br for the hydrolysis of NH2Br by OH- to NH3 and OBr- was found to be 7.5‚106 M-1 s-1. pH depression shifts the HOBr/ OBr- equilibrium to HOBr and also affects the ozone chemistry. The effect on ozone chemistry was found to be more important for bromate formation. For a given ozone exposure, the OH radical exposure decreases with decreasing pH. Therefore, for pH depression the overall oxidation capacity for a certain ozone exposure decreases which in turn leads to a smaller bromate formation.
a simplified mechanism for bromate formation by ozone and OH radicals (•OH) is shown. Table 1 shows a detailed list of the involved chemical reactions. Bromate formation occurs through a multistep oxidation of bromide by either ozone or OH radicals or a combination thereof. It is shown in Figure 1 that HOBr is either formed directly by oxidation of bromide (Br-) by ozone or by conversion of Br•, which is formed through oxidation of Br- by •OH. HOBr/OBr- can be oxidized to bromate by O3 or •OH or react with ammonia. Another possible reaction pathway to bromate includes the oxidation of the Br• radical by ozone to BrO• which disproportionates to bromite (BrO2-), which is quickly oxidized to bromate by ozone (7). To describe bromate formation, in natural waters, the oxidation system has to be fully quantified with respect to the concentrations of ozone and hydroxyl radicals. This information can be gained by a procedure, which was described in detail previously by the Rct concept (9). The Rct is defined as the ratio of the exposures [The exposures O3-ct and •OH-ct are defined as the integral of concentration vs time plots. The integral of the ozone concentration vs time plot corresponds to the area under this curve and can be calculated by commericially availability graphites computer programs.] of OH radicals and ozone. It can be determined by simultaneously measuring the decrease of an added ozone-resistant probe compound for OH radicals and the ozone concentration. In our systems p-chlorobenzoic acid (pCBA) was added as a probe compound. The decrease of pCBA can be described as (9)
(
ln
Introduction The formation of bromate during drinking water ozonation has been intensively studied since the early 1990s when bromate (BrO3-) was classified as potentially carcinogenic by the IARC (International Agency for the Research on Cancer) (1). Later, the maximum contaminant level for bromate in drinking waters has been set to 10 µg/L both in the United States and in the European Union (2, 3). To respect these new regulations, the ozonation processes have to be optimized in many cases to take into account bromate formation while disinfection is still guaranteed (4, 5). As a result, more refined prediction of BrO3- formation by kinetically based models is needed. Therefore, a detailed mechanistic understanding as described before for standard ozonation and advanced oxidation processes is necessary (6-8). We have shown previously that hypobromous acid (HOBr) is a key intermediate for bromate formation (6). In Figure 1 * Corresponding author phone: +41 1 823 52 70; fax: + 41 1 823 5210; e-mail:
[email protected]. † Present address: Metrohm AG, CH-9101 Herisau, Switzerland. 10.1021/es001502f CCC: $20.00 Published on Web 05/12/2001
FIGURE 1. Reaction scheme for bromate formation during ozonation of bromide-containing waters.
2001 American Chemical Society
)
∫
[pCBA] ) -Rct‚k°OH,pCBA‚ O3 dt [pCBA]0
(1)
If the logarithmic decrease of pCBA is plotted versus the ozone exposure, the Rct can be calculated from the slope of eq 1 (kOH,pCBA ) 5‚109 M-1 s-1 (10)). The influence of the measured ratio of the •OH and ozone exposures, Rct, on bromate formation was studied in the present work. In former studies, experiments were performed with high bromide concentrations (approximately 10 times higher than those found in natural waters) due to the lack of sensitive methods for bromate analysis. In the present study, improved analytical methods for bromate allowed to perform laboratory scale experiments with natural waters containing low bromide concentrations. The objective of our investigation was to elucidate several control options for bromate minimization (pH depression, ammonia addition, addition of an OH radical scavenger, HOBr scavenging) in connection to their effect on oxidant concentrations. These concentrations were then used to model bromate formation with a kinetically based model. VOL. 35, NO. 12, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2525
TABLE 1. Bromate Formation during Ozonation: Individual Reactions no.
reaction
k+,k- or K
(4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22)
Br- + O3 f OBr- + O2 Br- + •OH f Br• + OHHOBr + O3 f BrO2- + O2 + H+ OBr- + O3 f BrO2- + O2 OBr- + O3 f Br- + 2O2 HOBr + •OH f BrO• + H2O OBr- + •OH f BrO• + OHHOBr ) OBr- + H+ Br• + O3 f BrO• + O2 Br• + Br- ff HOBr BrO• + BrO• f BrO2- + OBrBrO2- + O3 f BrO3- + O2 NH4+ ) NH3 + H+ •OH + pCBA f products Br• + NOM f BrHOBr + NH3 f NH2Br + H2O NH2Br + OH- f OBr- + NH3 2NH2Br f NHBr2 + NH3 NHBr2 + NH3 f 2NH2Br
160 M-1 s-1 1.1‚109 M-1 s-1 0.01 M1 s-1 100 M-1 s-1 330 M-1 s-1 2‚109 M-1 s-1 4.5‚109 M-1 s-1 1.26‚10-9 M 1.5‚108 M-1 s-1 ∼108 M-1 s-1 5‚109 M-1 s-1 105 M-1 s-1 5‚10-10 M 5‚109 M-1 s-1 109 M-1 s-1 7.5‚107 M-1 s-1 7.5‚106 M-1 s-1 250 M-1 s-1 100 M-1 s-1
model ref
a a a a a a a a a a a a a a a a a
(28) (29) (28) (28) (28) (18) (18) (28) (6 ) (6 ) (18) (6 ) (10) b (23) c c c
a Reactions used for the model calculations. b Fitted by model calculations for bromate formation with [NOM] ) 2‚10-6 M. c Fitted by model calculations.
Experimental Section Reagents. All reagents were analytical grade. Ozone stock solutions of approximately 1.2 mM ozone were produced by passing O3-containing oxygen through ice-cooled redistilled water (11). Titrisol buffers (Merck) were used to calibrate the pH electrode (Ross, ATI Orion, Boston, MA). Two different types of natural raw waters were collected for ozonation experiments and stored at 4 °C after filtration through 0.45 µm filters: (i) water from Lake Zu ¨ rich (Zu ¨ rich, Switzerland, DOC ) 1.3 mg/L, alkalinity ) 2.4 mM) and (ii) water from River Seine after rapid sand filtration (Maisons-Laffitte, France, DOC ) 2.4 mg/L, alkalinity ) 3.9 mM). For some experiments, the water was buffered with 10 mM borate buffer. The pH of all waters was adjusted to the desired value by adding dilute sulfuric acid or sodium hydroxide. All water samples were spiked with 0.5 µM pCBA before ozonation. Water parameters such as pH, temperature, ozone dose, and concentration of bromide and ammonia are given in the results and discussion section. Experimental Setup. Batch type ozonation experiments were performed by injecting small volumes of the ozone stock solutions to 500 mL of the prepared water sample in a closed bottle equipped with a dispenser system (6). The bottle was kept in a thermostated water bath to maintain a constant temperature. Samples of 10 mL each for analysis of ozone and hypobromous acid were taken and were added to 1 mL acidified solutions of indigo and phenol, respectively, to quench the reaction. Samples for ion chromatographic analysis of bromide and bromate were quenched using indigo without phosphoric acid to avoid interference during the chromatographic separation. Rate Constants for HOBr. Kinetic tests for determination of rate constants for the reaction of hypobromous acid and some organic compounds were performed under pseudofirst-order conditions with respect to HOBr. The decrease of HOBr (1 µM-0.1 mM) was measured in excess of the substrate (at least 10-fold) in a 50 mM phosphate buffer (pH 2-6) at 20 °C. HOBr analysis was based on quantitative reaction with ABTS (2,2-azino-bis(3-ethylbenzothiazoline)-6-sulfonic acid - diammonium salt) forming a intensively green colored stable product ABTS•+ (12). Samples were taken with a dispenser system and added to a ready made mixture of ABTS and sulfuric acid and analyzed spectrophotometrically at 405 nm in 1-10 cm cuvettes after 10 min. This technique 2526
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 12, 2001
is suitable for reactions with half-lives t1/2 > 20 s. In Table 2 the investigated organic compounds are shown together with the results from kinetic measurements. Analyses. Stock solutions of ozone were analyzed by their UV absorbance (258 nm ) 3000 M-1 cm-1). Samples from ozonation experiments were analyzed as follows: ozone was quantified using the Indigo method (11). pCBA was determined by HPLC according to ref 9 (column: Licrospher 100, RP 18, 5 µm, eluent: 55% methanol and 45% 10 mM phosphoric acid at 1 mL/min, UV-detection at 234 nm). With a 200 µL injection, the detection limit was 0.025 µM. Bromide and bromate were analyzed simultaneously without further sample pretreatment using a sensitive ion chromatographic method with postcolumn reaction, with a detection limit of 0.1 µg/L bromate and 3 µg/L bromide (13). For bromine mass balance experiments, HOBr was determined after quenching the reaction by adding 0.5 mM phenol at pH 3. HOBr and other reactive bromine species are scavenged by phenol forming 2- and 4-bromophenol. These species are stable and were analyzed by HPLC (column: Merck Licrospher 100, RP 18, 5 µm; eluent: 50% of a 0.2% acetic acid and 50% methanol, 1 mL/min; sample volume: 10 µL; UV-detection at 225 nm, detection limit: 0.01 µM of each bromophenol). Ammonia concentrations were determined according to the colorimetric indophenol blue method adapted from (14). Phenol, sodium nitroprusside, and sodium hypochlorite were used as reagents in an alkaline citrate buffer. A manual standard addition procedure was used, which has been adapted to the different ammonia concentrations. The detection limit was 1 µg/L ammonia using a 10 cm cuvette for spectrophotometry. Computer Simulations. Computer simulations of reaction kinetics were performed using the program AQUASIM (15). It solves complicated systems of coupled chemical reactions in combination with the simulation of hydraulic systems. The simulation of bromate formation was based on the main reactions and the corresponding set of rate constants. Details are given in the text. Ozone concentration profiles from experiments were fitted by simulations. The hydroxyl radical concentration profiles were calculated by the program using the ozone concentrations and the Rct values, which were also determined during the experiments.
Results and Discussion Bromate Formation Mechanism: Role of Rct. To reduce bromate formation it is important to know the relative importance of ozone and OH radical reactions for two reasons: (i) OH radical reactions are difficult to influence by addition of chemicals (e.g. addition of scavenger) and (ii) pH depression, a control option for bromate lowers the ratio Rct of •OH exposure and ozone exposure (16). For an optimum modification of the ozonation process it is therefore necessary to evaluate the impact of the Rct on bromate formation. The influence of temperature, pH, alkalinity, and type and concentration of the dissolved organic matter (DOM) on the Rct was described previously (16). The Rct (ratio [•OH]/[O3]) is mainly controlled by these natural water parameters, which remain almost unchanged during an ozonation process. Usually, two phases are observed, an initial phase during which the ratio Rct is not constant and a secondary phase during which Rct remains fairly constant. In Figure 2, two typical profiles of the Rct as a function of the ozone exposure are shown. They were derived from batch ozonation experiments performed with water from Lake Zu ¨ rich and from River Seine. For both waters we observed a fast initial decrease followed by a constant value. The Rct values were determined by the graph shown in the inset of Figure 2. It shows the logarithmic decrease of pCBA versus the ozone exposure, which allows the direct calculation of the Rct from the slope according to eq 1 (9).
TABLE 2. Rate Constants and Calculated Pseudo-First-Order Reaction Rates for the Reaction of HOBr and O3 with Some Organic Compounds no. (23) (24) (25) (26) (27) (28) (29) (30) (31) (32)
formaldehyde benzene p-xylene phenol 2-butanonea cyclohexanonea oxalic acid malonic acid formic acid acetic acid
k′HOBr
kHOBr this study
organic compd
1.8‚10-3
M-1 s-1
1.8‚10-9
< 0.01 M-1 s-1 0.2 M-1 s-1 500 M-1 s-1 n.d. n.d. 40 M-1 s-1 30 M-1 s-1
