Grantee Research Project Results

Final Report: Kinetic-Based Models for Bromate Formation in Natural Waters

EPA Grant Number: R826835
Title: Kinetic-Based Models for Bromate Formation in Natural Waters
Investigators: Westerhoff, Paul
Institution: Arizona State University
EPA Project Officer: Hahn, Intaek
Project Period: September 1, 1998 through August 31, 2000
Project Amount: $99,500
RFA: Drinking Water (1998) RFA Text |  Recipients Lists
Research Category: Drinking Water , Water

Objective:

The objective of this research project was to develop an understanding of bromate formation and control in natural waters during ozonation. We will conduct detailed experiments on a single water source, but answers to key questions will be transferable to nearly all water treatment plants practicing ozonation. The single water supply will be Colorado River Water (CRW) because it is a major water supply for six southwestern states, and many water treatment plants using CRW are, or will be, practicing ozonation. The central hypothesis is that a kinetic-based understanding of, natural organic matter (NOM) reactions with hydroxyl radicals (OH^•) and Br (HOBr/OBr¯) over a range of temperatures is necessary to develop mechanistic models for bromate formation in bulk waters. The research study: (1) developed a comprehensive database of bromate concentrations, ozone residuals, and OH^• concentrations; (2) determined rates of reaction between Br species (HOBr and OBr¯) with NOM; (3) calibrated and verified a bromate formation mechanistic-based model that includes, dissolved organic carbon (DOC) reactions and temperature dependency; and (4) simulated bromate control measures necessary to meet proposed and future, maximum contaminant levels (MCLs) under a wide range of conditions (e.g., pH depression and ammonia addition).

Ozonation currently is considered one of the most effective microbial disinfection technologies (Finch, et al., 1994). The planning, design, and use of ozone in water treatment facilities has been growing exponentially since the early 1980s (Rice, 1997). Ozonation byproducts are a major concern with respect to pending regulations. Regulations and technologies are constantly changing, and the drinking water industry requires robust and flexible tools to balance microbial protection against byproduct formation. Mechanistic-based models have been widely used to understand and predict coagulation, sedimentation, filtration, and membrane separation processes, while helping researchers understand these complex processes. It is now possible to develop mechanistic-based models for ozonation byproduct formation by using data collected over the last 75 years from atmospheric and aquatic studies of ozone reactions. However, additional knowledge of NOM reactions with key intermediates and simultaneous quantification of two oxidants present during ozonation (molecular ozone and hydroxyl radicals).

Recent regulatory activities have resulted in the development of the Microbial and Disinfection Byproduct (M/DBP) cluster. The components of this cluster include the proposed Disinfectant/Disinfectant Byproducts (D/DBP) Rule (U.S. Environmental Protection Agency, 1994), the Interim Enhanced Surface Water Treatment Rule (IESWTR), and requirements to provide microbial protection. Included is a proposed regulation to set the MCL for bromate at 10 µg/L. The MCL will be based on a Practical Quantification Limit (PQL) using EPA Method 300.0 (Pfaff, et al., 1989; Gordon, et al., 1994). Future regulations for bromate may be lowered because 10 µg/L currently represents a 10^-4 excess cancer risk level and is higher than the normal regulatory level of 10^-6 (Wilbourn, 1993; Kurokawa, et al., 1992; Song, et al., 1996a).

Based on work in a organic-free laboratory on water and waters containing NOM, it is generally accepted that bromate formation occurs in three steps: (1) molecular ozone (O3) oxidizes bromide (Br¯) to aqueous bromine (HOBr/OBr¯); (2) OH· present during ozonation oxidize Br to bromite (BrO2¯); and (3) O3 rapidly oxidizes bromite to bromate (BrO3¯). NOM is involved in bromate formation in several ways: (1) affects ozone demand of water and rate of ozone decay; (2) affects OH^• concentrations; (3) reacts with HOBr/OBr¯; and (4) probably reacts with brominated radicals. Two widely accepted strategies for controlling bromate formation include pH depression or ammonia addition. Our understanding has been limited as to the efficacy of these two bromate control strategies in natural waters that contain NOM, because NOM affects the mechanisms of bromate formation.

Laboratory kinetic batch and continuous-flow, plus pilot-scale, ozonation was conducted with a single source water (CRW) to study bromate formation. Bromide, aqueous bromine (HOBr and OBr¯), bromate, and ozone residual were directly measured. OH^• concentrations were assessed indirectly through the use of a probe compound (parachlorobenzoic acid). Experiments were conducted at different initial pH, bromide, alkalinity, ammonia, and ozone dose conditions.

Additional batch experiments using bromine or chlorine separately also were undertaken to evaluate rate constants for the reaction between these compounds and NOM. Isolated NOM was used in these studies, from the Colorado and Suwannee Rivers. Limited experiments were conducted using preozonated NOM isolates. Experiments were conducted at different initial pH and temperature conditions.

Summary/Accomplishments (Outputs/Outcomes):

Ozonation of Colorado River Water: Ozone and Hydroxyl Radicals. In general, effects of individual water quality parameter on O3 exposure, RCT, as well as OH^• exposure, show trends that most water quality parameters have a negative impact on O3 exposure, except alkalinity (see Table 1). Water quality parameters such as pH, temperature, and DOC have a positive influence on RCT and a negative impact for bromide and alkalinity. These results in the OH^• exposure increase as water pH, temperature, and DOC increase, but decrease as bromide and alkalinity increase.

The value of RCT is not affected by water hydraulics when applying the same water quality and ozone treatment conditions. RCT remains the same in both bench-scale batch and pilot continuous ozonation processes using the same water quality and treatment processes (e.g., applying the same O3 dose to obtain the same O3 residuals) for waters with and without ammonia addition.

 

Table 1. Trends of O3 Exposure, RCT, and OH^• Exposure on Water Quality Parameters During Ozonation

Parameters	Level	O3-CT	RCT	HO•-CT
PH
Temperature
Bromide
Alkalinity
DOC

The hydraulics do not seem to affect the ratio of OH^• and O3 concentrations throughout the process; however, this only applies for the slow stage of ozonation. The fast stage of RCT for continuous flow ozonation cannot be determined due to unavailability of O3 residual and OH^• concentration profiles in the process. When water quality is changed as a result of coagulants addition, the water pH decreases and the RCT decreases. The difference of RCT in batch and pilot ozonation of water with coagulant addition is due to the difference of O3 residual profiles. Overall, O3 dose is probably key for controlling RCT because it is a function of the required O3 residual, the ozone key kinetics.

In this study, bench-scale batch, continuous flow, and pilot-scale kinetic ozonation varied the reaction. Individual parameters for examining the effect of various water quality conditions and treatment variables on O3 and OH· concentrations were investigated. The following conclusions were withdrawn from the results:

• By adding a trace amount of parachlorobenzoic acid (PCBA) into water, the concentrations of OH^• formation during ozonation process can be determined by the product of O3 concentrations and RCT.

• Ozonation of water with a set of water quality and treatment conditions, a two-stage RCT was defined, which is a fast stage followed by a slow stage.

• Fits of RCT data at variable pH results in the following regressions:
RCT (0-2 min): 3E-14*(pH)^6.5486
RCT (>2 min): 1E-13*(pH)^5.4762

• The values of RCT are affected by water quality parameters. In general, RCT values at fast stage are higher than at slow stage by a factor of two to three, and the values of RCT increase as water pH, temperature, and concentration of DOC increase, and decrease as bromide and alkalinity levels decrease.

Ozonation of Colorado River Water: Bromate Formation. Kinetics of O3 decomposition and OH^• generation, as well as bromate formation during ozonation process, are governed by water quality characteristics and treatment conditions. Among those parameters, ozone dose, initial bromide concentration, pH, temperature, and alkalinity have a positive effect on bromate formation, whereas DOC concentration and ammonia have an inverse effect on bromate formation. RCT is a ratio of OH^• exposure (or concentration) to O3 exposure (or concentration) during the ozonation process, and remains unchanged for a given water under fixed conditions. In addition, two RCT values (fast and slow) observed in the ozonation process are influenced by water quality characteristics and treatment conditions as well. Values of RCT show a linear relationship with bromate formation. In general, RCT values increase with increasing levels of pH, temperature, and DOC and increasing rate of bromate formation; however, DOC has an inverse impact on bromate formation (see Table 2). On the other hand, RCT values decrease as bromide and inorganic carbon concentrations increase, resulting in an increase of bromate formation. Ammonia acts as a treatment that efficiently reduces bromate formation. Increasing addition of ammonia enhances the reduction rate of bromate formation, while the RCT remains unchanged.

Table 2. Summary of the Relationship Among the Parameters, RCT and Bromate Formation for CRW Water Ozonation (CRW Baseline Conditions: O3: 3 mg/L; Br¯: 170 mg/L; pH 7.5; Temperature: 24°C; Alkalinity: 100 mg/L As CaCO3; DOC: 3 mg/L, O3: 4.5 mg/L for DOM Experiments)

 

Parameter	Values	RCT	Rate of Bromate Formation
Bromide	+	-	+
pH	+	+	+
Temperature	+	+	+
Alkalinity	+	-	+
Ammonia	+	unchanged	-
NOM	+	+	-

Experimental reactions involved in the minimization of bromate through the addition of ammonia are summarized below:

• Ammonia has little effect on bromate reduction at pH 6.5 (in comparison to raw CRW that has no detected ammonia), and less than 10 percent bromate reduction at 20 minutes is observed when a high dose of ammonia is applied.

• After 20 minutes of reaction, up to 60 percent and 85 percent of bromate reduction were achieved at pH 7.5 and 8.5, respectively when the ratio of NH4⁺/Br¯ is greater than 15. However, even when a very high dose of ammonia is applied, some bromate still forms. The additional bromate reduction is due to the reaction of ammonia with ozone, which results in less ozone residuals remaining in waters.

• Addition of ammonia does not affect ozone decomposition due to its slow reaction with ozone and OH^•, and therefore, RCT remains unchanged.

• Formation of intermediate bromine, which is oxidized from bromide by O3, is inhibited by reaction with ammonia to prevent further oxidation by O3 and OH^• through direct/indirect pathways.

• Bromate removal by ammonia addition is not efficient in waters that have low pH and/or already contain a high ammonia level. This is important for practical applications in water treatment facilities when considering bromate control using ammonia. Also, addition of ammonia probably cannot achieve bromate levels below MCL if waters contain a high level of bromide.

Bromine (and Chlorine) Reactions With NOM. Rate determination experiments were conducted with the addition of aqueous chlorine or bromine in various concentrations of different types of NOM (both unaltered and preozonated forms) under various pH and temperature conditions. The following specific conclusions were reached:

• An indirect UV absorbance method using ABTS as an indicator provides a useful tool in conducting kinetic chlorination and bromination.

• Both HOBr and OBr¯ (or HOCl and OCl¯) appear to participate in reacting with NOM, and the effect of pH, as well as temperature, does not significantly change the rate of reaction. Preozonated reduces mainly on oxidation reactivity of NOM, which occurs more favorable in chlorination than in bromination.

• Bromine reactions with NOM are nearly an order of magnitude faster than similar chlorine reactions.

• Bromine reaction rate constants for non-ozonated NOM were 30 to 130 M^-1s^-1, and 15 to 75 M^-1s^-1 for preozonated NOM. For non-ozonated NOM: K (M^-1s^-1) = 34*(UVA,m^-1) – 14.

• Chlorine reaction rate constants for non-ozonated NOM were 1 to 5 M^-1s^-1, and 1 to 2 M^-1s^-1 for preozonated NOM.

• Faster NOM HOBr reaction sites (µM), based on a K= 10⁵ M^-1s^-1, can be estimated by the following reaction: 0.153(UVA254) + 0.053. These sites are all considered to form Total Hypobromite Ions (TOBr). Slower NOM HOBr reaction sites have a K value of approximately 50 M^-1s^-1, and are of similar order of magnitude between pH 5 through 11. These sites are assumed not to produce TOBr.

Kinetic-Based Mechanistic Model. A bromate formation model was evaluated that predicted bromate formation at elevated pH well. The model was less accurate in predicting the short-term formation of bromate at lower pH levels. The effect of NOM (K = 15 M^-1s^-1) on the reaction/consumption of HOBr and OBr¯ was not negligible, accounting for approximately 30 percent reduction in bromate formation (3 mg/L DOC). Modeling of carbonate radical concentrations and scavenging by NOM was considered as a means of improving these predictions. The model was used to predict bromate formation under ozone contact conditions representative of chemical disinfection for cryptosporidium. The simulations suggest that a combination of pH depression and ammonia addition, not either one or the other techniques, would be effective in combination. Changes in pH can have issues associated with corrosion in water distribution systems. Ammonia addition offers potentially greater flexibility than acid addition, based on changing influent bromide levels or water temperatures.

Bromate control (pH depression and ammonia addition) were simulated for theoretical ozonation scenarios. Ozone residual and Rct was modeled as follows:
O3 + O3 3O2 K = 2e10*[OH-]
Rct (0-2 min): 3E-14*(pH)^6.5486
Rct (>2 min): 1E-13*(pH)^5.4762

Bromate formation ozone exposures (1E04 M-min) were compared. An ozone exposure of 1.3E-4 M-min was reported to be equivalent to 2 log inactivation of cryptosporidium (Pinkernell, von Gunten, 2001). This ozone exposure corresponds to roughly 10 minutes of contact time at pH 8.2. Simulated bromate formation between pH 7.4 and 7.8 with a range of ammonia doses is presented in Figure 1. At higher pH levels, increasing the ammonia to bromide molar ratios can decrease bromate formation to below the MCL of 10 µg/L. The slopes of bromate formation as a function of pH or ammonia to bromide ratios are approximately equal, and suggests that the two strategies are equally effective.

 

Figure 1. Model Predictions for Bromate Control Based on Equations 7-10 Through 7-12 (Bromide = 2.12E-6M, Alkalinity = 1 mm, DOC = 3 mg/L)

Conclusions:

Bromate formation is controlled by oxidant concentrations (hydroxyl radicals and molecular ozone), with less than 50 percent of the bromate formation attributable to OH^• mediated reactions. OH^• concentrations are influenced by water quality factors. As such, differences in OH^• concentrations in different waters may explain the range of reported bromate concentrations throughout the United States. OH^• appears to play a minimal role during disinfection, but are important during oxidation of micropollutants in water (pesticides, taste, and odors). Therefore, although reducing OH^• concentrations may reduce bromate formation, tradeoffs of micropollutant oxidation may exist.

Acid addition (pH depression) decreases OH^• concentrations, and shifts aqueous bromine towards hypobromous acid rather than hypobromite ion. Ammonia addition does not affect OH^• concentrations, but reacts with aqueous bromine species. Aqueous bromine reactions with NOM occur on a kinetic time-scale affecting bromate formation, potentially forming organo-bromine species. Future work should identify and quantify organo-bromine by-products from ozonation. Futhermore, future work should quantify the rates of reactions between bromine-containing radical species (e.g., BrO^•, Br^•) and NOM. Control of bromate formation may be feasible through the addition of chemicals that scavenge bromine-containing radical species.

Journal Articles on this Report : 1 Displayed | Download in RIS Format

Publications Views

Other project views:	All 9 publications	1 publications in selected types	All 1 journal articles

Publications

Type	Citation	Project	Document Sources
Journal Article	Westerhoff P, Chao P, Mash H. Reactivity of natural organic matter with aqueous chlorine and bromine. Water Research 2004;38(6):1502-1513.	R826835 (Final)	Abstract from PubMed Full-text: Science Direct Exit Other: Science Direct pdf Exit

Supplemental Keywords:

drinking water, oxidation, pathogens, human health, carcinogen., RFA, Scientific Discipline, Water, Health Risk Assessment, Environmental Chemistry, Analytical Chemistry, Drinking Water, microbial contamination, natural waters, public water systems, monitoring, predicting chemical concentrations, chemical byproducts, disinfection byproducts (DPBs), kinetics, database development, natural organic matter, analytical methods, bromate formation, brominated DPBs, carcinogenicity, treatment, microbial risk management, hydroxyl radicals, DBP risk management, water quality, drinking water contaminants, drinking water treatment, water treatment, drinking water system, ozonation

Relevant Websites:

http://ceaspub.eas.asu.edu/pwest/USEPA_Bromate_826835.htm

Progress and Final Reports:

Original Abstract

1999 Progress Report