Environ. Sci. Technol. 2003, 37, 2920-2928
Factors Influencing the Formation and Relative Distribution of Haloacetic Acids and Trihalomethanes in Drinking Water LIN LIANG† AND PHILIP C. SINGER* Department of Environmental Sciences and Engineering, CB# 7431, University of North Carolina, Chapel Hill, North Carolina 27599-7431
Various water quality and treatment characteristics were evaluated under controlled chlorination conditions to determine their influences on the formation and distribution of nine haloacetic acids and four trihalomethanes in drinking water. Raw waters were sampled from five water utilities and were coagulated with alum and fractionated with XAD-8 resin. The resulting four fractionssraw and coagulated water and the hydrophobic and hydrophilic extractsswere then chlorinated at pH 6 and 8 and held at 20 °C for various contact times. The results show that increasing pH from 6 to 8 increased trihalomethane formation but decreased trihaloacetic acid formation, with little effect on dihaloacetic acid formation. More trihalomethanes were formed than haloacetic acids at pH 8, while the reverse was true at pH 6. Hydrophobic fractions always gave higher haloacetic acid and trihalomethane formation potentials than their corresponding hydrophilic fractions, but hydrophilic carbon also played an important role in disinfection byproduct formation for waters with low humic content. The bromine-containing species comprised a higher molar proportion of the trihalomethanes than of the haloacetic acids. The hydrophilic fractions were more reactive with bromine than their corresponding hydrophobic fractions. Coagulation generally removed more haloacetic acid precursors than trihalomethane precursors. Waters with higher specific ultraviolet absorbance values were more amenable to removal of organic material by coagulation than waters with low specific ultraviolet absorbance values. Experimental evidence suggests that haloacetic acid precursors have a higher aromatic content than trihalomethane precursors.
Introduction Public health concern over the disinfection process was first raised over 25 years ago with the identification of chloroform and other trihalomethanes (THMs) as disinfection byproducts (DBPs) in chlorinated drinking water (1). Since that time, extensive investigations have been made on the formation of chlorination byproducts, with emphasis on THMs and, more recently, haloacetic acids (HAAs), the two most prevalent classes of byproducts resulting from chlorination in drinking water treatment. Although there are nine HAA * Corresponding author phone: (919) 966-3865; fax: (919) 9667911; e-mail:
[email protected]. † Current address: Greeley and Hansen LLC (Chicago), Richmond, VA. 2920
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 13, 2003
species, only five of them are regulated by the current Disinfectants/Disinfection By-Products (D/DBP) Rule due to limited formation and occurrence data for some of the species (2). The five HAAs are monochloro- and monobromoacetic acid, dichloro- and trichloroacetic acid, and dibromoacetic acid (ClAA, BrAA, Cl2AA, Cl3AA, and Br2AA, respectively). The formation of DBPs depends primarily on source water quality characteristics and on the location in the treatment process where disinfectants are added. Generally, all other factors being equal, fewer DBPs will be formed when the disinfectants are added later in the treatment process. The most important water quality parameters that influence the formation of DBPs include the nature and concentration of organic precursor materials, water temperature, pH, and conditions under which the disinfectant is used, such as the disinfectant dose, point of addition, contact time, and residual disinfectant concentration. In the presence of bromide, free chlorine (hypochlorous acid) rapidly oxidizes bromide to hypobromous acid which then, along with the residual hypochlorous acid, reacts with the precursor materials to produce mixed chloro-bromo substitution products. Bromine incorporation into HAAs parallels that of bromine incorporation into THMs (3, 4). Natural organic materials (NOM)se.g., humic substances, which are present to various degrees in all water supplies and constitute the major component of the total organic carbon (TOC) concentration in most watersshave been identified as the principal precursors in the formation of THMs and HAAs (5, 6). Reckhow et al. (7) have shown that halogenated DBP formation increases with the “activated (defined as electron-rich) aromatic” content of NOM. Many efforts have been made to correlate various fundamental characteristics of NOM (e.g. size, structure, functionality) to DBP formation, among which SUVA (specific ultraviolet absorbance, defined as ultraviolet absorbance at 254 nm (UV254) times 100 divided by the dissolved organic carbon concentration) has been demonstrated to be a good predictor of the aromatic carbon content of the NOM and the DBP formation potential of the water (8, 9). Aquatic NOM consists of both hydrophobic and hydrophilic organic material originating from the degradation and leaching of organic detritus within the watershed. The nature and distribution of these hydrophobic and hydrophilic materials differ, depending on the source materials and the biogeochemical processes involved in carbon cycling within the terrestrial and aquatic systems. In some water systems, THM levels exceed HAA levels (10, 11), while the reverse is true in others (11-13). The relative distribution of HAAs and THMs is believed to be influenced by the hydrophobic/hydrophilic distribution of NOM in the waters being chlorinated. Some researchers (9) have concluded that hydrophilic fractions of NOM are more significant precursors of THMs than HAAs based on the fact that in finished drinking water, where hydrophilic fractions of NOM tend to be dominant, greater yields of THMs relative to HAAs are obtained than from chlorinated hydrophobic humic substances. Others (14, 15) have suggested that the hydrophilic fraction produces a greater yield of HAAs relative to THMs than the corresponding hydrophobic fraction. Other factors that impact the distribution of HAAs and THMs include the pH of chlorination and distribution, temperature, contact time, type of disinfection scenario, whether coagulation is practiced prior to chlorination, and the potential for biodegradation of the HAAs. Coagulation has been reported to remove more hydrophobic carbon than 10.1021/es026230q CCC: $25.00
2003 American Chemical Society Published on Web 05/24/2003
FIGURE 1. Experimental schematic (* WU denotes water utility). hydrophilic carbon (16-18), resulting in a shift in the hydrophobic/hydrophilic distribution and hence the relative formation of HAAs and THMs upon subsequent chlorination. The effect of pH may be interpreted as a shift in chlorination mechanisms as the degree of protonation of the reacting species changes (7), and some of the HAA species are known to decompose at elevated pH values (10, 19). The reaction rates for HAA and THM formation are also different, HAAs tending to be formed faster than THMs (20). The primary objective of this study was to assess the impact of various water quality and treatment characteristics on HAA and THM formation and on the relative distribution of these two classes of DBPs under controlled chlorination conditions in different types of waters. The studies were conducted on five raw waters with different types of NOM, allowing for the following variables to be examined: chemical characteristics of NOM in the raw water, including SUVA and the hydrophobic/ hydrophilic carbon distribution; impact of coagulation prior to chlorination; pH of chlorination (pH 6 vs pH 8); bromide concentration (i.e., Br to TOC ratio); and chlorination contact time. In all cases, all nine bromineand chlorine-containing HAA species (HAA9), which include bromochloro-, bromodichloro-, dibromochloro-, and tribromoacetic acid (BrClAA, BrCl2AA, Br2ClAA, and Br3AA, respectively) in addition to the five regulated HAAs (HAA5) were measured along with the four THM species (THM4). This paper presents the results with respect to pH, hydrophobic versus hydrophilic carbon, and coagulation.
Experimental Methods All experiments were conducted on raw waters collected from five water utilities between February 1999 and March 2000. The water sources of the five utilities were as follows: White River (Indianapolis, IN), Lake Manatee Reservoir (Bradenton, FL), Mississippi River (East St. Louis, IL), Poquonnock Reservoir (Groton, CT), and South Fork Tolt River (Seattle, WA), respectively. These water utilities were selected to provide waters with different SUVA values as well as geographical diversity. A schematic of the experimental program is shown in Figure 1. A portion of the raw water was acidified to pH 2.0 using sulfuric acid and then passed through
an XAD-8 resin (Rohm and Haas, Philadelphia, PA) column to determine the hydrophobic/hydrophilic distribution of the NOM and to isolate each fraction, in accordance with the method of Thurman and Malcolm (21). The effluent from the column comprised the hydrophilic fraction. The hydrophobic carbon retained on the resin was eluted with 0.1 M NaOH in the reverse direction. This fraction was neutralized to pH 6-7 with sulfuric acid and diluted into deionized, organic-free water (DOFW) (Dracor Inc., Durham, NC) to the same TOC concentration as the hydrophilic fraction prior to use. Another portion of the raw water was coagulated with alum [Al2(SO4)3‚(14-16)H2O] in 2-L Phipps and Bird jars and settled for collection of the supernatant. The alum dose applied in the coagulation tests was the optimal dose for TOC removal as determined by preliminary jar tests (18). The pH for the coagulation tests was maintained above 6 with sodium carbonate to provide for effective particle (turbidity) removal. Prior to chlorination, all four water fractions derived from the original bulk water, i.e., the raw water itself, coagulated water, the hydrophobic fraction, and the hydrophilic fraction, were characterized by measuring TOC, dissolved organic carbon (DOC), UV254, and bromide concentration. If necessary, sodium bromide (NaBr) was spiked to some of the water fractions so that all four fractions had the same bromide/ TOC ratio. Samples for DOC and UV254 measurements were filtered through prerinsed 0.45 µm Supor-450 membrane filters (Gelman Sciences, Ann Arbor, MI) prior to analysis. Bromide ion concentration was determined using a Dionex ion chromatographic (IC) system (Sunnyvale, CA). Chlorination was performed on samples of each fraction buffered with 0.001 M Na2HPO4 at pH 6.0 and 8.0. Chlorine (sodium hypochlorite) stock solution was standardized by sodium thiosulfate titration in accordance with Standard Method 4500-Cl B (22). The chlorine dose for each fraction was determined through preliminary chlorine demand experiments such that the free chlorine residual was 1.0 mg/L as Cl2 after headspace-free storage for 24 h at 20 °C in the dark. The chlorine demand experiments were performed at pH 8.0 due to the higher chlorine demand at pH 8.0 compared to pH 6.0. Free chlorine residual was measured using a chlorine residual pocket colorimeter (Hach Company, Loveland, CO). All samples were chlorinated in 300-mL, chlorine demand-free, glass-stoppered BOD bottles and stored headspace-free at 20 °C in the dark. After contact periods of 1, 2, 4, 8, 24, and 72 h, samples at both pHs were collected headspace-free in 40-mL glass vials with polypropylene screw caps and Teflon-lined septa for subsequent analyses of THM4 and HAA9. The vials contained sufficient ammonium sulfate to quench the residual free chlorine. In the HAA9 vials, 20 µL of a 10 µg/L sodium azide solution was also added as a biocide if the samples were not analyzed within 24 h (23). All samples were refrigerated at 4 °C for no more than 2 weeks prior to analysis. Free chlorine residual was measured at the time of sample quenching. All samples were found to have a measurable free chlorine residual after 72 h. The liquid-liquid extraction, gas chromatographic procedure for the analysis of THM4 was developed after Standard Method 6232B (22). The THM4 calibration solutions were prepared from EPA Trihalomethanes Calibration Mix standard solution (Supelco Inc., Bellefonte, PA). Pentane was used as the extracting solvent and 1,2-dibromopropane was used as an internal standard. The HAA9 species were analyzed following derivatization with diazomethane using a micro liquid-liquid extraction gas chromatographic method which was developed by Brophy et al. (23) based on Standard Method 6251B (22) and EPA Method 552 (24). The method incorporated anhydrous magnesium sulfate as a drying agent to promote complete derivatization of the HAAs to their corresponding methyl esters. Methyl-tertiary-butyl-ether VOL. 37, NO. 13, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2921
TABLE 1. Characteristics of Water Fractions Prior to Chlorination DOC (mg/L)
UV254 (1/cm)
water and fraction raw water coagulated water hydrophilic fraction (67%)a hydrophobic fraction (33%)a
2.7 2.3 1.8 1.8
Indianapolis Water 2.8 0.087 2.3 0.063 1.8 0.042 1.8 0.076
3.1 2.7 2.3 4.2
23 23 ndb nmc
35 29 23 23
raw water coagulated water hydrophilic fraction (48%)a hydrophobic fraction (52%)a
8.1 3.9 3.9 3.8
Manatee Water 8.2 0.359 3.9 0.088 4.0 0.110 4.0 0.214
4.4 2.3 2.8 5.4
173 162 ndb nmc
173 162 173 173
raw water coagulated water hydrophilic fraction (57%)a hydrophobic fraction (43%)a
5.1 3.8 2.9 2.8
East St. Louis Water 5.0 0.163 3.8 0.102 2.9 0.073 2.8 0.101
3.3 2.7 2.5 3.6
59 61 41 ndd
82 61 41 41
raw water coagulated water hydrophilic fraction (56%)a hydrophobic fraction (44%)
3.4 2.1 1.9 1.9
Groton Water 3.3 0.119 1.7 0.035 1.8 0.042 1.9 0.075
3.6 2.1 2.3 3.9
