James M. Symons 1 , Gerald E. Speitel Jr2 Cordelia J. Hwang3, Stuart W. Krasner 3, Silvia E. Barrett3, Alicia C. Diehl2, and Rebecca Xia1
1Department of Civil and Environmental Engineering
University of Houston
Houston, Texas 77204-4791
2Department of Civil Engineering
University of Texas at Austin
Austin, Texas 78712
3Metropolitan Water District of Southern California
Water Quality Laboratory
700 Moreno Avenue
La Verne, California 91750-3399
American Water Works Association Research Foundation
6666 W. Quincy Avenue
Denver, Colorado 80235-3098
Disinfection practices in the U.S. drinking water industry are now evolving in response to several concerns and will continue to evolve over the next decade. In response to current and anticipated disinfection by-product (DBP) regulations, many utilities have begun to employ chloramines as a disinfectant, and others will do so in the future. Also, in response both to DBP regulations and to meeting the Surface Water Treatment Rule (SWTR) and Enhanced SWTR, other utilities will switch to ozone as the primary disinfectant and chloramines as the secondary disinfectant. A third possibility is the initial use of free chlorine for disinfection purposes to meet the SWTR, followed by the introduction of ammonia at some point in the treatment train to minimize further formation of DBPs.
Some known DBPs (e.g., trihalomethanes, haloacetic acids, and haloacetonitriles) associated with chlorination have been observed during chloramination as well; however, these chemicals are generally present at lower concentrations. A decreased dissolved organic halogen (DOX) concentration also is observed upon chloramination; however, a smaller percentage of the chemicals comprising the DOX has been identified for chloramination in comparison to chlorination. Except for cyanogen chloride, halogen-substituted DBPs preferentially formed from chloramination have not been identified. Furthermore, only limited work on the DBPs from the chloramination of ozonated water (be they halogen-substituted or not) has been performed. Thus, prior to this study, two key questions emerged as increased use of chloramines is taking place:
1) Why are significant quantities of known DBPs formed in some cases and
2) Are any of the currently unidentified chloramination DBPs of health and potential regulatory significance?
Formation of known DBPs may result from specific chemical characteristics of the water or the chloramination process. These parameters might include the total organic carbon (TOC) concentration, the bromide ion concentration, the pH, the chlorine to ammonia nitrogen ratio, the relative ratio of mono- and dichloramine, the chloramine dosage, the order of addition of chlorine and ammonia, and the intensity of mixing during this addition. Prior to this study, the importance of these parameters in the formation of known DBPs during chloramination was not well defined and needed detailed investigation.
To address the above issues, this project covered three primary aspects of work:
1. What chemical and operational factors influence DBP formation;
2. What known and unidentified DBPs are formed; and
3. What treatment steps can be implemented to lower the DBP concentrations?
The project research program consisted of laboratory and pilot-scale work, organized in a logical progression, starting from a basic investigation of the influence of specific water quality and operational parameters and progressing to identifying and implementing solutions to minimize DBP formation under practical treatment conditions.
The primary participants in this project were the University of Houston (UH) and the City of Houston, the University of Texas at Austin (UT) and the City of Austin, and the Metropolitan Water District of Southern California (MWDSC). Five other utilities across the country participated through a full-scale sampling program and provided water for limited laboratory-scale testing. These utilities were selected to cover various raw water characteristics and treatment conditions as well as to provide geographical diversity.
The project consisted of four main tasks. In the first task (1a), batch experiments were conducted on the three primary water sources, Lake Austin water (LAW), Lake Houston water (LHW) and California State Project water (CSPW). Using preformed chloramines, the batch experiments to determine DBP formation during chloramination were chosen to cover variable water chemistry conditions:
|pH: 6, 8, 10,|
|total chlorine residual after 48 hours: 1, 2, 4 mg/L, and|
|Cl2/NH3-N mass ratio (called Cl2/N ratio): 3/1, 5/1, 7/1,|
This task also included a study of variable mixing conditions, as well as sequential addition of chlorine followed by ammonia, each under five different water chemistry conditions (1b):
|low, medium, and high mixing energies with simultaneous addition of chlorine and ammonia, and|
|chlorine then ammonia with a 30 second delay, low, and medium mixing energies.|
The formation of DBPs were then measured after two days holding time to simulate passage through a distribution system (2-d simulated distribution system (SDS) DBPs).
In the second task (2), a pilot testing program on each of the primary water sources was conducted to confirm the findings of the batch studies in continuous-flow. The goal of this task was to provide insight into the expected behavior of full-scale plants. Whereas Task 1a was performed entirely on source waters, Task 2 studied source water and post-filter chloramination of conventionally treated water (i.e., coagulated or softened, settled and filtered) with and without source water ozonation.
In the third task (3), the scope of the project was expanded to include water sources in five other geographical locations; northeast, northwest, deep south (2) and mid-south. These other water sources were selected to cover a wide range of water qualities and operational characteristics. Operational data were collected from these five locations as well as finished water samples for analysis. Finally, source water from these five locations was shipped to the University of Texas, where selected, laboratory-scale batch study conditions were performed, three for each water.
For each condition in the first two tasks, four trihalomethanes (THMs) and DOX concentrations were determined for each sample collected and six haloacetic acids (HAAs) and two cyanogen halides (CNX) (cyanogen chloride and cyanogen bromide) were determined on selected representative samples. For Task 3, the complete suite of analyses were preformed on all full-scale and bench-scale tests.
The fourth task (4) consisted of development and application of analytical techniques for identifying currently unknown DBPs. These new analytical techniques were applied to selected representative samples collected throughout the study.
The results of this study confirm that DBP formation during chloramination generally does not pose a regulatory concern based on current drinking water regulations and probably will not cause a concern with the proposed Stage 1 regulations. Some problems may arise in meeting the proposed Stage 2 regulations for HAAs. Although chloramines limit the formation of THMs to concentrations generally below that of Stage 2 of the proposed Disinfectants/Disinfection By-Product (D/DBP) rule and trichloroacetic acid (TCAA) generally to concentrations below the detection limit (BDL) of the analytic method used, chloramines were not as effective in minimizing the formation of dihalogen-substituted HAAs (DCAA, DBAA, BCAA--DXAA).
Even though chloramines generally do not produce concentrations of most regulated chemicals that are of concern, formation of unregulated and uncharacterized halogenated chemicals, as measured by the DOX analysis, is significant (as high as 300 µg Cl-/L) under some conditions. Therefore, water utilities may want to consider concentrations of both specific regulated chemicals and DOX in selecting operating conditions for chloramination.
Some decrease in DBP formation may be observed through improved mixing at the point of chemical addition. Also, simultaneous addition of chlorine and ammonia, in comparison to delayed addition of ammonia, should reduce DBP formation, especially formation of THMs. In bench scale mixing tests, the decrease in DBP formation through improved mixing and simultaneous chemical addition did not exceed 50 percent based on 48-hr SDS tests; therefore, this approach to DBP control is most applicable to situations where modest decreases in DBP formation are sought. The possible benefits from this approach also are a function of the quality of the mixing and chemical addition schemes in current use.
System chemistry affects DBP formation far more than mixing. In general, the formation of DBPs decreases with increasing pH (up to pH 10 studied) and decreasing Cl2/N ratio (down to 3/1 studied). Therefore, manipulation of these two major operating variables can significantly impact DBP formation. Unfortunately, the general observations of the effect of pH and Cl2/N ratio on DBP formation may not hold for all waters near neutral pH (7 to 8.5), because of the complexity of haloamine chemistry over this pH range. Therefore, bench scale testing like that performed in Task 1a of this research is recommended as an initial step in investigating the impact of operating conditions on DBP formation. Further investigation at pilot scale also may be warranted if substantial changes in operating conditions are contemplated.
As noted above, decreasing the Cl2/N ratio, especially to low values such as 3/1, decreases DBP formation. Unfortunately, some water utilities have experienced problems in maintaining adequate microbiological quality in distribution systems at low Cl2/N ratios. Growth of nitrifying bacteria is a particular problem. Therefore, minimizing DBP formation and maintaining acceptable microbiological water quality in the distribution system may conflict with one another. Possible adverse water quality impacts should be considered in conjunction with a decrease in the Cl2/N ratio to low levels.
Any strategy aimed at controlling DBP formation through modification of pH and the Cl2/N ratio will have practical ranges of workable values that are specific to each situation. In some cases, the workable ranges may be inadequate to satisfactorily control DBP formation. In this study, the HAA6 concentration usually consisted of only dehalogenated acids (e.g., dichloroacetic acid). Conceivably, the HAA5 concentration in some waters could exceed the proposed Stage 2 regulations. Under these circumstances, preozonation followed by chloramination should be considered. This research showed that ozonation prior to chloramination decreased the formation of both HAAs and DOX.
In addition to pH and the Cl2/N ratio, two other system chemistry parameters may be important in DBP formation: bromide and alkalinity. This research shows that as the bromide concentration increases (up to 0.74 mg/L studied) DBP formation likewise increases, see Figure 5.13 as an example, and the speciation within the individual classes of DBPs (e.g., THMs) shifts toward the bromine substituted chemicals, see Figure 5.14 as an example. Therefore, water utilities that experience cyclical changes in the bromide concentration of their source water can expect an increase in DBP formation when the bromide ion concentration increases, and visa versa.
Monochloramine can react with organics via an acid-catalyzed mechanism to yield halogen-substituted organics. This reaction mechanism is catalyzed by proton donors such as carbonic acid and bicarbonate, the latter of which is a component of alkalinity and a common constituent of natural waters. Thus, as alkalinity increases, the rate of DBP formation also may increase. Utilities that have significant alkalinity (up to 165 mg CaCO3/L were investigated), especially those practicing or considering lime softening, may want to examine the effect of alkalinity removal on DBP formation. The effect of alkalinity on DBP formation was not formally part of this research; however, some very limited data from several pilot plant runs suggest that alkalinity may impact DBP formation.
Specific DBPs (e.g., THMs, HAAs, CNX) may comprise a very small percentage of the DOX concentration. Under such circumstances, water utilities may want to investigate their water in more detail to identify additional chemicals. This research examined a number of new analytical approaches for identifying additional chloramination DBPs. Ultrafiltration (UF) using DOX and TOC surrogates and liquid chromatography (LC) are methods that could be adopted by a research laboratory to provide general information about halogen-substituted DBPs. As with other MS investigations into the identification of chlorination DBPs, the ultimate goal is to develop analytical methods using more readily available instrumentation once unknown DBPs have been identified. The actual practice of this approach cannot be instituted, however, until more of the chloramine DBPs are identified and their health significance evaluated. This study has shown that an initial full-scan, low resolution LC-electrospray ionization (ESI)-mass spectrometry (MS) run can provide preliminary halogen content and molecular weight information. Subsequent, high resolution MS and MS-MS runs could then focus on peaks of interest to determine chemical composition and structure for DBP identification.
In summary, the following major conclusions can be made based on the results of this study.
|The results from the bench-scale chemistry studies, the pilot-scale studies, and the
studies of geographically diverse waters, generally agreed, giving confidence that the
findings of this study would be applicable to a wide variety of waters.|
|Over the range studied (1 to 4 mg/L), the total disinfectant residual after two days of
incubation had little influence on the resulting DBPs formed.|
|Controlling THMs to the levels of Stage 2 of the D/DBP rule using chloramination should
|Dihalogen-substituted HAAs (DXAA) dominated the 2-d SDS HAA6, implying that they might
not be well controlled by using chloramination.|
|Substantial quantities of 2-d SDS DOX were formed in all waters studied, particularly
when dichloramine was present.|
|Low percentages (commonly below 25 percent) of DOX could be accounted for by summing the molar concentration of the 12 2-d SDS DBPs measured in this study, indicating that many unidentifiable DBPs were being formed during chloramination.|
|In general, DBP formation increases as the pH decreases and the Cl2/N ratio
|The presence of bromide ion complicates the control of DBPs because of the complexity of
|When bromide ion is present CNBr is formed in addition to CNCl, thus increasing the CNX.
The base-catalyzed hydrolysis of CNX resulted in less CNX being present after two days of
incubation at higher pHs.|
|In the bench-scale batch tests, relative mixing energy had little influence on resulting
2-d SDS DBP concentrations, but simultaneous addition of chlorine and ammonia is
|Ozonation altered DBP precursors such that applying ozone prior to chloramination
resulted in lessened concentrations of resulting 2-d SDS DBPs.|
|In some chloraminated water, the <500 dalton ultrafiltration (UF) fraction
represented approximately 43 to 61 percent of the DOX.|
|In some of the other chloraminated waters, the two highest molecular weight fractions
(the 3 K to 10 K and >10 K) together represented approximately 39 to 55 percent of the
DOX. Thus, significant concentrations of halogen-substituted DBPs with very high molecular
weight also are possible.|
|UF provides a unique analytical tool to preliminarily ascertain which molecular weight
fraction is most significant for a site specific chloramination.|
|Dihalomethanes (dibromo-, bromoiodo-, and diido-) may be specific chloramination DBPs.|
|Monochloramine, not dichloramine, reacted with small model peptides.|
|UF, simultaneous distillation extraction gas chromatography&endash;mass spectrometry (SDE-GC-MS), liquid chromatography&endash;electrospray ionization&endash;mass spectrometry (LC-ESI-MS), and liquid chromatography-potassium iodide-ultraviolet detection (LC-KI-UV) are all techniques applicable to the study of chloramine DBPs.|
Overall, practicing conventional coagulation, adding well mixed chlorine and ammonia solutions simultaneously in the appropriate ratio, and keeping the pH in the distribution system (as represented by incubation pH in this study) as high as possible after chloramination at as low a Cl2/N ratio as possible should minimize overall DBP formation. Where needed, preozonation before chloramine addition should further decrease DBP formation.
Diehl, A. C., Speitel, G. E. Jr., Symons, J. M., Krasner, S. W., Hwang, C. J., and Barrett, S. E., "DBP Formation During Chloramination," Journal of the American Water Works Association, 92, (6), 76-90 (June, 2000).