Final Report: Analysis of Organic Byproducts from the use of Ozone/Chlorine and Ozone/Chloramines in Drinking Water Treatment

EPA Grant Number: R825364
Title: Analysis of Organic Byproducts from the use of Ozone/Chlorine and Ozone/Chloramines in Drinking Water Treatment
Investigators: Reckhow, David A.
Institution: University of Massachusetts - Amherst
EPA Project Officer: Hiscock, Michael
Project Period: January 1, 1997 through June 30, 1999
Project Amount: $355,795
RFA: Drinking Water (1996) RFA Text |  Recipients Lists
Research Category: Drinking Water , Water


The purpose of this project was to test, develop, and refine new and emerging analytical methods for nonvolatile organic disinfection byproducts (DBPs), and to identify new ozonation and mixed ozonation/chloramination byproducts. Special emphasis is placed on new aqueous-phase derivatization agents.

Summary/Accomplishments (Outputs/Outcomes):

Laboratory Studies: Analytical Method Development Using GC/MS. In early phases of this work, model carboxylic acids - known to be ozonation byproducts - were derivatized using hexylchloroformate, and detected by GC-ion trap-MS. Even though the procedure was simple, we could only get it to work for mono-, di-, hydroxy-, and chlorinated monoacids; it did not work adequately for alpha-ketoacids, which are very common in ozonated waters. In addition, this method did exhibit the expected sensitivities or low limits of detection needed for analysis of drinking water. Therefore, derivatization with ethylchloroformate was explored, making possible the detection of most alpha-ketoacids, but with very poor sensitivity and less satisfactory limits of detection than in the previous method. The derivatization yield of ketoacids with ethylchloroformate was found to be highly dependent on the composition of the reaction medium.

Because the detection limit was not as low as expected, another derivatization procedure was investigated. It was believed that pentafluorobenzylhydroxylamine (PFBHA) was a good alternative for chloroformates as a derivative in water because it currently is used for the detection of aldehydes and ketoacids. During this project, this analytical methodology was expanded and adapted for water samples based on the work of Yu, et al., Frazey, et al., and Spaulding, et al., who used a double derivatization with PFBHA and bis-(trimethylsilyl)-trifluoroacetamine (BSTFA) for air samples. It was found that aldehydes and ketoacids gave a strong linear response in the ppb range. Hydroxy-acids however could only be detected in the higher ppb and the ppm range. Therefore, it was thought that any future method using this dual derivatization approach would need to employ sample pre-concentration. Our recommendation is that water samples be concentrated on an anion exchange resin (e.g., Duolite A-7) before derivatization with PFBHA.

Laboratory Studies: Analytical Method Development Using LC/MS. Another analytical methodology that was studied involves the determination of carboxylic acids with an HPLC-MS system. The use of liquid chromatography allows the determination of highly polar/hydrophilic compounds, that cannot be determined using GC systems, and its combination with mass spectrometry can help in the identification of unknown products of chlorine and chloramination reactions. In our project, determination of carboxylic acids was accomplished by derivatization with 2-nitrophenylhydrazine. Under the proper conditions, 2-NPH only reacts with acids, avoiding interferences from ketones or aldehydes, compounds that usually are present in real samples. Additionally, the derivative forms a very stable negative ion, allowing its detection by negative ion electrospray mass spectrometry. This helps to reduce background noise from non-reacted reagents or byproducts of the reaction. These selective characteristics of the method were expected to allow the detection of acids (halogenated and non-halogenated) in very low concentrations and with high accuracy.

The derivatization procedure has proven to be simple and fast. A total of 32 acids, including keto-acids, were qualitatively determined by HPLC-MS after derivatization. Subsequent research led to the optimization of instrumental conditions and reaction conditions, to improve limits of detection. Under actual conditions, 100 ppb of propionic acid, a low-MW model byproduct has been detected using the MS-MS capability of the HPLC-MS system. This has been done without any sample concentration. The incorporation of broad-spectrum, sample pre-concentration into this HPLC/MS protocol would likely bring the limits of detection well down into the low ppb or even high ppt range. A well-targeted sample concentration scheme could lower this further.

Laboratory Studies: Characterization of Bulk Chlorinated NOM and Precursors. A range of bulk property tests were run on raw and treated waters, both with and without NOM fractionation. The purpose of this work was to characterize both the DBP precursors and to acquire some structural information on those high-molecular weight and otherwise recalcitrant DBPs.

Samples were subject to classical analyses (DOC, UV absorbance), preparative-scale resin extractions, and non-traditional analyses (fluorescence and pyrolysis-GC/MS). Pyrolysis GC/MS was performed both with and without simultaneous methylation (thermochemolysis). Results indicate that the methylation protocol helps in obtaining cleaner and more meaningful pyrochromatograms. Many structures are protected from over-degradation and interpretation of fragment origins becomes easier. However, the method still seems too harsh to preserve the structure of halogenated macromolecular NOM. For this reason, it doesn't provide much direct insight into unknown DBP structures.

Analysis of NOM by pyrolysis-GC/MS and fluorescence does provide some insight as to the nature of the DBP precursors. This can help in deducing plausible byproduct pathways and intermediate structures. For example, the pyrolysis and other spectral data showed that chlorination resulted in loss of aromatic structures, especially those that are hydroxylated. These data also showed that polysaccharide structures in NOM are little affected by chlorine. Based on this information, it would be unlikely to find important chlorination byproducts with structures resembling halogenated sugars, nor would it be very useful to adopt sugars as model compounds for DBP studies.

Laboratory Studies: Model Compound Experiments. Along with the analytical work, several model ozonation byproducts were chlorinated and chloraminated, and their chlorine demand and total organic halide (TOX) content were determined. From the set of compounds studied it was found that, in general, more unknown TOX is formed when chloramine is used as a disinfectant than when chlorine is used. Nevertheless, most of the compounds give rise to higher levels of halogenated acids (mostly as dichloroacetic acid) during chlorination than during chloramination. The comparison of TOX and chlorine demand provides important information on the reaction pathways and relative importance of oxidized products to halogenated ones.

After reaction with chlorine or chloramine, TOX, haloacetic acids (HAA) and pentane extractable compounds (trihalomethanes (THMs), acetonitriles and chloroacetones) were measured. From this analysis it has become clear that some compounds produce large amounts of "unknown TOX", whereas others give rise to compounds that can be almost fully characterized by analysis of the conventional chlorination byproducts. Some byproduct profiles allow one to postulate a priori the structure of major "unknown" halogenated byproducts.

A substantial set of experiments were conducted on the haloacetonitriles and haloacetamides. Dichloroacetamide (DCAD) was found to be the major long-lived intermediate in the degradation of dichloroacetonitrile (DCAN). Using careful laboratory studies, a multi-pathway kinetic model was proposed for the set of reactions that leads from DCAN to DCAD and on to DCAA. An N-chloro intermediate (DCAD-Cl) was also postulated. The analysis of DCAD proved to be challenging. It was noticed over the course of this work that DCAD will hydrolyse to form DCAA when samples are extracted and methylated for haloacetic acid analysis. Thus DCAD can create a positive bias in HAA measurements.

Field Studies. Field studies were conducted with the intent of applying some of the methods developed in this work, and documenting the persistence of ozonation DBPs into distribution systems. Many ozonation byproducts were found to degrade, yet others did persist even in systems practicing post-ozone filtration.

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

Other project views: All 8 publications 1 publications in selected types All 1 journal articles
Type Citation Project Document Sources
Journal Article Reckhow DA, Platt TL, MacNeill AL, McClellan JN. Formation and degradation of dichloroacetonitrile in drinking waters. Journal of Water Supply Research and Technology-Aqua 2001;50(1):1-13 R825364 (Final)
not available

Supplemental Keywords:

ozone, water, drinking water, chemicals, toxics, oxidants, organics, disinfection, oxidation, environmental chemistry, engineering, monitoring, analytical, surveys, measurement methods, northeast, Atlantic coast, Massachusetts, MA, Maine, ME, EPA Region 1., RFA, Scientific Discipline, Geographic Area, Water, Environmental Chemistry, Chemistry, State, Analytical Chemistry, Drinking Water, EPA Region, monitoring, alternative disinfection methods, public water systems, organic byproducts, exposure and effects, chemical byproducts, disinfection byproducts (DPBs), exposure, community water system, Massachusetts (MA), treatment, chlorine-based disinfection, microbial risk management, chloramines, DBP risk management, drinking water contaminants, water treatment, Region 1, drinking water system

Relevant Websites: exit EPA

Progress and Final Reports:

Original Abstract
  • 1997
  • 1998