You are here:
TECHNIQUES FOR ANALYZING COMPLEX MIXTURES OF DRINKING WATER DBPS
Citation:
Richardson, S D. TECHNIQUES FOR ANALYZING COMPLEX MIXTURES OF DRINKING WATER DBPS. Presented at Analytica Conference 2000, Munich, Germany, April 11-14, 2000.
Impact/Purpose:
(1) Use toxicity-based approach to identify DBPs that show the greatest toxic response. (2) Comprehensively identify DBPs formed by different disinfectant regimes for the 'Four Lab Study'. (3) Determine the mechanisms of formation for potentially hazardous bromonitromethane DBPs.
Description:
Although chlorine has been used to disinfect drinking water for approximately 100 years, there have been concerns raised over its use, due to the formation of potentially hazardous by-products. Trihalomethanes (THMs) were the first disinfection by-products (DBPs) identified and shown to cause cancer in laboratory animals, and are regulated by many countries. Recently, there has also been concern about haloacetic acids, which will be regulated in the United States following the promulgation of the Disinfectants/Disinfection By-product Rule in 1998. As a result of tightening regulations and increased concerns over chlorine DBPs, many drinking water treatment plants have changed their mode of disinfection to alternative disinfectants, including ozone, chlorine dioxide, and chloramine. However, DBPs from these disinfectants have not been thoroughly characterized, so it is not known whether these alternative disinfectants are safer or more harmful than chlorine. To solve this issue, the U.S. Environmental Protection Agency's (EPA's) National Exposure Research Laboratory has been applying a variety of analytical techniques to identify these unknown DBPs--with a focus on the alternative disinfectants (for which less is known), but also including chlorine DBPs. One reason that many DBPs have not been previously identified is that their spectra are often not found in the spectral libraries, such as the NIST or Wiley databases. Spectral matching against one of these databases is the easiest and most common way of identifying an unknown compound. But, when compounds are not present in the libraries, additional spectral information is required to solve their structures. Because drinking water extracts typically contain over 300 detectable compounds, a separation device (such as a gas chromatograph or liquid chromatograph) is beneficial to allow pure spectra to be obtained for these unknowns. For those compounds that can be extracted into an organic solvent, we use a combination of gas chromatography/mass spectrometry (GC/MS) techniques and GC/infrared (IR) spectroscopy. High resolution mass spectrometry limits the number of choices of possible atom combinations and typically points to a single empirical formula assignment for the unknown DBPs. Chemical ionization mass spectrometry is used to provide molecular weight information when molecular ions are not present in the electron ionization (EI) mass spectra. And, finally, GC/IR is used to provide functional group information for the unknown compounds. For example, we may know that an oxygen is present in the structure from high resolution EI-MS information, but we may not be sure whether the compound is a ketone, an alcohol, or an aldehyde. GC/IR will usually point to a definitive functional group, which aids in determining the exact structure of the compound. Another reason for the lack of information on DBPs is that many are believed to be polar in nature and cannot be easily extracted from water or analyzed by GC. To identify these compounds, we have utilized liquid chromatography (LC)/MS, in addition to using derivatizing agents such as pentafluorobenzylhydroxylamine (PFBHA) with GC/MS. Initially, capillary electrophoresis/MS was also investigated as a tool because it has been shown to work well for ionic and polar compounds. However, we found the separations to be poor, due to interfering salts from the drinking water. Conventional LC/MS, utilizing electrospray or atmospheric pressure ionization, presented difficulties also, as chemical backgrounds were difficult to overcome in the low molecular weight range that the DBPs are usually found. One solution we found for analyzing polar carbonyl DBPs was to first derivatize the drinking water sample with 2,4-dinitrophenylhydrazine (DNPH), then use solid phase extraction (which concentrates the trace levels of DBPs and removes salts), and analyze the unknown DBPs with LC/MS (with electrospray). We were successful in identifying a number of highly polar DBPs that were not easily analyzed by other methods. We are currently continuing to apply new derivatizing agents in order to broaden the classes of structures that can be identified. Using all of the spectroscopic techniques discussed, we have been able to identify more than 100 DBPs that were not previously known. Preliminary toxicity screening is now being conducted on those DBPs that were determined to be possible adverse health risks, and a U.S. nationwide occurrence study will provide occurrence and concentration information on those high-priority DBPs.