Investigations of NDMA Formation by Chlorination of Model Compounds

Abstract

This study has demonstrated that there is the potential for formation of NDMA from the chlorination or chloramination of various nitrogenous biological molecules. We examined the potential of a wide range of amino acids to act as precursors for NDMA formation. The amino acids that were shown to be precursors were: arginine, phenylalanine, histidine, tryptophan, tyrosine and l-glutamine. Purines and pyrimidines were also investigated in terms of their potential as precursors. It was found that uracil and guanine were positive as precursors for NDMA formation. Interestingly, the cyanobacterial toxin, cylindrospermopsin, was also found to act as a precursor for NDMA formation when chlorinated. Cylindrospermopsin structurally contains a uracil moiety in its molecule, which could be site of attack in NDMA formation. Experiments using chloramination additionally found that glycine, threonine, urea and the cyanobacterial toxin, microcystin-LR were precursors for NDMA formation. Microcystin-LR is a cyclic heptapeptide consisting of 7 amino acids including arginine and a structurally unique aromatic acid, ADDA. Mixes of amino acids were found to form NDMA more rapidly than when chlorinated singly. The situation of complex mixtures of amino acids and other precursors such as uracil, urea and cyanobacterial toxins in source waters is often encountered. The addition of dimethylamine or trimethylamine to the chlorination mixtures did not increase the number of amino acids that were shown to be precursors. Research into the development of a suitable analytical method for NDMA in disinfected drinking waters has culminated with a sensitive validated method utilising solid phase extraction combined with ammonia positive chemical ionization gas chromatography-mass spectrometry. This method is now established and is commercially available in Queensland Health Scientific Services. A limited survey of NDMA presence in Australian disinfected waters indicated that, with a few exceptions, the waters surveyed did not contain detectable levels of NDMA. These results however need to be regarded as indicative only as they utilised analytical methodology in development at the time. It is recommended that further, more extensive surveys be conducted using the latest sensitive analytical methods.