A method to differentiate total organic chlorine (TOCl), total organic bromine (TOBr) and total organic iodine (TOI), which involves carbon adsorption, pyrolysis and ion chromatographic detection, has been developed by previous researchers and applied to the analysis of water samples. However, incomplete recoveries from model compounds and interference from auxiliary gas (CO₂) were reported. The objective of this research was to determine the best TOX protocol for use with IC analysis. Two commercial analyzers (one using a pure O₂ carrier and one using O₂/CO₂ mixture) and three commercially available activated carbons (two coconut based and one bituminous coal based) were selected for this research. Chlorine, bromine and iodine containing model compounds and chlorinated water samples with varying levels of bromide and iodide were tested on both analyzers and three activated carbons. Results showed that the pyrolytic analyzer using pure O₂ and off-line IC combined with a standard TOX carbon (coconut based) achieved the most complete recovery of TOCl, TOBr and TOI for both model compounds and real samples. There was no obvious difference between the two analyzers when used in microcoulometric detection mode. The O₂/CO₂ pyrolytic system and off-line IC method showed incomplete recovery for some samples due to the condensation of halide ions and incomplete transfer line flushing. The use of CO₂ as an auxiliary gas resulted in interference in IC analysis, which made it necessary to purge dissolved CO₂ before IC analysis. The TOX method is moderately sensitive to nitrate rinse volume. The monohaloacetic acids were partly washed out during sample preparation. This problem was solved by modifying the nitrate rinsing solution. Complete recovery was achieved for all the selected compounds by this modified TOX protocol. Includes 13 references, tables, figures.