Si'erence, the colorimetric lest has an estimated accuracy of ±0.5 p.p.m. The titrimetric test has the same accuracy up to 200 p.p.m. When extended beyond that range by dilution of the sample, the estimated error in parts per million is 50/S. where V = milliliters of mercuric nitrate consumed, .V = normality of mercuric nitrate, and S = milliliters of sample water.If titration must be made on solutions which consistently exceed 1000 p.p.m. of chloride, it will be preferable to employ a higher concentration of mercuric nitrate.
EXPERIMENTAL DATAThe precision and accuracy of the colorimetric and titrimetric tests were evaluated by various laboratory personnel, using gravimetrically standardized solutions of c.p. potassium chloride in distilled water. The results are shown in Tables II and III.On completion of these analyses of pure chloride solutions, additional titrimetric tests were made on solutions contaminated with positive and negative ions apt to be encountered in industrial waters. No effort was made to overcome effects of the ions by modification of the titration. When used individually, 1000 p.p.m. of nitrate, sulfate, phosphate, magnesium, calcium, and aluminum did not significantly affect the titration. The heavy metal ions, zinc, lead, ferrous, nickel, and chromium, affected solution colors, but 100 p.p.m. concentrations of these ions did not reduce accuracy. Cupric ion was tolerable in concentrations up to 50 p.p.m. On the other hand, chromate and ferric ions ruined the titration in concentrations as low as 10 p.p.m. These two ions must be removed or otherwise counteracted before determination of chloride by this method. In the absence of inter-
CONCLUSIONSThe improved analytical methods described in this papei1 are applicable to the determination of chloride ion in all types of waters. They require only standard laboratory equipment and can be effectively applied by personnel with very limited laboratory experience. When implemented with slide-comparator equipment, the colorimetric test is ideally suited to field testing. The titrimetric method is independent of practically all common interferences.
ACKNOWLEDGMENTAcknowledgment is made to Genevieve Ziurys, who conducted most of the development tests concerned with the colorimetric method, and to Mary Q. Garner, who conducted the performance tests connected with the titrimetric method. The service of W. A. Taylor & Company in preparing a slide comparator for use with the colorimetric method, also is acknowledged.