Sampling and analysis considerations in evaluating levels of atmospheric lead

Abstract

The primary purpose of this investigation was to ascertain the accuracy and precision with which the total quantities of lead could be determined when only segments of the membrane filters used to collect the samples are employed for the analyses. A second objective was to compare the precision and relative sensitivities of the method of atomic absorption spectrophotometry and the wet chemical (dithizone) method for determining lead concentrations in the presence of other substances in urban environments. Experiments were conducted with samples of filters obtained from sampling stations in rural and urban areas in the United States. A total of 160 segments from 36 filters were analyzed by the method of atomic absorption and the wet chemical method. The estimated range of the quantities of lead per filter was from 274 to 4168 μg. Statistical analysis of the data (in the logarithmic domain) led to the conclusion that the precision with which quantity of lead on a filter may be estimated from the analysis of one segment is inconsistent from one filter to the next, and may lead to 95 per cent confidence limits as narrow as 97–103 per cent or as wide as 67–150 per cent, depending on whether the lower or upper 95 per cent limit of the estimated variance is applied. Furthermore, from the evidence at hand it could not be concluded that the estimate based on one 50 per cent segment is more precise or more accurate than that based on one 10 per cent segment from a filter. In that sense, determinations in two or more 10 per cent segments rather than one 50 per cent segment would lead to more precise estimates of the quantity of lead on a filter. The results suggest where conclusions have to be drawn from a single filter or from a limited number of filters, that the whole filter membrane rather than segments thereof be analyzed. When comparison of the performance of different analytical laboratories is the purpose, separate sections from a single filter paper should not be used to compare analytical accuracies; rather, the entire filter should be ashed and put into solution and aliquots sent to each laboratory for analysis. This approach will allow for comparison of the analytical method but does not take into account errors incurred in the preparatory procedures at different laboratories. The relative accuracy of the method of atomic absorption as compared with the wet chemical method of employing atomic absorption was better than that of the wet chemical method but with the evidence at hand, this suggested advantage could not be concluded as statistically significant at the 0.05 level.