Anal. Chem. 1907. 59. 945-950 945 of systematic error associated with interferences either in the determination of metals in one or more of the fractions or in the sequential extraction procedure itself.Precision was assessed by performing sequential microwave extractions on replicate (n = 31) sediment samples from California Gulch, CO (5CA). These samples were extracted in seven separate batches of four to six replicates by three technicians who had little or no experience working with this system. Data from these experiments are presented in Table I11 for major elements of Ca and Fe and minor elements Mn, Pb, and Zn. Excluding the exchangeable fraction, the average percent relative standard deviation (% RSD) for any particular fraction was about 11 % . The average % RSD for total metals as determined from the sum of the metals in each of the fractions was about 5%. In the exchangeable fraction Fe and P b average below the detectable limit for the determination of those metals by flame AAS, while the other three elements have rather large %RSD's ranging from 60% to 100%. This large variation is due in part to the fact that concentrations of metals in the exchangeable fraction are frequently being measured near the detectable limit for this procedure. This work shows that microwave heating techniques can be valuable in the study of the speciation of metals within binding fractions of sediments. Use of the microwave technique reduces the time required for fractionation from about 24 h to about 4 h, while at the same time producing results that are comparable to the conventional procedure. Recovery of iron from the residual fraction is virtually complete if the aqua regia/HF digestion is used rather than HNO,/HCl alone. The procedure appears to give the best results for minor and trace elements since the extraction rate experiments suggest that the efficiency of separation of these elements from one fraction to another is much greater than for major matrix elements. This paper reports results obtained from metal determinations made by flame AAS using direct calibration after correction of the absorbance for the signal from an appropriate reagent blank. Other than the correction for reagent blank, no special effort was made to identify and compensate for interferences. Accuracy and precision could possibly be improved by using the method of standard additions.
ACKNOWLEDGMENTThe authors thank the following undergraduate students at USC who also contributed to this work Cheryl Passarelli, Weliershaus, S.; Forstner, U. Environ. Techno/. Lett. 1982, 3, 199. Farmer, John G.; Gibson, M. J.; Loveli, M. A. Miner. Environ. 1983, 5 , 57. Brannon, J. M.; Engler, R. M.; Rose, J. M.; Hunt, P. G.; Smith, I., U.S.