AOAC Methods in Codex STAN 234 (Preliminary Methods Review)

J. ASSOC. OFF. ANAL. CHEM. (VOL. 65, NO 4, 1982)

981

CAPAR ET AL.:

taining a code that identified the commodity and fortification level. Two of the sponsoring laboratories, FDA and NMFS, analyzed 9-10 samples of each com– rnoorators participating, each commodity was analyzed by 10 different laboratories. Each collaboratnr also received bottles of cleaned ekctrolyte solution (d) and K2SO 4 solution (b) for 2 reasons: to expedite the study for those labo– ratories which did not have the necessary cleanup apparatus and to minimize the reagent blank differences between laboratories. Each collaborator also received a practice sample of f<',1d- and cadmium-fortified green beans or liver and vegetables, a copy of the method, reporting forms, special instructions, and a questionnaire to be completed and returned. Before beginning analysis of the collaborative samples, the collaborators were instructed to ,\naly?.e the practice sample and 3 reagent blanks. l'he practice sample result was required to be within a stated range, and the reagent blanks were required to contain <0.5 ng cadmium/ml. and <1.0 ng lead/mL as measured in the vol-

tammetric cell. All laboratories reported satis– factory blank levels for cadmium. Two labora– tories, 5 and 15, reported high lcild blanks. To evaluate the performance and precision of the various laboratories and their instrumenta– tion., we also asked the collaborators to pt'rform 3 standard additions after measuring the initial sample response, whether or not they routinely used automated instrumentation that quantitates in a different manner. Eighteen laboratories used differential pulse anodic stripping voltammetry (DPASV) at the hanging mercury drop electrode (HMDE) and 2 used linear sweep anodic stripping voltammetry (LSASV) at the composite mercury graphite electrode (CMGE). All analysts had at least 6 months of experience in voltammetric and trace element methodology and, of these, only 4 had less than 2 years of experil•nce. Three labora– tories reported using ASV daily for determining lead and cadmium; 6, weekly; 5, monthly; l, pe– riodically; and the rt•maining 5 laboratories, hardly ever. Furnaces from 7 different manu– facturers were used. All of the furnaces had been calibrated at least 6 months before the col– laborative study, except for 2, which had been calibrated less than a year before thf? study. Fourteen laboratories used quart;: beakers, 4 used Vycor beakers, and 3 used Pyrex beakers; I lab– oratory used both quart;: and Pyrex beakers. Four laboratories reported difficulty in ashing the green bean practice sample and/or the green bean collaborative sample. The source of the difficulty was a brown residue in the ash which, we believe, was iron and not unashed sample. The remaining 16 laboratories reported no dif– ficulty with the ashing. No significant rela– tionship was found between the reports of pre– cipitate formation and the am,)unts of time elapsing between sample dissolution and sample analysis. It is noted, however, that all Qf the laboratories reporting precipitates also reported results that compared favorably with the refer– ence values. Many of the laboratories offered helpful suggestions and, where practical, these were incorporated into the method as reported here. Results and Discussion Twenty laboratories completed this study and their results are shown in Tables I and 2. Results from Laboratory 7 were not included in our sta– tistical evaluation because this laboratory did not use the standard addition method as directed, and because they believed that their instrument was not capable of reliably determining cadmi-

eview