Metal constituents in orange juice have been determined by four atomic spectral methods-flame atomic absorption spectrometry, flame atomic emission/fluorescence spectrometry, dc plasma atomic emission spectrometry, and inductively coupled plasma atomic emission spectrometry. Substantial differences between methods were observed with regard to speed, convenience, precision, detection limits, and interpretation of analytical calibration curves near the detection limits. In general, however, all atomic spectral methods used in the comparison gave plausible agreement in the values obtained. Some of the present authors (1) have previously published a study on the use of flame atomic absorption spectrometry (FAAS) to determine several elements in orange juice. Since the time of that publication, direct current plasma atomic emission spectroscopy (DCPAES) and inductively coupled atomic emission spectroscopy (ICPAES) have become increasingly popular as sources for atomic emission spectrometry (2-8), and interest is growing in flame atomic fluorescence spectrometry (FAFS) as better instrumentation becomes available (9-12). It seemed of timely interest to compare the measurement of trace metals in a practical analytical problem such as the analysis of orange juice using these atomic spectral methods and to compare the results with already published values obtained by flame atomic absorption spectrometry (FAAS). In this paper, the authors compare values obtained by DCPAES, ICPAES, and FAFS with those from FAAS.Also in this paper, comparative analytical figures of merit using two sets of replicate samples dissolved on the one hand in dilute hydrochloric acid and on the other in dilute nitric acid are given. EXPERIMENTAL Apparatus. The ICPAES and FAFS systems used in this study have already been described (13,14). It should be noted that the signal of the FAFS system (wavelength modulated) consisted of the sum of atomic fluorescence and atomic emission.The DCPAES (Spectraspan II) system also has been described by the commercial supplier, Spectrametrics, Inc., and by others who have used the equipment. The Spectraspan II was modified in this study by the replacement of the Spectrajet II plasma (15,16) source with the recently designed Spectrajet III. The instrumental parameters and source conditions are listed in Table I. In Table II are listed the wavelengths used for the various elements.Reagents. Analytical grade reagents were used in preparing all standard solutions. Copper, iron, magnesium, manganese, and zinc metals were dissolved in acid and diluted to volume. Barium, calcium, potassium, rubidium, and sodium standard solutions were prepared from the dry carbonates. To provide a sample-matching environment for the trace elements-barium, copper, iron, manganese, rubidium, and zinc-a solution was prepared con-
