Statistical evaluation of the performance of a new programmed-scan monochromator for multielement determinations by atomic emission

Abstract

The reproducibility of a programmed-scan monochromator with stationary dispersion optics was evaluated by means of analysis of variance (ANOVA). The spectrometer used an optical multiplexer coupled with glass-fiber optic light-guides to a multiple entrance-slit spectrometer employing a photomultiplier as the detector. With this spectrometer, 15 emission intensity measurements at the lithium resonance line wavelength (670.7 nm) were collected for five rotations of the optiplexer mirror under four different emission situations: flame background emission at 670.7 nm, lithium emission from an acetylene—air flame in the absence of an ionization buffer, lithium emission from an acetylene—air flame in the presence of an ionization buffer, and tungsten lamp emission at 670.7 nm. For all four situations, the ANOVA results showed that instrumental changes which occurred during mirror rotation in the optiplexer were a significant source of signal variation compared with factors not associated with mirror rotation, i.e., photon shot noise, source fluctuation noise, and electronic drift. The actual magnitude of the signal variability introduced during mirror rotation, however, was found to be quite small, producing an average relative standard deviation of only 0.76% for the signal.