EXTENDED CALIBRATION RANGES FOR CONTINUUM SOURCE ATOMIC-ABSORPTION SPECTROMETRY WITH ARRAY DETECTION

Citation
Jm. Harnly et al., EXTENDED CALIBRATION RANGES FOR CONTINUUM SOURCE ATOMIC-ABSORPTION SPECTROMETRY WITH ARRAY DETECTION, Spectrochimica acta, Part B: Atomic spectroscopy, 51(9-10), 1996, pp. 1055-1079
Citations number
15
Categorie Soggetti
Spectroscopy
ISSN journal
05848547
Volume
51
Issue
9-10
Year of publication
1996
Pages
1055 - 1079
Database
ISI
SICI code
0584-8547(1996)51:9-10<1055:ECRFCS>2.0.ZU;2-V
Abstract
Computer modeling has been used to construct calibration curves and re lative concentration error plots for continuum source atomic absorptio n spectrometry with array detection and graphite furnace atomization. Model results are compared with experimental results obtained with a l inear photodiode array detector. The model uses a Lorentzian absorptio n profile convoluted with a rectangular entrance slit (25, 50, 100, 20 0, or 500 mu m wide) and detected with an array of pixels (each 25 mu m wide) using a high resolution spectrometer. Transient furnace signal s are modeled as triangular functions with a half-width of 2 s whose h eight and area are linearly dependent on concentration. With detector read noise limiting (characteristic of a photodiode array detector), t he best signal-to-noise ratios have been obtained with a 500 mu m entr ance slit width and wavelength integrated absorbance (i.e. integration of absorbance over the whole absorption profile). The shapes of the m odeled calibration curves agree well with those theoretically predicte d and those obtained experimentally. Useful calibration ranges approac hing six orders of magnitude of concentration have been achieved using a single calibration curve and integrating over a spectral region equ ivalent to 60 times the half width of the absorption profile (about 0. 16 nm for Cd at 228.8 nm). When concentration is normalized by the int rinsic mass, all elements give the same curve shapes with the inflecti on point, from a slope of 1.0 to 0.5 (on a logarithmic scale), determi ned by the stray light. A hyperbolic function has been developed which accurately fits the modeled and experimental data. With photon shot n oise limiting (characteristic of a charge coupled device), the signal- to-noise ratio is much less dependent on the entrance slit width. With a 25 mu m entrance slit width, wavelength selected absorbances (i.e. absorbances computed for selected pixels or wavelengths) have been use d to construct three calibration curves covering six orders of magnitu de of concentration.