Precision and noise in inductively coupled plasma atomic emission spectroscopy with charge-injection device detection

An inductively coupled plasma (ICP)-echelle optical system coupled with a charge-injection device (CID) detector was evaluated for precision and noise. Simultaneous collection of analyte wavelengths and simultaneous background correction have invalidated underlying assumptions in some traditional models. Signal flicker noise, a major limitation in single-channel detection, was eliminated by the time correlation of analyte and internal standard wavelengths. Once flicker noise was eliminated, most fluctuation was found to be caused by slight wavelength shifts along the axis of the detector, rather than by the theoretically predicted signal shot noise. This fluctuation can be corrected by increasing center subarray sizes or by employing well-matched internal standard lines. In either case, precision of 0.3 to 0.18% relative standard deviation (RSD) was achieved over a time span of 7 h. Background flicker noise was also eliminated by simultaneous background correction, demonstrating a multichannel advantage of ICP detection using charge-transfer devices. In this work, theoretical expressions were developed to incorporate major sources of noise based on the specific readout characteristics of an array detector. These expressions match experimental data extremely well for measurement of relative standard deviation in background (RSDB) and detection limits. These data have then been compared to the single-channel case, which describes photomultiplier tube (PMT) detection, as well as the theoretical limit dictated by background shot noise.