Information-Based Expert Systems for Atomic Emission Spectroscopy

Abstract

The development of the direct-current plasma echelle/CID spectroscopic system for atomic emission spectroscopy (AES) provides new alternatives for automated system control and data analysis. With this system, the concept of the “intelligent” spectrometer becomes tangible. The echelle/CID system simultaneously gathers a wealth of spectral information over a wide spectral region. The mechanical stability of the system and the absence of moving parts give rise to reproducible wavelength assignment. The large amount of spectral information acquired has led to the development of information-based expert systems for AES: automated qualitative analysis, semi-quantitative analysis, and an “on the fly” matrix-dependent line selection. These algorithms are effective in situations where there is a large variability among samples. The analytical power of these routines is heavily dependent on their utilization of the large database and the use of fundamental spectroscopic principles. Examples of the use of these algorithms in environmental monitoring, in the identification of chemical waste, in the analysis of geologic materials and steels, and in HPLC-AES are presented.

Instrumentation and automation

“…An expert system was proposed by Galil and Levinsky'l" to examine entire wastewater treatment stages and processes using process kinetics, mass balances, and specific calculations for each process and estimated the treatment cost for each process combination. Pomeroy et al 45 described the development of information-based expert systems to assist in analyzing a large amount of spectral information acquired from atomic emission spectroscopy. Examples of the use of this expert system in environmental monitoring applications were presented.…”

Assessment of a simple partial LTE model for semi-quantitative ICP-OES analysis based on one single element calibration standard

“…The emission signal (counts/s) from an ion line measured at l ij due to emitting analyte ions of a particular element is given by: TR  %ions  %excited,i  A ij  t  %light  ds (1) where TR is the analyte transport rate into the ICP (atoms/s), % ions is the percentage of atoms in the original sample that are converted into elemental ions in observation volume of the plasma, %excited,i is the percentage of analyte ions in the plasma observation volume that are in the excited state corresponding to the upper level of the transition that produced light at wavelength l ij , A ij is the Einstein transition probability (photons/s per excited ion), t is the time each analyte ion spends in the observation volume, %light is the percentage of light emitted from analyte ions at l ij that reaches the detector and ds is the number of counts per photon produced by the detector. For an ion line, the sensitivity (counts/s per ppb) can be described as the product:…”

“…One approach to semi-quantitative ICP-OES analysis is to first empirically measure sensitivities of all elements and emission lines of interest using a standard that contains all of the elements of interest at a known concentration one time. 1,2 Subsequently, the sensitivities could be used without recalibration for concentration measurements. Changes in any process described in eqn (2) may result in a change in sensitivity.…”

Influence of the operating parameters and of the sample introduction system on time correlation of line intensities using an axially viewed CCD-based ICP-AES system

“…As stated above, this can be easily obtained by applying long integration times and by using cooled detectors. The second condition requires the use of multichannel detection systems, such as CID [21], SCD [22] and CCD [23] devices, and the application of similar acquisition times for the analytical and reference lines. The third condition is probably the most difficult to reach, since time correlation between signals depends on several factors, such as the operating conditions [18], the energy and type (atomic/ionic) of the emission lines [18] and, possibly, on the analytical concentration and the matrix.…”