Electrothermal atomization—the way toward absolute methods of atomic absorption analysis

Considering the currently prevailing opinion that electrothermal atomizers are very susceptible to matrix interferences, an opinion which is contradictory to the optimistic forecasts of the first publications, the general status of the problem to date has been investigated.The first part of the paper deals with the ideal models of the two basic types of electrothermal atomizers, viz., those of the semi-enclosed and those of open type. The graphite cuvette and the Massmann furnace have been selected for discussion in their two commercial versions—the HGA and the CRA—of the first type of atomizers; the West filament and the combined atomizers—rod-in-flame and capsule-in-flame—of the second type. The models describing the distribution of atoms in the absorbing zone have been compared and the data obtained have been used to interpret the experimentally observed differences in the sensitivity of real atomizers.The second section of the paper discusses the conformity of real atomizers relative to ideal models and to the requirement of correctly recording the absorption by means of integrating the pulses. A marked time and space non-isothermality of the commercial atomizers has been established which makes it impossible to measure the integral absorption correctly. The graphite cuvette and the combined atomizers best meet the requirements of the ideal models. On the basis of the data obtained, the possible ways of bringing the semi-enclosed atomizers closer to ideal models have been explored. In this connection, the possibility of using a graphite platform for vaporizing samples in tube furnaces as well as using the temperature-controlled furnace and the pulse heated graphite furnace with a capacitance power supply has been examined. The application of the last method ensures time and space isothermality of the absorbing layer and reduces by 10–100 times the electric power input.The third section of the paper examines the thermochemical laws governing the possible chemical effects arising from the interaction between the analytes, on the one hand, and the furnace walls, the gaseous atmosphere and the matrix, on the other. A critical review has been carried out of the results of some recent publications on investigation of sample vaporization in graphite furnaces, which reveals the fact that the temperature of absorption appearance is rarely connected with the heat of vaporization of the free element. In most instances it is determined by the sum of the heat of decomposition of the non-volatile carbide and the heat of vaporization of the free element. Thermochemical examination of the stability of the compounds formed in the gaseous phase revealed the fact that besides forming monoxides, there is the possibility of forming monocyanides. In addition, the presence of large quantities of halogens causes partial binding of the analyte as gaseous monohalides, mainly monochlorides. In order to eliminate the latter effect, it has been suggested to employ higher atomization temperatures or to bind chlorine in a stable lithium monochloride. The efficiency of the proposed methods has been tested by experiment.Our research has shown that the main reasons for the unsatisfactory status of the problem concerning the matrix effects, are connected with the use of the amplitude (peak) method of recording the absorption, with the time and space non-isothermality of the absorbing layer of the commercial atomizers, and with the formation of gaseous monohalides. All these problems may be eliminated on the basis of theoretically proved and tested methods, some of which are discussed in this paper.