Atomic absorption spectroscopy. Stagnant or pregnant

Abstract

WHEN Foil Miller invited me to participate in the Silver Anniversary Symposium on Great Moments in Analytical Chemistry, he suggested that I may care to indulge in some personal reminiscences and comments regarding the development and present status of atomic absorption spectroscopy. I hope the ones I have selected may illuminate, if not answer, the problem posed by the title of may address. I realize that anyone who reminisces is usually so decrepit that his memory is totally unreliable. I shall, therefore, try to avoid too many errors of fact by restricting my comments largely to matters which are documented in reports of CSIRO, in correspondence, or in publications. My initial interest in atomic absorption spectroscopy was a result of two interacting experiences: one of the spectrochemical analysis of metals over the period 1939-1946; the other of molecular spectroscopy over the period from 1946-1952. The interaction occurred early in 1952, when I began to wonder why, as in my experience, molecular spectra were usually obtained in absorption and atomic spectra in emission. The result of this musing was quite astonishing: there appeared to be no good reasons for neglecting atomic absorption spectra; on the contrary, they appeared to offer many vital advantages over atomic emission spectra as far as spectrochemical analysis was concerned. There was the attraction that absorption is, at least for atomic vapours produced thermally, virtually independent of the temperature of the atomic vapour and of excitation potential. In addition, atomic absorption methods offered the possibility of avoiding excitation interference, which at that time was thought by many to be responsible for some of the interelement interference experienced in emission spectroscopy when using an electrical discharge as light source. In addition, one could avoid problems due to self-absorption and self-reversal which often make it difficult to use the most sensitive lines in emission spectroscopy. As far as possible experimental problems were concerned, I was particularly fortunate in one respect. For several years prior to these first thoughts on atomic absorption, I had been regularly using a commercial i.r. spectrophotometer employing a modulated light source and synchronously tuned detection system. A feature of this arrangement is that any radiation emitted by the sample produces no signal at the output of the detection system. This experience had no doubt prevented the formation of any possible mental block associated with absorption measurements on luminous atomic vapours. In an internal report for the period February-March 1952, I suggested that the same type of modulated system (Fig. 1) should be considered for recording atomic absorption spectra. The following extracts from that report may be of interest: “Assuming that the sample is vaporised by the usual methods, e.g. flame, arc, or spark, then the emission spectrum is ‘removed’ by means of the chopper principle. Thus the emission spectrum produces no output signal and only the absorption spectrum is recorded.” “For analytical work it is proposed that the sample is dissolved and then vaporised in a Lundegardh flame. Such flames have a low temperature (2000 K) compared to arcs