Abstract: Composition profiling—a comparison of surface analysis techniques vs methods involving the detection of sputtered species

Abstract

Most of the methods which are currently being used to obtain high depth resolution composition profiles in solids rely on the continuous removal of surface particles by sputter−etching as a microsectioning technique. Sputter−etching has the advantages of being able to microsection essentially all materials in a clean, controllable, relatively uniform manner, and of being compatible with the vacuum requirements of several elemental analysis procedures. There are two general methods by which the sputter−etch procedure can be used to determine composition profiles: (i) Analysis of the sputter−etched sample surface (Auger electron spectroscopy,1 x−ray photoelectron spectroscopy,2 appearance potential spectroscopy,3 low−energy ion scattering spectrometry4); and (ii) analysis of the sputtered particles following their ejection from the surface (secondary ion mass spectrometry,5 glow discharge mass spectrometry,6 analysis of impact radiation,7 glow discharge optical spectroscopy8). Each of these two general approaches to the problem has some advantages and disadvantages independent of the specific technique employed and these considerations will now be discussed. The methods which involve the detection of the emitted sputtered particles have an advantage in that the recorded data include information about the sputter−etch rate if all the major constituents of the sample are monitored. The data provided by surface analysis methods are independent of the sputter−etch rate and an independent measurement of the sputter−etch rate is needed to calibrate the depth scale. Furthermore, if the sputter−etch rate changes during the analysis, it is possible to extract the correct composition profile from the data provided by the sputtered species detection methods, whereas only an approximate profile can be obtained from surface analysis methods. A second basic difference between these two approaches to composition profiling arises from the fact that in general, the steady−state surface composition of a sputter−etched multiconstituent sample will be different from the bulk composition, whereas the composition of the steady state flux of sputtered species will be representative of the bulk composition. That is to say the data provided by the sputtered species detection methods are independent of the sputtering yields of the individual species, whereas the data provided by the surface analysis methods depend directly on the sputtering yields. Although this effect can be accounted for by a calibration factor for the surface analysis methods, the sputtered species detection methods have an advantage in quantitative analyses. The fact that the sensitivity of the surface analysis methods is independent of the sputter−etch rate, whereas the sensitivity of the sputtered species detection methods is proportional to the sputter−etch rate complicates a comparison of relative sensitivities. However, it is believed that the sputtered species detection methods have a somewhat higher sensitivity with the sputter−etch rates normally encountered in composition profiling work. Although some information about chemical bonding at a surface can be obtained from a study of the nature of molecular sputtered species with the sputtered species detection methods, most of the surface analysis methods are clearly more useful in obtaining information about chemical bonding and electronic structure. Whereas the molecular sputtered species do provide some information about surface chemistry, they also complicate the analysis for the sputtered species detection methods. The surface analysis methods are not influenced by molecular sputtered particles and consequently have an advantage in terms of the complexity of the data and calibration procedures. Finally, if the sputter−etching and the surface analysis can be performed simultaneously as opposed to sequentially in the surface analysis methods, it is suspected that the surface analysis methods will be somewhat less sensitive to the effects of residual gas in the vacuum system than the sputtered species detection methods. The reason for this statement is that two of the sputtered species detection methods (secondary ion mass spectrometry and analysis of impact radiation) are very sensitive to surface contamination, whereas in the other two methods (glow discharge mass spectrometry and glow discharge optical spectroscopy) the residual gas can contribute to the observed signal directly from the gas phase.