Optimizing thin film X‐ray spectra for quantitative analysis

Abstract

Energy dispersive X-ray detectors are frequently attached to electron microscopes to enable microanalysis to be performed, but because such detectors accept X-rays generated within an appreciable solid angle, the recorded spectra usually include some spurious contributions from the instrument. This paper describes instrumental modifications firstly to reduce and secondly to permit the subtraction of the residual extraneous contributions. The probable accuracy of this subtraction procedure is examined.Results are presented showing the effects of various instrumental modifications on spect.ra from thin specimens and demonstrate that by careful attention to experimental details it is possible to separate the spectrum due to the thin specimen alone from all other extraneous signals.Two test specimens and a number of test procedures for investigating the analytical performance of (scanning) transmission electron microscopes are presented.Sources of error and the method of their correction when the thin specimen Bremsstrahlung is used for quantitative analysis are discussed. I N T R O D U C T I O NIn transmission electron microscopes (TEMs) equipped with energy dispersive X-ray detectors even the best designed systems show some spurious contributions to the X-ray spectra. The recorded spectra include along with the X-rays from the thin specimens, Bremsstrahlung and characteristic X-rays from the instrument and, to a lesser extent, directly recorded electrons. To suppress the instrumental characteristic lines, a solution which has been frequently adopted is to construct part of the specimen holder of carbon or beryllium (Russ, 1977; Saubermann et al., 1981; Panessa et al., 1978); however, this will leave an unknown Bremsstrahlung background which may be high. In materials science where the intensity of the characteristic lines is of prime importance the aim of most workers has been to try to eliminate instrumental peaks, and X-rays from regions of the specimen not under analysis, by careful attention to collimation of the electron beam above the specimen (e.g. Goldstein & Williams, 1978; Nicholson et al., 19771. In the literature little attention has been given to the design of the parts of the instrument beneath the specimen, and to extraneous signals whose intensities are dependent upon the electron scattering introduced by the specimen itself.There are two areas of interest where knowledge of the Bremsstrahlung from the thin specimen alone is important. For analysis of specimens with organic matrices such as biological specimens, in which the matrix produces some characteristic lines which cannot be detected with an energy dispersive detector, the Hall method (1971) of quantitation is frequently used. The method uses an integrated region of the Bremsstrahlung to normalize characteristic line intensities far variations in specimen mass thickness and incident beam current. Thus it is

Improvement of quantitation of biological X‐ray microanalysis

“…In the present study, such extraneous radiation would have arisen primarily from the subsequent scattering of electrons initially scattered by the samples themselves (Nicholson et al, 1982) since the collimation of the electron beam in the EMMA-4 should not allow significant X-ray generation from areas of the specimen or supports outside the main electron beam ('hole' count < 1 c.P.s. ), and the backscattered electrons are deflected away from the detector by the focussing fields (Statham, 1980).…”

Preparation of cryosections with a modified Sorvall MT2B ultramicrotome and cryoattachment

“…Fine mounted pins seem to give fewer problems with static charging of the sections but damage the support films easily. This problem is particularly acute where, for X-ray analytical reasons (Nicholson, 1974;Nicholson et al, 1982), single hole mounts with large areas of unsupported film are used.…”

The continuum normalization method for quantification of X‐ray spectra in biological microanalysis. 1. Generalized bremsstrahlung production cross‐sections and analysis using standards

“…Clearly, the former should be high to reduce statistical errors, whilst the latter should be as low as possible, especially when no correction is to be made for this. Bremsstrahlung correction methods are not considered further here, but see Gupta & Hall (1979), Roomans & Kuypers (1980), Nicholson et al . (1982), Nicholson & Chapman (1983) and Hall & Gupta (1986) for discussions of correction for instrumental contributions to the bremsstrahlung in relation to CN quantification.…”