Natural Width of X-Ray Lines

Another consideration is the natural line width and satellite structure of the x-ray line used. Titanium (TiKa=4510.9 eV) has seen limited use (12) for non-destructive depth profiling, but the observed spectra are complicated by the TiKa satellite structure and the large natural line width of 2.0 eV (13). 

Figure 13.2—Simplified schematic of an atom showing the origin, and the Siegbahn nomenclature, of some fluorescence radiation processes caused by impact of a photon having a high energy. The position of the spectral line is not significantly influenced by the chemical combination in which the atom is found. For example, the Kat line from sulphur is observed at 0.5348 nm for S + and at 0.5350 nm for S°, yielding a shift of 1 eV, which is comparable to the natural line width for X-rays.

The width of the spectral line equals the sum of the widths of initial and final levels. Due to the short lifetime of highly excited states with an inner vacancy, their widths, conditioned by spontaneous transitions, are very broad. The other reasons for broadening of X-ray and electronic lines (apparatus distortions, Doppler and collisional broadenings) usually lead to small corrections to natural linewidth. 

The energy calibration was obtained with Si and S Ka fluorescence X-rays excited by means of an X-ray tube. Because the natural widths of the fluorescence X-rays exceed the experimental resolution at least by a factor of two, the response function of the spectrometer was determined from the narrow antipro-tonic transitions pHe(5g — 4/) and pNe(13p — 12o), lines which are not affected by the strong interaction. For the 1.7 keV Balmer a transition from pH, quartz crystals are the only possible choice for the Bragg crystal. The theoretical limit for the resolution of 180 meV was missed by a factor of 1.7 (Table 1). In the case of pD, a silicon crystal was used because of its higher reflectivity. Here, the theoretical limit for the resolution of 360 meV was reached [21]. 

Except when fitting peaks containing very few counts, the assumption of a gaus-sian peak shape is clearly inadequate. Figure 6.7 illustrates the principal effects contributing to the peak shape. If the resolution broadening caused by the x-ray spectrometer is deconvoluted from the spectrum, the characteristic x-ray line approaches a delta function with a width limited only by the natural line width. This narrow line is superimposed on the background from the specimen. Occasionally, when the characteristic photon is detected in the Si(Li) detector, not all... 

---