Determination of the Suitability of Photon-Induced X-Ray Fluorescence Analysis for the Quantitation of Selected Low-Atomic Trace Elements in Biological Materials.

The technique of photon-induced, energy-dispersive x-ray fluorescence spectroscopy was investigated in order to determine its suitability as a routine clinical procedure for the analysis of trace levels of titanium, calcium and magnesium in tissue. One part of the investigation consisted of Monte Carlo modeling of two systems, one employing an Fe-55 radionuclide excitation source, and the other using a 55-kVp x-ray tube and an iron secondary fluorescer excitation system. The ability of each simulation to predict relative responses for each element was assured by experimentally analyzing a series of starch pellets doped with low concentrations of the three elements. The experimental data yielded values for each element of minimum detectable concentrations for the experimental configurations. Geometrical and source constants were adjusted in the Monte Carlo simulations and the improvements in minimum detectable concentrations for the optimized configurations were predicted. For the experimental radionuclide system, the results indicated higher peak-to-background (signal-to -noise) ratios and lower minimum detectable concentrations than for the x-ray tube system. At the concentration levels found in tissue, however, only Ca analysis is feasible. The Monte Carlo results predicted reductions in the minimum detectable concentrations by as much as a factor of ten through proper choices of geometrical constants. With a properly designed system, the predicted minimum detectable concentrations for Ti, Ca and Mg were 0.24-, 0.78- and 277.- ppm, respectively. Tissue analysis for both Ti and Ca would be possible with such a system. However, the improvements in Mg detectability were not sufficient to allow its analysis at the levels found in tissue. Computer predictions on the effects of different radionuclide sources yielded only minor improvements in Mg detectability. For the experimental x-ray tube--secondary fluorescer system, the data yielded higher minimum detectable concentrations than for the radionuclide system, although Ca analysis was still feasible. This experimental system was more flexible than the radionuclide system, so most of the geometrical optimization was done empirically. Computer optimization of the apparatus consisted of increasing the detector size and resulted in only minor improvements. The predicted minimum detectable concentrations for Ti and Ca were 11.2 - and 38.5- ppm, respectively. Mg at 10,000 ppm was not detectable with this experimental system and thus no predictions of minimum detectable concentration could be made. The greatest increase in signal-to-noise ratios would be obtained by using a lower-energy x-ray tube with a Be window to allow emission of the low energy ((LESSTHEQ) 10 keV) x -rays. Such a source has the potential to lower the minimum detectable concentrations to levels below those of the radionuclide system.