Abstract
As an advantageous example of application of X-ray emission analysis using radioactive sources, we have investigated the rapid analysis of cement raw mixture. Having succeeded in quantitative analysis of lime, silica, alumina and ferric oxide in the samples, we are now constructing the simultaneous analyzer of the all elements for the detector in the computer control system of cement raw mix composition.
Of these developments, we discussed here the problems concerning the pulse height analysis for non-dispersive X-ray spectrometry; first, spectrometry of adjacent elements, then, stabilization of the spectrometer.
In the former, pulse height channel setting technique utilizing the cancel effect was tried. Namely, suppose the case when only the content of lower-Z-element B increases, the count-decrease of higher-Z-element A is canceled by the count-increase of B. This effect appears between two adjacent elements where X-rays of A is resonantly absorbed by B. Thus, by the channel where this effect is valid, content of A can be easily determined independently of B, when. Z difference of both elements is adequately large, or when excitation-detection-efficiency of A is higher sufficiently than B. In fact SiO2 was analyzed without any effect by MgO or Al2O3 in cement raw mix.
In the next problem, on examination, it was proved that level stability of a single channel pulse height analyzer is fairly good and the most probable variation in pulse height arises out of a proportional counter (gas amplification in the counter varies sensitively by temperature or pressure of the gas) . So the detecting methods of pulse height variation using pulse height analyzer were studied, after all the method of H, de Waard was adopted (except that vacuum tube diode was displaced by silicon diode in pumping and switching circuit) . As the remarkable and symmetrical peak was produced by calcium in the spectrum of our example, fortunately, the vibrating channel was set at this peak level for the detection of pulse height variation. An feed back control of pulse height was carried out with the improved DC amplifier circuit (feed back amplifier as a controller) . Because of the lack of the amplifier gain, resetting operation by addition of “integration” was not effective. However, the results (stabilization factor and response characteristics) were better than those of H. de Waard. Stabilization factor was about 33 using the peak of Ca (half width=26%, total count rate=170 cps) .
