The necessity of recognizing all events in X-ray detection

• Published in: Applied radiation and isotopes 2010-04, Vol.68 (4), p.561-565
• Main Authors: Papp, T., Maxwell, J.A., Papp, A.T.
• Format: Article
• Language: English
• Subjects: Computer-Aided Design; Equipment Design; Equipment Failure Analysis; Radiation Dosage; Radiometry - instrumentation; Reproducibility of Results; Sensitivity and Specificity; X-Rays
• ISSN: 0969-8043; 1872-9800
• DOI: 10.1016/j.apradiso.2009.10.048

In our work in studying properties of inner shell ionization, we are troubled that the experimental data used to determine the basic parameters of X-ray physics have a large and unexplainable scatter. As we looked into the problems we found that many of them contradict simple logic, elemental arithmetic, even parity and angular momentum conservation laws. We have identified that the main source of the problems, other than the human factor, is rooted in the signal processing electronics. To overcome these problems we have developed a fully digital signal processor, which not only has excellent resolution and line shape, but also allows proper accounting of all events. This is achieved by processing all events and separating them into two or more spectra (maximum 16), where the first spectrum is the accepted or good spectrum and the second spectrum is the spectrum of all rejected events. The availability of all the events allows one to see the other part of the spectrum. To our surprise the total information explains many of the shortcomings and contradictions of the X-ray database. The data processing methodology cannot be established on the partial and fractional information offered by other approaches. Comparing Monte Carlo detector modeling results with the partial spectra is ambiguous. It suggests that the metrology of calibration by radioactive sources as well as other X-ray measurements could be improved by the availability of the proper accounting of all events. It is not enough to know that an event was rejected and increment the input counter, it is necessary to know, what was rejected and why it happened, whether it was a noise or a disturbed event, a retarded event or a true event, or any pile up combination of these events. Such information is supplied by our processor reporting the events rejected by each discriminator in separate spectra. Several industrial applications of this quality assurance capable signal processor are presented.