Measurement and Monte Carlo prediction of diagnostic tungsten x-ray spectra

A hypothesis that diagnostic tungsten x-ray spectra can be accurately predicted via the Monte Carlo method is proven. A modified general purpose Monte Carlo radiation transport code is thoroughly benchmarked and presented for diagnostic radiology applications. The Monte Carlo code represents a new and powerful tool for simulating a variety of complex radiological physics applications. Detailed measurements of benchmark quality diagnostic tungsten x-ray spectra in the 30--150 kVp range are described. The benchmark spectra are presented in useful tabular form with a complete error analysis. Monte Carlo matrix weighting is introduced and validated for unfolding distorted detector pulse height distributions into x-ray spectra. A laser alignment technique for quantitative alignment of primary x-ray beams passing through pinhole collimators is presented. A method for characterizing the effects of anode pitting on primary x-ray spectra is described. A heuristic analysis of measured spectra and air kerma data is used to develop spectra-specific ion chamber calibration factors with better than two percent accuracy. Different methods for re-binning x-ray spectra are explored and the maximum energy bin widths that should be allowed in rendering diagnostic x-ray spectra without introducing significant errors in common diagnostic physics calculations is determined.