Study of unfolding methods for X-ray spectra obtained with CDTE detectors in the mammography energy range

Quality control parameters for an X-ray tube strongly depend on the accurate knowledge of the primary spectrum, but it is difficult to obtain it experimentally by direct measurements. Indirect spectrometry techniques such as Compton scattering can be used in X-ray spectrum assessment avoiding the pile-up effect in detectors. However, an unfolding method is required for this kind of measurements. In previous works, a methodology to assess primary X-ray spectra in the diagnostic energy range by means of the Compton scattering technique has been analysed. This methodology included a Monte Carlo simulation model, using the MCNP5 code, of the actual experimental set-up providing a Pulse Height Distribution (PHD) for a given primary spectrum. It reproduced the interaction of photons and electrons with the Compton spectrometer and with a High Purity Germanium detector. In this work, a CdTe detector is proposed instead of the HP Germanium. CdTe detector does not require a liquid nitrogen cooling system, but its resolution is poor for the same energy range and its efficiency comes down for energies greater than 55 keV being 70% at 90 keV. In despite of these disadvantages, CdTe detector has been considered due to its low cost and easy handling and portability. The model can provide a PHD and a Response Matrix, for different X-ray spectra, taken from the IPEM 78 catalogue. The primary spectrum can be estimated applying the MTSVD (Modified Truncated Singular Value Decomposition) and the Tikhonov unfolding method. Both unfolding methods cause some loss of information on the reconstructed primary spectra. In this paper, a comparison of the ability to obtain primary spectra using both MTSVD and Tikhonov unfolding methods has been done. As well a sensitivity analysis in order to test the proposed unfolding methods when they are applied to PHDs obtained with the MCNP model has been developed. A variation on parameters such as target materials and voltages over the mammography energy range has been considered. (authors)