Samenvatting

In the medical community, X-rays are an important tool in imaging the human body. In principle X-ray images are 2 dimensional images. By measuring such X-ray images at multiple angles, one can construct the 3 dimensional inner structure of the sample. This 3 dimensional structure is created by applying a computerized tomography algorithm (CT-reconstruction) on the multiple angle X-ray images. Currently, artifacts appear in the 3 dimensional structure because little or no spectral information is used in the CT-reconstruction. At Nikhef, research is carried out to study the possibility to improve CT-reconstruction by exploiting the spectroscopic capabilities of pixel detectors from the Medipix family.

In this thesis, an energy calibration is performed for one of these detectors, a Timepix detector. The calibration is performed using X-ray fluorescence emitted by several pure materials. A Timepix detector consists of four pixel devices, each consisting of 256 by 256 pixels. During this project, the Timepix detector is calibrated on a global (per device) and per pixel scale, using C++ analysis software.

The first version of the per-pixel analysis software had an efficiency of less than 75%, because the ranges were initialized on a global scale. The global range also caused several pixels to repeatedly fall out of range, leading to pixels without a calibration. Therefore the algorithm is enhanced during this project, leading to a new efficiency of at least 90%. Next to this the new algorithm is completely independent from any global parameters, leading to almost 100% of the pixels to be calibrated correctly.