Analysis and Design of the Si3 Sensor : A study of a silicon SPECT detector using Monte-Carlo methods

Sammanfattning

Single Photon Emission Computed Tomography (SPECT) scans are a type of medical imaging where a detector receives photons emitted from radioactive molecules that have been injected into a patient's body. Most SPECT scanners today use collimators to determine photon incident direction. These collimators result in large amounts of lost photons which lead to long exposure times and reduced resolution. An alternative that could lead to an increase in efficiency of up to 10^6 times is the Si3 sensor, a 3D silicon detector that uses Compton scattering to calculate the incident angle. The goal with this thesis is to quantify the Si3 sensor qualities by sensor and object geometries and properties, which is achieved through a Monte-Carlo simulation of the Si3 sensor. We examine the dependence of sensor width, sensor depth, sensor dead-space, and radiation source geometries on photon chain quality and proportion. We find that there is a large degree of dependence between the various geometries and properties. In a sensor with infinite depth, the highest possible proportion of photons with complete chains is approximately 60%, while it is approximately 24% for incomplete chains. Each isolated examined parameter provides interpretable graphs, but the final optimal sensor design depends on the real-world costs and values of the various factors. In the end, the optimal depth of the Si3 sensor likely exists in the 5cm-10cm interval.