with David Vonk, Honors Student, Sabinus Ekeh , graduate student
X-rays and electrons between about 4 MeV and 20 MeV are used to destroy malignant tumors. Successful treatment requires careful calculation of the dose distributions. This can be complicated by the presence of inhomogeneities such as bone and air passages.
Are simple algorithms based on linear propagation sufficient or should sophisticated Monte Carlo codes borrowed from particle physics (such as EGS4 + derivatives, Penelope and MCNP) be used?
We decide to look at electrons first because even for X-rays the most important process is the production of high energy electrons by Compton scattering. The minimum voxel size is 1mm even in stereotactic radiosurgery, the most demanding radiation treatment. This is many thousands of scattering mean free paths for electrons. Since the electrons are of such high energy the scattering angles are small and the multiple scattering gives a Gaussian distribution as predicted by the Fermi Eyges equation.
The size of the voxel is also of order the range of an electron of energy 400kV. This means that it is a waste of time following electrons below this energy and worrying about secondary processes such as Auger emission. A simplified Monte Carlo scheme including Compton Scattering, Brehmstrahlung and electron inelastic scattering should be sufficient for Medical Physics applications.
We also calculated dose distributions using different codes for various model inhomogeneities.
In collaboration with Los Alamos National Lab (Drs Ed McKigney and Rico del Sesto), funded by Los Alamos National Lab
Since 9/11/2001 there has been considerable interest in detecting concealed special nuclear materials (U 235, Pu 238). In principle these can be distinguished by gamma ray spectroscopy. Simple detectors using plastic scintillators do not have sufficient energy resolution, high purity Ge detectors are too costly and require liquid N2. We are part of a collaboration investigating new detector materials.
The detector resolution is given by
Where e is the average electron hole pair energy, E is the energy of the gamma ray and F is the Fano factor, the ratio of observed variance in the number of counts from a gammy ray to the expected variance from Poisson noise
We have shown using the Penelope Monte Carlo code that inelastic electron scattering is by far the most significant process contributing to the electron hole pair generation. Investigation of the cascade produced in these inelastic scattering collisions indicates the competition between plasmon (collective excitation) scattering and single electron scattering can control the Fano factor
Work with Raman Narayan (graduate student)