摘要：Monte Carlo simulation has been widely used in particle transport theory.Theaccuracy of Monte Carlo simulation of particle transport depends on the accuracy ofphysical processes used in the simulation.Previously some efforts have been done touse accurate and efficient methods to describe the Monte Carlo simulation of electrontransport in general purpose electron transport codes but still there is need to improvethe efficiency and accuracy of physical models used in Monte Carlo electron transportcodes.In Monte Carlo simulation of electron transport, electron bremsstrahlung andmultiple elastic scattering are the most time consuming and complicated phenomena.Therefore, in this work, accurate and efficient methods for electron bremsstrahlung andmultiple scattering have been developed and implemented in a general purpose MonteCarlo Program for Nuclear and Radiation Simulation called SuperMC.Focusing onthese aspects the main research and innovative work of this dissertation are followings.
Firstly, a fast and accurate algorithm for the Monte Carlo simulation of electronbremsstrahlung based on SuperMC has been developed.Efficient and accurate methodshave been used by which the angular distribution and energy of emitted bremsstrahlungphotons are sampled.A discussion of the sampling efficiency and accuracy of this newly developed algorithm is given.Photon energy is sampled according to scaled energyloss differential cross sections (DCSs) tabulated by Seltzer and Berger (1983).A novelhybrid model for photon angular distribution by low as well as high energy incidentelectrons has been developed.The model uses Tsai full form of angular distributionfunction with atomic form factors for high energy incident electrons, while for electrons with kinetic energy less than 500 keV, a simple, efficient and accurate analyticaldistribution function has been proposed.The later uses adjustable parameters determined from the fitting of numerical values of the shape functions tabulated by Kisselet al.(1983).The efficiency of sampling photon energy is at least 80％.Our angularsampling algorithm for high energy electron bremsstrahlung based on Tsai distributionfunction is more efficient (sampling efficiency at least 70％) in the useful photon energyrange than the sampling technique proposed by Alex F.Bielajew et al.(1989)
Secondly, an efficient and accurate model has been developed based onGoudsmit-Saunderson theory of multiple elastic scattering for angular distribution of electrons andpositrons.Differential cross sections of electrons and positrons by neutral atoms havebeen calculated by using Dirac partial wave program ELSEPA.The Legendre coefficients are accurately computed by using Gauss-Legendre integration method.Finally, a novel hybrid method which combines rejection sampling and fitting function forsampling angular distribution has been developed.The model uses efficient rejection
sampling method for low energy electrons (＜ 500 keV) and larger path lengths (103mean free paths).For small path lengths, a simple, efficient and accurate analytical distribution function has been proposed.The later uses adjustable parameters determinedfrom the fitting of Goudsmith-Saunderson angular distribution.The efficiency of rejection sampling algorithm is at least 50％ for electron kinetic energies less than 500 keVand longer path lengths (＞ 500 mean free paths).Monte Carlo Simulation results are
then compared with measured angular distributions of Ross et.al (2008).Comparisonshows that our results are in good agreement with experimental measurements.
The results showed that the proposed algorithm provides accurate and efficient methods for sampling energy and direction of bremsstrahlung photons and angulardistribution of multiple scattering.Hybrid model provides very accurate treatment ofbremsstrahlung photon angular distribution by low as well as high energy incident electrons.It is worth mentioning that the approaches used for sampling energy and directionof bremsstrahlung photons and angular distribution of multiple scattering can be usedfor a wide range of incident electron energies and materials and hence suitable for general purpose Monte Carlo simulations.
Keywords: Monte Carlo; Electron transport; SuperMC; Bremsstrahlung; Differentialcross section; Multiple scattering