Electron mass scattering powers: Monte Carlo and analytical calculations.

Values of electron mass scattering power, T/p, for various materials have been calculated by using the EGS4 Monte Carlo system and by integration of the Molière multiple-scattering distribution. The energy range covered is 0.5-100 MeV. Monte Carlo calculations test the concept of T/p "experimentally" and assess the contribution to electron mass scattering power from effects such as Moller scatter and energy-loss straggling. The Monte Carlo results agree within 2% with the analytical results calculated from Molière multiple-scattering theory at energies less than 20 MeV for high-Z materials and for energies less than 50 MeV for low-Z materials. At higher energies the Monte Carlo calculations include the effects of bremsstrahlung production which can significantly increase values of T/p. For low-Z materials and electron energies less than 60 MeV, the Monte Carlo calculated T/p values are generally 22% higher than those given by ICRU Report 35, while those for high-Z materials and energies less than 25 MeV are found to be consistent (within 1%) with ICRU Report 35. The effects of Moller scatter, which significantly affect T/p for low-Z materials, as well as bremsstrahlung effects, are included in the present Monte Carlo calculations. If the tabulated T/p data of ICRU Report 35 are modified to include the Moller scatter effect, then for energies less than 60 MeV they are generally 6% less than the present Monte Carlo data for low-Z materials as well as for copper. It is shown that T/p is a well-defined constant over an appropriate range of slab thickness except when bremsstrahlung effects are significant. It is found that T/p is proportional to E-n, where n is in the range of 1.5-2.0 for the energies considered here. The Monte Carlo calculations are shown to agree well with various relevant experimental measurements. Accurate T/p data, which should include the effect of Moller scatter, are necessary in electron-beam treatment planning, especially for a small field size. The choice of the depth step in the implementation of pencil-beam codes should not violate the slab-thickness limits for T/p data.