Recently a potential customer decided not to proceed with a purchase of GamBet. He had spoken to a a user of another Monte Carlo package who claimed that Penelope (the radiation physics core package of GamBet) underestimated electron backscatter ratios at low energy. Because such rumors are self-perpetuating, I decided to carry out calculations to put the matter to rest.
The low-energy validity limit for Penelope (as stated in the GamBet manual) is several hundred eV. Accordingly, I used results in the incident electron energy range of 0.5-6.0 keV for comparison. The paper A.M.D. Assa’d and M.M. El Gomati, Backscattering coefficients for low energy electrons, Scanning Microscopy 12, 185 (1998) describes measurements of the backscatter fraction η from carefully-prepared surfaces (10^{-10} mbar vacuum with ion bombardment cleaning). I prepared a set of GamBet runs for aluminum and silver targets. At the lowest energy, the electron penetration depth was less than 10 nm. Accordingly, I used a small surface element width of 20 nm. I initiated 50,000 electrons at the surface of the target at normal incident. To compare to the conditions of the experiments, I used GenDist to filter the particle escape file to include only electrons leaving the front face with kinetic energy >50 eV.
The two figures show comparisons of GamBet/Penelope results (colored squares) to the experimental results (X’s) for clean aluminum and silver surfaces. The code did an excellent job reproducing absolute values and data trends. An interesting fact is that the GamBet prediction for aluminum was slightly higher than the experimental value at the lowest energy. My conclusion is if a Penelope simulation reports a low value of η, it is either not set up correctly or not interpreted correctly. The irony of worrying excessively about these numbers is that the difference in η that results from a less-than-perfect surface (e.g., nanometer scale surface contamination) is far greater than the discrepancies in the graphs.
Here are links to the input files if you have GamBet and you would like to try the calculations: lowenergy.min, lowenergy_aluminum.gin and lowenergy_silver.gin. You can get a PDF copy of the Assa’d/El Gomati paper at http://www.aber.ac.uk/~ecmwww/journal/smi/pdf/smi98-19.pdf.
