Earlier today Marcus Mendenhall at Vanderbilt reported a bug with Penelope EM. We have observed a separate but perhaps related problem when 100 MeV protons are incident on a 2 micron layer of silicon. Using low energy EM processes, as one would expect, there is no difference between a target with one layer 2 microns thick and another with two layers each 1 micron thick. Using Penelope EM in place of low energy EM with no other changes there is a noticable discontinuity of track density at the imaginary boundary between the two identical material layers. The result is a bi-lobed display of delta electron tracks in the penelope case and a continuous tube of tracks in the low energy em case. If as we suspect there may be a simply geometrical issue, this could be a second manifestation of it. It may, of course, also be a totally separate problem. I have rated this as "major" because if the simple openGL track display is indicative, the density of deposited energy will be notably incorrect in the vicinity of the virtual boundary. However, since we see this in semiconductor-device size samples, it may not affect the majority of users.
The problem is reassigned to the Geant4 developer responsible for Penelope processes. Best wishes, Maria Grazia Pia
I am not able to reproduce with my code that impressive behaviour of electron tracks. Maybe the problem comes from some interference with geometry, given the highly miniaturization of your simulated setup.
