The Geant4 toolkit does not include solid state effects like electron/hole pairs, lattices, dispersion relations, etc. G4 treats all materials as amorphous continua with spatially uniform properties.
In particular the G4 ionization processes (which is where you might imagine e/h production to occur, tend to have a lower validity range in the tens of eV, well above the semiconductor bandgaps.
I maintain, and continue to develop, an add-on library called “G4CMP” (https://github.com/kelseymh/G4CMP.git) which is a preliminary attempt to support these sorts of things. It has some known limitations and missing features, but might possibly be useful for your case.