Space is a hostile environment for embedded electronic devices on board satellites. The high fluxes of energetic electrons that continuously impact these satellites may penetrate inside their electronic components and cause malfunctions. This is the case, for example, with the multipactor effect which corresponds to an electron avalanche in radiofrequency hardware, and with the ionizing dose effect that occurs in microelectronic components.Taking into account these effects requires high-performant 3D numerical tools, such as codes dedicated to the electron transport using the Monte Carlo statistical method, valid down to a few eV. In this context, the ONERA has developed, in collaboration with the CNES, the code OSMOSEE, dedicated to the transport of low energy electrons (10 eV – 2 keV) in aluminum. For its part, the CEA has developed for silicon the low-energy electron module MicroElec for the code GEANT4.To provide a better understanding of the transport of low energy electrons in solid, the aim of this thesis, in a collaborative effort between ONERA, CNES and CEA, is to extend those two codes to different materials. To describe the interactions between electrons, we chose to use the dielectric function formalism. Since the dielectric functions are obtained through the measurements of optical data, these functions enable to overcome the disparity of electronic band structures in solids, which play a preponderant role at low energy. The validation of the codes, for aluminum, silver and silicon, by comparison with measurements from the experimental set-up DEESSE at ONERA, showed the existence of two transport regimes in the low energy domain studied. At very low energy, the electrons are mainly deflected by nuclei, and, as a result, stay in the first few nanometers from the surface. At higher energy, the electrons go deeper into the solid and less electrons escape from the surface.This result enables us to better understand how the electron emission properties of solids strongly depend on its surface state, such as oxidation, contamination or roughness. These parameters, which may have a significant impact on the electron emission yield, are not usually taken into account in Monte Carlo transport codes, which only simulate ideally flat materials. The modeling of the surface roughness, using the new version of the code MicroElec that was developed during this thesis, has shown that it is possible to reduce the number of electrons emitted by a solid by nearly 80% by adding on top on its surface rough structures, with grooves or checkerboard patterns of great height and of small width. This result offers interesting prospects to limit the multipactor effect in radiofrequency hardware.