A MODEL FOR THE ELECTROLUMINESCENCE OF POROUS N-SILICON

At the present stage of the exploration of porous silicon, it seems impossible to explain all experimental observations in one unified model. It is necessary to put the main emphasis on some special aspects. We present a structure model for the electroluminescence (EL) of our porous silicon devices of the first generation that is consistent with the experimental observations we made on this material. The devices are fabricated by light assisted electrochemical etching of n-silicon. Semitransparent metal contacts are applied to allow electrical contacting. The photoluminescene (PL) observed from the samples seems to have its origin in small clusters of silicon with typical dimensions <50 Angstrom. The main intensity of PL is emitted from a layer located about 5 mu m below the surface of the sample. In contrast to this observation, we found that electroluminescence light is emitted from a 1-2 mu m wide layer at the top of the sample. In this layer we find crystalline silicon structures >100 K surrounded by oxide. We interpret these silicon structures as constricted wires that allow current transport. The constrictions should have dimensions that allow an increase of the effective band gap due to quantum size effects. It seems very likely that EL is generated in the constrictions where an increased band gap can be assumed. Following the smart quantum confinement model for PL, we think that the charges that are forced to pass the constrictions are captured by surface slates and the red luminescence observed is caused by a radiative recombination at these surface states. Other groups explain the luminescence mechanism by excitonic recombination in quantum structures or by special surface chemistry. Each of these alternatives is consistent with our structure model.