Infrared reflectance anisotropy spectroscopy of Si(111)-2X1: Surface excitons and polarons

The experimental data regarding the energy distribution of dangling-bond electronic states at the Si(111)-2x1 surface, with special attention to the gap between filled and empty states, are revisited. Clearly some of the results are not compatible. Theorists almost unanimously conclude in favor of a strong (about 0.25 eV) excitonic energy in the optical transitions between surface states. On the experimental side, on the other hand, conflicting results seem to support either large or negligible exciton energy. In order to reconcile all available data, it has been suggested that in highly doped samples a sizeable gap shrinking should occur. We have performed a reflectance anisotropy spectroscopy experiment in order to verify this hypothesis. The optical peak at 0.45 eV associated to dangling-bond transitions does not shift, in highly doped crystals (N-D=4x10(18) cm(-3)), within the experimental accuracy of 20 meV, thereby ruling out the hypothesis of a shrinking of the gap. Under the assumption that all experimental results are correct, an alternative explanation of this puzzling problem, based on the existence of surface polarons together with surface excitons, is suggested. Although very difficult to detect spectroscopically in a direct way, it is argued that polaronic states could play an important role in highly doped samples. A number of possible experiments to clarify the last point are suggested. Accurate theoretical calculations of the polaron energy would also be helpful.