We investigated the electronic structure of the Si(111)-7 x 7 surface below 20 K by scanning tunneling and photoemission spectroscopies and by density functional theory calculations. Previous experimental studies have questioned the ground state of this surface, which is expected to be metallic in a band picture because of the odd number of electrons per unit cell. Our differential conductance spectra instead show the opening of an energy gap at the Fermi level and a significant temperature dependence of the electronic properties, especially for the adatoms at the center of the unfaulted half of the unit cell. Complementary photoemission spectra with improved correction of the surface photovoltage shift corroborate the differential conductance data and demonstrate the absence of surface bands crossing the Fermi level at 17 K. These consistent experimental observations point to an insulating ground state and contradict the prediction of a metallic surface obtained by density functional theory in the generalized gradient approximation. The calculations indicate that this surface has or is near a magnetic instability, but remains metallic in the magnetic phases even including correlation effects at mean-field level. We discuss possible origins of the observed discrepancies between experiments and calculations.
