MetallicSi(111)−(7×7)-reconstruction: A surface close to a Mott-Hubbard metal-insulator transition

Abstract

Li adsorption at extremely low coverages ͑10 −3 ML and below͒ on the metallic Si͑111͒-͑7 ϫ 7͒ surface has been studied by ␤-NMR experiments ͑measurement of T 1 -times͒. Instead of increasing linearly with the sample temperature, as expected for a metallic system, the relaxation rate ␣ =1/T 1 is almost constant in between 50 K and 300 K sample temperature and rises considerably above. Comparison with T 1 -times around 900 K ͑ob-served with 6 Li-NMR͒ excludes adsorbate diffusion as the cause of the relaxation rate. Thus the almost temperature independent relaxation rate below 300 K points to an extremely localized and thus narrow band ͑width about 10 meV͒ which pins the Fermi energy. It is responsible for the metallicity of the ͑7 ϫ 7͒-reconstruction. Because of the steeply rising relaxation rate beyond 300 K this narrow band is located energetically within a gap ͑approximately 100-500 meV wide͒ in between a lower filled and an upper empty ͑Hubbard͒ band. Due to its extremely narrow width it can hardly be detected in photo electron experiments. In dynamical mean field theories based on Hubbard Hamiltonians this kind of density of states is typical for correlated electron systems close to a Mott-Hubbard metal-insulator transition.

Disentangling Surface, Bulk, and Space-Charge-Layer Conductivity inSi(111)−(7×7)

“…According to this comparison, Si111-7 7 is clearly not in the metallic range. This finding appears consistent with a recent temperaturedependent and surface sensitive NMR study of Si111-7 7, which suggested that the surface is close to a Mott-Hubbard-type metal insulator transition [13] and it could also be related to the strong electron-phonon coupling [14].…”

Conductivity of theSi(111)7×7dangling-bond state

“…It has also been reported that Si͑111͒7 ϫ 7 is likely a surface close to a Mott-Hubbard metal-insulator transition. 30,31 Accurate determination of the detailed adatom-state band structure as well as theoretical estimations of the conductance by much sophisticated model such as the Kubo formula is highly requested.…”

Correction of systematic errors in scanning tunneling spectra on semiconductor surfaces: The energy gap of Si(111)-7×7at 0.3 K

“…Another possible reasons for a gap at E F are structural distortions such as charge density waves. There is no theoretical estimation of the Mott-Hubbard or charge density wave gap for the Si(111)-(7x7) surface, and the experimental data from high-resolution photoemission [7], high-resolution electron-energy-loss spectroscopy [8], nuclear magnetic resonance [9] and transport measurements [10,11] hint to a gap between 0 eV and a few tenths of an eV at low temperatures. These values are well below the minimum gap reported from the published STS data.…”