We have measured optical emission from interface states formed by metal deposition on UHV-cleaved InP(110) and GaAs(110) surfaces by means of cathodoluminescence spectroscopy. Our study reveals discrete levels distributed over a wide range of energies and localized at the microscopic interface. Our results demonstrate the influence of the metal, the semiconductor, and its surface morphology on the energy distributions. The detailed evolution of optical emission energies and intensities with multilayer metal deposition exhibits a strong correlation between the deep gap levels, the Fermi level movements, and Schottky barrier heights. The results demonstrate that in general electronic states deep within the band gap continue to evolve beyond monolayer coverage into the metallic regime.
We report the first study of optical-emission properties associated with interface-state formation for metals on III-V semiconductor surfaces. Cathodoluminescence spectroscopy reveals discrete levels distributed over a wide energy range and localized at the microscopic interface. Our results demonstrate the influence of the metal, the semiconductor, and its surface morphology on the energy distributions. Evolution of spectral features with interface formation, particularly above monolayer metal coverage, is correlated with Fermi-level movements and Schottky-barrier heights.
We report temperature-dependent cathodoluminescence spectroscopy (CLS) and internal photoemission spectroscopy (IPS) studies of metal/molecular-beam epitaxy GaAs(100) interfaces. For Au, Cu, and Al deposited at 90 K on clean, ordered GaAs(100) surfaces under ultrahigh vacuum conditions, we observe low intensity, metal-induced optical transitions centered at 0.85 eV. Upon room temperature (RT) annealing, this emission feature increases in intensity, corresponding to an increased density of midgap states. For Al, an additional excitation-energy-dependent emission is evident at 1.2 eV due to states even more localized near the interface. RT IPS measurements for metals deposited on GaAs at 90 K give well-defined Schottky barriers. We obtain stable and reproducible Schottky barrier heights of 1.05 and 0.93 eV for Au/GaAs and Cu/GaAs, respectively. For Al/GaAs, we obtain a Schottky barrier height as low as 0.3 eV with significant, time-dependent variations up to 0.41 eV. After 400 °C annealing, the Schottky barrier height increases to 0.57 eV. These results demonstrate that the barrier height and the formation of metal-induced states are temperature dependent. The correlation between CLS and IPS results indicates that the barrier height depends on the presence of metal-induced interface states and their metal-specific interface reactions. The values of the Schottky barrier height determined here are in close agreement with those obtained from our earlier soft x-ray photoemission studies. Our CLS and IPS results emphasize the detailed role of interface states coupled with chemical interactions for Schottky barrier formation at metal/GaAs interfaces.
Soft x-ray photoemission spectroscopy measurements of clean, ordered InxGa1−xAs (100) surfaces with Au, In, Ge, or Al overlayers reveal an unpinned Fermi level across the entire In alloy series. The Fermi level stabilization energies depend strongly on the particular metal and differ dramatically from those of air-exposed interfaces. This wide range of Schottky barrier height for III-V compounds is best accounted for by a chemically induced modification in metal-alloy composition.
We report results of x-ray photoemission and cathodoluminescence spectroscopies studies of interface formation at metal–GaAs junctions. The results are interpreted by using a microscopic model of metal–semiconductor interfaces. Our low-temperature measurements and analyses show the validity of Schottky’s phenomenological description, thereby suggesting that metal-induced gap states and native defect mechanisms are not major factors in determining the Fermi-level energy at the low-temperature formed interface. Our room-temperature results show that a broad range of Fermi-level stabilization and the formation of two reaction-induced interface states are obtained upon metallization of GaAs(100) surfaces. These results strongly imply that the insensitivity of rectifying barrier height on metal work function results from metallization-induced atomic relaxations at the interface.
We report soft x-ray photoemission studies of metal/molecular-beam epitaxy (MBE)-GaAs(100) interfaces formed at low temperature. Our results indicate that rectifying barrier heights are proportional to the metal work function in accordance with Schottky’s original description of metal–semiconductor contacts. These results confirm the predictions of a self-consistent model of metal–semiconductor interfaces, and suggest that metal-induced gap states and native defect mechanisms are not major factors in determining the Fermi level energy at ‘‘ideal’’ interfaces. We attribute deviations from the ideal Schottky limit behavior observed for interfaces formed at room temperature to metallization-induced atomic relaxations (rather than electronic relaxations) occurring at metal–semiconductor contacts. We present a useful methodology for analyzing electronic properties at metal–semiconductor interfaces. The pronounced differences in barrier height formation between MBE vs melt-grown GaAs can evidence the role of deep states in controlling Schottky barriers at metal/melt-grown GaAs.
We have performed a photoemission study of the Schottky barrier obtained by depositing Ag onto UHV cleaved GaP(110) at room temperature. The barrier appears to be fully developed at a coverage of ∼2 monolayers and its height is 1.1 eV. This figure is to be compared with higher (lower) values obtained with metals having a larger (smaller) work function than Ag, for example Cu and Au (In and Al) deposited onto the same surface. The result corroborates the idea that the metal–GaP(110) interface is a good example of the Schottky limit in metal–semiconductor junctions. The effective work function model proposed by Woodall and Freeouf provides the best interpretation of the Schottky barrier height in this interface.
