The clear dependence of the Schottky barrier height (SBH) on the structure of a metalsemiconductor (MS) interface is described for a few high-quality epitaxial MS systems. Polycrystalline MS interfaces, for which such a dependence is often absent, are shown to display clear evidence for SBH inhomogeneity, in direct conflict with the Fermi level (FL) pinning concept they helped to create. Recent theoretical calculations show that many basic tenets of the interface state pinning models are unfounded. The redistribution of charge as a result of bonding at the MS interface seems to be the most important contribution to the interface dipole and, hence, the SBH.

The growth and chemical intermixing of submonolayer and a few monolayer thick Fe films on a Au(001) surface was studied by High Resolution Low Energy Electron Diffraction (HRLEED) technique. Through the analysis of the energy dependent angular profiles as a function of time, we obtained the distribution of islands and distribution of spacings during submonolayer growth. The interference of electron waves from different chemical elements in terraces at different heights in the surface contributes to the background intensity and broadening in the angular profiles of diffraction beams. A subsurface Fe capped by Au islands as a result of atomic place exchange was observed at the initial stage of monolayer growth. From the energy dependent angular profiles as a function of temperature, we determine the quantitative change of inhomogeneity length (∼20 Å) at the interface of ultrathin films at elevated temperatures due to intermixing.

We have observed that a build-up of gas-phase adsorbates on Si(001) followed by mild annealing at temperatures in the range 500 - 800°C can lead to a quasi-periodic array of line defects, which in STM images appear similar to lines of dimer vacancies aligned perpendicular to the surface dimer rows. These line defects can be eliminated by annealing at temperatures above 900°C, indicating that Ni contamination is not the cause. Adsorbed O2 or H2O followed by an anneal do not exhibit the same behavior.

Diffusion of impurity atoms or point defect species occurs simultaneously with segregation in materials that are inhomogeneous or consisting of heterostructures. Based on thermodynamic principles, a general diffusion-segregation equation (DSE) simultaneously describing the species’ diffusion and segregation behavior has been derived. The DSE treats the diffusion-segregation problems on the same fundamental level as the diffusion problems are treated by Fick’s Laws. A critical comparison of modeling diffusion-segregation phenomena using the DSE vs. previous formulations is given. The wide range of applications of the DSE is demonstrated by several examples.

The electrical, magnetic and mechanical properties of bulk nanocrystalline nickel produced by pulse-electrodeposition are reviewed and discussed in the context of nanoprocessing as a distinct form of grain boundary engineering to develop soft magnetic materials with improved performance characteristics.

The conductivity behavior found in polymerization-filled composites differs from that predicted by percolation theory and can be explained by a simple model predicting the existence of two characteristic values in such systems: the radius of filler particles and the distance of charge penetration into polymer matrix. Detailed mechanisms, responsible for such peculiarities in charge transfer are also discussed.

Ellipsometric measurements of surfaces of oxidized silicon give information on the optical properties, structure and composition of the interface between the silicon and oxide. From such measurements in ambient liquids with different refractive indices, some close to that of the oxide, we conclude that there is an interfacial layer about one nm thick at all oxide thicknesses. This layer is either a transition layer of partially oxidized silicon or a layer of silicon of higher absorption than bulk silicon. The oxide has the refractive index of vitreous silica at all thicknesses.

The initial stage of oxidation of 40wt-% NH4F treated Si(111) surface at 300 °C in dry oxygen with a pressure of 133 Pa and the subsequent oxidation at 600 and 800°C were studied. It was found from the analysis of Si 2p photoelectron spectra that non-uniform layer by layer oxidation proceeds at 300°C, while the layer by layer oxidation proceeds at 600 and 800°C. Furthermore, at these temperatures the interface becomes atomically flat with the progress of oxidation.

The growth of ultra-thin oxide films by the thermal oxidation of silicon has been studied by low and medium energy ion scattering spectroscopies (LEIS and MEIS) and X-ray photoelectron spectroscopy (XPS). To help elucidate the diffusional and mechanistic aspects of oxide growth we have used sequential isotope oxidation (18O2 followed by 16O2). LEIS demonstrates that both 18O and 16O atoms are on the silicon surface under our growth conditions. MEIS also distinguishes 18O from 16O and gives a depth distribution for both with high accuracy. Our results show that several key aspects of the Deal-Grove model (oxygen diffusion to the Si-SiO2 interface and oxide formation at the interface) are consistent with our results for 50Å films. For very thin oxide films (15Å or less), we found a mixed isotopic distribution in the film, demonstrating more complex oxidation behavior.

X-ray reflectivity has been used to study properties of thin SiO2 films on (100) Si prepared by plasma-enhanced chemical vapor deposition (PECVD) or by anodic oxidation. Reflectivity curves were analyzed to obtain information concerning the thickness, density (p) and interface roughness (σ) of these oxide films. We found that the PECVD films have typically a density of 0.98 psi with a sharp Si/SiO2 interface (σ < 0.2 nm) and a surface roughness of -0.5 nm. Anodic oxide films have a lower density (p = 0.88 psi) and larger interface and surface roughness σ - 0.5 nm. High temperature annealing of the anodic oxides up to 900 °C resulted in no appreciable densification but caused an increase in the interfacial roughness. Considerable variations were observed in the reflectivity curves from a PECVD oxide film annealed in the 650-1050 °C range.-Modeling of these results suggests the existence of a thin interfacial layer (-0.5 nm) of less density (∼0.9psi) at the original silicon/PECVD oxide interface.

This paper describes the preparation of silicon-based metal oxide semiconductor, MOS, devices, capacitors and field effect transistors, FETs, using deposited oxide dielectrics. A critical aspect of the device fabrication process is the way the Si-SiO2 interface is formed; e.g., either before, during, or after the oxide deposition. We have studied different methods of fabricating Si-SiO2 heterostructures, and have concluded that the implementation of independently controllable and sequential process steps for (i) interface formation, and (ii) oxide deposition consistently yields MOS devices with electrical properties that are superior to those of devices fabricated under other processing conditions which include specifically interface formation during the oxide deposition.

The adsorption reaction of O2 molecule with symmetric dimers on the Si(001)–2×1 reconstructed surface has been investigated by ab initio molecular orbital calculations. Detailed analysis of the lowest energy reaction path has revealed that there exists a metastable state in which O2 molecule adsorbs on silicon dimer without dissociation, the dissociation of O2 molecule requires large activation energy, and a silicon oxide and an isolated oxygen atom are produced after the reaction has been completed. The activation energy required for the conversion from the metastable state to the final products has been estimated to be 60.4 kcal/mol. This result suggests that a symmetric dimer on the Si(001)–2×1 surface is hardly oxidized at room temperature. This conclusion is consistent with the recent STM observations that the initial stage of oxidation starts from the dimer defect sites on the Si(001) surface. On the contrary, it has been found that no activation energy is required for the oxidation reaction by O atom.

Chemical/native oxides grown on Si(100) after several standard wet cleans are characterized by Angle-resolved X-ray Photoelectron Spectroscopy (ARXPS), and Auger Electron Spectroscopy using sputter depth profiles. Target Factor Analysis (TFA) was used to separate the Si LVV Auger peak into three components identified by their lineshapes and positions as Si, SiO2, and SiOx- Auger depth profiles were used to quantify the thickness of the oxides, the depth distribution, and amount of SiOx in the interface region. ARXPS was used to study the chemical state distribution in the native oxides as a function of depth. The depth distribution function from the Auger data was converted to an angle-resolved format for direct comparison to the angle-resolved XPS data. With this comparison, the SiOx lineshape is correlated to a 3:1 mixture of Si 3+ and Si 2+ oxidation states.

We have developed a novel solid-state method for forming interfacial silicon dioxide by a two step process of rf sputtering tin-doped indium oxide (ITO), a transparent conductor, onto Si substrates followed by annealing at 800°C in flowing ultra-high purity nitrogen. TEM, XRD, FTIR, and C-V measurements were used to characterize the microstructure and morphology of as-deposited and annealed samples. The silicon dioxide layer grows interfacially between Si and In2O3 and, for a 5 nm thick oxide, is planar on a 2–4 monolayer scale over hundreds of nm. The kinetics of the solid-state interfacial oxidation reaction were studied and a model has been developed.

Ballistic Electron Emission Microscopy (BEEM) is shown to be a versatile spectroscopic tool to investigate scattering phenomena of energetic electrons. Two examples are given for obtaining numerical values of scattering parameters. Thus the elastic and inelastic mean free paths are deduced from model fits to attenuation data for thin Pd films deposited on Si(111) and Si(100) substrates. In the second example, the quantum yield for electron-hole pair production through impact ionization of hot carriers injected into Si(111) is directly measured over an energy range from 1-7 eV.

PtSi/Si interfaces have been formed by depositing Pt layers on chemically cleaned, lightly doped, n-type Si (100) wafers in a UHV magnetron sputter-deposition system using ultra high purity Ar as the sputter gas, followed by ex-situ silicidation in N2 ambient utilizing a 3-step rapid thermal annealing (RTA) process. The polycrystalline PtSi layer, with oriented grains ranging in size from 50-100 nm, exhibits a columnar growth morphology. The PtSi/Si interface is planar with interface roughness in the order of 5 nm peak-to-peak. Auger depth profile shows uniform composition through the PtSi layer and a clean and chemically abrupt PtSi/Si interface.

Epitaxial p-InP/Au Schottky diodes were fabricated by evaporation of Au onto Zn doped epitaxial layers of InP grown by MOVPE, on a highly doped InP substrate. The reverse current-voltage (Ir-Vr) and 1 MHz capacitance-voltage (C-V) characteristics of the Au/p-InP diodes were measured in the temperature range 220-393 K. At all temperatures, soft reverse current-voltage characteristics were observed, which may be due to the decrease in the effective Schottky barrier height (øbr) with the increase of Vr. The voltage dependence of the reverse current was well described in terms of the interface layer thermionic emission (ITE) model which incorporates the effects of applied reverse voltage drop and the transmission coefficient across the interface layer and image force lowering of the barrier height into the thermionic emission theory. A self consistent iteration and least square fitting technique was used to obtain the zero bias barrier height (øbo) and interface layer capacitance (Ci) from the Ir-Vr data. Both, the Ir-Vr and the C-V data were analyzed under the assumption of reverse bias voltage independence of the charge trapped in the interface states, which was supported by our experimental data. The values of øbo obtained from the C-V measurements agreed well with those obtained from the Ir-Vr data for a value of 0.45 AK−2cm−2 for the effective Richardson constant (Aeff).

Au and Pd Schottky contacts to n-InP produce extremely high barrier heights and low leakage currents when deposition is on a substrate cooled to 77K. Extensive chemical and structural analyses indicate that this process causes the metal film to be continuous at 50Å, much better than in standard processing. Stoichiometry of InP near the surface is better maintained with this process. A thin P:O compound may exist at the interface which also contributes to a high barrier height.

Thermally stable Al/n-GaAs Schottky contacts, up to annealing temperature at 500 °C for 20 seconds, have been realized by sputter deposition from an Al target to (100) n-GaAs at a base pressure ∼2×10−7 Torr. The Schottky barrier height was 0.75 eV (0.9 eV) when using the I-V (C-V) method with an ideality factor of 1.09 for the as-deposited samples. The Schottky barrier height increased to 0.97 eV (1.06 eV) with an ideality factor of 1.07 after annealing at 400 °C for 20 seconds. This barrier height, 0.97 eV, is the highest value reported for Al/n-GaAs diodes. The interfacial stability between Al and GaAs has been examined by cross section transmission electron microscopy. A (200) dark field cross section transmission electron microscopy image of the contact after annealing at 600 °C showed that the (Ga,Al)As phase formed at the interface and the enhancement of the Schottky barrier height was due to the formation of this phase.

This paper briefly reviews the standard Transmission Line Model (TLM) commonly used to measure the specific contact resistance of a planar ohmic contact. It is proposed that in the case of a typical Au-Ge-Ni alloyed ohmic contact, a more realistic model would need to take into account the presence of the alloyed layer at the metal-semiconductor interface. An alternative is described which is based on three contact layers and the two interfaces between them, thus forming a Tri-Layer Transmission Line Model (TLTLM). Expressions are given for the contact resistance Rc and the contact end resistance Re of this structure, together with a current division factor, f. Values for the parameters of this model are inferred from experimentally reported values of Rc and Re for two types of contact.