Induced Gap States Research Articles

Thick sodium overlayers on GaAs(110).

We report density-functional theory calculations of the electronic structure, total energy, and forces for the Na adsorption on GaAs(110) using the local-density approximation of the exchange-correlation functional and ab initio pseudopotentials. Results are presented for coverages ranging from one adatom per substrate surface cell up to the thick overlayer limit. The atomic and electronic structure of the substrate is locally changed by the sodium adsorption on GaAs(110), depending on the coverage. In particular, we analyze the wave-function character of the states at the Fermi level, how it changes with sodium coverage, and we identify the formation of metal induced gap states (MIGS) at the interface. These MIGS are found to have mostly Ga dangling-bond character for all coverages. The calculated values of the p-type Schottky barrier and of the variation of photothreshold as a function of coverage are in good agreement with experimental data.

Theoretical approach to interfacial metal-oxide bonding

This paper reviews existing theoretical calculations of the electronic characteristics and of the adhesion energy at a metal-oxide interface, in the context of non reactive adhesion processes and two- or three-dimensional metal growth processes. Emphasis is put on the competition between metal-cation and metal-oxygen interfacial interactions, on the resulting charge transfers and on the Ferrni level position. Aside from the well-known image force and van der Waals contributions to the adhesion energy, the importance of kinetic and electrostatic terms associated with the Metal Induced Gap States and the interfacial dipole is stressed.

Interfacial atomic composition and Schottky barrier heights at the Al/GaAs(001) interface

The influence of different interfacial chemical compositions on the Schottky barrier heights across the Al/GaAs(001) interface is studied using first-principles local density functional calculations. The barrier heights are calculated for seven different interfacial chemical compositions, including the chemically abrupt As-terminated and Ga-terminated interfaces, and also several other interfaces related by Al↔Ga place exchange or containing As antisites. We find p-type barrier heights φp that vary by 0.4 eV, demonstrating a significant influence of the interface composition on the resulting barrier heights. The barrier height variation is explained by the different chemical bonding at the interface in the various cases. The metal induced gap states (MIGS) of two structures with different barrier heights are compared in order to demonstrate why such states do not result in barrier heights independent of interfacial chemical composition. It is thus suggested that the reason an experimental value of φp=0.65 eV is generally found for Al/GaAs is not due to several possible interfacial structures all having their Fermi level pinned by MIGS at the same value, but rather due to the thermodynamic preference for certain structures over others with significantly different barriers. This proposal offers a potential explanation for recent photoemission experiments that find the Fermi level position in the gap varying by several tenths of a volt as a function of the initial surface structure of the GaAs substrate: some of the samples are likely to have interfacial compositions which are metastable with respect to structures that give the standard barrier heights.

Electrical in situ characterisation of metal/gallium phosphide (110) Schottky contacts

Schottky barrier formation for metals on p-GaP(110) has been systematically investigated by electrical transport measurements, in order to clarify the dependence of Schottky barrier height on metal and interface properties. Metals were deposited onto clean cleaved GaP(110) in ultrahigh vacuum and the temperature dependent current-voltage ( I–V) and capacitance-voltage ( C–V) characteristics were recorded in situ. The Schottky barriers (Ф B) were determined for nine metals (Na, Mg, In, Ti, Sn, Sb, Ag, Au and Pt), covering a wide range of metal work function (Ф M) from 2.75 to 5.65 eV. The results reveal a weak dependence of Ф B on Ф M with S=dФ B/dФ M ≈ 0.14 , which is consistent with a strong Fermi level pinning by induced gap states.

Theory of the formation of Schottky barriers

The case of thick metallic layers on semiconductors is first treated to understand the origin of an eventual pinning of the Fermi level. The notion of metal induced gap states is precised and related to resonant dangling bond states. After having summarized simple rules governing the relaxation and reconstruction on free surfaces the formation of the interface is investigated for alkali and noble metals on GaAs(110). Calculations in the submonolayer range are discussed. The influence of correlation effects is examined and it is shown that a negative U situation is most likely to occur. The results of low temperature formation are also interpreted in some detail as well as the onset of metallization.

Energy distribution of interface state density in AlN/ n-GaAs heterostructures formed by laser chemical vapor deposition

AlN/ n-GaAs heterointerface formed by the Laser Chemical Vapor Deposition techniques have been studied by the quasistatic-low-frequency and Terman's-high-frequency methods. The results compare favourably with those obtained from high-low frequency, conductance and DLTS methods. All of these methods give essentially the same results with interface trap densities in the range of 10 12–10 13 eV −1 cm −2 with a “U”-shaped energy distribution having a minimum at about E c − 0.9 eV. The distribution is analyzed in terms of exponential density-of-state tails derived from conduction and valence band edges in the Disorder Induced Gap State (DIGS) model. The surface Fermi level is pinned at E c − 0.83 eV by the requirement that depletion and interface charges are equal and opposite, and the charge distribution profile makes this pinned value particularly stable. The surface potential V s can be moved at most 0.35 eV either side of the pinned level before breakdown occurs in the AlN dielectric film.

In-Situ Characterization of Compound Semiconductor Surfaces by Novel Photoluminescence Surface State Spectroscopy

The recently proposed novel photoluminescence surface state spectroscopy (PLS3) technique is applied for in-situ, non-destructive and contactless characterization of variously processed surfaces of GaAs, InP and InGaAs. Chemically etched, anodized and passivated surfaces, as well as the original as-received surface, give rise to U-shaped surface state density distributions with characteristic charge neutrality energy levels, EHO, which is consistent with the disorder induced gap state (DIGS) model. Annealing of as-received surfaces in hydrogen ambient leads to formation of discrete levels, possibly due to escape of As or P atoms. The effectiveness of a new UHV-based passivation scheme for InGaAs using an ultrathin MBE Si interface control layer (ICL) is also confirmed.

Study of the defects induced by low-energy (100 eV) hydrogen-ions on amorphous silicon dioxide

The low-energy (100 eV) hydrogen-ion bombardment effects on a-SiO2 have been investigated by using synchrotron radiation photoemission spectroscopies. The argon bombardment effects have also been studied, in order to discriminate between physical and chemical characters in the hydrogen/a-SiO2 interaction. Our results show that hydrogen treatment produces predominantly Si-H defects, which are observed to induce gap states in a-SiO2.

The Ba/Si(100)-2 × 1 interface: II. XPS, BIS and synchrotron PS studies of Schottky-barrier formation

In the preceding paper [Surf. Sci. 260 (1992) 97] we have reported the results of XPS and XAES studies of silicide formation at the Ba/Si(100) interface at elevated temperatures. In this paper we present the results of XPS, BIS, and synchroton PS studies on the same system but at room temperature. We find that Ba grows layer by layer on the Si substrate and that electron energy losses and final state effects are important for the correct interpretation of the data. Observed shifts in the Si 2p core levels are ascribed to a lowering of the Schottky barrier and are explained with the model presented by Mönch [Phys. Rev. Lett. 58 (1987) 1260], while measured shifts in the Ba 4d core levels are related to the reduction of the silicon work function. This model requires the presence of metal induced gap states (MIGS) which are indeed observed with BIS. Here we find increasing spectral intensity near the Fermi level upon Ba deposition. This intensity is assigned to Ba 5d states which probably fulfill the role of the MIGS in Mönch's model.

Schottky barriers and interface structure at silicide-silicon interfaces

Schottky barriers at metal-semiconductor interfaces have attracted much interest in recent years. One of the principal interests has centred on the mechanism for Fermi level pinning. The sililcide-silicon interface has been proposed as a system which is described by the metal induced gap states model. We have performed calculations on the NiSi 2/Si(111) type A and type B interfaces as well as the NiSi 2/Si(100) interface. In addition we have also studied the CoSi 2/Si interface. For the NiSi 2/Si(111) interface, we have further investigated the influence of point defects and hydrostatic pressure on the Schottky barrier height. Based on the results of our calculations we conclude that these interfaces do indeed subscribe to the MIGS model. We also present the results of some total energy calculations and discuss these with experimental observations.

Vibrations and Some Electronic Excitations of Adsorbed Alkali Metal Monolayers

We discuss briefly data obtained for adsorbed alkali metal monolayers with three different techniques (HREELS, UPS and SHG). The vibrational spectra provide support for the traditional picture of a fractionally charged adsorbate at low coverage and a neutral one at high coverage. We ascribe a vibrational overtone reported in previous work to contamination due to residual water vapour. UPS, uncharacteristically, gives little information of interest for low coverages except for substrates with a band gap and with a surface state in the gap. For Na on Si(100) we monitor the metallization of the surface region in the monolayer coverage range and observe metal induced gap states at higher coverages. For Cs on Cu(111) the optical second harmonic generation (SHG) shows a strong monolayer coverage dependence which we ascribe to optical transitions between surface bands and to surface barrier photoabsorption indicating the potential of SHG as a probe of the details of the surface electronic structure. .

Electronic structure of a metal-insulator interface

We present an analytical study of the electronic structure of a metal-insulator interface with special emphasis on the metal induced gap states (MIGS). It includes three steps: (i) a tight-binding approach of the dispersion relation and Green's function of insulators of NaCl or ZnS structure; (ii) a matching with free electron-like wavefunctions at the NaCl(100) or ZnS(110) surfaces, which yields the density and penetration depth of the MIGS as a function of the ionocovalent characteristics of the insulator and of the metal Fermi level; (iii) a self-consistent determination of the Fermi level position in a Thomas-Fermi approximation. The Schottky barrier height is derived under a simple analytic form and its dependence upon the metal work function is found in good agreement with experimental results.

LCAO‐Calculations of the Valence Band Structure of Modified Polar Semiconductor Surfaces: Hydrogen‐, Copper‐, and Ruthenium‐Covered Gallium Arsenide (001)

AbstractResults of local density first principles LCAO calculations for a 13 layer slab of free and modified As‐terminated (1 × 1) GaAs(001) surfaces are presented. The surfaces are covered by a monolayer of hydrogen, copper, or ruthenium. The surface calculation for the hydrogen covered surface gives hydrogen‐induced and GaAs surface states in agreement with angle‐resolved photoemission results within the energy range where they are found. For the systems GaAs(001):Cu and GaAs(001): Ru metal induced gap states are predicted with a density of 2 × 1015 cm−2. The calculated surface states indicate a higher Schottky barrier for the ruthenium than for the copper covered surface.

Hydrogen effects on a-SiO 2: A photoemission study

Negative bias stress experiments on a-Si:H/a-SiO 2 Thin Film Transistors show that the instability is predominantly related to charge injection in the gate insulator. We propose that charge injection is occurring in a defected region of the gate insulator adjacent to a-Si:H, caused by the effect of activated hydrogen on a-SiO 2 surface during deposition. To investigate the hydrogen influence, we have studied low-energy (100 eV) hydrogen-ion and Ar-ion bombardment effects on a-SiO 2 using synchrotron radiation photoemission spectroscopies. Comparing the two kind of treatments, we show that hydrogen treatment produces predominantly Si-H defects, which are observed to induce gap states in a-SiO 2 .

Electron excitation during anodic decomposition of III–V compounds induced by hole injecting species (Ce 4+)

Evidence is presented for electron excitation steps during the anodic decomposition of III–V compounds induced by hole injecting species such as Ce 4+. Their number depends on the nature of the compound. Thus, the behaviour of GaAs on the one hand and phosphorus containing compounds on the other are quite different. The nature of the decomposition intermediates connected with the existence of the electron excitation steps is related to the phosphorus induced gap states.

Investigation of interface states and fermi level pinning mechanisms with metals on InP(110) surfaces

Metal/InP(110) interfaces have been systematically studied using photoelectron Spectroscopy. This work aims at understanding the physical nature of interface states and Fermi level pinning mechanisms. Clear-cut experimental evidence shows that two types of interface states are responsible for Fermi level pinning. At the disrupted interfaces which occur at room temperature, a large amount of defects are generated during various interfacial processes and pin the surface Fermi levels around 0.95 eV above the valence band maximum (VBM). On the contrary, much fewer defects are present at the nearly perfect interfaces which are obtainable at reduced temperature. In these cases, the metal induced gap states play an overwhelming role, and the Fermi level is pinned around 0.75 eV above the VBM. It is also concluded that interplay between the two mechanisms should be taken into account in order to understand the process of Schottky barrier formation.

A photoemission study of the BiInP(110) interface

The room temperature adsorption of bismuth on clean cleaved p-type InP(110) surfaces has been studied with soft X-ray photoelectron spectroscopy in the coverage range 0.01-60 monolayers (ML): Line shape analysis of the overlayer (Bi 5d) and substrate (In 4d) core level emission reveals that the interface is abrupt. Bismuth is preferentially adsorbed at a single substrate site within the first layer. However, strong clustering takes place above this coverage. The evolution of the band bending indicates that different mechanisms may be important in the high and low coverage regimes. At low coverages the imperfections in the first bismuth layer cause high band bonding which is reduced as the first layer nears compleion. Above 20 ML the semi-metallisation of the overlayer leads to pinning in accordance with the metal induced gap state model.

A computer simulation of the recombination process at compound semiconductor surfaces and hetero-interfaces

The complex recombination process through quantum states at compound semiconductor surfaces and hetero-interfaces is analyzed in a unified manner on the computer, using the unified disorder induced gap state (DIGS) model. Recombination through uniformly distributed states at surfaces and hetero-interfaces, and that through U-shaped surface states at GaAs surfaces subjected to various surface treatments, are specifically analyzed. The result indicates that the effective surface recombination velocity is not constant, but is strongly dependent on the excitation intensity and the location of charge neutrality level, EHO. PL intensity enhancement after photochemical oxidation in water and sulfur treatments (Na2S, (NH4)2S) is shown to be not due to reduction of the durface states, but due to the generation of a fixed charge, whereas photochemical HCl treatment reduces the surface states significantly.

Antimony on indium phosphide: Electrical barriers, defects, and induced gap states

We describe an investigation of an unreactive interface, namely Sb on InP(110) clean cleaved surfaces. Transport techniques (I–V, C–V), Raman spectroscopy, and soft x-ray photoelectron spectroscopy have been used to study the Schottky barrier heights, overlayer structure, and interface interactions for Sb layers deposited on substrates held at room and low temperatures. The evolution of the overlayer structure at room temperature is complex, with an amorphous Sb layer growing on top of an ordered monolayer, until a thickness of ∼10 monolayers (ML), when the layer crystallizes. The evolution of the Schottky barrier is also complex but for thick layers we observe the lowest barriers on n-InP and the highest ever observed on p-InP. Possible mechanisms for the fact that the Fermi level appears pinned close to the conduction band are discussed.

Control of compound semiconductor–insulator interfaces by an ultrathin molecular-beam epitaxy Si layer

Based on the disorder induced gap state (DIGS) model, an attempt is made to control the insulator–semiconductor (I–S) interfaces of GaAs and In0.53Ga0.47As by an ultrathin surface-oxidized silicon (Si) interface control layer (ICL). A Si ICL is grown on GaAs or InGaAs by molecular beam epitaxy (MBE), and is partially oxidized. Then, a thick SiO2 or Si3N4 layer is deposited by in situ photo-CVD processes. An in situ XPS analysis confirms formation of the intended structures. In the GaAs structure, the state density in the midgap region is remarkably reduced. However, the interface Fermi level is blocked by a high density of interface states near the conduction band minimum, contrary to the recent reports of ‘‘complete unpinning.’’ These states are Si-derived intrinsic states, judging from the band line-up. On the other hand, no such states exist in the InGaAs structure and completely ‘‘unpinned’’ behavior with a very small hysteresis is realized after annealing. The result is interpreted in terms of successful pseudomorphic matching of Si ICL to InGaAs combined with excellent I–S matching between Si ICL and SiO2.