Unified Defect Model Research Articles

A unified mid-gap defect model for amorphous GeTe phase change material

Understanding the nature of the mid-gap defects in amorphous chalcogenide alloy-based phase change materials is crucial to guide the development of multi-level memories, selectors, and neuromorphic devices. Currently, however, the knowledge of the mid-gap defects is fragmented and the known structural and electrical characteristics of the defects cannot be unified in a single atomic model. In this work, a Ge–Ge chain defect model is built by artificial bond rearrangement in an amorphous GeTe network generated by atomic distortion, revealing a possible valence alternation mechanism in its formation. Atoms in the Ge–Ge chain are found to have a crystalline-like environment. The chain is found to have a negative-U property and to introduce mid-gap states. This model unifies the main known properties of the mid-gap defects in amorphous chalcogenide alloy-based phase change materials.

Surface barrier analysis of semi-insulating and n+-type GaAs(0 0 1) following passivation with n-alkanethiol SAMs

The surface Fermi level of semi-insulating and n+-type GaAs(001) was determined before and after passivation with n-alkanethiol self-assembled monolayers (SAMs) by X-ray photoelectron spectroscopy. Fermi level positioning was achieved using Au calibration pads integrated directly onto the GaAs surface, prior to SAM deposition, in order to provide a surface equipotential binding energy reference. Fermi level pinning within 50meV and surface barrier characteristics according to the Advanced Unified Defect Model were observed. Our results demonstrate the effectiveness of the Au integration technique for the determination of band-edge referenced Fermi level positions and are relevant to an understanding of emerging technologies based on the molecular–semiconductor junction.

Capacitance-voltage characterization of GaAs–Al2O3 interfaces

The authors apply the conductance method at 25 and 150°C to GaAs–Al2O3 metal-oxide-semiconductor devices in order to derive the interface state distribution (Dit) as a function of energy in the bandgap. The Dit is governed by two large interface state peaks at midgap energies, in agreement with the unified defect model. S-passivation and forming gas annealing reduce the Dit in large parts of the bandgap, mainly close to the valence band, reducing noticeably the room temperature frequency dispersion. However the midgap interface state peaks are not affected by these treatments, such that Fermi level pinning at midgap energies remains.

GaN Cleaning by Ga Deposition, Reduction and Re-Evaporation: An SXPS Study

The Ga deposition, reduction, and re-evaporation technique commonly used to produce clean n-GaN surfaces and Ag–GaN interface formation on the resultant surface, have been investigated by Soft X-ray Photoelectron Spectroscopy (SXPS) and current–voltage measurements. SXPS studies have indicated that Ga deposition produces a band-bending of ΔEk = + 1.0 eV to higher kinetic energy. Our results show this shift to be a partially reversible process: re-evaporation of the deposited Ga resulted in a Fermi shift of ΔEk = — 0.6 eV to lower energy. Ag deposition did not cause any further Fermi shift, indicating that the Fermi level is pinned (2.2 ± 0.2) eV above the valence band edge, possibly as a consequence of the cleaning procedure itself. Current voltage (I–V) measurements have shown a barrier height of 0.77 eV and an ideality factor of 1.6. Metal induced gap states and the unified defect model are discussed as possible barrier formation mechanisms.

Evolution mechanism of nearly pinning-free platinum/n-type indium phosphide interface with a high Schottky barrier height by in situ electrochemical process

Recently, unusually high Schottky barrier heights (SBHs) have been realized by our group on n-type InP and related materials by an in situ electrochemical deposition of Pt. In an attempt to understand the underlying mechanism of the SBH enhancement, this article investigates in detail the evolution process of the metal (Pt, Ni, Co, and Ag)-InP interface during the in situ electrochemical process, using current–voltage, capacitance voltage, deep level transient spectroscopy, x-ray photoelectron spectroscopy, Raman, atomic force microscope, and scanning electron microscope measurements. Pt deposition by the electrochemical process realized an oxide-free, defect-free, stress-free, and nearly pinning-free interface, whereas Pt deposition by conventional electron beam evaporation and sputtering processes as well as Ag, Ni, and Co deposition by the electrochemical process gave rise to stressed and pinned interfaces. The observed large process dependence of SBH can be explained by none of the metal induced gap state model, the unified defect model, and the effective workfunction model. It is explained here by the disorder induced gap state model.

DLTS study of GaAs MIS structures with plasma deposited insulator

The results of a DLTS study of GaAs MIS structures with plasma deposited Si3N4 as insulator are given. It is established that the activation energy and the capture cross section of traps decrease with reverse bias and at large biases correspond to the respective parameters of bulk traps. A correlation is found between the DLTS spectra and the maximum HF capacitance of the MIS structure in the “accumulation” region. Experimental data are explained in the framework of a “unified defect model” by the change of the potential relief in a semiconductor as a result of a high-resistivity layer formed near the interface. The interpretation of the DLTS spectra of the GaAs MIS structures is discussed. [Russian Text Ignored].

Surface photovoltage studies of n-type and p-type InP

Surface photovoltage spectroscopy (SPV) was used to study the initial stages of oxidation of single crystal InP(110) in an attempt to understand the nature and origin of the surface states that develop. Distinct surface states were seen to develop on n-type as the surface was exposed to oxygen. A surface state, associated with cleavage damage, was also observed on p-type. A detailed fit to the experimental data was made by using a model of the dependence of surface charge on photon energy. This was used to unfold the position and intensity of the states. States trailing into the band gap from the bulk bands were seen on both n- and p-types. The analysis also indicated that pairs of isolated states, a donor and an acceptor state, were produced. On p-type, these were present on the clean, cleaved surface while they developed with oxygen exposure on n-type. These states are consistent with the point defect states proposed by the unified defect model. The time response of the SPV signal was also recorded for these surfaces. They were analysed by careful fitting to a model describing the charging and discharging characteristics. This revealed that the midgap state on n-type had a fast and a slow component.

Metal-semiconductor contacts: electronic properties

Rectification in metal-semiconductor contacts was first described by Braun in 1874. We owe the explanation of this observation to Schottky. He demonstrated that depletion layers exist on the semiconductor side of such interfaces. The current transport across such contacts is determined by their barrier heights, i.e., the respective energy difference between the Fermi level and the edge of the majority-carrier band. Since Schottky had published his pioneering work in 1938 the mechanisms, which determine the barrier heights of metal-semiconductor contacts, have remained under discussion. In 1947, Bardeen attributed the failure of the early Schottky-Mott rule to the neglect of electronic interface states. The foundations for a microscopic description of interface states in ideal Schottky contacts was laid by Heine in 1965. He demonstrated that a continuum of metal-induced gap states (MIGS), as they were called later, derives from the virtual gap states of the complex semiconductor band-structure. Neither this MIGS model nor any of the many other monocausal approaches, the most prominent is Spicer's Unified Defect Model, can explain the experimental data. In 1987, Mönch concluded that the continuum of MIG states represents the primary mechanism, which determines the barrier heights in ideal, i.e., intimate, abrupt, and homogeneous metal-semiconductor contacts. He attributed deviations from what is predicted by the MIGS model to other and then secondary mechanisms. In this respect, interface defects, structure-related interface dipoles. interface strain, interface compound formation, and interface intermixing, to name a few examples, were considered.

Reaching consensus and closure on key questions, a history of success, and failure of GaAs surfaces and interfaces at the Proceedings of the Physics and Chemistry of Semiconductor Interfaces

It is difficult to make scientific progress without consensus and closure. Periodic and focused meetings such as the Proceedings of the Physics and Chemistry of Semiconductor Interfaces (PCSI) provide perhaps an ideal format for this. Two examples are given and examined. The first involves intrinsic pinning of the Fermi level at the surface of GaAs and other III–V semiconductors. This is explored as an example of successful consensus and closure at PCSI. The mechanism for extrinsic pinning of the Fermi level and Schottky barrier formation is given as a second and unsuccessful example. Rather than surveying all models and experiments related to this, this article focuses on the unified defect model and the advanced defect model in order to give specific examples of this failure and suggest approaches which might avoid these problems in the future. Because it is difficult to define ‘‘consensus and closure,’’ considerable space is allocated to the definition and meaning of these concepts.

Time dependence of the surface Fermi level of GaAs in atmosphere

This letter reports the time dependence of the surface Fermi level of GaAs grown by molecular-beam epitaxy and then exposed to atmosphere. The sheet resistance of sample structures for field effect transistors alternately increased, decreased, increased, and decreased to become nearly constant after about 500 h. These changes correspond to the surface Fermi level varying between 0.3 and 0.7 eV and finally settling 0.7 eV above the valence band maximum. Comparison between annealed and unannealed samples with low-temperature-grown GaAs layers showed that the pinning of the surface Fermi level at 0.7 eV above the valence band maximum is caused by arsenic antisite defects. The result supports the advanced unified defect model.

The effect of external excitation on the sulphur passivation of GaAs surfaces

A simple chemical treatment using alkaline sulphides as an effective and robust means of surface passivation of GaAs has been widely studied. A new method of external excitation in the form of light and heat during sulphorization is investigated in the present study. This method of passivation is found to be more effective. The stoichiometry and the electronic properties of the passivated GaAs surfaces have been characterized by X-ray photoelectron spectroscopy, photoluminescence and I-V measurements on Schottky contacts. The recently proposed advanced unified defect model is found to explain the apparently contradictory results of decrease in surface state density accompanied by an increase in the surface band bending.

Hall-effect measurement and a defect model analysis of surface-potential changes in GaAs

Recently, increases of the surface potential in n-type epitaxial GaAs after NH4OH treatment studied using Hall-effect measurements, were reported. In this work, complementary results for p-type samples show decreases in band bending, again in agreement with the predictions of the advanced unified defect (AUD) model for the surface-interface states. From this and other data, a working mathematical model based on the AUD model is developed. Probable values for its parameters are determined which suggest a physically reasonable mechanism to explain the surface-potential changes.

Surface Fermi-level changes in n-type GaAs determined from Hall-effect measurements

Changes in the surface Fermi-level position in n-type epitaxial GaAs samples are determined by Hall-effect measurements of the corresponding changes in the sheet concentrations and theoretical calculations of the surface depletion thickness. The changes are induced and reversed repeatedly by alternating wet chemical treatments in hydrogen peroxide and ammonium hydroxide. This is the first known use of hydrogen peroxide to restore the surface Fermi level to near its starting value and demonstrate the repeated variation of the surface Fermi level by ammonium hydroxide. The results agree with the predictions of the advanced unified defect model and with published reports of increased band bending on n-type material, rather than with conflicting reports of decreased band bending. The results also indicate that problems may exist with other techniques used to measure surface- potential changes.

Characterisation of CoSi2/- and TiSi2/n-GaAs Schottky barriers

The heights of CoSi2/- and TiSi2/n-GaAs Schottky barriers are determined, by various methods, to be 0.77 and 0.64 eV respectively. These values are discussed in view of the recently developed Schottky barrier theories such as the advanced unified defect model and the effective workfunction model. The thermal stability of these contacts is investigated and it is found that the CoSi2/n-GaAs barrier height remains reasonably stable up to 600 degrees C for 10 min anneals and up to 750 degrees C for 1 min anneals. Under oxygen-free conditions and/or for shorter anneal times, this limit value may probably be increased. Consequently, CoSi2 may be an interesting contact material for high-temperature applications such as self-aligned GaAs MESFET fabrication.

From synchrotron radiation to I-V measurements of GaAs schottky barrier formation

Through the use of synchrotron radiation photoemission spectroscopy (PES) and related techniques, we have gained detailed knowledge of Fermi level pinning, interfacial chemistry and disruption of GaAs for coverages up to several monolayers (ML). A link has been made between these data and that in the thick layer regime (hundreds of ML), which characterizes practical Schottky diodes. PES results for thin layers deposited at room temperature (RT) and low temperatures of about 80 K (LT) as well as thick films deposited at RT and annealed to higher temperatures are considered. At LT where GaAs disruption is minimized for thin films, metal-induced gap states seem to dominate the Fermi level pinning process except where GaAs metal reactions are strong. For RT thin and thick films, the effects of defects must be considered, and the advanced unified defect model (AUDM) is applied. In the AUDM the key defects are identified as the AsGa (double donor with levels at 0.75 and 0.5 eV above the valence band maximum) and the GaAs antisite (double acceptor) with the AsGa normally dominating due to the excess As which characterizes LEC GaAs crystals. The literature is reviewed and a number of phenomena are explained in terms of this model including the Fermi level position on MBE grown GaAs observed by Svensson et al. and the anomolously high Schottky barrier height (SBH) of thick Ga on n-GaAs observed by several groups. By performing electrical, TEM, and chemical studies of thick diodes and by evaluating the changes upon thermal annealing of diodes it is found that the AUDM successfully predicts the increase or decrease of barrier height on annealing.

Carbon and zinc delta doping for Schottky barrier enhancement on n-type GaAs

The growth of thin (50–100 Å), C or Zn δ-doped layers on n-type GaAs is shown to yield large enhancements in the effective Schottky barrier height (ΦB) of TiPtAu contacts subsequently deposited on the material. The incorporation of a single C δ-doped layer (p=1.5×1020 cm−3, 50 Å wide) within 100 Å of the surface leads to a barrier height of 0.93 eV, a significant increase over the value for a control sample (0.76 eV). The use of two sequential δ-doped layers leads to an apparent barrier height in excess of the GaAs band gap (ΦB=1.67 eV). This appears to be consistent with the predictions of a unified defect model. Zinc δ doping (p∼3×1018 cm−3) in a similar fashion produces barrier heights of 0.81 eV for one spike and 0.95 eV for two spikes.

Molecular-beam epitaxial growth of NiAl on GaAs(001)

High-quality, single-crystal overlayers of NiAl that create a Schottky barrier height of 0.90 eV have been grown on semi-insulating GaAs(001) by molecular-beam epitaxy and characterized by high-energy resolution x-ray photoelectron spectroscopy, low-energy electron diffraction, and high-energy x-ray photoelectron diffraction. Thin Al interlayers were grown between the NiAl and the substrate in order to test the possibility that an ultrathin AlxGa1−xAs layer of higher band gap than the substrate forms at the interface and is the cause of the high barrier height. It has been found that such a layer does indeed form when the Al prelayer is deposited near room temperature and annealed at 500 °C, or if the deposition is carried out at a substrate temperature of 500 °C. Moreover, the presence of this prelayer correlates with the appearance of a high barrier height at the interface. Virtually identical results are obtained when layers of Ni and Al are deposited sequentially at room temperature and then annealed at 500 °C to alloy and recrystallize the metal. The realization of a 0.9-eV barrier height occurs independently of the removal of any excess As from the interface, which according to the advanced unified defect model, is thought to cause the high barrier height by virtue of a change in the balance of antisite defects.

Photoemission study of the band bending and chemistry of sodium sulfide on GaAs(100)

Recently, there has been a great deal of interest in Na2S⋅9H2O as a passivating chemical treatment for GaAs surfaces. It has been shown that it reduces the high surface recombination velocity characteristic of GaAs surfaces, and may offer hope for ‘‘unpinning’’ the surface Fermi level. We have used photoemission spectroscopy to study the band bending and chemistry of these overlayers on n-type GaAs(100). Identically prepared samples show the characteristic increase in photoluminescence signal, and have also been characterized using surface conductivity measurements. We find using photoemission that the surface Fermi level of the treated wafer is still near midgap. We also observe the chemistry at the interface, and offer a possible explanation of the photoluminescence and surface conductivity data in terms of it and the advanced unified defect model.

Photoemission study of the band bending and chemistry of sodium sulfide on GaAs (100)

Recently, there has been a great deal of interest in Na2S⋅9H2O as a passivating chemical treatment for GaAs surfaces. It has been shown that it reduces the high surface recombination velocity characteristic of GaAs surfaces, and may offer hope for ‘‘unpinning’’ the surface Fermi level. We have used photoemission spectroscopy to study the band bending and chemistry of these overlayers on n-type GaAs (100). Identically prepared samples show the characteristic increase in the photoluminescence signal, and have also been characterized using surface conductivity measurements. We find using photoemission that the surface Fermi level of the treated wafer is still near midgap. We also observe the chemistry at the interface, and offer a possible explanation of the photoluminescence and surface conductivity data in terms of it and the advanced unified defect model.

Mo/GaAs Schottky barriers prepared by d.c. sputtering

Mo/GaAs Schottky barriers have been prepared by d.c. sputtering, for different values of the sputtering voltage. Current-voltage and capacitance-voltage measurements show that these barriers can have very good properties (near one ideality factor, very low inverse saturation current) if suitable preparation conditions are chosen. A detailed study of the current-voltage characteristics as a function of the temperature allows us to analyse the carrier transport mechanisms and to correlate them to the preparation conditions. The experimental results show that the behaviour of the Mo/GaAs Schottky barriers, prepared by d.c. sputtering, can be successfully explained on the basis of the unified defect model proposed for the GaAs schottky barriers.