Shlomo Mehari

Epitaxial NiInGaAs formed by solid state reaction on In0.53Ga0.47As: Structural and chemical study

Thin epitaxial layers of NiInGaAs formed by solid state reaction of Ni on (100) In0.53Ga0.47As are used as metal source and drain regions for In0.53Ga0.47As metal oxide field effect transistors. Here, the authors present a structural and chemical analysis of this phase. The stoichiometry of the layer was determined as Ni2In0.53Ga0.47As. Transmission electron microscopy revealed an abrupt interface and a detailed x-ray diffraction analysis showed that the layer is of a hexagonal lattice, which grows epitaxially with the orientation relations of {100}InGaAs||{100}NiInGaAs; ⟨011¯⟩InGaAs||[001]NiInGaAs. Only one domain can be observed in this epitaxial growth. Understanding the structure of these layers is a crucial step not only in their incorporation into InGaAs based devices but also a step toward novel devices.

Measurement of the Schottky barrier height between Ni-InGaAs alloy and In0.53Ga0.47As

The temperature dependence of the current-voltage characteristics of Ni-InGaAs alloy Schottky contacts to n-In0.53Ga0.47As was measured. Nearly ideal plots with an ideality factor close to unity were obtained. The Arrhenius curve across the wide temperature range of 80–300 K was perfectly linear, yielding a barrier height of 0.239 ± 0.01 eV. This value is substantially larger than previously reported. Conventional metal based Schottky diodes did not exhibit an ideal Schottky behavior. The ideal Schottky diode characteristics are attributed to the lack of oxidation and contamination of the interface between Ni-InGaAs and InGaAs.

Role of Transport During Transient Phenomena in AlGaN/GaN Heterostructure FETs

Transient phenomena in AlGaN/GaN heterostructure field-effect transistors are attributed to trapping and detrapping of electrons from deep localized states, and the measured activation energies are conventionally associated with the electron capture and emission processes. This standard interpretation ignores, however, transport between the two-dimensional electron gas and the trap. Using gated van der Pauw structures, we demonstrate that the transient behavior is determined by transport (at least for the trapping process). The measured activation energy is, therefore, a characteristic of the transport process rather than the emission-capture processes.

Density of Traps at the Insulator/III-N Interface of GaN Heterostructure Field-Effect Transistors Obtained by Gated Hall Measurements

We apply the gated Hall method to obtain the density of trap distribution, Dit(E), at the insulator/III-N interface of a heterostructure field effect transistor. It is shown that Dit is proportional to the difference between the steady-state and the high-frequency prediction of the gate-induced two-dimensional electron gas concentrations. The Dit profile at the SiNx/GaN interface in the energy range of 1.2-2.3 eV below the GaN conduction band is obtained as a demonstration of the method.

Comparison of Simulation and Measurement of Gate Leakage Current in Metal/Al 2 O 3 /GaN/AlGaN/AlN/GaN Capacitors

We compare the simulations and measurements of gate leakage current in metal-insulator-semiconductor capacitors fabricated on standard AlGaN/GaN transistor layers. In the simulations, a fixed charge density at the interface between the Al2O3 layer and the GaN cap layer was assumed for simplicity. The fixed charge layer (which depends on bias and device history) was obtained from capacitance-voltage measurements. The calculated tunneling current due to field emission from localized states in the GaN cap layer agreed with the experimental results. No additional leakage mechanisms had to be invoked.

Identification of Energy and Spatial Location of Electron Traps in AlGaN/GaN HFET Structures

We present a methodology, based on gated van der Pauw measurements, to identify the electron trap charging or discharging energy at different spatial locations in AlGaN/GaN heterostructures. The slow transient response of the 2-D electron gas concentration was recorded following different Schottky gate and substrate voltage pulses and at different temperatures. While recovering from a gate voltage stress, both trapping and detrapping processes are observed at different times and temperatures, demonstrating that dynamic rather than quasi-equilibrium considerations should be invoked to explain the data. Several distinct electron traps were identified, three traps located in the GaN buffer layer having detrapping activation energies of about 0.55, 0.45, and 0.21 eV, and an activation energy of about 0.64 eV was associated with trapping processes in the AlGaN barrier layer.

The Role of Barrier Transport and Traps in the Tradeoff Between Low OFF-State Leakage Current and Improved Dynamic Stability of AlGaN/GaN HFETs

A tradeoff behavior between low Schottky gate leakage current and improved dynamic stability of AlGaN/GaN heterostructure FETs was previously reported, and was attributed to variations in the metal/semiconductor interface properties. Here, we show that a tradeoff behavior is found in transistors and gated van der Pauw test structures that were fabricated on the same wafer, and underwent identical surface treatments. In the gated van der Pauw measurements, a slow transient response and a large variation in the 2-D electron gas concentration compared with equilibrium were detected, following a gate voltage stress in low leakage devices, but not in high leakage ones. Based upon our experimental observations, we argue that electron traps located in the AlGaN barrier layer are responsible for the tradeoff. The activation energies required to fill the empty barrier traps by electrons, obtained from the gated van der Pauw measurements, are 0.60 and 0.65 eV.

Gated van der Pauw measurements: A powerful tool for probing electron trapping effects in GaN HEMTs

Gated van der Pauw structures can be used to distinguish between different trapping effects in AlGaN/GaN HEMT layers, and evaluate trap density. Activation energies can also be obtained [2]. The absence of transistor access region effects greatly simplifies the interpretation of the data compared to transistor pulsed I-V experiments.

Study of the Ni-InGaAs alloy as an ohmic contact to the p-type base of InP/InGaAs HBTs

Following the silicide to silicon contact approach, Ni-InGaAs alloy was studied as an ohmic contact to p-type InGaAs. The Schottky barrier height of this system is similar to that of conventional metals such as Ti, but the interface is oxide free. The obtained specific contact resistivity to p-type material was substantially lower than that of standard Pt based metallic contacts. However, if employed for the fabrication of the base contact to HBTs, nickel thickness variations may degrade the performance of the base collector junction.