For an interface to be considered thermodynamically stable, the phases in contact must be in equilibrium with each other (connected by a stable tie-line) and have negligible mutual solubility on the phase diagram. The stability of Co based magnetic tunnel junctions (MTJs), with Co/MxO1-x/Co structures (M = Al, Gd, Hf, La, Mg, Si, Ti, Ta, Y and Zr), were evaluated with regard to these two conditions. Specifically, low temperature ternary isothermal phase diagrams were calculated and evaluated for the Co–M–O systems. All of these systems have at least one oxide in equilibrium with Co and thus have at least one thermodynamically stable tunnel barrier candidate for use in Co based MTJs. In light of the assumptions made in this analysis, along with the uncertainty in applying bulk enthalpy data to thin films, the current evaluation of interfacial stability serves as a first step in identifying suitable stable tunneling barrier materials in MTJs for detailed study.

Nickel is widely used in electrical contacts to InP, especially in NiAuGe ohmic contact to n-type InP. Several researchers have even suggested that the reaction of Ni with InP plays an important role in the ohmic nature of Ni-based contacts. However, numerous discrepancies are found in the literature concerning the Ni/InP reaction. In this study, an examination of the phase equilibria in the Ni-In-P system aids in an interpretation of bulk and thick film diffusion couples, and many of the apparent discrepancies in the literature are clarified. Finally, minor variations in the processing of the contacts (changes in the annealing gas and substrate cleaning procedure) are found in this study to have a less important role in altering the Ni/InP reactions than previous researchers have suggested.

A combined thermodynamic/kinetic methodology was presented to give a rational approach to the metallization of n-GaAs when a reciprocal system of GaAs-MδsGa-MsM'-MAs exists. Fortunately, for many quaternary systems consisting of Ga-M-M'-As, such a reciprocal system does exist. This methodology may be used either for Schottky enhancement or ohmic contacts to n-GaAs. Two examples were used to illustrate the approach and preliminary results were given.

Phase equilibria in the M-In-P (M = transition metal) ternary systems is determined through a combination of thermodynamic calculation and experimentation. Palladium is identified as a particularly attractive metal for contacts to InP. One useful feature of this system for contact design is the occurance of PdχInP ternary phases. Additionally, Pdln is found to be in equilibrium with InP and is identified as a promising contact material to InP because of its high melting point, low resistivity, and favorable properties for processing. Progress in the development of Pdln-based contacts to InP is briefly discussed.

Diffusion couple experiments were carried out for Mo/γ-TiAl at 900, 1000 and 1100°C for periods of time ranging from 121 to 553 hrs. Using the Boltzmann-Matano analysis, the two diagonal interdiffusion coefficients for the two phases formed in these couples were obtained assuming the cross diagonal terms to be negligible. These two phases are δ-(Mo, Ti)3Al and β-(Mo,Al)Ti. The three intrinsic diffusion coefficients were also obtained using the Darken-type relationships between the interdiffusion and intrinsic diffusion coefficients. The calculated interdiffusion coefficients from the three intrinsic diffusion coefficients are in reasonable accord with those obtained directly from the Boltzmann-Matano analysis.

A liquidus projection of the Al-rich Al-Li-Cu system is proposed. The proposed liquidus projection was based on DTA, X-ray diffraction, metallography, EPMA, SEM, and chemical analysis of 50 ternary alloys. Using these data and those reported in the literature, and thermodynamic models of Al-Li, AI-Cu and Cu-Li, a thermodynamic description of the Al-rich AI-Li-Cu system was developed. The calculated isothermal sections at several temperatures and the liquidus projection are in agreement with the experimental determinations. Combining the thermodynamic models and a Scheil-type equation, quantitative solidification paths were described. The calculated amount of primary solidification phase was compared to the experimental determination.

The interactions between Ni and Ni/In/Ni thin-films and GaAs were studied by SEM, SAM and AES. The presence of a molten phase was observed for both contacts after annealing at 820°C for 3 minutes. This behavior was rationalized in terms of the presence of a ternary eutectic reaction in the gallium-nickel-arsenic system. DTA confirmed the existence of the reaction:

at 810°C. In the case of Ni/In/Ni, melting was thought to occur because of the segregation of indium metal to the contact surface and the subsequent melting of the nearly ternary interfacial region. Upon further cooling the formation of NiGa, NiAs, InxGa1−xAs and an unspecified compound Ni-In was believed to occur. The contact was shown to be either sintered or alloyed, depending upon processing conditions.

The Wagner-Schottky model was used to describe the thermodynamic behavior of ordered intermetallic compound phases. To demonstrate the utility of the approach, the models developed for triple-defect B2 (and B32) and anti-structure L10 phases were used to describe the thermodynamic properties of β-AlLi and γ-TiAl respectively. Since any potential engineering materials to be developed on the basis of intermetallics will be multi-component systems, the methodology was extended to describe the thermodynamic properties of ternary intermetallics. The ternary Ti-Mo-Al system was used as an example for discussion. It is believed that the general topologies concerning the phase equilibria of Ti-M-AI with M being V, Nb, Ta, Mo and W are similar. The relative stabilities of the competing phases, i.e. BCC and HCP, in the mid-composition region of Ti-M-AI were discussed.

Solid-state reactions between niobium and gallium arsenide in both thin film and bulk forms were studied in the temperature range 400 to 1000 °C using transmission electron microscopy (TEM), metallography, scanning electron microscopy (SEM), and electron probe microanalysis (EPMA). Initially Nb4As3 and Nb5Ga3 formed at the interface and grew very slowly. Following an incubation period, NbAs and NbGa, nucleated and grew at rates several orders of magnitude faster than the initial phases. The resulting metastable diffusion path, Nb/NbGa3/NbAs/GaAs, was observed even after relatively long-term annealing and is believed to be kinetically stabilized. Formation of the other Nb–Ga binary compounds as predicted by the phase diagram was inhibited by nucleation and kinetic barriers. The observed reaction sequence is discussed considering the thermodynamics, kinetics, and possible growth mechanisms involved in the reaction.

Extensive new data and modeling in the In-Ga-As system has allowed the authors to reexamine the phase equilibria between the melt and the epitaxial solid. A detailed thermodynamic model was constructed with the following improvements: (1) The solid-solid interaction parameters were based on InAs-GaAs miscibility gap data, and (2) liquid-bulk solid, as well as liquid-epitaxial solid, tie-lines were used. Comparison of tie- lines from epitaxial systems and bulk systems demonstrated that strain energy is not the dominant factor in equilibrium growth of epitaxial solid films of In1-xGaxAs on any Ill-V binary substrate. Both the “lattice- pulling” effect and the “substrate-orientation” effect were shown to be caused by different quaternary equilibria at the In1-xGaxAs/InP interface, and not by film-substrate strain.

The chemical stability of interfaces between metals and GaAs was discussed in terms of reaction sequence and diffusion path concepts. The factors which determine interface morphology were also given. These general ideas can be applied to any interfacial reactions between two dissimilar materials.

Control of the structure and chemistry at the interfaces of compound semiconductors is essential for the commercial use of these materials in electronic and optical technologies. This can only be achieved when the governing thermodynamics and kinetics of interfacial reactions are understood. Based primarily on the experience of metal/Si interactions, however, a prevailing belief was born that thin-film reactions follow a separate set of thermodynamic and kinetic “rules” which are different from bulk reactions. The intent of our work has been to not only characterize metal/GaAs contact reactions but also to rationalize these reactions with equilibrium phase diagrams and bulk metal/GaAs diffusion couple experiments. Through this approach, a better understanding of thin-film and bulk differences has been obtained.

The Ir/GaAs system is used as an example. Phase formation and reaction kinetics were studied for 30 nm Ir films on (100) GaAs using TEM, XTEM, and AEM. Bulk diffusion between 0.25 mm thick Ir foil and (100) GaAs wafers was studied with SEM and electron probe microanalysis (EPMA). The diffusion paths and kinetics were the same for thin-film and bulk. The phase sequence Ir/IrGa/IrAs2/GaAs formed for all diffusion couples. Reaction kinetics were parabolic with an activation energy of 3.0 eV for both thin-film and bulk, and the data was colinear in an Arrhenius plot. Reacted layer morphology in both cases was layered. The effects of grain size, crystallographic texturing, and the relative diffusivities of the components on the reaction mechanisms in bulk versus thin-film reactions are considered.

The applicability of Nb as a Schottky barrier on GaAs depends to a large extent on the thermal stability of the contacts. In this study, bulk diffusion couple and phase diagram studies in addition to thin film studies were completed to understand the stability of and the reactions at the Nb/GaAs interface. Nb thin films were deposited onto GaAs substrates by dc magnetron sputtering and were annealed in the temperature range 300 to 1000°C. Analysis was done using plan-TEM and XTEM. The Nb/GaAs interface was found to break down into a series of binary compounds above 500°C. Bulk diffusion couples annealed at 600°C were analyzed using an electron microprobe. The stable sequence of phases formed in the couple, i.e., the diffusion path, was determined and was used to rationalize the observed compound formation in the thin film contact system.

Phase stabilities of iron alloys at low temperatures are strongly infuenced by magnetic effect. The appearance of certain type of equilibria is often due entirely to magnetic contribution to the Gibbs energy of the pertinent phase. The appearance of the stable and metastable equilibria in fcc(Fe,Ni) alloys are discussed in terms of the magnetic interaction.