J. M. Redwing

p-GaN surface treatments for metal contacts

The chemical bonding and electronic properties of wet, chemically treated p-GaN surfaces were studied using synchrotron radiation photoemission spectroscopy. Chlorine-based chemical bonding was identified on the conventional HCl-treated p-GaN surface, which is associated with a shift of the surface Fermi level toward the conduction band edge by ∼0.9 eV with respect to the thermally cleaned surface. Compared to the HCl-treated surface, the surface Fermi level on the KOH-treated surface lies about ∼1.0 eV closer to the valence band edge, resulting in a much smaller surface barrier height to p-type materials than the HCl-treated surface. The smaller surface barrier height to p-GaN after KOH treatment can lead to a lower contact resistivity and can play an important role in lowering the metal contact resistivity to p-GaN.

THERMALLY STABLE PTSI SCHOTTKY CONTACT ON N-GAN

Platinum silicide (PtSi) and Pt Schottky contacts on n-GaN have been investigated and compared. The PtSi contacts were formed on n-GaN by annealing a multilayer structure of Pt/Si with the appropriate thickness ratio at 400u2009°C for 1 h in forming gas. The barrier height of the as-formed PtSi contacts was found to be 0.87 eV capacitance–voltage (C–V), and remained unchanged after further annealing at 400 and 500u2009°C. Upon annealing at 600u2009°C for 1 h, the barrier height decreased to 0.74 eV (C–V), but the diodes remained well-behaved. The as-deposited Pt yielded a barrier height of 1.0 eV (C–V). Upon annealing at 400u2009°C for 1 h, the Pt diodes degraded and most of the diodes did not survive additional annealing at 400u2009°C for longer times. The electrical measurements and the Rutherford backscattering spectrometry results indicated that PtSi contacts are thermally much more stable than Pt contacts on GaN.

X-ray photoemission spectroscopic investigation of surface treatments, metal deposition, and electron accumulation on InN

The effects of surface chemical treatments and metal deposition on the InN surface are studied via synchrotron-based photoemission spectroscopy. Changes in the In 4d core level as well as the valence band spectra are reported. The surface Fermi level position, EF, relative to the valence band maximum was determined for both Au and Ti Schottky barriers. EF lies at an energy of 0.7 eV above the valence band maximum for Au deposited on annealed InN and 1.2 eV above the valence band maximum for Ti deposited on Ar-sputtered InN. These results that the surface Fermi level lays at or above the conduction band maximum when a value of InN band gap of 0.7–0.9 eV is assumed.

GaN Growth by Metallorganic Vapor Phase Epitaxy A Comparison of Modeling and Experimental Measurements

A model for the growth of gallium nitride in a vertical metallorganic vapor phase epitaxy (MOVPE) reactor is presented and compared to experimental growth rate measurements. For a mixture of nondilute gases, the flow, temperature, and concentration profiles are predicted using recent kinetic data. Growth rates are predicted based on simple reaction mechanisms and compared with those obtained experimentally. These comparative results show that the growth of GaN epi layers proceeds through an intermediate adduct of trimethylgallium and ammonia. Loss of adduct species due to oligmerization leads to the lowering of the growth rate. Quantification of this loss of reacting species is made based on experimentally observed growth rates. An apparent chemistry model is presented based on the salient features of GaN MOVPE. Process conditions are perturbed to obtain trends in growth rate and uniformity in order to demonstrate the utility of such a model in optimizing the GaN MOVPE process.

Carbon doping in metalorganic vapor phase epitaxy

Abstract Carbon in metalorganic vapor phase epitaxy has been studied as both an unintentional and intentional dopant. It is generally an acceptor in most compound semiconductors with a very low thermal diffusion coefficient. Carbon limits the ultimate purity of the grown material when present as an unintentional impurity. Carbon can originate from the metal-organic compounds used in the growth of the materials and the study of carbon incorporation can lead to insights into the underlying chemical mechanisms of the metalorganic vapor phase epitaxy (MOVPE) growth process. The intentional doping of carbon in MOVPE, and allied technologies such as MOMBE, has replaced several of the traditional dopants for applications in heterojunction devices. This paper will review and discuss several of these aspects of carbon doping: the introduction of carbon into the growing layer, the properties of carbon in binary and alloy semiconductors, and device applications of carbon doping.

A near‐field scanning optical microscopy study of the photoluminescence from GaN films

We have achieved spatially resolved photoluminescence from GaN films using a near‐field scanning optical microscope (NSOM). GaN films grown by hydride vapor phase epitaxy (HVPE) and metal‐organic vapor phase epitaxy (MOVPE) on sapphire substrates have been studied. We have performed spatial scans of topography, band edge, and yellow luminescence signals. Atomic force microscopy measurements were also made and compared with the NSOM topography. We have found spatial variations in photoluminescence characteristics at the submicron scale for both HVPE and MOVPE GaN. The observed enhancement of yellow luminescence at multiatomic step edges on the HVPE GaN surface suggests that the yellow luminescence is associated with chemical impurities incorporated during the growth of GaN films.

Photoelastic waveguides and the controlled introduction of strain in III‐V semiconductors by means of thin film technology

We have investigated the use of thin film technology to introduce controllable and thermally stable stress into semiconductor heterostructures. Two simple schemes are used. The first scheme is to use interfacial reactions between a metal and the substrate, such as Ni, Co, Pd, and Pt on GaAs/AlGaAs. The induced stress in the structure is reproducible and controllable because the volumetric change for a given reaction is fixed, as long as the deposited film is fully reacted to form a compound. The stability of the stress depends on the stability of the compound. In the case of Ni and Co on GaAs/AlGaAs, the induced stress is thermally stable up to 600u2009°C. Evaporated films and reacted films are usually under tension. The second scheme is to use rf sputtered W or WNi alloy films where W or WNi is sputtered onto a negative dc biased substrate. This scheme effectively provides highly compressed films. The thermal stability depends on the concentration of Ni in the WNi alloy. Using the two schemes above, we have ...

Study of the gas phase chemistry in the silicon doping of GaAs grown by metalorganic vapor phase epitaxy using tertiarybutylarsine as the group V source

Abstract Gas phase decomposition studies have been combined with metalorganic vapor phase epitaxy (MOVPE) growth and doping experiments to investigate the mechanism behind the Si doping of GaAs grown using tertiarybutylarsine (t-C 4 H 9 AsH 2 or TBAs) as the Group V source. The use of TBAs leads to an increase in the Si doping efficiency from both SiH 4 and Si 2 H 6 sources and this enhancement was proposed to originate from the homogeneous generation of Si-bearing intermediates, (t-C 4 H 9 )SiH 3 (TBSi) and SiH 3 AsH 2 , from reactions between SiH 4 and the pyrolysis products of TBAs. We have investigated the formation and role of these intermediates in this Si doping chemistry. Results of our decomposition study show that TBSi pyrolyzes in the gas phase at relatively low temperatures to form SiH 4 . Consequently, the use of TBSias a Si dopant source results in a Si doping efficiency and temperature dependence that is identical to SiH 4 . TBSi is therefore not a likely route to increased Si incorporation. Our studies of the co-decomposition of TBAs and SiH 4 or Si 2 H 6 show that SiH 3 AsH 2 is a co-pyrolysis product of TBAs and Si 2 H 6 , but not of TBAs and SiH 4 . Furthermore, there is no indication of gas phase interaction between TBAs and SiH 4 . The results of this study suggest that homogeneous reactions are not as important in this doping chemistry as had previously been considered and that surface processes may instead be responsible for the increase in Si doping observed with TBAs.

Photoluminescence studies of erbium-doped GaAs under hydrostatic pressure

The photoluminescence properties of metal-organic chemical vapor deposition GaAs:Er were investigated as a function of temperature and applied hydrostatic pressure. The 4I13/2→4I15/2Er3+emission energy was largely independent of pressures up to 56 kbar and temperatures between 12 and 300 K. Furthermore, no significant change in the low temperature emission intensity was observed at pressures up to and beyond the Γ-X crossover at ∼41 kbar. In contrast, AlxGa1−xAs:Er alloying studies have shown a strong increase in intensity near the Γ-X crossover at x∼0.4. These results suggest that the enhancement is most likely due to a chemical effect related to the presence of Al, such as residual oxygen incorporation, rather than a band structure effect related to the indirect band gap or larger band gap energy. Modeling the temperature dependence of the 1.54 μm Er3+ emission intensity and lifetime at ambient pressure suggested two dominant quenching mechanisms. At temperatures below approximately 150 K, thermal quenc...

Growth studies of erbium‐doped GaAs deposited by metalorganic vapor phase epitaxy using novel cyclopentadienyl‐based erbium sources

Erbium‐doped GaAs layers were grown by metalorganic vapor phase epitaxy using two new sources, bis(i‐propylcyclopentadienyl)cyclopentadienyl erbium and tris(t‐butylcyclopentadienyl) erbium. Controlled Er doping in the range of 1017–1018 cm−3 was achieved using a relatively low source temperature of 90u2009°C. The doping exhibits a second‐order dependence on inlet source partial pressure, similar to behavior obtained with cyclopentadienyl Mg dopant sources. Equivalent amounts of oxygen and Er are present in ‘‘as‐grown’’ films indicating that the majority of Er dopants probably exist as Er‐O complexes in the material. Er3+ luminescence at 1.54 μm was measured from the as‐grown films, but ion implantation of additional oxygen decreases the emission intensity. Electrical compensation of n‐type GaAs layers codoped with Er and Si is directly correlated to the Er concentration. The compensation is proposed to arise from deep centers associated with Er which are responsible for a broad emission band near 0.90 μm pres...