In this paper we describe the results of a study on the hydrogenation treatment of multicrystalline substrates by an RF-plasma with emphasis on discriminating between effects on the intra-grain material and grain boundary regions. For this purpose small mesa-type diodes were processed. Two types of multicrystalline material are being compared in this study. The main difference between these materials is their oxygen and metallic impurity content. The effects of the hydrogenation treatment were studied by means of I-V and DLTS-measurements. Finally, we will present data on small multicrystalline solar cells and the effect of hydrogenation on the main parameters of this device to illustrate the correlation and the differences between the measurements on small-scale diodes and the effects on the macroscopic device which is the solar cell.

We applied the ultrasound treatment (UST) to improve properties of poly-Si thin films on glass substrates for thin-film transistor applications. A strong decrease of the sheet resistivity in hydrogenated films subjected to UST was observed. UST improves the film homogeneity as monitored by spatially resolved surface photovoltage mapping. Studies of hydrogenated thin-film transistors demonstrated remarkable UST induced improvement in transistor characteristics, especially, a reduction of leakage current by as much as one order of magnitude. All these effects are explained in terms of UST enhanced hydrogenation of poly-Si film.

The effect of hydrogen passivation on photovoltaic performance of 1 MeV electron irradiated polycrystalline cast silicon solar cells is described. These cells were processed on cast p-type boron doped polycrystalline silicon substrates using standard technology. Passivation was made by low-energy hydrogen ion implantation on the front side. Cells performance was measured as a function of fluence, and it was found that the hydrogenated cell had the higher radiation resistance.

Defect behavior were studied using deep level transient spectroscopy and infra-red spectroscopy. It was shown that the concentration of vacancies (Ec −0,09 eV), divacancies (Ec −0,23 eV) and A-centers (Ec −0,18 eV) is significantly lower in hydrogenated samples. This consistency strengthens the belief that hydrogen interacts with vacancy-type defects to prevent formation of the secondary radiation defects. It is confirmed by IR-measurements.

The passivation by hydrogen of the shallow donors sulfur, selenium, and tellurium in GaAs was studied by infrared absorption spectroscopy (FTIR), capacitance-voltage (CV) depth profiling, and secondary ion mass spectroscopy (SIMS). Local vibrational mode (LVM) frequencies due to hydrogen-donor complexes agree with a microscopic model where the hydrogen atom is bound in the antibonding position to one of the donor’s neighboring gallium atoms. Our results suggest a hydrogen penetration depth much greater than the passivated region, however the infrared active region coincides with the passivated part of the material.

Hydrogen passivation effects in In doped n-CdTe upon exposure to rf hydrogen plasma have been studied by electrical and photoluminescence measurements. Shallow dopant passivation of approximately an order of magnitude at 150°C and 50% at 170°C is observed. No visual damage is seen. Reverse bias annealing effects are also studied. Results are discussed.

Deep electron traps in n-InP introduced during plasma exposure have been studied by means of isothermal capacitance transient spectroscopy (ICTS). Three electron traps, (Ec–0.21 eV), (Ec–0.34 eV) and (Ec–0.54 eV), which are designated E2, E3 and E4, respectively, are detected in n-InP treated with H2 plasma and by subsequent annealing. The E2 trap is induced by plasma exposure and the E3 trap is produced by thermal annealing. The E4 trap is generated by both plasma exposure and thermal annealing. These three traps are passivated with hydrogen atoms. The E2 trap density near the surface of hydrogen-plasma-treated samples is strongly enhanced by applying electric field because of dissociation of hydrogen from E2 trap. The E2 trap is annealed out with the activation energy of 1.5 eV and the attempt-to-escape frequency of 3.2 × 1014 s−1. Phosphine plasma treatment is effective in suppressing generation of these electron traps.

It has recently been shown that deuterium diffusion experiments can provide information on the deepening of the hydrogen acceptor level in the band gap of AlxGa1-xAs alloys with increasing x. In the present work, we report on the influence of hydrostatic pressure on deuterium diffusion in n-GaAs:Si. SIMS analysis reveals that the deuterium profiles in n-GaAs:Si are sensitive to hydrostatic pressure: the diffusion depth decreases and a plateau appears in the diffusion profile as the pressure is applied. The results are interpreted in terms of an increasing amount of the H- species as pressure is applied. This increase is mainly attributed to a deepening of the H acceptor level with respect to the bottom of the Γ conduction band of GaAs. Qualitatively, this effect is similar to the deepening of the H acceptor level in AlxGa1-xAs alloys as x increases.

GaAs grown at low temperatures by MBE has been doped with shallow impurities to ∼1019 cm−3. A layer doped with Be acceptors was completely compensated and the simultaneous detection of AsGa0 and AsGa+ confirmed that the Fermi level was close to midgap and that compensation was partly related to AsGa defects. There was no evidence for the incorporation of VGa in this layer. For Si-doped samples, more than 80 % of the donors were compensated and the detection of SiGa-VGa pairs by infrared localized vibrational mode (LVM) spectroscopy indicated that VGa were at least partly responsible. Increasing the Si concentration suppressed the incorporation of AsGa. Exposure of the Be-doped layer to a hydrogen plasma, generated a LVM near 2000 cm−1 which may be the stretch mode of a AsGa-H-VAs defect complex. For Si-doped layers, two stretch modes at 1764 cm−1 and 1773 cm−1 and a wag mode at 779 cm−1 relating to a H-defect complex were detected and we argue that the complex could be a passivated As antisite. The detection of characteristic hydrogen-native defect LVMs may provide a new method for the identification of intrinsic defects.

Deep-level formation upon plasma hydrogenation has been studied with n-GaAs grown by various methods. Four electron traps (EH0-EH3) were generated in As-rich n-GaAs crystals. No electron traps were observed in the LPE layer before and after hydrogenation. The hydrogen as well as excess arsenic defects are responsible for the formation of these deep levels. Two of the generated levels in our study, EH0/EH2, exhibit metastability and are identical to the M3/M4 levels reported by Buchwald et al. It can be speculated that both diffused hydrogen and already existing As antisite defects are responsible for the generation of the metastable defects.

The effect of hydrogenation on the photoluminescence (PL) of InP : Mg, InP : Zn and undoped n-InP is presented. An increase in the near band edge pl intensity due to passivation of non-radiative centers was observed in all the samples. A donor - acceptor pair transition was observed before hydrogenation in the InP : Mg sample and after hydrogenation in the InP : Zn sample due to the acceptor deactivation. In n-InP the enhancement of donor bound exciton after hydrogenation points to the absence of donor passivation.

We study the hydrogen-induced passivation of interface traps in n-GaAs:Si, grown by MBE at 580°C, with a buffer layer grown at 200 to 250°C (LT-buffer). In the as-grown samples, the LT-buffer contains As-precipitates, with a density of 4×1016 cm−3 and a diameter that depends on the LT-buffer growth temperature and takes values in the range 4–8nm. The epilayer/LT-buffer interface in the as-grown samples is characterised by interface related traps which dominate the electrical behaviour of the epitaxial layer. Secondary ion mass spectroscopy profiling of deuterated samples reveals that 2H has a nearly uniform concentration throughout the epilayer, while it accumulates at the interface. Hydrogenation induces a reduction of the interface trap concentration and a significant improvement of the carrier mobility values.

We discuss the application of state-of-the-art first-principles calculations to the problem of defects, impurities, and doping levels in semiconductors. Since doping problems are of particular relevance in wide-band-gap materials, we focus here on studies of ZnSe and GaN. For ZnSe, we discuss our latest insights in the influence of compensation and dopant solubility on the experimentally observed limitation of the free carrier concentration in p-type ZnSe. For GaN, we focus on the role of native defects in doping or compensation of the material, with particular emphasis on the n-type conductivity of as-grown GaN.

Gallium nitride (GaN) films grown by hydride vapor phase epitaxy on a variety of substrates have been investigated to study what role silicon and oxygen impurities play in determining the residual donor levels found in these films. Secondary ion mass spectroscopy analysis has been performed on these films and impurity levels have been normalized to ion implanted calibration standards. While oxygen appears to be a predominate impurity in all of the films, in many of them the sum of silicon and oxygen levels is insufficient to account for the donor concentration determined by Hall measurements. This suggests that either another impurity or a native defect is at least partly responsible for the autodoping of GaN. Additionally, the variation of impurity and carrier concentration with surface orientation and/or nucleation density suggests either a crystallographic or defect-related incorporation mechanism.

We report on the isolation properties of In0.75Al0.25N implanted with either N or O for several doses and post-implant anneal temperatures. Sheet resistance versus anneal temperature data are reported for the various implants with a maximum sheet resistance of <1×109 Ω/□ achieved for a high dose N-implant annealed at 600 or 700 °C and <5×108 Ω/□ achieved for a high dose O-implant annealed at 600 °C. These sheet resistances correspond to a greater than three order of magnitude increase over the as-grown values. The compensating defect level for the highest resistance N-implanted sample has an estimate ionization level 580 meV below the conduction band edge. Implant isolation of InAIN is also compared to oxygen implant isolation of InxGa1-xN — where only a 50 to 100 fold increase in sheet resistance is observed — to study the effect of Al in the isolation scheme.

Electronic defects in MOCVD-grown n-type GaN were characterized by conventional deep level transient spectroscopy (DLTS) and by photoemission capacitance transient spectroscopy (O-DLTS) performed on Schottky diodes. With DLTS two deep levels were detected with thermal activation energies for electron emission to the conduction band of 0.16 eV and 0.44 eV. With O-DLTS we demonstrate four new deep levels with optical threshold energies for electron photoemission of ∼ 0.87 eV, 0.97 eV, 1.25 eV and 1.45 eV. The O-DLTS apparatus and the measurement are discussed in detail. We also report characterization of the Au-GaN barrier height of the Schottky diode by internal photoemission.

Hydrogen is found to readily diffuse into InGaN, InAIN and InGaAIN epitaxial layers during plasma exposures at 170 – 250 °C for 40 sec - 30 min. The diffusivity of hydrogen is > 10−11 cm2-s−1 at 170 °C, and the native donor species are passivated by association with the hydrogen. Reactivation of these species occurs at 450–500°C, but the hydrogen remains in the material until ≥ 800 °C.

Based on extensive first-principles total-energy calculations we study the electronic structure, atomic geometry and energetics of atomic hydrogen in cubic GaN. All charge states of hydrogen (H+, H0, H-) are examined. For H- the gallium tetrahedral interstitial site is energetically most stable. All other sites are much higher in energy, indicating a high diffusion barrier for H- in GaN. H+ favors positions on a sphere with a radius of ≈ 1 Å and a nitrogen atom in the center. Among these positions the nitrogen antibonding site is energetically most stable. An unexpectedly large negative-U effect (U = —2.5eV) indicates that H0 is unstable.

Wide bandgap GaN very often shows a high electron concentration. Although several impurities such as O and Si have been identified, the concentration is not high enough to account for the number of free carriers. As a consequence native defects namely the nitrogen vacancies are widely considered to be present at high densities. Several calculations predict different energy levels of this strongly localized defect. We present photoluminescence experiments of wurtzite GaN and AlGaN layers under large hydrostatic pressure to search for localized defects within the questionable energy range of 3 .0 to 3 .8 eV above the valence band edge.

The optical absorption in gallium nitride (GaN) films was studied by photoconductivity (PC) spectroscopy at various temperatures. At all the measured temperatures, the photoconductivity per incident photon increases with photon energy hv from 1.8 eV to 3.0 eV approximately as exp(hv/E0). Surprisingly, the measured photoconductivity tail to the infrared becomes more pronounced in magnitude at lower temperatures. We suggest from the data that the photoconductivity is dominated by majority carriers, and that the quantum efficiency η for conducting electron and hole generation by across-gap excitation increases with increasing temperature.

Free and bound exciton luminescences as well as donor-acceptor pair recombination of GaN epitaxial layers on 6H-SiC and sapphire substrates were investigated using time integrated and time resolved photoluminescence measurements at low temperatures. Lifetimes are determined for the donor bound exciton at 3.4722eV and for two acceptor bound excitons with energies of 3.4672eV and 3.459eV. Luminescences between 3.29eV and 3.37eV are identified as due to excitons deeply bound to centers located near the substrate-epilayer interface.