The efficacy of implementing terrestrial-based photovoltaics is dictated by trade-offs in device performance, cost, and reliability. Presently, the highest efficiency polycrystalline CdS/CdTe superstrate solar cells utilize back contacts containing copper as an intentional dopant. Accelerated stress data correlates copper diffusion from this contact with performance degradation. Degradation at the device level exhibits two characteristic modes that are influenced by CdTe surface treatments prior to contacting. Rapid degradation associated with a rapidly decreasing open-circuit voltage can occur in cases where processing favors stoichiometric CdTe surfaces. Slower degradation associated with roll-over is typified by treatments favoring the presence of Te at the back contact. The chemical composition and extent of Te-rich contact interfaces is revealed by transmission electron microscopy. Deep-level transient spectroscopy of NP etched and non-etched devices show Te-rich conditions are necessary for the detection of deep-acceptor CuCd defect levels at (Ev +0.28 to 0.34 eV). Low keV cathodoluminescence measurements show that these defects can be found localized at the back surface of CdS/CdTe devices.

Copper trace analysis using Transient Ion Drift (TID) combined with a Rapid Thermal Annealing (RTA) process is investigated. A double pulse method is implemented to allow unambiguous identification of the copper-induced capacitance signal. Use of a mercury probe as sensing Schottky barrier enhances the flexibility of the method and allows mapping of the contaminant. The method is evaluated on quantitatively contaminated silicon wafers and compared to Total X-ray fluorescence (TXRF).

It is shown that in Czochralski grown material, the RTA is sufficient to dissolve most copper atoms into interstitial sites independently of their initial configuration. As a result, both, the surface and bulk contamination can be monitored by RTA/TID with a bulk detection limit close to 1011cm-3.

In Float Zone material mapping of the quenched interstitial copper revealed the existence of defect reactions involving presumably vacancy clusters.

The recent work in the field is reviewed and some new applications are reported. In particular we review some aspects where new contributions have been added. The mechanism of bubble formation when He is implanted into silicon is described till the supersaturation of vacancies (void formation). The void evolution has been described considering direct coalescence or Ostwald ripening. Applications such as gettering, lifetime control, and more recently nanosensors for interstitials are critically discussed.

Single crystal wafers of <0001>-oriented 6H-SiC were irradiated at different temperatures using a variety of ion species. The disorder on both the Si and C sublattices has been studied in situ using a combination of ion beam analyses in multiaxial channeling geometry. The fraction of the irradiation-induced defects surviving simultaneous recovery processes decreases with decreasing ion mass and with increasing irradiation temperature. Some of the Si and C defects are well aligned with the <0001> axis and the rate of C disordering is higher than that of Si disordering. Three recovery stages in Au2+-irradiated 6H-SiC have been identified.

Spatial ordering of InAs quantum dots was attained by using misfit dislocations generated in a metastable InGaAs layer by means of thermal annealing. Influence of quantum dot positional ordering and dot proximity to dislocation arrays on carrier dynamics was studied by timeresolved photoluminescence. Substantially narrower inhomogeneous broadening from the ordered quantum dots was observed. Excitation intensity dependence of the photoluminescence intensity and carrier lifetime indicates stronger influence of nonradiative recombination for the ordered quantum dot structures. Numerical simulations allow estimating electron and hole capture rates from the quantum dots to traps located either at the quantum dot interfaces or in the vicinity of the quantum dots.

High dose He implantation, followed by a thermal annealing, is a suitable technique for metal gettering. Nevertheless, a strong interaction between the dopants and the defect layer has been evidenced. This can largely influence the dopant distribution. In order to study this interaction, p and n-type samples uniformly doped were implanted with helium (40 keV, 5×1016 He+.cm-2) and furnace annealed for various times and temperatures. In this paper, we shed light on the evolution of the dopant segregation. Using isochronal treatment, we found a large dependence of the dopant gettering phenomenon upon annealing temperature. Moreover, stability of the gettered fraction is observed for isothermal annealing. This study permits also to investigate the origin of the trapping mechanism involved for both boron and phosphorus.

We present a theoretical investigation of the effects of dislocations, impurities and dopants on the thermal conductivity of GaN layers. It is shown that the experimentally observed decrease of the room-temperature thermal conductivity with increasing doping density is a result of enhanced phonon relaxation on silicon dopant atoms. Scattering of acoustic phonons on free carriers plays a relatively minor role in GaN. The functional dependence of the thermal conductivity on doping density is in good agreement with experiment. A developed model can be used for thermal budget calculation in high-power density GaN devices.

A thermodynamic model for the site preference of dilute solutes in ordered compounds was developed recently. The model is extended here in two ways to help interpret experimental results from this laboratory. (1) A preference for solutes to occupy sites in ‘sinks’ such as grain boundaries rather than regular lattice sites is included using a simple model. A wide range of site preference behavior is found; for example, a solute may switch between substitutional and sink sites with changing composition. (2) The effect of an attractive interaction between solute and defect on the site preference is examined. Conditions are established under which a solute can be stabilized by association with a defect on a site where the solute would otherwise not be found.

Ion-implantation has been an interesting topic on impurity-doping in GaN. Raman measurement is a strong tool for the characterization of semiconductors. We have investigated the Raman scattering spectra in Be-implanted GaN epilayers. In as-implanted GaN, new Raman bands at ∼310, ∼360, 669 cm-1 appeared. From phonon-dispersion curves for hexagonal GaN, the ∼300 cm-1 and 669 cm-1 bands were assigned to the highest acoustic-phonon branch and the optical-phonon branch at the Brillouin zone boundaries, respectively. Two sharp bands at 168 and 199 cm-1 were observed in the Raman spectra of Be-implanted GaN after post-implantation annealing treatments. We tentatively assign these two bands to Be-related local vibrational modes.

We have measured the thermal activation energies of electrically active defects in 4H- and 6H-SiC implanted with either Al or B ions, using Thermal Admittance Spectroscopy. The net acceptor concentrations were monitored using room temperature low frequency Capacitance-Voltage measurements. The substrates were n/n+ epilayers. The implantations plus annealing produced p-type layers that were acceptable for characterization. The specimens were annealed in Ar at 1600 ° C, after which Ni Schottky diodes were fabricated on the specimens. Annealing times were 5, 15, 30, and 60 min. In some of the specimens, a shallow level was found that did not correspond to known levels. As the annealing progressed, energy shifts were noted for some of the detected levels. In some specimens, the implanted p-type impurity and the n-type residual dopants in the substrate were simultaneously detected. Measurements of electrically active ptype species were compared to “control” specimens implanted with Ar. From this comparison, we conclude that at least one shallow donor level is introduced into the bandgap by the implantation process, and is not annealed out. The defects associated with the implantation may affect actual device performance of diodes by destabilizing the lattice occupation of the implanted dopant atoms (energy shift with annealing), and act as lifetime killers.

We have investigated the electronic structures of p- or n-type doped ZnO and ZnS based on the basis of the analysis of the results obtained by ab initio electronic band structure calculations. We find the doping problems of the two materials. They at highly N-doping levels have the narrow impurity bands at the top of the valence band, where holes are very localized by repulsion effects. As a solution to the doping problems described above, we proposed a codoping method: The deliberate codoping of the donors with N acceptors is essential for the enhancement of N-acceptor incorporation and delocalization of the N-impurity states with the electrostatic energy gain in p-type highly doped ZnO and ZnS.

Silicon nanowires (SiNW) grown using Si monoxide source materials are mainly oriented in the <112> direction, and some in the <110> direction. These directional features may be understood by postulating that growth of the SiNW is governed by lateral advancement of {111} plane layers that are stepped, and the role of particular kinds of dislocations providing perpetuate steps.

Deep level transient spectroscopy (DLTS) studies of both p-type (uid) and n-type (Sidoped), lattice-matched, 1.05 eV bandgap InGaAsN grown by molecular-beam epitaxy (MBE) are reported, and the results are compared to previous measurements of similar materials grown by metal-organic chemical-vapor deposition (MOCVD). In MBE-grown p-type InGaAsN, two majority-carrier hole traps were observed: H3' (0.37 eV) and H4' (0.51 eV), and no evidence was found for the presence of minority-carrier electron traps. These two traps appear to be similar to two levels, H3 (0.48 eV) and H4 (0.5 eV), previously characterized in MOCVD-grown InGaAsN. In MBE-grown n-type InGaAsN, we observed a shallow distribution of electron levels, E1' (0 < EA < 0.35 eV), as well as a deep electron trap E4' (0.56 eV) and a deep hole trap H5' (0.71 eV). E1' appears to be coincident with a superposition of two levels observed in MOCVD-grown InGaAsN, a shallow distribution termed E1 (0 < EA < 0.20 eV) and a discrete (though broadened) level E3 (0.34 eV). Further, E4' appears to be similar in character to a level observed in MOCVD-grown material, E4 (0.82 eV), although a disparity in activation energy exists. This disparity may be due to a temperature-dependent capture cross-section for one or both levels, a possibility that is currently under investigation. In contrast, H5' appears to have no analogue in MOCVD-grown material and thus may be unique to the MBE growth technique.

A focused ion beam system is applied to investigate the dose dependence of the shape of Ge channeling implantation profiles in Si and SiC at two very different dose rates (1011 and 1018 cm-2 s-1), and for implantation temperatures between room temperature and 580 °C. The competing influence of dose rate and temperature observed is explained in terms of intracascade defect relaxation. For the different implantation temperatures, the time scale for defect reduction is estimated. The results obtained for Si are compared with those for SiC.

The effects of hydrogen on the deep level luminescence in Si, ZnO and diamond were studied by means of cathodoluminescence. It is well known that most of the deep levels in Si are passivated by hydrogen. Scratch lines on Si surface, which do not show any characteristic luminescence, obtain so-called D-line luminescence by hydrogen plasma treatment. It indicates that only nonradiative defects are passivated but D-line luminescence is not passivated by hydrogenation. Contrarily, typical ZnO crystal shows the green emission, which is attributed to the point defects or impurities. Hydrogen completely passivates this green emission, and in turn, enhances the band edge emission. The effect of hydrogen passivation disappeared by the annealing at temperatures higher than 600 °C. Hydrogen behaves more peculiarly in diamond. The hydrogenated diamond film shows the characteristic emission around 2.3 eV in photon energy. Since it disappears by oxidization treatment, this emission is attributed to hydrogen at the subsurface region. The detailed study indicated that hydrogen in diamond has bistable states.

Isotopically controlled silicon multilayer structures were used to measure the enhancement of self- and dopant diffusion in extrinsic boron doped silicon. 30Si was used as a tracer through a multilayer structure of alternating natural Si and enriched 28Si layers. Low energy, high resolution secondary ion mass spectrometry (SIMS) allowed for simultaneous measurement of self- and dopant diffusion profiles of samples annealed at temperatures between 850°C and 1100°C. A specially designed ion-implanted amorphous Si surface layer was used as a dopant source to suppress excess defects in the multilayer structure, thereby eliminating transient enhanced diffusion (TED) behavior. Self- and dopant diffusion coefficients, diffusion mechanisms, and native defect charge states were determined from computer-aided modeling, based on differential equations describing the diffusion processes. We present a quantitative description of B diffusion enhanced self-diffusion in silicon and conclude that the diffusion of both B and Si is mainly mediated by neutral and singly positively charged self-interstitials under p-type doping. No significant contribution of vacancies to either B or Si diffusion is observed.

Injection of point defects into GaAs/AlxGa1-xAs heterostructures by oxidation of AlxGa1-xAs is investigated. The blueshift of the PL emission from GaAs quantum wells is measured as a function of rapid thermal annealing conditions in as-grown and thermally oxidized samples. Contrary to published data for point defect injection by oxidation of GaAs, oxidation of AlxGa1-xAs appears to reduce the interdiffusion of the quantum wells. The PL peak shifts after annealing have been fit assuming Al diffusion in the quantum wells and solving the singleparticle Schrödinger equation. These fits show a reduction in group-III diffusivity of over an order of magnitude with an oxidized layer of AlxGa1-xAs on the surface.

In GaAs, (Si,H) complexes are efficiently dissociated at 300 K by photons with energies above 3.5 eV. Their optical cross-section is 10-19-10-18 cm2. This dissociation is the result of an electronic excitation of the Si-H bond of the complex from a bonding state to an antibonding state. (Si,H) and (S,H) complexes in AlGaAs alloys are also dissociated under UV illumination with optical cross-sections similar to GaAs. In passivated 2D AlGaAs-GaAs heterostructures, the evolution of the extra sheet carrier concentration at low photon densities presents a loss of free carriers attributed to the filling of surface states. In AlGaAs and in 2D AlGaAs-GaAs heterostructures, the replacement of hydrogen by deuterium in the complexes shows that the (Si,D) and (S,D) complexes are significantly more stable than the (Si,H) and (S,H) complexes as previously found in GaAs:Si,H.

Iron was introduced in InP by ion implantation with the aim of obtaining a high concentration of substitutional, electrically active, deep level impurities. A substrate temperature higher than 200 °C was maintained during implantation in order to reduce damage accumulation and Fe defect reactions. The lattice position of the implanted Fe atoms and its modification during annealing treatments was studied by means of Proton Induced X-ray Emission (PIXE) in channeling conditions and correlated with the ion induced damage measured by different techniques. The results show that a high fraction of substitutional Fe atoms is present after the implantation. This fraction is progressively reduced during thermal treatments by increasing the annealing temperature, with the formation of inactive Fe aggregates, probably in the form of small Fe-P complexes.

Features of a process of delamination of crystalline silicon layer from silicon wafer along hydrogen platelet layer formed by microwave plasma hydrogenation are described. The process involves first making a buried layer of nuclei for hydrogen platelets. Ion implantation of inert or low-soluble gases is used to form the layer. The nuclei are microbubbles that appear along Rp plane of implanted ions. Results for argon, xenon, and krypton implantation are compared. Wafers thus processed with a dose of 1015cm-2 are then hydrogenated with a microwave plasma. During hydrogenation, an atomic hydrogen diffuses into the silicon wafer and collects onto internal surfaces of the microbubbles. Then the hydrogen increases the internal surface of the microbubbles by growing a platelet type extensions to the microbubbles. The extensions grow preferably along the buried layer plane. A silicon layer above the layer of grown platelets were delaminated through pre-bonding/cut/post-bonding sequence as in the Smart-cut process. The plasma hydrogenation of the trap layer may be used as a step in a process of fabricating of SOI wafers with a very thin top crystalline silicon layer. Also, implant doses needed to form the microbubble trap layer are much lower than doses of direct implantation of hydrogen in the Smart-cut process. Temperature range of 200°C to 400°C during the hydrogenation process allows effectively grow extended hydrogen platelets from the nuclei. Mechanisms of nucleation of platelets as extentions of microbubbles are suggested. Control of hydrogen outdiffusion/platelet growth with thermal trajectory during plasma processing is discussed.