The structural and optical properties of erbium-doped silicon-rich silica samples containing 12 at. % of excess silicon and 0.63 at. % of erbium are studied as a function of annealing temperature in the range 600-1200°C. Indirect excitation of Er 3+ ions is shown to be present for all annealing temperatures, including annealing temperatures well below 1000°C for which no silicon nanocrystals are observed. Two distinct efficient ͑ tr Ͼ 60% ͒ transfer mechanisms responsible for Er 3+ excitation are identified: a fast transfer process ͑ tr Ͻ 80 ns͒ involving isolated luminescence centers ͑LCs͒, and a slow transfer process ͑ tr ϳ 4 -100 s͒ involving excitation by quantum confined excitons inside Si nanocrystals. The LC-mediated excitation is shown to be the dominant excitation mechanism for all annealing temperatures. The presence of a LC-mediated excitation process is deduced from the observation of an annealing-temperature-independent Er 3+ excitation rate, a strong similarity between the LC and Er 3+ excitation spectra, as well as an excellent correspondence between the observed LC-related emission intensity and the derived Er 3+ excitation density for annealing temperatures in the range of 600-1000°C. The proposed interpretation provides an alternative explanation for several observations existing in the literature.

Cu2ZnSnS4 (CZTS) is an alternative material to Cu(In,Ga)Se2 (CIGSe) for use in thin film photovoltaic absorber layers composed solely of commodity elements [1,2]. Thus, if similar material quality and performance can be realized, its use would allow scale-up of terrestrial thin film photovoltaic production unhindered by material price or supply constraints. Here we report on our research on the deposition of CZTS by RF sputtering from a single CZTS target and co-sputtering from multiple binary sources on Mo-coated glass. We find some samples delaminate during post-sputtering furnace annealing in S vapor. Samples on borosilicate glass (BSG) delaminate much more frequently than those on soda-lime glass (SLG). We investigate the influences of the formation of frangible phases such as MoS2 at the CZTS/Mo interface and residual and thermal mismatch stress on delamination. We implicate fracture in a layer of MoS2 as the mechanism of delamination between the Mo and CZTS layers using scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy (XPS). Wafer curvature measurements show significant (˜400 MPa) deposition stress for minimally optimized Mo deposition; however nearly stress-free Mo layers with good adhesion can be deposited using a multi-step Mo deposition recipe. Co-sputtering CZTS adds 100 MPa of stress on both BSG and SLG, however delamination is nearly absent for samples deposited on low-stress Mo layers. We investigate metallic diffusion barrier layers to prevent the formation of MoS2 at the interface. Lastly we discuss the importance of removing Mo oxide by sputter etching before CZTS deposition and its effects on adhesion and series resistance.

In Scanning Transmission Electron Microscopy (STEM) the High-Angle Annular Dark-Field (HAADF) signal increases with atomic number and sample thickness, while dynamic scattering effects and sample orientation have little influence on the contrast. The sensitivity of the HAADF detector for a FEI F30 transmission electron microscope has been calibrated. Additionally, a nearly linear relationship of the HAADF signal with the incident electron current is confirmed. Cross sections of multilayered samples for contrast calibration were obtained by focused ion-beam (FIB) preparation. These cross sections contained several layers with known composition. A database with several pure elements and compounds has been compiled, containing experimental data on the fraction of electrons scattered onto the HAADF detector for each nanometer of sample thickness. Contrast simulations are based on the multi-slice formalism and confirm the differences in HAADF-scattering contrast for the elements and compounds. TEM offers high lateral resolution, but contains little or no information on the thickness of samples. Thickness maps in energy-filtered transmission electron microscopy, convergent-beam electron diffraction and tilt series are so far the only methods to determine thicknesses of particles in a transmission electron microscope. We show that the calibrated HAADF contrast can be used to determine the thicknesses of individual nanoparticles deposited on carbon films. With this information the volumes of nanoparticles with known composition were determined.

Cu2ZnSnS4 (CZTS) is a promising alternative for Cu(In,Ga)Se2 (CIGS) absorber layers in thin film solar cells and is comprised of commodity elements which will enable scale-up of chalcopyrite panel production unhindered by elemental supplies and costs. Various CZTS synthesis methods, especially sulfurization of stacked metal or metal sulfide layers, are being studied and have led to cell efficiencies up to 6.7% [1]. Here we report our studies of CZTS thin film synthesis via room temperature sputtering from a single CZTS target and co-sputtering from Cu2S, ZnS and SnS2 binary targets, both followed by sulfurization between 500 C - 600 C using either elemental sulfur vapor or in-situ generated H2S. Sputtering from sulfur-containing targets is designed to increase the sulfur content in the precursor films to promote stoichiometry. We report on the effects of processing including deposition on soda-lime and borosilicate glasses and deposition of Na-containing layers on film morphology (AFM/SEM), composition (EDS), phase (XRD), grain size (XRD/EBSD), grain boundary structure (EBSD), optical (spectroscopic ellipsometry) and electrical properties. Processing conditions producing desirable Zn-rich/Cu-poor films are identified [1]. The formation of MoSe2 at Mo/CIGS interface is believed to promote Ohmic contacts, but in CZTS we associate excessive formation of frangible MoS2 with film delamination from Mo/borosilicate glass substrates. Strategies for preventing delamination including adhesion layers are investigated and discussed. P-N junctions are formed with CdS/ZnO using chemical bath deposition and sputtering, and I-V characteristics are reported. Schottky junctions are formed and C-V measurements are used to determine the doping in the CZTS absorber layers.[1] H. Katagiri, et al., MRS Symp. Proc. 1165 1165-M04-01 (2009).