Next Generation High-Efficiency Low-cost Thin Film Photovoltaics

Our goal is to develop novel growth approaches for producing low cost solar cells that have efficiency comparable to those made from high-cost, single-crystalline materials. Our efforts to date focus on using ion beam assisted deposition (IBAD) to control the grain boundary alignment in polycrystalline silicon thin films. The boundaries between highly aligned grains have smaller defect densities, which lead to higher carrier mobilities and lifetimes. Our approach can be extended to other thin film PV systems. In this initial period, we have produced highly aligned Si films by using a biaxially textured template layer of CaF2. We have chosen CaF2 as a candidate material due to its close lattice match with silicon and its suitability as an ion beam assisted deposition (IBAD) material. We show that the IBAD produces biaxially textured CaF2 at a thickness of ~10 nm and, with the addition of an epitaxial CaF2 layer, has an in-plane texture of ~15°. Deposition of a subsequent layer of Si aligns on the template layer with an in-plane texture of 10.8°. The additional improvement of in-plane texture is similar to the behavior observed in more fully characterized IBAD materials systems. A germanium buffer layer is used to assist in the epitaxial deposition of Si on CaF2 template layers and single crystal substrates. These experiments confirm that an IBAD template can be used to biaxially orient polycrystalline Si. Anticipating improved control of the texture with the IBAD process, we also investigated the ultimate limit of seeded orientation of silicon by growing silicon on CaF2 that is epitaxially grown on single crystal yttria-stabilized zirconia (YSZ) with an extremely high degree of orientation. This approach yielded Si films with in-plane orientation of about 0.7°, showing that the heteroepitaxial growth of Si on highly oriented seeds will produce a very high degree of texture. Introduction The solar cell market is currently dominated by crystalline silicon modules due to their high efficiency and reliability. The cost of electricity produced by crystalline silicon panels however is still significantly higher than the cost of today’s grid power ($0.27/kW.hr vs. $0.06/kW.hr). As a result, in order to increase the deployment of solar power, there is a need for new technologies to both decrease the cost and increase the efficiency of photovoltaics (PVs). Thin-­‐film solar cells offer the opportunity to dramatically lower the price of solar energy by using small amounts of materials and low-­‐cost manufacturing technologies. Ultimately, the efficiency of inorganic thin film solar cells is fundamentally limited by the fact that the active layers are polycrystalline and therefore by definition contain defects. In particular, recombination of photogenerated carriers at grain-­‐boundaries is extremely detrimental to the efficiency of thin film solar cells.[1] Thus, a technology that drastically reduces the density of minority carrier recombination sites at the grain-­‐ boundaries of inorganic thin films would clearly be a step-­‐out innovation as it would allow thin film PVs to approach the performance of single crystal devices at a fraction of the cost. Such devices constitute the next generation thin film PVs. Here we propose to develop ion-­beam based processing strategies that reduce grain boundary misalignment in polycrystalline inorganic thin films. Reducing the relative misalignment of neighboring grains will lead to lower recombination rates at grain-­boundaries. We will thus create polycrystalline solar cells with power conversion efficiencies approaching those of their single crystal counterparts. In this exploratory proposal we concentrate on Si PVs because they constitute a well-­known and controlled system and because of their relevance in the marketplace. The process we will develop however is quite general and could be applied to many other polycrystalline thin films systems, such as CuInGaSe2 (CIGS) or Cu2ZnSnS4 (CZTS). Solar cell efficiency is a strong function of minority carrier lifetime, since photogenerated carriers that recombine before reaching the p-n junction do not contribute to photocurrent. Grain boundaries in polycrystalline silicon films provide electron traps that act as recombination centers that reduce minority carrier lifetimes [1]. This recombination is a function of the grain boundary structure. In particular, the high dislocation density of high angle grain boundaries result in higher recombination rate than low angle grain boundaries. Dimitriadis et al. showed that the effective carrier lifetime increases as the dislocation density decreases [2], and in an elegant experiment using electron beam induced current contrast ratios in polycrystalline silicon films, Seifert et al. showed that recombination is a strong function of grain boundary defect density [3]. Grain boundaries can be described as having both out-of-plane and in-plane misorientation known as tilt and twist, respectively. Both types of misorientation result in defect densities that lead to recombination. The degree of tilt and twist in a thin film grain boundary population reflects the crystallographic texture of the film. Biaxial texture, which has a preferred crystallographic direction for both out-of-plane and inplane directions, can decrease both twist and tilt misorientation between grains. One way to develop biaxial texture is application of an ion beam during the initial stages of nucleation of a thin film. This ion beam assisted deposition (IBAD) process uses a low energy (< 1keV), inert (Ar) ion beam to develop in-plane texture in a growing thin film during concurrent physical vapor deposition of the desired source material. The ion beam is aligned along a particular crystallographic direction at an oblique angle relative to the desired out-of-plane growth direction. The ion beam sputters away unfavorably oriented crystallites and allows favorably oriented crystallites to survive and grow. If the correct channeling angle is selected then bi-axial texture can be developed.