Solar Cells最新文献

The microstructure and morphology of polycrystalline thin film CuInSe₂ were studied extensively in the compositional range 17–32 at.% Cu. The grain size varied with substrate temperature, copper content, and in variable ways with substrate type, and ranged in size from 0.1 to 5.0 μm. The morphology of copper-rich films appeared additionally to depend on the resident nucleation and growth of the Cu_2−δSe binary compound. A microstructural model of polycrystalline thin film CuInSe₂ is presented and suggests that the intergranular microstructure is dominated by the compositional and substrate temperature dependence of Cu_2−δSe precipitation at grain boundaries and free surfaces. The intragranular microstructure of the near-stoichiometric grain is a phase-separated mixture of ordered chalcopyrite and disordered sphalerite, with $\text{Cu}{}_{\mathrm{x}}\text{Se}\text{(x=0.5, 1.0, 1.5, 2.0)}$ minority phase inclusions. Off-stoichiometric copper-poor film compositions additionally contain isolated grains of the chalcopyrite-variant ordered-vacancy compound CuIn₂Se_3.5. The potential ramifications of the microstructure on the device performance include a reduction in the photo-active volume, carrier transport across phase boundaries, and dependence of transport parameters on the crystallite size.

Environmental testing data are presented and discussed in relation to the suitability of CuInSe₂ (CIS) as a durable photovoltaic material and to the Interim Qualification Tests and Procedures for Terrestrial Photovoltaic Thin-Film Flat-Plate Modules (IQTP). Groups of modules having no significant change after ten humidity-freeze cycles are reported. Heating during module packaging or during environmental testing to temperatures above those normally encountered by modules in outdoor service may introduce a temporary power loss; the power recovers with time to near the initial power. Data indicate that temperature alone, rather than temperature combined with humidity, causes the temporary power loss and that CIS is not inherently sensitive to humidity. Hermetic seals are not in general necessary for CIS materials. The IQTP may improperly indicate poor performance if the temporary power loss is not considered in electrical performance testing between different sections of the environmental test procedures and at the end of all environmental tests. Data are not available to validate accelerated testing as a means of predicting long-term in-service performance; however, correlations between outdoor and accelerated testing are seen.

Research on CuInSe₂ and CdTe thin film solar cells is discussed. CuInSe₂ was deposited by selenization of Cu/In layers and was used to make a 10% efficient CuInSe₂/(CdZn)S cell. Characterization of the reaction mechanisms is described. The open-circuit voltage $\text{V}{}_{\mathrm{<mtext>oc</mtext>}}$ of CuInSe₂/(CdZn)S cells is dominated by recombination in the space charge region, so increasing the band gap or decreasing the width of this region should increase $\text{V}{}_{\mathrm{<mtext>oc</mtext>}}$. Increasing the band gap with a thin Cu(InGa)Se₂ layer at the CuInSe₂ surface has demonstrated increased $\text{V}{}_{\mathrm{<mtext>oc</mtext>}}$ with collection out to the CuInSe₂ band gap. A post-deposition treatment and contacting process for evaporated CdS/CdTe cells was developed and high efficiency cells were made. Several steps in the process, including a CdCl₂ coating, a 400 °C heat treatment, and a contact containing copper are critical. ZnTe films were deposited from an aqueous solution as a contact to CdTe.

Large-area hydrogenated amorphous silicon solar cells with a two-stacked p-i-n junction tandem structure are practical solar cells with high conversion efficiency and reliability. We attained a total-area efficiency of 10.05% for a $\text{30}\text{cm}\text{× 40}\text{cm}$ tandem submodule. We found that the initial degradation of these tandem cells was about 15% on the initial value. Using these results, we discuss the feasibility of a stable 10% efficient a-Si photovoltaic module.

Computer simulations of two-junction, concentrator tandem solar cell performance show that IR-sensitive bottom cells are required to achieve high efficiencies. Based on this conclusion, two novel concentrator tandem designs are under investigation: (1) a mechanically stacked, four-terminal GaAs/GaInAsP (0.95 eV) tandem, and (2) a monolithic, lattice-matched, three-terminal InP/GaInAs tandem. In preliminary experiments, terrestrial concentrator efficiencies exceeding 30% have been achieved with each of the above tandem designs. Methods for improving the efficiency of each tandem type are discussed.

The theoretical performances of ideal single- and multijunction cells are compared at 100 × concentration under a range of cloudless-sky conditions. The sensitivities of device performance to cell temperature and spectral variations are shown to depend on the number of junctions (one, two or three), the way in which the junctions are connected (series, parallel or independent), and the band gaps of the devices. The average performances of all of the multijunction devices surpass that of a single-junction GaAs device, but the inconsistency in performance of some of the multijunction devices is significant for large variations in cell temperature and incident spectrum. The choice of band gap and connection scheme is more important than the number of junctions in determining the consistency of device performance.

At a sufficiently high photon flux the density of light-induced defects in amorphous hydrogenated silicon (a-Si:H) can be made to saturate within a few hours. The saturated defect density $\text{N}{}_{\mathrm{<mtext>sat</mtext>}}$ is independent of light intensity, of temperature and of further illumination over a wide range of conditions. Therefore, $\text{N}{}_{\mathrm{<mtext>sat</mtext>}}$ can serve as a robust criterion for the rapid evaluation of the stability of a-Si:H. In this paper we show that $\text{N}{}_{\mathrm{<mtext>sat</mtext>}}$ is correlated with the rate of defect buildup during light-soaking, so that the entire defect history and the useful life of a particular sample may be inferred from $\text{N}{}_{\mathrm{<mtext>sat</mtext>}}$. So far, we have measured values of $\text{N}{}_{\mathrm{<mtext>sat</mtext>}}$ between 4 × 10¹⁶ and 2 × 10¹⁷ cm⁻³ in electronic-grade a-Si:H. $\text{N}{}_{\mathrm{<mtext>sat</mtext>}}$, and thus the rate of defect buildup, rises with the value of the band gap and with the hydrogen content; it drops with increasing deposition temperature. $\text{N}{}_{\mathrm{<mtext>sat</mtext>}}$ is not correlated with either the initial defect density or the Urbach energy. We demonstrate an application of the correlation between the saturated defect density and the optical gap to predict the long-term performance of solar cells.

This paper gives an overview of the fiscal year 1990 research activities and results under the Solar Radiation Research Task of the Photovoltaic Advanced Research and Development Project at the Solar Energy Research Institute. The activities under this task include developing and applying measurement techniques, instrumentation, and data analysis and modeling to understand and quantify the response of photovoltaic devices to variations in broad-band and spectral solar radiation.

High-density, non-porous, highly photoconductive films of amorphous hydrogenated germanium (a-Ge:H) showing minimal microstructure were prepared using the r.f. glow discharge method out of a gas plasma of GeH₄ and H₂. These films, deposited onto substrates mounted on the powered electrode of a diode reactor, showed remarkable improvement over codeposited material taken from the unpowered electrode. Films were also prepared under the systematic variation of substrate temperature, discharge power, and dilution of the plasma by H₂. For the reactor geometry used, dilution of the plasma is found to be essential to the preparation of high-quality a-Ge:H. The conditions for the preparation of optimized a-Ge:H material were quite different from those found to produce optimized a-Si:H in this reactor. We assert this to be the principal cause of the finding of inferior properties for a-SiGe:H alloys when compared with a-Si:H.