Publications

Abstract not provided.

The Sandia Photovoltaic Program conducted research in crystalline-silicon solar cells between 1986 and 2000 for the U.S. Department of Energy. This period saw rapid improvements in the fundamental understanding of c-Si materials and devices, improvements in c-Si PV manufacturing and control, and a rapid expansion of c-Si PV manufacturing capacity. Crystalline-silicon technology has provided the basis for PV to emerge as a serious option for global energy needs. The c-Si cell research at Sandia examined c-Si materials, devices, processing, and process integration. This report summarizes research conducted in this program over the past 15 years.

Abstract not provided.

The availability of polysilicon feedstock has become a major issue for the photovoltaic (PV) industry in recent years. Most of the current polysilicon feedstock is derived from rejected material from the semiconductor industry. However, the reject material can become scarce and more expensive during periods of expansion in the integrated-circuit industry. Continued rapid expansion of the PV crystalline-silicon industry will eventually require a dedicated supply of polysilicon feedstock to produce solar cells at lower costs. The photovoltaic industry can accept a lower purity polysilicon feedstock (solar-grade) compared to the semiconductor industry. The purity requirements and potential production techniques for solar-grade polysilicon have been reviewed. One interesting process from previous research involves reactive gas blowing of the molten silicon charge. As an example, Dosaj et all reported a reduction of metal and boron impurities from silicon melts using reactive gas blowing with 0{sub 2} and Cl{sub 2}. The same authors later reassessed their data and the literature, and concluded that Cl{sub 2}and 0{sub 2}/Cl{sub 2} gas blowing are only effective for removing Al, Ca, and Mg from the silicon melt. Researchers from Kawasaki Steel Corp. reported removal of B and C from silicon melts using reactive gas blowing with an 0{sub 2}/Ar plasma torch. Processes that purify the silicon melt are believed to be potentially much lower cost compared to present production methods that purify gas species.

Sub-wavelength periodic texturing (gratings) of crystalline-silicon (c-Si) surfaces for solar cell applications can be designed for maximizing optical absorption in thin c-Si films. We have investigated c-Si grating structures using rigorous modeling, hemispherical reflectance, and internal quantum efficiency measurements. Model calculations predict almost ∼ 100 % energy coupling into obliquely propagating diffraction orders. By fabrication and optical characterization of a wide range of ID & 2D c-Si grating structures, we have achieved broadband, low (∼ 5 %) reflectance without an anti-reflection film. By integrating grating structures into conventional solar cell designs, we have demonstrated short-circuit current density enhancements of 3.4 and 4.1 mA/cm2 for rectangular and triangular 1D grating structures compared to planar controls. The effective path length enhancements due to these gratings were 2.2 and 1.7, respectively. Optimized 2D gratings are expected to have even better performance.

InGaAsN alloys are a promising material for increasing the efficiency of multi-junction solar cells now used for satellite power systems. However, the growth of these dilute N containing alloys has been challenging with further improvements in material quality needed before the solar cell higher efficiencies are realized. Nitrogen/V ratios exceeding 0.981 resulted in lower N incorporation and poor surface morphologies. The growth rate was found to depend on not only the total group III transport for a fixed N/V ratio but also on the N/V ratio. Carbon tetrachloride and dimethylzinc were effective for p-type doping. Disilane was not an effective n-type dopant while SiCl4 did result in n-type material but only a narrow range of electron concentrations (2-5e17cm{sup -3}) were achieved.

Crystalline silicon continues to be the dominant semiconductor material used for terrestrial photovoltaics. This paper discusses the scientific issues associated with silicon photovoltaics processing, and cell design that may yield cell and module performance improvements that are both evolutionary and revolutionary in nature. We first survey critical issues in ''thick'' crystalline silicon photovoltaics, including novel separations processes for impurity removal, impurity and defect fundamentals, interface passivation, the role of hydrogen. Second, we outline emerging opportunities for creation of a very different ''thin-layer'' silicon cell structure, including the scientific issues and engineering challenges associated with thin-layer silicon processing and cell design.

Surface texturing of Si to enhance absorption particularly in the IR spectral region has been extensively investigated. Previous research chiefly examined approaches based on geometrical optics. These surface textures typically consist of pyramids with dimensions much larger than optical wavelengths. We have investigated a physical optics approach that relies on surface texture features comparable to, or smaller than, the optical wavelengths inside the semiconductor material. Light interaction at this are strongly dependent on incident polarization and surface profile. Nanoscale textures can be tuned for either narrow band, or broad band absorptive behavior. Lowest broadband reflection has been observed for triangular profiles with linewidths significantly less than 100 nm. Si nanostructures have been integrated into large ({approximately}42 cm{sup 2}) area solar cells, Internal quantum efficiency measurements in comparison with polished and conventionally textured cells show lower efficiency in the UV-visible (350-680 mu), but significantly higher IR (700-1200 nm) efficiency.

Application of nanotechnology and advanced optical structures offer new possibilities for improved radiation tolerance in silicon solar cells. We describe the application of subwavelength diffractive structures to enhance optical absorption near the surface, and thereby improve the radiation tolerance.

This paper summarizes recent progress in the development of back-contact crystalline-silicon (c-Si) solar cells and modules at Sandia National Laboratories. Back-contact cells have potentially improved efficiencies through the elimination of grid obscuration and allow for significant simplifications in the module assembly process. Optimization of the process sequence has improved the efficiency of our back-contact cell (emitter wrap through) from around 12% to near 17% in the past 12 months. In addition, recent theoretical work has elucidated the device physics of emitter wrap-through cells. Finally, improvements in the assembly processing back-contact cells are described.

Boron-doped back-surface fields (BSF`s) have potentially superior performance compared to aluminum-doped BSF`s due to the higher solid solubility of boron compared to aluminum. However, conventional boron diffusions require a long, high temperature step that is both costly and incompatible with many photovoltaic-grade crystalline-silicon materials. We examined a process that uses a relatively low-temperature aluminum-alloy process to obtain a boron-doped BSF by doping the aluminum with boron. In agreement with theoretical expectations, we found that thicker aluminum layers and higher boron doping levels improved the performance of aluminum-alloyed BSF`s.

Multicrystalline-silicon (mc-Si) materials and cells feature large areal variations in material and junction quality. The regions with poor device quality have been predicted to have more recombination current at forward bias than a simple area-weighted average due to the parallel interconnection of the good and bad regions by the front junction. The authors have examined the effect of gettering on areal inhomogeneities in large-area mc-Si cells. Cells with large areal inhomogeneities were found to have increased non-ideal recombination current, which is in line with theoretical predictions. Phosphorus-diffusion and aluminum-alloy gettering of mc-Si was found to reduce the areal inhomogeneities and improve large-area mc-Si device performance.

This paper summarizes the U.S. Department of Energy R&D program in crystalline-silicon photovoltaic technology, which is jointly managed by Sandia National Laboratories and National Renewable Energy Laboratory. This program features a balance of basic an d applied R&D, and of university, industry, and national laboratory R&D. The goal of the crystalline-silicon R&D program is to accelerate the commercial growth of crystalline-silicon photovoltaic technology, and four strategic objectives were identified to address this program goal. Technical progress towards meeting these objectives is reviewed.

Phosphorus diffusions are used in the fabrication process for nearly all crystalline-silicon (c-Si) photovoltaic solar cells to form the emitter of the solar cell. These phosphorous diffusions are also well known to have beneficial gettering benefits, i.e., deleterious metallic impurities are gettered from the bulk of the c-Si substrate into the phosphorous doped layer. In this study, we examined the effect of oxidations performed after the phosphorus diffusion. We were particularly interested in using the oxidation to passivate the surface of the phosphorus diffusion. Post-diffusion oxidations or moderate temperature steps in oxidizing ambients are also commonly found in commercial fabrication sequences of c-Si solar cells. we found that the bulk lifetime was degraded in Czochralski (Cz) silicon due to the post-diffusion oxidation unless there was a gettering agent present during the oxidation. Possible explanations for these results are presented at the end of the paper.

Fabrication of high-efficiency silicon solar cells in an industrial environment requires a different optimization than in a laboratory environment. Strategies are presented for process development of high-efficiency silicon solar cells, with a goal of simplifying technology transfer into an industrial setting. The strategies emphasize the use of statistical experimental design for process optimization, and the use of baseline processes and cells for process monitoring and quality control. 8 refs.

Fabrication of high-efficiency silicon solar cells requires processing technology capable of maintaining long bulk carrier lifetime and low surface recombination. Development of long-lifetime processing techniques using experimental designs based on statistical methods is described. The first three experiments investigated pre-oxidation cleans, phosphorus gettering, and a comparison of different phosphorus diffusion sources. Optimal processing parameters were found to depend on type of silicon material. 2 refs., 2 figs., 2 tabs.

Photovoltaic concentrator collectors are an attractive option for utility-scale photovoltaic power plants. This paper reviews the current status of photovoltaic concentrator collector and cell development. Included in the review is a discussion of the economic motivation for concentrators, a summary of recent concentrator collector and cell development, and a description of a major new program to accelerate development and commercial introduction of concentrator collectors. 21 refs., 1 fig., 3 tabs.

An AlGaAs window layer is used in high-efficiency GaAs solar cells to reduce carrier recombination at the front surface. Free surfaces of III-V semiconductors have a high density of surface states that serve as recombination sites and create a depletion region at the front surface. We have performed a theoretical investigation of front-surface recombination that includes the effect of a surface space-charge layer. It was found that the surface space-charge layer can have a profound effect on front-surface recombination for thin or lightly doped window layers. 15 refs., 5 figs., 1 tab.