Efficient two-junction monolithic cascade solar cells have been fabricated in two dif-ferent III-V materials systems. The cells are grown by metalorganic chemical vapor deposi-tion (MOCVD). Two-junction efficiencies as high as 25% (1-sun, AM2) are projected from the individual GaInAs (1.15 eV) and A1GaAs (1.72 to 1.75 eV) subcell efficiencies obtained in a cascade configuration. Factors influencing actual cascade efficiency and reproducibility are discussed. Metal interconnection of the AlGaAs and GaInAs subcells during post-growth processing yields two-terminal cascades with power conversion efficiencies (n) as high as 13.6% at 1-sun, airmass two (AM2) , and 11.1% at 1-sun, airmass zero (AMO). Three-terminal AlGaAs (1.72 eV)/GaInAs (1.15 eV) cascades have shown 11 of 12.6% (1-sun, AMO). Alternatively, insertion of a novel high-conductance junction between the cascade subcells during the MOCVD growth of the structure reduces the complexity of processing to that required of a single-junction cell. Voltage addition has been demonstrated, with n of 10.0% (1-sun, AM2) in the AlGaAs (1.72 eV)/GaInAs (1.15 eV) system. The A1GaAs (1.72 eV)/GaAs lattice-matched pair allows demonstration of proof-of-concept structures; tunnel-junction-connected A1GaAs (1.72 eV)/ GaAs cascades show n as high as 13.9% at 1-sun, AM2, and three-terminal, two-junction A1GaAs (1.72 eV)/GaAs (1.42 eV) systems as high as 23.0% at 1-sun, AM2.

This paper describes experience gained in developing manufacturing procedures for high efficiency gallium arsenide solar cells, to be used to provide electrical power for satellites. Details are given of the manufacturing methods selected, development of support infrastructure, and results achieved.

High efficiency is the key to large scale applicability of photovoltaic systems. Detailed cost analysis done by Electric Power Research Institute (EPRI) indicates that greater than 15% efficient modules will be required at a cost of less than 50c/watt in order for PV to compete with conventional energy sources. There are various approaches being investigated for a commercially viable PV product including low cost, low efficiency thin film approach using amorphous silicon and CnInSe2, and high cost, high efficiency approach using GaAs material and sophisticated tandem and concentrator cell designs. Silicon solar cells, which are the focus of this paper, provide an intermediate path for achieving the cost and efficiency goals. It has been more than thirty years since silicon solar cells were first discovered, but the drive for high efficiency has only recently become intense due to the balance of system (BOS) cost. Figure 1 shows a road map of silicon cell efficiency improvement since 1950, when the cell efficiency was only 5%. During the period 1960-1975, only a small improvement in cell efficiency was realized because major emphasis was on radiation hardness and space applications. Drive for terrestrial applications started in 1975 with the initial emphasis on low cost, therefore, only a moderate improvement in cell efficiency was observed during 1975-1980. In the early 1980s, when the importance of balance of system cost was recognized, the emphasis shifted to high efficiency and since then the progress in silicon cell efficiency has been phenomenal (Figure 1). Five years ago, 20% efficient cells seemed unattainable, but today we have already seen 22% (AM1) efficient silicon cells. Major advances in cell efficiency in the 1980s came from the emphasis on improving V c and optical design of I silicon cells. Improved material quality also contributed a ot to the current rapid progress. The purpose of this paper is to review and summarize some of those approaches that resulted in AM1 silicon cell efficiencies in the range of 18-22%.

Recent work with high-efficiency solar cells has led to improvements in optical efficiency to the point where short-circuit currents of 40.8 mA/cm2 have been measured under terrestrial sunlight in 380 micron thick cells; this is over 90% of the theoretical maximum for silicon. Three parts of the design of these cells contribute to the high currents: First, careful grid design, surface texture, and a double-layer antireflection coating reduce the reflection loss to approximately 4% over the entire solar spectrum. Second, the use of high-purity float-zone silicon, coupled with an effective back-surface-field structure and an aluminum back surface reflector, is responsible for the efficient collection of carriers generated by infrared light, which can penetrate through the entire cell. Internal quantum efficiencies of .85 at 1050 nm and .56 at 1100 nm have been achieved. Finally, the use of a thin emitter, the optimization of the emitter doping level and the passivation of the front surface have allowed internal quantum efficiencies exceeding 0.95 over the entire visible range. The possibility of achieving quantum efficiencies greater than 1 for ultraviolet light is also discussed.

A new type of silicon concentrator solar cell has been developed. It is called the point-contact cell because the metal semiconductor contacts are restricted to an array of small points on the back of the cell. The point contact cell has recently demonstrated 22 percent conversion efficiency at one sun and 27.5 percent at 100 suns.

The value of a passive, maintenance-free, renewable energy source was immediately recognized in the early days of the space program, and the silicon solar cell, despite its infancy, was quickly pressed into service. Efficiencies of those early space solar arrays were low, and lifetimes shorter than hoped for, but within a decade significant advances had been made in both areas. Better performance was achieved because of a variety of factors, ranging from improvements in silicon single crystal material, to better device designs, to a better understanding of the factors that affect the performance of a solar cell in space. Chief among the latter, particularly for the mid-to-high altitude (HEO) ana geosynchronous (GEO) orbits, are the effects of the naturally occurring particulate radiation environment. Although not as broadly important to the photovoltaic community at large as increased efficiency, the topic of radiation damage is critically important to use of solar cells in space, and is a major component of the NASA research program in space photo-voltaics. This paper will give a brief overview of some of the opportunities and challenges for space photovoltaic applications, and will discuss some of the current research directed at achieving high efficiency and controlling the effects of radiation damage in space solar cells.

A high-throughput full-automated photovoltaic cell/module manufacturing plant has been constructed in Chitose, Japan. The production line consists of a solar cell fabrication line and a module assembly line, both controlled by computer. In the cell line, cells are transferred by conveyor during the fabrication processes such as formation of PN and BSF junction by ion-implantation, lamp anneal, metallization and quality checks except for wet processes which include chemical etching. In the module line, two free-flow conveyors are used for sequence assembly of the modules. For tab connection and cell arrangement, robots are used. Modules are assembled with super-straight type lamination and framing processes, and then stocked in the stocker with final electrical checks including high voltage. The plant is completely operated by computer control. Sequencers set at each equipment in the process are controlled by four medium computers, followed by overall central computer control. Using the situation data in each equipment, the plant condition is strictly controlled on real time. The plant is entirely under the control of two operators at the factory central control desk.

To demonstrate technical viability of photovoltaic modules in central, grid connected energy systems, ARCO Solar, Inc. has designed, installed and is operating two photovoltaic power plants on the megawatt scale. These systems use two-axis tracking. The first generation plant in Lugo (Hesperia), California, with a nominal rating of one MWpk (DC)" was installed in 1982 in the Southern California Edison Company grid. The second system, rated at 6.4 MWDk (DC), is located in the Carrisa Plain in California and connected to the Pacific Gas and Electric Company grid. Based on the cost and performance data from these installations, an assessment of the current status and future needs of large scale photovoltaic energy systems is made. With each new system, improved techniques of design, installation and system integration have been developed. Expectations have been confirmed as to the performance and adaptability of solar cells, especially the ease of incremental increases in capacity when needed. Modular photovoltaic systems have been found to be easy to build and operate, and to be highly reliable. Prologue: Technological advancement usually requires good science and logical engineering. In the main, faith, persistence and feel are also required. Rule: The balance-of-system costs for photovoltaic energy systems equal photovoltaic module costs. Photovoltaic systems have progressed to their current stage of high promise because of faith, persistence, feel and belief in this rule.

As part of the U. S. National Photovoltaic Program, the Jet Propulsion Laboratory's (JPL's) Flat-Plate Solar Array (FSA) Project has conducted a comprehensive 10-year research activity addressed to understanding the reliability attributes of terrestrial flat-plate photovoltaic modules, and to developing the technology required to achieve 30-year life (1). This paper provides an overview of the reliability issues and progress, and highlights the design, analysis, and test tools generated to achieve the high levels of reliability necessary for future large-scale terrestrial applications. Much of the technology is also directly applicable to the large space arrays currently under consideration for use in extraterrestrial applications.

The Amorphous Silicon Research Project (ASRP) was established at the Solar Energy Research Institute (SERI) in 1983 and is responsible for coordinating all U.S. Department of Energy (DOE) research activities in amorphous silicon photovoltaics. The strategy, objectives, and research directions of the project have been established by a five-year research plan, which was published in 1984. The technical plan of the ASRP is organized into two principal activities: multidisciplinary activities and fundamental research activities. Near-term objectives (early 1987) for single-junction cells and submodules are efficiencies of 12% (for an area of 1 cm2) and 8% (for 1000 cm2), respectively; for multi junction cells, the near-term goal is an efficiency of 13% (for 1 cm2).

The Solarex Thin Film Division started commercial production of amorphous silicon solar cells for consumer applications in February 1984. Since that time, Solarex has produced more than twenty different types of amorphous silicon solar modules with sizes varying from 0.5cm2 to 1000cm2. The modules contain from 2 to 30 cells connected in series by sequentially patterning the various layers in the solar-cell structure. Conversion efficiencies as high as 10.9% have been obtained with superlattice p layers in 1.16cm2 p-i-n and efficiencies up to 8.7% have been obtained in large area modules (active area = 864cm2).

In this paper, we give some details of the first continuous multichamber in-line system for the production of inexpensive Amorphous Silicon solar cells. GSI's innovative design ensures maximum production rates and performance from each panel; the versatility in design allows new device concepts to be incorporated and prevents obsolescence.

A process for fabricating a-Si photovoltaic devices utilizing continuous deposition onto a plastic substrate has been developed. The advantages that spurred the development of this process include cost, ease of handling, shipping weight, steady state processing conditions, and the feedback control available in continuous processing. The process is easily scalable. Questions remain on the effect of plastic substrates on ultimate efficiency and stability.

Density of midgap defect states (DOS) in amorphous silicon (a-Si:H) is a fundamental material parameter in determining the transport properties. It is shown in this paper that the commonly believed model DOS, consisting of one major mid-gap defect, namely a Si dangling bond, is probably not correct, and is logically inconsistent with the position of the Fermi level. The dangling-bond DOS is also inconsistent with data from space-charge-limited-current, with experimentally determined device field profiles, and with some data on photo-degradation. In contrast, we propose that a DOS model proposed by Adler, based on negative corelation energy and consisting of primarily doubly charged and empty dangling bonds (T.- and Ta÷ states), is a more accurate representation of actual DOS in undoped, high quality a-Si:H. This model is shown to be consistent with all the experimental data on material and device properties, including photo-degradation and doping.

There has been what has appeared to be an insurmountable barrier holding back the progress of amorphous photovoltaics. Conventional amorphous photovoltaic cells, despite claims of initial efficiency, lack stability to the point of reducing their efficiency to no more than 5-6%. We will describe how true efficiencies of 13% and over are now practical, showing virtually no degradation. We have now developed materials that can bring efficiencies to 20%. Equally important, by the use of a proprietary continuous web production process, multilayered cells are produced with very high yields. Based upon several years of production experience, we can now realistically discuss how amorphous photovoltaics can become competitive to conventional fuels.

The research and development of amorphous silicon solar cells at Kanegafuchi Chemical Industry Co. Ltd.(Kaneka) are introduced. In 1981, a wide gap a-SiC and a-SiC/a-Si hetero-junction solar cells were developed through joint research with Osaka University. Then flexible amorphous solar cell utilized the a-SiC/a-Si cell were developed and industrialized for the consumer market in 1983. To reduce the cost, a large capacity of deposition system were developed and built at Sakamoto plant in 1985. As to the instability of a-Si solar cell, a new and stable tandem cell has been developed which shows noble stability under both sun light and thermal conditions.

Thin film tandem devices are progressing toward required performance levels for large scale power plant use. Two different approaches, 4-terminal thin-film-silicon-hydrogen/copper indium diselenide hybrid, and thin-film-silicon-hydrogen/thin-film-silicon-germanium hydrogen series tandems, are emerging as the leading contenders. Each of these employs thin-film-silicon-hydrogen as the top device, and each has different technological problems limiting performance. These differences are explored and contrasted. Particular emphasis is placed upon spectral mismatch issues using outdoor data from representative modules. Evaluation in terms of energy delivery rather than peak efficiency is discussed.