SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Si reduction, purification and sheet generation work has been concentrated on gaining information about a reduction process combined with purification (higher purity arc furnace with gas blowing and gradient freezing), transport process with purification and polycrystal sheet growth potential (SiF2), plastic deformation for sheet generation, and float zone sheet recrystallization.
Ion implantation and thermal annealing techniques for processing junctions and back surface layers in solar cells are discussed. Standard 10 keV (31)p(+) junction implants and 25 keV (11)B(+) back surface implants in combination with three-step furnace annealing are used for processing a range of silicon materials and device structures. Cells with efficiencies up to 16.5% AM1 are being produced, and large-area terrestrial cells with implanted junctions and back fields being fabricated in pilot production exhibit average efficiencies in excess of 15% AM1. Thermal annealing methods for removal of the radiation damage caused by implantation should be replaced by transient processing techniques in future production. Design studies have been completed for solar cell processing implanters to support 10 MW/yr and 100 MW/yr production lines, and analyses indicate that implantation costs can be reduced to approximately 1 cent/watt.
High-risk, high-payoff improvements to a baseline process sequence of simultaneous junction formation of silicon solar cells are discussed. The feasibility of simultaneously forming front and back junctions of solar cells using liquid dopants on dendritic web silicon was studied. Simultaneous diffusion was compared to sequential diffusion. A belt furnace for the diffusion process was tested.
Single wall carbon nanotubes used or this study were synthesized using the pulse laser vaporization technique. The as-produced SWNTs were collected from the condensed region on the quartz tube outside the furnace and purified by modification of the previously reported procedure. Preparation of the SWNT-P3OT composite solutions was performed using a series of mixing and sonication steps as previously demonstrated for other SWNT-polymer systems. Device fabrication involved the use of commercially-obtained, high quality (i.e., less than 10 W/sq.) ITO-coated polyethylene terapthalate (PET) substrates. Initially, an intrinsic layer of pristine P3OT is spray deposited (approx. 1-2 mL of the 15 mg/mL solution) onto the masked, 1 sq in. active area of the substrate. This is followed with spray deposition of the SWNT-P3OT composite solutions at similar volumes. As seen completion of the solar cell occurs when aluminum contacts (typical thicknesses of l000A) are applied to the IT0 and SWNT-P3OT composite film layers. The homogeneous distribution of SWNTs in a polymer matrix is dependent upon the ability of the polymer chain to associate with the SWNT superstructure. Stable composite dispersions of 0.1 % and 1.0% w/w SWNTs in P3OT were produced and analyzed in this study. Optical spectroscopy was conducted to observe the SWNT doping level effects on the absorption properties and infer potential electronic interactions between dopant and polymer. Shown is an overlay of these spectra where the pristine P3OT shows strong optical absorption at energies less than 2 eV. As expected, variation in absorption properties for the SWNT-P3OT composites is observed as the doping level increases. Through modification of the SWNT doping level in the polymer, it is possible to alter the absorption pattern of these composite materials. In fact, at these relatively low doping levels, the P3OT shows a significant enhancement in absorption through the near-IR and visible regions. The gray curve for purified SWNTs is offset from the other three for clarity, but indicates the typical complex pattern observed for this spectral range.
A process of metallurgical grade (MG) silicon production is presented which appears particularly suitable for photovoltaic (PV) applications. The MG silicon is prepared in a 240 KVA, three electrode submerged arc furnace, starting from high grade quartz and high purity silicon carbide. The silicon smelted from the arc furnace was shown to be sufficiently pure to be directionally solidified to 10 to 15 kg. After grinding and acid leaching, had a material yield larger than 90%. With a MG silicon feedstock containing 3 ppmw B, 290 ppmw Fe, 190 ppmw Ti, and 170 ppmw Al, blended with 50% of off grade electronic grade (EG) silicon to reconduct the boron content to a concentration acceptable for solar cell fabrication, the 99% of deep level impurities were concentrated in the last 5% of the ingot. Quite remarkably this material has OCV values higher tham 540 mV and no appreciable shorts due to SiC particles.
Efforts to demonstrate that the dendritic web technology is ready for commercial use by the end of 1986 continues. A commercial readiness goal involves improvements to crystal growth furnace throughput to demonstrate an area growth rate of greater than 15 sq cm/min while simultaneously growing 10 meters or more of ribbon under conditions of continuous melt replenishment. Continuous means that the silicon melt is being replenished at the same rate that it is being consumed by ribbon growth so that the melt level remains constant. Efforts continue on computer thermal modeling required to define high speed, low stress, continuous growth configurations; the study of convective effects in the molten silicon and growth furnace cover gas; on furnace component modifications; on web quality assessments; and on experimental growth activities.
The high speed solar wind, which is associated with coronal holes and unipolar interplanetary magnetic field, has now been observed in situ beyond 0.3 a.u. and at latitudes up to 80 degrees. Its important characteristics are that it is remarkably steady in terms of flow properties and composition and that the ions, especially minor species, are favored in terms of heating and acceleration. We have proposed that the high speed wind, with its associated coronal holes, forms the basic mode of solar wind flow. In contrast, the low speed wind is inherently non-stationary, filamentary and not in equilibrium with conditions at the coronal base. It is presumably the result of continual reconfigurations of the force-free magnetic field in the low-latitude closed corona which allow trapped plasma to drain away along transiently open flux tubes. Observations of high speed solar wind close to its source are hampered by the essential heterogeneity of the corona, even at sunspot minimum. In particular it is difficult to determine more than limits to the density, temperature and wave amplitude near the coronal base as a result of contamination from fore- and back-ground plasma. We interpret the observations as indicating that the high speed solar wind originates in the chromospheric network, covering only about 1% of the surface of the sun, where the magnetic field is complex and not unipolar. As a result of small-scale reconnection events in this 'furnace', Alfven waves are generated with a flat spectrum covering the approximate range 10 kHz to 10 Hz. The plasma is likely to be produced as a result of downwards thermal conduction and possibly photoionization at the top of the low density chromospheric interface to the furnace, thus controlling the mass flux in the wind. The immediate source of free (magnetic) energy is in the form of granule-sized loops which are continually carried into the network from the sides. The resulting wave spectrum is such that energy can be efficiently transferred to the ions within a few solar radii of the base of the corona, favoring heavy species and creating stable, fast solar wind.
The absorption spectrum of Si I in the wavelength region between 1500 and 1900 A has been photographed at high resolution. The silicon vapor was produced in a 122-cm-long King furnace at 1800 to 2300 C. Forty-two Rydberg series have been observed from the ground state terms 3p(2)3P and 1D to terms associated with the 3pns and 3pnd configurations. All of the series from these configurations with J less than 4 have been extended with the 3pnd 3D(3) levels reaching n = 56. Numerous perturbations have been observed. This laboratory work has provided the basis for extending the identification of silicon lines in the solar spectrum. Nearly all lines found in the laboratory spectrum are also found in rocket spectrograms of the solar chromosphere. More than 300 lines have been attributed to Si I. The excellent correlation between laboratory and solar Si I lines is illustrated.
Solar cells made from directionally solidified silicon cover 57% of the photovoltaic industry's market [1]. One major issue during directional solidification of silicon is the precipitation of foreign phase particles. These particles, mainly SiC and Si3N4, are precipitated from the dissolved crucible coating, which is made of silicon nitride, and the dissolution of carbon monoxide from the furnace atmosphere. Due to their hardness and size of several hundred micrometers, those particles can lead to severe problems during the wire sawing process for wafering the ingots. Additionally, SiC particles can act as a shunt, short circuiting the solar cell. Even if the particles are too small to disturb the wafering process, they can lead to a grit structure of silicon micro grains and serve as sources for dislocations. All of this lowers the yield of solar cells and reduces the performance of cells and modules. We studied the behaviour of SiC particle depots during float-zone growth under an oxide skin, and strong static magnetic fields. For high field strengths of 3T and above and an oxide layer on the sample surface, convection is sufficiently suppressed to create a diffusive like regime, with strongly dampened convection [2, 3]. To investigate the difference between atomically rough phase boundaries and facetted growth, samples with [100] and [111] orientation were processed.
Stirling Technology Company (STC) developed the RG-350 convertor using components from separate Goddard Space Center and U.S. Army Natick SBIR contracts. Based on the RG-350, STC commercialized a product line of Stirling cycle generator sets, known as RemoteGen(TM), with power levels ranging from 10We to 3kWe. Under SBIR agreements with Glenn Research Center, the company refined and extended the capabilities of the RemoteGen convertors. They can provide power in remote locations by efficiently producing electricity from multiple-fuel sources, such as propane, alcohol, gasoline, diesel, coal, solar energy, or wood pellets. Utilizing any fuel source that can create heat, RemoteGen enables the choice of the most appropriate fuel source available. The engines operate without friction, wear, or maintenance. These abilities pave the way for self-powered appliances, such as refrigerators and furnaces. Numerous applications for RemoteGen include quiet, pollution-free generators for RVs and yachts, power for cell phone towers remote from the grid, and off-grid residential power variously using propane, ethanol, and solid biomass fuels. One utility and the National Renewable Energy Laboratory are evaluating a solar dish concentrator version with excellent potential for powering remote irrigation pumps.
A process for the continuous growth of crystals by the Czochralski method, suitable for producing single silicon crystals for use in solar cells was studied. Continuous growth is the growth of 100 Kg of single silicon crystals, 10 cm in diameter, from one container. A furnace with continuous liquid replenishment of the growth crucible, accomplished by a melt-down system and a liquid transfer mechanism, with associated automatic feedback controls was developed. Elements of the transfer system were further developed and tested during actual transfer runs. Considerable simplification of the heating element of the transfer tube was achieved. Accuracy and reliability of the temperature sensor, which is part of the power input control system for the transfer tube, was improved. Electrical and thermal effectiveness were increased while assembly of the transfer tube system was further simplified.
Equipment developed for the manufacture of over 100 kg of silicon ingot from one crucible by rechanging from another crucible is described. Attempts were made to eliminate the cost of raising the furnace temperature to 250 C above the melting point of silicon by using an RF coil to melt polycrystalline silicon rod as a means of rechanging the crucible. Microprocessor control of the straight growth process was developed and domonstrated for both 4 inch and 6 inch diameter. Both meltdown and melt stabilization processes were achieved using operator prompting through the microprocessor. The use of the RF work coil in poly rod melting as a heat sink in the accelerated growth process was unsuccessful. The total design concept for fabrication and interfacing of the total cold crucible system was completed.
Some hypotheses to explain both of these features are advanced and the possible implications for solar cell performance are touched upon. The multiple ribbon growth system has shown a number of flaws with respect to the reliability of the basic furnace design. These definitely need to be rectified before any significant demonstration of multiple ribbon growth can proceed. The cartridges, however, have performed quite well. The work on 3" cartridge design and automatic controls has proceeded nearly on schedule and the report contains a detailed description of the approach and the equipment to be used for automatic control of ribbon growth.
In the photovoltaics industry, the largest market share is represented by solar cells made from multicrystalline silicon, which is grown by directional solidification. During the growth process, the silicon melt is in contact with the silicon nitride coated crucible walls and the furnace atmosphere which contains carbon monoxide. The dissolution of the crucible coating, the carbon bearing gas, and the carbon already present in the feedstock, lead to the precipitation of silicon carbide, and silicon nitride, at later stages of the growth process. The precipitation of Si3N4 and SiC particles of up to several hundred micrometers in diameter leads to severe problems during the wire sawing process for wafering the ingots. Furthermore the growth of the silicon grains can be negatively influenced by the presence of particles, which act as nucleation sources and lead to a grit structure of small grains and are sources for dislocations. If doped with Nitrogen from the dissolved crucible coating, SiC is a semi conductive material, and can act as a shunt, short circuiting parts of the solar cell. For these reasons, the incorporation of such particles needs to be avoided. In this contribution we performed model experiments in which the transport of intentionally added SiC particles and their interaction with the solid-liquid interface during float zone growth of silicon in strong steady magnetic fields was investigated. SiC particles of 7μm and 60μm size are placed in single crystal silicon [100] and [111] rods of 8mm diameter. This is achieved by drilling a hole of 2mm diameter, filling in the particles and closing the hole by melting the surface of the rod until a film of silicon covers the hole. The samples are processed under a vacuum of 1x10(exp -5) mbar or better, to prevent gas inclusions. An oxide layer to suppress Marangoni convection is applied by wet oxidation. Experiments without and with static magnetic field are carried out to investigate the influence of melt convection on the distribution of particles and their incorporation into the crystal. The field strengths applied by a superconducting magnet are 1T, 3T, 4.5T, and 5T. The increase in field strength dampens the melt flow, and so this study provides comparative data to the crystal growth experiment to be carried out onboard the sounding rocket mission TEXUS 51, where purely diffusive growth condition will be achieved under microgravity conditions.
Several experimental and projected Czochralski crystal growing process methods were studied and compared to available operations and cost-data of recent production Cz-pulling, in order to elucidate the role of the dominant cost contributing factors. From this analysis, it becomes apparent that substantial cost reductions can be realized from technical advancements which fall into four categories: an increase in furnace productivity; the reduction of crucible cost through use of the crucible for the equivalent of multiple state-of-the-art crystals; the combined effect of several smaller technical improvements; and a carry over effect of the expected availability of semiconductor grade polysilicon at greatly reduced prices. A format for techno-economic analysis of solar cell production processes was developed, called the University of Pennsylvania Process Characterization (UPPC) format. The accumulated Cz process data are presented.
Screened electrodes made from fluorocarbon activated copper paste and silver fluoride activated copper paste, tape adhesion and scratch tests were studied. Experiments were conducted with variations in past parameters, firing conditions, including gas ambients, furnace furniture, silicon surface and others. A liquid medium intended to provide transport during the carbon fluoride decomposition, is incorporated in the paste.
Open-circuit voltages as high as 0.645 V (AM0, 25 C) have been obtained by a new process developed for low-resistivity silicon. The process utilizes high-dose phosphorus implantation followed by furnace annealing and simultaneous oxide growth. The effect of the thermally grown oxide is a reduction of surface recombination velocity; the oxide also acts as a moderately efficient antireflection (AR) coating. Boron-doped, float-zone silicon with resistivities from 0.1 to 1.0 (omega)(cm) has been processed according to this sequence; results show that the highest open-circuit voltage is obtained with 0.1(omega)(cm) starting material. The effects of Auger recombination and bandgap narrowing caused by high doping concentrations in the n(+)junction region have been investigated by implanting phosphorus over a wide range of dose levels. The effects of emitter-phosphorus concentrations tailored to optimize electric fields in the emitter have also been investigated.
Solar cells fabricated from HEM cast silicon yielded up to 15% conversion efficiencies. This was achieved in spite of using unpurified graphite parts in the HEM furnace and without optimization of material or cell processing parameters. Molybdenum retainers prevented SiC formation and reduced carbon content by 50%. The oxygen content of vacuum cast HEM silicon is lower than typical Czochralski grown silicon. Impregnation of 45 micrometers diamonds into 7.5 micrometers copper sheath showed distortion of the copper layer. However, 12.5 micrometers and 15 micrometers copper sheath can be impregnated with 45 micrometers diamonds to a high concentration. Electroless nickel plating of wires impregnated only in the cutting edge showed nickel concentration around the diamonds. This has the possibility of reducing kerf. The high speed slicer fabricated can achieve higher speed and longer stroke with vibration isolation.