The EFG employs drawing (Figure IV.B.l.a.18) a ribbon of silicon through a carbon die which is fed by capillary action. The problem with the EFG method is that molten silicon which reacts with almost all substances combines with the die material. Carbon has been the most successful material used to date. However, even with it, the surface of the ribbons has pieces of silicon carbide imbedded in it. Also, there are numerous lattice defects in the crystal's surface and the general problem of twinning. Several other methods of fabricating silicon sheets are being studied under ERDA/JPL contracts. One such method involves producing silicon on a (Figure IV.B.l.a.19) molten layer of tin similar to the process which is used to produce plate glass. Another technique being investigated is dipping (Figure IV.B.l.a.20) in which a carbon coated ceramic sheet is dipped into molten silicon and then withdrawn. The resulting sheet is hoped to have a single crystalline layer of silicon on its surface. Another proposed method of forming silicon sheets involves forming a silicon plate between two rollers (Figure IV.B.l .a.21). The thick slab of silicon is heated to its elastic temperature as it is squeezed between the rollers into a thin sheet. It is doubtful that the resultant material will be a single crystal which is necessary for high efficiency solar cells. It is also believed to suffer from similar reaction problems as the EFG technique. (c) GALLIUM ARSENIDE While the earth's crust contains an abundant supply of gallium (15 ppm), gallium does not occur in concentrated quantities but as a trace element in ores of Al, Zn, Cu, and coal. At present gallium is obtained from bauxite which has an average concentration of 50 ppm as a by-product in the production of aluminum. The recovery process is inefficient and yields 1 percent of the gallium present. Present estimates indicate that the process can be improved to recover as much as 30 percent of the gallium present. Improvement of the process will increase initially the cost of gallium from a recent price of $600/kg. Mineral producers regard the production statistics of gallium as proprietary data and therefore production data is estimated (Table IV.B.l.a.3). Current estimates place the U.S. production of gallium in excess of one ton and world production at more than 12 tons. With the increased usage and expanding market in GaAs light emitting diodes, the production and consumption are continually increasing. From the known reserves of bauxite, the U.S. reserves of gallium are estimated to be from 2,700 to 8,400 tons, and the world reserves are 115,000 tons. Therefore, a more precise estimate of the gallium required for a GaAs solar cell array will require sufficient research and development to define an acceptable GaAs cell. The preliminary
RkJQdWJsaXNoZXIy MTU5NjU0Mg==