ployed.” The SPS studies presumed significant advances in processes and changes in materials. Boeing and Arthur D. Little examined the potentials of automating present methods, e.g., gang-sawing of Czochralski-grown crystals, with evolutionary process improvements, projecting that the SPS estimates could be approached, but perhaps not reached. If the JPL estimate is extrapolated to 18,000 MW per year (the level required by the SPS reference program), the resulting figure is about $1.35 per watt, compared with about $0.25 per watt for the SPS estimates (for silicon). If the SPS estimates had been at this higher value, the tradeoff of whether to use concentrators would have been different; concentrators would have been included in the reference silicon system with resulting mitigation of the higher costs. The result would have been an SPS system 10% to 20% more expensive. Potentials for reducing space transportation costs have been debated for decades. The original target for the space shuttle (for a fully reusable design) was less than $100 per pound, in something like 1970 dollars. Current estimates for shuttle are about $1000 per pound. Shuttle critics tend to forget the complete reusability of the early concepts, not to mention inflation. The low costs predicted by the SPS studies resulted from economies of scale (larger vehicles with higher performance), economies of rate, complete reusability, all low-cost propellants, and maturity of the technology permitting fast turnaround. It is difficult to convince someone familiar with the difficulties of turning the shuttle orbiter around in two months that turnarounds for advanced vehicles might be as short as two days. But it can happen. A different and more constructive light can be shed on these arguments by calculating intrinsic costs and then considering the nature of advances in the state of the art needed to achieve these. The intrinsic cost of the SPS solar array is readily estimated. The cost of the raw material, semiconductor grade (SeG) silicon, is roughly $ 100/kg. The unit mass of silicon in the SPS array was 0.24 kg/m2 and the estimated output 180 W/m2. About 60% of this output reached a utility grid as useful power. The cost figures to 13 cents per watt on the array, or $220 per kW at the grid. To reach a lower limit realistic figure, we should include a value-added multiplier of about two for costs of other materials, labor and capital equipment, plus a profit of 15% or so for the manufacturer. This yields about 30 cents per watt on the array, not much different than the SPS estimates. Now about the technology: This scenario requires that the process for making solar cells from SeG silicon result in little or no waste, that the production plant and machinery to produce 18,000 MW per year cost no more than five to ten billion dollars, and that the plant employ no more than five to ten thousand people. Note that the annual sales of this industry, at 30 cents per watt, are over five billion dollars, certainly a large and presumably mature industry. These figures are not, at least at first blush, unimaginable, but they do represent a yet-to-be-invented process, i.e., one without waste, and a degree of automation and scale-up that is far beyond anything presently conceived in plant design. This assessment is, of course, much the same as that made by the NAS, but in a positive light. Although we don't know how to do it today, it is not an unreasonable target. A similar estimate can be made for space transportation. The energy of an object in GEO orbit is about 16 kWh/kg, and the (retail) cost of thermal energy today, in the form of oil, is roughly $0.05/kWh. The efficiency of chemical rocket propulsion to space, for practical vehicles, is only about 5% because the attainable jet velocity is much less than orbital velocity. A value-added factor of two is again appropriate, at
RkJQdWJsaXNoZXIy MTU5NjU0Mg==