Presented at the XXIth Space Photovoltaic Research and Technology Conference (SPRAT-2009), Cleveland, OH, October

Transcription

1 Presented at the XXIth Space Photovoltaic Research and Technology Conference (SPRAT-2009), Cleveland, OH, October SOLAR POWER FROM SPACE: SEPARATING SPECULATION FROM REALITY Geoffrey A. Landis NASA Glenn Research Center Cleveland, OH BACKGROUND AND HISTORICAL DEVELOPMENT With the increases in energy cost and recent interest in finding ways to produce energy with reduced emission of greenhouse gasses, there has been renewed interest in the concept of producing power using solar panels in space, and then beaming this power downward to provide electrical power for use on the Earth. This concept, called the "Solar Power Satellite," was first proposed by Peter Glaser in 1968 [1], and, in revised and updated form, has been proposed many times since [2-5] as a possible solution to the energy crisis. It is the purpose of this paper to examine this concept on a physics basis, reviewing the basic concepts in order to assess the current and future feasibility without adopting either a advocating or critical stance. The Satellite Power System was studied in the late 1970s by NASA in collaboration with the Department of Energy, producing a conceptual "reference" design [3] for a system. This design was analyzed and critiqued by an Office of Technology Assessment study [4]. The concept has gone by several different names and acronyms, starting as the "Solar Power Satellite" (SPS) or "Satellite Solar Power System" (SSPS), and more recently studied under the name "Space Solar Power" (SSP). Figure 1 shows a summary of the 1980 "Reference Design" for such a solar power satellite. The baseline satellite concept produces about 10 GW of electrical power at the Earth, using a large (10 km by 15 km) solar array located in Geosynchronous orbit. The power is transmitted to the Earth by a microwave beam at 2.45 GHz, and a large (approximately 100 square kilometers) rectifying antenna (or "rectenna") array at Earth receives the beamed microwave power and converts it into DC electrical power. The early study also looked at several alternate technologies for both the energy conversion and the power transmission, and made attempts to predict the In 1995, NASA headquarters initiated a "Fresh Look" study of solar power satellites, which did not revise the original concepts, but started with a clean sheet of paper to re-think the basic concepts and come up with a new design [5]. Some of the initial concepts examined included use of low Earth orbit instead of geosynchronous orbit, gravity gradient stabilized structures, sun-synchronous orbits, and use of large- area Fresnel lenses to focus light onto panels of concentrator cells. Later evolution of the Fresh Look study (which evolved into the Space Solar Power Exploratory Research and Technology (SERT) program [6-8] in the time frame) returned to the geosynchronous orbit design, but continued to look at new options for satellite design, including alternative power transmission methods such as laser beaming. In addition to these NASA studies, there have been several non-nasa studies of the concept [9], as well as a number of studies that have been proposed to use power-beaming technology for other applications, including inspace applications such as satellites, electric-propulsion, and lunar bases [10]. It's worth noting that in 2004, the North American electrical energy generation was 4730 Terawatt-hrs, about 45 percent of which was generated by coal-fired plants. At a typical production cost of 5 cents per kw-hr, this

2 represents a revenue of 230 billion dollars per year for the potential market in North America alone. Thus, although the concept requires mega-engineering on a scale that would dwarf all previous space projects, the potential revenue from it is extremely large. Figure 1: Conceptual diagram of 1980 Solar Power Satellite "Reference Design" [3]. POWER BEAMING FUNDAMENTALS The ability to use solar power generated in space for terrestrial use requires beaming the power from the source, in space, to the user on the ground. This is done by converting the electrical power to electromagnetic radiation, beam the radiation across free space, and collecting it at a receiver that converts electromagnetic radiation to electrical power. In the 1980 study [3] (and most of the subsequent studies), it was proposed to do this using radio-frequency ("RF") beams at a frequency of 2.45 GHz (that is, microwaves in the Industrial, Scientific and Medical Band). This was chosen because of the high demonstrated conversion efficiency of electrical power to microwave energy, and because of the transparency of the atmosphere to microwave radiation. Efficiency of Microwave transmission In the "ideal" case, RF power beaming efficiency can be quite high. At these frequencies, magnetron tubes can convert DC to RF efficiency at transmitter efficiency of 90% or better, and a rectenna array (consisting of an array of GaAs Schottky diodes and quarter wave antennas) can convert RF power back to DC at receiver efficiency that has been demonstrated as high as 86%. The product of these two efficiencies yields an overall "potential" transmission efficiency, DC in to DC out, of ~77% In the real world, however, it is very easy to produce much degraded efficiency. For example, Neville Marzwell estimated the following potential losses in a Real World estimated RF Link efficiency [11]: 2

3 DC to RF efficiency: 0.90 EMC and diplexing: Subarray aperture efficiency 0.95 Subarray failures 0.96 Amplitude errors and taper quantization Phase errors 0.97 Electronic beam steering scan loss Clear air absorption Propagation impairments (scattering) Coupling efficiency (divergence) Polarization mismatch Rectenna aperture efficiency 0.95 (reflections) EMC filtering Rectenna RF-DC conversion efficiency 0.86 Table 1: Estimated link efficiencies for a realistic microwave transmission (from Marzwell 2007 [11]) Multiplying each of these losses out yields a much lower value of 37.5% for the DC-to-DC efficiency, and even this has not accounted for any losses of the transmitted beam not captured by receiver, nor additional losses in power management and distribution either at the satellite or the user side of the system. The reference design study assumed that the transmitter would be a phased array consisting of many millions of individual small transmitters operating in phase. The phase would be controlled by a "beacon" at the ground site, and hence if the beacon is not present, the beam would fail to be in phase, and hence would not be focused. The requirement for a cooperative link at the ground station functions as an additional safety factor; in he event of an error, the transmitter is not capable of beaming high intensities to sites other than the ground station. Spot Diameter The minimum spot diameter of any transmitted electromagnetic beam is set by the diffraction limit. The Airy Diffraction disk is the diffraction pattern produced from a uniformly illuminated circular aperture [12, 13]. This area contains 84% of the beam energy. The "tails" of the diffraction pattern outside this have little power, with a fall-off in power of approximately P ~ exp[-r 2 ]. The diameter of the Airy diffraction pattern is: d spot = 2.44 λx/d transmit (1) where d transmit is the diameter of transmitter aperture, x the source to receiver distance, and λ the wavelength. This is the fundamental physics of electromagnetic power transmission a beam can do worse than this (i.e., a larger spot size) but cannot be focused into a smaller size. We can use this equation to calculate the minimum spot size for a beam originating at geosynchronous orbit. For this case, the distance x= km. At a frequency of 2.45 GHz, the wavelength is 122 m (12.2 cm). The diffraction-limited spot size is: d spot = 10.7/d transmit (for d in kilometers) (2) Thus, for a 1 km diameter transmitter, the minimum spot size on Earth is 10.7 km. This explains why the 1980 reference design (and most successive design concepts) is so large: it is necessarily large because of the beam spot size. The transmitter or receiver diameter is reduced in size directly proportional to wavelength. 3

4 Figure 2: The Airy diffraction pattern (from [12]). Changing to other wavelengths has a direct effect on the size of the system. Assuming that both the transmitter and receiver diameters are reduced proportionately, the spot size for transmitting from geosynchronous orbit is shown in table 2. F d (transmit) d (receiver) 2.45 GHz 1 km 10.7 km 5 GHz 0.73 km 7.3 km 10 GHz 0.51 km 5.1 km 100 GHz 0.16 km 1.6 km Table 2: Aperture size vs. frequency for a beam transmitting from geosynchronous orbit, for case of transmit aperture diameter = 10x receive aperture Minimum power level for microwave beaming These receiver sizes, in turn, mean that there is a minimum power level at which the system needs to operate. This is partly an economic issue: At lower power intensity the transmitter and receiver costs will be too expensive (i.e., it would not be economically feasible to build a 1 km diameter transmitting antenna in space to transmit a power of only 1 kw), but it is also an effect of the physics of the receiving antenna: rectennas are inefficient at low 4

5 power intensities. A minimum intensity at the receiver is required in order for the diode rectifiers to convert power at high efficiency. Diffraction limits thus mean that SPS is inherently big: small scales are not feasible It is clear that the minimum feasible power scales directly with the wavelength, or 1/microwave frequency. Assuming minimum intensity of 100 W/m 2 (100 MW/km 2 ), this minimum power is: 2.45 GHz: 9 GW 5 GHz: 5.3 GW 10 GHz: 1.6 GW 100 GHz: 0.5 GW This shows that higher frequencies allow much smaller system sized. microwaves have a higher absorption by water vapor: 10 GHz won't penetrate rain 100 GHz needs receiver on mountaintop Can a large aperture be synthesized from many small transmitters? However, the higher frequency Instead of making a single large aperture, the idea is sometimes suggested that a small spot size can be synthesized from a large array of small apertures. This is effectively a suggestion that a large aperture can be "synthesized" using the interferometric aperture synthesis approaches which is done, for example for radiotelescopes such as the Very Large Array ("VLA"). The short answer to this question is no. While high observational resolution can be achieved by aperture synthesis, this is not effective for transmission. A large transmitting array synthesized from many small arrays can beam to a smaller area, but the power density beamed into that area does not increase-- in essence, the extra beam power is lost into side lobes. This effect is known as the "Thinned Array Curse" [13]. Power transmission by Laser Beaming Power can also be beamed with a laser beam, using a photovoltaic array as a receiver [10, 14] Laser power beaming still uses electromagnetic radiation, and hence is still subject to the electromagnetic limitations on beam size due to diffraction. However, since laser radiation has much shorter wavelengths, smaller transmitter and receiver sizes can be used. An example of 1 μm is chosen in the near IR, close to the response peak for a high-efficiency silicon photovoltaic receiver. Shorter wavelengths will reduce the size proportionately, but will require more expensive receiving arrays. Assuming minimum intensity of 500 W/m 2 for optical (producing an output equivalent to solar illumination, 1 sun intensity), and assuming a transmitting lens diameter of 2.3 meters, a laser beaming system operating at 1μm will produce a 23 m diameter minimum spot size on the ground, for an area of km 2. This results in a minimum power level of 208 kw. Laser transmission removes problem of inherently large sizes, but lasers have their own problems. First, laser efficiencies are considerably lower than microwave efficiencies, for lasers with good coherence. High power semiconductor diode lasers arrays are highly efficient (50% conversion efficiency or higher), but are not mutually coherent-- the net result of a high-power laser diode array is that it will have the diffraction pattern characteristic of a flashlight, not the narrow diffraction-limited spot size of a laser. Existing technology lasers might have efficiency approaching ~ 40% (for example, for a diode-pumped alkali) A second problem is that PV converter efficiencies are also low. The conversion efficiency is better than solar conversion efficiency, because the beam can be made monochromatic at a wavelength tuned to the optimum conversion wavelength of the cell, but is still lower than rectenna conversion efficiencies. 50% conversion efficiency is a reasonable efficiency. 5

6 For laser transmission, clouds are now a problem. In addition, eye safety is now a problem. Overall, use of a PV array for power receiving eliminates the signal advantage of space solar power, of putting the PV array in space, for cloud-free power. Laser-transmitted space power has less power per solar array area than ground solar Despite these disadvantages, laser transmission also has advantages, The laser power converters are PV arrays, and so can operate both on the beamed laser power, as well as on transmitted laser power. Thus, laser power satellite could "fill in" power to already existing solar arrays. If a ground solar installation has already paid for the ground infrastructure, space photovoltaic power looks much more attractive! The ground infrastructure now gets double use for both solar and laser power. This may provide an evolutionary route to space power from ground power. However, ground solar arrays typically do not cover 100% of the land area. This wastes laser power (unless ground arrays redesigned). Clouds and eye safety are both still a problem. Figure 3: Artist's conception of a concept for a solar power satellite in Geosynchronous Earth Orbit. (This conceptual diagram varies from the Reference design in that the artist shows two microwave beaming transmitters, servicing two different spots on the Earth, rather than a single transmitter.) 6

7 WHY PUT SOLAR IN SPACE? The critical question is, why should the solar arrays be placed in space? Why not just put the same solar cells on the ground? The proposed rationale for space solar is that, unlike the ground, space has solar power 24 hours per day, 365 days per year. How much more power can be achieved by putting the cells in space? Three reasons are usually given for why more power is achieved for space solar: 24 hour sunlight space has continuous sunlight, 24 hours per day. Space solar arrays are (or can be) directly sun pointing Space has no atmosphere, and no clouds The availability of 24-hour sunlight inherently gives a factor of two advantage; somewhat more in winter, less in summer. Sun-pointing Space solar arrays are (or can be) directly sun pointing. The added benefit of tracking depends on pointing assumptions. For solar arrays located not too far from the equator, calculating the added power using the cosine approximation gives an added factor of π/2 in the average power on a solar array. This is exactly true only ignoring atmospheric effects (discussed in the next subsection), but is roughly true for a fixed solar array tilted south at the latitude angle. It is also worthwhile to note that, in fact, ground solar arrays can be made sun pointing. This is slightly less efficient in the use of land area, but allows more effective utilization of the solar cell area. No atmosphere This comprises two advantages: the sunlight is not filtered by atmosphere, and space has no clouds. The solar intensity is 1.37 W/m2 in space, compared to 1.0 kw/m2 (AM1.5 standard). The AM1.5 standard only applies for high sun angles near noon; there is more atmosphere filtering at higher sun angles, later in the day. However, despite the higher intensity, solar cell efficiency in space is less than that achieved on ground by roughly factor of 0.8 Accounting for both intensity and efficiency, solar cells produce about 10% more power at AM0 (space) than AM1.5 (ground spectrum). Clouds also are an issue for ground solar. For a fair comparison of space and ground solar installations in the near term, however, we need to compare space solar power with the best ground solar locations, not the average locations. From NREL data on average solar power, for the best locations in the southwest United States, the annual average solar radiation is 12 kw-hr/m2, corresponding to 1 kw/m2 for 12 hrs/day This is, essentially, zero cloud loss. Summation: how much more power do you get by putting the cells in space? Space has solar power 24 hours per day: added power = factor of 2 Directly sun pointing: added power = factor π/2 (if ground arrays are not sun pointing). Higher intensity in space: added power ~ factor 1.1 Space has no clouds: No added power (over cloud-free ground locations). The bottom line is that a solar array in space produces a factor of 1.1π = 3.5 times more energy/area than the solar arrays at the best ground locations. 7

8 The direct corollary to this is that, unless the DC to DC transmission efficiency of the power beaming link is at least 29%, the net power output from the space system will be less than the energy output from the same array placed (at the optimum location) on the ground. In the most optimistic case of transmitter efficiency-conversion efficiency 90%, Airy disk capture of 84%, and rectenna conversion efficiency of 86%, the ground DC power is 65% of the power produced in space. This reduces the above numbers to an improvement of 2.25 for the space array compared to the ground array. For the case that the ground array is tracking, the second factor reduces to one, and a space array produces 2.2 times more power per unit area than a tracking array at the best site on the ground. Incorporating optimistic transmission factor of 65%, this reduces to 1.4 times more power. Future Evolution The discussion so far has compared space location of a solar array with the best locations on the ground. This is appropriate for the initial phases of solar power, since the first implementations of large-scale power production by photovoltaics will, of course, be at the best locations, and not at the words. However, looking into the longer term, not all sites on the ground are best sites. Long-distance transmission lines can transmit power on the ground for some distance, but there will be large losses for transcontinental transmission-- it is not feasible to power produce in the Mojave Desert and use it in New York. Ground solar is worse by a factor of two for areas of the US outside of southwest, and as much as a factor of 2.5 worse for New England, a significant electrical market. (It's worth noting, however, that the ground receivers for space solar power also have significant location constraints, most notably a requirement for large areas of land in order to incorporated "keep-out" zones near the beam, and thus may not be able to be located near large cities in the northeast in any case.) A caveat on this calculation is that it has implicitly assumed balance-of-systems cost of ground solar array (e.g., land cost) and space solar array (e.g., rectenna land area) are comparable. In addition to the potentially higher total amount of power produced, space solar power may have other virtues. Most notable of these is that the power is continuous, 24 hour power, rather than power peaked at noon and dropping to zero at sunrise and sunset. At the moment, however, 24-hour power is not an asset, since during the night-time hours, the power demand is very low, and power available at night sells at very low price. Ground solar produces power that is moderately well matched to the (early afternoon) peak demand. Nevertheless, as solar capacity grows, this production curve will be increasingly mismatched to the demand and eventually solar will need to provide power outside peak solar hours. The transition of solar power from peak to a requirement for power outside of the mid-day peak is typically expected to occur when ground solar reaches ~ 10-15% of the energy market. (In the US, this represents about 300 Billion dollars per year total, although the price-break occurs earlier in the areas where solar is most effectively used). At this point, the continuous availability of power from space becomes an asset. IS GEOSYNCHRONOUS THE RIGHT ORBIT? This analysis has assumed that a power satellite would be in Geosynchronous Earth Orbit (GEO). GEO is a location where the beaming station remain stationary (with respect to the Earth) over the equator at the longitude of the receiver. This maximizes the utility of the station at a given receiver site. It is interesting to look at the possibility of other orbits, in particular, lower orbits that would allow a shorter distance to beam, and hence smaller spot sizes and a smaller system size. Only GEO orbit puts satellite over ground station with 100% usage fraction, and hence any lower orbit will have an immediate disadvantage that it will be out of direct beaming line of sight of the ground station for much of the time. In addition to the non-stationary nature of lower orbits, another difficulty of low orbit is that these orbits will have to be non-equatorial if we want to get power to northern hemisphere users. Thus, low-orbit view factors are 8

9 low; for example, for an orbital altitude of 1000 km, the time in view above ten degrees of elevation is only 12.6 minutes, twice a day. This results in a total use fraction of 25.2 minutes out of 24 hours, which is too low a usage fraction to be economically feasible. One possible solution would be to make multiple ground stations for the power, receiving power at whatever location is in sight of the ground station, and likewise multiple power satellites, so that a satellite is available over each customer at any time. However, to make this work for low orbits would require a large number of ground stations dotted uniformly around the world, including in many locations where there are few customers for the power, such as the Pacific Ocean. The cost of such a system is several orders of magnitude higher than the baseline, since the number of satellites is much higher. It is difficult to make this economic case. An alternative would be to reconsider the proposal to send power directly to northern-hemisphere ground sites, and to put the power beaming satellites into a low equatorial orbit, and beam only to sites on, or near, the equator. Users at sites distant from the equator would then have to either be served by transmission lines, or else by secondary beams passing through microwave relay satellites. This reduces the number of power satellites and receiver stations considerably; for example, approximately 24 satellites in 1000-km orbit could provide continuous power to service sites at or near the equator. Although this is a larger number of satellites than the number required in geosynchronous orbit, the transmitting aperture of each satellite is 25 times smaller. Another approach is to adapt the space power concept to the price structure of the electric power market, by redesign of solar power satellite to maximize the amount of power available to service peak power markets, instead of baseload. If this could be done, it would roughly doubles the revenue in terms of $/kw-hr. Analyses of this approach have been summarized in an earlier paper, "Reinventing the Solar Power Satellite", NASA [15, 16]. THE ECONOMIC CASE The economic return for space solar power requires return on investment. If a SPS is to be commercially viable, it must charge the utilities to which it is selling power price (per kw-hr) less than the utility's cost of generating new power. Note that it is important to beat the utilities' cost, not the customer's electric cost. This cost may include the cost for externalities (e.g., possible penalties to be imposed for generating with coal), if any. As a minimum, even if operating cost is zero (i.e., small compared to the capital) and operating lifetime is infinite, the invested money must be returned. Quick & dirty economics It is instructive to calculate example numbers for return on investment case. The rate of return must be at least: $/yr > P*Capital cost *(1+P) Construction time (3) (Where P is the required rate of return on investment). This must be divided by the power produced per year to get the dollars per kw-hr. Power produced is Thus: kw-hr/yr = 8700 hr/yr*1,000,000kw/gw*power(gw) (4) $/kw/hr > P*Capital cost($m)*(1+p) Construction time /8700*Power(GW) (5) Let us assume an investment rate of return of 8.25% (P =0.0825), a construction time of 5 yrs. Then the factor (1+P) Construction time = 1.49, which raises the effective interest rate to 12.3%. The equation then is: $/kw-hr > 0.014*Capital cost($b)/power(gw) (6) This can also be expressed as a limit on the capital cost, as a function of the selling cost of electricity: Capital cost($b) < 70.8 ($/kw-hr)*power(gw) (7) 9

10 Assuming a power level of 5 GW and a selling price of 5 /kw-hr, this is: Capital cost < $17.5 B This analysis has assumed an infinite lifetime of the assets, and no associated operating costs. If, instead, the loan is to be paid off in a finite amount of time, a slightly higher rate of return is required. For 30-year payoff of 8.25% on the loan, the payment is 1.09 times higher rate; for a 25-year payoff, the yearly payment is 1.15 times higher. If operating costs are not negligible, they need to be added directly to the cost per kilowatt-hour. It is useful to divide the maximum allowable capital cost ($17.5 B for the infinite lifetime/zero operations cost case) by the power produced, 5 GW. This shows that, to be sell electricity at a price of $0.05/kW-hr, the total cost of the system must not exceed $3.50/watt. This is an extremely challenging target, since it must include not only the solar arrays, but also all of the in-space structure as well as the ground assets. BEAMING POWER TO SPACE: A FIRST STEP TO SPS? The economic calculation has so far assumed that the market for beamed power is in fact the terrestrial electric-power market. The market on Earth, however, is currently served by relatively low cost generating capacity. It is much more likely that the first steps to demonstrating and using beamed power would be for ground-to-space or space-to-space power beaming [17]. Probably the best market is to start with in-space power beaming [13]. Currently, electrical power in space has an effective price tag that is 10,000 times price of power on the ground. It makes sense to beam power to the place where it is expensive, from the place where it's cheap. Figure 4: Earth to space power beaming concept, showing possible receivers in satellites, electric-propulsion transfer vehicles, and lunar bases [13]. SUMMARY POINTS TO PONDER Producing solar power in space to be sent to the surface of the Earth for use by consumers on the ground ("Satellite Solar Power," or SPS) has been proposed as a means of solving the problem of electrical supply on the surface of the Earth with a renewable, low-carbon-emission technology. The fundamental physics are feasible, but economical feasibility is as yet an open question. Although a space location for the solar panels gets more sun 10

11 than a ground location, the bottom line numbers show that it is not that much more solar energy than the best ground locations. The added power mostly from 24-hr sunlight, but much of the power may thus be produced when the need is low. Electromagnetic beam diffraction limits mean that SPS is inherently big. 5 GW minimum for 5GHz power beam --Higher frequency won't penetrate rain. Switching to laser transmission allows very small satellites, but loses big on efficiency To a rough approximation, solar arrays in space produce 3.5 times more power than non-tracking array on the ground array, or 2.2 times more than tracking array on the ground. Accounting for transmission losses, this reduces to 1.63 times more power/solar-cell area. If cost of solar arrays is less than 61% of the total SPS cost, array is better on ground Space solar comes in after ground solar is economically competitive in best markets GEO is the only reasonable orbit choice. Other orbits aren't over the receiver continuously; storage is not competitive A final question is, can space cell cost equal cost of cheap terrestrial cells? This is yet to be demonstrated. Space cells have stringent requirements. ACKNOWLEDGEMENT An earlier version of this was presented at the MIT Space Solar Power Workshop, May 15-17, REFERENCES 1. P.E. Glaser, Power from the Sun: Its Future, Science Vol. 162, (1968). 2. P.E. Glaser, F. P. Davidson, and K. I. Csigi (eds.), Solar Power Satellites: The Emerging Energy Option, John Wiley & Sons Inc (June 1994; second edition 1998). 3. NASA, Satellite Power System Concept Development and Evaluation Program System Definition Technical Assessment Report, U.S. Department of Energy, Office of Energy Research, Report DOE/ER/ , Dec U. S. Office of Technology Assessment, Solar Power Satellites, J. Mankins, "A Fresh Look At Space Solar Power: New Architecture, Concepts, and Technologies," paper IAF-97-R.2.03, 48th International Astronautical Congress, October , Turin, Italy. In: Acta Astronautica, Vol. 41, 4-10, August-November 1997, pp J. Howell and J.C. Mankins, "Preliminary results from NASA's Space Solar Power Exploratory Research and Technology Program," 51st International Astronautical Congress, Rio de Janeiro, Brazil, J.E. Dudenhoefer and P.J. George, Space Solar Power Satellite Technology: Development at the Glenn Research Center An Overview, paper NHTC , 34th National Heat Transfer Conference, Pittsburgh, Pennsylvania, August 20 22, NASA Technical Memorandum NASA/TM H. Feingold and C. Carrington, "Evaluation and comparison of space solar power concepts," 53rd International Astronautical Federation Congress. In: Acta Astronautica. Vol. 53, 4-10, August-November 2003, pp doi: /s (03) Proceedings of the SPS '97 Conference, Space Power Systems: Energy and Space for Humanity, Montréal, Canada, August , organized by the Canadian Aeronautics and Space Institute and the Sociéte des Électriciens et Électroniciens France. 11

12 10. G.A Landis, "Space Power by Ground-based Laser Illumination," IEEE Aerospace and Electronics Systems, Vol. 6 No. 6, 3-7, Nov Presented at 26 th Intersociety Energy Conversion Engineering Conference, Boston MA, Aug , Proc. 26 th IECEC, Vol. 1, p N. Marzwell, "The Challenge of Space Solar Power," MIT Space Solar Power Workshop, May 15-17, Swinburne Centre for Astrophysics and Supercomputing, "COSMOS: the SAO Encyclopedia of Astronomy" ( 13. R.L. Forward, "Roundtrip Interstellar Travel Using Laser Pushed Lightsails," J. Spacecraft and Rockets, Vol. 21, No. 2, Mar-Apr 1984, pp See also G.A. Landis, "Photovoltaic Receivers for Laser Beamed Power," 22nd IEEE Photovoltaic Specialists Conference, Las Vegas NV, Oct. 1991, Vol. II, NASA Report CR (1991). 15. G.A. Landis, "Reinventing the Solar Power Satellite," NASA Technical Memorandum TM (2004). 16. G.A. Landis "Re-evaluating Satellite Solar Power Systems for Earth," IEEE 4th World Conference on Photovoltaic Energy Conversion, Waikoloa, HI, May 7-12, G.A. Landis, "An Evolutionary Path to SPS," Space Power, Vol. 9 No. 4, pp (1990). 12