NATIONAL AERONAUTICS AND SPACE ADMINISTRATION Technical Report 32-1573 Photovoltaic Solar Array Technology Required for Three Wide-Scale Generating Systems for Terrestrial Applications: Rooftop, Solar farm, and Satellite Paul A. Berman (NASA-CR-1283 8 1 ) PHOTOVOLTAIC SOLAR ARRAY N72-3 30 6 l TECHNOLOGY REQUIRED FOR THREE WIDE SCALE GENERATING SYSTEMS FOR TERRESTRIAL APPLICATIONS: P.A. Berman (Jet Propulsion Unclas Lab.) 15 Oct. 1972 27 p CSCL 110A G3/03 44953 JET PROPULSION LABORATORY CALIFORNIA INSTITUTE OF TECHNOLOGY PASADENA, CALIFORNIA October 15, 1972 h ~ ' !
Transcript
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION
Technical Report 32-1573
Photovoltaic Solar Array Technology Required forThree Wide-Scale Generating Systems for
Terrestrial Applications: Rooftop,Solar farm, and Satellite
Paul A. Berman
(NASA-CR-1283 8 1 ) PHOTOVOLTAIC SOLAR ARRAY N72-3 3 0 6 lTECHNOLOGY REQUIRED FOR THREE WIDE SCALEGENERATING SYSTEMS FOR TERRESTRIALAPPLICATIONS: P.A. Berman (Jet Propulsion Unclas
Lab.) 15 Oct. 1972 27 p CSCL 110A G3/03 44953
JET PROPULSION LABORATORY
CALIFORNIA INSTITUTE OF TECHNOLOGY
PASADENA, CALIFORNIA
October 15, 1972
h ~ ' !
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION
Technical Report 32-1573
Photovoltaic Solar Array Technology Required forThree Wide-Scale Generating Systems for
Terrestrial Applications: Rooftop,Solar Farm, and Satellite
Paul A. Berman
JET PROPULSION LABORATORY
CALIFORNIA INSTITUTE OF TECHNOLOGY
PASADENA, CALIFORNIA
October 15, 1972
C
Prepared Under Contract No. NAS 7-100National Aeronautics and Space Administration
f
Preface
The work described in this report was performed by the Guidance andControl Division of the Jet Propulsion Laboratory.
Three major Options for wide-scale generation of photovoltaic energy forterrestrial use are considered: (1) Rooftop Array, (2) Solar Farm, and (3)Satellite Station. The Rooftop Array would use solar cell arrays on the roofsof residential or commercial buildings; the Solar Farm would consist of largeground-based arrays, probably in arid areas with high insolation; and theSatellite Station would consist of an orbiting solar array, many square kilo-meters in area. The Technology Advancement Requirements necessary foreach Option are discussed, including cost reduction of solar cells and arrays,weight reduction, resistance to environmental factors, reliability, and fabrica-tion capability, including the availability of raw materials. The majority ofthe Technology Advancement Requirements are applicable to all three Op-tions, making possible a flexible basic approach regardless of the Optionsthat may eventually be chosen. No conclusions are drawn as to which Optionis most advantageous, since the feasibility of each Option depends on thesuccess achieved in the Technology Advancement Requirements specified.
JPL TECHNICAL REPORT 32-1573vi
Photovoltaic Solar Array Technology Required forThree Wide-Scale Generating Systems for
Terrestrial Applications: Rooftop,Solar Farm, and Satellite
I. IntroductionIn a previous report (Ref. 1) the author discussed
general areas for investigation that could significantlyreduce the cost of photovoltaic solar arrays to be usedfor wide-scale terrestrial solar-to-electrical power gen-eration. In the present report, the same considerationshave been used with a broader philosophical scope tosuggest an approach for solar array development ap-plicable to three major 'systems for using solar photo-voltaics to generate electrical power for terrestrial pur-poses. These systems include two Earth-based concepts(Rooftop Array and Solar Farm) as well as a SatelliteStation.
The principal Technology Advancement Require-ments to attain a viable solar array for wide-scale ter-restrial use are (1) a reduction in the dollars-per-wattcost of solar arrays, as fabricated and installed, by ap-proximately three orders of magnitude over those pres-ently experienced in the space program and (2) a dras-tic increase in production capability of such arrays and
systems, so that many square miles of arrays can befabricated and installed in a routine manner. ThreeOptions for achieving the objectives of wide-scale ter-restrial power generation by photovoltaic systems aredefined in the following section. Each of these Optionshas major Technology Advancement Requirementsother than those associated with the solar arrays. Op-tions 1 (Rooftop Array) and 2 (Solar Farm), for exam-ple, require an efficient means of energy storage for atruly self-contained energy generation system (however,this will not be required if photovoltaic energy con-version is to be used as a supplementary power sys-tem). Option 3 (Satellite Station) requires a method oftransmitting and converting the generated power intoa usable Earth-based power distribution center andtechnology for the insertion, deployment, and main-tenance of large area solar arrays in synchronous orbit.The solar array problem, however, is common to allthree Options and the success of any or all Optionsis predicated upon satisfaction of the Technology Ad-vancement Requirements for the solar array. TheseTechnology Advancement Requirements are describedin detail in this report.
JPL TECHNICAL REPORT 32-1573 1
II. Definition of Three OptionsThe three Options to be discussed in this report
are defined as follows:
Option 1. Rooftop solar generatorResidential and commercial buildings(Rooftop Array)
Option 2. Large-area photovoltaic solar generatorLarge flat land areas(Solar Farm)
Option 3. Satellite solar power stationLarge arrays in synchronous orbit(Satellite Station)
These Options are described below.
A. Rooftop Array
The Rooftop Array is an Earth-based system usingthe roofs of buildings, both residential and commercial,as an area upon which solar cell arrays are to bemounted. The arrays then provide electrical power forthe building in either a supplementary manner, beingaugmented as needed by power from more conven-tional sources, or an independent manner by meansof suitable electrical energy storage devices. The latterapproach requires (1) a considerable effort to econom-ically produce energy storage devices with the re-quired storage capacity and minimum maintenance re-quirements, and (2) a probable major effort in im-proving the utilization of electrical power so as todrastically reduce the overall electrical power demandsof the building (because array area and hence gen-erating capacity is limited). The supplementary ap-proach would circumvent these requirements, butwould impose a requirement for essentially a dual (re-dundant) system to supply the differential betweenthe power that is available at the moment and thepower that is needed. This dual system requirementis economically less attractive, since the costs assoc-iated with the backup system (generators, transmis-sion lines, maintenance, etc.) must be considered aspart of the total system costs. Both the supplemen-tary and independent approaches would, however, sig-nificantly reduce the demands on natural resourcesand the pollution by-products (particulate, radioactive,and thermal) associated with the conventional meansof electrical power generation.
B. Solar Farm
The Solar Farm, the large-area Earth-based centralgenerating station of Option 2, has many of the charac-teristics of the Rooftop Array. If the Solar Farm is tooperate in an independent mode, a low-cost, low-main-tenance storage system is also required. More optionsmight exist, however, for the large storage systems ofthe Solar Farm (e.g., pumped-water storage dams) thanfor the small storage system of the Rooftop Array.Also, because of the "sharing" nature of the generatedpower, the demand loads would be somewhat smootherthan the sharper demand peaks and valleys of the in-dividual systems of the Rooftop Array. In the supple-mentary approach, the smoothing effect of user shar-ing associated with this Option would also be more ad-vantageous, but the cost disadvantage of maintaininga redundant backup system would still exist.
The Solar Farm, or Farms, interconnected, wouldrequire transmission lines to the user for both theindependent and supplementary approaches, whereasthe independent approach for the Rooftop Array wouldhave no such requirement. In both the Earth-basedsystems, improvements in the efficiency with whichelectricity is used (energy conservative systems) wouldgreatly enhance the success probability of the inde-pendent approach since it would reduce the require-ments imposed on the power generating and storagecapacity of the system. For the Rooftop Array, thegenerating capacity is expected to be a limiting fac-tor, while for the Solar Farm, the storage capacitymight be a limiting factor.
C. Satellite Station
The large-scale space satellite generating station ofOption 3 has probably received the greatest amountof study of the three Options listed (Refs. 2-7), mostlikely because the system has some extremely tanta-lizing advantages, although it represents an almostscience-fiction-like undertaking in resource commit-ments and technology advancements. It is probablyonly due to NASA's commitment to the Space Shuttle,with its expected drastic decrease in payload costs,that Option 3 can be considered at all since the weightof the proposed system will be 18-45 million kg (40-100 million lb)!
The major advantages of the Satellite Station are:
(1) The system receives full sunlight, unattenuatedby atmospheric absorption (including clouds) forall but a 1.2-h interval every 24 h for 25 days
JPL TECHNICAL REPORT 32-15732
before and after equinox in the 35,600-km (22,300-mi) synchronous orbit proposed. Thus, by usingmultiple stations, the need for storage systemsis eliminated, even for the independent approach.
(2) Since, in the most optimistic case, the Earth-based systems will receive less than an averageof 6 usable sun hours per day (and even this isattenuated by atmospheric absorption of usablephotons) while the Satellite Station will receiveclose to 24 h of (unattenuated) sunlight per day,the solar array area of the latter system wouldbe only 10-20% that of an Earth-based arrayproducing equivalent power.
(3) While the Earth-based arrays must take into ac-count such factors as wind, dust, sand, precipi-tation, etc., the Satellite Station arrays wouldnot be exposed to such conditions, although theywould be subject to ionized particle irradiation.
The problems associated with the Satellite Station,aside from those discussed in detail in the followingsection on Technology Advancement Requirements,are in transportation (e.g., launch and insertion intoorbit) of the system, which weighs 18-45 million kilo-grams (40-100 million pounds), erection and mainte-nance (e.g., attitude control) of the system, which pres-ently entails 33 km2 (13 mi2 ) of array area plus 133km2 (52 mi2 ) of solar concentrator area, and the trans-fer of the generated power to usable Earth-based pow-er (e.g., through direct current-to-microwave conver-sion, transmission, collection, and reconversion).
The discussion above makes no attempt to suggestadoption of any one of the Options, but is intendedto supply some overall perspective. Each of the Op-tions has many advantages and disadvantages, andselection of the most desirable Option must await theresults of a concentrated effort to attain the requiredtechnology advances and an assessment of the degreeof success achieved in accomplishing these advances.
III. Rationale for Concentration onSolar Array Development
Each of the three defined Options, integrated intoa viable system, is quite complex, the complexity in-creasing according to the Option number assigned (i.e.,the Rooftop Array is least complex, the Satellite Sta-tion, most complex); however, all three Options haveone common primary technology requirement, namely,
large scale, economical fabrication of reliable solararrays. If this primary need is not satisfied, none ofthe Options is viable.
Over the past 10 years, the economic viability ofsolar photovoltaics has been evaluated several times;however, the divergence between the optimistic andpessimistic evaluations is about three orders of magni-tude. At the recent Ninth Photovoltaics Specialist Con-ference of the Institute of Electrical and Electronics En-gineers, still another series of economic forecasts waspresented. Once again there was a discrepancy ofseveral orders of magnitude between the optimisticand pessimistic forecasters. It therefore appears thatthe forecasting of the eventual cost for photovoltaicsolar power conversion has not come very far in thelast decade, no matter how wise in the ways of econ-omics the forecaster may be (applying such factors asamortized capitalization investment, inflationary trendson interest rates, etc.). There is really no appropriatecost data to use, since the fabrication of photovoltaicsolar arrays has been, without exception, on an ex-tremely small scale, nonautomated (in an industrialcontext), high-reliability, custom-made basis. Hence,the first order of business is to provide the requiredcost information.
Since financial and personnel budgets are alwaysfinite, and in fact usually quite limited, it is impera-tive to direct the available resources into areas thatwill provide the greatest number of options. Indeed,this was NASA's philosophy in supporting the SpaceShuttle, which will provide a tremendous number offuture options, at the expense of some interesting butvery specific missions. This rationale should also pre-vail in developing technology needed for electric pow-er from solar energy.
The major technology effort over the near termshould therefore be the development of materials, tech-niques, and processes for large-scale, low-cost produc-tion of photovoltaic solar arrays. This emphasis wouldmaintain the greatest program flexibility by concen-trating on the solar array aspects until it is establishedwhich Technology Advancement Requirements can besatisfied and the relative economics of doing so. Thisinformation would then be used to determine whichof the three defined Options are feasible. Parallel low-level feasibility studies on the other critical aspectsof the three Options could also be begun at this time,or at some time in the future as determined by a mile-stone event in the solar array program (that is, at
JPL TECHNICAL REPORT 32-1573 3
such time as certain feasibility criteria are met). Suchparallel studies might consist of:
(1) Methods of energy storage (Options 1 and 2).
(2) Conversion to microwave power (Option 3).
(3) Microwave beam forming and collecting (Op-tion 3).
(4) Conversion to commercial electrical energy (allOptions).
(5) Launch and assembly in space including tele-operators (Option 3).
(6) Station keeping and attitude control (Option 3).
(7) Maintenance in space (Option 3).
(8) Environmental/ecological effects of systems (allOptions).
A document entitled Solar Energy R & D PolicyAssessment, prepared by E. L. Ralph, Heliotek Divi-sion of Textron, for submission to the Committee onEnergy R & D Goals established by the Federal Coun-cil on Science and Technology presents one set ofpossible costs and time schedules for various phasesof the three Options defined above. These are sum-marized in Table 1 and indicate a greater expenditureof money than the present author would have estimated.As early as Phase B, expenditures of $60 million foreach of the three Options is estimated by Ralph.Phase C requires $300 million for Options 1 and 2 andover $2 billion for Option 3. Phase D requires $1 bil-lion for Options 1 and 2, and $8 billion for OptionC. Therefore it is obvious that, at this time, one willhave to pick and choose areas of investigation carefullyso as to make the greatest impact for the least expendi-ture of money and manpower. Again, it is the presentauthor's belief that this can be accomplished by di-recting our resources toward development of the solararray.
One of the major objectives of this report is to showthat most of the solar array Technology AdvancementRequirements are at least qualitatively important toall three defined Options, although they may varyin priority among the Options. If this is indeed thecase, a very logical development program can then
be generated to maintain flexibility with respect toall three Options. Such a program would generate abaseline series of solar arrays, in an iterative process,using the most sophisticated applicable technologiesavailable at the time of each iteration. The resultsobtained through analysis of these baseline arrays willprovide information on the cost-vs-performance trade-offs associated with materials and process modifica-tions, based on the boundary conditions associatedwith each of the three Options. That is, what perform-ance penalties must one pay in accepting a less sophis-ticated technology, and what, if any, are the economicgains in doing so? The magnitude and direction (i.e.,positive or negative) of the resultant balance mightbe different for each of the Options. This method-ology is the only way in which one can ascertain, forexample, whether an inexpensive process for solar cellcontact deposition really gives rise to an economicallymore attractive system than a more controlled and ex-pensive process. The Option 3 (Satellite Station) canbe expected to require the most technically sophisti-cated array because of the significant cost of the otheraspects of the system and the more limited amenabilityto maintenance and replacement.
IV. Technology Advancement Requirements
A. Background
The techniques for fabricating and installing very-high-reliability photovoltaic solar arrays are well known.Photovoltaic solar arrays have been the backbone ofelectrical power generation for almost all unmannedspacecraft and have successfully operated in near-Earthspace and at distances closer to the Sun than Venusand further away from the Sun than Mars, for extendedperiods of time. In these cases, however, the cost wasonly a secondary consideration, while reliability andend-of-mission performance were the major criteria.Each mission, or series of missions, used custom-builtarrays designed by engineers who felt that their par-ticular design was the only one that would satisfythe mission requirements. This resulted in dollars-per-watt ratios of between $200 and $1000 per watt (de-pending upon what factors are put into the deriva-tion of the ratio: for example, development costs, typeapproval models, documentation, orientation mech-anisms, etc.). This situation is analogous to spending$200,000 to develop and fabricate an automobile thatwill win a specific, highly competitive race as op-posed to a $2000 compact automobile capable of get-ting its owner down to the corner grocery store.
JPL TECHNICAL REPORT 32-15734
Table 1. Projected program costs and schedule for three Optionsa
Schedule,Cost, $Mb yearsyears
Option 2Solar Farm
Schedule,Cost, $M yearsyears
Option 3Satellite Station
Schedule,Cost, $M yearsyears
Pre-phase APreliminary design and feasibilityassessment. Conceptual design ofalternative approaches. Identificationof critical system parameters.
Phase AEstablish system feasibility and mostdesirable approaches. Assessment oftechnical advances needed. Grosscost and schedule projection.
Phase BPreliminary design of preferred system.Detail assessment of requirementsincluding resource, manufacturing andtest requirements. Preliminary systemcost and schedule projection. Allprecommitment objectives completed.
Phase CFinal definition: Freezing of concepts,approaches, designs, schedules, andcosts of program. Intensive develop-ment of operational systems. Initiationof testing.
Phase DFinal development and operationalphase. Operational system componentsdeveloped. Demonstration hardwarefabricated and extensively tested.Prototype built. Commercial readinessachieved and competitive positionascertained.
None
None
60
300
1,000
None
None
3
3
5
0.1
3
60
300
1,000
1.5
1.5
3
3
5
3
20
60
2,100
1.5
1.5
3
3
8,000 8
Total 1,360 11 1,364 14 10,183 17
aExtracted from Research Paper No. 135, Heliotek Div. of Textron, Inc., "Solar Energy R & D Policy Assessment" by E. L. Ralph.b$M = millions of dollars.
The overall requirements, then, are those that willenable us to reduce the $1000-per-watt figure down toa more reasonable $1.00 per watt and the cost persquare meter of array from $100,000 to $100.00. Fur-thermore, technology needs to be developed to fabri-cate not square meters of array but square kilometers,and the generation not of watts but of 10,000s of mega-watts. This could result in a major impact on exist-ing industries or the creation of new industries. Forexample, to cover 2.6 km 2 (1 mi2 ) of area with single-crystalline silicon cells 0.25 mm thick, assuming a wast-
age of 50% of the silicon (which, unfortunately, is op-timistic at present), represents six times the yearlyproduction of such silicon in the United States (Ref.8). (This of course assumes that one wants to use single-crystalline silicon in the fabrication of the arrays, whichmay not be true.) Thus the required Technology Ad-vances are formidable, but the very fact that the dol-lars-per-watt numbers are so high and that so littlehas been done to significantly reduce them can onlymake one optimistic that progress can be made to-ward this end.
JPL TECHNICAL REPORT 32-1573
Phase
Option 1Rooftop Array
5
B. General Considerations
In the past, cost reduction has been of only second-ary importance in the space program, where the capa-bility of the solar array to successfully satisfy the mis-sion power requirements within the constraints of themission boundary conditions has been of prime im-portance. Fabrication capability, that is, the capabil-ity of fabricating very large quantities and areas ofsolar arrays, has also not been a major concern (e.g.,there has been no need for concern that there wouldnot be enough single-crystalline silicon produced inthe United States to satisfy the needs of the space pro-gram). Weight reduction, likewise, has not been of primeimportance, although recent development efforts havebeen extended toward achieving significant weight re-duction because of mission requirements such as thoseproposed for ion propulsion utilization. Oddly enough,the advent of the Space Shuttle, which would signifi-cantly reduce the payload cost, might actually inhibitsignificant further efforts in weight reduction of solararrays for the normal class of missions. This, of course,would not be the case for the Option 3 satellite powersystem, which requires 33 km2 (13 mi2 ) of solar arrayarea plus 133 km2 (52 mi2) of concentrator area witha proposed system weight of 18-45 million kg (40-100 million pounds). Thus, in this case, transportationcosts are expected to be a significant portion of thetotal system cost.
The fact, then, that major efforts have not been ex-pended to provide improvement of the factors so criti-cal to the viability of the photovoltaic system forwide-scale terrestrial use certainly does not detractfrom the expectation that required improvements canindeed be achieved. In addition, basic technologiesappropriate to fabrication of solar arrays have alreadybeen established and the problem areas clearly deline-ated, so that there is a firm baseline from which toextrapolate toward major improvements. The Tech-nology Advancement Requirements previously dis-cussed are of an engineering nature and do not re-quire fundamental breakthroughs. Therefore, immedi-ate and significant near-term improvements can beexpected to occur. A fundamental breakthrough inthe area of the photovoltaic conversion, for example bydevelopment of a high-efficiency, thin-film gallium ar-senide cell or an organic photovoltaic converter,would most certainly increase the probability of suc-cess of photovoltaics for terrestrial applications, butsuccess does not at this time appear to hinge uponsuch developments.
Table 2. Relative priority of technical requirements and needs(highest priority = 1)
Option 1 Option 2 Option 3Technology Advancement Rooftop Solar Satellite
The specific Technology Advancement Require-ments are listed in Table 2 and ranked in order ofpriority with respect to the three Options previouslydefined. The priority is defined as increasing with de-creasing numerical value; that is, priority 1 is of great-est importance, priority 4 is of little importance. TheTechnology Advancement Requirements and the pri-ority ranking are discussed below.
1. Cost Reduction. Low cost is a prime criterionfor Options 1 and 2, the two Earth-based systems.Some sacrifices in the other areas of high power den-sity and light weight can be tolerated if costs can bekept very low. For Option 3, cost is important for
JPL TECHNICAL REPORT 32-15736
the solar cell modules and arrays but not as impor-tant as achieving high power density and light weight,because of the costs involved in other aspects of thesystem, such as transportation and insertion into orbit,erection and maintenance of the array, conversion tomicrowave power, beam forming, microwave collection,reconversion to electrical power, attitude control, etc.Reference 1, which discusses at length the studies andtradeoffs that could significantly reduce the cost ofsolar arrays, was primarily directed at Earth-based sys-tems (e.g., Options 1 and 2) to be used as supplementalpower and therefore not requiring an energy storagesystem.
At present, as previously mentioned, solar arrayscost about $200 to $1000 per watt, produce about 90W/m2 and about 4 W/kg. If we take the higher fig-ure, this results in about $90,000 per square meter.Of this, the solar cells represent about 10-20% of thecost, or about $100 per watt for the cells alone. Witha little imagination, one can easily see the array costs,exclusive of the solar cells, decreasing three orders ofmagnitude from approximately $90,000 per square meterto about $90 per square meter by using a substrateof Kapton or some even less expensive material withprinted circuits for interconnections deposited ontothe substrate in an economical manner. Reduction ofthe solar cell costs by three orders of magnitude re-quires somewhat more imagination, especially if weconstrain ourselves to the use of single-crystalline sili-con, which presently costs about $0.30 per gram, andwhich in ingot form (cylindrical in geometry) resultsin wastage of 75% or more of the silicon by the timeit is cut into rectangular blanks having a thickness ofapproximately 0.3 mm. Thus, in the future, one canenvision the cost of the solar cells as being the princi-pal contributor to the cost of the solar array.
The costs of solar arrays must be reduced by ap-proximately three orders of magnitude from the pres-ent level. To this end, careful consideration must begiven to the areas discussed below.
a. Higher-efficiency cells. High efficiency, or, as acorollary, high power density, is obviously importantfor all Options, since it directly affects the array sizeand weight necessary to achieve a specified poweroutput. It is most important, however, for Option 3(Satellite Station) since even with an efficiency of 18%(based on normal measurement conditions), the solararray alone is proposed to be 33 km2 (13 mi2 ), with133 km2 (52 mi2 ) of solar concentrator area (Ref. 7).Present-day solar cells have an efficiency of only about
11.5% and a thickness of about 0.3 mm, as opposed tothe 0.05-mm-thickness cells proposed for Option 3. Atpresent, cells of this latter thickness have never beenmade, but even cells having a thickness of 0.1 mmexhibit significantly decreased efficiency (about 9-10%)so that Option 3 is predicated on an approximate 100%improvement in solar cell efficiency. High-efficiencyarrays also have a high priority for the Rooftop Arraysince the area of this system is limited and thereforethe highest power density is desired to meet the user'sneeds. It is assumed that for Option 2 (Solar' Farm),the area would not be critical, since land that is notuseful for any other purpose might be used. Thismight well be the case, because land areas havingthe greatest insolation are usually arid and not am-enable for farming or even desirable for living.
It has been tacitly assumed that the solar cells fromwhich the arrays are fabricated would be made fromsilicon; however, the band gap of gallium arsenide issuch that higher theoretical efficiency could be ob-tained, although investigations in the early 1960s didnot prove this to be true in practice. Gallium arsenide,being a direct-band semiconductor, has a very sharplight absorption edge, and all the usable hole-electronpairs are created in one or two micrometers of material,as opposed to silicon, which absorbs usable light atdepths greater than 200 micrometers. The techniquefor fabricating thin-film gallium arsenide cells, for ex-ample by vapor-phase or liquid-phase epitaxy, mightresult in a very-high-efficiency, ultra-lightweight cellwhich, because of the small amount of material re-quired per unit area, might also be very economical.This area of investigation could be quite costly andrequire a good deal of time to pursue but, if success-ful, could have a favorable effect upon all three Op-tions, especially if high power density is required.Furthermore, with respect to the Satellite Station, sincesolar concentrators are an integral part of the system,with their attendant increase in cell operating tem-perature, gallium arsenide with its more desirable high-temperature characteristics presents an additional ad-vantage, as would also be the case if concentratorswere to be used in the Earth-based systems.
In all cases, the cost of fabricating higher efficiencycells must be traded off against the overall systemscosts. For example, for Option 2 (Solar Farm), it mightbe advantageous from a systems point of view to ac-cept a lower efficiency and utilize a far less expensivegrade of silicon to achieve a lower overall dollar-per-watt figure.
JPL TECHNICAL REPORT 32-1573 7
Higher silicon solar cell efficiencies can be achievedby improving the materials and processing involvedin the fabrication of the cells, such as (1) fabricationof very shallow, high-quality electrical junctions, (2) de-crease of surface recombination velocity through care-ful blank preparation, diffusion, and surface passiva-tion, (3) utilization of low-resistivity silicon with highminority carrier diffusion length, combined with fabri-cation processes that do not adversely affect theseparameters, and (4) improvement of reflection proper-ties of the portion of the array blanket between thejunction and the incoming solar energy. Improvementsin efficiency might also be obtained through the useof thin-film gallium arsenide photovoltaic converters,but this would require a major research effort.
b. Lower cost cells. In general, cell costs must bedecreased by approximately three orders of magni-tude. A significant portion of these cell costs is associ-ated with the use of ultrapure, single-crystalline sili-con, which presently costs approximately $0.30 pergram, and which, in cylindrical ingot form, results inwastage of more than 75% of the silicon by the timeit is cut into rectangular blanks having a thickness ofapproximately 0.3 mm. Ingot utilization can be im-proved by making use of the natural cylindrical geo-metry of the ingot to fabricate disk-shaped cells, whichcould be utilized with individual conical concentratorsto achieve still greater cost effectiveness.
A second technique for better silicon utilization, nowbeing investigated*, is a process capable of growingsilicon ribbons having the proper thickness and aerealdimensions for fabrication of large area solar cells. Inthis process, there would be no wastage of the siliconmaterial in the slabbing and cutting operations, andno loss of silicon that could not be cut into rectangles.Furthermore, the silicon would not require lappingand etching to remove mechanical damage inducedby the cutting and lapping operations but would havea high-quality, damage-free surface. The use of largearea blanks can also be expected to reduce the overallcosts of the cell on a unit area basis if the cell pro-cessing is modified to optimally integrate the largearea blanks.
Another possible cost reduction for the Earth-basedsystems would be to reduce the purity of the siliconused in the solar cells (Ref. 8). The single-crystalline
*"Development of Thick Film Silicon Growth Techniques," JPLContract No. 953365 with Tyco Laboratories, Inc., Waltham,Mass. Contract initiated on Feb. 17, 1972.
8
silicon costs about $300 per kilogram. High-purity sili-con costs about $10 per kilogram, and low-purity metal-lurgical silicon costs only about $0.45 per kilogram.As the quality of the silicon is reduced, the efficiencyof the resultant cell can also be expected to be reduced,so that the investigation of using lower-quality siliconfor fabrication of solar cells would probably not beapplicable to the Satellite Station, which requires veryhigh efficiency. However, the reduction of the materialcosts between one and two orders of magnitude couldhave a favorable impact on the array costs of the Earth-based systems if, as assumed, the array costs becomeso low that the cost of the cells determine the costof the array.
A fourth possibility for decrease in the cost of thecells is the investigation of thin-film solar cells. Twomaterials, cadmium sulfide and gallium arsenide, comeimmediately to mind, with a third possibility beingcadmium telluride. Significant effort has already beenexpended on cadmium sulfide, with only limited suc-cess (Refs. 9-11); however, one wonders whether thisconcept should be completely abandoned. The majorreasons for disenchantment with cadmium sulfide werecell instabilities. One instability could probably beavoided by proper sealing or encapsulation of the cellduring fabrication. The second instability was morefundamental and due to the mobility of the copper inthe cell, which is really a Cu, CdS cell. Whether thisinstability can be controlled is debatable, but possiblyfurther low-level funding would not be inappropriate.Very little work has been done on gallium arsenidesince the early 1960s and very, very little work hasbeen done on thin-film gallium arsenide (Ref. 12). Asdiscussed previously, gallium arsenide has a band gapmore theoretically optimal to photogeneration by solarenergy and absorbs almost all usable photons withinone or two micrometers. Thus, a thin-film gallium ar-senide cell could have a very high efficiency, andsince little gallium arsenide material would be usedper unit area, it could be highly economical. As statedabove, however, a considerable research effort wouldbe necessary before feasibility could even be deter-mined, and with only limited resources one might notconsider this to be an 'appropriate allocation of sig-nificant resources at this time, especially if very largecost reductions can be achieved in the reasonably nearfuture by improved array fabrication techniques.
In general, the use of highly automated, large-batchor continuous-belt processing techniques is requiredfor the low-cost production of solar cells, and this isa major Technology Advancement Requirement.
JPL TECHNICAL REPORT 32-1573
C
c. Improved array fabrication techniques. Fabrica-tion of solar arrays for space use has been, without ex-ception, on a very-small-volume, custom-built basis,and it has therefore been uneconomical to pursue ser-iously the high-volume, automated manufacturing tech-niques that are so prevalent throughout industry ingeneral. Furthermore, cost considerations have notplayed a major role in the design of space-type solararrays, where highest priority has been given to en-suring the reliable operation of the array within ratherdiverse mission constraints. Therefore, the major costreductions can be expected to be achieved by im-proved fabrication techniques in the near term, witha significant effort in production and manufacturingengineering, especially since the array, exclusive ofthe cost of the solar cells, now account for 80-90% ofthe total cost.
The improvement of array fabrication techniques willrequire materials ahd fabrication investigations for sub-strates, printed circuitry, wiring, module interconnec-tion, and cell laydown, and a considerable effort mustbe placed on automating the processes involved. In-expensive techniques for applying the protective layeror coverglass directly onto the cell, or, even better,the completed array, would greatly simplify the pro-cess. Automated pulse soldering techniques or parallelgap welding are good candidates for performing inter-connections. Techniques for inexpensive cell laydownonto the substrate should be developed, and one shouldthink not only of bonding by means of adhesives andepoxy, but also of possible mechanical attachment,perhaps using the interconnections themselves to holdthe cells to the substrate. Work at the NASA LewisResearch Center on encapsulating cell modules inFEP Teflon (Refs. 13 and 14) also appears to be en-couraging and should probably be pursued with greateremphasis.
The use of large-area cells (for example, cells fabri-cated from large-area thick-film silicon ribbons) wouldsignificantly reduce the number of required intercon-nections (and also yield higher packing factors) andwould therefore reduce the complexity of array fab-rication.
So many options appear to be available for fabricat-ing arrays more economically than is done at presentthat one would be surprised if effective cost reductionscould not be made.
d. Large-area cells and arrays. The use of large-area cells presents an overall cost advantage in array
fabrication if the cells on a unit basis are no more ex-pensive than the smaller area cells, because the celllaydown and interconnection for fewer large-area cellsshould be significantly less expensive than for a greaternumber of small-area cells. Where large-area cells would,of course, also present significant advantages for thefabrication of the Option 3 (Satellite Station) arrays, inthis case it would be expected to be of secondary im-portance, as against requirements of high power densityand light weight that are critical for Option 3. Further-more, the present techniques for cell blank sizing willresult in a tradeoff between minimizing thickness andmaximizing area due to breakage factors. Large areaarrays, on the other hand, have prime importance forOptions 2 and 3, where many square kilometers ofarray area would be required. For Option 1, the areaof the array would have to be only large enough tocover the roof. Three methods for fabricating large-area cells are as follows:
(1) Large-area cells can be achieved by cutting thesilicon ingot into large disk-shaped blanks, ratherthan rectangular blanks, as is presently the case,thus making use of the natural cylindrical geom-etry of the ingot, which can be grown with dia-meters as large as 7.6 cm. These cells, however,having a poor panel packing factor comparedwith that of rectangular cells, would probablynot be appropriate for Option 3 and possiblynot for Option 1.
(2) Large rectangular cells can be obtained by a sec-ond method, namely, by slicing the ingot so thatthe major axis of the rectangle is parallel to theingot growth axis, thus making use of the lengthof the ingot rather than the width (which is ofsmaller dimension).
(3) The third, and most tantalizing, method for achiev-ing large-area cells is growth of large-area, rec-tangular-shaped silicon ribbons. This latter meth-od has the further advantage of allowing verythin large-area cells to be fabricated; this is nottrue for method 2, which requires a tradeoffamong the breakage, thickness, and area associ-ated with the process. Furthermore, method 2involves yet another tradeoff: the size of the blankas against ingot utilization, because of the geo-metric constraints associated with cutting rec-tangles from cylinders. This, of course, wouldnot be true for method 3.
JPL TECHNICAL REPORT 32-1573 9
e. Use of concentrators. One way of avoiding the ex-pense of fabricating solar cells and arrays is to use in-expensive concentrators that would significantly reducethe cost per resultant power output unit. The propo-nents of Option 3 (Satellite Station) propose using aconcentrator system to achieve a 3-to-1 concentrationratio (Ref. 7). This results in the need for developinga cold-mirror concentrator (to minimize array tempera-ture) that can be fabricated, transported, and erectedat a cost less than 1/4 that of an installed solar cellarray of the same area (since the concentrator systemproposed for the Satellite Station is four times the areaof the array and results in a net power increase ofabout 100%).
Because of the temperature rise associated with con-centrating the solar energy in this manner for Option3, and the fact that solar cells decrease in efficiencyas temperature is increased, it is estimated that thecell efficiency will drop from the assumed 18% to 11.7%.The decision to use such a system implies that thecost penalties involved in fabricating and erecting asolar concentrator four times the area of the array andthe efficiency penalty due to the cell and array heat-ing are more than offset by the 100% increase in powergenerated by the array.
The desirability of using concentrator systems withsolar arrays for the Earth-based systems (Options 1and 2) is apparently less clear-cut. It was suggestedby the author in the early 1960s and later in 1966(Refs. 15-17) that the use of simple, inexpensive, coni-cal concentrators in conjunction with disk-shaped cellsfabricated from a centerless-ground silicon ingot (tak-ing advantage of the natural cylindrical geometry ofthe ingot) could achieve significant reduction in costper watt of a solar array system. This certainly appearsto be the case for light normally incident upon thecell surface. In actual use, however, the light wouldnot be necessarily normally incident unless a solartracking system is utilized. Thus, a tradeoff studywould be required to determine the cost effectivenessof solar concentrators for optimally oriented station-ary systems as against solar-oriented systems that wouldtrack the Sun, and both of these against unconcentratedsystems. This tradeoff study would probably providedifferent answers for different geographical locations,and the optimal concentration ratio could also varywith geographic location. To obtain a valid answer, areasonable estimate of the relative costs of the solararray and the concentrators would be required, andthese numbers are simply not available at this time.In the extreme cases, the use of solar concentrators
could result in a less cost-effective system than an un-concentrated system or, conversely, the use of solarconcentrators could be the only mechanism by whichsolar photovoltaics could be economically competitivewith the more conventional means of electrical powergeneration. It thus appears that more careful scrutinyis required for the use of solar concentrators (some ad-ditional details are given in Ref. 1).
Some effort should certainly be devoted to the useof solar concentrators systems for Options 1 and 2;it is already being considered as an integral part ofOption 3. Thus the relative priority is ranked as 1 forOption 3 (Satellite Station) since it is an integral partof the proposed system and ranked either 2 or 4 forthe Earth-based systems, depending on whether con-centrator systems are required to render the Earth-based systems economically viable.
If concentrator systems are indeed determined tobe useful, the arrays must be kept compatible withsuch systems. For the Satellite station, this means pri-marily that the additional heating that results fromthe solar concentration does not adversely affect thearray components or the array itself. Again, for theEarth-based systems, the situation is more complex.Not only must the array and components be compat-ible with the increased temperatures, but the impactof the concentrators on the array performance with re-spect to sand, sedimentation, dust, and precipitationmust also be considered, as must the environmentaleffects on the concentrating material. The use of largeconcentrators or smaller individual concentrators is stillanother tradeoff topic to be studied.
f. Orientation mechanismsltechniques. Here, again,the case is more clear-cut for the Satellite Station thanfor the Earth-based systems. For the Satellite Station,it would certainly be advisable to provide solar track-ing and orientation for the solar arrays, so as to achievemaximum efficiency; this is even more important withsolar concentrators. With the Earth-based systems, atradeoff must be made between the costs of providingthe capability for solar tracking, including the implicitrequirement that the array be rigidized in some man-ner, plus the solar tracking and orientation mechan-isms required, as against the costs of an unorientedarray to achieve the same average daily, monthly, oryearly power output from the array.
At this time; and for some time to come, this ques-tion cannot be answered, since the cost associated withall the aspects that need to be considered are com-
JPL TECHNICAL REPORT 32-157310
pletely unknown. The answer might also be expectedto vary according to geographic location. One aspect,however, does seem to be intuitively obvious: if it isdetermined that even an unconcentrated solar arraywould require provision for solar tracking and orien-tation of the array, then it is reasonable that the great-est cost effectiveness can be achieved by using inex-pensive solar concentrator systems to enhance poweroutput per unit cell. Here, again, the Earth-based sys-tems rank this Technology Requirement either 2 or4 for reasons similar to those outlined above.
2. Weight reduction. This is of greatest importancefor the Satellite Station, in order to reduce the trans-portation system costs for putting the array into orbit.At present, these costs are about $700 per kilogram,which of course would be greatly reduced by the useof the Space Shuttle. The costs of transportation andinsertion into orbit, however, are still expected to beof major concern since the station weight is estimatedto be 18-45 million kilograms (40-100 million pounds).For Option 1 (Rooftop Array), it is simply a matterof logistics, that is, getting the array up on the roofand deploying it, so that light weight would be de-sirable. For the Solar Farm, weight is not expectedto be so critical a parameter, and the low-cost re-quirement would probably dictate a reasonably lowweight in any case.
The proponents of the Satellite Station estimate thata weight reduction by a factor of about 50 is neces-sary to ensure the feasibility of the economics involved.The Earth-based systems are less stringent as to thetechnology need for light weight, but even in thesecases, weight reduction is desirable for transportationand installation of the arrays. Whereas current Marinertechnology utilizes solar arrays capable of approxi-mately 22 W/kg, and roll-out array feasibility of 66W/kg has been demonstrated, the literature on satel-lite solar power stations suggests that this figure mustgrow to 950 W/kg. Technology needed to approach sucha requirement includes higher-efficiency cells, lighter-weight substrates and mechanisms, and cells that canbetter resist the detrimental effects of ionized particleradiation.
a. Higher-efficiency cells. At present a Mariner-class solar cell array weighs 6.5 kg/m2 . The use ofsimilar weight, higher-efficiency cells would propor-tionately decrease the panel area and therefore theweight required to achieve a given power output. Whilethe absolute value of weight reduction would not beso great for lightweight arrays, the percentage weight
reduction would be the same, assuming the higher-efficiency cells weighed no more (i.e., if the cell ef-ficiency were doubled, the array area and weightwould be halved). Furthermore, as lightweight arraydevelopment proceeds, the cells could become thedominant weight factor, and it would therefore be im-portant not only to achieve higher cell efficiencies,but also lower cell weight, primarily by reducing thecell thickness and eliminating requirements for solder.Higher cell efficiencies will be achieved by methodspreviously discussed.
b. Radiation resistance. The radiation resistance re-quirement is applicable only to the Satellite Station.The rationale for this Technology Requirement is iden-tical to that discussed for high-efficiency cells, sinceradiation degradation directly affects the power-pro-ducing capability of the array and hence the requiredarray area and weight. Of particular concern for theSatellite Station would be the effects of solar flares(Ref. 18), which occur sporadically. It has been observedin the past that one major solar proton event can in-troduce as many protons into the near-Earth space en-vironment as would be accumulated in a relativelyquiet 5-year period. Thus a major solar flare protonevent occurring the day after the satellite system iserected in space could result in an immediate de-crease of power output capability that would be verysignificant.
Two primary approaches to minimizing the adverseeffects of radiation degradation are:
(1) Design of solar cells and arrays that are inherent-ly radiation-insensitive.
(2) Design of solar arrays that are capable of an-nealing out damage caused by radiation. Forstate-of-the-art silicon solar cells, this approachrequires a temperature of about 400°C. Becauseof the high emissivity of the faces of the solarpanel, very large power inputs would be re-quired to heat the arrays or array segments tothe required temperature. Furthermore, even ifthe required 400°C temperature could be achieved,significant attention must be paid to matchingthe thermal expansion properties of the'arraystack materials (i.e., substrate, adhesives, cells,contacts, interconnections, and coverglasses) toavoid failure-inducing stresses due to the verydrastic thermal excursions. An alternative to theuse of state-of-the-art silicon cells would be the
JPL TECHNICAL REPORT 32-1573 11
use of lithium-doped silicon cells that exhibitsignificant annealing at temperatures around60°C (Refs. 19-24).
For Approach 1, the fabrication of cells in which thebase region width is considerably smaller than theminority carrier diffusion length would achieve thedesired result. In addition, a shift in the spectral re-sponse of the cell toward shorter wavelengths, by sig-nificantly reducing the depth of the p-n junction,would achieve greater radiation resistance.
Radiation resistance could also be theoreticallyachieved with gallium arsenide solar cells, since mostof the usable photons are absorbed very close to thep-n junction, and hence long minority carrier diffusionlengths are not required to collect these photons. Thisis to be contrasted with the absorption of photons insilicon, which can generate and collect solar-generatedminority carriers as far away as 250 /m from the junc-tion. Since the effect of penetrating radiation is pre-dominantly the significant reduction of base-regionminority carrier diffusion length, the absorption of us-able photons at the junction should, at least in theory,render GaAs cells relatively operationally immune tothe effects of penetrating radiation until very highfluences are accumulated. Again, there is no majorprogram that would develop such cells to technologyreadiness.
On the optimistic side, silicon cells having a thick-ness of only 0.05 mm should also be relatively opera-tionally immune to the minority carrier diffusion lengthdegradation until high fluences are accumulated, sincethe initial (unirradiated) diffusion length would be afactor of 2 or 3 greater than the thickness of the cellitself.
c. Lightweight substrate and mechanism. Significantimprovements in solar cell array weight have beenmade as a result of the NASA and Air Force Roll-Out Array Programs and the NASA Large Area SolarArray Program (Refs. 25-28) and could be used asbaselines for extrapolation toward even lighter weights.Whereas Mariner-class arrays produce 22 W/kg, roll-out arrays have demonstrated specific power of 66 W/kg.These programs utilize lightweight substrates such asKapton or stretched tape attached to a lightweightberyllium frame upon which the cells are mountedand interconnected. For the Earth-based systems, theuse of exotic materials such as beryllium would notbe advisable because of the increased costs, and inthis case some less expensive metal or plastic could
be used to achieve a semirigid or rigid array. Investi-gations are required into the various means of fabri-cating the array and appropriate substrates so as tominimize both weight and cost.
The protective layer for the Earth-based systemswould be installed either at the array fabrication facil-ity or on-site, whichever is more economically feas-ible. This would appear advantageous from the pointof view of cost and weight, as well as storage andtransportation. For the Solar Farm, the array couldbe rolled out and tied down in some manner at thesite of operation to achieve rigidity and resistance towinds. This could be accomplished by attachment tosome sort of inexpensive type of structure, such asplastic rods, or the array could be rigidized by chemicalor pneumatic means.
For the Rooftop Array, the array could be flexibleor attached to an inexpensive frame, as in the SolarFarm; chemically rigidized on the site; or attacheddirectly to the roof. Another approach for the Earth-based systems would be to modify the array systemused on the JPL Large Area Solar Array Program,which utilized a frame with stretched tape to providethe substrate upon which the cells were mounted.Such an array structure would not be as convenientfor storage and transportation as a flexible roll-up arraysystem, but might greatly facilitate installation.
Enough options exist for the Earth-based systemsthat, with a reasonable effort, a very economical andconvenient system to satisfy this Technology Require-ment could be evolved. The same is true for the Satel-lite station; however, the latter problem is greater andrequires much more effort to achieve success. Basically,the proponents of the Satellite Station anticipate usinga Kapton substrate and an FEP Teflon protective layerover the cells (Ref. 7).
The array erection and deployment techniques willbe most critical for the Satellite Station since this mustbe performed in space, preferably in an automated man-ner (but the involvement of manual assistance by as-tronauts should not be ruled out). Whether the opera-tion is automatic, manual, or a combination of thetwo, it could be an important engineering problem.Again, the results of the NASA and Air Force Roll-Out Array programs and the NASA Large Area SolarArray Program could be used as a baseline; however,significant modification would be required since theSatellite Station system involves erection and deploy-
JPL TECHNICAL REPORT 32-157312
ment of 33 km2 (13 mi2 ) of array plus 133 km2 (52 mi2 )of concentrator.
The erection/deployment techniques appropriate tothe Earth-based systems, which are, of course, criticalto the program, should be far less difficult to achievein practice than for the Satellite Station system, asdiscussed above.
3. Life extension. The usable lifetime of a solar ar-ray system will have a direct bearing on the overallcost per kilowatt-hour of the system. Thus provisionsfor replacement, rework, or refurbishing the solar ar-rays could favorably affect the economics involved insolar-electric power generation if the usable lifetimeof the system is significantly extended by so doing.As a preliminary design goal, a lifetime of approx-imately 30 years would not seem inappropriate. Thelife of solar arrays is generally governed by effectsthat cause deterioration of the conversion efficiencyof the arrays through obscuration of the solar cells,through changes in the physics of the solar cell, orthrough deterioration of the cell-to-cell contacts. Tech-nology needed to meet the requirement of long lifeinvolves the resistance of arrays to ionized particle ra-diation, ultraviolet radiation, humidity, wind, dust, pre-cipitation, and temperature cycling.
For the Earth-based systems, it is to be expectedthat routine maintenance, replacement, and refurbish-ing will be economically feasible and will be usedto increase the overall lifetime of the array system.The Satellite Station should be designed for minimummaintenance on a routine basis but should be amen-able to repair in case of catastrophic events such asmeteorite showers or major equipment malfunctions.Since the solar array has no moving parts (exclusiveof orientation and deployment mechanisms) and op-erates at relatively low temperatures, the array shouldbe inherently long-lived if properly designed for theapplicable environmental conditions.
a. Radiation resistance. The Technology Advance-ment Requirements for radiation resistance apply onlyto the Satellite Station. So far as radiation degradationof the array output is concerned, the concept of thesatellite power system is considerably different, andmore favorable, than the concept normally used forspace missions. In normal space missions, once thearray power falls below a certain specified design valuethere is no longer enough power to operate the as-sociated electronics required for the mission; the mis-sion must then be ended. With the satellite solar pow-
er system, however, there is no such sharp terminationof the mission, and even a severely degraded solararray could continue to supply power to the receivingstation. The 30-year life-span requirement, then, be-comes somewhat arbitrary, since at the end of 30 years,the array would not simply turn off, but would con-tinue to produce power, possibly at a lower rate. Anadditional advantage is that, for penetrating radiation,the power degrades as the logarithm of the fluence;that is, for a constant-radiation environment it wouldtake 100 years to degrade the array by the same per-centage as was lost during the first 10 years (assum-ing no capability for annealing out radiation damage).The Technology Advancement Requirements for im-proving radiation resistance have been discussed aboveunder "Weight Reduction."
b. Resistance to ultraviolet exposure. Degradationof solar arrays because of exposure to the ultravioletcomponent of the solar spectrum is usually reducedor eliminated by the use of ultraviolet reflecting filtersdeposited onto the solar cell coverglass. Space-typesolar arrays are presently designed so that there is littleor no adverse effect from ultraviolet exposure; however,since different materials and techniques are expectedto be used in the improved arrays, especially for theEarth-based system arrays, where inexpensive plasticsmight be used, particular attention must be given tothe effects of ultraviolet radiation on these new materialsand components. Materials interposed between the cellsurface and the solar radiation must be resistant toloss of transparency, and materials used to bond onecomponent to another must be resistant to embrittle-ment resulting from such exposure.
Many materials are extremely sensitive to exposureto ultraviolet (short-wave-length) light. Some materialsbecome embrittled by such exposure; others lose theiroptical transparency; still others suffer both types ofdamage. If the protective coatings, coverglasses, or ad-hesives interposed between the cell surface and theincoming solar radiation lose transparency, this losswould adversely affect the solar cell light-generatedcurrent, and consequently the cell or array efficiency.If materials such as epoxies or adhesives became em-brittled by ultraviolet exposure, this degradation couldcause loss of mechanical integrity of the array andpossibly result in decoupling or delamination of pro-tective layers, coverglasses, or cells from the substrate.Such an effect would, of course, be minimized if pre-dominantly mechanical means were used to attach all,or major portions, of the array blanket.
JPL TECHNICAL REPORT 32-1573 13
Since the Earth's atmosphere absorbs a considerableamount of ultraviolet light (short-wavelength photons)contained in the solar spectrum, the Earth-based sys-tems might be expected to be somewhat less sensitiveto this parameter than the Satellite Station, which re-ceives the total solar spectrum, unattenuated by theEarth's atmosphere. This, assumption, however, maybe too simplistic, since the total environments of in-terest consist of many components that are not neces-sarily simply superimposed upon one another as in-dependent variables, but that might indeed be depen-dent variables. Thus, the cumulative effects of two en-vironmental conditions could well be of greater mag-nitude than the sum of the two separately taken. Forexample, a combination of high humidity plus ultra-violet light exposure might result in a degradationgreater than the sum of exposure to ultraviolet aloneplus humidity exposure alone.
For the Earth-based systems, it might be more econ-omical to use a protective layer of a low-cost material,such as an inexpensive plastic, which experiences someamount of degradation from ultraviolet light, and toreplace this layer periodically, rather than to use a moreultraviolet-radiation-resistant but expensive material,assuming provisions are made for simple replacement(e.g., by mechanical attachment). This tradeoff wouldprobably not be applicable to the Satellite Station,since it is anticipated that periodic protective layerreplacement would be technologically and economic-ally expensive to perform in space.
c. Resistance to humidity. A considerable effort hasbeen expended by JPL to determine the effects ofhumidity (as well as other environmental conditions)on solar cell behavior, using both electrical and mech-anical criteria (Refs. 29-32). The urgency of such in-vestigations was clearly delineated by internal com-munications within NASA organizations indicating thatsevere degradation occurred in titanium-silver solarcell contacts, which proved to be a result of exposureto humidity. It should be emphasized that these deg-radations were the result of exposure to the then-nor-mal storage conditions (where humidity was not con-trolled) and not a result of specific testing to inducesuch failures. The problem of solar cell contact degra-dation in humid environments has been extensivelystudied, and it appears that the degradation mechan-ism is a result of a corrosive reaction between thetitanium and the silver in the presence of moisturewhich permeates through the rather thin silver layerto the titanium-silver interface. The JPL investigationsindicated that this condition could be greatly allevi-
ated through the application of a solder coating overthe contact metals, which acts as a physical barrier tomoisture permeation. The JPL studies also indicated,however, that the solder-coating could have adverseeffects in other environments, particularly for deepthermal shocks to temperatures about -196°C, due tothe thermal coefficient of expansion mismatch be-tween the solder and the silicon (Refs. 29, 30, and 32),and to storage at temperatures of approximately 150°C,probably because of the interaction between the tincomponent of the solder and the silver at this tem-perature (Refs. 29 and 30).
A second approach to increasing resistance of titan-ium-silver solar cell contacts to the detrimental ef-fects of humidity was reported several years ago byAEG Telefunken (Ref. 33): palladium was used in thetitanium-silver contact system to minimize or elimin-ate the corrosive reaction. Under JPL funding, twosolar cell manufacturers (Heliotek, a Division of Tex-tron, Sylmar, Calif., and Centralab, a Division of Globe-Union, Inc., Milwaukee, Wis.) fabricated and suppliedcells utilizing the palladium-containing contact systemfor JPL evaluation. This evaluation (Ref. 31) indicatedlittle or no advantage to these particular cells in ahigh-humidity, 80°C environment. Further extensiveanalysis of these cells by JPL indicated extremely poorprocess control used by the manufacturers to producethese cells, both in the deposition of the contact ma-terials and in the very high variation in the amountof palladium actually incorporated into the system.(In some cases, analysis showed no measurable quan-tity of palladium.) Hence, the results of the JPL evalu-ation are probably not indicative of those that couldbe achieved with an optimized process.
The problem of contact degradation as a result ofhumidity exposure is extremely important with respectto the Earth-based systems, where considerable hum-idity and precipitation exposure for long periods oftime can be expected to occur, and such exposure,especially when compounded with exposure to hightemperatures, could be disastrous. The problem issomewhat less serious for the Satellite Station, sincethe only humidity exposure would occur during storage,and this, of course, can be circumvented by carefulcontrol of the storage environment. The humidity prob-lem could also be circumvented for all Options byencapsulation of the cells/modules/arrays in a man-ner similar to that presently being investigated byNASA-Lewis Research Center utilizing FEP Teflon(Refs. 13 and 14). In any case, enough options exist
JPL TECHNICAL REPORT 32-157314
to lead one to believe that with a reasonable effortthis Technology Advancement could be acomplished.
d. Resistance to wind/dustlprecipitation. The arraysmust be protected from the adverse affects of wind,dust, and precipitation for the Earth-based systems.This can probably most conveniently be accomplishedby a means of encapsulating the array, somewhatsimilar to the method being investigated by NASA-Lewis Research Center using FEP Teflon (Refs. 13and 14); however, perhaps some less expensive alter-native material could be used. The replacement orrefurbishing of a protective layer to guard against theadverse affects of these environmental factors shouldalso be considered. Such a layer must be highly trans-missive with respect to the usable photons of theEarth-surface solar spectrum and must be compatiblewith the other technology requirements. This Tech-nology Advancement Requirement is not applicableto the Satellite Station, since the storage conditionscan be carefully controlled.
e. Capability of withstanding temperature variations.Many spacecraft have been successfully designed towithstand both large numbers of temperature cyclesand deep temperature cycles (thermal shocks). Sinceradical departures in array design and materials areanticipated in the future, care must be exercised toensure that this capability is not compromised. Deg-radation resulting from large numbers of temperaturecycles is usually associated with fatique mechanisms,while deep temperature cycle degradation usually re-sults from thermal expansion coefficient mismatchesbetween the composite materials of the solar arraystack, which can cause very significant stresses.
Resistance to large numbers of temperature cycles.For the Earth-based systems, very large numbers oftemperature cycles can be expected due to seasonalvariations, diurnal variations, and climatic fluctuationsthroughout the day. In these cases, the thermal varia-tions could be combined with variations in humidity(the problems of which were previously discussed).The components of the array should not interact ad-versely with one another (for example, by severe ther-mal coefficient expansion differences) in a way detri-mental to the array performance, and the materialsused should not become fatigued by such repeatedtemperature cycles.
For the Satellite Station, the number of temperaturecycles is considerably reduced. The satellite wouldbe exposed to full sunlight most of the time, except
for a 1.2-h interval every 24 h for 25 days before andafter equinox in the 35,600-km (22,300-mi) synchron-ous orbit proposed by the proponents of this system.Thus the number of cycles is not nearly so severe asfor the Earth-based systems, but the temperature ex-cursions in going from full sunlight with a 3-to-1 con-centration ratio to a 1.2-h orbital night would bevery large and rapid with respect to time-rate of tem-perature change, because of the low thermal mass ofthe array.
Resistance to large temperature excursions. As dis-cussed above, the Satellite Station would undergo verysevere and rapid thermal excursions during the 1.2-horbital night, experienced every 24 h for 25 days be-fore and after equinox. It has been the experience ofpanel designers that the failure mechanisms resultingfrom a large number of rather shallow temperaturecycles can be quite different from those resulting froma smaller number of very deep temperature cycles.In the former, the failure is generally a fatigue mech-anism, and in the latter, the failure mechanism is dueto thermal expansion coefficient mismatches of the com-ponent materials of the array. Deep temperature cycleswould be much less likely for the Earth-based systemsthan for the Satellite Station.
Technology Requirements to solve temperature vari-ation problems include careful selection of the materials,with respect to fatigue characteristics and thermal expan-sion coefficient matches, and proper integration of thematerials to form the solar array blanket (e.g., mech-anical design of interconnectors to ensure compliancyand adequate stress-relief). Enough materials and de-sign alternatives appear to be available to satisfy theserequirements with a reasonable amount of effort. Fur-thermore, the considerable progress being made instress analysis and modeling appropriate to solar arraydesign (Refs. 34 and 35) provides a good baselinefrom which advancements can be made.
4. Reliability. As discussed above under "Life Ex-tension," reliability in the three Options being consid-ered has a somewhat different connotation than re-liability in normal solar arrays for space application.In normal space-type arrays, certain finite boundaryconditions, primarily in array power output, are im-posed, and once the array no longer satisfies theseboundary conditions, which are very specific, the ar-ray can be considered as having "failed." Thus, forexample, under the normal concept of reliability wewould demand that the array produce an amount ofpower equal to or greater than a specified value for
JPL TECHNICAL REPORT 32-1573 15
the 30-year design goal lifetime. Furthermore, reliabil-ity considerations presently used tacitly assume thatthere is no capability for repair, rework, or replace-ment of defective components, but that the deployedarray, as originally designed and fabricated, must sur-vive and satisfy the boundary conditions.
Clearly, the highest reliability (as presently definedin our space program) is desired in the environmentsto which the arrays in these Options will be exposed;however, there will very likely be a major tradeoffbetween reliability and the costs of array fabrication.That is, there will generally be some adverse cost im-plications as a result of increasing reliability. It mighttherefore be economically advantageous, on a long-term systems basis, to accept a somewhat lower reli-ability and a requirement that the degraded elementsbe repaired, replaced, or refurbished, rather than todemand that the array operate for the 30-year designgoal without such maintenance requirements.
The Earth-based systems, being easier to maintain,could be expected to benefit especially from such amaintenance concept, whereas, for the Satellite Station,replacement, refurbishment, and rework would be con-siderably more difficult. Therefore, the reliability andstability requirements have the greatest priority forthe Satellite Station, with a somewhat lower priorityfor the Earth-based systems. For the Solar Farm, someperiodic and routine maintenance is expected, so thatdefective parts could be replaced or refurbished andthe array could be cleaned periodically. For the Roof-top Array, the homeowner or the service contractorwould be required to get up to the roof, locate thedefective part, and replace or rework it; therefore,reliability and stability are somewhat more importantfor this Option than for the Solar Farm.
The requirement for reliability and stability for allOptions implies that the worst-case and nominal-caseenvironmental conditions be accurately known, and thisin itself would require a significant effort.
a. Definition of the environment. One of the firstorders of business must be the definition of the anti-cipated environments. Oddly enough, the space en-vironment appropriate to the Satellite Station is prob-ably most easily defined, with the exception of thesporadic solar flare protons, about which there is verylittle information, and the possibilities of meteoriteshowers, which are also sporadic and hence unpre-dictable. The need does exist, however, to define thisenvironment as best we can, using the considerable
body of existing information on the space environmentand perhaps performing additional selected experi-ments to fill in any remaining gaps.
The definition of the Earth-based systems' environ-ments will vary with geographical location. While agood deal of information on certain aspects of the en-vironment is available through the United StatesWeather Service, very little is known of how wind,dust, wind-borne particles, sedimentation rates etc.,may affect solar arrays. Thus, the existing informa-tion and its applicability to solar array operationsmust be evaluated, and the type of additional infor-mation that is required must be determined. Sincethe real combined effect of all the environments canbe discovered only by actually operating and measur-ing the arrays installed in representative geographicallocations, a program for accomplishing this must beinitiated. Such a program would define nominal andworst-case levels for the environmental componentsand develop testing procedures (preferably acceleratedtests) to be applied to candidate array systems.
This Technology Requirement is of major impor-tance; a significant effort involving literature searches,compilation of data, statistical analysis, computer mod-eling, laboratory testing, and on-site testing and eval-uation will be required.
b. Characterization of cell, module, and array. Thecell, module, and array electrical and mechanicalcharacteristics must be determined, principally as afunction of the environment (including insolation) inwhich the system is to be operated. Design informa-tion may initially require a significant amount ofcharacterization, possibly even down to microscopicevaluations of the crystalline perfection of the solarcell. Such a study could utilize analytical techniquespresently used by JPL and other organizations toevaluate and modify the cell, module, and array de-sign. The development effort will require comprehensivecharacterization; for actual production of solar arrays,however, the minimum characterization measurementsto achieve a desirable system must be determined andused because the degree of characterization is reflectedin the total system cost. Since the Earth-based systemsare amenable to rework and replacement, the charac-terization required for these arrays should be less thanthose required for the Satellite Station, where such op-erations would have to be accomplished in space. Typi-cal techniques and processes for characterization aredescribed in detail in Refs. 18-24 and 29-35.
JPL TECHNICAL REPORT 32-157316
c. Maintenance. Because there will likely be sometradeoff between reliability and array fabrication costs,as discussed above, considerable attention must begiven to possible cost reductions and increase of over-all system lifetime associated with maintenance tech-niques to replace, refurbish, or rework degraded por-tions of the array. This appears to be particularly ap-propriate for the Earth-based systems and should noteven be ruled out for the Satellite Station.
In the ideal case, for any Option, no maintenancewhatsoever would be required; however, it is to beexpected, especially for the Earth-based systems, thatsome maintenance will be required, such as periodiccleaning of the surfaces and/or replacement of ele-ments that have been damaged by such factors assand or pebble abrasion, excessive wind loading, etc.Perhaps the space environment, other than the radia-tion problem previously discussed, would be morebenign in this respect, meteorites being the only otherelement one could envision as causing physical dam-age to the array. In this case, enough redundancy anddiode protection devices might be designed into thesystem so that maintenance would not be required,except for very highly improbable events, such aslarge meteorite showers impinging directly on thespacecraft and/or arrays.
Because a homeowner or building manager mightbe reluctant to do his own rooftop maintenance, aminimum-maintenance requirement might be moreimportant for the Rooftop Array than for the SolarFarm, where routine maintenance would be expectedto be performed by semiskilled technicians. However,material and fabrication costs for both Earth-basedsystems must be traded off against maintenance coststo achieve greatest overall cost effectiveness; if the de-graded elements of the array could be located and re-placed very cheaply, it might be most economical toaccept some measure of maintenance requirement forboth systems.
Technology for simple replacement of defective ar-ray segments is especially needed in the Satellite Sta-tion, since this will have to be accomplished in space;the replacement-simplicity requirement would prob-ably be least stringent with respect to the Solar Farmand somewhat more important for the Rooftop Array,so that the person responsible for the building couldreplace defective segments himself, should he so de-sire, or hire a semiskilled operator to do it. For allthree Options, the replacement of solar cells must berelatively simple for the systems to be viable.
A simple, accurate means of locating defective ordegraded elements would be required for the Satel-lite Station and would also be very important for theSolar Farm, since, in this array, areas of the order ofsquare miles would be involved. For the Rooftop Ar-ray, some system of monitoring array segment perform-ance within the dwelling could perhaps be utilized,especially since the array area involved is not nearlyso large as the others, and location of defective seg-ments might therefore in practice be simpler.
5. Fabrication capability. The requirement for totalcombined power output of solar arrays for terrestrialapplications is in the thousands of megawatts ratherthan the kilowatt range. In addition to the obviousproblem of fabricating arrays for the Solar Farm andthe Satellite Station, even if done piecemeal, thereare production problems in terms of the large de-mand for silicon solar cells for all three Options. Cellproduction in the United States would have to in-crease by at least five orders of magnitude. More-over, the energy consumed in producing silicon cellsand solar arrays must be considered in relation tothe energy that can be extracted from the system.The development of simplified automated cell and ar-ray production techniques and the development anduse of low-cost, readily available materials are signif-icant Technology Advancement Requirements if photo-voltaic solar arrays are ever to become feasible pro-ducers of commercially available electrical energy.
a. Cell fabrication techniques. Improvements in cellfabrication techniques are discussed in detail in Ref.1. The first consideration must be the choice of solarcell material on the basis of (1) economy, (2) amenab-ility to fabrication of high-efficiency, stable solar cells,and (3) availability. Silicon was recommended as thebaseline cell material since it satisfies requirement (2)and is the most well-understood semiconductor mater-ial. Silicon is the second most abundant element onEarth, but the availability of silicon in the ultrapure,single-crystalline form presently used for solar cellfabrication is very limited. For example, to cover onesquare mile of area with single-crystalline silicon cells0.25 mm thick, assuming a wastage of 50% of the sili-con (which unfortunately is optimistic with respect topresent-day fabrication techniques), represents sixtimes the yearly production of such silicon in theUnited States (Ref. 8). Ultrapurity single-crystallinesilicon costs about $0.30 per gram. As discussed pre-viously, the use of high-purity or metallurgical gradesilicon (as opposed to ultrapurity single-crystallinesilicon) would help solve the problems of economy
JPL TECHNICAL REPORT 32-1573 17
and availability but might result in lower efficiencycells. The increased demand for ultrapurity single-crystalline silicon should also make more silicon avail-able at lower cost because of the savings and automa-tion improvements of high-volume production. Moreefficient use of the silicon-for example, by growingit in the form of large-area silicon films-would offerfurther improvement. The availability of other solar cellmaterials that may be used must also be considered. Forexample, some proponents of cadmium sulfide advo-cate reclaiming the cadmium, which is in limited sup-ply, from cells that are no longer usable (Ref. 36).Reclamation of materials used in the Earth-basedsystems should be considered, while this, would prob-ably not be appropriate for the Satellite Station.
The use of large-area cell blanks should presentsignificant advantages for cell processing, since fewerpieces must be handled to achieve a required totalcell area. To optimally integrate large-area cell blanksinto the cell processing,. severe modifications and im-provements in production techniques are required, butsuch modifications and improvements are required inany case to significantly improve the quantity andeconomics associated with cell fabrication. All process-ing steps must be greatly improved and automated,including blank sizing, junction formation (diffusion,epitaxial growth, ion implantation, etc.), electricalcontact (or contact-interconnect) formation (evapora-tion, electroforming, plating, silk screening, etc.), anti-reflective coating, measurement, and all associatedpre- and post-operation cleaning/etching treatments."Endless-belt" processing steps are to be preferred.
b. Array fabrication techniques. The key to improvedfabrication techniques lies in total automation of theprocesses, a drastic departure from the traditionalconcept of solar arrays as a highly specialized custom-built product to the consideration of solar arrays asa mass-market, very-high-volume product. This re-quires the iinvolvement of a new body of talent withexpertise in the area of such mass-market, high-volumefabrication. It requires the merging of the experienceof the aerospace community in the design, testing,and evaluation of the array, with the experience ofthe high-volume production-oriented industrial com-munity, hitherto untapped in the fabrication of solararrays. The design engineer must ensure minimummaterial costs and compatibility of all materials andprocesses with one another, as well as maximum reli-ability, allowable tolerances, and efficiency in antici-pated environments. The manufacturing engineer mustensure high-volume production capability, minimum
equipment costs, automation of processes, minimumreject rate, minimum use of highly specialized equip-ment and controls, maximum process simplicity, maxi-mum rework capability, and maximum materials andequipment utilization. In actuality this start with cellfabrication, integrated into module fabrication, sub-strate fabrication (material, techniques, size, weight,printed circuitry), cell/module laydown and intercon-nection techniques and materials (e.g., welding, mech-anical bonding, chemical bonding, mechanical attach-ment), protective layer attachment (coverglass, inte-gral coverglass, plastics, Teflon, spray-on, roll-on,mechanical, etc.), characterization (mechanical/elec-trical), and rework as required.
It is imperative that automation of all processes bedeveloped and coupled with low-cost, readily avail-able materials.
V. Discussion and ConclusionsThe Technology Advancement Requirements repre-
sent a 3-order-of-magnitude reduction in costs and a5-order-of-magnitude increase in production capability.While these approximate order-of-magnitude numbersappear formidable, such cost reductions and capacityincreases have been achieved with reasonable regular-ity in other areas. Furthermore, the fundamental tech-nologies of fabricating solar arrays and their compo-nents are rather well understood, so that we have afirm baseline from which to work. The TechnologyAdvancements do not require major breakthroughs (al-though, of course, major breakthroughs would increasethe probability of success) or the use of exotic mater-ials, but rather are predicated on significant extensionsof existing technology. The approach toward achiev-ing the Technology Advancements will center aroundengineering and manufacturing improvements and notaround major commitments to fundamental research(as was the case at the inception of the nuclear re-actor electrical generator development and will bethe case with the development of a fusion-processgenerator) and, therefore, the success probability shouldbe reasonably high.
The author's personal bate noire is estimating theeventual costs of a solar array system to provide large-scale electrical power generation for terrestrial appli-cations, mainly because there are no hard data fromwhich to draw sound conclusions. Qualitatively, how-ever, there appears to be good reason for considerableoptimism. Silicon, which is proposed as the baselinematerial for fabrication of the solar cells, is the second
JPL TECHNICAL REPORT 32-157318
most abundant element on Earth, and therefore thereis certainly no lack of material: it is simply a questionof processing this material inexpensively into a formusable for photovoltaic energy conversion.
Despite all the problems outlined in the Techno-logy Advancement Requirements, it seems almost in-conceivable that such a simple thing as a solar arraysubstrate with printed circuit interconnections and wir-ing upon which cells are mounted in some simple,economical manner, and over which some inexpen-sive protective layer is positioned, having no movingparts and using no exotic materials, cannot be madefor a few dollars a square meter rather than the thou-sands of dollars per square meter experienced in thespace program. History gives us many examples ofitems which were once prohibitively expensive butwhich have now become so inexpensive that they havebecome disposable. (Aluminum was once so expensive
to produce that regal crowns were fabricated fromthis exotic and precious material-we now use it forwrapping our leftovers.) It is almost more difficultfor the author to believe that economically viablesolar arrays cannot be built rather than that theycan be built.
The same philosophy can be applied to the otherproblems mentioned in Section III (i.e., electrical ener-gy storage devices, satellite-to-Earth power transmis-sion, attitude control, etc.). Since no major fundamentalbreakthroughs are required, but rather drastic modi-fication and improvement of existing technologies, itmight appear that successful attainment of the goal ofwidespread, pollution-free generation of electrical pow-er could fail to be achieved only because of a lackof inventiveness, resourcefulness, or commitment ofresources to achieve that goal.
JPL TECHNICAL REPORT 32-1573 19
References
1. Berman, P. A., Considerations With Respect to the Design of Solar Photo-voltaic Power Systems for Terrestrial Applications, Technical Report32-1556. Jet Propulsion Laboratory, Pasadena, Calif., June 15, 1972.
2. Glaser, P. E., "Power From the Sun," Mech. Eng., Vol. 91, No. 3, Mar. 1969.
3. Gaucher, L. P., "Energy Sources of the Future for the United States," J.Solar Energy Soc., Vol. 9, 1965.
4. Goubau, G., "Microwave Power Transmission From an Orbiting SolarPower Station," J. Microwave Power, Vol. 5, No. 4, Dec. 1970.
5. Robinson, W. J., Jr., The Feasibility of Wireless Power Transmission foran Orbiting Astronomical Station, NASA Technical Memorandum TM-X-53806. National Aeronautics and Space Administration, Washington, D.C.,May 1969.
6. Brown, W. C., "The Receiving Antenna and Microwave Power Rectifica-tion," J. Microwave Power, Vol. 5, No. 4, Dec. 1970.
7. Ralph, E. L., and Benning, F., "The Role of Solar Cell Technology in theSatellite Solar Power Station," Conference Record of the 9th IEEE Photo-voltaic Specialists Conference, held at Silver Spring, Md., May 1972,Institute of Electrical and Electronics Engineers.
8. Ralph, E. L., "Large Scale Solar Electric Power Generation," presentedat the Solar Energy Society Conference, NASA Goddard Space FlightCenter, Greenbelt, Md., May 1971.
9. Cadmium Sulfide Thin-Film Solar Cell Review, NASA Technical Memor-andum TM-X-52920. NASA Lewis Research Center, Cleveland, Ohio,Apr. 1970.
10. Palz, W., et al., "Analysis of the Performance and Stability of CdS SolarCells," Conference Record of the 8th IEEE Photovoltaic Specialists Con-ference, held at Seattle, Wash., Aug. 1970, Institute of Electrical andElectronics Engineers.
11. Bernatowicz, D. T., and Brandhorst, H. W., "The Degradation of Cu2 S-CdS Thin Film Solar Cells Under Simulated Orbital Conditions," Con-ference Record of the 8th IEEE Photovoltaic Specialists Conference, heldat Seattle, Wash., Aug. 1970, Institute of Electrical and ElectronicsEngineers.
12. Thin Film GaAs Photovoltaic Solar Energy Cells, RCA Final Report,Contract NAS-8510, Aug. 1967.
13. Broder, J. D., et al., Recent Results of FEP Solar Cell Cover Develop-ment," Conference of the 9th IEEE Photovoltaic Specialists Conference,held at Silver Spring, Md., May 1972, Institute of Electrical and Elec-tronics Engineers.
14. Investigation of FEP Teflon as a Cover for Silicon Solar Cells, NASACR-72970, Contract NAS 3-14398. Lockheed Palo Alto Research Labora-tory, Palo Alto, Calif., Aug. 1971.
JPL TECHNICAL REPORT 32-157320
References (contd)
15. Berman, P. A., High Efficiency Silicon Solar Cells, Report No. VIII,Final Report, U.S. Army Electronics Laboratories contract DA 36-039-SC-907, covering the period from June 1962 to July 1964. Heliotek, a Divisionof Textron, 1964.
16. Berman, P. A., "Design of Solar Cells for Terrestrial Use," presented at theSolar Energy Conference, Boston, Mass., Mar. 1966; published in SolarEnergy, Vol. 11, Nos. 3 and 4, 1967.
17. Berman, P. A., and Ralph, E. L., "Improved Solar Cells for Use in Con-centrated Sunlight," Proceedings of the 18th Annual Power Sources Con-ference, U.S. Army Signal Corps, Ft. Monmouth, N.J., May 1964.
18. Berman, P. A., Effects of Solar Proton Flares on the Power Output ofSolar Cells Having Various Configurations, Technical Report 32-1251.Jet Propulsion Laboratory, Pasadena, Calif., Feb. 15, 1968.
19. Berman, P. A., "Summary of Results of JPL Lithium-Doped Solar CellDevelopment Program," Conference Record of the 9th IEEE Photovol-taics Specialists Conference, Silver Spring, Md., May 1972, Institute ofElectrical and Electronics Engineers.
20. Proceedings of the Fourth Annual Conference on Effects of LithiumDoping on Silicon Solar Cells, Technical Memorandum 33-491, edited byP. A. Berman. Jet Propulsion Laboratory, Pasadena, Calif., Sept. 15, 1971.
21. Berman, P. A., Effects of Lithium Doping on the Behavior of Silicon andSilicon Solar Cells, Technical Report 32-1514. Jet Propulsion Laboratory,Pasadena, Calif., Feb. 1, 1971.
22. Proceedings of the Third Annual Conference on Effects of Lithium Dop-ing on Silicon Solar Cells, Technical Memorandum 33-467, edited byP. A. Berman and J. Weingart. Jet Propulsion Laboratory, Pasadena,Calif., Apr. 1, 1971.
23. Berman, P. A., "Status of Lithium Solar Cell Investigations at the JetPropulsion Laboratory," Conference Record of the 8th IEEE Photovol-taic Specialists Conference, held at Seattle, Wash., Aug. 1970, Instituteof Electrical and Electronics Engineers.
24. Proceedings of the Conference on Effects of Lithium Doping on SiliconSolar Cells, Technical Memorandum 33-435, edited by P. A. Berman.Jet Propulsion Laboratory, Pasadena, Calif., Aug. 15, 1969.
25. Wolff, G., "The Flexible Roll-up Solar Array Flight Experiment," Con-ference Record of the 9th IEEE Photovoltaic Specialists Conference,held at Silver Spring, Md., May 1972, Institute of Electrical and Elec-tronics Engineers.
26. Flexible Rolled Up Solar Array, Hughes Aircraft Final Report, ContractAF 33(615)68-C-676. Hughes Aircraft Company, Nov. 1970.
27. Rollup Sub-Solar Array, General Electric Final Report, Contract No.JPL 952314. General Electric Company, Oct. 1970.
28. Lightweight Solar Panel Development, Boeing Final Report, ContractNo. JPL 952571. The Boeing Company, July 1970.
JPL TECHNICAL REPORT 32-1573 21
References (contd)
29. Berman, P. A., and Yasui, R. K., Effects of Storage Temperatures onSilicon Solar Cell Contacts, Technical Report 32-1501. Jet PropulsionLaboratory, Pasadena, Calif., Oct. 15, 1971.
30. Berman, P. A., and Yasui, R. K., Supporting Data Package for TR 32-1541,Effects of Storage Temperatures on Silicon Solar Cell Contacts, Tech-nical Memorandum 33-497. Jet Propulsion Laboratory, Pasadena, Calif.,Oct. 15, 1971.
31. Yasui, R. K., and Berman, P. A., Effects of High-Temperature, High-Humidity Environment on Silicon Solar Cell Contacts, Technical Report32-1520. Jet Propulsion Laboratory, Pasadena, Calif., Feb. 15, 1971.
32. Moss, R., and Berman, P. A., Effects of Environmental Exposures onSilicon Solar Cells, Technical Report 32-1362. Jet Propulsion Laboratory,Pasadena, Calif., Jan. 15, 1969.
33. Fisher, H., and Cereth, R., "New Aspects for the Choice of ContactMaterials for Silicon Solar Cells," Conference Record of the 7th IEEEPhotovoltaic Specialists Conference, held at Pasadena, Calif., Nov. 1968,Institute of Electrical and Electronics Engineers.
34. Salama, A. M., Rowe, W. M., and Yasui, R. K., Stress Analysis and De-sign of Silicon Solar Cell Arrays and Related Material Properties, Tech-nical Report 32-1552. Jet Propulsion Laboratory, Pasadena, Calif., Mar. 1972.
35. Butterworth, L. W., and Yasui, R. K., Structural Analysis of Silicon SolarArrays, Technical Report 32-1528. Jet Propulsion Laboratory, Pasadena,Calif., May 1971.
36. Boer, K. W., "Large Scale Use of Photovoltaic Cells for Terrestrial SolarEnergy Harvesting," Conference Record of the 9th IEEE PhotovoltaicSpecialists Conference, held at Silver Spring, Md., May 1972, Instituteof Electrical and Electronics Engineers.
![An Adaptive Solar Photovoltaic Array Using Model-Based ......2) The local hot spot in the shaded part of the solar PV array can damage the solar cells [7]. There are several approaches](https://static.documents.pub/doc/80x56/5f155d1cad09a87aff0b5cff/an-adaptive-solar-photovoltaic-array-using-model-based-2-the-local-hot.jpg)
