VC01 Solar collection and conversion
Depth of coverage
At its target depth of L2. What the depth levels mean.
What this layer covers
Cells, concentrators, optics, thermal conversion, degradation, sun pointing of the collector
Where it ends
Ends at the DC or electrical output terminals of the collector
Solar collection: where it stands
Solar collection turns sunlight into electricity with solar cells on the satellite. The main reference is a 1978 US paper design. It traced every loss from sunlight to grid and reached about 7 per cent overall with cells of 17.3 to 20 per cent efficiency. That assumed radiation damage to the cells could be partly reversed by heating them, a process called annealing (CLM-0974). Tests for Boeing in 1977 and 1979 recovered a fifth to a half of the radiation loss in a few bare silicon cells by annealing (CLM-1118, CLM-1119). No glass-covered cell was annealed (CLM-1119). The ASTRAL thin-cell concept is modelled to exceed 25 per cent in geostationary orbit, the high orbit where a satellite stays over one point on Earth, but no device had been made by 2025 (CLM-0200). This site's own calculation, at medium confidence, puts the area of solar cell blanket at 2.3 to 6.5 km2 per gigawatt at the array, without concentrators. Cell efficiency sets most of that range, and the upper efficiencies are modelled or assumed, not achieved (CLM-0975).
Scope. A solar power satellite starts with a very large array of solar cells, sometimes with mirrors, called concentrators, that put extra sunlight on the cells. This section covers those cells and mirrors, how radiation ages them, and how the array keeps facing the Sun. It ends at the electrical terminals of the array. Wiring beyond that point is covered in Moving the power around the satellite.
What is settled. The arithmetic of the old reference designs holds. The 1978 US reference design traced every loss from sunlight to the grid as a chain of boxes, and the chain multiplies out at every step. For silicon cells it runs from 70.81 GW of intercepted sunlight to 5 GW at the utility busbar, the point where power passes to the grid. That is 7.06 per cent overall, or 6.97 per cent for its gallium aluminium arsenide (GaAlAs) option, a type of gallium arsenide cell (CLM-0974). The design used cells of 17.3 to 20 per cent efficiency. The chain implies a solar constant of 1353 W/m2 (CLM-0974). The solar constant is the sunlight power per square metre in space at Earth's distance from the Sun. Boeing's later sizing states 1353 W/m2 outright (CLM-1018). A 1981 programme chain is the same, box for box (CLM-1018).
What the arithmetic shows. A 2025 Chinese design's 12.3 per cent sunlight-to-grid figure, 1 GW from 6 km2 of array, holds when recalculated from its published values at a solar constant of 1361 W/m2. It is a design target, not a measurement (CLM-0692). The 1978 chain implies a slightly lower 1353 W/m2 (CLM-0974).
This site's own calculation sweeps the ranges of cell efficiency, ageing, pointing and sunlight in the sources held here. It gives 2.3 to 6.5 km2 of blanket, the thin sheet that carries the cells, per gigawatt of direct current at the array, for flat arrays with no mirrors (CLM-0975). Only cell efficiency moves the answer by more than about 15 per cent. Per gigawatt at the busbar the areas are about 1.8 times larger. Confidence is medium, because the higher cell efficiencies are modelled or assumed, not achieved (CLM-0975).
What is uncertain. Cell efficiency is the main variable, and the evidence for the high values is modelled or assumed. The ASTRAL concept, an ultra-thin two-layer cell, is modelled to pass 25 per cent in geostationary orbit at one sun (CLM-0200). Geostationary orbit is the high orbit where a satellite stays over one point on Earth. One sun is ordinary sunlight with no concentration. ASTRAL is also modelled to give 1 kW/m2 at 50 degrees C under less than ten suns, that is, up to ten times ordinary sunlight. The concentrator, mounting and radiator are still undefined (CLM-0202). A 2 micrometre front coating would replace the usual 100 micrometre cover glass, the glass layer that shields a cell from radiation. The model gives a shielding-mass saving of more than 30 times, for "decades" of life with no stated years (CLM-0431). No ASTRAL device had been made by 2025 (CLM-0200).
The older designs ran ahead of their hardware. Boeing's 1977 silicon design assumed an 18 per cent cell with no test data (CLM-1114). With mirrors, its array was calculated to lose 41 per cent of its output over 30 years, on extrapolated radiation data (CLM-1115, CLM-1116).
Annealing means heating the cells to heal radiation damage. The tests of it were small. In March 1977 five bare cells regained 20 to 40 per cent of their lost power by laser, and 44 and 54 per cent by electron beam (CLM-1118). Boeing's July 1977 trade, for a 10 GW satellite with a projected 18 per cent cell, then found annealing the least-cost way to hold output for 30 years. It favoured flat arrays without mirrors, partly because mirrors also degrade and cannot be annealed (CLM-1117). The 1978 ageing factors assume annealing (CLM-0974). In 1979 a laser restored about half the damage in six bare 50 micrometre cells. An oven restored 90 to 100 per cent, but that figure came from earlier work. No glass-covered cell was annealed, because the test found no suitable way to put a cover glass on so thin a cell (CLM-1119).
Gallium arsenide has the same gap. A 1977 study predicted a 20 per cent cell by 1990 and assumed it would stay close to 20 per cent for 30 years, especially if it could be annealed. The report said it could not support the end-of-life value with numbers (CLM-1120). In 1980 Rockwell put the best laboratory cells at about 18 per cent and expected 20 per cent by 1990 (CLM-1085).
Also in 1980, the US Department of Energy put the energy payback of a silicon satellite at 6.38 years, against 1.30 years for gallium aluminium arsenide (CLM-1038). Energy payback is the time a plant takes to generate the energy used to build it. The silicon figure rests almost wholly on an assumed energy cost of making the cells. At the low end of the stated range, or with solar power used to make the cells, it falls to 1 to 2 years. The gallium figure rests on one reference that the report says needs closer checking (CLM-1038).
What is missing. Measured results for the higher efficiencies are missing: those figures are modelled or assumed, not achieved (CLM-0975). Four sources would help: test data from a built ASTRAL-type device, an annealing test on a cell with its cover glass, long data from orbit on how thin cells age, and a recent study of the energy used to make space cells.
Read next. Efficiency of power beaming at range. Collector area and waste heat: what the designs state. Which assumption drives the cost spread. Moving the power around the satellite. Getting rid of the waste heat. From direct current to beam.
Records
32 of the 37 records tagged to this layer are published. The rest are not yet published. A record is published only after a second reader has checked it and it has passed every publication check. What the gate checks.
| Record | In plain language | Tier | Confidence |
|---|---|---|---|
| CLM-0200 | A new solar cell design, tested only in computer models, is predicted to convert more than 25% of sunlight to electricity in orbit. | T2 | MED |
| CLM-0202 | In simulation, this new cell design can produce a strong 1 kW per square metre of power while staying at a moderate 50 degrees C. | T2 | MED |
| CLM-0431 | This cell design might need over 30 times less shielding mass, from a coating 50 times thinner, than today's space solar cells, while still lasting for decades, according to computer models. | T2 | MED |
| CLM-0508 | An independent university research group has modelled a solar-cell design that could need over thirty times less protective shielding weight than today's space solar cells, but the design exists only in computer simulation. | T2 | MED |
| CLM-0539 | A UK consultancy, working for the European Space Agency and using the same design as its reference point, actually published a full weight breakdown -- something the original 2019 paper never did. It comes out lower than the headline claim, at about 0.70 kilowatts per kilogram rather than 1, even though it leaves out several categories of weight altogether. It also rates one of the satellite's power-conversion stages as barely more than an idea on paper, and its input notes say the lightest case for the radio equipment is an expert's judgement, not something anyone has measured or built. | T4 | MED |
| CLM-0692 | The Chinese design's own summary table adds up: the parts weigh what the total says they weigh. Its claimed 12.3 per cent efficiency is sunlight-to-grid, which checks out against the area of the panels and the strength of sunlight in space. The antenna is a long thin strip, 100 metres by 10.5 kilometres, at 5.8 gigahertz. | T2 | HIGH |
| CLM-0880 | The Chinese design runs its solar panels at about 500 volts and steps up to a 5,000 volt main bus. It is the only modern satellite power distribution architecture recorded here. | T2 | MED |
| CLM-0922 | In 2001 a National Research Council committee found that the basic trade study for the solar array had never been done, and that the mass and cost of getting the array to orbit dominates whatever the array itself costs. Their summary is that making the array free would not fix it. | T3 | HIGH |
| CLM-0923 | The 2001 committee says a solar power satellite several square kilometres across would need tens to hundreds of kilometres of cable in total, and that cable mass is one of the costs of using less efficient cells. | T3 | MED |
| CLM-0924 | In four parts of NASA's programme the 2001 committee found an analysis missing or a goal without a stated basis: for solar arrays, for how much of the assembly work robots should do, for structures, and for thermal control. Reading these as one pattern, targets set before the study that would tell you what to aim for, is an inference; the committee makes each finding separately. | T3 | MED |
| CLM-0928 | In 2011 an international study group said the last whole-system analysis of space solar power was done in 1997. The American National Research Council had already found the underlying studies missing in four technology areas in 2001. If these are the same gap, which is a reading and not something either body says, it was four years old when the NRC found it and still open ten years after that. | T4 | MED |
| CLM-0974 | The 1978 reference design shows every loss between sunlight and the power grid as a chain of numbered boxes, and the boxes multiply out exactly: about 7 per cent of the sunlight hitting the satellite reaches the grid. A rival 1980 design from Rockwell does the same sum and lands within about one per cent of its own published answer. | T3 | HIGH |
| CLM-0975 | Depending on which published assumptions you take, a satellite needs between about 2 and 6 square kilometres of solar blanket for every gigawatt it generates. Nearly all of that spread comes from how efficient the cells are; degradation and pointing each shift it by ten per cent or so, and the exact sunlight value by under one per cent. | T0 | MED |
| CLM-1018 | Three years of study left the efficiency chain exactly where 1978 put it, and Boeing's own version says out loud the solar constant the 1978 figure only implied. | T3 | HIGH |
| CLM-1038 | In 1980 the Department of Energy calculated that a silicon space solar satellite would take six years to pay back the energy that built it, four times longer than coal or nuclear, and that almost all of that was the energy cost of making the silicon cells; with gallium arsenide cells it was a little over a year. | T3 | HIGH |
| CLM-1076 | In 1980 the electronics that would condition the satellite's power weighed, at the state of the art, more than the whole satellite was allowed to; the design needed them fifty times lighter, and the arrays to run at fifty times the voltage anyone had flown. | T3 | HIGH |
| CLM-1085 | By 1980 both contractors' satellites had grown about 2.3 to 3.5 times heavier per kilowatt than the 1974 sketch; the gallium arsenide cell assumed for 1990 was two points above the best in any laboratory, and the silicon array would lose almost a third of its output in thirty years unless it could be annealed in orbit. | T3 | HIGH |
| CLM-1106 | The satellite's dark side charges to thousands of volts in the geostationary plasma, and the plastic backing of the solar cells would hold a few thousand volts across itself, within a factor of two of breaking down; the study's fix was a more conductive backing or a bleed path. | T3 | MED |
| CLM-1114 | The 1977 silicon design assumed an 18 per cent cell that did not yet exist, doubled the sunlight on it with cheap plastic mirrors, and got about 258 watts per square metre of blanket at the start of life, at a cell temperature of over 100 degrees. | T3 | HIGH |
| CLM-1115 | Mirrors that doubled the light only added 40 per cent to the output, because the cells ran hot and the mirrors were not flat; after thirty years of radiation the array was down to 59 per cent of where it started, and the mirrors helped even less. | T3 | HIGH |
| CLM-1116 | Thirty years of radiation behind a hair-thin glass cover was expected to take a third of the cell's power; a cover three times thicker would give about 16 per cent more power after thirty years but weigh twenty thousand tonnes more across the satellite. The design took the thin cover and planned to anneal. | T3 | HIGH |
| CLM-1117 | The 1977 study found the cheapest way to keep a silicon satellite at full power was to heal the radiation damage in place rather than build it bigger, that healing works best without mirrors because mirrors cannot be healed, and that a cheaper cell mattered more than a better one. That is why the reference system went to flat arrays and an annealing plan. | T3 | HIGH |
| CLM-1118 | In 1977 five radiation-damaged cells were healed with a laser or an electron gun, getting back between a fifth and a half of what they had lost; the electron gun did better and the laser missed parts of the cells. It was a first try, and everyone said so. | T3 | HIGH |
| CLM-1119 | Two years on, a laser could put back about half the radiation damage in a paper-thin cell in a few seconds, and an oven could put back nearly all of it in twenty minutes; but nobody could yet put a glass cover on such a cell that survived the heat, so the thing the satellite would actually carry had not been annealed. | T3 | HIGH |
| CLM-1120 | In 1977 the gallium arsenide option rested on a 20 per cent cell nobody had made, an assumption that it would still be near 20 per cent after thirty years because it heals more easily, and on films a few thousandths of a millimetre thick, because there was not enough gallium for anything thicker. | T3 | MED |
| CLM-1121 | Half a kilogram per square metre for the working blanket, about half of it the silicon itself; per kilowatt delivered on the ground the 1977 satellite came to six or seven kilograms, and the blanket alone to about two. | T3 | HIGH |
| CLM-1274 | NASA said its programme had built a test solar panel that it expected to lead to arrays more than three times lighter per watt than the best of the day, kept its power beams to about a tenth of midday sunlight for safety, and found that the very high voltages assumed in 1979 would not work in space. | T5 | MED |
| CLM-1280 | In July 2023 Europe's space agency asked researchers for ideas on the safety questions around beaming power from space: effects on people, wildlife, the atmosphere, aircraft and the carbon cost of launches. Ideas were due within four weeks, and the best would compete for jointly funded research projects. | T6 | HIGH |
| CLM-1316 | The world makes about 900 tonnes of raw gallium a year, almost all of it in China, and could make about 1,700 tonnes. Gallium is one ingredient of the high-efficiency solar cells some space designs assume. | T1 | HIGH |
| CLM-1317 | There is no reliable world total for germanium output, the material many space solar cells are built on, and its price roughly doubled in 2025. | T1 | HIGH |
| CLM-1318 | The world refines about 1,100 tonnes of indium a year, most of it in China. | T1 | HIGH |
| CLM-1418 | The standard ground sunlight used to test solar cells adds up to about 1,000 watts per square metre, or 1 kilowatt, on a surface tilted towards the sun, under one fixed set of atmospheric conditions. It is a reference value, not what any site receives. | T0 | MED |