The solar-pumped laser array

Why a laser and not a wire

The architecture that throws its losses away as light.

The Sun is an extended source, and brightness conservation forbids collimating sunlight directly into an interplanetary beam. The light has to be converted to coherent light. That conversion has losses, and in space a loss is a radiator.

Run the obvious chain — collect sunlight, convert to electricity, drive fiber lasers — at about 20 percent end to end, and delivering 100 gigawatts means collecting 500 and rejecting 400, most of it at the laser, where the gain medium sits near room temperature. That radiator is on the order of 100 square kilometers and 320,000 tons, roughly 600 times the collector.

And the radiator is the same size at 1 AU. In the electrically driven architecture Mercury’s location advantage is worth about one percent of system mass, and the array might as well orbit Earth. That is the sorting rule applied to the export product itself, and it is the correction that most changes the Mercury case.

The escape is the one the sorting rule predicts: use an architecture whose losses are rejected optically rather than thermalized. It is far less efficient end to end, so it needs a much larger collector — and the collector is exactly what Mercury’s 6.7-fold sunlight buys.

The laser

Real hardware at roughly 100 watts.

The device is a solar-pumped laser: a concentrator floods a doped crystal — neodymium-doped yttrium aluminum garnet, Nd:YAG, emitting at 1.064 micrometers — and the crystal lases directly off sunlight, with no electricity in between. The gain medium absorbs only its pump bands, and a cold mirror returns the rest to space. That selective absorption is the physical embodiment of the sorting rule, and it is why the solar-pumped route rejects so little heat.

This is real hardware rather than a concept. JAXA has studied solar-pumped lasers for space power since the late 1990s, and groups at PROMES-CNRS in France and in Japan have run Nd:YAG and cerium-codoped Ce:Nd:YAG rods for two decades. The demonstrated numbers land almost exactly where the architecture places them: solar-to-laser conversion around 3 to 5 percent, system slope efficiency touching 5, at the correct 1,064-nanometer line, in devices that have run continuously for a full day. The 5-to-8-percent target is a real stretch above the record and not a fantasy above it.

The gap is scale. The best solar-pumped lasers built to date put out on the order of 100 watts from a rod pumped by a few square meters of mirror. The Mercury array needs 100 gigawatts — achieved not by one enormous laser but by a field of collectors each feeding its own gain element, which turns the problem into the next one.

The phased array

A field of lasers is not a beam until it is phase-locked.

To act as one aperture, the outputs must be phase-locked — held in step to a fraction of a wavelength so the wavefronts add coherently in the far field. That is a phased array, and it steers by controlling relative phase rather than by moving any mirror.

The phasing is demonstrated, at a scale far below this. Coherent beam combining of fiber lasers is mature: a 100-kilowatt Nd:YAG array was phase-combined in the laboratory in 2009, combined ultrafast fiber systems now reach 10 kilowatts, and directed-energy programs scale to hundreds of channels held to a hundredth of a wave using stochastic-gradient and internally-sensed interferometric control.

So coherent phasing of many laser channels into a near-diffraction-limited beam is proven at roughly 100 kilowatts and tens to hundreds of elements. The Mercury array is a million times the power and asks for millions of elements phased across kilometers of structure in sunlight. That is a scale problem stacked on a metrology problem, not a new principle.

One architectural consequence the source analysis does not draw out: independent solar-pumped oscillators are not mutually coherent, so the real array is a master-oscillator/power-amplifier chain — one reference oscillator seeding a field of amplifiers — rather than a field of independent lasers. That is what the metrology requirement is actually describing.

The photonics

Two problems, and only one of them is on the array.

On the array, the load-bearing component is the spectrally selective optic: the dichroic mirror set that admits the pump bands to the gain crystal and reflects everything else back to space unabsorbed. This is what keeps the thermalized fraction small and lets the radiator collapse. The underlying coatings are ordinary cold-mirror technology; doing it over 137 square kilometers at Mercury’s flux is a manufacturing problem rather than a physics one.

The harder photonics problem lives on the sail, and it belongs only to the interstellar beam line rather than to the power-delivery array. A gram-scale sail pushed to a fifth of light speed must reflect a gigawatt per square meter while absorbing less than one part in 100,000 — anything more and it cooks. Worse, the band moves: as the sail recedes at 0.2c the beam redshifts by a factor of 1.2247, so a 1.06-micrometer laser arrives as 1.06 at ignition and 1.30 at burnout, and the sail must hold that low absorptivity across a 22 percent wavelength band, at several hundred kelvin, in a membrane about 100 nanometers thick, strong enough to survive 10,000 g.

The state of the art is close enough to be encouraging. In 2025 a Delft group fabricated a pentagonal photonic-crystal reflector 60 by 60 millimeters and 200 nanometers thick, with over a billion nanoscale holes placed by neural-topology optimization, cutting modeled fabrication cost per square meter by a factor of 9,000 — the highest-aspect-ratio nanophotonic element made to date. Caltech’s light-sail group works the companion problem of Doppler-broadened, self-stabilizing metasurface sails.

The sail line is dedicated hardware, not borrowed. What the interstellar shot inherits from Mercury is the collector field, the film industry that makes it, the metrology that holds it in phase, and the operators. The beam line it must build itself. See the sail.

Demonstrated against needed

Every subsystem exists; none at scale.

Every subsystem exists in the laboratory at the correct wavelength and on the correct physics. Not one exists within several orders of magnitude of the scale the array asks for, and two of them — the solar-pumped efficiency and the sail’s absorptivity across a moving band — are genuine open items rather than pure scale-ups. That is the honest state of it: a machine with no missing principle and no built precedent.

The gate with no sponsor

Bench-scale work, answerable this decade.

Three of the four technologies Mercury needs are being built by somebody else for their own reasons. High-power solar-electric propulsion arrives in the 2040s, paid for by cislunar logistics. Off-Earth thin-film manufacture at scale arrives in the 2040s and 2050s, paid for by orbital power build-out. High-power phased optical arrays with sub-wavelength metrology arrive in the same window, paid for by strategic directed-energy programs.

The fourth gate is Mercury’s own and has no sponsor: an optically rejecting laser conversion architecture above about 5 percent efficiency. The electrically driven beam is radiator-dominated and therefore location-indifferent, so Mercury’s entire advantage lives in an architecture whose losses leave as light. Solar-pumped laser efficiency is currently in the low single digits and no program anywhere is aimed at improving it, because there has been no application.