How Laser Sintering of Lunar Regolith Works

Lunar construction sounds like science fiction until you break it into its actual steps: condition the material, sinter it with precision, stack layers into walls, and put mass only where mass earns its place. Here is how Lunar Forge does each one - and why the details are where this problem is actually won.

Every build starts with feedstock. Raw regolith is collected at the construction site and passed through filtration and sieving to control particle size distribution (PSD) - our process feeds the sintering head with conditioned material in the 75-150 µm range - before any energy is applied.

This step looks unglamorous. It is the most important step in the chain.

Particle size distribution governs how regolith absorbs laser energy, how the melt pool forms, and how layers bond to each other. Uncontrolled feedstock produces unpredictable sintering - voids, weak bonding, inconsistent density. Conditioned feedstock produces repeatable, engineering-grade results. Lunar Forge conditions first, always, because structural material demands predictable inputs.

Conditioning also protects the machine itself: a filter-first process keeps abrasive fines away from optics and mechanisms, one of several ways feedstock discipline pays for itself downstream.

Vertical geometry is simple: stack thousands of horizontal rings and fuse them into one solid. Each layer follows the same six-beat cycle:

1. Degrain and feed. Sieved regolith (75-150 µm) flows to the sintering head.
2. Deposit the bead. The head traces the wall contour, laying down a thin bead 20-40 mm wide.
3. Laser sinter. The laser melts the bead - and the top of the layer beneath it.
4. Fuse and cool. The melt pool penetrates 1-3 mm into the previous layer, fusing the interface. Cooling in vacuum is mainly radiative.
5. Index up. The head rises 5-15 mm to the next layer height.
6. Repeat. Thousands of layers become a monolithic wall - not stacked bricks, but one continuous fused structure.

That interlayer remelt in step 4 is the entire ballgame. A wall is only as strong as its weakest interface, and controlled remelt depth is what makes a stack of layers behave as a single solid.

Closed-loop control runs the whole cycle. Sensors read the melt pool in real time and continuously adjust laser power, traverse speed, and remelt depth to the material actually in front of the beam - because regolith varies by region, by depth, by scoop, and open-loop systems fail when the feedstock changes. The system sinters the regolith it has, not the regolith a model assumed. In-process sensing doubles as quality assurance: dimensional accuracy is verified as the wall grows, not inspected after the fact.

Why vertical is easy - and overhangs are harder. A vertical wall is self-supporting: every layer sits fully on the one below. Overhangs and domes require corbeling - stepping each layer inward - and in lunar conditions the self-support angle is roughly 50-60°. This is why our structures favor vertical shells, buttressed footings, and sintered caps: geometry chosen to work with the physics, not against it.

Here is the insight that changes lunar construction economics: a thick, high-performance wall does not need to be solid sintered material.

Reactor shielding needs mass. Thermal buffering needs mass. But mass does not need to be laser-fused - it just needs to be contained. So Lunar Forge builds thick structures as a sandwich shell:

• Outer shell - sintered regolith, 8-12 cm typical
• Loose regolith fill - raw regolith placed by rover, providing the bulk mass for radiation shielding and thermal inertia
• Inner shell - sintered regolith, 8-12 cm typical
• Periodic sintered ties - internal cross-members that connect the shells, control wall bowing, and provide mounting points

The laser only sinters the thin shells and ties. The fill - the overwhelming majority of the wall's mass - is simply placed, orders of magnitude faster than sintering and at near-zero energy cost. The result: 70-85% less laser energy than a solid sintered wall, with equal shielding mass per unit area, superior thermal insulation, and superior impact resistance - loose regolith absorbs micrometeorite strikes and dissipates energy in ways a solid wall cannot.

The lift-by-lift build sequence:

1. Sinter the inner shell to the next lift height
2. Place fill - rover deposits loose regolith between shells
3. Sinter the outer shell to the same height
4. Repeat in 0.5-1 m lifts to full height
5. Overfill and settle - slight overfill allows for settlement from vibration and moonquakes
6. Sinter the cap - a final fused layer with penetration collars, sealing the wall and locking down dust

This is the architecture behind our anchor product: reactor housings - radiation shielding and containment - along with heat-management structures and load-bearing bases. Strength where you need it, minimum energy, built for the Moon.

Sintered under lunar vacuum, the walls perform as engineering material, not experiment:

• Compressive strength: 200-345 MPa - in the range of high-performance concrete
• Density: ~1.6-1.9 g/cm³
• Thermal cycling: designed for -170°C to +110°C, the full lunar day-night swing
• Vacuum stable: no binders, no water, no outgassing, no weakening - the vacuum is part of the recipe, not an obstacle
• Typical geometry: 300 mm minimum wall thickness, 500-800 mm footings where loads demand

And not every structure needs a wall. For roads and dust mitigation, the same system surface-sinters a 1-2 cm crust that locks down loose regolith along traffic routes - fast, low-energy, and operationally transformative, because dust is the most pervasive hazard in lunar operations. Landing pads get engineered sintered surfaces that eliminate plume ejecta, the high-velocity dust blast every unprepared landing throws at everything nearby.

One system. One process. Recipes matched to the structure.

Wall height is a kinematic question, and the answer scales with the platform:

• Rover arm - structures to 2-3 m: reactor housings, berms, foundations
• Mobile gantry - taller builds to roughly 10 m
• Climb-and-build systems - tens of meters, for the structures a growing surface economy will eventually demand

The process does not change. The reach does.

We publish our failure modes, because a construction company that cannot name its risks has not done the engineering:

• Interlayer delamination from thermal cycling → controlled remelt depth, uniform particle size, optimized thermal profiles
• Dimensional drift across thousands of layers → closed-loop height control, vision metrology, periodic calibration
• Optics contamination from dust and spatter → filter-first feedstock, airflow isolation, self-cleaning optics
• Thermal stresses in large structures → buttresses, expansion joints, gradual heat management
• Fill settling in sandwich walls → overfill margins and sintered caps, topped off where needed

Every mitigation above is a design feature of the system, not a procedure bolted on afterward.

Lunar Forge's operational logic follows a strict sequence, because on the Moon, energy is the constraint that rules all others:

1. Reactor housing first. Forge systems arrive operating on their own solar power - independent of any surface reactor, with radioisotope heat carrying them through the lunar night - and sinter the shielding and housing for the customer's fission power system.
2. Reactor online. The customer's reactor installs into prepared infrastructure and begins delivering continuous power.
3. Everything else. With surface power available, the build-out accelerates: pads, roads, berms, and foundations, each cheaper and faster than the last.

This sequence is why our power architecture matters. A construction system dependent on reactor power cannot build the first reactor site. Ours can - and does it first.

Lunar Forge builds fully autonomous construction robots. Adaptive sintering is not remote-controlled printing - the closed-loop system that reads the melt pool and adjusts laser parameters thousands of times per build is the machine making its own decisions, in real time, against material no operator could evaluate fast enough. Excavation, conditioning, deposition, and sintering run as an autonomous cycle.

Humans supervise from Earth. The Moon's proximity - seconds of signal delay, not the minutes of Mars - means our operators watch builds in real time, review quality data, and intervene when judgment calls arise. But supervision is the exception path, not the operating model: the robots build, and they build through the lunar day whether anyone is watching or not.

This is what construction before crews arrive actually requires. Infrastructure must exist before people do - that is the point of infrastructure - and only autonomous systems can deliver it.

If you are landing hardware on the Moon in the next five years, the process above is your site preparation supply chain:

• Shielding and housing built before your reactor arrives - sandwich-shell mass at a fraction of the energy budget
• Pads ready before your lander cadence scales
• Roads and crusts wherever your operations move
• All of it from local material, none of it on your launch manifest