Four Critical Unknowns: What ArXiv Research Tells Us About Asteroid Mining and Processing

Phase 0 of Project Dyson has 46 research questions. Four of them are classified as critical priority and share a common thread: they address fundamental physical processes that have never been demonstrated at the scales we need. We conducted a systematic literature review across arxiv to assess what is known, what is not, and where the real risks lie.

The short version: the literature is more encouraging than expected for excavation and silicon purification, but deeply sobering for anchoring reliability and industrial-scale metallurgy.

1. Regolith Excavation: Bennu Changes Everything

Question: How does asteroid regolith behave during microgravity excavation? (rq-0-6)

Until September 2023, every model of asteroid regolith behavior was based on remote sensing data and theoretical predictions. Then OSIRIS-REx dropped 120 grams of Bennu onto a Utah desert, and we got ground truth.

Lauretta et al. (arxiv:2404.12536) describe a material that is remarkably weak—highly porous, friable carbonaceous rock that crumbles under gentle handling. Bulk density of just 1.19 g/cm³ confirms what Murdoch et al. (arxiv:1306.1764) predicted in their microgravity regolith simulations: asteroid surface material is governed by cohesive forces (van der Waals, electrostatic) rather than gravity.

What this means for excavation:

• The good news: material strength is orders of magnitude below terrestrial soils, so excavation power requirements are modest
• The bad news: excavated particles do not settle back down. Active capture systems are mandatory
• The ugly: fine dust generation is virtually certain, and electrostatically charged particles will create persistent debris clouds

Nallapu et al. (arxiv:1701.07547) have optimized bucket wheel designs specifically for asteroid excavation, showing the concept is viable. But a critical finding from Frizzell & Hartzell (arxiv:2509.04074) complicates matters: lateral impulses on vacuum-exposed granular surfaces cause bulk dilation that propagates far beyond the disturbance zone. Every time an excavation tool contacts the surface, it may destabilize material meters away.

Confidence level: Medium. The physics is understood, excavation tools are designed, but no one has actually excavated anything in sustained microgravity. Parabolic flights give 20-25 seconds—we need experiments lasting hours.

2. Anchoring: Philae's Ghost Haunts Every Design

Question: Which anchoring technology works reliably across C-type, S-type, and M-type asteroids? (rq-0-7)

This is the question where reality bites hardest. The only time humanity tried to anchor to a small body—Philae landing on comet 67P in 2014—the harpoon system failed to deploy, the backup ice screws could not engage, and the lander bounced for two hours before settling in a shadowed crevice (Baranyai et al., arxiv:1604.04414). The mission was severely compromised.

Our mining robots face a harder problem: they must anchor and re-anchor thousands of times over a 5-year life while resisting continuous excavation reaction forces.

The three candidate technologies each have fatal gaps:

Microspine arrays (Ervin et al., arxiv:2502.12347) show the most promise. Passive compliant arrays adapt well to irregular surfaces and provide both normal and shear force resistance. But they need surface asperities to grip—fine regolith on C-type asteroids may not provide them.

Harpoon systems offer deep substrate penetration but are inherently single-use per deployment. Philae proved they have critical single-point failure modes. Redeployable designs add mass and complexity.

Gecko-inspired adhesives are the highest-risk option for primary anchoring. Dust contamination rapidly degrades adhesive performance, and electrostatically charged regolith will adhere to adhesive surfaces within minutes of operation.

Walsh's comprehensive review of rubble pile structure (arxiv:1810.01815) underscores the core difficulty: these asteroids are loosely bound aggregates with 20-40% porosity. The surface you are trying to anchor to may simply not be there.

Our recommendation: A hybrid multi-modal system with microspine arrays as primary, rotary drill anchors as secondary, and electromagnetic anchors for M-type asteroids. Total mass: 85-135 kg per robot (2.4-5.4% of mass budget).

Confidence level: Medium-Low. No anchoring technology has been tested on asteroid analog materials in microgravity, and no systematic comparison exists.

3. Microgravity Metallurgy: An Eight-Order-of-Magnitude Gap

Question: Can microgravity metallurgy scale from lab experiments to industrial production? (rq-0-11)

This is the question that keeps us up at night. The Material Processing Station needs to process 50,000 tonnes of raw asteroid material per year. The largest metal processing experiment ever conducted in microgravity handled about 100 grams.

That is a gap of roughly eight orders of magnitude.

D'Angelo et al. (arxiv:2102.09815) demonstrated a gravity-independent powder-based additive manufacturing process that works in microgravity—a genuine achievement. But this is precision manufacturing of small components, not bulk smelting of raw ore.

The solidification science is actually promising. Phase-field simulations (Dejmek et al., arxiv:0401250) predict that microgravity solidification can produce more uniform microstructures than terrestrial processing. Todaro et al. (arxiv:2008.04485) showed that ultrasound can substitute for gravity-driven convection in controlling grain structure—a potential breakthrough for microgravity metal quality.

But the fundamental challenges at scale remain unsolved:

• Melt containment: Electromagnetic containment works for grams. Containing tonnes of liquid metal at 1,500-1,800°C requires magnetic systems of unprecedented scale
• Slag separation: On Earth, slag floats because it is less dense than metal. In microgravity, it stays mixed. No alternative separation method has been demonstrated
• Thermal management: Processing 50,000 tonnes/year generates roughly 500 MW of waste heat. With no convective cooling, radiator requirements are measured in thousands of square meters

Our probability estimate for achieving 50,000 tonnes/year in pure microgravity: 15-25%. The contingency plan—incorporating centrifugal gravity sections into the modular station—should be carried as a parallel development track. This adds an estimated $2-5B and 100,000-200,000 kg to the station.

Confidence level: Low. The fundamental physics challenges are understood, but solutions are purely theoretical at industrial scale.

4. Silicon Purity: The Metallurgical Route Surprise

Question: What silicon purity is achievable in microgravity, and is it sufficient for solar cells? (rq-0-15)

This question produced our most encouraging finding. The conventional wisdom is that solar-grade silicon requires 99.9999% (6N) purity, achievable only through energy-intensive chemical processes like the Siemens method. In microgravity, without gravity-assisted zone refining, reaching 6N seemed nearly impossible.

Then we found Fornies et al. (arxiv:2101.08019).

Their work on Upgraded Metallurgical Grade silicon (UMG-Si) demonstrates that the metallurgical route alone—without chemical purification—can produce silicon at 4N-5N purity that makes functional solar cells at 15-18% efficiency. This was validated in mass production on commercial cell lines.

That changes the calculus entirely. Instead of chasing 6N purity with unproven microgravity zone refining, we can target 4N-5N using methods that are more adaptable to space:

1. Directional solidification (no gravity dependence): gets us from 2N to 3N-4N
2. Plasma purification (no gravity dependence): removes boron and phosphorus, reaching 4N-5N
3. Zone refining (gravity-dependent, TRL 1-2 in microgravity): optional polish to 5N-6N if developed

There is a catch: asteroid-derived silicon will be enriched in iron, nickel, and cobalt—the very impurities most damaging to cell efficiency (arxiv:2106.15926). The metallurgical purification chain must be validated specifically for asteroid feedstock compositions.

An additional finding softens the purity blow further: if radiation degrades cells by 20% over their lifetime regardless of initial purity (arxiv:1012.0717), then the difference between a 20% cell degraded to 16% and an 18% cell degraded to 14.4% is just 1.6 percentage points. Chasing 6N purity may yield diminishing returns in the L4/L5 environment.

Confidence level: Medium. The metallurgical route is proven on Earth. Adapting it to microgravity and asteroid feedstock is a tractable engineering challenge, not a fundamental physics barrier.

Summary: Where We Stand

Question	Literature Coverage	Key Risk	Confidence
Regolith excavation	Strong (Bennu data)	Debris cloud management	Medium
Anchoring reliability	Moderate (Philae + lab)	No proven multi-surface solution	Medium-Low
Metallurgy scaling	Weak (gram-scale only)	8 order magnitude gap	Low
Silicon purity	Strong (UMG-Si proven)	Asteroid feedstock validation	Medium

The research validates two of our core assumptions (excavation is feasible, silicon cells are achievable) while flagging two areas where contingency planning is essential (anchoring needs a hybrid approach, metallurgy may need artificial gravity sections).

All four questions have been updated to "investigating" status with "partially-resolved" resolution. The full literature analyses, including paper-by-paper reviews and research gap inventories, are available in each question's research folder.

What Comes Next

Each question's analysis identifies specific experiments that would most reduce uncertainty:

1. ISS excavation experiment: Sustained microgravity regolith excavation lasting hours, not seconds
2. Parabolic flight anchoring comparison: Systematic testing of all three anchoring approaches on C/S/M-type analog materials
3. ISS metallurgy scale-up: Graduated experiments from 100g to 10kg batch sizes
4. ISS float zone trial: Single most impactful experiment—zone refine silicon in microgravity and measure achievable purity

These four experiments would collectively cost a fraction of the Phase 0 budget while dramatically reducing the risk of committing to the wrong architecture.