Four More Unknowns: Electrostatic Charging, Zero-G Refining, Cryogenic Storage, and Spectral Algorithms

Our first literature review tackled the four critical-priority unknowns in Phase 0: regolith excavation, anchoring reliability, metallurgy scaling, and silicon purity. This second review addresses four high-priority engineering questions that, while not existential threats, carry significant design risk. One of them produced a finding that may partially retire the question before any experiment is conducted.

1. Electrostatic Charging: The Invisible Saboteur

Question: How do electrostatic charging effects degrade mining robot mechanisms? (rq-0-9)

Mining robots operating on asteroid surfaces face continuous exposure to solar wind plasma, UV radiation, and triboelectric charging from regolith contact. Every articulating joint, every anchoring mechanism, every sensor surface is a target for charged dust accumulation.

The literature reveals the physics is well-understood but the specific mechanism contamination problem is unstudied.

Quan et al. (arxiv:2405.10744) model dust charging on airless bodies and show that photoelectron emission from solar UV creates predictable charge states that vary with mineral work function. Particles with certain work functions levitate higher and longer—and these are precisely the sub-micron particles most dangerous to precision mechanisms. Rosenfeld & Zakharov (arxiv:1706.09664) explain why: even surfaces with negligible average charge density contain submicron "chaotic spots" with intense local electric fields, sufficient to loft particles smaller than 0.1 microns. Standard seals designed for visible dust will not stop these particles.

On the mitigation side, Hao et al. (arxiv:2505.24074) demonstrate that low-energy electron beams can remove dust from surfaces even at -123C—a promising active cleaning technique for mining robot mechanisms. And Shin et al. (arxiv:2101.12485) provide a framework for predicting triboelectric charging between any two materials based on their thermoelectric properties, enabling systematic material selection to minimize charge transfer at mechanism interfaces.

The gap: Nobody has studied dust intrusion into enclosed articulated joints in a plasma environment. The physics of surface charging is mature; the engineering of mechanism protection is virgin territory.

Confidence level: Medium. We can predict the charging environment. We cannot yet predict mechanism degradation rates.

2. Zero-G Zone Refining: The Question That May Answer Itself

Question: What is the optimal zone refining process in zero-g? (rq-0-12)

This is where the literature produced a surprise. Zone refining—passing a molten zone along a silicon ingot to segregate impurities—is the traditional route to solar-grade (6N) silicon. In microgravity, the process is complicated by altered Marangoni convection, changed meniscus geometry, and the need for containerless electromagnetic processing. Wünscher et al. (arxiv:1102.3800) provide the terrestrial baseline for float zone silicon growth, and the deviations in microgravity are expected to be significant.

But then we looked at the UMG-Si data.

Forniés et al. (arxiv:2101.08019) demonstrate that upgraded metallurgical grade silicon at 4N-5N purity—produced through the metallurgical route alone, without any zone refining—achieves 20.76% cell efficiency in commercial production. Méndez et al. (arxiv:2102.11571) confirm the economics: 20% lower emissions, 25% better energy payback. This isn't a laboratory curiosity; it's mass-production validated.

The purification pathway that avoids zone refining entirely:

1. Directional solidification (gravity-independent): 2N to 3N-4N
2. Plasma purification (gravity-independent): removes B/P, reaches 4N-5N
3. Zone refining (gravity-dependent, needs microgravity R&D): optional polish to 5N-6N

Given that radiation at L4/L5 degrades all cells by ~20% regardless of initial purity (per rq-0-25), the difference between a 20% UMG cell degraded to 16% and a hypothetical 22% zone-refined cell degraded to 17.6% is just 1.6 percentage points. That may not justify the R&D investment.

This doesn't kill the question entirely—the metallurgical route still needs validation with asteroid-derived feedstock (high Fe/Ni/Co), which has different segregation chemistry than terrestrial quartzite. But it transforms the question from "how do we achieve the impossible" to "how do we optimize something that already mostly works."

Confidence level: Medium. The metallurgical route is proven on Earth. The asteroid feedstock validation is the remaining uncertainty.

3. Cryogenic Boiloff: Where Arxiv Runs Out

Question: How do we manage cryogenic propellant boiloff at L4/L5? (rq-0-30)

This was the most frustrating literature search of the four. Cryogenic fluid management is one of NASA's most active research areas, with decades of work on zero-boiloff storage, propellant depots, and MLI thermal performance. But almost none of it appears on arxiv—it lives in AIAA conference proceedings, NASA Technical Memoranda, and the journal Cryogenics.

What we did find was useful. Aasen et al. (arxiv:2501.18451) establish that thermal network models predict LH2 tank heat ingress with ~1% accuracy, and that the penalty for cooling to 20 K (vs. 77 K for nitrogen) increases heat ingress by less than 26%. Cheng et al. (arxiv:2508.21802) present ARCTIC, a lightweight AI framework for autonomous cryogenic fluid management validated against NASA test data—exactly the kind of system needed for autonomous station operation.

The metal hydride literature (Makridis et al., arxiv:1304.1959; Pasini et al., arxiv:1307.0890) offers an intriguing alternative: solid-state hydrogen storage that eliminates boiloff entirely. MgH2 nanocomposites achieve 7.6 wt% theoretical capacity. The problem is mass: storing 1 tonne of hydrogen requires 13-50 tonnes of hydride material, making it impractical for bulk propellant but potentially useful as an emergency buffer.

Our back-of-envelope thermal analysis suggests that with a 99%-effective deployable sunshield and 60-layer MLI, a 100-tonne LH2 tank at L4/L5 would see 50-200 W of heat ingress—manageable with active cryocoolers consuming 100-500 kW. Without the sunshield, heat ingress jumps to 5-10 kW and boiloff reaches 150-400 kg/day—catastrophic for a 70-130 tonne/year production rate.

Bottom line: The sunshield is the single most critical component of the cryogenic storage system. Everything else is engineering detail.

Confidence level: Low-Medium. The physics is straightforward, but the specific L4/L5 thermal environment, MLI long-duration performance, and cryocooler reliability at this scale are all uncharacterized.

4. Spectral Algorithms: The Best-Covered Question

Question: How do we validate asteroid composition algorithms before deployment? (rq-0-5)

Of the four questions, this one has the richest literature and the clearest path forward.

Mahlke et al. (arxiv:2203.11229) have built a state-of-the-art probabilistic asteroid taxonomy from 2,983 observations across 17 classes. Their system handles partial spectral data and quantifies uncertainty—exactly what an autonomous prospecting satellite needs. Korda et al. (arxiv:2210.01006) demonstrate that convolutional neural networks predict mineral composition from VNIR spectra within 10 percentage points of ground truth, including a critical finding: space weathering causes systematic olivine depletion in spectra that algorithms must correct for.

The OSIRIS-REx mission provides both the instrument benchmark (OVIRS at 17.8 kg and 8.8 W per Reuter et al., arxiv:1703.10574) and, for the first time in history, ground-truth validation data (Bennu samples per Lauretta et al., arxiv:2404.12536). We now have comprehensive remote spectra AND laboratory mineralogy of the same asteroid material—the ingredients for the first true algorithm validation.

Rizos et al. (arxiv:1903.11559) demonstrate spectral clustering tools validated on Ceres that map surface compositional variability, and Masiero et al. (arxiv:1406.6645) show that even simple NIR albedo measurements can sort asteroids into three distinct compositional groupings.

The gap is not algorithmic capability but validation methodology. No one has conducted a blind test where algorithms predict composition without access to ground truth. The pathfinder satellite phase (5 units before full constellation) is the natural venue for this, but only if the validation framework is designed now.

Our recommendation: Conduct a blind validation study using Bennu OVIRS data as input and returned sample mineralogy as ground truth. This can be done today, costs nothing beyond researcher time, and would establish the first quantitative accuracy baseline for asteroid composition algorithms.

Confidence level: Medium-High. The tools exist. The validation data exists. The gap is procedural, not fundamental.

Summary: A Mixed Bag, With One Gift

Question	Papers Found	Key Finding	Confidence
Electrostatic charging	8	Physics understood; mechanism engineering untested	Medium
Zero-g zone refining	6	UMG-Si may make zone refining optional	Medium
Cryogenic boiloff	6	Sunshield-critical; arxiv coverage thin	Low-Medium
Spectral algorithms	9	Tools ready; validation methodology needed	Medium-High

The zero-g zone refining finding is the standout. If the UMG metallurgical route validates for asteroid feedstock, it removes a major R&D dependency from the critical path. The cryogenic boiloff question highlighted a limitation of arxiv-centric literature review—for some engineering topics, the relevant literature lives in agency technical reports and conference proceedings.

All four questions have been updated to "investigating" status with research folders containing full paper-by-paper analysis. The research gaps identified in each analysis point directly to the experiments that would most reduce uncertainty.

What Comes Next

The highest-impact next steps from this review:

1. Blind Bennu validation study: Use existing OSIRIS-REx data to validate composition algorithms. Zero additional cost.
2. Triboelectric mechanism testing: Measure charge transfer between regolith simulants and mechanism materials in vacuum. Requires plasma chamber facility.
3. UMG-Si asteroid feedstock validation: Process silicon with Fe/Ni/Co impurity profiles through the metallurgical purification chain. Can be done in existing terrestrial facilities.
4. Sunshield thermal modeling: Model the L4/L5 thermal environment and optimize sunshield geometry for cryogenic storage. Computational study only.

Three of these four steps are achievable with existing facilities and data. The literature review's job is done when it points to experiments, and these experiments are actionable.