# Lunar resources Lunar resources are the materials the [[Moon]] already holds that could be used in place rather than lifted from [[Earth]]: water ice cold-trapped in the polar shadows, [[Oxygen|oxygen]] bound as oxide in every grain of [[Regolith|regolith]], the [[Iron|iron]], [[Titanium|titanium]] and [[Aluminium|aluminium]] of the mare basalts and highland anorthosites, bulk soil as [[Construction_engineering|construction]] and shielding feedstock, and — the one that draws the headlines — [[Helium-3|helium-3]] implanted by the solar wind. This article gives the first four an honest section each and then puts its analytical weight on the fifth, because helium-3 is the claim in this cluster that most needs auditing: the [[Nuclear_fusion|fusion]] [[Physics|physics]] is sound, the [[Geometry|geometry]] is unforgiving, and several numbers in wide circulation are wrong by factors a reader can check with arithmetic. ## Microsims — three.js <iframe src="https://wikitube-3d-microsims.netlify.app/Lunar_resources.html" width="100%" height="620" frameborder="0" loading="lazy" sandbox="allow-scripts allow-same-origin" title="Lunar resources — three.js microsim"></iframe> **`Lunar_resources` (three.js).** This sim makes the helium-3 arithmetic fall out of geometry rather than out of a lookup table: a cutaway regolith grain carries a 30 nm implanted rind — exaggerated 100× and labelled as such — inside a window of fixed 200 µm width, so dragging the grain-size slider down visibly shrinks the grains while the rind does not move, and the grade rises. Take the mean grain size from 160 µm to 20 µm and the HUD's live `[He-3]` readout climbs from 3.50 ppb through 8.00 ppb at the 70 µm default to 27.9 ppb, while the derived areal density — helium-3 atoms per cm² of grain surface — stays identical across all three, which is the proof that this is a genuine surface model and not a fitted curve. Solar-wind tracers are ray-cast against the grain cluster, so 100% of early implantation lands on sunward faces and only impact gardening turns the grains over, which makes the maturity argument something measured rather than asserted. Switch to the site-scale view, run the excavator, and the HUD totals tonnes moved against grams recovered, and thermal energy in against fusion energy out. The verdict genuinely swings across the sourced range instead of being pre-cooked: **6.5× net positive at 8 ppb and a 700 °C release temperature, marginal at 1.4 ppb, and net negative at 1.4 ppb with the release temperature pushed to 900 °C.** At its defaults it prints the number that matters most — **16.3 years and 26.2 km² of continuous mining to win one tonne of helium-3**. The ledger it draws is deliberately partial, and its own header says so: it weighs the thermal energy needed to heat the processed regolith against the fusion energy of the product, with no capital energy, no hauling, beneficiation, cryogenic separation or transport, and no heat recuperation. It is a tool for thinking about the arithmetic, not a resource assessment. ## Overview: a resource is not a reserve Almost every disagreement here reduces to one distinction from terrestrial [[Mining_engineering|mining engineering]]. A **resource** is material inferred present; a **reserve** is the part demonstrated extractable under stated grade, depth, recovery, minability and an [[Energy|energy]] budget that closes. On the [[Moon|Moon]] the two differ by orders of magnitude, and that gap is where the optimism lives — [[Accuracy_and_precision|precision]] in a headline is no substitute for saying which you mean. The resources also split physically. Water ice, oxides and metals are *bulk* materials whose grade is set by [[Chemistry|chemistry]] and [[Geomorphology|geological]] history; solar-wind volatiles — [[Hydrogen|hydrogen]], [[Carbon|carbon]], [[Nitrogen|nitrogen]], [[Helium-4|helium-4]] and [[Helium-3|helium-3]] — are *surface* materials painted onto grain exteriors by an external flux. Everything economically strange about helium-3 descends from that difference. ## Water ice at the poles The Moon's axis is tilted about 1.5° to the ecliptic, so polar crater floors never see the [[Sun|Sun]]. Diviner measured annual maximum temperatures in these permanently shadowed regions below roughly 110 K, the coldest floors under 40 K — enough to suppress [[Phase_transition|sublimation]] for billions of years. LCROSS's 2009 impact into Cabeus gave the strongest direct evidence: **5.6 ± 2.9 wt%** water ice in the excavated [[Regolith|regolith]] of one site, an uncertainty half the value and no basis for a polar average. Orbital [[Sensor|spectrometers]] map exposed ice patches; SOFIA found molecular water at 100–412 ppm in sunlit Clavius. Water is life support, shielding and — by [[Hydrogen_production|electrolysis]] into [[Hydrogen|hydrogen]] and [[Oxygen|oxygen]] — [[Rocket_propellant|rocket propellant]]: the shortest chain from ore to product on the Moon, and the least [[Estimation_theory|characterised]], since no lander has returned a bulk grade from a mining depth. ## Oxygen bound in regolith oxides [[Oxygen|Oxygen]] is the most abundant element in the [[Crust_(geology)|lunar crust]], roughly 40–45% of regolith by mass, all of it chemically bound — mostly [[Silicon_dioxide|silica]] and oxides of [[Iron|iron]], [[Aluminium|aluminium]], [[Magnesium|magnesium]], [[Calcium|calcium]] and [[Titanium|titanium]]. Freeing it is a [[Chemical_engineering|chemical-engineering]] problem, not a prospecting one, and the [[Process_engineering|routes]] are established: hydrogen reduction of ilmenite (FeTiO₃ + H₂ → Fe + TiO₂ + H₂O, then electrolysis of the water), [[Chemical_reaction_engineering|carbothermal reduction]], and molten-regolith or molten-salt electrolysis. All run hot and all are [[Energy_engineering|energy-hungry]], tying them to power supply rather than ore grade. Oxygen is most of the mass of both a propellant load and a [[Breathing_gas|breathing-gas]] supply, so making it locally deletes the largest term from an Earth-launched [[Logistics|logistics]] budget — the whole argument for [[In_situ_resource_utilization|in-situ resource utilisation]]. ## Metals: iron, titanium, aluminium Mare basalts carry high [[Iron|iron]] and [[Titanium|titanium]]; ilmenite, the iron–titanium oxide of the high-titanium maria, is at once the best oxygen feedstock, the best metals feedstock and the best helium host. Highland anorthosites, dominated by plagioclase feldspar, supply [[Aluminium|aluminium]] and [[Calcium|calcium]], with [[Chromium|chromium]], [[Nickel|nickel]] and [[Sulfur|sulfur]] minor. Regolith also holds free metallic iron — meteoritic grains and the nanophase iron of space-weathered agglutinates — separable magnetically with no [[Metallurgy|metallurgical]] step. As [[Materials_science|materials science]] this is unremarkable: ordinary [[Iron_ore|ores]], poorer than many terrestrial ones. What matters is the transport differential: the comparison is never a mine on [[Earth|Earth]] but the delivered cost of the same [[Alloy|alloy]] or [[Steel|steel]] landed from it. ## Bulk regolith as construction feedstock The least glamorous lunar resource is the most certain. Regolith heaped over a habitat is radiation and micrometeoroid shielding, and [[Density|bulk density]] is the only property required. Sintered or microwave-fused soil makes pads, roads and blocks and feeds additive [[Manufacturing|manufacturing]] of [[Ceramic_engineering|ceramic]] and [[Structural_engineering|structural]] elements. No beneficiation, no grade, no reserve calculation — which is why every [[Construction_engineering|construction]] architecture from a buried habitat to a [[Stanford_torus|Stanford torus]] assumes it, and why it never appears in the pitches. ## Helium-3 is on the Moon, not in it The [[Moon|Moon]] has no global magnetic field and no atmosphere, so for ~4 billion years the solar wind — a [[Plasma_(physics)|plasma]] streaming off the [[Sun|Sun]] — has struck bare mineral surfaces directly. Solar-wind helium arrives as [[Alpha_particle|doubly charged ions]] near 1 keV per nucleon, about 4 keV per [[Ion|ion]] at wind [[Velocity|speeds]] of 300–800 km/s, and stops within tens of nanometres. [[Materials_science|Atom-probe tomography]] of real lunar ilmenite puts the maximum at **30–50 nm with a tail to ~200 nm**; a [[Mathematical_model|model]] of ³He into ilmenite at 1 keV/amu gives **26 nm**, code output rather than [[Accuracy_and_precision|measurement]]; space-weathered rims reach ~**120 nm**. There is no reservoir at depth. Every ³He [[Atomic_mass|atom]] sits in a rind on the *outside* of a grain tens to hundreds of micrometres across — a skin of order 0.1% of its radius. > **Helium-3 concentration scales with grain SURFACE AREA, not with volume.** For a grain of diameter *D* and rind thickness *d*, the rind is a volume fraction 1 − (1 − 2*d*/*D*)³, which for *d* ≪ *D* is 6*d*/*D*: halve *D* and the grade doubles. That is [[Geometry|geometry]], not [[Chemistry|chemistry]], and it makes fine, mature, ilmenite-rich mare [[Regolith|soil]] the ore and coarse fresh ejecta the waste — about 88% of the helium in Apollo 11 soil sat in the sub-100 µm fraction. Cameron's 1992 assessment fixes both controls: [[Titanium|TiO₂]]-rich samples are helium-rich, because helium is retained in ilmenite rather than the surrounding silicate [[Crystal_structure|lattices]]; and maturity enters through the ferromagnetic-resonance index Iₛ/[[Iron|FeO]], with (Iₛ/FeO)·TiO₂ predicting grade better than TiO₂ alone. Grade is roughly independent of depth in the [[Porous_medium|regolith column]], because impact gardening homogenises it while [[Heat_transfer|thermal]] [[Diffusion|diffusion]] does nothing at lunar surface temperatures — which is why maturity is the right variable: a mature soil has had every grain face turned to the Sun. Retention is unsettled too; the same study found only ~0.5‰ of the *expected* [[Helium-4|⁴He]] in the vesicle it probed. ## What the samples actually measure There is no such thing as "the" lunar helium-3 abundance. A source printing one number without a sample name and a range is not reporting a [[Accuracy_and_precision|measurement]]. | Sample or class | ³He, ppb by mass | Source | |---|---|---| | Apollo 11 bulk soil **10084** | **11.8 mean; range 9.22–17.9** | Olson 2021 | | Category I, high-Ti mare | **15.1** | Slyuta et al. 2007 | | Category II, moderate-Ti mare | **8.0** | Slyuta et al. 2007 | | Category III, low-Ti mare | **5.7** | Slyuta et al. 2007 | | Category IV, **highlands** | **3.1** | Slyuta et al. 2007 | | Mare Tranquillitatis working figure | "at least **13**"; undisturbed Ti-rich **20–30** | Schmitt et al. 2006 | | Sunlit terrain, commonly cited | **1.4–15** | Wikipedia refs, `[UNVERIFIED at primary]` | The defensible range is **1.4–20 ppb**, mature high-[[Titanium|titanium]] mare soils at the top and highlands at the bottom, exactly as the [[Geometry|geometry]] predicts. The [[Estimation_theory|estimation]] consequence is severe: the highlands are most of the [[Moon|Moon's]] area and sit at the low end, so any global figure applying a mare grade everywhere is inflated before a single further assumption. ## Resource, reserve, and a factor of a thousand | Estimate | Value | Method | Source | |---|---|---|---| | Low | **660,000 t** (nearside 3.7×10⁸ kg, farside 2.9×10⁸ kg) | Chang'E-1 [[Microwave_engineering|microwave]] radiometry to a regolith-thickness map | Fa & Jin 2010 | | Canonical | **~1,000,000 t** | Uniform ~10 ppb over the whole surface, 3 m depth, **no minability discount** | Wittenberg, Santarius & Kulcinski 1986 | | High | **2.47 Mt total; 1.28 Mt nearside** | Four TiO₂ categories (3.1–15.1 ppb) × regolith thickness **4.4–10.1 m** | Slyuta et al. 2007 | | **Mineable reserve** | **7,041 t** (2,726 t high-Ti + 4,315 t medium-Ti) | **50% minability**, 3 m, one *defined region*: Mare Tranquillitatis | Cameron 1992 | | Reserve (Schmitt) | **≥5,000 t**, ~25 t/km² to 3 m | 84,000 km² of highest-grade Mare Tranquillitatis | Schmitt et al. 2006 | The million-tonne figure is not fabricated: 3.79×10⁷ km² × 3 m × 1700 kg/m³ ≈ 1.9×10¹⁷ kg of [[Regolith|regolith]], which at 10 ppb is 1.9×10⁶ t. The inflation is the leap from resource to reserve, and its size is the last two rows — **Cameron's mineable figure is three orders of magnitude below the headline.** Four things drive the high end: [[Mathematical_model|assumed]] regolith depth, where 3 m against Slyuta's 4.4–10.1 m is a 1.5–3.4× lever; applying mare or high-Ti grades globally instead of area-weighting highlands at 3.1 ppb; quoting a resource with no minability discount and letting readers hear a reserve — the commonest move in this literature, and the one Cameron alone refuses; and speculative polar enhancement, where Schmitt's "may triple… above 70° latitude due to cold trapping" is a hypothesis, not a [[Sensor|measurement]], and the up-to-50 ppb polar figure downstream of it has never appeared in a returned sample. ## Auditing "25 tonnes powers the United States" The most quoted sentence in the subject is Harrison Schmitt's, from Space.com on 30 June 2000: *"The equivalent of a single space shuttle load or roughly 25 tons could supply the entire United States' energy needs for a year."* **As stated it is false by a factor of about seven** — and no step that produced it was dishonest, which is what makes it worth tracing. Start from the [[Energy|energy]] content, computed rather than quoted. One tonne of ³He is 1.997×10²⁹ nuclei; at the D–³He [[Binding_energy|Q-value]] of 18.353 MeV that is **5.87×10¹⁷ J = 0.587 EJ = 163 TWh thermal = 18.6 MW·yr per kilogram**, matching the ~19 MW·yr/kg of the [[Mining_engineering|mining]] literature. So 25 t is **14.7 EJ ≈ 13.9 quads thermal** against US primary energy consumption of **~94–96 quads**: about **15%** of a US year, not all of it. The mutation runs in three steps. Schmitt's 1988 paper states the load-bearing assumption openly — "High efficiency (70–80 percent) in energy conversion due to direct conversion of charged particles to electricity." Cameron in 1992 wrote that 25 t "contains sufficient energy to supply U.S. requirements of **electrical** energy for one year at the 1991 rate of use," defensible under that assumption. The Space.com version dropped the word *electrical*, and once it is gone a [[Thermodynamics|thermal]] quantity is being compared with a primary-energy one. Corrected: - US [[Electrical_grid|electricity]] generation ≈ 4,200 TWh ≈ 15 EJ **electrical**; 25 t → 4,080 TWh electric **only at 100% conversion**. - At Schmitt's own 70–80% direct conversion you need **31–36 t**; at a conventional 40% thermal cycle — the [[Second_law_of_thermodynamics|second law]] on a heat engine — **62 t**. - **One tonne alone** = 163 TWh thermal → 65–130 TWh [[Electric_power_transmission|delivered]] electric → **about 1.5–3% of one year of US electricity.** Four assumptions sit in every version: (a) 70–80% direct [[Energy_transformation|energy conversion]] of the charged [[Nuclear_fusion|fusion]] products, never demonstrated at reactor scale; (b) complete burn-up; (c) thermal energy silently equated with delivered electricity; (d) zero energy cost for moving and processing the ~10⁸ tonnes of regolith one tonne of ³He requires. ## A units slip worth correcting The 2021 *Joule* paper of De Temmerman, Chuard and Rudelle is the best sceptical treatment of ³He as a [[Nuclear_fuel|fusion fuel]] — source for terrestrial supply of ~18 kg/y from tritium [[Beta_decay|beta decay]], a US strategic reserve near 25 kg, and mantle leakage of "not even a few kilograms per year." It also carries a plain [[Accuracy_and_precision|units]] error: its text gives [[Regolith|regolith]] ³He abundance as **"about 30 micrograms per gram."** Thirty µg/g is 30 **ppm** — three orders of magnitude too high for helium-3, and about right for *total* [[Helium|helium]], overwhelmingly [[Helium-4|⁴He]]. That it is a text slip and not a data error is provable from the paper itself: its own ~1 Mt lunar total is consistent with ~10 ppb, whereas 30 ppm implies an inventory a thousand times larger than any published [[Estimation_theory|estimate]]. Cite the conclusions, not that sentence. ## Getting it out: release temperature and throughput Helium [[Diffusion|degasses]] on heating, and the curve is measured. For Apollo soil 10084, about **50% of the implanted ³He is released by 500 °C and about 95% by 800 °C**; De Temmerman et al. give 300–900 °C as the window with 700 °C sufficient; [[Molecular_dynamics|molecular-dynamics]] modelling of ilmenite puts the optimum near 1000 K (727 °C) at ~74% cumulative release. So: **most helium comes off between 500 and 800 °C, and ~700 °C recovers the bulk.** A non-thermal route exists too — 70–96% of implanted ⁴He removed by mechanical agitation without bulk [[Heat_transfer|heating]] — which, if it holds for ³He, deletes the ledger's largest term. Either way the product needs [[Cryogenics|cryogenic]] [[Fractional_distillation|separation]] from a stream overwhelmingly ⁴He and hydrogen, assayed by [[Helium_mass_spectrometer|mass spectrometry]]. The [[Thermal_engineering|thermal]] arithmetic is what the ledger turns on. At 8 ppb — Slyuta's moderate-[[Titanium|Ti]] mare value, deliberately not the optimistic end — one tonne of regolith holds **8 milligrams** of ³He, so one gram needs **125 tonnes** of soil at perfect recovery and one tonne needs **1.25×10⁸ tonnes** before losses. Soil specific heat runs near 771 J/(kg·K) at 300 K, rising toward a [[Thermodynamics|Dulong–Petit]] limit near 1130 J/(kg·K); heating ~1.4×10⁸ tonnes through ~700 K costs of order 10¹⁷ J against the 5.87×10¹⁷ J of fusion [[Energy|energy]] recovered. Net positive — the microsim measures 6.5× at 8 ppb and 700 °C — but a ratio, not a margin, and it inverts *inside* the sourced range: at 1.4 ppb with a 900 °C release it goes negative. At 1,000 t/h continuously that is **16.3 years** and **26.2 km²** stripped to 3 m for one tonne. Throughput, not grade, is the binding [[Limiting_factor|limiting factor]]. ## Where it stands in 2026 Essentially all commercial helium-3 is the beta-decay daughter of tritium ([[Half-life|half-life]] 12.32 ± 0.02 y, ~5.5%/y), harvested from [[Nuclear_engineering|nuclear-weapons]] reservoirs — in the US from NNSA operations at Savannah River, a [[Cold_War|Cold-War]] inheritance rather than an industry. GAO's audit of the 2008 shortage put annual US extraction capacity at **8,000–10,000 L/y** against average sales of ~30,000 L/y over 2003–2009, and found the failure was governance, not geology. Federal demand has since fallen to a projected **under 6,000 L/y**. None of this resembles the [[National_Helium_Reserve|reserve]] logic governing [[Helium_production_in_the_United_States|helium production]] from [[Natural_gas|natural gas]], because ³He is co-produced with nothing and cannot be [[Helium_storage_and_conservation|stockpiled]] into existence. Against that background, the **Bluefors–Interlune agreement of 16 September 2025** commits Interlune to supply up to **10,000 L/y of lunar helium-3 from 2028 to 2037**, up to 100,000 L in all — roughly the *entire* historic annual US extraction capacity, and a purchase commitment against a resource nobody has extracted, from a body nobody has mined, by a process never run outside a laboratory. Two points about that demand deserve stating plainly. The near-term driver is not [[Nuclear_fusion|fusion]] but [[Dilution_refrigerator|dilution refrigerators]] for [[Quantum_computing|quantum computing]], plus [[Neutron_detection|neutron detection]] and [[Helium_cryogenics|low-temperature]] research, with [[Hyperpolarization_(physics)|hyperpolarized]] gas imaging now a shrinking niche. And a fridge's ³He charge is a **one-time inventory cost**: the isotope circulates in a sealed loop through the [[Superfluidity|superfluid]] [[Superfluid_helium-4|⁴He]] dilute phase and is not consumed, so a fleet of machines is a stock to fill once, not an annual burn. Quoting fridge demand in "kilograms per year" as though it were fuel turns a capital charge into a perpetual revenue line — a conflation lunar-mining pitches make routinely. The market-size and per-litre price figures behind such projections trace, on audit, only to secondary compilations, and are not reproduced here. ## The case for and the case against The case against is arithmetic, not prejudice. The grade is 8 milligrams per tonne at a representative mare value, an ore-to-product mass ratio near 10⁸. The published totals are resources; the one carefully computed *reserve* is 7,041 t in a named region. The famous [[Energy|energy]] claim is off by seven as stated and works only under an unbuilt conversion technology. And D–³He has never been ignited: it needs roughly an order of magnitude more triple product than D–T in a [[Plasma_(physics)|plasma]], and it is not truly [[Aneutronic_fusion|aneutronic]] — D–D side reactions put ~1% to ~20% of the power into [[Neutron|neutrons]] depending on fuel ratio, temperature, and whether bred tritium is burned or extracted. The case for rests on assumptions that are stateable and testable, which is more than most speculative resources manage. Direct conversion is physically real here: D + ³He → [[Proton|p]] (14.64 MeV) + [[Alpha_particle|α]] (3.71 MeV) puts *all* the primary yield into charged particles, so 70–80% conversion is a serious [[Nuclear_engineering|engineering]] target rather than a fantasy, merely an undemonstrated one — and the commonly quoted 14.68/3.67 split is only the crude mass-number approximation to it. The same reaction underwrites the [[Fusion_rocket|fusion rocket]] literature. The ore sits at the surface: no drilling, no overburden. [[Sun|Sunlight]] is free and continuous for two weeks at a time, exactly what a heat-driven process wants. Beneficiation by grain size and ilmenite separation is ordinary mineral [[Chemical_reaction_engineering|processing]] and could lift feed grade by nearly an order of magnitude, which the microsim deliberately does not model. And the same plant liberates [[Hydrogen|hydrogen]], water, [[Nitrogen|nitrogen]], [[Carbon|carbon]] and [[Helium-4|⁴He]] from the same heated soil — the [[In_situ_resource_utilization|ISRU]] synergy that may carry the [[Systems_engineering|system]] economics whether or not any ³He reaches a reactor. Elsewhere the isotope is more abundant and less reachable: [[Jupiter|Jupiter]] dwarfs the lunar inventory, and neither [[Mars|Mars]] nor [[Venus|Venus]] offers the airless, unmagnetised surface that lets the Moon collect solar wind at all. Every one of those propositions is testable, none has been tested on the Moon, and the burden of proof sits with whoever quotes a single number to three significant figures. ## Sources Figures in the helium-3 sections are drawn from the portal's audited source review; the four non-helium-3 sections rest on the primaries listed at the end, which lie outside that audit and are marked accordingly. - **Wittenberg, L. J.; Santarius, J. F.; Kulcinski, G. L. (1986).** "Lunar Source of ³He for Commercial Fusion Power." *Fusion Technology* **10**(2), 167–178. DOI [10.13182/FST86-A24972](https://doi.org/10.13182/FST86-A24972). — Founding paper of the lunar-³He proposition and origin of the ~10⁹ kg figure; the internal resource number is `[UNVERIFIED at primary]`, attributed secondhand via Olson 2021 and Santarius' own presentations. - **Cameron, E. N. (1992).** *Helium Resources of Mare Tranquillitatis.* WCSAR-TR-AR3-9207-1. [PDF](https://fti.neep.wisc.edu/fti.neep.wisc.edu/pdf/wcsar9207-1.pdf). — The most rigorous Wisconsin assessment; source of He(total)/³He = 2600, of the ilmenite and (Iₛ/FeO)·TiO₂ maturity controls, and of the only true *reserve* in the literature: 7,041 t at 50% minability. - **Schmitt, H. H. (1988).** "Economic Geology of Lunar Helium-3." *Second Symposium on Lunar Bases*, Houston. [NTRS 19890005478](https://ntrs.nasa.gov/api/citations/19890005478/downloads/19890005478.pdf). — States the 70–80% direct-conversion assumption explicitly; essential for auditing every downstream "X tonnes powers Y" claim. - **Schmitt, H. H. (2000).** Space.com, 30 June 2000. [FTI copy](https://fti.neep.wisc.edu/fti.neep.wisc.edu/gallery/pdf/space_com063000.pdf). — Primary text of the "25 tons could supply the entire United States' energy needs for a year" claim audited above. - **Schmitt, H. H.; Henley, M. W.; Kuhlman, K.; Kulcinski, G. L.; Santarius, J. F.; Taylor, L. A. (2006).** "Lunar Helium-3 Fusion Resource Distribution." In *Return to the Moon*. [LPI decadal copy](https://www.lpi.usra.edu/decadal/leag/DecadalHelium3.pdf). — Source of the "at least 13 ppb" working figure, the ~25 t/km² and ≥5,000 t recoverable estimates, and the polar cold-trapping *hypothesis*. - **Slyuta, E. N.; Abdrakhimov, A. M.; Galimov, E. M. (2007).** "The Estimation of Helium-3 Probable Reserves in Lunar Regolith." *LPSC XXXVIII*, abstract 2175. [PDF](https://www.lpi.usra.edu/meetings/lpsc2007/pdf/2175.pdf). — The 3.1/5.7/8.0/15.1 ppb TiO₂ bins, the 2.47 Mt high-end total, and the point that estimates built from *enriched fractions* rather than bulk regolith are unsuitable. - **Fa, W.; Jin, Y.-Q. (2010).** "Global inventory of Helium-3 in lunar regoliths estimated by a multi-channel microwave radiometer on the Chang-E 1 lunar satellite." *Chinese Science Bulletin* **55**(35), 4005–4009. DOI [10.1007/s11434-010-4198-9](https://doi.org/10.1007/s11434-010-4198-9). — The low-end, remote-sensing-constrained total of 6.6×10⁸ kg. - **Johnson, J. R.; Swindle, T. D.; Lucey, P. G. (1999).** "Estimated solar wind-implanted helium-3 distribution on the Moon." *Geophys. Res. Lett.* **26**(3), 385–388. DOI [10.1029/1998GL900305](https://doi.org/10.1029/1998GL900305). — Clementine-based distribution map, highest in farside maria and high-TiO₂ nearside maria; deliberately gives no global total. - **Olson, A. D. S. (2021).** "Lunar Helium-3: Mining Concepts, Extraction Research, and Potential ISRU Synergies." AIAA ASCEND 2021. [NTRS 20210022801](https://ntrs.nasa.gov/api/citations/20210022801/downloads/AIAA%20ASCEND%202021%20Paper_211018.pdf). — Best modern compilation of the Apollo numbers, the 10084 range, the 500/800 °C release curve, the 26 nm ilmenite implantation model, the ~19 MW·yr/kg figure and the mechanical-agitation alternative. - **Greer, J.; Rout, S. S.; Isheim, D.; Seidman, D. N.; Wieler, R.; Heck, P. R. (2020).** "Atom probe tomography of space-weathered lunar ilmenite grain surfaces." *Meteorit. Planet. Sci.* **55**(2), 426–440. DOI [10.1111/maps.13443](https://doi.org/10.1111/maps.13443). — Atomic-scale implantation profile (maximum 30–50 nm, tail to ~200 nm) and the ~0.5‰ retention shortfall. - **Guo, Z.; et al. (2024).** *Commun. Earth Environ.* **5**, 426. DOI [10.1038/s43247-024-01590-6](https://doi.org/10.1038/s43247-024-01590-6). — Space-weathered rim thickness up to ~120 nm, with nanophase iron and vesicles in the top 50 nm. - **Kou, J.; et al. (2021).** *Minerals* **11**(3), 319. DOI [10.3390/min11030319](https://doi.org/10.3390/min11030319). — Molecular-dynamics and SRIM modelling of ³He release from ilmenite: optimum near 1000 K, cumulative release ~74%. - **De Temmerman, G.; Chuard, D.; Rudelle, J.-B. (2021).** "The helium bubble: prospects for ³He-fuelled nuclear fusion." *Joule* **5**(6), 1312–1315. DOI [10.1016/j.joule.2021.05.001](https://doi.org/10.1016/j.joule.2021.05.001). — The essential sceptical counterweight (~18 kg/y terrestrial supply, ~25 kg US reserve, ~1 Mt lunar reserve ≈ 600 ZJ); **contains the 30 µg/g units error corrected above — cite the conclusions, not that sentence.** - **U.S. Government Accountability Office (2011).** *Managing Critical Isotopes.* **GAO-11-472**, 12 May 2011. [gao.gov/products/gao-11-472](https://www.gao.gov/products/gao-11-472). — US extraction capacity 8,000–10,000 L/y against ~30,000 L/y average sales 2003–2009, and the stewardship finding. - **Congressional Research Service (2010, rev. 2011).** *The Helium-3 Shortage: Supply, Demand, and Options for Congress.* **R41419**. [PDF](https://www.congress.gov/crs_external_products/R/PDF/R41419/R41419.8.pdf). — Stockpile history and the government/commercial price spread. - **U.S. Energy Information Administration.** *U.S. energy facts explained.* [eia.gov](https://www.eia.gov/energyexplained/us-energy-facts/). — The ~94–96 quads of US primary energy consumption against which the 25-tonne claim is audited. - **Bosch, H.-S.; Hale, G. M. (1992).** "Improved formulas for fusion cross-sections and thermal reactivities." *Nucl. Fusion* **32**(4), 611–631. DOI [10.1088/0029-5515/32/4/I07](https://doi.org/10.1088/0029-5515/32/4/I07). — The reactivity parameterisation behind the ~1–20% D–³He neutron-power range quoted above; note that D–³He is still rising at the top of the fit's validity, so no "peak temperature" may be cited from it. *Outside the audited scope — the four non-helium-3 sections:* Colaprete et al. (2010), "Detection of Water in the LCROSS Ejecta Plume," *Science* **330**, 463–468, DOI [10.1126/science.1186986](https://doi.org/10.1126/science.1186986), for the 5.6 ± 2.9 wt% Cabeus figure; Paige et al. (2010), *Science* **330**, 479–482, DOI [10.1126/science.1187726](https://doi.org/10.1126/science.1187726), for permanently-shadowed-region temperatures; Li et al. (2018), *PNAS* **115**(36), 8907–8912, DOI [10.1073/pnas.1802345115](https://doi.org/10.1073/pnas.1802345115), for surface-exposed polar ice; Honniball et al. (2021), *Nature Astronomy* **5**, 121–127, DOI [10.1038/s41550-020-01222-x](https://doi.org/10.1038/s41550-020-01222-x), for 100–412 ppm molecular water at Clavius. Regolith bulk density, particle-size distribution and specific-heat values used by the microsim derive from Carrier's geotechnical compilations and Diviner-based thermophysical fits and are `[UNVERIFIED]` against primaries in this pass. <!-- CRAFT-LINK:START g12 --> *Built to the [[WT!Three_js_Microsim_Master_Class|three.js Master Class]]. Mining context: [[WT!Space_Mining_In_Minnesota]] · energy context: [[WT!Energy_Center_of_Excellence]].* <!-- CRAFT-LINK:END --> ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Lunar_resources) : [Wikitube](https://en.wikitube.io/wiki/Lunar_resources) ## Previous hub tags Hubs: `Life_Physics`. Portals: [[PORTAL_Helium-3]], [[PORTAL_Energy]].