Frontier
← All research
Frontier

Research Brief 02: Getting Water Out of a Carbonaceous Asteroid: Methods, Energy, and What a First Demonstration Needs

Frontier (thefrontier.space) — Research Programme Date: 2026-10-29 Status: Published, v1.0 Scope: How much water is actually locked inside carbonaceous (C-type) asteroid material, the physical methods proposed for getting it out, the energy each one costs, what has been demonstrated in laboratories and prototypes to date, and what a 2–5 kg payload could realistically prove on a first Frontier mission.

1. Executive summary

Asteroid water is the commodity that every commercial asteroid plan, including Frontier's, ultimately rests on. It is also the resource with the most laboratory evidence behind it and the least flight evidence. No spacecraft has ever extracted a gram of water from an asteroid. What we have is returned samples from two C-type bodies, a century of meteorite chemistry, and a handful of vacuum-chamber prototypes.

The key findings of this brief:

  • The water is real, but it is not ice. In carbonaceous material the water is chemically bound inside clay-like phyllosilicate minerals (serpentine, saponite). CI chondrites, the most altered class, hold roughly 20 ± 5 wt% water; CM chondrites average about 7 wt% (range 1.9–10.5 wt%); CR chondrites are below about 9 wt% [1][2][3]. The two asteroids we have samples from are at the low end of this: Ryugu grains gave 6.8 mass% total H₂O by thermal analysis [4], and Bennu samples are "hydrated (0.84–0.95 wt% H)", with roughly 80 vol% phyllosilicates and up to about 2 wt% of loosely held interlayer water [5][6][7].
  • Getting it out means heating rock to 400–770 °C. Thermogravimetric work on CM chondrites shows adsorbed water leaves below 200 °C, iron-oxyhydroxides release OH/H₂O at 200–400 °C, phyllosilicates dehydrate and dehydroxylate at 400–770 °C, and carbonates break down at 770–900 °C [8]. Any process that stops at a few hundred degrees recovers only a fraction of the inventory.
  • Energy is the dominant cost. Our own arithmetic from published constants (section 4) puts the thermodynamic minimum at roughly 9–16 MJ (2.5–4.5 kWh) per kilogram of water from CI- to CM-like regolith, before losses. The one published small-spacecraft design study that costed this used 2 Wh/g (7.2 MJ/kg) for microwave extraction, from a range of 2–17 Wh/g measured in NASA Marshall simulant tests [9]. A bucket-wheel excavation study found heating "the most energy intensive portion of the process" [10].
  • Three method families exist; none has flown. Concentrated sunlight ("optical mining", TransAstra, tested to TRL 5 at Colorado School of Mines) [11][12]; microwave heating (NASA MSFC lunar and carbonaceous-simulant tests) [9]; and conductive heating of excavated or cored material with a cold trap (Honeybee Robotics' WINE, vacuum-chamber hop on 31 December 2018) [13][14]. Every one needs a cold trap to condense the vapour [9].
  • The hard part may not be heat but handling. OSIRIS-REx sank 48.8 cm into Bennu's surface before its thrusters fired [15]; Philae's harpoons never fired and it bounced [16]; Bennu throws centimetre-scale particles off its own surface at about 3 m/s [17]. Anchoring, containing vapour in vacuum, dust, and organic contamination of the extraction hardware are open engineering problems.
  • Recommendation. Frontier's first mission (Brief 01) goes to a dry S-type, so water extraction cannot be tested there. A separate 2–5 kg thermal-extraction package should be developed and tested first in vacuum chambers and on suborbital or ISS flights with carbonaceous simulant (about $95/kg) [18], then flown as a secondary on whichever Frontier or partner mission first reaches a C-type body. The specific test matrix is in section 7.

2. How much water is there, and in what form

2.1 Meteorite classes

Carbonaceous chondrites are the meteorite analogues of C-complex asteroids. Water in them is not free ice; it is bound as hydroxyl (OH) groups and interlayer molecules inside phyllosilicates formed when liquid water altered the parent body billions of years ago [1][2]. The main classes relevant to Frontier:

Class Typical water content Form Source
CI (Ivuna-type) ~20 ± 5 wt% (18–20 wt% commonly quoted; Orgueil up to 15 wt% in older work) Serpentine- and saponite-type phyllosilicates, OH and interlayer H₂O [1][2]
CM (Mighei-type) ~9 wt% H₂O⁺ in reviews; measured falls and finds 1.9–10.5 wt%, average 7.0 wt% Serpentine, tochilinite [1][3]
CR (Renazzo-type) < 9 wt%; matrix-rich fractions of CR2 EET 92159 ~11 wt% Fe-rich hydrated amorphous silicate plus nanometre phyllosilicates [1]

Two cautions from the meteorite literature. First, terrestrial contamination: CM finds absorb atmospheric water, and the Lee et al. (2023) study was explicitly designed to separate indigenous from terrestrial water [3]. Second, "water" measured by heating can include SO₂ from sulphide breakdown and organics decomposing in the same temperature window, though thermogravimetric values correlate well with independent methods [8].

2.2 Returned samples: Ryugu and Bennu

The two sampled C-type asteroids are drier than their meteorite analogues, which matters for any business case.

  • Ryugu (Hayabusa2). Bulk composition is CI-like [19]. A ~1 mg grain gave 6.8 mass% total H₂O and 5.5 mass% CO₂ by thermal analysis, noted as "less H₂O than CI chondrite" [4]. Water held loosely between clay layers is ≤ 0.3 wt% [20]. Brief 01 already cited the 6.8 wt% versus 13–20 wt% CI comparison, with interlayer water apparently lost to space [21].
  • Bennu (OSIRIS-REx). The returned sample is "hydrated (0.84–0.95 wt% H) and carbon-rich (4.5–4.7 wt% C)" [5]. X-ray diffraction finds Mg-rich phyllosilicates at about 82 vol%, matching the 80–90 vol% predicted from orbit, and estimates the clay interlayer spaces hold "up to ~2 wt% water", notably more than Ryugu [6][7]. Pre-return spectroscopy had put the hydrogen in surface phyllosilicates at 0.71 ± 0.16 wt%, consistent with CM-like rather than CI-like hydration [22].

A rough conversion: if all of Bennu's 0.84–0.95 wt% hydrogen were present as H₂O or OH (hydrogen is 11.2% of water by mass), the water-equivalent would be about 7.5–8.5 wt%. That is our arithmetic, not a measured number; some of the hydrogen is in organics, so the recoverable water is lower. The practical planning figure for a Bennu- or Ryugu-like target is therefore 5–8 wt% recoverable water, not 20 wt%.

3. Extraction methods

Every proposed method does the same three things: bring the hydrated mineral to several hundred degrees, let the vapour escape, and condense it somewhere cold. The published review of options (Calla, Fries & Welch 2018, citing Wiens et al.) lists funnels and conveyors, focused light, and microwaves, and states that "for any of these techniques, a cold trap needs to be attached in order to condense the vapor and collect liquid water" [9].

3.1 Optical mining (concentrated sunlight)

TransAstra's Apis architecture captures a small asteroid or boulder in an inflatable bag and uses reflectors to focus sunlight on it; the concentrated light causes thermal spalling, breaking the surface into fragments and releasing water from the hydrated minerals in the heated spalls [11][23]. NASA selected the "Mini Bee" flight demonstration (about 250 kg, testing against a synthetic asteroid made of CI simulant) for a NIAC Phase III award in 2019 [24][25]; we found no record of it having flown.

The most recent published ground results are Broslav, Dreyer & Sercel (2025), Acta Astronautica 235, which tested optical mining at Colorado School of Mines "in a relevant vacuum environment (TRL 5)" with a compositional simulant called Nectar. The paper reports excavation rates and water production rates as functions of beam irradiance, with TGA-MS on the mined material, and flags two problems: mineral morphology can exacerbate spalling, and "organic vapor deposition hampers excavation and water production rates" [11]. The yield figures themselves are behind a paywall and are not quoted here (open question 1).

Advantages: no digging, no moving parts in contact with regolith, sunlight is free. Disadvantages: needs a bag or enclosure to contain spall and vapour, needs the asteroid (or a boulder) to be small enough to capture, and organics vaporised from the rock coat the optics.

3.2 Microwave heating

Microwaves couple energy directly into bound water rather than heating the bulk rock, and penetrate low-conductivity regolith [9][26]. NASA Marshall (Ethridge and Kaukler) extracted water from cryogenic lunar-permafrost simulant and then "repeated [the experiment] with a simulated carbonaceous chondrite asteroid. After exposure to microwaves, thermal decomposition occurs and released water vapour was captured in a cold trap. This recovered all water from the sample plus an additional 3 ml of water which had been chemically bound in the clay materials" [9]. Measured efficiencies in Mars-simulant tests were 2, 7 and 17 Wh per gram of water depending on conditions [9]. More recent lunar work at 2.45 GHz and 400–800 W found uniform heating of icy simulant at −80 °C [26].

The Calla et al. small-spacecraft design assumed 2 Wh/g, seven 200 W solid-state microwave probes (1,400 W total), and 15 net operating days to extract about 250 kg of water [9]. Note that its spacecraft came out at 372.8 kg dry and $113.6 million per unit, and the authors concluded that over one hundred such craft would be needed for profitability [9]: this is a design study, not a plan.

3.3 Mechanical excavation plus conductive heating, with a cold trap

The alternative to heating the asteroid in place is to dig or core material and heat it inside the spacecraft. Honeybee Robotics' World Is Not Enough (WINE) prototype, funded through NASA SBIR with UCF and Embry-Riddle, demonstrated the full chain in a vacuum chamber on 31 December 2018: mine and heat regolith, capture water as ice in a cold trap, move it to a tank, heat it to steam, and hop [13][14]. The simulant was hydrated artificial asteroid regolith analogous to Ceres/C-type material and was physically dry; the water came out of the phyllosilicate crystal structure, and it "contained large amounts of dissolved carbon dioxide, metals, and organic matter" [14][27]. Honeybee's PVEx corer heats the core through the double wall of its auger [14]. Honeybee is now part of Blue Origin [28].

NASA's earlier Robotic Asteroid Prospector (NIAC) study tested auger extraction and thermal-vacuum distillation on frozen simulant slurry, reaching TRL 4; a pressure difference of about 500 Pa was enough to drive vapour to the trap [29].

A 2023 modelling study of thermal extraction from regolith makes the containment point explicitly: water vapour "moves away from a heating device in the soil", and if the geometry fails to contain it "the vapor may diffuse away through the regolith and be re-frozen somewhere else" [30].

3.4 Cold-trap capture

All three families end at a cold trap. In the NASA permafrost experiments "virtually all of the water ice sublimed and was efficiently captured in the cold trap (approximately 99% efficient, based on weight measurements)" [31]. In the Calla design the cold trap was budgeted at 1 kg, 140 mm diameter by 160 mm tall [9]. The trap must be colder than the vapour's condensation point at the local pressure and shadowed from the Sun; on a spinning spacecraft or asteroid that is an attitude-control constraint, not just a thermal one.

3.5 Comparison

Method Where the heat goes Needs Best ground result Main risk
Optical mining Surface of a bagged rock Reflectors, inflatable bag, cold trap TRL 5 vacuum tests on Nectar simulant [11] Organic vapour fouling optics; needs capturable body
Microwave Directly into bound water Power (hundreds of W to kW), probes in contact or borehole, cold trap All water plus 3 ml bound water from carbonaceous simulant [9] Power budget; coupling varies with mineralogy
Excavate + conductive heating Inside spacecraft oven/auger Excavator or corer, anchoring, oven, cold trap Full chain demonstrated (WINE, 2018) [13] Anchoring and regolith handling in microgravity

4. Energy per kilogram of water

No published source gives a clean "MJ per kg of asteroid water" figure, so we derive one from cited constants and show the working.

Inputs:

  • Specific heat of chondritic rock: a standard model used in planetesimal thermal studies is c_p = 800 + 0.25 T − 1.5×10⁷ T⁻² J kg⁻¹ K⁻¹ (Neumann et al. 2012), which is roughly 0.9 kJ kg⁻¹ K⁻¹ over 500–1,000 K [32].
  • Phyllosilicate dehydration is endothermic with a latent heat of about 417 kJ per kg of hydrated silicate (Wakita & Sekiya 2011), with the reaction near 773 K [33].
  • Latent heat of vaporisation of water: about 2,257 kJ/kg [34].
  • Temperature window for full dehydroxylation: 400–770 °C [8]; start from ~250 K regolith, so ΔT ≈ 800 K to reach 770 °C.

Arithmetic (ours):

  1. Sensible heat to raise 1 kg of regolith by 800 K: ≈ 0.9 × 800 ≈ 720 kJ.
  2. Dehydration enthalpy: serpentine is about 13 wt% water, so 417 kJ per kg of silicate ≈ 3.2 MJ per kg of water released.
  3. Vaporisation: 2.26 MJ per kg of water.
Regolith water content Sensible heat per kg water + dehydration + vaporisation Total (thermodynamic minimum)
20 wt% (CI-like) 720 / 0.20 = 3.6 MJ 3.2 MJ 2.3 MJ ≈ 9 MJ/kg (2.5 kWh/kg)
7 wt% (CM / Ryugu / Bennu-like) 720 / 0.07 = 10.3 MJ 3.2 MJ 2.3 MJ ≈ 16 MJ/kg (4.4 kWh/kg)

Real systems will be worse: heat lost to space, incomplete reaction, trap inefficiency, and the fact that you cannot heat only the hydrated grains. The NASA Marshall microwave range of 2–17 Wh/g is 7–61 MJ/kg [9], bracketing our minimum from above, which is consistent.

What that means for power. The IAU nominal solar irradiance at 1 au is 1,361 W/m² [35]. A 1 m² concentrator at 50% end-to-end efficiency delivers ≈ 680 W, so producing 1 kg of water from CM-like material at 16 MJ/kg takes about 6.5 hours of continuous sunlight (our arithmetic). The Calla design's 1,400 W of microwave power for 250 kg in 15 net days is the same order of magnitude [9]. A tonne of water per year is therefore a 1–2 kW continuous power problem, which is a realistic size for a 200 kg-class spacecraft (AstroForge's DeepSpace-2 is specified at 1.7 kW [36]), but not for a cubesat.

5. Microgravity problems: anchoring, regolith handling, contamination, spin

5.1 Anchoring and regolith behaviour

The flight record is short and sobering:

  • OSIRIS-REx at Bennu (20 October 2020) sank 48.8 cm into the surface before its back-away burn and left an 8 m crater; models show regolith with cohesion of 25 Pa "provides little resistance" to the sampler while 125 Pa limits penetration [15][37]. The surface behaved like a fluid.
  • Philae at comet 67P (2014): the harpoon anchors never fired, the lander bounced, and only part of its planned science was completed [16].
  • Hayabusa2 at Ryugu: the surface was so bouldery that the team "cannot find a 100 percent safe place to touch down" and had to shrink the landing ellipse dramatically [38].

Anchoring research since then favours drilling-based or multi-point force-closure anchors over single-shot harpoons. An ultrasonic cross-drilling anchor achieved 60–250 N holding force in media from tens of kPa to 60 MPa, but only on a bench [39]. NASA's microspine drill/anchor was the first self-contained system shown to drill inverted and could in principle work in microgravity [40]. Parabolic flights remain the main way to test regolith shear and impact behaviour at asteroid-level gravity [41][42].

For an excavation approach, every newton of digging force must be reacted by anchors; for optical mining, the bag replaces the anchor; for microwave probes, the probe must be pushed into or held against the surface. Frontier should regard "can we stay attached and push on this surface?" as the first thing to demonstrate.

5.2 Contamination: in both directions

  • Rock contaminating the hardware. The optical-mining tests found that organic vapour deposition from the heated simulant "hampers excavation and water production rates" [11]. WINE's extracted water carried dissolved CO₂, metals and organics [14]. Any water meant for electrolysis or life support needs purification; water for steam propulsion may not.
  • Hardware contaminating the science. Hayabusa2 and OSIRIS-REx invested heavily in contamination control for sample science [43]. A Frontier mission that both prospects and extracts must decide which comes first, because extraction hardware outgassing will corrupt volatile measurements made afterwards.
  • Dust. Bennu ejects centimetre-scale particles at ~3 m/s in repeated events, some of which temporarily orbit the asteroid [17]. Ejecta from any excavation activity "could degrade or damage instrumentation and reduce visibility and communications" [44]. Optics, radiators and the cold trap are the vulnerable surfaces.

5.3 Spin

Brief 01 showed that most sub-100 m NEAs rotate in minutes, and that on 2022 OB5 surface centrifugal acceleration is nearly 100 times gravity [45]. On such bodies nothing can rest on the surface; extraction would have to be by capture (bagging the whole object, the TransAstra approach) or not at all. On the slow rotators (Bennu 4.3 h, Ryugu 7.6 h, Apophis 30.6 h [45]) surface operations are possible, but a sun-tracking concentrator or a sun-shaded cold trap on the surface will see the Sun sweep across the sky every rotation, so the extraction duty cycle is at most about half. Calla et al. assumed 50% occultation for exactly this reason [9].

6. Who is working on this

Organisation Approach Status (as found) Source
TransAstra (USA) Optical mining, Apis / Mini Bee TRL 5 ground tests published 2025; NIAC Phase III flight demo selected 2019, no flight found [11][24]
Colorado School of Mines, Space Resources Program Vacuum test facility for optical mining (Dreyer co-author) Active [11]
Honeybee Robotics (Blue Origin) WINE: corer + oven + cold trap + steam hop Vacuum-chamber demo Dec 2018; NASA SBIR [13][14][28]
University of Central Florida (Exolith/CLASS; Metzger) Simulants, WINE partner, solar-thermal oven at kilogram scale Exolith simulant production transferred to Space Resource Technologies [14][18][25]
NASA Marshall (Ethridge, Kaukler) Microwave extraction with cold trap, incl. carbonaceous simulant Lab proof-of-concept [9]
Karman+ (USA) Excavate regolith, extract water for orbital refuelling; first mission to dry Apophis will "deploy our tests on the surface" $20 million seed Feb 2025; mission 1 launch mid-2028; heating method not published [46][47][48]
AstroForge (USA) Metals, not water; relevant for bus (200 kg, 1.7 kW, EP) and lessons DeepSpace-2 on IM-3 [36]
Space Resource Technologies (USA) CI (Orgueil-type) simulant, $95/kg Commercial [18]

Karman+'s own site describes the goal (water for refuelling, or split into H₂ and O₂) and the plan to "excavate regolith", but we could not find any published description of its heating method [46][47]. Its mining spacecraft mass is also unpublished (unknown).

7. What a 2–5 kg demonstration can realistically test

A first Frontier spacecraft to Apophis (Brief 01) will find no water to extract. The purpose of a small extraction package on that mission, or on a partner's C-type mission, is therefore to retire handling risks, not to make water. We propose a staged programme.

Stage A: Ground (vacuum chamber), 2027. Oven-and-cold-trap unit sized for 100–500 g of CI simulant ($95/kg [18]) and, if obtainable, CM simulant. Measure water yield versus peak temperature at 300, 500 and 770 °C against the Garenne thermogravimetric steps [8], energy per gram actually consumed, trap efficiency, and the composition of condensate (CO₂, organics, metals, following WINE [14]). Target: confirm the 9–16 MJ/kg floor of section 4 and establish a real-system multiplier.

Stage B: Microgravity (parabolic flight, New Shepard, or ISS), 2027–2028. Blue Origin's New Shepard single and double lockers carry 11.34 and 22.68 kg [49]; parabolic flights are the established way to test regolith behaviour at reduced gravity [41][42]. Test only the mechanical steps: feeding grains into a heated cell without gravity, vapour transport to the trap, and ice capture. Prices are quote-only (unknown).

Stage C: Flight package, 2–5 kg, on a C-type mission, 2029+. For scale, Hayabusa2's MINERVA-II rovers were 1.1 kg each and MASCOT 9.8 kg [50][51], so a 2–5 kg unit is a credible secondary payload. Candidate contents:

  1. A 0.5–1 kg heated sample cell with a scoop or probe that accepts a few grams of surface material, and a Peltier- or radiator-cooled trap with a mass sensor (quartz crystal microbalance) to measure captured water to milligram precision.
  2. A contact force/penetration sensor to measure regolith cohesion and bearing strength at the point of contact, directly addressing the OSIRIS-REx and Philae experience [15][16].
  3. A miniature near-infrared sensor on the heated cell's exhaust to confirm H₂O (2.7–3 µm) is what came off.

What it can prove: that bound water can be released and captured in space from a real asteroid, the per-gram energy, and whether the surface can be pushed on. What it cannot prove: throughput, anchoring at engineering loads, or long-duration storage of water in deep space, which one profile correctly calls "unproven at any scale" [48].

Cost. Not published for any comparable unit. Order-of-magnitude anchors: the Calla design budgeted a 1 kg cold trap and seven 200 W microwave units [9]; Karman+ says it keeps missions under budget by qualifying "off-the-shelf components from other industries" [52]; AstroForge built a 120 kg spacecraft for $3.5 million [53]. A 2–5 kg package built the same way plausibly costs a low single-digit number of millions including test campaigns. We flag this as an estimate, not a quote.

Partner options. Honeybee/Blue Origin (WINE lineage) [13][28]; Colorado School of Mines (vacuum facility) [11]; UCF / Space Resource Technologies (simulant, Exolith heritage) [18][25]; NASA NIAC or SBIR as funding routes used by both TransAstra and Honeybee [13][24]; and ride-along on Karman+'s or AstroForge's buses, both of which already plan surface tests [36][47].

8. Open questions

  1. What were the Nectar yields? The Broslav, Dreyer & Sercel (2025) water-production rates versus irradiance are paywalled [11]. Frontier should obtain the paper before designing a concentrator.
  2. How much of Bennu's hydrogen is recoverable water? The sample gives 0.84–0.95 wt% H [5] and up to ~2 wt% interlayer water [6], but no bulk thermogravimetric H₂O figure comparable to Ryugu's 6.8 mass% [4] has been found by us. Unknown.
  3. Does microwave coupling work on real asteroid mineralogy? The MSFC carbonaceous test is reported second-hand [9]; the primary report is a scanned NTRS document we could not machine-read.
  4. What is Karman+'s extraction method? Not published [46][47].
  5. Can a sub-5 kg trap stay cold on a slow rotator's surface? Thermal design against a 4–30 h day-night cycle has not been published for any small package.
  6. What do the organics do to a reusable oven? Vapour deposition degraded optical mining [11]; the equivalent fouling of microwave probes and conductive ovens is unquantified.
  7. Is there any C-type target Frontier can reach before 2032? Brief 01 found none below 7 km/s round trip [45]; the extraction package may need to fly on someone else's mission first.

9. References

  1. Garenne, A. et al. (2014), "The abundance and stability of 'water' in type 1 and 2 carbonaceous chondrites (CI, CM and CR)," Geochimica et Cosmochimica Acta 137. https://www.sciencedirect.com/science/article/abs/pii/S0016703714002130
  2. Palme, H. et al. (2022), "The composition of CI chondrites and their contents of chlorine and bromine," Meteoritics & Planetary Science (CI water ~20 ± 5 wt%). https://onlinelibrary.wiley.com/doi/10.1111/maps.13720 ; see also Wikipedia, "CI chondrite" (18–20 wt%). https://en.wikipedia.org/wiki/CI_chondrite
  3. Lee, M. R. et al. (2023), "The water content of CM carbonaceous chondrite falls and finds, and their susceptibility to terrestrial contamination," Meteoritics & Planetary Science (1.9–10.5 wt%, average 7.0 wt%). https://onlinelibrary.wiley.com/doi/10.1111/maps.14099
  4. Homma, Y. & Motomura, K., "Bulk chemical composition of samples recovered from asteroid Ryugu," Rigaku Journal (6.8 mass% H₂O, 5.5 mass% CO₂). https://rigaku.com/resources/rigaku-journal/chemical-composition-asteroid-ryugu-samples
  5. Lauretta, D. S. et al. (2024), "Asteroid (101955) Bennu in the laboratory: Properties of the sample collected by OSIRIS-REx," Meteoritics & Planetary Science (0.84–0.95 wt% H, 4.5–4.7 wt% C). https://onlinelibrary.wiley.com/doi/10.1111/maps.14227
  6. King, A. J. et al. (2026), "Low temperatures and high water/rock ratios in asteroid (101955) Bennu's history based on X-ray powder diffraction of returned samples," Meteoritics & Planetary Science (~82 vol% phyllosilicates; up to ~2 wt% interlayer water). https://onlinelibrary.wiley.com/doi/10.1111/maps.70169
  7. Tasker, E. J., Connolly, H. C. & Tachibana, S. (2026), "The science from asteroid sample return missions," Contemporary Physics / arXiv:2604.22182. https://arxiv.org/abs/2604.22182
  8. King, A. J. et al. (2021), "Thermal alteration of CM carbonaceous chondrites: mineralogical changes and metamorphic temperatures," arXiv:2102.07634 (summarising Garenne et al. 2014 TGA temperature steps). https://arxiv.org/pdf/2102.07634
  9. Calla, P., Fries, D. & Welch, C. (2018), "Asteroid mining with small spacecraft and its economic feasibility," arXiv:1808.05099 (microwave 2–17 Wh/g; MSFC carbonaceous simulant test; cold trap; 1,400 W; $113.6 million/unit). https://arxiv.org/abs/1808.05099
  10. "Optimized Bucket Wheel Design for Asteroid Excavation," arXiv:1701.07547 (heating most energy-intensive step). https://arxiv.org/pdf/1701.07547
  11. Broslav, M., Dreyer, C. & Sercel, J. (2025), "Optical mining of carbonaceous chondrite simulants: Testing and lessons learned in asteroid mining research," Acta Astronautica 235. https://www.sciencedirect.com/science/article/pii/S0094576525002334
  12. AZoMining, "Laser-Driven Optical Mining Study Reveals How Beam Control and Mineral Makeup Unlock Asteroid Water Extraction" (Nectar A/B simulants). https://www.azomining.com/News.aspx?newsID=18470
  13. Honeybee Robotics, "The World Is Not Enough Demonstrates The Future of Space Exploration." https://www.honeybeerobotics.com/news-events/the-world-is-not-enough-demonstrates-the-future-of-space-exploration/
  14. Wikipedia, "World Is Not Enough (spacecraft propulsion)" (31 Dec 2018 test; cold trap; dissolved CO₂, metals, organics; NASA SBIR; UCF and Embry-Riddle). https://en.wikipedia.org/wiki/World_Is_Not_Enough_(spacecraft_propulsion)
  15. Wikipedia, "OSIRIS-REx" (sank 48.8 cm; 8 m crater; 121.6 g sample). https://en.wikipedia.org/wiki/OSIRIS-REx
  16. "Analysis of a Mechanical Anchoring System for Landing on Asteroids" (Philae harpoon failure, Ulamec et al.). https://www.researchgate.net/publication/372324308_Analysis_of_a_Mechanical_Anchoring_System_for_Landing_on_Asteroids
  17. Lauretta, D. S. et al. (2019), "Episodes of particle ejection from the surface of the active asteroid (101955) Bennu," Science. https://iac.es/en/science-and-technology/publications/episodes-particle-ejection-surface-active-asteroid-101955-bennu ; NASA OSIRIS-REx, "Explaining Bennu's Mysterious Particle Events." https://www.asteroidmission.org/?p=21129
  18. Space Resource Technologies, "CI Carbonaceous Chondrite Simulant" ($95.00/kg). https://spaceresourcetech.com/products/ci-carbonaceous-chondrite-simulant
  19. Yokoyama, T. et al. (2022), "Samples returned from the asteroid Ryugu are similar to Ivuna-type carbonaceous meteorites," Science 379. https://eprints.lib.hokudai.ac.jp/dspace/handle/2115/90313
  20. Astronomy & Astrophysics (2025), study of 186 Ryugu grains (interlayer water ≤ 0.3 wt%). https://www.aanda.org/articles/aa/full_html/2025/03/aa53270-24/aa53270-24.html
  21. Carbonne, D. et al. (2023), Hayabusa2 Sample Return Symposium abstract S23-02. https://curation.isas.jaxa.jp/symposium/abstract/2023/S23-02_Carbonne.pdf
  22. Praet, A. et al. (2021), "Hydrogen abundance estimation and distribution on (101955) Bennu," Icarus (0.71 ± 0.16 wt% H). https://www.sciencedirect.com/science/article/abs/pii/S0019103521001111
  23. Space.com (2015), "Asteroid-Mining Plan Would Bake Water Out of Bagged-Up Space Rocks." https://www.space.com/30582-asteroid-mining-water-propulsion.html
  24. NASA, "NASA Invests in Tech Concepts Aimed at Exploring Lunar Craters, Mining Asteroids" (Mini Bee NIAC Phase III, 2019). https://www.nasa.gov/news-release/nasa-invests-in-tech-concepts-aimed-at-exploring-lunar-craters-mining-asteroids/
  25. Space.com, "No Digging Required: Space Mining on the Moon and Beyond May Be Solar Powered" (Mini Bee ~250 kg; CI simulant by Exolith; UCF 2–5 cm samples and kilogram-scale solar-thermal oven). https://www.space.com/moon-asteroid-space-mining-with-concentrated-sunlight.html
  26. "Massive Water Production from Cryogenic Icy Lunar Regolith by a Microwave Heating Method" (2.45 GHz, 400/800 W, −80 °C). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12308065/
  27. The Space Resource, "Honeybee Robotics Demonstrates WINE" (Ceres/C-type simulant). https://www.thespaceresource.com/news/2019/1/honeybee-robotics-demonstrates-wine
  28. Factories in Space, "Honeybee Robotics (Blue Origin)." https://www.factoriesinspace.com/honeybee-robotics
  29. NASA NIAC, "Robotic Asteroid Prospector (RAP)" final report (thermal-vacuum distillation TRL 4; ~500 Pa). https://ntrs.nasa.gov/api/citations/20190001162/downloads/20190001162.pdf
  30. "Thermal Extraction of Volatiles from Lunar and Asteroid Regolith in Axisymmetric Crank–Nicolson Modeling," arXiv:2306.03776 / J. Aerospace Engineering. https://arxiv.org/pdf/2306.03776
  31. NASA, "Extraction and Capture of Water from Martian Regolith" (cold trap ~99% efficient). https://ntrs.nasa.gov/api/citations/20160010258/downloads/20160010258.pdf
  32. "Modeling carbon outgassing from chondritic planetesimals," arXiv:2605.29914 (c_p model after Neumann et al. 2012). https://arxiv.org/pdf/2605.29914
  33. "Growth and thermal evolution of icy planetesimals," arXiv:2512.07978 (dehydration latent heat 417 kJ/kg after Wakita & Sekiya 2011; T ≈ 773 K). https://arxiv.org/pdf/2512.07978
  34. Nagwa, "Latent heat of vaporization of water" (≈ 2,257 kJ/kg at 100 °C, 1 atm). https://www.nagwa.com/en/videos/975156972395
  35. IAU 2015 Resolution B3, "Nominal values for selected solar and planetary quantities" (S_N = 1,361 W/m²). https://arxiv.org/pdf/1605.09788
  36. AstroForge, "Odin't: A Complete Debrief of Our Deep Space Mission," 6 Mar 2025 (Odin 120 kg, 180–190 W; DeepSpace-2 200 kg, 1.7 kW, electric propulsion). https://www.astroforge.com/updates-collection/odint-mission-debrief
  37. "Modified granular impact force laws for the OSIRIS-REx touchdown on the surface of asteroid (101955) Bennu," MNRAS 507 (2021). https://academic.oup.com/mnras/article/507/4/5087/6356586
  38. Science News, "Two daring spacecraft aim to bring asteroid dust back to Earth." https://www.sciencenews.org/article/asteroid-sample-missions-hayabusa2-ryugu-osiris-rex-bennu
  39. "An asteroid anchoring method based on cross-drilling geometric force closure of ultrasonic drill" (60–250 N). https://www.researchgate.net/publication/347468123_An_asteroid_anchoring_method_based_on_cross-drilling_geometric_force_closure_of_ultrasonic_drill
  40. NASA NTRS, "Microgravity Drill and Anchor System." https://ntrs.nasa.gov/search.jsp?R=20130014134
  41. "Simulating regoliths in microgravity," MNRAS 433 (2013). https://academic.oup.com/mnras/article/433/1/506/1035055
  42. "Regolith behavior under asteroid-level gravity conditions: low-velocity impact experiments," arXiv:1810.01459. https://arxiv.org/pdf/1810.01459
  43. JAXA ISAS, "Keeping out the Earth: Protecting the asteroid Ryugu sample from contamination." https://cosmos.isas.jaxa.jp/keeping-out-the-earth-protecting-the-asteroid-ryugu-sample-from-contamination/
  44. "Ejecta orbital and bouncing dynamics around asteroid Ryugu," IAC-19. https://iafastro.directory/iac/archive/browse/IAC-19/C1/3/53757/
  45. Frontier Research Brief 01, "Near-Earth Asteroid Target Shortlist for a First Prospecting Mission" (2026-10-28), sections 2.3 and 3. https://thefrontier.space/research
  46. Karman+, home page and news (water for refuelling or split into H₂/O₂; Prospector software; Q3 2025 build update). https://www.karmanplus.com/
  47. Karman+, "K+ Mission 1 -> 99942 Apophis," 21 Jul 2026. https://www.karmanplus.com/k-mission-1-99942-apophis/
  48. Space Bandits, "Karman+: Asteroid water for orbital refueling" (nothing built or flown; storage and transfer unproven at any scale). https://spacebandits.io/startups/karman-plus
  49. Blue Origin, "Payloads on New Shepard" (Single Locker 11.34 kg, Double Locker 22.68 kg, Full Stack 68 kg). https://www.blueorigin.com/payloads/
  50. "Operation results of MINERVA-II twin rovers onboard Hayabusa2 asteroid explorer," IAC-19 (≈1.1 kg each). https://iafastro.directory/iac/archive/browse/IAC-19/A3/4A/52276/
  51. "The MASCOT lander aboard Hayabusa2: The in-situ exploration of NEA (162173) Ryugu," Planetary and Space Science (9.8 kg; 16 h). https://www.sciencedirect.com/science/article/abs/pii/S0032063321000398
  52. Factories in Space, "Karman+" (off-the-shelf components; budget under $10 million). https://factoriesinspace.com/karman-plus
  53. The Space Review, "The new wave of asteroid mining ventures," 17 Mar 2025 (Odin $3.5 million). https://www.thespacereview.com/article/4955/1

Additional context consulted: Zheng et al. (2024), "Kinetic analysis of dehydration/dehydroxylation from carbonaceous chondrites by in situ heating experiments under an infrared microscope," Meteoritics & Planetary Science (https://onlinelibrary.wiley.com/doi/full/10.1111/maps.14259); NASA MSFC, Ethridge & Kaukler, "Microwave Extraction of Water from Lunar Regolith Simulant" (https://ntrs.nasa.gov/api/citations/20070018802/downloads/20070018802.pdf, scanned, not machine-readable by us); University of Arizona, "Asteroid Bennu Comes From a Long-lost Salty World" (https://lpl.arizona.edu/news/spotlight/asteroid-bennu-comes-long-lost-salty-world-ingredients-life); Wikipedia, "Joel C. Sercel" (https://en.wikipedia.org/wiki/Joel_C._Sercel).

Frontier research briefs are written to be checked. Every number above links to its source; if you find one that is wrong or out of date, tell us at https://thefrontier.space/research.