# Aneutronic fusion Aneutronic fusion is [[Nuclear_fusion|nuclear fusion]] in which the released [[Energy|energy]] emerges almost entirely as charged particles rather than as [[Neutron|neutrons]], and the whole case for it rests on one fact about electric charge: a charged product feels a magnetic [[Force|force]], so it can be confined, steered and decelerated, while a neutron feels nothing and simply leaves. Everything the field promises to [[Power_engineering|power engineering]] descends from that asymmetry — and so does everything it is accused of overselling. ## Microsims — three.js <iframe src="https://wikitube-3d-microsims.netlify.app/Aneutronic_fusion.html" width="100%" height="620" frameborder="0" loading="lazy" sandbox="allow-scripts allow-same-origin" title="Aneutronic fusion — three.js microsim"></iframe> **`Aneutronic_fusion` (three.js).** The sim is a magnetised cylinder carrying real SI units internally — a 0.5 m first-wall radius, 2 m between the end plates — in which every product is launched with the energy the code derives from a nuclear mass table rather than a typed-in constant and then pushed by an exact Boris rotation, so the charged products spiral on their true gyroradii and stream to the direct-conversion rings while the neutrons, feeling no field whatever, fly dead straight through the wall and leave a strike mark. Drag the field slider and watch the gyroradius fall as 1/B: at 1 T a 14.66 MeV [[Proton|proton]] needs 55 cm of room inside a 50 cm vessel and arrives at the wall as heat, while at 12 T it is 4.6 cm and every proton reaches the converter. The live HUD quantity is the neutron power fraction, accumulated from the events the sim actually fires rather than asserted — 0.80 for D–T, exactly 0.0000 for the D–³He primary channel alone, and 0.06–0.08 the moment you switch the D–D side reactions on. That number moving off zero is the honest heart of the sim; p–¹¹B is deliberately omitted because no verifiable reactivity curve was available to drive it. ## What the label means, and why anyone wants it By the usual convention, a reaction counts as aneutronic when less than 1% of the released [[Energy|energy]] is carried by [[Neutron|neutrons]]. The threshold is agreed rather than found in [[Physics|physics]], but it separates two entirely different machines. Charged products — a [[Proton|proton]], an [[Alpha_particle|alpha particle]], a helion — are [[Ion|ions]] in a [[Plasma_(physics)|plasma]] and obey the Lorentz [[Force|force]]. They gyrate about the field of a [[Superconducting_magnet|superconducting magnet]] on a radius √(2mE)/(qB), stay inside the vessel, heat the fuel by collision, and can in principle be decelerated against an electrostatic grid to yield [[Voltage|voltage]] and [[Electric_current|current]] directly. Nothing in that chain needs a boiler, and nothing in it needs [[Heat_transfer|heat transfer]] at all. Neutrons do none of it. Having no charge they cannot be steered by any practical field, so they leave in a straight line and deposit 14 MeV wherever they strike. In the first wall that means displacement damage and transmutation: [[Materials_science|structural materials]] such as [[Steel|steel]], [[Tungsten|tungsten]] and reduced-activation [[Alloy|alloys]] are activated into radionuclides with their own [[Half-life|half-lives]], embrittled, and exposed to [[Fatigue_(material)|fatigue]] and [[Fracture_mechanics|fracture]] mechanisms with no fission analogue at this spectrum, losing [[Strength_of_materials|strength]] over the life of the plant. The [[Energy|energy]] is not lost, but recovering it means thermalising the neutrons in a [[Lithium|lithium]]-bearing blanket, raising steam, and accepting the [[Second_law_of_thermodynamics|Carnot]] ceiling of a [[Thermal_engineering|thermal cycle]] before a watt reaches [[Electric_power_transmission|transmission]]. A neutron budget is therefore at once a [[Materials_science|materials]] problem, a [[Reliability_engineering|reliability]] problem, a [[Safety_engineering|safety]] problem and an efficiency problem — which is why [[Nuclear_engineering|nuclear engineering]] cares about the label at all. ## The reactions, with the product energies done properly The reactions any D–³He design must account for, with Q values and correct two-body product energies at zero centre-of-mass [[Energy|energy]]: | Reaction | Q (MeV) | Products | |---|---|---| | D + T → n + ⁴He | 17.589 | **n 14.03**, α **3.56** | | D + ³He → p + ⁴He | 18.353 | **p 14.64**, α **3.71** | | D + D → T + p (≈50%) | 4.033 | T 1.01, p 3.02 | | D + D → ³He + n (≈50%) | 3.269 | ³He 0.82, **n 2.45** | The split, not the Q value, is what a [[Nuclear_engineering|reactor designer]] budgets — the neutron number sizes the blanket and the [[Safety_engineering|shielding]], the charged number sizes the converter — and the figures in wide circulation are wrong. "D–T → 14.07 MeV n + 3.52 MeV α" and "D–³He → 14.68 MeV p + 3.67 MeV α" are not measurements; they are the naive mass-*number* split, 4/5 and 1/5 of Q, assuming a mass-3 and a mass-1 body divide the [[Energy|energy]] in inverse proportion to their mass numbers. Multiply 17.589 by 4/5 and 14.07 falls out exactly; multiply 18.353 by 4/5 and so does 14.68. The table values are recomputed relativistically from CODATA-2022 masses. They differ because [[Atomic_mass|nuclear masses]] are not integers — each product's [[Binding_energy|binding energy]] appears as a mass defect — and because a 14 MeV [[Neutron|neutron]] carries kinetic energy that is not negligible against its rest mass. An intermediate convention, the non-relativistic momentum split E₁ = Q·m₂/(m₁+m₂), gives D–T as 14.05 / 3.54: what the microsim above computes from its own [[Atomic_mass|mass table]], and what many teaching texts print. ITER's habitual "14.1 / 3.5" is a defensible rounding of the right number. "14.07 / 3.52" is a precise statement of the wrong one, and the [[Accuracy_and_precision|precision is what makes it misleading]]. One caveat runs the other way. In a burning [[Plasma_(physics)|plasma]] the reactants are not at rest: [[Kinetic_theory_of_gases|thermal motion]] at tens of keV [[Doppler_effect|Doppler]]-broadens these lines, so what [[Neutron_detection|neutron diagnostics]] record is a distribution, not a line. Quoting product energies to 0.01 MeV is over-precise for a reactor, however correct as two-body kinematics. ## The catch: a D/³He plasma is also a D/D plasma This is the heart of the matter, and where most writing on the subject stops being honest. D–³He is aneutronic **in its primary channel only**. The [[Nuclear_fuel|fuel]] contains deuterium, and deuterium fuses with itself whether or not the designer wants it to. Both D–D branches run at once — D + D → T + p at 4.033 MeV and D + D → ³He + n at 3.269 MeV, near-equal [[Probability|branching ratios]] — and the second emits a 2.45 MeV [[Neutron|neutron]] outright. Worse, the first breeds tritium — [[Radioactive_decay|radioactive]] [[Hydrogen|hydrogen]], 12.32-year [[Half-life|half-life]] — inside the [[Plasma_(physics)|plasma]], and it burns with the surrounding deuterium to give the full 14.03 MeV D–T neutron. A D/³He machine thus contains a D–T reactor running quietly in its own exhaust. Whether that matters turns on whether the bred tritium is allowed to burn or is pumped out first — a [[Process_design|process-design]] and [[Systems_engineering|systems]] choice, not a law of [[Physics|physics]]. The published range is roughly **1% to 20% of fusion power as neutrons**, and the authors disagree not through carelessness but because each assumes something different: | Source | Figure | What is assumed | |---|---|---| | UW Fusion Technology Institute (UWFDM-935) | "on the order of 1%" | D–D side reaction only; implicitly no tritium burn | | Razin et al. (PPPL) | f_P = 0.07 at 10% D; "as low as 5%" with tritium suppressed | Explicit fuel ratio and tritium handling | | Razin et al. (same paper) | f_P = 0.33 | The neutron fraction **of the D–D reactions themselves** | | Parisi, Diallo & Meschini | "roughly half" the power from D–D and secondaries | Low-temperature thermal plasma | | De Temmerman et al. | no figure; notes 15% D / 85% ³He mixes proposed to suppress neutrons | — | Recomputing the [[Kinetic_theory_of_gases|Maxwellian]] channel rates from the Bosch–Hale parameterisation — primary reactions plus an optional in-situ tritium burn — reproduces the whole span from a single [[Mathematical_model|model]]: | T_i (keV) | 50:50 D:³He, tritium burns | 50:50 D:³He, tritium removed | 3:1 ³He:D, burns | 3:1 ³He:D, removed | |---|---|---|---|---| | 20 | **21.3%** | 4.4% | — | — | | 60 | **6.3%** | 1.1% | ~2.2% | ~0.39% | | 100 | **5.5%** | 1.0% | ~2.0% | ~0.34% | Three things follow. First, **every quoted neutron fraction is meaningless without its fuel ratio, its temperature and its tritium-handling assumption** — those are not corrections to a headline number, they *are* the number. Second, the Wisconsin "about 1%" and the PPPL "as low as 5%" are both defensible under their own premises and are not in conflict: Wisconsin counts the D–D branch alone, PPPL counts a plasma in which bred tritium contributes. Third, the widely repeated 0.33 is a misreading. [[Expected_value|Averaged]] over its two branches D–D releases 3.651 MeV of which 1.22 MeV is neutron, giving 0.335 — a property of the **D–D reaction**, not of a D/³He plasma, where D–D is a minority channel. At 70 keV in a stoichiometric mix, D–³He is about 82% of all events; the neutron channels are a handful of reactions carrying a disproportionate share of the escaping [[Energy|energy]]. The fraction is a ratio of channel rates, so [[Density|density]] cancels out of it; the fuel ratio and the temperature do not. D–³He is therefore *low*-neutron, not *no*-neutron: a few per cent of fusion power at reactor-relevant temperature, one per cent only if bred tritium is extracted promptly, above 20% if the [[Plasma_(physics)|plasma]] runs cold. No one of these figures survives being quoted alone. ## Why it is harder than D–T even before the neutrons The barrier is electrostatic. Fusion needs two nuclei to tunnel a [[Coulomb's_law|Coulomb]] barrier whose height scales with the product of the charges: for D–T that product is Z₁Z₂ = 1×1, for D–³He it is 1×2. In the Bosch–Hale formulation this appears as the Gamow constant B_G, 34.38 keV^½ for D–T against 68.75 keV^½ for D–³He — a factor of two in B_G, four in Gamow energy, and an exponential penalty in the tunnelling [[Probability|probability]] that no amount of [[Superconducting_magnet|magnet]] engineering can repeal. Proton–[[Boron|boron]]-11, at Z₁Z₂ = 5, is worse again. The reactivities make the penalty concrete. ⟨σv⟩ averages cross-section times relative [[Velocity|speed]] over the [[Probability_density_function|Maxwellian]], folding barrier and temperature into one number. Computed from the Bosch–Hale fit and checked against that paper's own tabulated values by [[Numerical_integration|numerical]] reproduction: - **D–T peaks at ⟨σv⟩ = 8.95×10⁻²² m³/s near T_i ≈ 67 keV** — a genuine interior maximum, well inside the fit's validity. - **D–³He is still rising at 190 keV**, the top of validity, where it reaches only 2.68×10⁻²² m³/s. Its true maximum lies above the parameterisation's range, so a "D–³He peak temperature" citing Bosch–Hale is not a legitimate citation. - At 10 keV the ratio is stark: D–T 1.136×10⁻²² against D–³He 2.126×10⁻²⁵, a factor of 534. Even at the D–T peak the ratio is still 8.9. Two more corrections belong here. The frequently seen "D–T reactivity ≈ 1.1×10⁻²² m³/s" is the value at 10 keV, not the peak — the peak is eight times larger. And the *fit* [[Estimation_theory|uncertainty]] is not the *data* uncertainty: Bosch–Hale reproduce their own D–³He points to 2.5%, but the underlying R-matrix cross-section carries about 10% absolute uncertainty against 3% for D–T. The D–³He rate is simply less well known. Then there is radiation. Bremsstrahlung from the [[Electron|electrons]] scales as n_e Σ n_i Z_i² √T_e, so a fully stripped [[Helium-3|³He]] nucleus at Z = 2 radiates four times as hard as a hydrogenic [[Ion|ion]] at the same [[Density|density]], and the temperature D–³He demands is several times the D–T optimum. At 60–100 keV the radiated [[Photon|photon]] flux is a charge against the [[Energy|power]] balance, not a rounding error. The neutron-fraction table above explicitly *excludes* bremsstrahlung, burnup, fuel dilution and non-thermal effects: it is a reaction-rate calculation, not an engineering power balance. ## The other candidates Proton–boron-11 is the genuinely aneutronic case in principle: p + ¹¹B → 3 ⁴He, releasing about 8.68 MeV entirely into [[Alpha_particle|alpha particles]], with no [[Neutron|neutron]] in the primary channel and no radioactive [[Nuclear_fuel|fuel]] on either side. [[Boron|Boron]] is abundant and cheap on [[Earth|Earth]], which deletes at a stroke the supply objection that dominates D–³He and makes [[Mining_engineering|mining]] a non-issue. It is also harder than anything else here: Z₁Z₂ = 5 puts the reaction in the hundreds-of-keV regime, and a Z = 5 [[Chemical_element|element]] radiates ferociously, so the bremsstrahlung balance rather than the barrier is the standing objection to p–¹¹B as a source of [[Electric_current|electricity]]. There is a data problem too, and it deserves stating rather than papering over. Bosch–Hale does not cover p–¹¹B, and the published reactivity fits disagree by **tens of per cent** — the Nevins–Swain 2000 curve against the later Sikora–Weller 2016 cross-section re-evaluation, which raised ⟨σv⟩ substantially. When two curves for one reaction differ that much, every downstream power balance inherits the disagreement, and any [[Mathematical_model|model]] that picks one silently asserts more than it knows. The microsim above omits p–¹¹B for that reason, and the omission is a finding, not a gap. ³He + ³He → ⁴He + 2p releases about 12.86 MeV with no deuterium in the [[Nuclear_fuel|fuel]] at all and is aneutronic without qualification — but it burns two scarce nuclei per event, hotter still. A thought experiment, not a design. ## The fuel problem, which is the binding one D–³He needs [[Helium-3|helium-3]], and there is almost none. Essentially all commercial ³He is the [[Beta_decay|beta-decay]] [[Decay_product|daughter]] of tritium harvested during weapons-stockpile maintenance; there is no economically significant primary terrestrial source, and the [[Helium|helium]] in [[Natural_gas|natural gas]] carries ³He at parts per billion of the helium itself. The US Government Accountability Office recorded American extraction capacity of 8,000–10,000 L/y against average sales near 30,000 L/y over 2003–2009 — about 1.1–1.3 kg/y of production against three to four times that in demand, most of it now going to [[Neutron_detection|neutron detectors]] and to [[Dilution_refrigerator|dilution refrigerators]] rather than to [[Nuclear_fusion|fusion]]. De Temmerman and colleagues put total terrestrial supply from tritium [[Radioactive_decay|decay]] at about 18 kg/y and the entire US strategic reserve at about 25 kg. Set that against consumption. One tonne of ³He fully burned releases 5.87×10¹⁷ J = 0.587 EJ = 163 TWh thermal, or 18.6 MW·yr per kilogram. On Schmitt's own 2006 figure — 100 kg powering a 1000 MWe plant for a year, implying about 54% net conversion — **a single gigawatt-electric station would eat roughly 100 kg of ³He a year, about five times the entire world's annual production.** The [[Limiting_factor|limiting factor]] for D–³He is not the [[Physics|physics]]. It is the fuel. Which is why the [[Moon|Moon]] appears in every serious proposal. [[Sun|Solar]]-wind ³He implanted in the [[Regolith|regolith]] over billions of years gives a resource in place of order a million tonnes — but at 1.4–20 ppb by mass, meaning of order 10⁸ tonnes of regolith processed per tonne of product, and the defensible *reserve* figures sit three orders of magnitude below the headline. The supply picture belongs to [[Helium-3]]; the mining arithmetic, with its ranges and its resource-versus-reserve distinction, belongs to [[Lunar_resources]], to [[In_situ_resource_utilization|in-situ resource utilisation]] and to [[WT!Space_Mining_In_Minnesota|the space-mining track]]. Aneutronic fusion is the demand side of that ledger, and the ledger does not balance. ## Direct energy conversion: the actual prize The reason to accept all of the above is direct energy conversion. Charged products leaving a magnetic nozzle can be decelerated against an electric field, surrendering kinetic [[Energy|energy]] as [[Voltage|potential difference]] and feeding the [[Electrical_grid|grid]] with no working fluid, no turbine and no [[Second_law_of_thermodynamics|Carnot]] limit. It is an ion accelerator run backwards: an [[Energy_transformation|energy transformation]] with no heat step. In a fuel whose products are all charged, nothing in the [[Energy_engineering|plant]] needs a [[Thermal_engineering|thermal island]] — which is exactly why the surviving few per cent of [[Neutron|neutrons]] force a small one anyway. The same logic drives the [[Fusion_rocket|fusion rocket]] literature. The efficiency assumed for that step is load-bearing and almost always left implicit. Schmitt's 1988 economic geology of lunar ³He states it outright: **"High efficiency (70–80 percent) in energy conversion due to direct conversion of charged particles to electricity."** Every downstream "one tonne of ³He powers X" claim rests on that assumption, and no direct converter has been demonstrated at reactor scale. Audit one claim and the dependence surfaces: 25 tonnes of ³He is 14.7 EJ, about 13.9 quads thermal, against US primary energy consumption of roughly 94–96 quads — so the much-repeated "25 tonnes could supply the entire United States' energy needs for a year" is off by a factor of about seven as stated. Cameron's 1992 original said *electrical* energy, which is nearly defensible; the claim broke when the word was dropped. Aneutronic fusion is worth pursuing on its merits and the [[Emerging_technologies|technology]] is genuinely young — but its [[Power_engineering|power-engineering]] case should always carry the conversion assumption, exactly as its neutron fraction should always carry the fuel ratio. ## Sources - **Bosch, H.-S.; Hale, G. M. (1992).** "Improved formulas for fusion cross-sections and thermal reactivities." *Nuclear Fusion* **32**(4), 611–631. DOI [10.1088/0029-5515/32/4/I07](https://doi.org/10.1088/0029-5515/32/4/I07). The standard R-matrix parameterisation; Table VII the coefficients, Table VIII the tabulated reactivities. Source of the Gamow constants, the peak reactivities, the validity ranges (0.2–100 keV for D–T and D–D, 0.5–190 keV for D–³He) and the fit-versus-data uncertainty distinction. An erratum exists `[UNVERIFIED for volume/page]`. - **CODATA 2022 fundamental constants**, NIST, [physics.nist.gov/cuu/Constants](https://physics.nist.gov/cuu/Constants/Table/allascii.txt). The masses from which the corrected product energies, and the p–¹¹B and ³He–³He Q values quoted here, are derived. - **Ongena, J. (2022).** *EPJ Web of Conferences* **268**, 00011. DOI [10.1051/epjconf/202226800011](https://doi.org/10.1051/epjconf/202226800011). Independent confirmation of the Q values: D–T 17.59, D–³He 18.35, D–D 3.27 and 4.04 MeV. - **Razin, Y. S.; et al.** IAC-12,C4,7-C3.5,10 (Princeton Plasma Physics Laboratory). Source of f_P = 0.07 at 10% deuterium, the "as low as 5%" claim with tritium suppressed, and the f_P = 0.33 figure routinely misquoted as a plasma neutron fraction when it is the neutron fraction of the D–D reactions themselves. - **Parisi, J.; Diallo, A.; Meschini, S.** arXiv:2504.09869. Finds that at lower temperature roughly half the fusion power in a D–³He plasma comes from D–D and secondary reactions — the high end of the honest range. - **University of Wisconsin Fusion Technology Institute, UWFDM-935.** The "on the order of 1% of the energy produced" figure, counting the D–D side reaction and implicitly assuming no tritium burn. - **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 sceptical counterweight: ~18 kg/y terrestrial supply, ~25 kg US strategic reserve, and the 15% D / 85% ³He suppression proposal. Its text contains a units error on regolith ³He concentration — cite the conclusions, not that sentence. - **Schmitt, H. H. (1988).** "Economic Geology of Lunar Helium-3." *Second Symposium on Lunar Bases*, [NTRS 19890005478](https://ntrs.nasa.gov/api/citations/19890005478/downloads/19890005478.pdf). Where the 70–80% direct-conversion assumption is stated explicitly, and therefore the document to check before believing any downstream energy claim. - **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 careful version of the "25 tonnes" claim — 25 t of ³He against one year of US *electrical* energy — and the 7,041 t mineable reserve that the million-tonne headline is not. - **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). The founding paper of the lunar-³He proposition. `[UNVERIFIED at primary — citation metadata confirmed, text not accessible; the ~10⁹ kg attribution is secondhand.]` - **U.S. Government Accountability Office (2011).** *Managing Critical Isotopes*, [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. - **Nevins, W. M.; Swain, R. (2000)** and **Sikora, M. H.; Weller, H. R. (2016)** — the two p–¹¹B reactivity treatments that disagree by tens of per cent. `[UNVERIFIED]` — flagged in the microsim's own build notes as the reason p–¹¹B is not simulated; neither fit was checked against a primary source here, which is exactly the point. - **NRL Plasma Formulary.** Used only as an independent cross-check on the Bosch–Hale reactivities coded into the microsim: agreement within 5% at 5, 10, 20, 50 and 100 keV for D–T, D–D and D–³He. <!-- COMPENDIUMLINK:BEGIN g19 — generated from _registry/plans/THURY_COMPENDIUM_SECTIONS.md; do not hand-edit inside --> *Linked from the [[WT!Thury_Hydrodynamics_Compendium|Thury Hydrodynamics Compendium]], section 14, Helium-3 and fusion.* <!-- COMPENDIUMLINK:END --> ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Aneutronic_fusion) : [Wikitube](https://en.wikitube.io/wiki/Aneutronic_fusion) ## Previous hub tags Hubs: `Life_Physics`. Portals: [[PORTAL_Helium-3]], [[PORTAL_Energy]].