Scalable production of nuclear battery alpha emitters using fusion neutrons

Abstract: Nuclear batteries powered by alpha decay have been deployed successfully for over 60 years, on a worldwide $<sup>{238}$Pu supply of kilograms per year. We show that the 14 MeV neutrons of a single deuterium-tritium fusion plant can produce alpha emitter battery fuels up to tons per year, in three classes: fuels with completely new production pathways ($<sup>{236}$Pu, $<sup>{227}$Ac, $<sup>{210}$Pb), fuels previously proposed whose scarce feedstock the same pathways now breed at scale ($<sup>{232}$U, $<sup>{228}$Th), and the established $<sup>{238}$Pu. OpenMC simulations of actinide channels in a tokamak blanket give, per GW yr of fusion: 11 to 57 kg of $<sup>{236}$Pu, whose chain releases 18 GJ per gram over a century, ending at stable $<sup>{208}$Pb, plus up to 5.2 t of co-product $<sup>{238}$Pu; up to 1.4 t of $<sup>{231}$Pa from thorium, and, from channel with $<sup>{231}$Pa feedstock, up to $\sim$15 t of $<sup>{232}$U or $\sim$122 kg of $<sup>{210}$Pb, with $<sup>{227}$Ac produced at 21 g/yr per tonne of $<sup>{231}$Pa. Neutron capture also upgrades $<sup>{241}$Am to a $<sup>{242}$Cm/$<sup>{242m}$Am/$<sup>{241}$Am/$<sup>{238}$Pu blend with up to 10 times higher power density. The same $<sup>{236}$Pu and $<sup>{232}$U also serve as proliferation safeguards: the $<sup>{237}$Np, $<sup>{232}$Th, and $<sup>{231}$Pa channel products are self-protecting, the plutonium by $<sup>{236}$Pu and $<sup>{238}$Pu decay heat and the 2.6 MeV gammas from $<sup>{208}$Tl content, and similarly the uranium from its $<sup>{232}$U. Many of these fuels ($<sup>{236}$Pu, $<sup>{232}$U, $<sup>{228}$Th, $<sup>{227}$Ac) have an order of magnitude higher power and energy density than current alpha emitters, and at human spaceflight-relevant doses the $<sup>{227}$Ac and $<sup>{210}$Pb chains need less shield mass than $<sup>{238}$Pu or $<sup>{241}$Am above a few hundred watts. Fusion neutrons could therefore enable nuclear batteries at the kilowatt to megawatt scale and unlock new possibilities for power sources requiring exceptionally high energy density.

• The paper uses OpenMC neutronics and depletion simulations to show that a 1.5 GW fusion plant could produce alpha emitters such as 236Pu, 227Ac, 232U, and 210Pb at rates from kilograms to tonnes per year.
• The modeled fuels offer major energy-density and shielding advantages, with 236Pu delivering 18.1 W/g initially and 210Pb providing a low-neutron, comparatively lightweight option for higher-power batteries.
• The paper identifies substantial deployment barriers—including isotope separation, actinide extraction, radon management, safeguards, economics, and the need to validate simulation results experimentally.

Overview and motivation

This paper examines the use of 14 MeV deuterium–tritium (D-T) fusion neutrons as a production engine for alpha-emitting radioisotope fuels for nuclear batteries (radioisotope thermoelectric generators, RTGs, and their dynamic-converter descendants). The central claim is quantitative: a single fusion power plant at the gigawatt-fusion scale could produce alpha battery fuels at ton-per-year rates, in three categories — fuels with entirely new production pathways (${}^{236}$Pu, ${}^{227}$Ac, ${}^{210}$Pb), fuels proposed in 1960 whose scarce feedstock can now be bred at scale (${}^{232}$U, ${}^{228}$Th), and the established ${}^{238}$Pu as a co-product. The author, affiliated with Marathon Fusion, supports the claim with OpenMC neutronics and depletion simulations of actinide channels inserted into tokamak blankets, using ENDF/B-VIII.0 cross sections.

The physical motivation rests on the energy density hierarchy of alpha decay chains. The ${}^{236}$Pu chain, a nine-step sequence ending at stable ${}^{208}$Pb, releases 45 MeV of locally deposited heat per atom — 18 GJ/g, or 0.020% of the fuel rest mass, roughly a fifth of the fraction released by ${}^{235}$U fission and eight times the 5.6 MeV of ${}^{238}$Pu. This is a strong claim: a decay chain, with no reactor, delivers energy densities approaching a meaningful fraction of fission fuel.

Decay chain properties

The paper compares six chains in detail. The ${}^{227}$0Pb chain (5.8 MeV average, ending at stable ${}^{227}$1Pb) is nearly gamma-free; its shielding is set by bremsstrahlung from the 1.16 MeV ${}^{227}$2Bi beta, and its spent form is stable lead, avoiding the long-lived ${}^{227}$3U and ${}^{227}$4Np terminal daughters of the ${}^{227}$5Pu and ${}^{227}$6Am chains. The ${}^{227}$7Ac chain releases 35 MeV, terminating at ${}^{227}$8Pb with its hardest line at 832 keV, a factor of three below the 2.6 MeV ${}^{227}$9Tl gamma of the ${}^{210}$0Pu chain. Specific power spans three orders of magnitude across fuels: ${}^{210}$1Th exceeds 100 W/g at one year versus 0.1 W/g for ${}^{210}$2Am. Fresh ${}^{210}$3Pu delivers 18.1 W/g, remaining above 2 W/g for a century and exceeding ${}^{210}$4Pu by a factor of eight in integrated energy over 500 years. Only the ${}^{210}$5Pu/${}^{210}$6U/${}^{210}$7Th and ${}^{210}$8Ac chains sustain multi-W/g output over 100-year horizons; every single-alpha parent falls an order of magnitude below. The trade-off is gamma output: the high-energy-density chains emit gamma power at the 0.1 W/g scale, four orders of magnitude above ${}^{210}$9Pu.

Production pathways and blanket performance

Production rests on three 14 MeV threshold ${}^{232}$0 reactions: ${}^{232}$1Np${}^{232}$2 at 0.47 b feeding ${}^{232}$3Pu via 22.5 h ${}^{232}$4Np; ${}^{232}$5U${}^{232}$6 at 0.89 b breeding ${}^{232}$7Np feedstock; and ${}^{232}$8Th${}^{232}$9 at 1.49 b breeding ${}^{228}$0Pa, the feedstock for ${}^{228}$1U, ${}^{228}$2Pb, and (by alpha decay, at 21 g/yr per tonne) ${}^{228}$3Ac. The blanket architecture places a 5 mm (variable) actinide channel against the tungsten first wall, ahead of a ${}^{228}$4Li-enriched FLiBe breeder; the pure-FLiBe baseline gives TBR = 1.22 via ${}^{228}$5Be${}^{228}$6.

Key quantitative results for a 1.5 GW${}^{228}$7 plant include:

Channel (operating point)	Product	Rate	Co-product	Total power
NpO${}^{228}$8, 5 mm	${}^{228}$9Pu	16 kg/yr	673 kg/yr ${}^{238}$0Pu	2.9 GW
NpO${}^{238}$1, 30 mm	${}^{238}$2Pu	85 kg/yr	7.8 t/yr ${}^{238}$3Pu	19 GW
${}^{238}$4U, 50 mm	${}^{238}$5Np	1.23 t/yr	—	7.2 GW
${}^{238}$6Th, 100 mm (Pa retained)	${}^{238}$7U	804 kg/yr	${}^{238}$8Pb 6.3 kg/yr	3.0 GW
Solid ${}^{238}$9Pa, 200 mm	${}^{236}$0Pb	183 kg/yr	${}^{236}$1U 22.5 t/yr	78 GW

These are striking figures. The ${}^{236}$2Pu co-production alone, at 0.7 to 7.8 t/yr, is three orders of magnitude above current global production and hundreds of times NASA's 1.5 kg/yr target. The ${}^{236}$3Np channel is unambiguously a fusion-fission hybrid: fission events outnumber ${}^{236}$4 by up to 85, and fission heat dominates the energy product. The neptunium channel reaches criticality at ~26.5 mm as metal; the paper therefore constrains thickness by volumetric heating (1044 MW/m³ at 5 mm, comparable to a PWR pellet; HFIR-like at 50 mm) rather than criticality, with ${}^{236}$5 held to 0.68 at the 30 mm NpO${}^{236}$6 operating point. Depletion simulations show 87% of the ${}^{236}$7Np consumed in 30 years, and the ${}^{236}$8Pu inventory peaking near 61 kg at year five — so economic ${}^{236}$9Pu production requires extraction cycles short compared to its 2.86 yr half-life. The paper concedes this directly: isotopically pure ${}^{208}$0Pu is not realistically achievable at scale, with fresh plutonium only 1.1–2.4% ${}^{208}$1Pu (raisable to 6.2 at% with a Li${}^{208}$2O absorber layer, at no TBR cost).

A significant secondary result is that the daughters of impure ${}^{208}$3Pu are extremely pure in activity terms: uranium stripped from held plutonium is 99.98% ${}^{208}$4U by activity, and the thorium fraction is 99.999% ${}^{208}$5Th. The paper argues this reframes the purity problem — the low atomic fraction of ${}^{208}$6Pu may be irrelevant if the goal is ${}^{208}$7U and ${}^{208}$8Th production. Flight-grade ${}^{208}$9Pu, by contrast, requires aging: the ${}^{235}$0Pu content decays by a factor of ten every 9.8 years, reaching the historical 2 ppm specification after 37 years with 75% of the ${}^{235}$1Pu retained, yielding 5.8 t/yr of flight-grade material in steady state.

The ${}^{235}$2Ac route is notable for its dependence on stockpiling. The world's entire declared separated ${}^{235}$3Pa stock is 125 g (UKAEA, 1961); a 1.5 GW${}^{235}$4 plant at the tritium self-sufficiency limit breeds that much in under half an hour. But ${}^{235}$5Ac recovery at kilogram-per-year rates requires tonne-scale ${}^{235}$6Pa stockpiles held outside the blanket for years, since in-blanket neutron burning to ${}^{235}$7U outpaces alpha decay to ${}^{235}$8Ac by four orders of magnitude. The paper is explicit that ${}^{235}$9Pa is not available at present and that the ${}^{238}$0Pa-fed blankets presume decades of thorium-blanket accumulation.

Economics

Using a revenue-per-fusion-neutron framework, the paper finds ${}^{238}$1Pu worth ${}^{238}$2 per source neutron, about 14 times the value of the same neutron sold as electricity. At 100 kg/yr, the hypothetical ${}^{238}$3Pu market is ${}^{238}$4_\mathrm{fus}${}^{238}$5{}^{236}238{}^{238}6{}^{238}238{}^{238}7{}^{238}Pu</sup></sup></sup>atthousandsoftimestheNASAdemand.</p><h2class=′paper−heading′id=′shield−mass−and−the−low−power−high−power−inversion′>Shieldmassandthelow−power/high−powerinversion</h2><p>ConvertingtheNelsonetal.shieldanalysisintototalshieldmassfor100WPu</sup></sup></sup> at thousands of times the NASA demand.</p> <h2 class='paper-heading' id='shield-mass-and-the-low-power-high-power-inversion'>Shield mass and the low-power/high-power inversion</h2> <p>Converting the Nelson et al. shield analysis into total shield mass for 100 W{}^{238}$8 sources at 10 mrem/h contact dose, the paper establishes a regime inversion with source power. Below a few hundred watts, the neutron-limited single-alpha fuels (${}^{238}$9Pu, ${}^{227}$00Am) are lightest — 8–11 kg in minimum-mass LiH, or unshielded entirely at 100 W and 100 mrem/h — while the gamma chains require 20 kg (${}^{227}$01Ac) to 385 kg (${}^{227}$02Th) of depleted uranium. Above a few hundred watts the ordering reverses: photon attenuation costs 1.5–3 cm of DU per decade of power versus 13–19 cm of moderator per decade for fast neutrons, so shield mass grows roughly ten times faster for the neutron-limited fuels. At 1 MW${}^{227}$03 the ${}^{227}$04Ac chain needs a quarter of the ${}^{227}$05Am shield mass. The ${}^{227}$06Pb chain is the standout at moderate power: 30 kg of shield at 100 W${}^{227}$07, no ${}^{227}$08 neutron emission from lead metal at all, and a shield growing only 1 cm of DU per decade of power. This inversion is the paper's strongest design-level claim: it implies nuclear batteries at kilowatt-to-megawatt scale favor the high-energy-density chains precisely where RTG fuel supply has historically forced a switch to fission reactors.

Radon separation of the ${}^{227}$09Pu chain

The 2.6 MeV ${}^{227}$10Tl gamma has historically disqualified ${}^{227}$11Pu. The paper's mitigation exploits the single noble-gas link in the chain: ${}^{227}$12Rn (55.6 s half-life) can emanate from porous fuel at fractions of 0.5–0.7 when grain size approaches the ~50 nm recoil range, carrying the gamma-emitting lower chain segment out while the upstream four-alpha segment (22 MeV) stays as heat. Four implementations are sketched: a shielded back-end plate-out chamber for ground installations (~2–5 t of tungsten or lead, a fixed cost growing one tenth-value layer per decade of inventory); overboard venting as a ~40 km/s ion stream in deep space, where the 10.6 h ${}^{227}$13Pb delay places the gamma source ~10${}^{227}$14 km behind the vehicle; a boom-mounted decay chamber with a ~10 kg shadow shield; and a passive central-getter design that reduces external shielding by 30–55%. The paper notes plainly that the 4.0 s ${}^{227}$15Rn of the ${}^{227}$16Ac chain is too short-lived for these techniques, so that chain cannot shed its gammas, and that venting is safe only in vacuum far from life.

Limitations and open questions

The paper is candid about its constraints. The ${}^{227}$17Pu chain's safety for crewed spaceflight is identified as the largest open question, contingent on demonstrating ${}^{227}$18Rn emanation fractions near 0.9 at ~100 W/cm³ — a level approached but not demonstrated in targeted-alpha-therapy ${}^{227}$19Pb generators. The neptunium blanket requires engineered subcritical margin (${}^{227}$20 = 0.68 at the preferred operating point) and continuous molten-salt extraction at scale, neither of which exists. The safeguards analysis is self-described as "a first pass": the ${}^{227}$21U channel breeds essentially pure ${}^{227}$22Pu at one significant quantity every 20 hours at 50 mm — faster than any previously analyzed fusion breakout configuration — and the paper concludes that the ${}^{227}$23U channel is "relatively uninteresting" for batteries and of much higher proliferation concern, whereas the ${}^{227}$24Th and ${}^{227}$25Np channels appear manageable through ${}^{227}$26U and ${}^{227}$27Pu self-protection. The heuristic that ~100 ppm of ${}^{227}$28U or ${}^{227}$29Pu renders material self-protecting is acknowledged as an assumption, weakened by the decades-long ingrowth of ${}^{227}$30Tl and by the possibility of chemical removal of daughters. All yields are OpenMC tallies with ENDF/B-VIII.0 data; the 14 MeV isomeric branching of ${}^{227}$31Am lacks measurements. Finally, the economics omit all costs and assume product separability and market absorption.

Conclusion

The paper establishes, through systematic OpenMC blanket modeling, that fusion neutrons can supply alpha battery fuels at rates of kilograms to tonnes per year per plant, including three chains with no prior scalable production route and two 1960s-proposed fuels whose feedstock bottleneck the same blankets remove. The combination of order-of-magnitude gains in power and energy density, a shield-mass inversion favoring the new chains above a few hundred watts, and radon-based gamma management constitutes a coherent case that fusion neutrons could support nuclear batteries from watts to megawatts thermal. The results remain simulation-level: the decisive open questions are experimental — radon emanation at full activity, extraction chemistry on hot actinide streams, sustained safeguards over 30-year irradiations, and launch licensing for fuels without flight heritage.