Fusion Transmuters: Neutron-Driven Isotope Production

• Pure transmuter fusion systems are fusion-driven neutron sources that optimize D–T reactions to prioritize isotope transmutation over net electricity generation.
• They employ tailored blanket designs to harness 14.1 MeV neutrons for targeted nuclear reactions, enabling applications such as medical isotope production, fusion breeding, and waste transmutation.
• These systems achieve economic viability through optimized neutron economy and hybrid breakeven criteria, allowing lower fusion gain requirements with diversified product streams.

Pure transmuter fusion systems are fusion-driven neutron sources whose primary product is transmuted isotopes rather than electricity. In the most explicit recent definition, a pure transmuter operates its plasma specifically to maximize D–T neutron output into a blanket engineered for targeted nuclear reactions, and its revenue stream is dominated by isotope sales rather than electricity; closely related arXiv usage extends the concept to breeder blankets for $^{233}\mathrm{U}$, subcritical molten-salt transmutators for transuranic destruction, and muon-catalyzed fusion neutron sources optimized for isotope production rather than net power export (Parisi et al., 10 Dec 2025, Manheimer, 2022, Tanner et al., 2021).

1. Definition and conceptual scope

The unifying feature of a pure transmuter is that neutrons are treated as the primary economic or strategic output. In D–T systems this usually means exploiting 14.1 MeV neutrons in a blanket that contains selected feedstocks, tritium-breeding materials, and sometimes neutron multipliers. The system may still generate and internally use electricity—for separation, pumping, and control—but it does not require net electric output to be viable if the value per neutron of the transmuted product is sufficiently high. This is the central distinction from electricity-first fusion plants, which rely on net-electric sale and therefore require substantially higher fusion gain and tighter recirculating-power margins (Parisi et al., 10 Dec 2025).

Across the literature, the same design philosophy appears in several technically distinct forms. One branch targets medical radioisotopes and precious metals from D–T blankets. Another branch uses the 14 MeV source to breed $^{233}\mathrm{U}$ from $^{232}\mathrm{Th}$. A third uses distributed D–T neutron sources to drive subcritical transuranic destruction in molten salts. A fourth treats muon-catalyzed D–T fusion as a neutron source whose lack of external heating relaxes heat-flux constraints. These are not identical systems, but they share the same mission logic: neutron production is valuable in its own right (Manheimer, 2022, Tanner et al., 2021, Parisi et al., 26 Nov 2025).

Archetype	Neutron source	Primary product or mission
Blanket transmuter	Externally heated D–T fusion	Medical isotopes; gold; optional electricity
Fusion breeder	D–T core with breeder blanket	$^{233}\mathrm{U}$ and tritium
Molten-salt transmutator	Distributed 14 MeV D–T sources	TRU destruction
$\mu$CF transmuter	Muon-catalyzed D–T cycles	High-value isotopes

2. Neutron physics and blanket transmutation

The fundamental source reaction is

$\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$

with total fusion energy

$E_\mathrm{fus}=17.6~\mathrm{MeV},$

of which fractions $f_\alpha=1/5$ and $f_n=4/5$ are carried by $\alpha$ particles and neutrons, respectively. The neutron birth rate is

$^{233}\mathrm{U}$0

Because the neutron energy is 14.1 MeV, D–T systems can drive threshold reactions such as $^{233}\mathrm{U}$1, $^{233}\mathrm{U}$2, and $^{233}\mathrm{U}$3 that are weak or inaccessible in fission spectra. This is the basis for high-specific-activity isotope production and for fast-spectrum chrysopoeia in $^{233}\mathrm{U}$4 (Parisi et al., 10 Dec 2025, Rutkowski et al., 17 Jul 2025).

Blanket design is therefore a neutronic optimization problem. Feedstock selection is governed by the intended reaction channel, the energy dependence of $^{233}\mathrm{U}$5, parasitic channels such as $^{233}\mathrm{U}$6, chemical separability of the product, and tritium-breeding constraints. High-value radioisotope pathways often favor $^{233}\mathrm{U}$7 or $^{233}\mathrm{U}$8 because the proton number changes, enabling chemical rather than isotopic separation. By contrast, $^{233}\mathrm{U}$9 pathways are important both for product formation and for neutron multiplication. In blanket accounting, the thermal-power multiplication is written

$^{232}\mathrm{Th}$0

with $^{232}\mathrm{Th}$1 typically $^{232}\mathrm{Th}$2–$^{232}\mathrm{Th}$3 when neutron slowing, $^{232}\mathrm{Th}$4, endothermic $^{232}\mathrm{Th}$5, gamma absorption, and leakage are summed (Parisi et al., 10 Dec 2025).

For a feedstock with number density $^{232}\mathrm{Th}$6 and cross section $^{232}\mathrm{Th}$7, the production rate is

$^{232}\mathrm{Th}$8

In the thin-blanket limit,

$^{232}\mathrm{Th}$9

where $^{233}\mathrm{U}$0. This makes blanket thickness, feedstock density, and wall loading direct control parameters for throughput. When subcritical fission multiplication is added, the total neutron availability per fusion neutron becomes

$^{233}\mathrm{U}$1

and the generalized production rate is $^{233}\mathrm{U}$2 (Parisi et al., 10 Dec 2025).

The same fast-neutron logic underlies the mercury-to-gold concept. The desired $^{233}\mathrm{U}$3 channel has a threshold at about $^{233}\mathrm{U}$4, so a thin first-wall mercury layer can intercept the still-hard D–T spectrum, produce $^{233}\mathrm{U}$5 and $^{233}\mathrm{U}$6, and simultaneously act as a neutron multiplier for downstream tritium breeding. In one modeled ARC-class geometry, a $^{233}\mathrm{U}$7 Li–Hg channel at $^{233}\mathrm{U}$8 at% Hg and $^{233}\mathrm{U}$9 at% Li, with both species $\mu$0 enriched in $\mu$1 and $\mu$2, yielded a total tritium breeding ratio of approximately $\mu$3, with $\mu$4 from the inner LiHg layer and $\mu$5 from the outer FLiBe blanket (Rutkowski et al., 17 Jul 2025).

3. Operating regimes and hybrid breakeven

Pure transmuters are defined as much by operating regime as by hardware. In the isotope-production framework, the decisive quantity is not electric breakeven but value per neutron. The literature explicitly contrasts electricity-only fusion, for which meaningful margins generally require $\mu$6–$\mu$7 and often $\mu$8–$\mu$9, with transmuters that can be economically viable at $\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$0–$\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$1 because select products raise value per neutron by orders of magnitude above electricity (Parisi et al., 10 Dec 2025).

This is formalized by replacing a pure power-balance criterion with a hybrid breakeven condition. The revenue rate is written

$\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$2

with

$\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$3

The corresponding full-plant condition is framed in terms of net present value,

$\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$4

rather than engineering breakeven alone. A hybrid “engineering breakeven” is then defined by an electric-equivalent product power and the threshold

$\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$5

Within this formalism, $\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$6 can drive $\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$7 to extremely small values once $\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$8 exceeds very small thresholds, whereas for $\mathrm{D}+\mathrm{T}\rightarrow \alpha + n,$9 at $E_\mathrm{fus}=17.6~\mathrm{MeV},$0, $E_\mathrm{fus}=17.6~\mathrm{MeV},$1 is roughly halved relative to electricity-only operation (Parisi et al., 10 Dec 2025).

At low gain, the flagship example is $E_\mathrm{fus}=17.6~\mathrm{MeV},$2: a $E_\mathrm{fus}=17.6~\mathrm{MeV},$3 system could fulfill global $E_\mathrm{fus}=17.6~\mathrm{MeV},$4 demand with $E_\mathrm{fus}=17.6~\mathrm{MeV},$5. At higher gain, co-generation becomes the dominant architecture. For $E_\mathrm{fus}=17.6~\mathrm{MeV},$6, present gold prices and $E_\mathrm{fus}=17.6~\mathrm{MeV},$7 lower the required $E_\mathrm{fus}=17.6~\mathrm{MeV},$8 for viability from electricity-only values of $E_\mathrm{fus}=17.6~\mathrm{MeV},$9–$f_\alpha=1/5$0 to $f_\alpha=1/5$1–$f_\alpha=1/5$2, and in a $f_\alpha=1/5$3 example with $f_\alpha=1/5$4 and $f_\alpha=1/5$5B (Parisi et al., 10 Dec 2025).

Muon-catalyzed fusion yields an even sharper contrast between neutron value and energy breakeven. Because no external heating is required, the effective $f_\alpha=1/5$6 is formally infinite from the wall-loading standpoint. The wall load is

$f_\alpha=1/5$7

rather than

$f_\alpha=1/5$8

for externally heated systems. In the corresponding hybrid-gain model, the required number of catalyzed fusions per muon falls from $f_\alpha=1/5$9 for electricity-only operation to approximately $f_n=4/5$0 for $f_n=4/5$1, $f_n=4/5$2 for $f_n=4/5$3, $f_n=4/5$4 for $f_n=4/5$5, and $f_n=4/5$6 for $f_n=4/5$7 (Parisi et al., 26 Nov 2025).

4. Principal implementations and case studies

Recent work presents pure transmuters not as a single reactor class but as a family of neutron-economy architectures ranging from few-megawatt isotope sources to gigawatt-class co-generators and subcritical breeder systems (Parisi et al., 10 Dec 2025, Parisi et al., 4 Nov 2025, Rutkowski et al., 17 Jul 2025, Manheimer, 2022, Tanner et al., 2021, Parisi et al., 26 Nov 2025).

Pathway	Representative figure	Stated significance
$f_n=4/5$8	$f_n=4/5$9–$\alpha$0	Could fulfill global $\alpha$1 demand with $\alpha$2
$\alpha$3	About $\alpha$4	Compatible with tritium breeding and electricity production
$\alpha$5	About $\alpha$6 $\alpha$7 atoms per 14 MeV neutron	One breeder can fuel about 10 thermal reactors of equal neutron power, or about 5 of equal total power
$\alpha$8 in $\alpha$9CF	$^{233}\mathrm{U}$00, up to $^{233}\mathrm{U}$01	Ten times current global supply in the abstract
TRU molten-salt transmutation	Mid-$^{233}\mathrm{U}$02	Approximately $^{233}\mathrm{U}$03 TRU incineration

The $^{233}\mathrm{U}$04 case is the canonical few-megawatt pure transmuter. For a toroidal device with $^{233}\mathrm{U}$05, aspect ratio $^{233}\mathrm{U}$06, elongation $^{233}\mathrm{U}$07, and first-wall flux $^{233}\mathrm{U}$08, the blanket is sized by $^{233}\mathrm{U}$09, and the pathway is attractive because $^{233}\mathrm{U}$10 changes $^{233}\mathrm{U}$11, enabling rapid chemical separation and high specific activity. More broadly, high-energy fusion neutrons were shown to support many clinically important products—among them $^{233}\mathrm{U}$12, $^{233}\mathrm{U}$13, $^{233}\mathrm{U}$14, $^{233}\mathrm{U}$15, $^{233}\mathrm{U}$16, $^{233}\mathrm{U}$17, $^{233}\mathrm{U}$18, $^{233}\mathrm{U}$19, $^{233}\mathrm{U}$20, $^{233}\mathrm{U}$21, $^{233}\mathrm{U}$22, $^{233}\mathrm{U}$23, $^{233}\mathrm{U}$24, $^{233}\mathrm{U}$25, $^{233}\mathrm{U}$26, $^{233}\mathrm{U}$27, $^{233}\mathrm{U}$28, $^{233}\mathrm{U}$29, $^{233}\mathrm{U}$30, and $^{233}\mathrm{U}$31—with representative outputs such as $^{233}\mathrm{U}$32 for $^{233}\mathrm{U}$33, $^{233}\mathrm{U}$34 natural and $^{233}\mathrm{U}$35 enriched for $^{233}\mathrm{U}$36, and $^{233}\mathrm{U}$37 natural and $^{233}\mathrm{U}$38 for $^{233}\mathrm{U}$39 with enriched $^{233}\mathrm{U}$40 (Parisi et al., 10 Dec 2025, Parisi et al., 4 Nov 2025).

At the large-scale end, chrysopoeia is the best-developed electricity-compatible blanket concept. In an ARC-class $^{233}\mathrm{U}$41 tokamak with $^{233}\mathrm{U}$42, $^{233}\mathrm{U}$43, elongation $^{233}\mathrm{U}$44, triangularity $^{233}\mathrm{U}$45, a two-layer blanket containing a $^{233}\mathrm{U}$46 Li–Hg channel produced $^{233}\mathrm{U}$47 of $^{233}\mathrm{U}$48 in transport and $^{233}\mathrm{U}$49 in coupled transport-plus-depletion, with $^{233}\mathrm{U}$50 of $^{233}\mathrm{U}$51 removed as impurity. The normalized output is about $^{233}\mathrm{U}$52, and because the $^{233}\mathrm{U}$53 channel also acts as a neutron multiplier, the scheme was presented as compatible with tritium self-sufficiency and with essentially unchanged electricity production (Rutkowski et al., 17 Jul 2025).

Fusion breeding uses the same neutron resource for a different product. In the thorium chain,

$^{233}\mathrm{U}$54

with $^{233}\mathrm{U}$55 about $^{233}\mathrm{U}$56 minutes and $^{233}\mathrm{U}$57 about a month. One realistic blanket geometry was reported to produce about $^{233}\mathrm{U}$58 $^{233}\mathrm{U}$59 atoms, as well as the necessary tritium atom, from every 14 MeV neutron; because each $^{233}\mathrm{U}$60 fission releases about $^{233}\mathrm{U}$61, one breeder can fuel about $^{233}\mathrm{U}$62 thermal reactors of equal neutron power, or about $^{233}\mathrm{U}$63 of equal total power (Manheimer, 2022).

A different branch of the field treats fusion neutrons as a controllable source for waste transmutation. In the molten-salt transmutator concept, up to $^{233}\mathrm{U}$64 tunable D–T sources are distributed in a phyllotaxis pattern around a subcritical cylindrical core $^{233}\mathrm{U}$65 in inner diameter and $^{233}\mathrm{U}$66 in inner length, with a $^{233}\mathrm{U}$67 graphite reflector. Using MCNP with SCALE/ORIGEN-S and AI-assisted source control, the system was studied at operating points up to $^{233}\mathrm{U}$68 $^{233}\mathrm{U}$69; the paper reports $^{233}\mathrm{U}$70 TRU transmuted per $^{233}\mathrm{U}$71 for SNF-like compositions, $^{233}\mathrm{U}$72 for a general TRU mix, and an illustrative $^{233}\mathrm{U}$73 static-phase run with about $^{233}\mathrm{U}$74 transmuted (Tanner et al., 2021).

Muon-catalyzed fusion extends the pure-transmuter paradigm to nonthermal neutron sources. For the $^{233}\mathrm{U}$75 chain, a spherical blanket constrained by $^{233}\mathrm{U}$76 and loaded with $^{233}\mathrm{U}$77 of $^{233}\mathrm{U}$78 was found capable of producing up to approximately $^{233}\mathrm{U}$79 of $^{233}\mathrm{U}$80 at a muon rate of $^{233}\mathrm{U}$81 and $^{233}\mathrm{U}$82. The abstract characterizes this as ten times current global supply (Parisi et al., 26 Nov 2025).

5. Engineering constraints, materials, and product handling

The central engineering constraint in D–T transmuters is the coupling between neutron flux, heat load, and material limits. For externally heated systems,

$^{233}\mathrm{U}$83

so low $^{233}\mathrm{U}$84 increases wall heat per neutron. This motivates both higher $^{233}\mathrm{U}$85 and architectures that decouple heat-handling area from neutron-intercept area. One proposed strategy is the magnetic mirror, defined by a heat-flux spread factor

$^{233}\mathrm{U}$86

with feedstock-burn-rate enhancement

$^{233}\mathrm{U}$87

Another is asymmetric neutron wall loading using spin-polarized D–T fuel: in a modeled tokamak, a “parallel” polarization mode captured two-thirds of neutrons on only $^{233}\mathrm{U}$88 of wall area versus $^{233}\mathrm{U}$89 for isotropic emission, yielding an approximately $^{233}\mathrm{U}$90 FBR increase relative to unoptimized loading (Parisi et al., 10 Dec 2025).

Blanket materials must reconcile feedstock reactivity, tritium breeding, neutron multiplication, corrosion, activation, and downstream separation. For medical-isotope blankets, designers favor large $^{233}\mathrm{U}$91 at 14.1 MeV, low $^{233}\mathrm{U}$92, and element-changing channels that simplify chemistry. For gold production, the first wall and structures in the modeled design used $^{233}\mathrm{U}$93 W, $^{233}\mathrm{U}$94 and $^{233}\mathrm{U}$95 V–4Cr–4Ti layers, and a $^{233}\mathrm{U}$96 Eurofer97 blanket tank, with the Li–Hg channel at $^{233}\mathrm{U}$97. The same work also emphasized mercury toxicity, vapor control, compatibility of structural materials with liquid Hg, Li, or Li–Hg at $^{233}\mathrm{U}$98, gamma shielding for magnets, and the influence of residual $^{233}\mathrm{U}$99 on $^{232}\mathrm{Th}$00 impurity and cooldown time (Rutkowski et al., 17 Jul 2025).

Separation chemistry is not ancillary; it is often the key enabler of economic transmutation. Element-changing pathways permit rapid chemical extraction at rates $^{232}\mathrm{Th}$01, which is particularly important for short-lived parents and for high specific activity. The isotope-production papers emphasize rapid extraction for $^{232}\mathrm{Th}$02, gas handling for noble gases such as $^{232}\mathrm{Th}$03, and established rare-earth and copper separations for $^{232}\mathrm{Th}$04, $^{232}\mathrm{Th}$05, and $^{232}\mathrm{Th}$06. In the gold system, the inner Li–Hg loop circulates to a downstream processing module, where gold is selectively removed by gettering or alloying onto materials such as copper or tantalum; because $^{232}\mathrm{Th}$07 has $^{232}\mathrm{Th}$08 and $^{232}\mathrm{Th}$09 has $^{232}\mathrm{Th}$10, the processing schedule must allow decay to $^{232}\mathrm{Th}$11 before capture or during hold-up (Parisi et al., 10 Dec 2025, Rutkowski et al., 17 Jul 2025).

Tritium logistics differ by scale. For megawatt pure transmuters, external tritium purchase can be economically viable because the value per neutron in selected pathways is high and $^{232}\mathrm{Th}$12/$^{232}\mathrm{Th}$13 channels are only weakly parasitic for breeding. For larger co-generators, self-sufficient breeding with $^{232}\mathrm{Th}$14 becomes favored, and lithium inventory and enrichment become major costs. In the molten-salt waste-transmutator literature, safety is tied instead to deliberate subcriticality, electronic shutdown by source turnoff, and passive dump to a boronated drain tank through a freeze plug (Parisi et al., 10 Dec 2025, Tanner et al., 2021).

Muon-catalyzed transmuters relax one of the hardest constraints because no external heating is present. With $^{232}\mathrm{Th}$15, the corresponding wall heat flux is about $^{232}\mathrm{Th}$16 from $^{232}\mathrm{Th}$17-heating alone, and the blanket geometry can be sized by $^{232}\mathrm{Th}$18. The engineering burden then shifts toward muon stopping, cycling, target optical depth, and chemical processing of the product chain rather than to absorbed heating power (Parisi et al., 26 Nov 2025).

6. Misconceptions, controversies, and scaling outlook

A persistent theme in the literature is that pure transmuters should not be evaluated by the same criteria as grid-oriented pure fusion plants. One breeder-oriented paper argues that breeder requirements are “relaxed by at least an order of magnitude” relative to pure fusion and that commercial pure fusion is “extremely unlikely to be in this century” on tokamak pathways, whereas breeder deployment “not too long after midcentury” is plausible. In that view, the pragmatic route for fusion is to value the neutron first and electricity second (Manheimer, 2022).

There is also a terminological ambiguity. In one strand, “pure transmuter” means a fusion blanket whose revenue stream is dominated by isotope sales. In another, it means a breeder that supplies fissile fuel to thermal reactors. In a third, it denotes a subcritical waste-burning transmutator driven by compact neutron sources. These usages are compatible at the level of neutron economics, but they correspond to different system objectives, regulatory frameworks, and product chains. A plausible implication is that “pure transmuter fusion system” functions more as a mission class than as a single canonical reactor type.

The most speculative extension of the term appears in a chemonuclear paper that proposes metal-like hydrides and electron-donor mixtures as “pure” fusion-transmutation media producing fast $^{232}\mathrm{Th}$19 ions with suppressed prompt $^{232}\mathrm{Th}$20 and neutron emission. That work introduces a thermodynamic multiplier

$^{232}\mathrm{Th}$21

into nuclear rates, claims $^{232}\mathrm{Th}$22–$^{232}\mathrm{Th}$23 in some regimes, and suggests radiation-less multi-$^{232}\mathrm{Th}$24 channels and $^{232}\mathrm{Th}$25-missing positron annihilation. The same paper explicitly notes, however, that coupling $^{232}\mathrm{Th}$26 to nuclear tunneling in this way is not part of standard nuclear reaction theory in solids and that the quantitative claims require stringent, reproducible, multimodal verification. Within the current literature, this remains controversial rather than established (Ikegami, 4 Mar 2025).

The scaling outlook is correspondingly bifurcated. On the conservative side, recent D–T blanket analyses present a staged pathway: near-term megawatt-scale pure transmuters producing medical isotopes at $^{232}\mathrm{Th}$27–$^{232}\mathrm{Th}$28, followed by gigawatt-class co-generators producing electricity and gold at $^{232}\mathrm{Th}$29–$^{232}\mathrm{Th}$30, with fleet-level extrapolations to terawatt deployment. One study estimated that the entire gold market, approximately $^{232}\mathrm{Th}$31B/year, could support about $^{232}\mathrm{Th}$32 fusion capacity at fleet-averaged $^{232}\mathrm{Th}$33; the entire $^{232}\mathrm{Th}$34 market supports about $^{232}\mathrm{Th}$35 at $^{232}\mathrm{Th}$36, and $^{232}\mathrm{Th}$37 supports about $^{232}\mathrm{Th}$38 at $^{232}\mathrm{Th}$39 (Parisi et al., 10 Dec 2025). On the engineering side, the decisive open problems remain blanket materials under fast-neutron fluence, tritium handling, isotope-specific separation flowsheets, source reliability, and transport-validated neutronic optimization. The concept is therefore best understood not as a settled reactor design, but as a broad reorientation of fusion around the direct value of 14 MeV neutrons.