Low Energy Nuclear Reactions

Speculative If a condensed-matter nuclear effect exists at the scale its proponents claim — which no funded programme has confirmed, and section 7 says what would settle it — the prize is a heat source that runs on hydrogen isotopes and needs no plasma, no magnets and no gigawatt-scale plant. The specific figure the field puts on it is 80 to 400 watts of thermal power per kilogram of loaded metal, sustained, from fuel that is abundant and carries no long-lived radioactive inventory. Against a mainstream fusion programme that after decades of effort has produced net electricity nowhere and draws more than $1 billion a year in US public funds, a bench-top box at that power density would be an energy source with no existing category. That is the entire reason a thrice-buried idea is being funded again.

Established What is actually banked today is a bound, not a payoff. Two privately funded programmes — Google's roughly $10 million, three-laboratory effort and the 45,000-square-foot Project Callisto, between them running more than 200,000 calorimeter-hours across 500-plus experiment types — measured hydrogen loading, calorimetry and nuclear yields to milliwatt sensitivity and found no excess heat. The value delivered so far is the honesty of that null: a tightened upper limit and a public method, which is a real product even when the effect is absent.

Frontier Who would collect the payoff is the sharpest distributional fact here: nobody yet, because there is no verified output to sell. The mainstream fusion programme draws more than $1 billion a year in US public funding; the entire low-energy nuclear reactions field, public and private, has attracted a few tens of millions of dollars across a decade. If the effect is real the gain is large and the fuel is abundant; if it is not, the money bought a rigorous exclusion, which is the second-best outcome and the one the record currently supports.

Speculative The condition, stated as a number, is the ratio the field has never closed under blinded conditions: excess heat and helium-4 produced together at 23.8 MeV per helium-4 atom. Until that single measurement exists, replicated and pre-registered, every upside sentence above rests on a premise the bottleneck list in section 5 shows has failed reliable replication for thirty-seven years.

Established The subject is whether nuclear reactions can occur in a room-temperature solid at rates and yields that ordinary nuclear physics says are impossible. In March 1989 Martin Fleischmann and Stanley Pons announced that a palladium cathode electrolysing heavy water (D₂O) produced more heat than the electricity put in, and attributed it to deuterium nuclei fusing inside the metal lattice. The claim was extraordinary because two deuterons must overcome the Coulomb barrier — the electrostatic repulsion between two positive charges — which at room temperature is suppressed by an astronomically large factor relative to the temperatures inside the Sun or a tokamak. A lattice-mediated route around that barrier would be new physics.

Established Within a year the claim collapsed as a reproducible result. Major laboratories — Caltech, MIT, Harwell — failed to see the heat or the commensurate neutrons, and the field acquired a reputation as the textbook case of what Irving Langmuir called pathological science, a topic this corpus treats separately in Pathological Science. The name migrated from “cold fusion” to the more careful “low-energy nuclear reactions” (LENR) and “condensed-matter nuclear science,” partly to escape the stigma and partly because proponents no longer claim the mechanism is necessarily d–d fusion.

Frontier What proponents actually claim divides into three separable assertions, and they do not stand or fall together. The first is anomalous excess heat: more thermal energy out than chemical and electrical energy in, sometimes for days. The second is a nuclear signature: helium-4 produced in proportion to that heat, or tritium, or transmuted elements, or charged-particle tracks. The third is that the effect can be a practical energy source. A brief that flags all three the same way has not done its job; the evidence for each is different and each gets its own flag below.

Frontier The single most important physical anchor is the energy accounting. The reaction d + d → ⁴He releases 23.8 MeV, and if excess heat is genuinely nuclear and genuinely this reaction, then the heat and the helium-4 must appear together at that ratio — roughly 2.6 × 10¹¹ helium-4 atoms per second per watt of excess power. This is the field's own most testable prediction, and whether it has ever been met to modern standards is the question the whole brief turns on.

Speculative Proponents also report power densities that, taken at face value, would already be useful. A 2024 review compiles claims of 80–400 watts per kilogram of loaded nickel-hydrogen powder, specific energies of 100–500 eV per deuterium atom, and in one case 1.5 GJ per mole of hydrogen — figures far above any chemical process. The same review concedes in its own text that “low reproducibility of experimental results” remains the field's defining problem, which is the tension this brief exists to hold.

Frontier The label itself smuggles in the disputed part. “Cold fusion” names a mechanism — room-temperature d–d fusion — that the evidence does not support, while “low-energy nuclear reactions” names a phenomenology, anomalous heat and particles, that is still contested; using the two interchangeably, as the press routinely does, quietly asserts the mechanism the data have not earned.

Established The strongest evidence on the record is negative, it is recent, and it was produced by people who wanted the effect to be real. Between 2015 and 2019 Google funded a roughly $10 million, three-laboratory programme after Matthew Trevithick approached the chemist Curtis Berlinguette at the University of British Columbia; a team of around thirty researchers pursued three independent lines of attack and published the summary in Nature in 2019 under the title “Revisiting the cold case of cold fusion.”

Established All three Google lines came back null. The palladium team could exceed the 0.875 deuterium-to-metal loading ratio proponents said was necessary only twice in dozens of attempts, and found along the way that hydrogen enters palladium nanoparticles preferentially at vertices rather than faces — a real materials-science result. The calorimetry team tested 420 nickel-powder samples and saw excess heat in none. A Lawrence Berkeley group led by Thomas Schenkel reproduced the Los Alamos pulsed-plasma-discharge tritium experiment and, over weeks, saw no enhanced tritium. The programme's own conclusion was that it found “no evidence” of anomalous effects that cannot otherwise be explained prosaically — while noting, honestly, that the parameter space is large and “there remains much interesting science to be done.”

Established A second and larger private null landed the week this brief was written. Project Callisto, an independent effort of more than 100 people on a 45,000-square-foot campus, ran from 2015 to 2024 and released its reports on 8 September 2026. It logged over 200,000 calorimeter-hours across more than 500 experiment types, using closed-cell calorimeters with about ±2% steady-state stability and a most-sensitive instrument (a Setaram MS80) with a noise floor near 5 milliwatts. It set decision thresholds in advance — 5% excess heat “potentially real,” 10% “definitive” — and ran matched controls, light-water against heavy-water electrolyte and platinum against palladium cathodes. Across more than 40 attempted replications of Fleischmann–Pons and 60-day steady-state runs, no sample exceeded the margin of error.

Established Callisto also did the unglamorous work of explaining the false positives, which is where the field's history actually lives. Its reports attribute apparent anomalies to input-power measurement error, changes in a cell's heat-transfer coefficient — electrolyte foaming alone could produce up to a 60% overestimate of output — recombination of hydrogen and oxygen, combustion of PEEK plastic components, and instrument drift. A pushed-loading track using gas pressures up to 600,000 psi to force deuterium into palladium likewise reported no anomalous heat. Two well-funded programmes, different methods, same answer.

Frontier The strongest surviving pro-effect result is the heat–helium-4 correlation, and it is genuinely interesting precisely because it is a correlation rather than a bare heat claim. Melvin Miles at the Naval Air Warfare Center, China Lake, reported in the early 1990s that of 33 electrolysis experiments, 18 of 21 showing excess heat also showed helium-4, while 12 of 12 with no excess heat showed none. The measured helium-4 production fell in the range 10¹⁰ to 10¹² atoms per second per watt, with the best-controlled sample close to the theoretical 2.6 × 10¹¹ for d + d → ⁴He + 23.8 MeV. Miles put the odds that the correlation was random at one in 750,000.

Frontier The correlation's weakness is contamination, and Miles was candid about it. Atmospheric helium is 5.2 ppm, and his system's background helium-4 was 4.5 ± 0.5 ppb, attributed to atmospheric helium diffusing through 50 cm of rubber vacuum tubing. The measured signals sat above that background, but a correlation whose absolute levels are near a diffusion-driven background is exactly the kind of result that demands blinded, sealed-system replication — which, in more than thirty years, has not been done to a standard that ended the argument. This is the pivot of the whole subject: a suggestive result nobody has been able to kill and nobody has been able to confirm.

Established Two US Department of Energy reviews, in 1989 and 2004, reached the same verdict fifteen years apart. The 2004 panel of 18 reviewers split roughly evenly on whether the excess-heat evidence was persuasive, but about two-thirds found the evidence for a nuclear origin not conclusive, and only one reviewer was convinced of it. No reviewer recommended a focused federal programme. They did recommend that individual proposals in materials characterization and particle detection be judged on their merits, and warned that “more experiments similar in nature” to the previous fifteen years were “unlikely to advance knowledge.”

Frontier The institutional posture changed in 2023, which is the single most important recent development. ARPA-E, the US energy-research agency, launched a $10 million LENR Exploratory Topic funding eight teams — the University of Michigan (two awards), Amphionic, the Energetics Technology Center, Lawrence Berkeley National Laboratory, MIT, Stanford and Texas Tech — explicitly to establish reference measurements. Director Evelyn Wang framed the goal without advocacy: “does this area show promise, and if so, how? Or can we conclusively show that it does not?” Berkeley's $1.5 million project under Schenkel fires a low-energy ion beam below 500 eV at deuterium-loaded palladium and quantifies any nuclear events with a calibrated diagnostic suite. This is a debunk-or-confirm mandate with a null-result path built in, which is not how the field was funded before.

Frontier Separately, and often confused with cold fusion, NASA Glenn has published measured nuclear reactions in deuterated metals — but it is not an excess-heat claim. In two 2020 Physical Review C papers, Bruce Steinetz, Theresa Benyo, Vladimir Pines and colleagues irradiated ErD₃ and TiD₂ with 2.9 MeV bremsstrahlung, photodissociated some deuterons into neutrons, and observed 2.45 MeV neutrons from D(d,n)³He plus 4–5 MeV neutrons they attribute to electron-screened Oppenheimer–Phillips stripping. The measured rate was about 1.5–1.6 × 10³ neutrons per second — a real, peer-reviewed nuclear signature and a negligible power level. NASA states plainly that “the reaction rates need to be increased substantially to achieve appreciable power levels,” and that the process is driven by an external gamma beam, not by spontaneous room-temperature fusion.

Established The original 1989 replication failures were specific, not vague. Caltech and MIT calorimetry found no excess heat, and several early positive claims were traced to recombination of evolved hydrogen and oxygen inside sealed cells — the same artifact Callisto re-catalogued 35 years later, which is why the null is durable rather than merely old.

Frontier Public funding outside the United States is small and mostly open. The European Union backed the CleanHME and HERMES projects around 2020 at roughly $10 million combined, and MIT's Florian Metzler and George Washington University's David Nagel argue the evidence “has piled up” — a proponent framing this brief reports as a claim, not a finding.

Speculative Gas-loading heat claims are the live commercial frontier and the least openly checkable. Japanese groups at Tohoku University and firms such as Clean Planet report excess heat from hydrogen or deuterium gas cycled over nickel-palladium nanostructures; these underlie the 80–400 W/kg figures, and almost none has a blinded, third-party calorimetric verification.

Frontier The genuinely open question is narrow and it is the one this brief keeps returning to: does anomalous excess heat exist above the bounds the null programmes have now set? Google and Callisto set upper limits at the milliwatt-to-percent level in the systems they tested; they did not test every material, every loading protocol, or every triggering method proponents favour. A null in the tested space does not exclude an effect in an untested corner, and honest sceptics and honest proponents agree on that much. What separates them is a prior about how large that untested corner is.

Frontier Electron screening in a lattice is real, measured, and the strongest physical toehold the field has. Accelerator experiments and the NASA work both show that electrons in a metal reduce the effective Coulomb barrier between deuterons, raising low-energy fusion cross-sections above the bare-nucleus value. The disputed quantity is the screening energy: values of a few hundred eV are measured, but some LENR models require screening one to two orders of magnitude larger to reach observable heat, and no first-principles calculation delivers that. The gap between the screening that is measured and the screening that would be needed is, quantitatively, the frontier.

Speculative The heat–helium-4 correlation is the open experimental question that a decisive test could close in one campaign. If a blinded, sealed-system, multi-laboratory calorimetry programme measured excess heat and helium-4 rising together at 23.8 MeV per atom, the nuclear-origin question would be settled affirmatively; if the same programme found heat with no commensurate helium, the effect would be real but not d–d fusion; if it found neither, the correlation would join the list of contamination artifacts. All three outcomes are informative, which is what makes it the decisive test.

Speculative Transmutation and charged-particle claims are a weaker and messier frontier. US Navy researchers at China Lake reported charged-particle tracks in CR-39 solid-state detectors from palladium–deuterium co-deposition, and various groups report anomalous isotopic shifts. These claims are harder to contaminate-proof than heat and harder to quantify than helium, and none has an independent, blinded replication at the level the heat–helium question now has. They belong on the frontier list as reported observations, not as established phenomena.

Handwave No accepted theory predicts where or when the effect should appear, and this is the deepest problem. Widom–Larsen theory proposes heavy surface electrons capturing to make ultra-low-momentum neutrons; Peter Hagelstein and others propose lattice-phonon coupling that spreads the 23.8 MeV among many nuclei to suppress the expected energetic products. Both are constructed to fit the reported absence of the gamma rays and neutrons that ordinary fusion would produce, rather than to predict a new observation in advance. A theory that explains only why the tell-tale signatures are missing is doing the work of an excuse, and the field's own honest members say a predictive mechanism is the missing piece.

Frontier The absence of the expected radiation is itself the sharpest anomaly. If deuterons were fusing at rates sufficient to make watts of heat, ordinary branching ratios would produce a lethal neutron and gamma flux; experimenters are alive, so if there is heat, the reaction is not ordinary d–d fusion. Sceptics read this as proof there is no nuclear reaction; proponents read it as proof the reaction is novel. The datum is the same; the priors do the arguing, which is the signature of a question that only a decisive experiment resolves.

Speculative Anomalous isotopic and elemental shifts are reported but poorly bounded. Several groups claim transmutation products in spent cathodes; without a measured production rate tied to an energy balance, these are observations in search of a controlled measurement rather than evidence of a new reaction.

Frontier The size of the untested parameter space is itself contested. Google and Callisto tested many but not all materials, loadings and triggers; a proponent's prior that the effect hides in an untested corner and a sceptic's prior that a real effect would have shown up somewhere by now are both defensible, and only more measurement moves them.

Established Reproducibility is the binding bottleneck, and it is not a slogan — it is a specific chain of measurement failures the null programmes documented. Excess heat is inferred by subtracting a large input power from a large output power; when both are of order tens of watts and the claimed anomaly is a few percent, small systematic errors dominate. Callisto's catalogue — input-power error, heat-transfer-coefficient drift, electrolyte foaming worth up to 60% overestimation, hydrogen–oxygen recombination, combustion of cell materials — is a list of ways to manufacture a false positive, each of which has appeared in the literature as a claimed effect.

Established Materials variability defeats the experiment before calorimetry begins. The reported effect, when reported, depends on the specific batch, surface, and impurity profile of the palladium, and proponents themselves say only certain cathodes “work.” Google could reach the required deuterium loading only twice in dozens of tries. An effect that depends on an uncontrolled and unidentified material variable is, operationally, an effect nobody can turn on, and until that variable is named the experiment is not an experiment but a search.

Frontier There is no control variable, which is the methodological heart of the problem. A clean experiment turns the effect on and off by changing one known parameter. The field's best surrogate is the light-water/heavy-water and platinum/palladium substitution that Callisto and Miles both used — and in the well-controlled versions, the substitution turned the claimed heat off, which is evidence against the effect, not for it. A trigger that reliably produced heat on demand would end the debate in an afternoon; its absence is why the debate is in its fourth decade.

Speculative Helium measurement is a specialist bottleneck of its own. Distinguishing a genuine helium-4 signal from atmospheric leakage requires sealed metal systems, mass spectrometry against known backgrounds, and blinding — because atmospheric helium at 5.2 ppm will diffuse through most seals and gloves. Miles's 4.5 ppb background is a warning, not a disqualification, but it means the decisive measurement needs metrology that most electrochemistry labs do not have, a dependency section 6 returns to.

Handwave The theory bottleneck is that no model outputs a number a device engineer could use. Without a predicted power density as a function of loading, temperature and material, there is nothing to optimise toward and no way to distinguish a failed run from a run in the wrong regime. Every other energy technology in this corpus — fission, fusion, photovoltaics — had a quantitative theory before it had a product. LENR would be the first to invert that order, which is not impossible but is a reason for the weak flag.

Frontier Loading instability is a concrete, quantified obstacle. Callisto reached a 95% deuterium-to-palladium ratio only with arsenic additives and watched it decay within hours; if the effect requires sustained high loading, the material will not hold still long enough to test it cleanly, which converts a physics question into a metallurgy one.

Established The scale of the numbers is the underlying trap. Inferring a few percent anomaly by differencing two tens-of-watts power figures means a 1% calibration error can masquerade as the entire claimed effect, which is why milliwatt-stable, closed-cell calorimetry is a precondition and not a refinement.

Frontier Blinding itself is hard here, which is under-appreciated. The experimenter usually knows which cell holds deuterium and which holds ordinary hydrogen, so genuine blinding requires automating both the assignment and the analysis; the field has rarely met that bar, and it is one reason the two funded null programmes carry more weight than the historical literature.

Frontier The subject's dependencies are unusual: it waits on almost nothing that other research is producing, and everything on a result nobody is producing. Fusion waits on materials and tritium; LENR waits on a single reproducible positive that would convert it from a claim into a research programme. That is why its critical dependency is written in section 14 as a scientific result rather than an industrial capacity.

Established Its concrete dependency is metrology, not physics. The decisive measurement needs calorimeters stable to milliwatts over weeks, mass spectrometry for helium against a characterised background, and neutron and charged-particle spectroscopy calibrated to known sources — the toolkit the corpus treats under Metrology Infrastructure. NASA's and Berkeley's programmes succeed as science because they brought that toolkit; the historical failures largely did not.

Speculative A softer dependency is materials characterization at the nanoscale. If the effect is real and material-dependent, identifying which surface, defect or nanostructure hosts it depends on the same electron-microscopy and surface-science tools that materials discovery uses. The 2024 nanomaterials review is an argument that this dependency is now satisfiable; its own admission of low reproducibility is an argument that satisfying it has not yet produced a controllable effect.

Frontier Institutionally it depends on a publication venue that a sceptic will read. The 2019 Nature paper and the 2020 Physical Review C papers matter as much for where they appeared as for what they found; a positive result in a mainstream journal is the dependency proponents themselves name as the field's next required step, and it has not been met for a heat claim.

Frontier It also depends on people, not only instruments. The credible modern results come from nuclear physicists and precision chemists entering the field with mainstream instruments; the dependency is as much the migration of that expertise, which ARPA-E funding now subsidises, as any single measurement.

Established None of these dependencies is exotic. The calorimetry, helium mass spectrometry and neutron spectroscopy all exist in national metrology and nuclear-physics laboratories today, which is why ARPA-E routed its awards to institutions that already hold the instruments rather than to the historical cold-fusion groups.

Frontier The decisive experiment is a pre-registered, blinded, multi-laboratory calorimetry campaign with a stated excess-heat threshold that also measures helium-4 production commensurate with that heat at 23.8 MeV per helium-4 atom. It is decisive because it collapses the field's three separable claims into one measurement: heat with commensurate helium confirms a nuclear origin, heat without helium proves a real but non-fusion effect, and neither confirms the null the two big private programmes have already indicated. Every outcome advances knowledge, which is the property the 2004 DOE panel said the field's repetitive experiments lacked.

Established The calorimetry half of this experiment is already running, which is what makes the moment unusual. Project Callisto's 200,000 calorimeter-hours with pre-declared 5% and 10% thresholds and matched light-water controls is exactly the blinded, thresholded protocol the decisive test requires — and it returned a null. ARPA-E's eight teams are running the reference-measurement version now. What none of them has yet done is the helium closure on a positive, because none has produced the positive.

Frontier The helium-4 half is the piece the field has never executed to a standard that ended the argument. Miles's one-in-750,000 correlation was neither blinded nor sealed against a 4.5 ppb atmospheric background, and it has not been independently repeated with modern sealed-system mass spectrometry. A single well-instrumented laboratory that reproduced the correlation under blinding, or excluded it, would resolve thirty-seven years of dispute; that this has not happened is the strongest evidence that the correlation is fragile.

Frontier The electron-screening cross-section measurement is a second, cleaner experiment that is already producing numbers. Berkeley's sub-500 eV ion-beam bombardment of deuterated palladium, and the NASA bremsstrahlung work, both measure how far a lattice enhances low-energy fusion. These experiments cannot by themselves make an energy source — NASA's measured 1.5–1.6 × 10³ neutrons per second is far below any power relevance — but they quantify the one piece of physics everyone agrees is real, and they set the ceiling on how much a lattice can plausibly help.

Speculative Charged-particle detection with solid-state track detectors is a third experiment, and it is the one most in need of blinding. The Navy CR-39 track claims would be strong evidence if they survived a protocol in which the analyst does not know which detector faced a deuterium cell and which a control. A coordinated multi-lab effort at the Naval Surface Warfare Center, Indian Head, with NIST, the Army Research Laboratory and the US Naval Academy, is positioned to do exactly this kind of cross-checked detection.

Handwave The experiment nobody can specify is the one that would matter most for a product: a demonstration of watts of excess heat sustained for a month, third-party-instrumented, with the cell sealed and the fuel accounted for. Proponents point to individual runs; none has been turned into a reproducible, on-demand demonstration that a sceptical engineer could operate. Until the trigger is identified, this experiment cannot be designed, only hoped for, which is why it carries the weakest flag in the section.

Frontier The natural experiment already running is the funded null programme itself. Each additional milliwatt-sensitive, controlled campaign that reports no excess heat tightens the upper bound and shrinks the untested corner where the effect could hide; two such programmes now agree. If a decade of well-instrumented searching continues to return nulls, the honest reading shifts from “unconfirmed” toward “excluded at the claimed scale,” and the funded, null-publishing structure is what makes that shift legible rather than merely rhetorical.

Speculative A tritium replication is the fourth test, and it is comparatively decisive because tritium is a hard nuclear signature. Google's reproduction of the Los Alamos pulsed-discharge experiment saw no enhanced tritium; a positive, blinded tritium result above the credible background would be difficult to explain prosaically, which is why it belongs on the list even though it, too, has so far returned null.

Speculative There is no engineering discipline here yet because there is no confirmed effect to engineer, so this section describes what building a device would require if the effect were real. The first requirement is a reproducible trigger: a specified material, loading protocol and stimulus that produces heat on demand with a stated success rate. No such specification exists, and everything downstream is conditional on it.

Speculative The second requirement is heat extraction at a useful temperature. Electrolysis cells operate near the boiling point of water; a competitive heat source needs output temperatures high enough to drive a turbine or a useful process, which is why gas-loading and higher-temperature nickel-hydrogen approaches attract the commercial claims. The 80–400 W/kg figures proponents cite, if real and sustainable, would be in the range of a low-grade industrial heat source, not a power plant.

Speculative The third requirement is scaling without losing control of the anomaly. An effect that appears in one cathode in dozens does not scale to a fleet; the manufacturing problem is identical to the reproducibility problem, one step later. If the hosting material variable were identified, this would become an ordinary materials-qualification task of the kind Additive Manufacturing Qualification and the fusion-materials programme in Commercial Fusion already run; without it, there is nothing to qualify.

Frontier One near-term engineering use is real and non-commercial: a compact nuclear-reaction source for space. NASA's interest in lattice-confinement reactions is partly as a potential companion to the fission and radioisotope options in Advanced Nuclear Propulsion, because a deuterated metal packs deuterons about a billion times denser than a magnetically confined plasma. That is an engineering advantage in fuel density only; at present reaction rates it delivers no usable power, and NASA says so.

Handwave Control and safety instrumentation would be the fourth requirement. A heat source whose mechanism is unknown cannot be qualified for uncontrolled release of energy; even granting the effect, the absence of a predictive model means no engineer could certify a runaway cannot occur, which is a licensing barrier before it is a market one.

Established Muon-catalysed fusion is the proof that room-temperature nuclear fusion is physically possible — and the proof of why it is not an energy source. A muon replacing an electron shrinks the deuterium molecule enough for the nuclei to fuse at room temperature; this is measured, uncontested physics. It fails as energy because each muon catalyses only a few hundred fusions before decaying, less than the energy cost of making the muon. It is the honest adjacent case: nuclear reactions in a cold condensed system are real, and the barrier is a specific, quantified accounting problem.

Established Hydrogen-in-metals is a mature materials science that LENR sits inside. Palladium hydride, metal-hydride storage and hydrogen embrittlement are well understood, and Google's vertices-not-faces loading result is a contribution to that field regardless of the fusion question. The corpus treats the storage side in Hydrogen Economies; the point here is that the substrate of the cold-fusion claim is otherwise ordinary, well-characterised matter.

Frontier Accelerator and electron-screening physics is the legitimate research neighbour. Low-energy nuclear cross-sections, screening in plasmas and solids, and the Oppenheimer–Phillips stripping NASA invokes are all mainstream nuclear physics; the ARPA-E and NASA programmes are, in effect, importing that community's instruments into the LENR question. This is the adjacency most likely to yield a durable result, positive or negative.

Handwave Sonofusion is the cautionary adjacent claim. The 2002 report of fusion in cavitating deuterated acetone (“bubble fusion”) briefly looked like a cousin of cold fusion and was retracted amid a misconduct finding. It is adjacent as a methodological warning: an extraordinary nuclear claim in a table-top system, unreplicated, is the pattern Pathological Science anatomises, and the LENR field is judged against it fairly or not.

Speculative Commercial fusion and aneutronic fusion are the energy-system neighbours a reader will want placed. LENR and the approaches in Commercial Fusion and Aneutronic Fusion share only a fuel and an ambition; the physics, the evidence base and the funding scale are different by orders of magnitude, and conflating them flatters LENR and insults fusion. They are adjacent on the map and not on the bench.

Frontier Superconductivity is a methodological neighbour worth naming. Like the claims in High Temperature Superconductors, LENR is a solid-state anomaly whose credibility rises and falls on reproducible, instrumented measurement rather than on theory; the difference is that high-temperature superconductivity delivered a reproducible signal within months and LENR has not in decades.

Established Pyroelectric-crystal fusion is the other honest adjacent demonstration. A heated tantalate crystal can field-ionise and accelerate deuterons enough to fuse and emit 2.45 MeV neutrons — a genuine table-top nuclear reaction demonstrated in 2005 — and, like muon catalysis, it consumes far more energy than it releases, which is precisely the accounting a practical LENR device would have to beat.

Established The field's defining institutional fact is exclusion. After 1989, mainstream journals largely stopped reviewing cold-fusion papers, funding agencies declined focused programmes in both 1989 and 2004, and the US Patent Office resisted patents premised on the effect. Proponents built parallel institutions in response: the International Conference on Condensed Matter Nuclear Science (ICCF), the Journal of Condensed Matter Nuclear Science, and the LENR-CANR document archive. These preserved the work and also insulated it from the outside criticism that improves it.

Frontier The 2023 ARPA-E programme is the most consequential institutional change in thirty years, because it re-attaches the question to mainstream review. By funding national labs and research universities to make reference measurements with a stated null-result path, ARPA-E converts a stigmatised topic into a fundable one without endorsing its claims — the “prove or debunk” framing is the mechanism. Whether the eight teams find an effect matters less institutionally than that their negatives will now be publishable and their positives, if any, credible.

Established The privately funded null programmes are a second institutional innovation. Google and Project Callisto demonstrate a model this corpus should notice: a well-resourced, methods-first search whose declared product is a bound and a published protocol, not a breakthrough. A field that has been unable to police its own false positives benefits more from a credible, funded null than from another unreplicated positive, and the honestly-flagged assessment is the point of the exercise.

Frontier Internationally, the topic has small public programmes and larger commercial secrecy. The European Union funded the CleanHME and HERMES projects around 2020 at roughly $10 million combined; Japanese work at Tohoku University and companies such as Clean Planet and Mitsubishi Heavy Industries has reported gas-loading heat, largely outside blinded independent verification. Commercial secrecy is rational for a firm and corrosive for a science, and the field's credibility gap is partly a governance gap: the strongest claims are the least openly checkable.

Established Patent and publication exclusion had teeth. The US Patent Office's reluctance to grant cold-fusion patents and mainstream journals' de facto refusal to review the papers are documented features of the field's history, and they are the institutional reason the work accumulated in a parallel literature that outside referees never stress-tested.

Frontier The timing is itself a datum. Callisto's pre-registered reports arriving in September 2026, months into the ARPA-E awards, mean the field now has two independent, methods-first nulls on the public record within one research generation — a density of credible negative evidence it never had in its first thirty years.

Frontier The clearest ethical hazard is investor and public harm from overclaiming. The Andrea Rossi “E-Cat” episode — a claimed megawatt LENR reactor that led to litigation and no verified device — is the field's standing example of how an unfalsifiable heat claim becomes a vehicle for fraud. Honestly-flagged assessment is itself an ethical instrument here: labelling a claim speculative protects the people a vendor's superlative would otherwise reach.

Speculative The opportunity-cost question cuts both ways and deserves a number. The entire LENR field has cost a few tens of millions of dollars over a decade against more than $1 billion a year for mainstream fusion; on that ratio, a funded search for a low-probability, high-payoff effect is a defensible portfolio bet, not a diversion. The ethical failure would be either to forbid the search or to sell its unproven output as a solution to the energy transition.

Frontier There is a subtler ethic about how science treats anomalies and the people who chase them. Fleischmann and Pons were serious electrochemists whose careers were destroyed; the field's exclusion was partly earned by their premature press conference and partly a social punishment that outran the evidence. A mature research culture can hold both that the claim is probably wrong and that suppressing its careful investigation is its own error — which is precisely the posture the ARPA-E and private null programmes now model.

Established Dual-use and safety concerns are minimal, and saying so is part of an honest assessment. The reported effects, if real, are low-power and produce little penetrating radiation; there is no weaponisation pathway and no meltdown mode. The risks here are epistemic and financial, not physical, which distinguishes LENR from most of the energy systems in this category.

Speculative Honest labelling has a specific beneficiary. The reader most exposed to harm is the non-specialist investor or policymaker who cannot themselves distinguish a sealed, blinded calorimetric result from a vendor's press release; the flag system is, in this subject, a consumer-protection instrument as much as an epistemic one.

Frontier Media amplification is part of the hazard. A single unreplicated positive attracts more coverage than a decade of careful nulls, and that incentive gradient — toward the exciting claim and away from the boring exclusion — is exactly what an honestly-flagged brief exists to correct.

Speculative If the effect were real at the claimed power density, the civilizational consequence would be distributed, fuel-abundant heat — and even that is smaller than the rhetoric. A bench-top source at 80–400 W/kg is industrial process heat and off-grid power, not a replacement for the grid; the deep decarbonisation case still runs through the electricity system this corpus treats under Planetary Scale Energy Systems. The honest upside is a new distributed heat option, conditional on a measurement nobody has made.

Frontier The more likely civilizational value is methodological, and it is being delivered now. LENR is turning into a live demonstration of how a scientific culture should handle a stigmatised anomaly: fund a credible, blinded, null-publishing search, let the bound tighten, and change the assessment as the evidence moves. That template — a product of honestly-flagged assessment rather than of any nuclear effect — is transferable to every fringe-adjacent claim, and is arguably the subject's real contribution.

Speculative The asymmetry of outcomes is the civilizational reason to keep watching. The cost of the search is a rounding error against the energy budget; the payoff, if the improbable holds, is large; and the null, if it holds, retires a thirty-seven-year controversy and frees attention. Few open questions have a payoff matrix this favourable to simply doing the decisive experiment well, which is the argument for finishing it rather than either believing it or dismissing it.

Handwave Claims that a confirmed effect would end the geopolitics of energy are unbounded. Even a real, distributed heat source would face the manufacturing, control and grid-integration constraints every energy technology in this category faces; the leap from a bench result to a civilizational reordering is assertion, and belongs flagged as such.

These horizons track the funded measurement programmes and the decisive heat–helium closure, not any product roadmap:

• 10 yr: Frontier The ARPA-E teams report reference measurements and the private nulls are digested; expect the upper bound on excess heat to tighten further. Speculative A single blinded, sealed-system replication of the heat–helium-4 correlation is executed by someone — the one experiment that could flip the assessment either way. Handwave A reproducible, on-demand watt-scale demonstration a sceptic can operate; nothing in the record makes this likely, and it is listed to be honest about what a product would require.
• 25 yr: Speculative Either the decisive experiment has confirmed a nuclear-origin effect and LENR becomes a small materials-and-nuclear research programme, or a decade of funded nulls has shifted the honest reading to “excluded at the claimed scale.” Frontier Electron-screening physics yields durable cross-section results regardless of the excess-heat verdict.
• 50 yr: Handwave Any claim that LENR is contributing to primary energy at this horizon presupposes the reproducible trigger that does not exist; it is not extrapolation from any measurement now on the record. Speculative The more probable 2070s outcome is that LENR is a settled entry in the history of anomalies, cited as a case study in Scientific Revolutions.
• 100 / 250+ yr: Handwave Beyond useful forecasting. The one durable observation is structural: a room-temperature nuclear heat source, if it were ever real, would be so consequential that its absence from the record after this long is itself strong evidence about its likelihood.

• Depends on Nothing on this map produces a result low-energy nuclear reactions waits on. The binding constraint is not a missing technology from another brief but a missing measurement: a reproducible positive that has never been produced. Honestly, no typed depends-on edge to another brief is claimed.
• Requires (not on this map) First, the scientific result the whole subject turns on: excess heat measured together with helium-4 at 23.8 MeV per atom, reproducibly and under blinding — the one datum no programme has produced. Second, a first-principles mechanism that predicts where and how strongly a lattice lets deuterons penetrate the Coulomb barrier, rather than a theory built after the fact to explain missing radiation. Third, the institutional instrument that would make either credible: a pre-registered, blinded, multi-laboratory calorimetry protocol with declared thresholds, which Callisto and ARPA-E have begun but not closed on a positive. Fourth, the industrial metrology — calorimeters stable to milliwatts over weeks with sealed helium mass spectrometry — without which a positive cannot be trusted. Fifth, and only if all of the above turn positive, a market: there is at present no buyer because there is no verified output to sell. None of the five is shortened by capital, and the first is not shortened by anything except doing the experiment.
• Enables If the first requirement were ever met, LENR would enable a distributed heat option relevant to Hydrogen Economies and to off-grid power, and a fuel-dense reaction source of interest to Advanced Nuclear Propulsion. No typed enabling edge is claimed, because an enabling relation conditional on an unconfirmed effect is a hope, not an edge.
• Adjacent Electron-screening and low-energy nuclear physics; hydrogen-in-metals materials science; precision calorimetry under Metrology Infrastructure; and on this map Commercial Fusion, Aneutronic Fusion, Advanced Fission and, as method rather than physics, Pathological Science.

Established “Cold fusion was debunked in 1990, and that is the end of it.” The 1989–1990 failures to replicate were real and decisive against the original strong claim, but they did not prove a universal negative, and serious measurement has continued. The honest statement is narrower and stronger: no one has demonstrated a reproducible, on-demand excess-heat effect, and two well-funded modern searches found none — which is a bound, not a metaphysical impossibility.

Speculative “It has been proven and is being suppressed.” This is the opposite error and the more dangerous one for a reader. The heat–helium correlation is suggestive and unresolved, not proven; the reason it is unresolved is that no one has replicated it under blinding, not that a cabal forbids it. ARPA-E is now paying national laboratories to look. A result that survived a blinded multi-lab test would not be suppressible, and its continued absence is evidence about the claim, not about the institutions.

Frontier “NASA proved cold fusion works.” NASA measured real nuclear reactions in deuterated metals — about 1.5 × 10³ neutrons per second under a 2.9 MeV gamma beam — and explicitly framed the work as not cold fusion: it uses an external beam, energetic deuterons and electron screening, and produces no net energy. Citing lattice-confinement fusion as vindication of Fleischmann–Pons reverses what NASA actually said.

Handwave “The E-Cat and similar devices are commercial LENR reactors.” The most publicised commercial claim, Andrea Rossi's E-Cat, produced litigation and no independently verified device; commercial gas-loading claims from several firms remain largely outside blinded verification. A vendor's power figure with no sealed, third-party measurement is a marketing number, and the corpus flags it accordingly.

Speculative “The helium correlation settles the nuclear question.” Miles's one-in-750,000 statistic is a real result and the field's best, but it was measured against a 4.5 ppb atmospheric-helium background in unblinded conditions and has not been independently reproduced to modern standards. It is the reason to do the decisive experiment, not a substitute for having done it.

Established “Excess heat, if measured, must be nuclear.” The null programmes catalogued a specific set of non-nuclear ways to measure heat that is not there — foaming worth up to 60% overestimation, recombination, combustion of cell components, and power-accounting error. Most historical positive claims are consistent with one of these, which is why the field's own careful members insist on calorimetry that closes every prosaic channel before invoking a new force of nature.

Frontier “A funded programme means the establishment now believes in cold fusion.” ARPA-E's mandate is explicitly agnostic — “show promise, or conclusively show that it does not” — and the two large private programmes were built to look for the effect and reported not finding it. Funding a rigorous search is the opposite of endorsement; it is the mechanism by which a stigmatised claim is finally made to answer for itself.