Cold Fusion & LENR: From Nanogaps to Megawatts?

October 15, 2026

Cold fusion didn’t end in 1989, but can it scale from lab experiments to megawatts? Former Los Alamos nuclear chemist Dr. Edmund Storms believes the answer depends less on building a bigger reactor than on engineering the microscopic structures inside its active material. In our October 12 interview, he connects decades of reported nuclear effects to tiny gaps in solids where, he argues, unusual electron behavior could enable fusion. His proposed path runs from milliwatt measurements and carefully prepared powders to gas-fed generators and modular power systems. The crucial link is still unresolved: can researchers reliably manufacture enough of those active sites to turn a laboratory effect into controllable, useful power?

From Los Alamos to the Cold Fusion Controversy

For Storms, the story begins with excitement rather than rejection. When Fleischmann and Pons announced cold fusion in 1989, he was working at Los Alamos National Laboratory, where the prospect of a compact nuclear energy source immediately connected with existing problems. Among them was providing power for use in space. He remembers colleagues redirecting their attention toward the new claim, attracted by the possibility that nuclear energy might be produced in an apparatus radically different from the systems they already understood.

His background helps explain that response. Storms earned a chemistry degree from Penn State in 1953, followed by a master’s and doctorate in radiochemistry from Washington University in St. Louis, completing his Ph.D. in 1958. Over roughly three decades at Los Alamos, he worked on high-temperature chemistry, refractory materials, carbides, nuclear fuels, and materials relevant to nuclear power and space propulsion. His 1967 monograph, The Refractory Carbides, belonged to that earlier scientific career. Cold fusion didn’t introduce him to nuclear materials; it presented a new problem in materials he was already equipped to investigate.

That combination of disciplines remains central to his thinking. Storms argues that LENR—low-energy nuclear reactions—can’t be understood by considering nuclear physics while treating the surrounding material as incidental. Composition, preparation, hydrogen uptake, and microscopic structure all enter his explanation. His criticism of conventional fusion is correspondingly forceful: he questions the cost, complexity, and maintenance implications of large, high-temperature systems. But the productive comparison in his account isn’t simply big machines versus small ones. It’s whether an unusual material environment could provide a fundamentally different route to the reaction.

Thirty-seven years after the announcement, Storms still believes the initial controversy has consequences. He attributes some resistance to scientific stigma and some to the economic disruption he thinks successful cold fusion would cause. Those are his interpretations of the field’s history, not measurements of its underlying mechanism. The engineering question stands apart from them. Whatever a researcher thinks about the controversy, a reproducible material and a transferable experiment would offer something more useful than another argument about what happened in 1989.

The Evidence: A Pattern, Not a Single Experiment

When I asked which experiment he considered the strongest evidence for LENR, Storms resisted choosing a single dramatic result. He looks instead for “a pattern of events,” assembled from observations made across decades. Helium measurements in deuterium experiments occupy a particularly important place in that pattern, alongside reports of excess heat, tritium, and transmutation. He regards their relationships as evidence of a nuclear process rather than a collection of unrelated anomalies. His confidence rests on the combined behavior, not simply on one sample becoming unexpectedly hot.

Our discussion of CR-39 detectors illustrates both the appeal and the difficulty of interpreting that evidence. Storms describes a plastic in which energetic particles can damage the material locally. Subsequent treatment with concentrated sodium hydroxide etches the affected regions into pits that can be examined. The resulting tracks preserve information about interactions that would otherwise be invisible. But he emphasizes that interpreting their shapes and origins is the hard part. A detector response doesn’t, by itself, identify the entire reaction sequence that produced it.

In his current explanation, some reported neutrons arise from a secondary process rather than directly from the initial fusion event. He proposes an energetic hydrogen-4 intermediate that interacts with palladium or another nearby element, triggering transmutation and neutron emission. That connects the detector discussion to the broader pattern he sees in the literature. It also separates three questions that are easy to blur together: what a detector recorded, which particle caused the response, and what nuclear mechanism produced that particle.

The interview doesn’t present a single experiment measuring every claimed signature simultaneously, nor does it supply the calibration records, product concentrations, backgrounds, and run statistics needed to independently assess each result. It presents Storms’s synthesis of the evidence and the explanation he believes connects it. That distinction matters without diminishing the question he’s trying to answer. Why should differently prepared samples, investigated through different methods, show behaviors he considers related? His answer begins with a feature that may be present in some pieces of material and absent from others.

Nanogaps and the Nuclear Active Environment

Storms calls that feature the nuclear active environment, or NAE. It’s the special local structure that, in his model, distinguishes material capable of supporting cold fusion from ordinary material that cannot. Simply choosing a particular metal wouldn’t be enough. Two nominally similar samples could behave differently because only one contains the required environments, or because their concentrations differ. This makes sample preparation much more than a preliminary step. It becomes part of the proposed mechanism itself.

The central physical obstacle is the Coulomb barrier: the repulsion between positively charged nuclei. Storms proposes that electrons in a suitable microscopic gap could reduce that repulsion sufficiently for a nuclear interaction to occur. He isn’t describing a miniature version of a high-temperature plasma. He’s proposing that the surrounding material creates an electronic environment different enough to change what the nuclei can do. The gap matters because of the conditions he believes it permits, not because empty space alone supplies an explanation.

In his account, the ordinary crystal environment constrains electron behavior. Small gaps or voids interrupt that environment, potentially allowing electrons to participate differently. He describes the relevant spaces as larger than an ordinary vacancy but still extremely small. The interview doesn’t establish an operating gap width, an optimum electron concentration, or a fabrication tolerance. Those aren’t minor missing dimensions on an otherwise finished drawing: they’re among the properties researchers would need to determine to turn his proposed environment into an engineered material.

The manufacturing implication is substantial. More metal doesn’t necessarily mean more useful activity; what matters in his model is how much of the right structure the metal contains. A large sample with few suitable sites could be less informative than a small sample containing many. The next step would therefore be to connect a measurable structural feature with reproducible behavior. Researchers would need to recognize the relevant environment, distinguish it from inactive structures, and learn how to create it deliberately rather than depend on a fortunate preparation.

Electrons, Momentum, and the Hydrogen-4 Puzzle

Storms’s explanation has changed over time. His earlier nanogap-based work included the hydroton concept, while our conversation places greater emphasis on electron emission and hydrogen-4. He says observations of electrical current emerging from active material helped drive that revision. After repeating electron-emission work, he interpreted the current as originating in a nuclear process rather than ordinary thermionic emission or a work-function effect. That is his interpretation of the measurements; the interview doesn’t supply the complete experimental record needed to establish their origin independently.

The proposed role of the electrons addresses another issue he considers fundamental: conservation of momentum. Storms argues that an explanation ending with helium alone leaves an accounting problem, and he assigns roles to emitted electrons and a hydrogen-4 intermediate in resolving it. In the sequence he describes, the electrons carry energy in one direction while hydrogen-4 moves in the other. The point is to connect the electrical signal with the nuclear explanation, not merely add an unrelated secondary observation to the list of anomalies.

Hydrogen-4 is also where his account explicitly departs from the nuclear-data interpretation raised in our discussion. He describes earlier experiments involving deuterium ions with energies around 50 MeV—50 million electron volts—and a conventional interpretation involving an extraordinarily short lifetime, on the order of 10⁻²² seconds. Storms disputes whether the observed neutrons establish the intermediate and decay process attributed to them. He argues that a different inferred mass and behavior would better fit the cold-fusion and transmutation observations he is trying to explain. These are the experimental interpretations he challenges, not new nuclear properties established by this interview.

That disagreement has practical consequences for his generator concept. The identity, energy, and subsequent behavior of reaction products would affect what leaves the active region and what happens to surrounding material. His explanation therefore calls for more than an excess-heat measurement: it calls for a convincing connection between heat, electron emission, and the proposed intermediate. The interview doesn’t resolve that connection with a complete reaction equation or a direct hydrogen-4 measurement. It identifies a specific part of the model that further experiments would need to test.

What Counts as LENR—and What Controls Its Power?

The original Fleischmann–Pons electrolytic cell is only one experimental format discussed in the interview. Our conversation also covers gas-loaded powders, plasma-assisted experiments, thin films, solid-state approaches, and the boundary with lattice-confinement fusion. Storms’s organizing principle is that the apparatus used to introduce hydrogen isn’t necessarily the apparatus feature that determines whether his proposed reaction occurs. That doesn’t establish that every experiment bearing one of these labels shares a mechanism. It explains how he sorts a field whose techniques have diversified considerably.

In his model, electrolysis and gas discharge can be different routes for introducing hydrogen into a solid. He emphasizes that a hydrogen molecule outside the material must be dissociated before the relevant species can enter and participate in the process he describes. Once inside, the hydrogen responds to the material’s environment rather than retaining a memory of its delivery method. The important question shifts from whether the experiment uses a liquid cell or a gas system to whether the resulting material contains the required active sites.

Transport then becomes central. Storms describes hydrogen moving through the material, sometimes under a concentration gradient, until it encounters a suitable nuclear active environment. He argues that the delay lies in finding the site rather than in a slow fusion step after arrival. Translated into engineering terms, his account contains two separate variables: the number of suitable sites and the rate at which hydrogen reaches them. Increasing either could increase power in his model, but increasing one doesn’t necessarily compensate for an inadequate supply of the other.

That distinction also shapes his interpretation of high deuterium loading. He accepts the reported behavior of earlier electrochemical studies while challenging the explanation assigned to it. His emphasis is on how increasing deuterium content changes the activation energy for diffusion, making it easier for deuterium to reach active environments. When I compared this picture with catalysis, he cautioned that borrowing the word doesn’t necessarily add insight. The useful question isn’t which familiar label fits. It’s which measurable property of loading, transport, or structure predicts the observed power.

Why Five Milliwatts Matter

One of the interview’s most concrete technical details is surprisingly modest: 5 milliwatts. Storms says he designed and built a calorimeter accurate to 0.005 watts, allowing him to investigate signals that other researchers would dismiss as uninformative. He contrasts this with instruments whose users needed roughly 100 milliwatts, or 0.1 watts, before they could rely on a result. The two quoted levels differ by a factor of 20, although his account isn’t a standardized comparison of complete instrument uncertainty budgets.

The significance, in his view, is that material regarded as inactive in one apparatus could still be studied in another. Small effects would provide information about preparation, loading, and operating conditions before a team attempted a high-power device. He argues that low-power activity is easier to obtain than high-power operation, making small samples a more productive place to investigate the controlling variables. His proposed development strategy is to learn from weak signals rather than require every early experiment to produce an immediately impressive energy output.

Five milliwatts is his stated measurement accuracy, not the output of a particular successful reactor. It isn’t a gain ratio, a duration of operation, or evidence that an apparatus powers itself. For a transferable experiment, those distinctions would have to remain visible in the data: what went into the system, what came out, how long the behavior continued, and how confidently the difference could be measured. In the engineering argument emerging from the interview, measurement capability comes before scale because it determines which material changes can be evaluated at all.

Computing and AI enter our conversation as possible ways to help a small research community make better use of limited resources. Storms responds positively to the broader opportunity, but the substance of his answer returns to materials and measurement. He doesn’t describe an AI system that has solved the problem. Instead, he emphasizes that enlarging an experiment without understanding what needs to change can miss the point entirely. “It’s what’s in the material that matters,” he says. Better tools become useful when they help identify and control that material’s relevant properties.

Designing a Gas-Fed LENR Generator

When the discussion turns to practical power, Storms favors a gas-fed solid rather than treating the original electrolytic cell as the necessary commercial architecture. His operational sketch involves placing active material in a chamber, removing the air, introducing a mixture of deuterium and helium, and heating until the proposed operation becomes self-supporting. He expresses confidence that he could design a generator around this approach. What he describes, however, is a proposed system architecture, not a machine demonstrated during the interview.

He identifies temperature and electrical current as potential controls. Temperature changes influence hydrogen transport, while current through the material could move hydrogen by electromigration in response to the associated voltage gradients. In his explanation, those changes would influence how quickly hydrogen reaches the active environments and therefore how much power they produce. This is a control principle, not yet a detailed specification: the conversation doesn’t establish an operating temperature, gas pressure, current density, or measured response time for the proposed generator.

The interaction between those controls raises an important engineering question. In a self-heating concept, temperature would be both an operating condition and a consequence of power production. Likewise, electrical current intended to influence hydrogen movement would itself be an input to the apparatus. Evaluating the proposed control scheme would therefore require separating its influence on the reaction from the energy it supplies. Stable operation would need to be measured through changes in operating conditions, rather than inferred simply from the existence of a control knob.

Storms also favors deuterium over light hydrogen because of the reaction products he anticipates, especially his concern about tritium accumulation in light-hydrogen systems. That preference follows from his model; it isn’t a demonstrated safety assessment of a finished device. Separately, he suggests that electron-emission current might supply the control circuits while the reaction continues. He describes the available current as limited, potentially useful for controls rather than a source of megawatts of direct electricity. Producing heat, regulating it, powering the controls, and delivering useful electrical output remain distinct engineering tasks.

The Next Decisive Experiments

Asked where substantial new funding should go, Storms proposes two routes aimed at the active material. The first is sintering fine powders. In his description, neighboring particles recrystallize as the material consolidates, but mismatched crystal structures can leave small gaps where they meet. He argues that particle size influences the gaps that form, making it a preparation variable worth investigating systematically. The objective wouldn’t initially be the largest possible heat output. It would be to determine which material preparations reliably produce the behavior his model predicts.

He describes a practical comparison: characterize fine particles, separate them into size groups, prepare materials from those groups, and test them under similar conditions. Chemical precipitation, he says, can produce fine black powders, but doesn’t automatically deliver the ideal particle size. He lacked the funding to pursue the investigation he wanted. The question is therefore more specific than whether powders can work. It’s whether a reproducible relationship can be established between a controlled particle-size distribution, the resulting structure, and measured heat production.

His second route is deliberate nanofabrication, drawing on techniques developed for the computer industry. Rather than depend on gaps emerging from the assembly of particles, he wants researchers to explore structures whose geometry and concentration are controlled more directly. The two approaches would attack the same proposed variable from different directions: one through material formation, the other through designed features. Storms believes relevant expertise already exists outside the LENR community, although he acknowledges the expense, learning process, and trial and error involved in bringing those capabilities into this research.

Neither proposal comes with an established optimum gap width, guaranteed material yield, or itemized research budget. The tens of millions of dollars raised in our exchange were part of my hypothetical funding question, not a costed program he presented. What makes the proposals useful is their specificity. They point toward comparisons in which preparation and structure could be connected with measured behavior. The decisive outcome would be a relationship another laboratory could reproduce—or a result that meaningfully narrows which structures and interpretations remain plausible.

Can LENR Really Scale to Megawatts?

Storms’s strongest qualification appears when I ask about scale. He connects output to the amount of material and its concentration of active environments, but acknowledges that researchers don’t know how high that concentration can be made. Consequently, he can’t specify how much material a generator would require. This is the missing engineering relationship between nanogaps and megawatts. Without reproducible output per unit of material, the proposed power level can’t be translated into a defined quantity of active material, much less a finished generator design.

His vision is nevertheless concrete at the architectural level. He imagines separate generators producing perhaps a megawatt each, potentially assembled into an array of a couple thousand units. Taking his example as an illustration, 2,000 one-megawatt modules would total 2 gigawatts. Shutting down one module would remove 0.05% of the nominal capacity, assuming the others continued unaffected. That arithmetic explains his interest in modularity. It is not measured plant performance, evidence that a one-megawatt unit exists, or a forecast of what a commercial installation would deliver.

The discussion also leaves the distinction between heat and usable electricity unresolved. It doesn’t establish conversion efficiency, auxiliary consumption, duty cycle, or tested power density, so the hypothetical array can’t simply be described as a two-gigawatt electrical station. Storms anticipates that material might eventually lose activity through a buildup of transmutation products and proposes replacing it. That introduces another specification an engineering program would need: service life. A sample’s output, stability, replacement interval, and resulting material inventory would all influence whether modular operation is practical.

For Storms, the larger attraction is generation close to where energy is used. Individual units could serve businesses, while arrays could combine their output without depending on a single source. Our discussion turns to wind farms and solar installations as examples of distributed components, and he describes his own experience with rooftop solar in Santa Fe. A dependable LENR source, in his vision, could address the periods when solar generation isn’t available. But that advantage remains conditional on the underlying unit delivering repeatable, controllable power—not merely on arranging many proposed units together.

The Knowledge That Has to Survive

The interview’s final turn is from equipment to people. Fleischmann and Pons brought cold fusion into public view; Eugene Mallove helped sustain advocacy and community around a field that faced enduring resistance. I remembered Mallove as someone who encouraged my own work and whom I’d spoken with many times. Storms sees these contributions as part of a history whose significance extends beyond the laboratory. Whatever conclusions future researchers reach, the experimental record and the people who preserved it are inseparable from the opportunity to investigate it again.

Storms is also clear about his present circumstances. At the time of our conversation, he says he is 95 and a half and no longer has access to experimental facilities. He still has his accumulated knowledge and database, and credits Jed Rothwell with making that information accessible and searchable online. His role now is to explain what he believes he has learned and encourage others to investigate it. “All I can do now is be an advocate for what I’ve discovered,” he says, hoping people will listen rather than reject the work outright.

That gives the technical discussion an immediate purpose. A successor doesn’t inherit only a proposed nuclear mechanism. There are questions about particle preparation, hydrogen loading, transport, calorimeter performance, and the difference between a weak response and an inactive sample. Preserving the record matters because these details determine what another experimenter can actually attempt. The useful inheritance is not a demand for agreement, but enough information to reconstruct the work, assess the interpretations, and decide which next experiment would reveal something new.

Cold fusion’s future, in this account, isn’t settled by either the controversy of 1989 or the confidence of one researcher. Storms offers a materials-centered explanation, two proposed experimental routes, and a modular energy vision whose central uncertainty is explicit. The question is whether the active structure he describes can be manufactured and its behavior reproduced. The road from nanogaps to megawatts doesn’t begin with a bigger reactor. In Storms’s proposed approach, it begins with a material another laboratory can make—and a result that laboratory can measure for itself.