In July 1996, Italian physicist Giuliano Preparata reported more than 300 watts of apparent excess heat from a cold-fusion cell containing a palladium wire just 50 micrometers thick. The reported output approached 60 kilowatts per cubic centimeter—roughly 180 times the average power density of conventional nuclear fuel. Although the extraordinary result remains independently unverified, it raised a provocative question: Could cold fusion produce heat at power densities exceeding those of nuclear fission?

The Cold Fusion Experiment Behind the Claim

Preparata’s experimental work appeared in the Journal of Electroanalytical Chemistry in 1996. Working with Marco Scorletti and Marco Verpelli, he investigated modified Fleischmann–Pons electrochemical cells designed to absorb unusually high concentrations of deuterium into palladium. The researchers used heavy water containing lithium deuteroxide, thin palladium cathodes, platinum anodes, and a temperature-controlled calorimeter. Their objective was to determine whether electrochemical loading could produce thermal energy beyond the electrical and chemical energy accounted for in the experiment. The journal paper described two palladium runs reporting approximately 25 and 30 watts of excess heat.

Those results became more striking when normalized to the palladium’s volume. The first run used a reported palladium volume of 0.00815 cubic centimeters and produced approximately 25 watts of excess heat, corresponding to 3.07 kilowatts per cubic centimeter. The second used 0.00835 cubic centimeters and produced approximately 30 watts, corresponding to 3.59 kilowatts per cubic centimeter. The authors reported power densities of at least 3 and 3.6 kilowatts per cubic centimeter, respectively. They also stated that other experiments had produced power densities 10 to 20 times larger, although those results weren’t fully documented in the journal article.

The higher-power experiments appeared in Preparata’s presentation at the Sixth International Conference on Cold Fusion, held in Japan in October 1996. His paper, “Everything You Always Wanted to Know About Cold Fusion Calorimetry,” described a subsequent experimental campaign conducted that July. A graph identified as Experiment 85, dated July 7, showed apparent excess power rising above 300 watts during a measurement record extending to approximately 16,000 seconds. A platinum control experiment, Experiment 86, dated July 8, showed excess power remaining close to zero. Preparata concluded that his group had obtained several hundred watts of excess heat and volumetric power densities of 50 to 100 kilowatts per cubic centimeter.

The distinction between the two publications matters. The peer-reviewed journal paper provided explicit numerical cathode volumes and reported excess powers in the tens of watts. The conference paper extended the claim into the hundreds-of-watts range, but its most dramatic results were presented primarily through graphs rather than a complete numerical record of electrical inputs, thermal outputs, and uncertainties. The conference experiment therefore supplied the basis for the famous fission-like comparison without providing the same level of quantitative documentation needed for a fully independent reconstruction.

The Engineering Significance of a Hair-Thin Wire

The remarkable power-density figures depended on the palladium cathode’s exceptionally small volume. Preparata’s apparatus descriptions identified a wire approximately 50 micrometers in diameter and 250 centimeters long, wound into a helix around quartz supports. Its diameter was only 0.05 millimeters, considerably smaller than a typical human hair. Despite extending 2.5 meters in length, the cylindrical wire contained approximately 0.00491 cubic centimeters of palladium. With palladium’s density near 12 grams per cubic centimeter, the metal weighed only about 59 milligrams.

That tiny volume transformed hundreds of watts into an enormous volumetric figure. An apparent excess of 300 watts distributed through 0.00491 cubic centimeters corresponds to approximately 61 kilowatts per cubic centimeter. Preparata’s stated range of 50 to 100 kilowatts per cubic centimeter would correspond to roughly 245 to 491 watts of excess heat using the same nominal geometry. This explains why later summaries frequently cite approximately 60 kilowatts per cubic centimeter. The figure is a calculation based on inferred thermal power and the assumed active metal volume, not an independent measurement of nuclear energy generation.

The geometry also introduces a complication. The earlier journal paper assigned palladium volumes of 0.00815 and 0.00835 cubic centimeters to its two principal runs, substantially larger than the 0.00491 cubic centimeters calculated from the nominal 50-micrometer, 250-centimeter wire. At the same diameter, those larger volumes would require approximately 4.15 and 4.25 meters of palladium wire. Different experiments may have used different cathode lengths, but the published descriptions don’t fully reconcile the figures. A precise engineering reconstruction must therefore preserve the published volumes for the earlier runs rather than assume that every experiment used identical geometry.

The wire’s surface area provides another important perspective. A 250-centimeter-long wire with a 50-micrometer diameter has approximately 3.93 square centimeters of lateral surface area. If 300 watts of heat were generated within the palladium and transferred through that surface, the average heat flux would be approximately 76 watts per square centimeter, or 0.76 megawatts per square meter. That’s a substantial thermal load, but it’s within the general engineering range encountered in boiling and high-performance liquid cooling. The volumetric power density was extraordinary largely because the heat-generating material was so thin; the surface heat flux was far less exceptional.

Comparing Cold Fusion with Nuclear Fission Fuel

The expression “fission-like power density” refers to thermal output per unit volume, not to nuclear fission occurring inside palladium. Preparata’s 1996 conference paper described power densities reaching what he called “incredible values” of 50 to 100 kilowatts per cubic centimeter. The explicit comparison with conventional reactor fuel became prominent in subsequent technical summaries, particularly those by physicist David Nagel and science writer Steven Krivit. Their discussions placed Preparata’s experimental claim alongside established nuclear-reactor engineering benchmarks.

A representative pressurized-water reactor described in International Atomic Energy Agency engineering literature has an average linear heat-generation rate of approximately 17.5 kilowatts per meter of fuel rod. Its uranium dioxide fuel pellets are approximately 8.19 millimeters in diameter. Using the fuel material itself as the normalization volume produces an average thermal power density of approximately 332 watts per cubic centimeter. Against that benchmark, Preparata’s reported 50 to 100 kilowatts per cubic centimeter would be approximately 150 to 300 times greater. The frequently cited value of 60 kilowatts per cubic centimeter would be about 180 times the representative reactor-fuel average.

The comparison becomes even larger when the entire reactor core is used as the denominator. The same representative reactor has an average core power density of approximately 102 watts per cubic centimeter, including coolant passages, structural materials, and spaces between fuel rods. Preparata’s claimed 50 to 100 kilowatts per cubic centimeter would be roughly 490 to 980 times that figure. However, the two denominators represent different physical quantities. The reactor-core value includes much more than the uranium dioxide fuel, while Preparata’s calculation uses only the palladium metal. Comparing the two without explaining the difference would exaggerate the apparent engineering advantage.

Surface heat flux narrows the comparison considerably. The representative reactor fuel rod operates at an average surface heat flux of approximately 0.584 megawatts per square meter, compared with approximately 0.76 megawatts per square meter for Preparata’s nominal 300-watt palladium wire. The values are of the same engineering order, even though their volumetric power densities differ by roughly two orders of magnitude. Commercial fuel rods also operate continuously within large, engineered cooling systems, whereas Preparata’s reported result came from a laboratory electrochemical cell. The comparison is meaningful as a measure of inferred heat generation per unit solid volume, but it doesn’t establish equivalent reactor performance, operating life, or scalability.

The Theory Behind Preparata’s Experiment

Preparata approached cold fusion through quantum field theory and condensed-matter physics. Working with researchers including Emilio Del Giudice, he developed a framework based on quantum electrodynamic coherence, in which large populations of particles could exhibit collective electromagnetic behavior. His 1995 book, QED Coherence in Matter, presented the broader theoretical framework. A 1994 paper coauthored with Martin Fleischmann and Stanley Pons applied related concepts to cold fusion in palladium deuteride. Rather than treating deuterium atoms inside the metal as independent particles, the theory proposed that collective interactions could produce physical conditions substantially different from those described by conventional models.

One proposed mechanism involved electron screening of deuterium nuclei. Under ordinary conditions, positively charged deuterons repel one another through the Coulomb force, making fusion at room temperature extraordinarily improbable. Preparata’s framework proposed that coherent electronic behavior could substantially reduce the effective repulsive barrier. The theoretical work discussed screening effects on the order of 100 electronvolts and suggested that collective oscillations within palladium could alter nuclear interaction probabilities. The proposed mechanism wasn’t ordinary thermonuclear fusion occurring at unusually low temperature; it depended on a fundamentally different description of the condensed-matter environment.

The theory also attempted to explain why a nuclear-scale energy release might appear primarily as heat rather than energetic radiation. Conventional deuterium-deuterium fusion predominantly produces either tritium and a proton or helium-3 and a neutron, with reaction energies of several million electronvolts. Preparata’s model instead emphasized a proposed reaction producing helium-4 and releasing approximately 23.8 million electronvolts. Collective interactions in the surrounding metal were hypothesized to transfer the released energy into electronic excitations and eventually heat, avoiding the large prompt radiation signatures expected from ordinary fusion channels. This proposed energy-transfer mechanism was essential to explaining the claimed calorimetric results.

A separate part of the theory addressed deuterium loading. Preparata proposed that a longitudinal electrical potential along a thin palladium conductor could change the chemical potential and distribution of absorbed deuterium. The approach became associated with the “Preparata effect.” The long, thin cathode was intended to sustain an electrical potential difference along its length while participating in electrolysis, potentially allowing unusually high deuterium concentrations within the metal. Later experiments explored deuterium-to-palladium atomic ratios approaching or exceeding one. The theory motivated the experimental configuration, but the available record doesn’t establish that it quantitatively predicted the eventual 50-to-100-kilowatt-per-cubic-centimeter result before the measurements.

Inside the Calorimeter

Preparata’s team used isoperibolic calorimetry, a method that estimates heat generation by monitoring temperature differences between an experimental cell and its controlled surroundings. The journal paper described a cylindrical electrochemical cell approximately 5 centimeters in diameter and 28 centimeters high, immersed in a constant-temperature bath. Platinum resistance thermometers measured temperatures at several locations, while an electrical heater delivered known calibration pulses. The investigators inferred excess thermal power from the difference between the heat indicated by their thermal model and the energy expected from the electrical and chemical inputs.

Calibration was a central feature of the experimental method. Rather than relying entirely on a fixed heat-loss coefficient determined before operation, the researchers periodically applied calibration-heater pulses during the runs. The heater generally operated in alternating six-hour energized and de-energized periods, allowing the investigators to update their estimate of the cell’s thermal dissipation. Representative calibration powers were approximately 1.30 watts for the platinum blank, 1.23 watts for one palladium run, and 1.59 watts for another. The authors reported typical excess-power uncertainties of approximately 0.7 watt for the blank and 2 watts for the palladium experiments.

The calorimeter nevertheless presented substantial technical challenges. Heat was generated primarily near the thin cathode rather than uniformly throughout the electrolyte, creating temperature gradients and potentially different heat-transfer pathways. Electrolysis produced gas bubbles that could change convection and mixing. The cell was open, allowing deuterium, oxygen, and water vapor to escape, but the outgoing gas stream wasn’t directly measured in the journal experiments. The investigators estimated the relevant chemical and gas-related contributions. Changes in gas evolution, evaporation, and temperature distribution could also change the relationship between the measured temperatures and the actual heat leaving the system.

A preliminary attempt to cross-check the results using heat-flux calorimetry revealed a significant discrepancy. At one reported operating point, electrical input was approximately 87 watts. The heat-flux measurement indicated about 170 watts of thermal output, while the simultaneous isoperibolic analysis indicated approximately 135 watts, with a stated uncertainty of 5 watts. Both methods suggested heat output above electrical input, but they disagreed by roughly 35 watts. Preparata attributed the difference partly to heat escaping with gases and imperfect cell sealing. The result was intriguing because both methods indicated an anomaly, but it also showed that the two calorimetric approaches hadn’t achieved close quantitative agreement.

The Nuclear Evidence That Was Missing

The principal scientific difficulty is that excess heat, even if measured correctly, doesn’t automatically establish nuclear fusion. Preparata’s 1996 high-power experiments weren’t accompanied by simultaneous quantitative measurements of helium-4, neutrons, gamma rays, tritium, or other nuclear products linked to the reported thermal output. The journal paper explicitly stated that the outgoing gas stream wasn’t analyzed. Without such measurements, the investigators couldn’t determine whether the amount of nuclear material produced was consistent with the inferred energy release. The reported calorimetry therefore remained separate from direct evidence of nuclear reactions.

The expected nuclear production can be estimated from Preparata’s proposed helium-4 reaction. If each reaction released approximately 23.8 million electronvolts, a sustained 300-watt nuclear heat source would require roughly 79 trillion reactions every second. Over one hour, it would produce approximately 280 quadrillion helium-4 atoms, corresponding to about 10 microliters of helium gas at standard temperature and pressure. That’s a small gas volume, but it could be investigated with appropriately designed containment and sensitive noble-gas mass spectrometry. A quantitative relationship between accumulated helium and integrated excess energy would provide a far stronger test of the proposed nuclear mechanism than calorimetry alone.

The energy inventory also makes the reported result significant. A genuine 300-watt excess maintained for one hour would release approximately 1.08 megajoules of energy. The nominal palladium wire contained only about 59 milligrams of metal, far too little material to supply that energy through ordinary stored chemical processes. A correctly measured megajoule-scale excess would therefore require an explanation beyond the wire’s finite chemical-energy inventory. However, the conclusion depends on establishing the excess heat independently. A systematic error in electrical-power accounting or thermal calibration wouldn’t be limited by the amount of chemical energy stored in the palladium.

Later Italian experiments attempted to connect excess heat with nuclear products. In 2002, researchers associated with ENEA published “Rapporto 41,” reporting helium-4 production and excess enthalpy in a different palladium configuration. Their apparatus used a thin palladium stripe approximately 100 centimeters long, 50 micrometers wide, and 2 micrometers thick, rather than the earlier cylindrical wire. The reported thermal anomalies were generally measured in milliwatts, not hundreds of watts. In one case, approximately 19 milliwatts of excess heat was compared with a helium-derived nuclear-power estimate of about 240 milliwatts. The discrepancy and the substantially different apparatus mean the ENEA work can’t be treated as an independent reproduction of Preparata’s 1996 high-power-density experiment.

How Nagel and Krivit Framed the Result

Physicist David Nagel helped preserve Preparata’s experimental claim in later technical reviews of cold fusion. In his 2017 report, “Evidence of LENR,” Nagel described the 1996 experiment as involving approximately 500 watts of maximum electrical input and more than 300 watts of maximum excess thermal power. The distinction is important because the reported 300 watts represented heat above the investigators’ calculated baseline, not the cell’s total thermal output. Nagel’s account emphasized the unusually large power density resulting from the small palladium volume and placed the experiment within a broader collection of reported low-energy nuclear reaction results.

Nagel’s earlier review, “The Status of Cold Fusion,” published in 1998, also discussed experimental evidence accumulated after the original 1989 Fleischmann–Pons announcement. His later treatment of Preparata’s result is particularly useful because it connects the experimental numbers with their engineering implications. A few hundred watts isn’t extraordinary in ordinary thermal engineering, but hundreds of watts attributed to only a few thousandths of a cubic centimeter of palladium is an entirely different claim. The distinction between absolute output and volumetric normalization is central to understanding why Preparata’s experiment continued to attract attention.

Science writer Steven Krivit presented the comparison more visually. His 2005 presentation, “What Really Happened With Cold Fusion and Why Is It Coming Back?,” included a graphic titled “Cold Fusion Volumetric Power Densities.” It compared reported cold-fusion results, including Preparata’s, with nuclear-reactor fuel-rod power density. The graphic offered an accessible way to communicate the magnitude of the reported result, but the comparison remained dependent on Preparata’s original calorimetry and the volume assigned to the active palladium. It didn’t represent an independent experimental measurement of the claimed thermal output.

The distinction between the original experimental claim and its later interpretation is important for the history of the phrase “fission-like densities.” Preparata’s 1996 conference paper explicitly reported 50 to 100 kilowatts per cubic centimeter, but the available research didn’t identify the phrase “fission-like” in his principal 1996 publications. Nagel and Krivit subsequently drew attention to the comparison with conventional nuclear fuel. Their accounts helped communicate the claimed engineering magnitude, but they didn’t establish the underlying nuclear mechanism or independently verify the experiment. The comparison remains mathematically meaningful only to the extent that the original excess-power measurements and cathode-volume assumptions are correct.

Reproducibility and the Unresolved Questions

Preparata’s strongest experimental argument involved the apparent consistency between palladium cells and their platinum controls. The 1996 journal paper described two palladium runs producing approximately 25 and 30 watts of excess heat, while the platinum blank remained near zero. The later conference presentation reported hundreds of watts from a palladium experiment and a corresponding platinum control with little apparent excess. Preparata also described a broader experimental campaign involving roughly 50 tests with similar cells. However, the published record doesn’t establish how many of those tests produced excess heat, how many failed, or whether the highest-power results could be reproduced reliably under identical conditions. The distinction is important: repeating an experimental procedure isn’t the same as repeatedly obtaining the same result.

Independent replication presents a more difficult question. The published literature contains numerous experiments involving palladium-deuterium electrochemistry, including reports of excess heat by Fleischmann and Pons, researchers at SRI International, and several Japanese and Italian groups. Some investigators reported excess thermal power ranging from fractions of a watt to tens of watts, while others reported substantially larger transient effects. But those experiments generally used different electrode geometries, loading methods, electrolytes, and calorimeters. They can’t automatically be counted as replications of Preparata’s particular configuration. The research reviewed for this story hasn’t established an independent experiment reproducing his reported 50-to-100-kilowatt-per-cubic-centimeter power density with comparable apparatus, electrical accounting, calorimetric documentation, and sustained operation.

Several technical uncertainties complicate the original result. Preparata’s calorimeter relied on temperature measurements and a calibrated thermal model rather than directly collecting and measuring every heat-loss pathway. The cell was open, electrolysis generated gases, and the outgoing gas stream wasn’t measured directly. Changes in evaporation, bubble-driven convection, temperature gradients, or the distribution of electrical heating could affect the inferred thermal output. The electrical arrangement also requires careful accounting because current passing longitudinally through a thin palladium wire can produce substantial resistive heating independently of the electrochemical current. A complete replication would need to measure both electrical circuits, determine where their energy was dissipated, and verify the heat balance using an independent calorimetric method. The approximately 35-watt disagreement between the two calorimetric methods reported in Preparata’s earlier work illustrates why these questions matter.

The unresolved issue isn’t whether 300 watts of heat can be removed from a thin palladium wire, or whether dividing that power by a few thousandths of a cubic centimeter produces a remarkable number. Both are straightforward engineering propositions. The unresolved issue is whether the reported thermal output genuinely exceeded every identifiable electrical and chemical energy contribution, whether the excess originated in the palladium, and whether the effect could be reproduced independently. A modern test could address these questions with a sealed electrochemical system, continuous mass and gas accounting, independently calibrated heat-flow calorimetry, precision measurements of every electrical input, and simultaneous helium and radiation monitoring. If such an experiment repeatedly produced hundreds of watts of verified excess heat from a similarly small palladium volume, Preparata’s fission-like power-density claim would become a substantially stronger experimental finding. Until then, it remains an extraordinary reported result rather than an established nuclear-energy capability.

Thirty Years Later, an Extraordinary Claim Remains

Nearly three decades after Preparata’s 1996 experiments, the reported power densities remain among the most striking claims in cold-fusion research. The numbers themselves aren’t difficult to understand. A few hundred watts of excess thermal power, attributed to less than five thousandths of a cubic centimeter of palladium, yields approximately 60 kilowatts per cubic centimeter. That’s roughly 180 times the average volumetric thermal power density of the uranium dioxide fuel in a representative commercial nuclear reactor. The comparison doesn’t require an unconventional interpretation of physics. It’s an engineering calculation based on reported excess heat and cathode volume. What remains unresolved is whether the calorimetry established that excess heat with sufficient accuracy, whether the energy originated in the palladium, and whether the result represented a nuclear process rather than an unrecognized experimental effect.

Preparata’s work also illustrates the distinction between extraordinary power density and practical energy production. A commercial nuclear reactor generates thousands of megawatts of thermal power through engineered fuel assemblies that operate continuously for years. Preparata’s apparatus reportedly generated hundreds of watts during laboratory experiments lasting hours, using an electrochemical system that required substantial electrical input. Even if the reported excess heat were independently confirmed, the experiment wouldn’t establish a commercially useful energy source. Engineers would still need to demonstrate reliable initiation, sustained operation, controllable output, efficient heat extraction, material durability, and reproducibility across many devices. The reported power density would be an important physical characteristic, but it wouldn’t by itself resolve those engineering challenges.

The experiment nevertheless presents a testable proposition. Modern calorimetry, precision electrical instrumentation, sealed electrochemical systems, and sensitive nuclear-product analysis could provide a much more complete examination of Preparata’s original claim. A replication could measure every electrical input, continuously account for escaping gases and water vapor, independently determine total thermal output, and monitor helium production alongside excess heat. It could also establish whether the effect depends on deuterium loading, longitudinal current, cathode geometry, or particular material conditions. Repeated demonstrations of hundreds of watts of verified excess heat, accompanied by quantitatively consistent nuclear products, would provide compelling evidence that an unusual energy-producing process was occurring. Repeated failures under well-characterized conditions would substantially weaken the original claim.

For now, Preparata’s reported result occupies an unusual position in the history of nuclear-energy research. The experiment produced numerical claims that can be compared directly with established reactor-engineering benchmarks, yet the underlying measurements haven’t achieved the independent verification necessary to establish a new source of nuclear energy. David Nagel and Steven Krivit helped bring attention to the extraordinary scale of the reported volumetric power densities, but their later comparisons couldn’t resolve the experimental uncertainties. The central question remains much as it was in 1996: Did a hair-thin palladium wire really produce hundreds of watts of unexplained heat, or did the calorimetry make an ordinary energy balance appear extraordinary? Until that question is settled experimentally, cold fusion’s fission-like power densities remain a remarkable scientific claim—and an unanswered engineering challenge.

References

Original experimental research

• Isoperibolic Calorimetry on Modified Fleischmann–Pons Cells — G. Preparata, M. Scorletti, and M. Verpelli. Journal of Electroanalytical Chemistry, Vol. 411, pp. 9–18 (1996). The principal peer-reviewed experimental paper reporting excess heat and volumetric power densities of approximately 3–3.6 kW/cm³. Full-text PDF.
• Everything You Always Wanted to Know About Cold Fusion Calorimetry — Giuliano Preparata. Proceedings of the Sixth International Conference on Cold Fusion, pp. 136–143 (1996). The central source for the reported 50–100 kW/cm³ power densities and the July 1996 experiments producing hundreds of watts of apparent excess heat. The link opens the conference proceedings.

Preparata’s theoretical foundations

• Possible Theories of Cold Fusion — M. Fleischmann, S. Pons, and G. Preparata. Il Nuovo Cimento A, Vol. 107, pp. 143–156 (1994). Discusses theoretical explanations for cold fusion, including collective effects and electron screening.
• QED Coherence in Matter — Giuliano Preparata. World Scientific (1995). Preparata’s book developing his quantum electrodynamic coherence framework, including its proposed application to cold fusion.

David Nagel and Steven Krivit

• Evidence of LENR — David J. Nagel (2017). Reviews experimental evidence for low-energy nuclear reactions, including Preparata’s reported hundreds-of-watts excess heat and unusually high volumetric power density.
• The Status of Cold Fusion — David J. Nagel (1998). A retrospective review of cold-fusion research and experimental developments during the decade following the Fleischmann–Pons announcement.
• What Really Happened With Cold Fusion and Why Is It Coming Back? — Steven B. Krivit (2005). Presentation containing the graphic Cold Fusion Volumetric Power Densities, comparing reported cold-fusion results with conventional nuclear-reactor fuel-rod power density.

Subsequent experimental research

• Experimental Evidence of ⁴He Production in a Cold Fusion Experiment — A. De Ninno, A. Frattolillo, A. Rizzo, E. Del Giudice, and G. Preparata. ENEA Technical Report RT/2002/41/FUS (2002), commonly known as Rapporto 41. Reports later Italian experiments investigating excess heat and helium-4 production in palladium.

Additional research archives

• LENR-CANR Scientific Library — An archive of experimental papers, conference proceedings, and technical reports concerning cold fusion and low-energy nuclear reactions, including Preparata’s publications.
• New Energy Times — Steven Krivit’s research and reporting archive, containing historical cold-fusion publications, technical commentary, and presentations.