A. Introduction
B. The Reaction Process
C. Neutrons
D. Direct Detection of Charged Particles and Gammas
E. Search for Accumulation of Tritium and Helium
F. Unconventional Explanations
G. Search for Products of Cold Fusion in the Earth

BACK to Contents


The nuclear fusion of deuterium has been studied intensively for over 40 years. The reaction between two low energy deuterium nuclei has been observed to proceed in three ways:

(a) D + D -->3He (0.82 MeV) + n (2.45 MeV)

(b) D + D -->T (1.01 MeV) + p (3.02 MeV)

(c) D + D -->4He + gamma (23.85 MeV)

The reactions (a) and (b) have been studied down to deuteron energies of a few keV and the cross sections (production rates) found to be equal to within 10%. In the interaction of deuteron beams with heavy ice or metal deuteride targets, almost one 2.45 MeV neutron is produced (with an accompanying 3He for every triton (with an accompanying proton). This near-equality of neutron and proton branches (production rates) of the D+D reaction, shown in Figure 3.1, is a reflection of the basic symmetry of nuclear forces between proton and neutron, disturbed only slightly at the MeV energies of the emerging particles by the Coulomb interaction, which is not symmetrical between proton and neutron. The cross section (production rate) for reaction (c) is on the order of 107 lower than the first two reactions.

All nuclear reactions at low energies between two deuterons are retarded by Coulomb repulsion between the positively charged nuclei--the penetration of the repulsive Coulomb barrier changes rapidly with bombarding energy: for instance the measured cross section for reaction (b) changes from 0.2 microbarns at 2.7 keV to 35 millibarns at 100 keV. But the ratios for the three reactions appear to be constant below 100 keV.

Any fusion between deuterium nuclei must lead to detectable fusion products. For reaction (a) neutrons are the most easily detected product, by direct counting. For (b) the protons or tritons can be detected by direct counting, and the accumulated tritium could also be identified by its radioactivity, albeit with lower sensitivity for few-hour or few-day experiments, in view of the 12-year half life of T. Neutron counting is perhaps again the most useful technique, since neutrons must be produced by the energetic tritons interacting with deuterons in the material at the expected rate of 1 neutron for every 10,000 to 50,000 tritons. Reaction (c) leads to readily detectable high energy gamma rays and 4He; the latter may be identified by mass spectroscopic measurements, whose sensitivity is low--though the 1011 levels implied by 1 watt of heat should be readily observable.

In the following we summarize the experimental evidence on these fusion products. First we discuss the plausibility of reactions at room temperature and the issue of whether the constancy of the three reaction modes is a reasonable extrapolation to such low energies. Then the data on neutrons, charged particles, gamma rays and tritium are summarized. Finally, some comments are included on unconventional explanations, and geochemical evidence is summarized on proposed cold fusion in the interior of the earth.

We have used published material, where available, or material prepared For publication and presented at formal meetings or as preprints distributed without restriction as to citation. We have also benefited from data submissions in response to Panel questions and from reports from the national laboratories, all of which are available to the public in the DOE reading room. It is important to include not only positive results, that claim the detection of fusion products, but also the negative ones, that have attempted to replicate the experimental procedure of the former and failed to detect anything above background, at a level of sensitivity substantially better than the positive results.


Fig 3.1

Figure 3.1: Neutron & Proton Branches of the D+D Reaction

Click on thumbnail to see full-size image.




Fusion reactions can occur only if during a nuclear collision, the Coulomb barrier is surmounted or, at low energies, penetrated so that the nuclei approach each other within about 10-5 nanometers. This distance is some 10,000 times smaller than the typical separations of atoms in ordinary matter. The penetration of the barrier at low energies takes place through a well-understood quantum mechanical phenomenon called tunneling that allows fusion to occur in collisions far less violent than might be required otherwise.

In the thermonuclear fusion that occurs in stars and in laboratory "hot fusion" experiments, high temperatures, (tens of millions of degrees or more) provide 'he violent collisions required to produce fusion. However, in the cold fusion interpretation of the experiments, it is claimed that the penetration of the barrier through quantum mechanical tunneling has become so effective as to allow fusion to occur even at room temperatures. Further, some experimenters claim that the nuclear process is changed by an unspecified mechanism so as to alter dramatically the nature of the reaction products. Each of these claims must be understood as separate and equally surprising.

Some simple calculations illustrate how remarkable is the claim of fusion at room temperatures. The fusion rate for the two deuterium nuclei in a deuterium molecule (where their separation distance is even closer than when embedded in a metal) results in one fusion per year in a solar mass of deuterium. Further, the fusion of protons and deuterons is calculated to be 109 times faster than the D&@nbsp;+ D reaction (although it is still extraordinarily slow). No mechanism is known by which these rates could be enhanced by the 40-50 orders of magnitude required to agree with the reported observations.

One mechanism invoked for enhancing cold fusion rates is screening the electrostatic repulsion by "heavy" electrons. Endowing the electrons with a hypothetical mass would indeed enhance fusion rates sufficiently to agree with most cold fusion claims [KOO]. It is also true that there are "heavy fermion" materials whose thermodynamic properties at very low temperatures are characteristic of quasiparticles with masses many times those of a free electron. However, this phenomenon is understood as involving long-wavelength Extcitations in which strong correlations modify the electron wave function near the Fermi surface. As such, heavy fermions extend over many lattice sites. Because the tunnelling in nuclear fusion occurs at distances smaller than one lattice site, only the short-wavelength "bare" electron excitations are relevant for screening, and cannot enhance the fusion rate significantly.


B.1. The D + D Branching Ratios

The relative rates of reactions (a), (b), and (c) are called the branching ratios and are a crucial issue in the discussion of some cold fusion claims. These reactions have been studied in laboratory experiments using accelerators for deuteron energies above a few keV [KRA]; the smallness of both the reactions (a) and (b) cross sections prevents reliable measurements at lower energies. The ratio between these two rates exhibits a weak energy dependence and is near 1.0 at the lowest energies, as seen in Figure 3.1. Data from muon-catalyzed D + D fusion [BAL], which probes an even lower energy range, is still consistent with nearly equal rates.

A change in the known branching ratios of more than one million, in favor of reaction (b) over reaction (a), would be required to explain experiments that claim to observe high fusion rates (either through heat or tritium production) without a corresponding high neutron flux. As cold fusion is thought to occur at energies on the order of a few electron volts (eV), this is not directly ruled out by the data discussed above. However, no mechanism is known for inducing such a rapid energy-dependence in the branching ratio. The Oppenheimer-Phillips process involving the Coulomb break-up of the deuteron has sometimes been invoked in this regard. However, this process is not effective at low energies in the D + D system.

B.2. The Gamma Branch.

Some researchers have hypothesized that the D + D --> 4He + gamma (23.85 MeV) reaction, which is ordinarily some 107times weaker [BARN] than reactions (a) and (b) in which two fragments are produced, somehow dominates in cold fusion situations. To be consistent with the lack of neutrons, a very large enhancement of the gamma branch by a factor somewhere in excess of 1013 would be required. We know of no way whereby the atomic or chemical environment can effect such an enhancement, as this ratio is set by nuclear phenomena and is on a length scale some 104 times smaller than the atomic scale.

If there were such an enhancement, high-energy electromagnetic radiation (photons, positrons, or fast electrons) would be observed. While direct coupling to the lattice through unspecified mechanisms has been invoked to suppress such radiation, any such coupling must occur through the electromagnetic field and would result in some observable high-energy radiation.

B.3. Anticipated Secondary Yields from Fusion Products.

i.   Neutrons from tritium. The tritons produced in reaction (b) have an energy of 1.01 MeV. This energy must be lost in the immediately surrounding material, which in an electrolytic cell is either the Pd electrode saturated with deuterium, or heavy water. The tritons have insufficient energy to cause nuclear reactions with Pd.


    However, they will also bombard the deuterium in the surrounding material. The T + D reaction is a rich source of 14-MeV neutrons, with a cross section that reaches 5 barns (i.e., 5x1014cm2) at 0.12 MeV, then falls to about 0.7 barns at 0.5 MeV, and reaches slightly below 0.3 barns at 1 MeV. For the 1.01 MeV tritons from the D + D reaction an average cross section is about 1.2 barns. For tritons that are stopped in PdD this translates into a neutron yield between 1.5 and 2x10-5 neutron per triton; for tritons stopping in heavy water there are about 9x10-5 neutron per triton.

ii.   Coulomb excitation of Pd by protons. The even Pd isotopes (104, 106. 108, 110) with abundances of 11, 27, 26, 12% have first-excited 2-states at 555, 512, 434, 374 keV and cross sections between 0.5 and 0.8 barns. The cross sections for Coulomb excitation are in the vicinity if 20 to 50 millibarns and thus the yields of gamma rays expected are 2 to 5x10-6 per proton. In palladium the half thickness for absorption oF these gamma rays is about 4 mm, in water it is several cm.

In terms of power, there must be about 108/sec secondary (14 MeV) neutrons per watt of fusion, even if direct neutron production (reaction a) is completely suppressed and all the reaction goes into tritium production (reaction b). Under these conditions there must also be slightly under 107 secondary gamma rays per second, of well defined energies, in the 500 keV range.

B.4. The p + D Reaction

It has been suggested that an alternative fusion process could be the reaction

p + D -->3He + gamma (5.49 MeV)

for which the gamma ray production (observability) factors are still overwhelmingly small at room temperature, but somewhat less so than for the D + D process [KOO]. This reaction produces a readily observable gamma ray and if it is to account for 1 watt of heat, then it should also produce3He in observable concentrations.



C.1. Detection.

Neutrons are a major product of D + D fusion. Neutrons are convenient particles to detect, since they interact only with the nuclei of atoms and so can emerge from reaction vessels of substantial size unscathed and without having lost energy. Similarly, large counters can be used without the problem of thin entrance windows, since neutrons enter into the volume of the counter without difficulty. Some simple facts about neutron detection are summarized in Appendix 3.A.


C.2. Initial Claims

The University of Utah (UU) group in its initial publication (FLE] claimed the detection of neutrons from D + D by virtue of the gamma ray emitted by the capture of the moderated neutron in the water bath surrounding the electrolytic cells. A narrow peak in the pulse-height spectrum from the NaI scintillator was published, narrower than is possible with this type of detector, and with internal inconsistencies in the energy scale as pointed out by a group at MIT [PET]. The photopeak at 2.2 MeV obtained at MIT from 252Cf spontaneous fission neutrons moderated in water and radiatively captured on protons is accompanied by other peaks from natural background that enable one to calibrate the energy. Successive interchange between UU and MIT groups in the scientific literature has demonstrated that the claimed detection of neutrons by the proton capture gamma ray at UU was an artifact of the experimental apparatus.

The original publication from Brigham Young University (BYU) [JON-1] presented the detection of neutrons as experimental evidence for the existence of cold nuclear fusion. The neutrons were detected in a two-stage neutron counter--first by the proton recoil in organic scintillator, followed within a few tens of microseconds by a signal from the capture of the moderated neutron on boron viewed by the same photomultipliers. This double detection of a single neutron serves substantially to reduce the ambient background due to gamma rays, although there remains background in the experiment due to gamma rays and to real neutrons from cosmic rays and other sources.

BYU has been working in collaboration with other groups, notably at Los Alamos National Laboratory (ME4], and also with a group at Yale University. The original claim of neutron detection five standard deviations above the background is somewhat reduced in statistical strength if one considers the degrees of freedom that are fixed by the presentation of a peak in one of a number of experiments and at a particular energy, and also the possible fluctuation in the cosmic-ray neutron background. Ordinarily, however, such a result can be improved through improved shielding or by moving to an underground site. Another group at BYU [PIT] electrolyzed 3 M LiOD on Pd cathodes and detected about 2.1x10-3 counts per second as compared to 1.2x10-3 for light water cells.

Typical of the experiments with greater sensitivity is work by the group at Sandia National Laboratory [SCH], in which a site was found with substantially less background and results presented for a limit on neutrons produced in electrolytic fusion. Similar results from the Frejus tunnel in France were also presented in Santa Fe [DEC]. However, in a BYU collaboration at the Grand Sasso Laboratory deep underground [BER] involving "electrolytic infusion of deuterons into titanium" cathodes, the detection of 350±51 counts of 2.5 Mev neutrons

*Care is needed as the rate of cosmic ray neutrons can fluctuate by 20% or more with variations in barometric pressure or with solar activity.


indicates consistency with their initial results, though unfortunately their background is 100-1000 times higher than previous measurements of the neutron level in the Grand Sasso Laboratory [ALEK] so that the excess neutron yield reported is still comparable to the background and the advantages of the underground site are not fully realized.

At the Santa Fe workshop, Moshe Gai of Yale presented results obtained in collaboration with Brookhaven National Laboratory, in which no neutrons were detected from electrolytic cells above the detection limit of 10-25 n per D-D pair/sec [GAI]. In a recent @preprint Salamon et al. from the University of Utah [SAL] report measurements of neutrons in the Pons Laboratory over 67 hours with several electrolytic cells operating. They observed no neutrons above background and quote a limit of 1 neutron per second emitted by any of the cells, comparable to the limits in Table 3.1.

Many claims have been made for the production and detection of neutrons produced in electrochemical cells, but these claims have mostly been withdrawn or moderated by the discovery of difficulties with the counter--particularly with the BF counters used. Some counters are sensitive to humidity; others to microphonic noise (vibration), or to other afflictions. Some results on reported neutron fluxes, compared to the flux reported by the BYU group, are summarized in Table 3.1.

C.3 Dry Fusion.

Results presented in April 1989 by a group at Frascati (DEN] opened a different area of investigation for the study of D + D cold nuclear fusion. In this work, deuterium gas at 60 atmospheres pressure was allowed to contact titanium lathe turnings in a stainless steel reaction vessel, and the temperature of the sample was varied either by heating or by cooling. No neutrons were observed from the hydriding reaction, at room temperature or at elevated temperature, when viewed by a nearby BF counter. However, after cycling to liquid nitrogen temperature (77°K) bursts of counts were observed--typically on the order of 20 counts per burst over a period of 60 microseconds. One set of counts, obtained by cycling to nitrogen temperature, shows neutrons mainly in these bursts.

A different type of neutron emission was also claimed by the Frascati group [DEN] following warming from nitrogen temperature over one weekend. The observed count rate follows a bell-shaped curve, rising to a peak of 300 neutrons per 10-minute counting interval, over some 5 hours. These important experimental results provoked great effort toward verification, both at Frascati and elsewhere. A recent personal communication from M. Martone at Frascati indicates that neither the burst results nor the continuous neutron emissions from the D-Ti system or from any other dry fusion activity at Frascati have been confirmed. In addition, electrochemical cells produced no observable neutrons, and their operation was terminated in July.

A group at LANL [MEN] has conducted dry fusion work with Ti and Pd, and has presented results both at the Santa Fe meeting and in a preprint. This group at LANL uses high-efficiency systems that moderate any fast neutrons emitted from experimental cells, detecting the moderated thermal neutrons in 3He gas counters. Some bursts of neutron counts are observed 3000-5000 seconds after the sample


is removed from liquid nitrogen, at sample temperatures of about -30°C. These bursts, of up to 100 neutrons, are seen in about 30% of the samples tested. An attempt to reproduce this effect at Sandia National Laboratory yielded negative results [BUT].

In a very recent report, a Yale-BYU-BNL collaboration using a different detection scheme, with multiple liquid scintillation counters with an overall efficiency essentially the same as that of the Los Alamos group, and incorporating counters to monitor cosmic ray showers, has examined four cells with conditions reproducing those af the Los Alamos experiment [RUG]. They see no bursts larger than five neutrons in 103 hours of temperature cycling with these cells. Two bursts of five were seen in background runs, and two more in the runs with the pressurized Ti-O cells, all of them coincident with cosmic ray events.

Finally, a conference report [IYE] from the Bhabha Atomic Research Center [BARC] in India provides text and tabulated results from several groups at BARC. Fig. 1 of the BARC report shows counts from neutron detectors observing a large electrolytic cell, with an estimated 2x107 neutrons in the 5 minutes following an overpower trip of the electrolyzer. Tritium and neutrons are observed at BARC from cathodes fabricated of PdAg alloy as well as from pure Pd.

Figure 2 of the BARC report shows dry fusion 3He neutron-counter output during a gradual increase of temperature of 20 g of Ti while deuterium gas was being pumped off. It is also commented that samples could be loaded with deuterium gas at 1 atmosphere and 900°C, and that "one such disc shaped button loaded on Friday 16th June began emitting neutrons on its own, almost 50 hours after loading. It produced (about) 106 neutrons over an 85-minute active phase. The background neutron counter did not show any increase in counts over this time."

C.4. Fracto-Fusion.

In 1986 a group at the Institute of Physical Chemistry in Moscow reported that when a single crystal of LiD was Fractured by a device powered by an air gun, a few neutrons appeared to be produced [KLU]. These were attributed to internal fields associated with fractures in the material. Recently the same group reported that when titanium chips were agitated in a drum with heavy water and deuterated polypropylene, using steel balls and vibration at 50 Hz, neutrons were observed for a few minutes at a rate of 0.31 ± 0.13 counts/sec [DER]. After a few minutes no neutrons were seen.

Neutron emission during plastic deformation of deuterium-containing solids under pressure is reported by Yaroslavskii [YAR] from "rheological explosions" induced by rotating the anvils of a press on a rock sample to which grains of beryllium bronze and D2O have been added. Typically, 1000 pulses are detected in a burst, interpreted by the experimenter as arising from 106 neutrons.

These effects are not fully understood at present, the experiments are difficult and need to be repeated by others. If confirmed, a quantitative understanding of such effects could lead to interesting physics. The scale of energy in internal field experiments is still orders of magnitude higher than the one relevant to cold fusion.


C.5. Secondary Neutron Yields

There are problems of consistency between the numbers of tritium atoms detected in some of the experiments discussed above and the number of neutrons. The BARC abstract reads, "The total quantity of tritium generated corresponds to about 1016 atoms suggesting a neutron to tritium branching ratio less than 10-8 in cold fusion." But, as discussed above there should be at least one neutron per 10,000 tritons, if the observed tritium were originating from fusion. Hence the neutron yield is 1000 times less than expected.

C.6. Summary

All the experimental measurements of neutrons associated with cold fusion give upper limits that are much smaller than that consistent with the reported excesses of heat production if resulting from D + D fusion.



D.1. Charged Particle Searches

A few experiments [POR, PRI, REH, SUN] to measure the 3 MeV protons and/or the 1 MeV tritons produced in the reaction, D + D -- T +p, have been reported; they are summarized in Table 3.2. A variety of different methods has been used, but the lowest limit on charged-particle production appears to be that set by Price using plastic track detectors. Their setup was designed so that the light water control cell matched the heavy water cell as closely as possible. Electrolysis was performed for 13 days, and the cathode stoichiometry was determined to be H(D)/Pd = 0.8.

Both cells showed track production rates that were consistent with the alpha-particle emission rate for native Pd foils due to trace (ppm) impurities of the natural 238U and 232Th decay chains; however, no tracks due to protons with energies between 0.2 and 3 MeV or tritons with energies between 0.2 and I MeV were found. From these data Price [PRI] set limits on the fusion rate of less than 0.002 per cm3 per second. This value results in an upper limit of 8.3x1026 fusions per D-D pair per second. This is about an order of magnitude lower than the limits obtained using Si surface barrier (SSB) techniques.

A limit on the fusion rate of 0.028 per cm3 per second or 1.2x10-24 fusions per D-D pair per second was obtained by Ziegler [ZIE] using a SSB technique. Porter [POR] used a SSB detector to view the back of a 76 micron thick Pd foil cathode in a heavy water electrolysis cell. They obtained an upper limit of 6x10-25 protons per D-D pair per sec at the 2 sigma level; chemical analysis of their electrolytes showed no evidence for anomalous increases in tritium concentrations.

Sundqvist et al. [SUN] also used a SSB technique to detect protons. The detector was placed close to Pd foil cathodes that were thin enough to allow all the protons produced to escape from the foil. All of their runs gave a null result within the statistical errors, resulting in a fusion rate of -2.1 (±2.2)x1024, if a bulk process is assumed.


Recently, Rehm [REH] has reported using a proportional Counter to search for charged particles from electrolytic cells with Pd and Pt electrodes in 0.1 M LiOD in D2O. They obtained an upper limit of 4x10-23 fusions per D-D pair per second, not as low as the limits using the other methods.

In summary, a variety of experimental techniques has been used in searches for charged particles; all of them set very low limits on fusion occurring via the D + D --> T + p reaction. Most of these results set limits that are considerably less than Jones' [JON-1] maximum value of 1±0.2 x 10-23 fusions per D-D pair per second for the D + D --> 3He + n channel obtained from neutron measurements.

The upper limit of Price [PRI] Of 8x10-26fusions per DD pair per second is below the average rate inferred from the neutron measurements of Jones or even of Menlove [MEN]. The low limits which the searches for charged particles (either protons or tritons) place on their production is thus inconsistent with the reported production of either neutrons or tritium via the cold fusion reaction.

D.2. Gamma-Ray Searches

A rare branch of the D+D reaction proceeds through capture, in which a 23.85-MeV gamma ray is emitted. Similarly, the p+D reaction is associated with a 5.49 MeV gamma ray. In several published searches no gamma rays that would be associated with the D+D or p+D capture reactions were seen. They include a report by Henderson [HEN] who cites limits around 10-23/sec 24-MeV gamma rays emitted per deuteron in various cells. Porter [POR] reports no 5.5 MeV gamma rays--though no absolute limit is quoted. They also comment on the absence of K X-ray production from Pd. Greenwood [GRE] reports limits of 10-23 for gamma rays above 1.9 MeV.

Other negative results are quoted in the Santa Fe abstracts without quantitative detail. From the data of Lewis et al., [LEW] it can be calculated that fewer than 4x10-25 5.5-MeV gammas are emitted per second per deuteron in a Pd cathode, and fewer than 2x10-23 24-MeV gammas are emitted per second per deuteron. Negative results from measurements of gamma rays were also reported recently from Utah on electrolytic cells in the Pons Laboratory over 831.5 hours during which several cells were operating [SAL].

D.3. Cluster Ion Fusion.

Another effect reported recently is that of Beuhler et al. [BEU] who accelerated singly charged clusters of D2O, on the order of 100 molecules, to voltages up to 325 keV onto a deuterated Ti target. They obtained evidence of the D + D --> p + T reaction, and more recently of the n+3He branch, at a rate of about 1 per 1011 clusters.

These effects are not fully understood at present, the experiments are difficult and need to be repeated by others. If confirmed, a quantitative understanding of such effects could lead to interesting physics. The scale of energy in cluster ions experiments is still orders of magnitude higher than the one relevant to cold fusion.




As has already been noted, all D+D fusion reactions produce either tritium or helium. Besides the direct detection of tritium or helium nuclei when they are produced, as discussed above, searches have also been made for the accumulation of tritium or helium gas in some of the palladium cathodes claimed to produce excess heat. These searches for accumulation of tritium or helium are discussed below.

In such experiments it is important to determine the initial tritium content of the heavy water and recognize that the electrolysis of the heavy water will enrich the naturally occurring tritium in the heavy water. Also, it should be noted that the detection of accumulated tritium or helium is generally less sensitive than direct detection. For example, the detection of tritium by measurement of its beta decay is inherently a less sensitive probe of the D+D reaction than the direct measurement of neutron or charged particle production. About 107 tritium atoms give 1 decay by beta emission per minute. The tritium content of normal water is about 10-18 relative to hydrogen but, as discussed in Appendix 3.B, the normal manufacturing of heavy water also enriches tritium. Heavy water currently being sold gives between 120 and 180 disintegrations per minute per milliliter (dpm/ml) from tritium decay. Thus it is important to determine the initial tritium content of the heavy water and to account for the electrolytic enrichment.

E.1. Null Experiments with no Excess Tritium.

Most reports to date of excess tritium in electrolytic cells can be accounted for by the electrolytic enrichment process. This includes experiments at ANL [3RE, RED-2], BNL [DAV, MCB, WIE], Cal Tech [LEW], CRNL [SCO], INEL [LON-2], LLNL [ALD], NRL [ERI], ORNL [FUL, SCO], Sandia [NAR], SRL [RAN], Texas A & M [MAR], and Utah [WAD, WIL]. The levels of tritium reported by Fleischmann and Pons are also consistent with this interpretation.

E.2. Tritium Bursts.

A few experimenters report occasional irreproducible amounts of excess tritium in D2O samples from their electrolytic cells after days of operation. This includes observations by Storms [STO] at Los Alamos, and Fuller [FUL] and Scott [SCO] at Oak Ridge National Laboratory (ORNL). The ORNL experiments show single cases of excess tritium of short duration, after which a cell returns to background level. Storms reports excess tritium, 100 times background, in two cells out of 70.

E.3. Closed Cells - Correlation with Excess Heat.

Four groups [MCB, MCC, SCO, MAR] have looked for tritium production in closed electrolytic cells. The initial tritium in the cell is that contained in the heavy water, and any contained in the electrodes. These experiments detect the excess tritium generated from the electrolytic process except for that which may be contained in the Pd cathode. In general, the deuterium inventory in the cathode is negligible compared with the D2O. Only that tritium formed within the cathode and which remains there because of slow diffusion is unaccounted for. In these experiments the total excess tritium formed in the D2O is less than


104 T atoms/sec. If this tritium is produced by the D+D reaction, then the maximum excess power (cold fusion power) is 10 milliwatts. In one experiment [WAD] in an open cell there was a heat burst reported of 35 watts for 90 minutes (187,000 joules). No excess tritium above the electrolytic enrichment was measured after the burst. Clearly the heat burst does not come from the D+D reaction.

E.4. High Levels of Tritium.

Two groups [PAC, IYE] find tritium at levels of 1012 to 1014 T atoms/ml D2O after periods if electrolysis of the order of hours. This amount of tritium cannot be produced by electrochemical enrichment with the D2O volume reductions reported. The results of the Bockris [PAC] group at Texas A&M for cells in which excess tritium was found are given in Table 1 of their paper. Excess tritium is not found in all of their cells. A listing of cells in which no excess tritium was found is given in their Table 4. The Bockris cells are 0.1 M in LiOD and have nickel anodes. They precipitate nickel oxide during the electrolysis; some nickel is also electroplated out on the palladium cathode. In one experiment, A8, the specific activity of the D2 gas produced by the electrolysis was measured. It is 100 times that of the electrolyte.

Wolf et al [WOL-1] at Texas A&M looked for neutron production in Bockris type cells. An upper limit to the production rate is 1 neutron/second, which is 10-10 less that of the tritium production rates reported with similar cells by the Bockris group [PAC]. This large discrepancy from the equal production rates for neutrons and tritons required by the branching ratio in the fusion reaction, discussed in section II.B., is inconsistent by a factor of 10,000 to 100,000, even with the secondary neutrons that must accompany the tritons produced from nuclear fusion.

The most extensive and systematic search for tritium in the electrolysis of D2O with Pd cathodes has been carried out by Martin [MAR] at Texas A&M. He has used both open and closed cells. His cathodes come from either Johnson & Mathey, a major supplier, or Hoover and Strong, who supplied the cathodes to the Bockris [PAC] group. He has operated cells with Pt, Ni wire and Ni gauze (obtained from Bockris) anodes. In a recent communication Martin has reported that, in a recent cell, tritium has been found at a level approximately 50 times the initial specific activity of the heavy water (cell M-1, Table 1) [WOL-2]. This amount of excess tritium is larger than what can be expected from electrolytic enrichment. Wolf [WOL-2] now concurs with the analysis of this Panel that none of the tritium found in any of the Texas A&M experiments is produced by the known D+D reaction.

The BARC [IYE] group have found amounts of tritium comparable to the Bockris group in the D2 electrolyte from cells in which electrolysis was carried out for a few days with currents varying between 1 to 100 amperes. There is again a factor of 1000 internal inconsistency between the measured neutron yields and the secondary neutrons that have to be there if this tritium was produced by fusion--even if one assumes a drastic modification of the branching ratio in the D+D reaction.


E.5. Summary of Tritium Results.

Some experiments have reported the production of tritium with electrolytic cells. The experiments in which excess tritium is reported have not been reproducible by other groups. These measurements are also inconsistent with the measured neutrons on the same sample. Most of the experiments to date report no production of excess tritium. Additional investigations are desirable to clarify the origin of the excess tritium that is occasionally observed.

E.6. Searches for Helium.

One branch of the D+D reaction produces 3He and a neutron, while another, more rarely, produces 4He and a gamma ray.

Thus far, there are no reports of accumulated helium, either 3He or 4He. Early reports of helium by Pons and Fleischmann were latter retracted. The Panel has been informed that new measurements of helium have been carried out on palladium cathodes (0.38 gm) provided by Professor Pons to several laboratories. The University of Utah arranged with Battelle Northwest to have five of their Pd electrodes examined for helium. Battelle distributed a portion of each Pd electrode to seven different laboratories. These results, although available, have not yet been released.

Other reported searches for helium have yielded only negative results. These searches involve sensitive mass spectroscopy, for instance by Lewis [LEW]. Both the gaseous products of electrolysis and the Pd electrode were examined. Helium is expected to be trapped in the palladium lattice, so the Pd electrodes were melted after being used in the electrolysis of heavy water and the evolved gases analyzed: no helium was found above a limit of 8x1011 He atoms per cm3 of Pd. A similar limit was set by the Harwell (U.K.) group [WIL]. Although such limits are not as stringent as the ones set by direct counting of particles, the levels of helium corresponding to the background are still several thousand times below that corresponding to 1 watt of D+D fusion for one hour.

Searches for helium were made at Lawrence Livermore National Laboratory on palladium cathodes provided by the Texas A&M group (0.017 gm and 0.018 gm). These small wire cathodes were claimed to produce excess heat generation of 40 mW. To the level of 3x105 3He atoms and 5x108 4He atoms in the sample, there was no helium generation in the wires. This is many orders of magnitude below the level required with 43 mW of fusion power for 100 hours as claimed [HOL, THO].



F.1. D + D Reactions

The data on fusion products, even where positive results are reported, give rates far below those that would be expected from the levels of heat reported in some electrolysis experiments. Some proposals invoke mechanisms where the reaction heat from the D + D --> 4He process would go entirely into lattice heat, rather than a photon [WAL, HAGE]. Analogies have been made with the internal conversion process, and with the Mossbauer effect. Neither of these analogies is applicable to 4He, as discussed in the next two paragraphs.


Internal conversion allows an atomic electron of an excited nucleus to carry off the reaction energy instead of a photon. This process is understood quantitatively--it is dominant in heavy atoms with tightly bound inner electrons and for low energy (less than 1 MeV) photons. In helium the atomic electrons are loosely bound and the photon is 23.85 MeV--there cannot be any appreciable coupling between the photon and the atomic electrons, and internal conversion or any related process cannot take place at anywhere near the rate that would be required. The proposal of Walling and Simons invokes enhancement of internal conversion by electrons of high effective mass appropriate to the solid; as we have discussed above, such band structure effects cannot play the role of real high-mass electrons either in screening at sub-atomic distances or in the internal conversion process at MeV energies. Furthermore, although Walling initially reported 4He in an appropriate amount to explain claims of excess heat, this result was due to atmospheric contamination.

In the Mossbauer effect the momentum of a very low energy (below 100 keV) photon, but not its energy, is taken up by the entire lattice in a coherent mode. The process cannot be relevant to the present process.

More generally, there are numerous reactions analogous to the D+D or p+D fusion process in which gamma rays of comparable energy are emitted from low energy nuclear reactions (thermal-neutron capture gamma rays). The cross sections for capture have been studied carefully and quantitatively; they are essential to the operation of fission reactors. If there were any anomalous processes in which a capture gamma ray were totally suppressed in favor of direct conversion into lattice heat, this would have almost certainly been noticed as a discrepancy in cross sections with major implications for the operation of reactors. After four decades of extensive study of the processes relevant to the operation of fission reactors the possibility is remote that an entirely new process, that could dominate these nuclear reactions, would have remained hidden.

F.2. Other Fusion Reactions

In addition to the D+D and p+D reactions discussed thus far, there are several other nuclear reactions that would be substantially exothermic if they could take place at a reasonable rate at low energy. Among these are deuterons fusing with 6Li, 7Li, 16O, as well as various Pd isotopes. The reaction rates for these processes are again governed by the Coulomb barrier and for fusion at low temperatures this becomes even more overwhelming than for D+D. The process on Li isotopes, where the nuclear charge is three, is relatively the most favorable, but even these would give fusion rates that are enormously suppressed even in comparison with that for the D+D reaction; the rates are some forty orders of magnitude slower. The D + 6Li and p + 7Li reactions would not produce neutrons or direct gamma rays--all the energy would be in alpha particles (4He nuclei), but these in turn would cause Coulomb excitation of the Pd. No such gammas have been seen.




Products of low-level cold fusion have been inferred to be produced by natural geologic processes [JON-1, JON-2]. The 3He:4He ratio is anomalously high in volatile emissions from deep-source volcanoes such as those in Hawaii, Iceland, and Yellowstone [LUP, KUR, MAM]; anomalous T is also suggested by fragmentary data [OST, JON-3], and production of other radiogenic products such as 36Cl have been predicted [KYL]. Although the high 3He values have previously been considered relics from early earth processes, presence of anomalous T or anomalous 36Cl (beyond that due to bomb tests) would be definitive evidence of natural cold fusion deep within the earth. Implications would be major for geophysical problems such as heat-flow modelling, element-distribution with depth, and composition of the Earth's core.

Some isotope geochemists see no evidence for naturally occurring cold fusion [CRA]. However, several government and university labs are searching for evidence of such fusion processes as recorded by volcanic volatile emmissions [JON-3, KYL, GOF, LOC, QUI], independently of laboratory fusion experiments. Such geologic studies could add to understanding the behavior of volcanic emissions. No rigorous results are yet available, but experiments proposed or underway at Brigham Young, Los Alamos, Lawrence Livermore, New Mexico Tech, and the U.S. Geological Survey (Denver) should yield data within 6 months to 1 year.



Careful experiments have been carried out to search for the expected products of cold fusion. The measured products are many orders of magnitude lower than what would be expected from the heat production reported in electrolysis. In many experiments, no products are detected. Some experiments report neutrons or tritium at a much lower level--however, the rates of these two fusion products (measured in the same experiments) are inconsistent with each other, again by large factors. In particular, reported tritium production is accompanied neither by the one 2.45-MeV neutron per T observed in all other low-energy D+D fusion, nor by the 10-514-MeV neutron per T that would be produced by the 1.01-Mev T itself in the D-rich environment.

The neutron bursts reported in some experiments are not reproducible by other experimenters, or even by those who report them. While some mechanism might produce small bursts of hot fusion (e.g. high voltage internal fields associated with fracture of the material at certain temperatures), the present experimental evidence is not readily reproducible, and the phenomenon does not appear to be related to cold fusion as postulated in the heat production experiments.

If there were such process as room temperature fusion it would require:

  1. a major enhancement of quantum mechanical barrier penetration, which has been extensively tested against measurements (such as the systematics of spontaneous fission and alpha radioactivity lifetimes and those of nuclear cross sections);

  2. drastic modifications of branching ratios in the D+D reaction; and

  3. if all fusion proceeded to 4He and lattice heat a hitherto undiscovered nuclear process.




AUTHORS Reference Neutrons per DD pair per seca Yield Normalized to Jones et al. [JON-1] neutronsb

Jones et al. [JON-1] 1x10-23 c 1
Mizuno et al. [MIZ] 5x10-23 5
Williams et al. [WIL]   <0.5
Alber et al. [ALBE] <4x10-24 <0.4
Broer et al. [BRO] <2x10-24 <0.2
Lewis et al. [LEW] <2x10-24 <0.2
Schriber et al. [SCH]   <0.2
Kashy et al. [KAS] <1x10-244 <0.1
Gai et al. [GAI] <2x10-25 <0.2
De Clais et al. [DEC]   <0.1

Assuming that neutrons are produced throughout the volume of Pd or Ti.
For comparison one watt of heat production by D + D fusion would correspond to 0.9x1012 in these normalized neutron yield units.
This fusion rate is reported by Jones et al. [JON-1] for run 6. The average fusion rate for all runs [JON-1] is a factor of 6 less.




AUTHORS Reference Protons per DD pair per seca Yield Normalized to Jones et al. [JON-1] neutronsb

Rehm et al. [REH] <4.0x10-23 <4.0
Schrieder et al. [SCH] <3.1x10-24 <0.31c
Sundqvist et al. [SUN] <2.0x10-24 <0.2
Ziegler et al. [ZIEI <1.2x10-24 <0.12c
Porter et al. [POR] <6.7x10-25 <0.07
Price et al. [PRI] <8.3x10-26 <0.008

assuming that particles are produced throughout the volume of Pd.
For comparison one watt of heat production by D + D fusion would correspond to O.9x1012 in these units.
Rehm et al. comment that the choice of the low-energy cutoff (e.g. 1 MeV) [ZIE] "restricts the emission angle of the protons with respect to the foil to a small cone representing only a few of the total solid angle." [sic] This effect seems to have been neglected in the efficiency calculations for the limits quoted by these authors.



BACK to Contents