exist. Just too many unknown and, therefore, uncontrollable variables are present. However, there are quite a number of conditions that are now known not to produce positive results. Some of these insights are examined.

An important special condition that must exist for any of the nuclear reactions to be initiated is the presence of high, local concentrations of deuterium, as previously suggested. 1.295 This view is widely accepted. Such high concentrations have been proposed to be associated with regions of stress,264,296 to be enhanced by surface or bulk impurities, and to be associated with phase changes. 118 On the other hand, the deuterium concentration can be reduced by excessive microcracking or by normal hydrogen in the metal. 297 Generally, these high concentrations are expected to occur at or near the surface, especially during electrolytic charging or ion bombardment. These considerations as well as many observations suggest that tritium production occurs mainly at the surface for both titanium and palladium. Heat and neutron production are not so easily located in this respect. Nevertheless, many workers still propose that the heat reaction occurs within the entire volume of the electrode. Consequently, they report heat production as energy per cubic centimetre or energy per mole. This gives the impression of much greater energy release than has actually occurred from the small cathode volume used in the cell. This approach also underestimates energy density within the few active regions.

From the first, the deuterium-deuterium (D-D) fusion reaction was thought to be the most likely source of nuclear products, although other possibilities have been proposed. This reaction has two branches that have been measured at high energies, yielding an almost equal probability. These branches are

and

D+D = 'He (0.82 MeV) + n (2.45 MeV)

D+D=H (3.02 MeV) + (1.01 MeV). A third branch,

D+D = 'He + y (23.5 MeV),

is possible, but it has a very low probability in plasma reactions. Considerable confusion and disbelief has been created by the lack of neutrons, consistent with detected tritium and heat. Data now clearly show the following:

1. No experiment has detected sufficient neutrons or tritium to account for the reported heat. 298 Although neither tritium nor neutrons are usually detected during heat production, there is evidence that *He is produced. 165,222,228

2. While there is a clear relationship between neutron and tritium production, occasionally neutrons are produced without any tritium being found.

In order for these and other apparently conflicting observations to be explained, even to a limited extent, an important hypothesis is proposed. At least three nuclear reactions are proposed to occur within a metal deuteride lattice. One reaction produces the major source of heat, the second produces mainly tritium with a few neutrons giving a neutron-to-3H (n/3H) ratio near 10-9, and the third appears to produce mainly neutrons. Which of these operates at any time depends on the special conditions that exist at that time. Of course, more than one of these reactions could occur at the same time but at different locations in the metal. Additional reactions have been suggested, but the evidence is less compelling. Detailed experimental justification for this suggestion and a model that combines the various mechanisms and nuclear reactions is developed in this review. The experimental studies for heat, tritium, neutron, gamma-ray and ‘He production are discussed.

IL EXCESS HEAT PRODUCTION

Four major logic levels are associated with understanding the source of excess heat. The first level asks the question

1. Is the excess heat caused by errors in the
measurement?

This possibility is discussed in Sec. II.A. Should the answer be "no," the question at the next level is

2. What is the reaction that produces excess heat?

As the amount of observed excess energy increases beyond a certain level, the probability for a nuclear origin increases as well. At some level, there is no possibility of evoking a chemical source without major conflict with basic chemical experience. This experience shows that there is an upper limit to the energy that can be obtained from a chemical bond. Thus, there is an upper limit to the energy that can be obtained from the limited quantity of chemicals that exist in a cell. This assertion is discussed in Sec. II.B.

If the amount of heat is sufficiently large and nuclear products are found, the question becomes

3. What nuclear reaction produces the excess
energy?

Conventional fusion theory predicts a trivial fusion rate at room temperature. Consequently, if a significant fusion rate does occur, it must be caused by an

unexpected phenomenon involving special conditions that exist in a periodic array of metal atoms. Several theories have addressed this aspect but are not discussed in this review.

The question at the next logic level is

4. What is the new phenomenon, how can it be initiated in the lattice, and what nuclear reactions are influenced by the phenomenon?

At this level of acceptance, we are dealing with an approach normally applied to a mature science. As yet, the cold fusion field has not reached this level in some people's minds.

Efforts to attribute the heat to a chemical source have taken three forms

1. The calorimeter has a positive bias because energy accounting has not been properly made. 2. Energy is accumulated in some chemical form during the initial charging and released later. 3. An unknown reaction releases the energy. The latter source is proposed to be either mechanical or chemical.

II.A. Errors in Calorimetric Measurements

One needs to appreciate that the technique of calorimetry is very highly developed, and has been used with increasing accuracy for >100 yr. While an individual may make a mistake or an apparatus may fail in some respect, these problems are not common within the field of expertise. Nevertheless, all measurements contain some error. Error analysis depends on the type and design of the calorimeter used. However, in all cases, the errors can be divided into two groups: (a) random errors that relate to the precision of the various individual physical measurements needed to arrive at the excess heat and (b) systematic errors that occur because some energy has been ignored in the calculation due to design defects or unknown chemical reactions. Both types of errors are normally revealed by studying cells containing H2O as the electrolyte, using platinum as the cathode instead of palladium, or inactive cells containing both D2O and palladium. If such cells show no excess heat above that which can be accounted for by considering all energy entering and leaving the cell, the calorimeter is generally considered to be accurate to the degree that input power equals measured power.

The most important systematic errors involve energy that leaves the cell. When gases escape, there is an uncertainty in the chemical energy carried away by the resulting hydrogen and oxygen because of possible partial recombination within the cell. However, all of the workers reporting excess heat are aware of this problem, and, when tested, recombination is found to be

300

negligible at currents >30 mA/cm2 (Ref. 299) and as long as the electrodes are kept below the fluid surface. 139.143.1 Dissolved hydrogen and oxygen can be recombined on the anode, but the rate is trivial. 30 Furthermore, many workers now place a recombiner within the cell so that no gas need leave, and all of the chemical energy remains in the cell. An additional factor is the transport of D2O vapor out of the cell with the gases. 139.212 Heat associated with this loss is small when the cell temperature and current are low but become significant when both are large. This quantity needs to be included in the heat calculations when the highest accuracy is desired. Some LiOD + D2O(/) mixture is also carried away as an aerosol within the gas stream. This effect increases as the current is increased but is small under most conditions. Of course, none of these processes is important when a closed cell is used.

Energy deposited within a cell is in the form of an electric current and is calculated by multiplying the current passing through the cell by the voltage measured between the electrodes. This energy can be determined with very high precision provided both current and voltage are measured as a function of time. Once the cell is assembled and turned on, energy flow is monitored, and any subsequent energy entering the cell will be detected within limits determined by known errors. Although some workers have charged the palladium at low current without monitoring the energy during this time, most excess heat has been found after a complete accounting was made from the time the cell was first turned on.

Besides the electric current, chemical reactions and physical strain may add energy to the cell.

II.B. Chemical Sources of Excess Heat

Typical cells contain a very limited number of elements and amount of material. Therefore, the total possible excess mechanical or chemical energy contained in these materials or produced by interaction also must be equally limited. Only energy above the lowest possible energy states of the initial constituents of the cell is available to be added to energy that is generated within the cell. The only additional substance that is normally added is D2O to replace that which is lost by electrolysis from an open, unsealed cell. Deuterium oxide is generally acknowledged as being in its lowest chemical state under these conditions. Furthermore, all of the construction materials are chemically inert to each other when allowed to remain in contact over a long time. Therefore, for a chemical reaction to occur, some form of energy must be supplied from the outside. In this case, this energy is supplied by the electric current.

Only three current-induced chemical reactions are known to have the potential to affect the energy balance to a significant extent. These are the formation of

D2 and O2 gases from the heavy water (see Sec. II.E.3), formation of PdD, at the cathode with the release of O2 gas, and formation of Pd-Li alloy at the cathode by reaction with lithium ions. Peroxide formation has been proposed but not detected.

141

The formation of B-PdD by a reaction between palladium metal and deuterium gas 34.301.302 and chemisorption on internal cracks 303 gives off heat. This source of heat is discussed in detail by Godshall et al. An observed increase in cathode temperature during the initial reaction to form PdD and a decrease during removal of deuterium is consistent with expectations. 219,229 An initial cooling of the cell has also been reported 149,216,219 that is not caused by the Peltier effect. 148 However, as the palladium cathode becomes saturated with deuterium, this relatively small energy perturbation decreases, and the temperature of the cell returns to a baseline value.

Elemental palladium and lithium form an alloy with heat release. However, the reaction in a cell is between palladium and Li* ions in solution. This reaction requires energy that, again, is supplied by the flowing current. Consequently, a slight temperature reduction would result to the extent that such an alloy formed. This temperature reduction is not observed because the rate of formation is limited by the diffusion rate of lithium in palladium, a very small number.

213

In summary, any chemical reaction that requires additional Gibbs energy supplied by the flowing current to occur will appear as a loss of energy from the cell. Should such a normally unstable compound be formed but go unnoticed, a heat effect would occur when the cell current was subsequently decreased or stopped, thereby allowing the compound to decompose. Only one such heat effect has been reported, but at a level much too small to suggest this effect could be the cause of significant excess heat. Thus, critiques that explain the excess heat as being due to the formation of hypothetical compounds are not consistent with observation. 128.304 An analysis of possible chemical reactions was developed soon after the first announcement of the cold fusion effect and should be consulted for more detail. 3,305

307

The palladium can contain some strain energy that would be released during hydriding. 306 While the amount of this energy is difficult to quantify, it is expected to be small, and it would be released and measured only during the initial hydriding process. Excess heat lasting weeks has been detected even when fully annealed, strain-free palladium was used. Should strain energy be introduced and subsequently released during hydriding, no excess heat would be seen because the energy would be part of the measured heat balance within the calorimeter. Consequently, the proposed behavior of strain energy, either initial or induced, is not consistent with observation.

Another approach can be taken to put the chemical-mechanical source of heat in perspective. The

amount of excess energy reported can be compared to various chemical and physical processes. As much as 50 MJ/cm3 of palladium has been reported to be released during the life of a cell.212 This amount of energy is sufficient to melt 641 cm3 or 7690 g or 72 mol of palladium at 1825 K (Ref. 308), and its production would require 1700 g of hydrogen to be burned. If released during a short time, this energy would raise a typical cell (<1 kg of glass and heavy water) to > 12000 K. Of course, only a few cells have achieved this much excess energy production, and, fortunately, it occurred over many weeks. Nevertheless, these examples of common chemical and physical processes show how difficult it is to attribute this amount of excess energy to a nonnuclear process. The only other alternative is to argue that a large positive and variable bias exists in all calorimetry measurements giving positive results. This suggestion is not supported by any other behavior pattern in the measurements.

II.C. Examples of Experimental Results for Excess Heat Production

Table I compares the various reported heat measurements, both positive and negative, and lists the accuracy for the calorimeter when a value is reported. Each entry generally represents several examples of positive and/or negative results. The presence of excess heat was usually not claimed until the accuracy limit was exceeded. On average, the calorimeters that produced negative results have lower accuracy than those that gave evidence of excess heat. Nevertheless, they should have been sufficiently sensitive to see heat if it had been produced. The absence of heat is proposed to be caused by unfavorable conditions that existed in the palladium cathode. This aspect is discussed in Sec. II.E.2.

Although all of these results have potential importance, a few studies need to be examined in detail because they are unique. In general, the heat production rate reported in these studies is so large that dismissing the results as being caused by calorimeter errors is not reasonable. Furthermore, the heat was made under a variety of conditions so that if it is caused by a chemical effect, a variety of very energetic chemical reactions must be assumed. Such an assumption is very difficult to justify.