----
 

Charles Beaudette's Review of Wilson Critique
in Excess Heat: Why Cold Fusion Research Prevailed,
2nd edition, 2002

Excerpt from "Chapter 9: The Critics II"
Excerpt from "Chapter 14: Validation"
Excerptfrom "Appendix: The Wilson Critique"

 

from "Chapter 9: Critics II" (pages 117-118):
The Wilson Critique

A more aggressive and most revealing critique emerged from a group led by R. H. Wilson at General Electric Co., Schenectady, New York.* They submitted a paper criticizing the calorimetry work in the original article by Fleischmann and Pons.13 The Wilson paper was published in July 1992 and was followed directly by Fleischmann and Pons’s response.14

These papers were prepared in the proper manner that included peer review before the publisher accepted them. They got to the heart of the excess power question. Whether cold fusion was a science or not hinged on precisely what was in these three papers: is there, or is there not, an anomalous source of energy which appears as an excess of heat in the Fleischmann and Pons experiment? These papers were the proper battleground. The passage of three years had allowed for the critics to find themselves and their weapons, so that the two sides engaged each other properly fitted out with the necessary information.

An overview of the Wilson critique shows it to be quite limited. The authors found it necessary to qualify their comments with, “appear,” or “possible,” or “probable,” or “potential,” which reduces the value of the criticism. It is interesting that they stated, “. . . inadequate mixing within the cell does not appear to be a problem,” without further discussion.†

The Wilson team summarized their paper in the following words.

We evaluate the data and methods of Pons, Fleischmann and coworkers and, where sufficient data are available, conclude that they overestimate significantly the excess heat . . . While our analysis shows their claims of continuous heat generation to be over stated significantly, we cannot prove that no excess heat has been generated in any experiment.15

In their response, Fleischmann and Pons pointed out that the Wilson calculations still showed excess heat after taking into account their corrections, in one case at the 50% level, far above the uncertainty floor. The Wilson report was not negative. It was supportive in that there was still excess heat after all the criticism Wilson could muster.

The argument between Fleischmann and Pons and the Wilson group was over the manner of computing excess heat energy flow. Regarding the burst of excess energy shown in the original paper (see Figure 4.2), Wilson said, “the ‘burst’ data [Fleischmann and Pons] present is not greatly reduced by the corrections that we describe.” They also state that, “. . . the possible recombination of oxygen and deuterium within the cell is apparently eliminated...” So in three crucial areas, that of the recorded burst of energy, the uniformity of temperature within the cell, and the possible recombination of gasses, the Wilson critique supported the Utah chemists’ techniques and claims.16

Wilson’s report also supported the claim of the existence of anomalous power. The authors allowed that several of the cells still showed significant power even after their values were recalculated. In one cell, after Wilson’s recalculation, the power amounted to four watts per cubic cm. of palladium and the total amount for the run amounted to four megaJoules of energy. These quantities were beyond what chemical reactions can provide. That the Wilson team at GE did not follow up the Fleischmann and Pons defense with further analysis is a pity.

* It should be noted that GE had a financial incentive to reach a negative conclusion. They wanted to back out of a research contract.

 

† A précis of the report is included in the appendix.

 

14. Fleischmann, M., and S. Pons, "Some Comments on the Paper Analysis of Experiments on Calorimetry of LiOD/D2O Electrochemical Cells, R. H. Wilson, et al.," (Journal of Electroanalytical Chemistry, vol. 332, 1992), pp. 33-53.
15. Wilson, R. H., et al., "Analysis of Experiments on the Calorimetry of LiOD-D2O Electrochemical Cells," (Journal of Electroanalytical Chemistry, vol. 332, 1992), p. 1.
16. Ibid., p. 2.


from page 188, Chapter 14: Validation
R. H. Wilson

R. H. Wilson et al. at General Electric published a critique of the initial full length paper by Fleischmann and Pons which we discussed in Chapter 9, p. 117. Wilson comes into this court (of validation) as a reluctant witness, brought to the bar by the bailiff: he and his cohort insist there is no such thing as excess heat.

In 1991 Wilson et al. recalculated the cell performance as presented by Fleischmann and Pons to take into account what they felt were several technical oversights in the original paper. Wilson still found that one Fleischmann and Pons cell generated approximately 40% anomalous power compared to the power put into the cell. This amounted to 736 milliwatts. This level of anomalous power was more than ten times larger than the error levels associated with the data.

pages 357-360:
Appendix: The Wilson Critique

The following is a continuation of the Wilson, et al., critique from Chapter 9.

The considerable power levels that Fleischmann and Pons measured in their seminal article (July 1990) claiming anomalous power generation were reviewed in Chapter 4. During the following year R. H. Wilson, J. W. Bray, P. G. Kosky, and H. B. Vakil, of General Electric Co., Schenectady, New York, and F. G. Will, of the Department of Chemical Engineering, University of Utah, offered a substantial critique of that article.

The Wilson group was active in cold fusion experimental work from the beginning. Their paper was submitted for publication in June 1991 and accepted for publication that December. A copy of their manuscript would then have been sent to the original paper’s authors for preparation of a response. Fleischmann and Pons responded, and the two papers were published together in July 1992, in the normal manner of professional publications.

Wilson summarized his critique as follows.

We evaluate the data and methods of Pons, Fleischmann and coworkers and, where sufficient data are available, conclude that they overestimate significantly the excess heat. This is in part because in their calibration they did not include calculation of the change in input electrochemical power to the cell resulting from the calibration heater power. An additional significant overestimate of excess energy occurs when the calibration is made at cell temperatures above 60°C, owing to the increased evaporation of heavy water during the calibration.*
The Wilson critique had two purposes. It discusses Fleischmann and Pons’s seminal paper, and it reports on research with its own electrolytic cells, none of which produced excess heat. The concern here is with that part of the report that discusses the Fleischmann and Pons paper. (The question of the significance of failed experiments was evaluated in the Chapter 8, page 107.)

Wilson’s criticism of Fleischmann and Pons’s paper was limited to the following concern expressed in the paper’s abstract “. . . in their calibration they did not include calculation of the change in input electrochemical power to the cell resulting from the calibration heater power.” Their criticism of inadequate data treatment concerns the consequences of using the calibration pulse to determine cell performance. The impact is a reduction in the amount of calculated excess power reported in the original Fleischmann and Pons article. Wilson accordingly recalculates the generated power for the Fleischmann and Pons cells.

The recalculated excess power in one case amounts to minus 0.43 watts.** This result implies that there is a thermodynamic black hole in that cell which swallows 0.43 watts of power thus causing energy to disappear from existence. This unexplained disappearance is fully as remarkable as the original Fleischmann and Pons announcement where they claimed the unexplained appearance of power. The Wilson critique is incomplete on this point.

Fleischmann and Pons respond as follows to these criticisms of their calculations.

The central assumption in the paper by Wilson is that one can assume the systems to be in a steady state at the point in time at which they are calibrated . . . and at which the values . . . are to be evaluated. In point of fact there is no such steady state . . . as can be seen from . . . the paper by Wilson, . . . The magnitudes of [those] terms . . . are in fact comparable to those of the corrections . . . introduced in deriving the heat transfer coefficients . . .†

In other words, a principal assumption of the Wilson paper introduces an error equal in size to other significant corrections being proposed.

Consequently, Fleischmann and Pons note that the integral form of the differential equation must be used for computation, as follows,

It is well known in many fields of research that accurate values of the parameters of the differential equations which model the systems can only be obtained by comparing the integrated forms of the equations with the experimental data. (p. 38)

Fleischmann and Pons continue by cautioning on the limitations of this technique. The process of integration will smooth values over time thus reducing mathematical errors in the computed coefficients.

After identifying the various methods used to calculate results from the cell data as methods one through six, Fleischmann and Pons offer the following assessments of Wilson’s work.

Wilson does not deal with any of these evaluations: they regard Method 2, which was outlined . . . as “very complicated and very difficult to follow in detail.” However, this method, together with low pass filtering, . . . is the standard method of modern data processing. (p. 40)
The filters which have been used . . . take full account of the evaporation of the [electrolyte] . . . the assertion that we did not take this into account can be seen to be incorrect . . . We observe that the results of the independent investigation using filtering were presented to the group at GE during 1991; their omission of reference to this work shows that they also reject this method of data processing in addition to Method 2. (p. 40)

Clearly, the Wilson team had not achieved a review of calorimetric data reduction; they were not evaluating each of the possible ways to get the optimum information out of the cell data reading. It was up to Fleischmann and Pons to explain to the Wilson team what needed to be done.

Fleischmann and Pons provided a tour of the six computational methods and then got down to the nitty-gritty.

We are therefore reduced to examining the claim that the method put forward by Wilson, Method 5, provides an accurate means of evaluating [excess power]. The authors imply that as the results obtained by their Method 5 differ from those obtained by our own approximate method, Method 1, it is our method which must be judged to be incorrect. (p. 40)

Fleischmann and Pons’s conclusion is reached after several pages of equations and tables.

We conclude that [the two methods] are comparable but that they give the [heat] balance at different [times of the cell’s operating cycle]. (p. 44)

After rejecting the possibility of maintaining the electrolyte level perfectly constant, they promptly continue to a way of dealing with this intrinsic artifact.

The answer lies in making the [equation’s] term . . . part of the evaluation [calculation] and this in itself dictates the strategy that the whole [period of the variables] be fitted to the integrated forms of the . . . equations which model the calorimeters, i.e. it dictates the use of methods such as Methods 2 and 4. It is not surprising that such methods can give precise results as a matter of routine. (p. 47)

Fleischmann and Pons touch upon two summary items that are of interest to us. One concerns the general accuracy of calorimetry, and the other concerns the results of the Wilson heat generation effort.

The information on this issue [of calorimetric accuracy] which was contained in our original paper and in the related papers has been ignored by Wilson. They have also ignored the fact that we showed that it is possible to achieve at least 99% heat accuracy by the methods we have used . . . ; we have never claimed an accuracy better than ±1% or ±1 milliwatt whichever is the greater. (p. 47)

They gave the following assessment of Wilson’s recalculation regarding their cells’s performance.

. . . They also do not discuss the fact that even on the basis of their own evaluations[,] the excess [power] for a 0.2 cm diameter x 10 cm length Pd cathode polarized at 128 milliamperes for each square centimeter of surface area has reached [approximately] 50% of the [power] input after 15 days of [operation] . . . Presumably they believe that the errors have now reached 50% to explain away these effects? It should be noted the these [power] outputs are of the order 4 watts/cubic cm . . . and that over the duration of the experiment shown[,] the total [energy] released is of the order of 4 MegaJoules per cubic cm . . . which hardly lies in the province of chemistry. (p. 47)

Wilson does not mention their own recognition from their own calculations of the existence of excess energy as presented in the Fleischmann and Pons original paper. They are also unwilling to face up to the implication that the amount greatly exceeds what can be explained by measurement inaccuracies.

* Wilson, R. H., et al., “Analysis of Experiments on the Calorimetry of LiOD-D2O Electrochemical Cells,” (Journal of Electroanalytical Chemistry, vol. 332, 1992).

** Wilson, R. H., et al., “Analysis of Experiments on the Calorimetry of LiOD-D2O Electrochemical Cells,” (Journal of Electroanalytical Chemistry, vol. 332, 1992), p. 10.

† Fleischmann, M., and S. Pons, “Some Comments on the Paper Analysis of Experiments on Calorimetry of LiOD/D2O Electrochemical Cells, R. H. Wilson, et al.,” (Journal of Electroanalytical Chemistry, vol. 332, 1992), p. 38.