DEBATE BETWEEN DOUGLAS MORRISON and STANLEY PONS

We reply here to the critique by Douglas Morrison [1] of our paper [2] which was recentlypublished in this Journal. Apart from his general classification of our experiments into stages 1-5, we find that the comments made [1] are either irrelevant or inaccurate or both. In the article "Comments on Claims of Excess Enthalpy by Fleishmann and Pons using simplecells made to Boil" Douglas Morrison presents a critique [1] of the paper "Calorimetry of the Pd-D2O system: from simplicity via complications to simplicity" which has recently been publishedin this Journal [2]. In the introduction to his critique, Douglas Morrison has divided the time-scale of the experiments we reported into 5 stages. In this reply, we will divide our commentsinto the same 5 parts. However, we note at the outset that Douglas Morrison has restricted hiscritique to those aspects of our own paper which are relevant to the generation of high levels ofthe specific excess enthalpy in Pd-cathodes polarized in D2O solutions i.e. to stages 3-5. Byomitting stages 1 and 2, Douglas Morrison has ignored one of the most important aspects of ourpaper and this, in turn, leads him to make several erroneous statements. We therefore start ourreply by drawing attention to these omissions in Douglas Morrison's critique. Stages 1 and 2 In the initial stage of these experiments the electrodes (0.2mm diameter x12.5mm length Pd-cathodes) were first polarised at 0.2A, the current being raised to 0.5A instage 2 of the experiments. We note at the outset that Douglas Morrison has not drawn attention to the all important "blankexperiments" illustrated in Figs 4 and 6 or our paper by the example of a Pt cathode polarised inthe identical 0.1M LiOD electrolyte. By ignoring this part of the paper he has failed tounderstand that one can obtain a precise calibration of the cells (relative standard deviation0.17%) in a simple way using what we have termed the "lower bound heat transfer coefficient,(kR')11", based on the assumption that there is zero excess enthalpy generation in such "blankcells". We have shown that the accuracy of this value is within 1 sigma of the precision of thetrue value of the heat transfer coefficient, (kR')2, obtained by a simple independent calibrationusing a resistive Joule heater. Further methods of analysis [3] (beyond the scope of the particularpaper [2]) show that the precision of (kR')11 is also close to the accuracy of this heat transfercoefficient (see our discussion of stage 3). We draw attention to the fact that the time-dependence of (kR')11, (the simplest possible way ofcharacterising the cells) when applied to measurements for Pd-cathodes polarised in D2Osolutions, gives direct evidence for the generation of excess enthalpy in these systems. It is quiteunnecessary to use complicated methods of data analysis to demonstrate this fact in a semi-quantitative fashion. Stage 3 Calculations Douglas Morrison starts by asserting: "Firstly, a complicated non-linearregression analysis is employed to allow a claim of excess enthalpy to be made". He has failedto observe that we manifestly have not used this technique in this paper [2], the aim of which hasbeen to show that the simplest methods of data analysis are quite sufficient to demonstrate theexcess enthalpy generation. The only point at which we made reference to the use of non-linear regression fitting (a technique which we used in our early work [4] was in the section dealingwith the accuracy of the lower bound heat transfer coefficient, (kR')11, determined for "blankexperiments" using Pt-cathodes polarised in D2O solutions. At that point we stated that theaccuracy of the determination of the coefficient (kR')2 (relative standard deviation ~1.4% for theexample illustrated [2], can be improved so as to be better than the precision of (kR')11 by usingnon-linear regression fitting; we have designated the values of (kR') determined by non-linearregression fitting by (kR')5. The values of (kR')5 obtained show that the precision of the lowerbound heat transfer coefficient (kR')11 for "blank experiments" can indeed be taken as a measureof the accuracy of (kR'). For the particular example illustrated the relative standard deviation was~ 0.17% of the mean. It follows that the calibration of the cells using such simple means can beexpected to give calorimetric data having an accuracy set by this relative standard deviation inthe subsequent application of these cells. We note here that we introduced the particular method of non-linear regression fitting (of thenumerical integral of the differential equation representing the model of the calorimeter to theexperimental data) for three reasons: firstly, because we believe that it is the most accurate singlemethod (experience in the field of chemical kinetics teaches us that this is the case); secondly,because it avoids introducing any personal bias in the data treatment; thirdly, because it leads todirect estimates of the standard deviations of all the derived values from the diagonal elements ofthe error matrix. However, our experience in the intervening years has shown us that the use ofthis method is a case of "overkill": it is perfectly sufficient to use simpler methods such as multi-linear regression fitting if one aims for high accuracy. This is a topic which we will discusselsewhere [3]. For the present, we point out again that the purpose of our recent paper [2] was toillustrate that the simplest possible techniques can be used to illustrate the generation of excessenthalpy. It was for this reason that we chose the title: "Calorimetry of the Pd-D2O system: fromsimplicity via complications to simplicity". Douglas Morrison ignores such considerations because his purpose evidently is to introduce acritique of our work which has been published by the group at General Electric [5]. We willshow below that this critique is totally irrelevant to the recent paper published in this Journal [2].However, as Douglas Morrison has raised the question of the critique published by GeneralElectric, we would like to point out once again that we have no dispute regarding the particularmethod of data analysis favoured by that group [5]: their analysis is in fact based on the heattransfer coefficient (kR')2. If there was an area of dispute, then this was due solely to the fact thatWilson et al introduced a subtraction of an energy term which had already been allowed for inour own data analysis, i.e. they made a "double subtraction error". By doing this they derivedheat transfer coefficients which showed that the cells were operating endothermically, i.e. asrefrigerators! Needless to say, such a situation contravenes the Second Law of Thermodynamicsas the entropy changes have already been taken into account by using the thermoneutral potentialof the cells. We will leave others to judge whether our reply [6] to the critique by the group at GeneralElectric [5] did or did not "address the main questions posed by Wilson et al." (in the words ofDouglas Morrison). However, as we have noted above the critique produced by Wilson et al [5]is in any event irrelevant to the evaluations presented in our paper in this journal [2]: we haveused the self-same method advocated by that group to derive the values of the excess enthalpy given in our paper. We therefore come to a most important question: "given that DouglasMorrison accepts the methods advocated by the group at General Electric and, given that wehave used the same methods in the recent publication [2] should he not have accepted thevalidity of the derived values?" Stage 4 Calculation Douglas Morrison first of all raises the question whether parts of the cellcontents may have been expelled as droplets during the later stages of intense heating. This isreadily answered by titrating the residual cell contents: based on our earlier work about 95% ofthe residual lithium deuteroxide is recovered; some is undoubtedly lost in the reaction of this"aggressive" species with the glass components to form residues which cannot be titrated.Furthermore, we have found that the total amounts of D2O added to the cells (in some cases overperiods of several months) correspond precisely to the amounts predicted to be evolved by (a)evaporation of D2O at the instantaneous atmospheric pressures and (b) by electrolysis of D2O toform D2 and O2 at the appropriate currents; this balance can be maintained even at temperaturesin excess of 90 degrees C [7] We note here that other research groups (eg [5]) have reported that some Li can be detectedoutside the cell using atomic absorption spectroscopy. This analytic technique is so sensitivethat it will undoubtedly detect the expulsion of small quantities of electrolyte in the vapourstream. We also draw attention to the fact that D2O bought from many suppliers containssurfactants. These are added to facilitate the filling of NMR sample tubes and are difficult(probably impossible) to remove by normal methods of purification. There will undoubtedly beexcessive foaming (and expulsion of foam from the cells) if D2O from such sources is used. Werecommend the routine screening of the sources of D2O and of the cell contents using NMRtechniques. The primary reason for such routine screening is to check on the H2O content of theelectrolytes. Secondly, Douglas Morrison raises the question of the influence of A.C. components of thecurrent, an issue which has been referred to before and which we have previously answered [4].It appears that Douglas Morrison does not appreciate the primary physics of power dissipationfrom a constant current source controlled by negative feedback. Our methodology is exactly thesame as that which we have described previously [4]; it should be noted in addition that we havealways taken special steps to prevent oscillations in the galvanostats. As the cell voltages aremeasured using fast sample-and-hold systems, the product (Ecell Ethermoneutral, bath)I will give themean enthalpy input to the cells: the A.C. component is therefore determined by the ripplecontent of the current which is 0.04%. In his third point on this section, Douglas Morrison appears to be re-establishing the transitionfrom nucleate to film boiling based on his experience of the use of bubble chambers. Thistransition is a well-understood phenomenon in the field of heat transfer engineering. A carefulreading of our paper [2] will show that we have addressed this question and that we have pointedout that the transition from nucleate to film boiling can be extended to 1-10kW cmin thepresence of electrolytic gas evolution. Fourthly and for good measure, Douglas Morrison once again introduces the question of theeffect of a putative catalytic recombination of oxygen and deuterium (notwithstanding the fact that this has repeatedly been shown to be absent). We refer to this question in the next section;here we note that the maximum conceivable total rate of heat generation (~ 5mW for theelectrode dimensions used) will be reduced because intense D2 evolution and D2O evaporationdegasses the oxygen from the solution in the vicinity of the cathode; furthermore, D2 cannot beoxidised at the oxide coated Pt-anode. We note furthermore that the maximum localised effectwill be observed when the density of the putative "hot spots" will be 1/delta where delta is thethickness of the boundary layer. This gives us a maximum localised rate of heating of ~ 6nW.The effects of such localised hot spots will be negligible because the flow of heat in the metal(and the solution) is governed by Laplace's Equation (here Fourier's Law). The sphericalsymmetry of the field ensures that the temperature perturbations are eliminated (compare theelimination of the electrical contact resistance of two plates touching at a small number ofpoints). We believe that the onus is on Douglas Morrison to devise models which would have to betaken seriously and which are capable of being subjected to quantitative analysis. Statements ofthe kind which he has made belong to the category of "arm waving". Stage 5 Effects In this section we are given a good illustration of Douglas Morrison's selectiveand biased reporting. His description of this stage of the experiments starts with an incompletequotation of a single sentence in our paper. The full sentence reads: "We also draw attention to some further important features: provided satisfactory electrodematerials are used, the reproducibility of the experiments is high; following the boiling todryness and the open-circuiting of the cells, the cells nevertheless remain at a high temperaturefor prolonged periods of time (fig 11); furthermore the Kel-F supports of the electrodes at thebase of the cells melt so that the local temperature must exceed 300 degrees C". Douglas Morrison translates this to: "Following boiling to dryness and the open-circuiting ofthe cells, the cells nevertheless remain at high temperature for prolonged periods of time;furthermore the Kel-F supports of the electrodes at the base of the cells melt so that the localtemperature must exceed 300 degrees C". Readers will observe that the most important part of the sentence, which we have underlined, isomitted; we have italicised the words "satisfactory electrode materials" because that is the nub ofthe problem. In common with the experience of other research groups, we have had numerousexperiments in which we have observed zero excess enthalpy generation. The major causeappears to be the cracking of the electrodes, a phenomenon which we will discuss elsewhere. With respect to his own quotation Douglas Morrison goes on to say: "No explanation is givenand fig 10 is marked 'cell remains hot, excess heat unknown'". The reason why we refrainedfrom speculation about the phenomena at this stage of the work is precisely because explanationsare just that: speculations. Much further work is required before the effects referred to can beexplained in a quantitative fashion. Douglas Morrison has no such inhibitions, we believemainly because in the lengthy section Stage 5 Effects he wishes to disinter "the cigarette lightereffect". This phenomenon (the combustion of hydrogen stored in palladium when this is exposed to the atmosphere) was first proposed by Kreysa et al [8] to explain one of our earlyobservations: the vapourisation of a large quantity of D2O (~ 500ml) by a 1cm cube palladiumcathode followed by the melting of the cathode and parts of the cell components and destructionof a section of the fume cupboard housing the experiment [9]. Douglas Morrison (in commonwith other critics of "Cold Fusion") is much attached to such "Chemical Explanations" of the"Cold Fusion" phenomena. As this particular explanation has been raised by Douglas Morrison,we examine it here. In the first place we note that the explanation of Kreysa et al [8] could not possibly haveapplied to the experiment in question: the vapourisation of the D2O alone would have required~1.1MJ of energy whereas the combustion of all the D in the palladium would at most haveproduced ~ 650J (assuming that the D/Pd ratio had reached ~1 in the cathode), a discrepancy of afactor of ~ 1700. In the second place, the timescale of the explanation is impossible: thediffusional relaxation time is ~ 29 days whereas the phenomenon took at most ~ 6 hours (wehave based this diffusional relaxation time on the value of the diffusion coefficient in the alpha-phase; the processes of phase transformation coupled to diffusion are much slower in the fullyformed Pd-D system with a corresponding increase of the diffusional relaxation time for theremoval of D from the lattice). Thirdly, Kreysa et al [8] confused the notion of power (Watts)with that of energy (Joules) which is again an error which has been promulgated by criticsseeking "Chemical Explanations" of "Cold Fusion". Thus Douglas Morrison reiterates the notionof heat flow, no doubt in order to seek an explanation of the high levels of excess enthalpyduring Stage 4 of the experiments. We observe that at a heat flow of 144.5W (corresponding tothe rate of excess enthalpy generation in the experiment discussed in our paper [2] the totalcombustion of all the D in the cathode would be completed in ~ 4.5s, not the 600s of the durationof this stage. Needless to say, the D in the lattice could not reach the surface in that time (thediffusional relaxation time is ~ 10s) while the rate of diffusion of oxygen through the boundarylayer could lead at most to a rate of generation of excess enthalpy of ~ 5mW. Douglas Morrison next asserts that no evidence has been presented in the paper about stagesthree or four using H2O in place of D2O. As has already been pointed out above he has failed tocomment on the extensive discussion in our paper of a "blank experiment". Admittedly, theevidence was restricted to stages 1 and 2 of his own classification but a reference to anindependent review of our own work [10] will show him and interested readers that such cellsstay in thermal balance to at least 90 degrees C (we note that Douglas Morrison was present atthe Second Annual Conference on Cold Fusion). We find statements of the kind made byDouglas Morrison distasteful. Have scientists now abandoned the notion of verifying their factsbefore rushing into print? In the last paragraph of this section Douglas Morrison finally "boxes himself into a corner":having set up an unlikely and unworkable scenario he finds that this cannot explain Stage 5 ofthe experiment. In the normal course of events this should have led him to: (i) enquire of uswhether the particular experiment is typical of such cells; (ii) to revise his own scenario. Instead,he implies that our experiment is incorrect, a view which he apparently shares with Tom Droege[11]. However, an experimental observation is just that: an experimental observation. The factthat cells containing palladium and palladium alloy cathodes polarised in D2O solutions stay athigh temperatures after they have been driven to such extremes of excess enthalpy generation does not present us with any difficulties. It is certainly possible to choose conditions which alsolead to "boiling to dryness" in "blank cells" but such cells cool down immediately after such"boiling to dryness". If there are any difficulties in our observations, then these are surely in theprovince of those seeking explanations in terms of "Chemical Effects" for "Cold Fusion". It iscertainly true that the heat transfer coefficient for cells filled with gas (N2) stay close to those forcells filled with 0.1M Li0D (this is not surprising because the main thermal impedance is acrossthe vacuum gap of the Dewar-type cells). The "dry cell" must therefore have generated ~120kJduring the period at which it remained at high temperature (or ~ 3MJcmor 26MJ(mol Pd)-).We refrained from discussing this stage of the experiments because the cells and procedures wehave used are not well suited for making quantitative measurements in this region. Inevitably,therefore, interpretations are speculative. There is no doubt, however, that Stage 5 is probablythe most interesting part of the experiments in that it points towards new systems which meritinvestigation. Suffice it to say that energies in the range observed are not within the realm of anychemical explanations. We do, however, feel that it is justified to conclude with a further comment at this point intime. Afficionados of the field of "Hot Fusion" will realise that there is a large release of excessenergy during Stage 5 at zero energy input. The system is therefore operating under conditionswhich are described as "Ignition" in "Hot Fusion". It appears to us therefore that these types ofsystems not only "merit investigation" (as we have stated in the last paragraph) but, morecorrectly, "merit frantic investigation". Douglas Morrison's Section "Conclusions" and some General Comments In his section entitled "Conclusions", Douglas Morrison shows yet again that he does notunderstand the nature of our experimental techniques, procedures and methods of data evaluation(or, perhaps, that he chooses to misunderstand these?). Furthermore, he fails to appreciate thatsome of his own recommendations regarding the experiment design would effectively precludethe observation of high levels of excess enthalpy. We illustrate these shortcomings with anumber of examples: (i) Douglas Morrison asserts that accurate calorimetry requires the use of three thermalimpedances in series and that we do not follow this practice. In point of fact we do have threeimpedances in series: from the room housing the experiments to a heat sink (with twoindependent controllers to thermostat the room itself); from the thermostat tanks to the room(and, for good measure, from the thermostat tanks to further thermostatically controlled sinks);finally, from the cells to the thermostat tanks. In this way, we are able to maintain 64experiments at reasonable cost at any one time (typically two separate five-factor experiments). (ii) It is naturally essential to measure the heat flow at one of these thermal impedances and wefollow the normal convention of doing this at the innermost surface (we could hardly dootherwise with our particular experiment design!). In our calorimeters, this thermal impedance isthe vacuum gap of the Dewar vessels which ensures high stability of the heat transfercoefficients. The silvering of the top section of the Dewars (see Fig 2 of our paper [2] furtherensures that the heat transfer coefficients are virtually independent of the level of electrolyte in