286) Colorado-2 in Boulder and over the Internet


Ludwik Kowalski; 3/23/2006
Department of Mathematical Sciences
Montclair State University, Upper Montclair, NJ, 07043

The first message in the “Mizuno type experiments” was posted on March 15. This was only one week ago. Today my In box has 150 messages with that subject line. That means about 20 messages per day, on the average. I did not have time to read all of these messages very carefully in Boulder. But I will do this now, before deleting them from the Inbox. That might take several days. Being retired I have a lot of time. It is good to know that deleted messages will be preserved in the archive of the CMNS list. The purpose of this unit it collect ideas inspired by these messages.

1) X1 asked “Have you tried other metals e.g. steel wire?” That would be an interesting project after a RoD setup (reproducible results on demand) becomes available. Naudin was able to generate excess heat using anodes and cathodes made from several different metals. Iorio’s cathode and anode were tungsten rods.

2) X2 suggested that we preserve the electrolyte for future examinations. We did this. he also wrote; “ You might consider purchasing some more of the same production lot and keep them intact in a safe place, in case they contain some particular impurity useful for the reaction.” I guess we would do this if our resources were not limited. Platinized mesh is expensive. Tungsten cathodes and potassium carbonate from the same supplier are probably identical enough. We would not start this investigation without what we learned from Naudin, and others, about good reproducibility. They said “do as we did and you should have similar results; they did not say that generation of excess heat depends on impurities. Only good chemists should deal with experiments whose outcomes depend on presence or absence of some contaminants.

3) JR wrote (Jed Rothwell gave me permission to quote anything he posts): “I think you need a more rigorous test to exclude the possibility of droplets, but Mizuno has done countless tests with closed cells from which droplets cannot escape, and he has found copious excess heat, so he has ruled out this possibility.” In one case we covered the cell with a fine stainless screen and still observed excess heat. But doing this only once is not sufficient. In future experiments the open container should always be covered with such screen. Only steam should be allowed to escape; liquid drops and droplets should not be allowed to splash out of the cell. Why not? Because this creates illusion of excess heat. Fortunately, under favorable red-orange plasma, the amount of water lost due to splashing was less than 2% of what was lost by evaporation. This can be tolerated when the COPs are larger that 1.2 or so. Test with excessive splashing should be stopped at once because they are indicative of unfavorable plasma.

4) As soon as I finished composing the above, a new message appeared in my Inbox. It was from Jed. Here is a quote from his message; I agree with it 100%. “Much can go wrong with this experiment. You should not declare victory until you have done it many times, using many different calorimeter types and configurations.” That is what should be done before we try to publish a paper. Today is 17th anniversary of the famous press release about the discovery of cold fusion. As indicated at the end of unit # 271, we should learn from negative consequences of premature publishing.

5) I am really happy that other researchers help us to identify potential errors and suggest methods to estimate them. The escaping invisible droplets was a good example. A message from Michel Jullian (MJ), that appeared in my Inbox at about the same time, as from JR, is another example. Below is his entire message (quoted with permission):

“Richard: Insulating layer not visible!? That layer I have been ranting about is nothing but the conspicuous floating foam of bubbles I can see on all the photographs, what else? Let me explain again my concern, hopefully more clearly: I suspect that when the liquid surface is calm (no heat-radiating splashes) this floating blanket of electrolytic gas bubbles must be a fairly good thermal insulator, much better in any case than the NON-insulation on the naked liquid surface which happily let radiated/conducted/convected heat escape during ohmic heater calibration (an ohmic heater makes no bubbles).

If you now consider that _most_ of the non-evap heat loss from an uncovered hot water tank is emitted by said liquid surface, there is no doubt that the non-evap heat losses will be significantly affected by a foam layer appearing due to electrolysis. They won't be 60W any more, but may be half of that, maybe less, in any case this error should be evaluated, if only to demonstrate that it can be neglected if such is the case. This is the point of the temperature decay measurement I was suggesting, not _instead_ of ohmic calibration, but _in addition to it_, so the error can be evaluated.

Does it make sense now Richard? (Please don't hesitate to criticize or ask for precessions) M.J.

PS Jed, the droplets question must be addressed indeed (a screen as used in Colorado-2 should be fine, two screens would be totally droplet-proof), but the missing mass due to dissociation is not a problem at all, on the contrary it leads to UNDERestimating the COP, as it is more than compensated by consumed dissociation energy (cf earlier posts in this thread by Scott Little and myself)”

It is a real pleasure to read such messages. They are very helpful. We are lucky to be part of a group of researchers (on a restricted Internet list) who are no less interested in transforming our protoscience into real science.

6) And here is another piece from JR. “Actually, the worst thing that can go wrong with this experiment is when the cell explodes. On the plus side, when Mizuno's cell exploded, it produced at least 411 times more than input energy, and probably several thousand times more energy than any chemical reaction with these materials could produce. It was anomalous -- no question about that! On the minus side . . . you can see why Mizuno has largely abandoned this line of research. If the cell can do this, I suppose there is no reason to think it cannot produce two or three orders of magnitude more energy. If that happens to you, you will, at least, die happy. Your last thought will be: ‘Hey, it worked!’ "

7) Let me summarize what has been stated in several messages today and yesterday. Being a retired teacher I want to do this in the form of a formal lecture.

A) Under ideal situation non-evaporative losses are zero. In that case Pt=Pe, provided there is no excess heat. And Pt>Pe when excess heat is generated. The Pe is the average electric power and the Pt is the average thermal power. The Pe can be calculated from the W*h received during a test and the Pt can be calculated from the amount of water evaporated during that test. Note that 1W*h=3600 J and that Pd=L*m/t, where L is the latent heat of water evaporation (2260 J/g), m is the mass of the evaporated water (grams), and t is the duration of the test (seconds). The COP is defined as the ratio Pt/Pe.

B) In reality the non-evaporative losses might not be negligible, for example, 10% of received energy might be escaping through the walls of the vessel (conduction, Pd), with the rising hot air (convection, Pc, and as the electromagnetic radiation (Pr). In other words, Pt=Pv + Pd +Pc+Pr. The COP is still defined as Pt/Pe but determining the Pv and Pe alone is no longer sufficient to to calculate the COP.

C) The sum of Pd, Pc and Pr is what we call the non-evaporative losses. How do we measure such losses? We do this by using the ohmic heater, for several different values of Pe, such as 400 W, 600 W, 800 W and 1000 W. The ohmic heater is inside our cell with the electrolyte. We know that no excess heat is produced in the ohmic heater. We measure two quantities, Pe and Pv. The always positive difference, between the Pe and Pv, gives us the rate of non-evaporative losses.

D) In a large vessel, and without stirring the electrolyte, the temperature is not uniform; it might be 100 C near the source (between the electrodes or near the coils of the ohmic heater) and considerably lower neat the walls, for example 75 C or 85 C, depending on the Pe, the size of the cell and other factors. That is why the (Pd+Pc+Pr) should be determined for several Pe. The dependence of (Pd+Pc+Pr) on Pe is called the "calibration curve." In the ideal case, the (Pd+Pc+Pr) is zero for any Pe.

E) Next comes an important ASSUMPTION --> The rate of non-evaporative losses (Pd+Pc+Pr), determined by using the ohmic heater, are the same as the (Pd+Pc+Pr) during the plasma electrolysis. That is where a systematic error can possibly be made in the evaluation of COP. Let me consider several specific situations.

a) In Paris-1 experiments the measured (Pd+Pc+Pr) was the same at each sufficiently large Pe. That indicates that the entire electrolyte was boiling. Boiling provided natural stirring and the temperature of the vessel wall was also 100 C, at any Pe. But was the entire electrolyte boiling during the plasma electrolysis? Pierre told me that it was, at least when their heater was turned on during the electrolysis (correcting for the input heat supplied by the ohmic heater is trivial when the W*h delivered to the heater are measured during a test).

b) And what about cases in which the ohmic heater was off during the electrolysis? The answer depends on the size of the vessel and on thermal isolation around its walls. In our case (a beaker whose capacity was 2 liters) the measured temperature of the electrolyte, near the walls, was below the boiling point, for example, 80C or 90C, depending on the Pe. That is why our (Pd+Pc+Pr) was not the same at each Pe.

c) Pierre also told us that many measurements of COP were made without turning the ohmic heater on. As far as I know, the temperature of the electrolyte, near the walls, was not measured. As in the Colorado-2 experiment the electrolyte was not stirred. I suspect that under such conditions the (Pd+Pc+Pr) was not the same at every Pe. Was the dependence of non-evaporative losses on Pe taken under consideration in calculations of COPs? A good referee is likely to ask such question? My question is slightly different. I know that the average COPs at 300 V and at 350 V were essentially the same as in Colorado-2. In Paris-1, on the other hand, the COPs were found to be voltage-dependent. Can this be due the unjustified assumption that non-evaporative loses are the same at every Pe? This question must be answered before we say that COP increases with the applied voltage, as stated in the Paris-1 paper at ICCF12.

d) One of the question asked by Michel had to do with the Pc (rate of convection losses). Is it possible that, for some reason, the Pc during the ohmic heater calibration is not the same as the Pc during the plasma electrolysis. An error in the determination of Pc would lead to an error in the evaluation of COP. A good referee is likely to ask this question, unless we anticipate it and discuss the possibility.

e) Another question asked by Michel had to do with Pd (rate of conduction losses). Is it possible that, for some reason, Pd during the ohmic calibration is not the same as Pd during the plasma electrolysis. An error in the determination of Pd would lead to an error in the evaluation of COP. A good referee is likely to ask this question, unless we anticipate it and discuss the possibility.

F) In the last case Michel speculated about a possible cause of differences in Pd. Suppose, he said, that a thin invisible layer of gas is formed over the electrolyte during the plasma electrolysis but not during the ohmic heater calibration. That would reduce the Pd (during the electrolysis) and would lead to the overestimation of COPs. What kind of experiment can be performed to rule out possibilities of significant errors postulated in (d) and (e) above? Jed Rothwell suggested that other ways of determining the Pt (average rate of releasing thermal energy) should be used, for example, a flow calorimeter or a bomb calorimeter. I think that using a bomb calorimeter is a better idea. The flow calorimeter calls for a closed vessel (with a heated recombiner etc.) in a constant temperature environment. That seems to be more demanding that a bomb calorimeter used to evaluate caloric contents of foods.

G) Another good suggestion made by Jed was to determine the (Pd+Pc+Pr) from the cooling curve. Consider a brick at 90 C. The temperature decreases exponentially according to the so-called Newton's law of cooling. Suppose its thermal capacity is 50 calories per degree. Also suppose that the temperature goes down at the rate 0.2 deg/s. That would mean that (Pd+Pc+Pr)=50*0.2=10 cal/s = 41.9 watts. At a lower temperature the rate of cooling may be, for example, 0.1 deg/s. That would be an indication that the non-evaporative losses become two times smaller. The thermal capacity of a vessel can easily be determined by using the ohmic heater. It is simply a matter of delivering a know number of w*h and measuring the resulting increase in the temperature.

Jed suggested measuring several values of (Pd+Pc+Pr) at low temperatures, such as 20 and 30 C, and extrapolating to the actual temperature of the cell's wall. I do not think that this is practical; the extrapolation uncertainty is likely to be unacceptable. But measuring the (Pd+Pc+Pr) at several higher temperatures, for example, between 80 and 90 C will offer a reliable extrapolation toward the wall temperature, such as 95 C. Water, however, is also evaporated below the boiling point and this fact must be accounted for, as I mentioned in an earlier message. Ideally the (Pd+Pc) determined from the cooling curve, and the (Pd+Pc+Pr) determined as in Paris-1 experiments, must be identical. In reality a discrepancy of 1 W, or so, would be acceptable because the (Pd+Pc+Pr) are usually close to 50 W. I am thinking about cases in which COPs are larger than 1.2.

The bomb calorimeter test and the cooling curve test must be performed before we are ready to write the paper. I would not be surprised to learn that such tests are being conducted right now in Paris-2 experiments.

Appended on 4/7/06:

A reader of this unit, David Dow, was confused by the paragraph F above. Here is my reply to him: “Dave, you probably confused the Pd, which stands for the rate at which thermal energy is lost by conduction, with the same symbol for the chemical element. A thin invisible layer of foam, suspected to be formed on top of the the liquid during the electrolysis, might decrease conductive losses. Presumably, no such additional layer is formed when the ohmic heater is used. Under such scenario the COP > 1 might be an illusion. I will now make a little change in the unit #286, to avoid another confusion. In fact I will probably add an appendix to that unit, but not today. Thanks for your help."