This website contains other cold fusion items.
Click to see the list of links

271) Are we now confirming French EE results?


Ludwik Kowalski; 3/16/2006
Department of Mathematical Sciences
Montclair State University, Upper Montclair, NJ, 07043


Introduction and tentative conclusions:
The controversy about the excess energy (EE), described in the item #270, remains unresolved. Let me tell you about a recent attempt to resolve it. The idea of repeating the experiment under the protocol used in Paris was discussed with Pierre Clauzon during the 12th International Conference on Cold Fusion (ICCF12) in Japan. He, and other researchers, including Mizuno himself, said that our kind of plasma was very different to what they were working with. About a months ago Pierre invited me to work with him at CNAM in Paris. Come to learn how to produce desired orange-red plasma, he wrote. My suggestion was that he comes to the US. A confirmation in another laboratory, I said, would be more convincing than repeating an experiment in my presence. That is why Pierre came to work in Richard Slaughter’s lab in Boulder, Colorado. He brought his own electrodes and his own potassium carbonate. During the first four days we had many difficulties. But on the day #4 the situation changed and we started confirming French results.

We learned how to recognize favorable conditions under which results become very reproducible. It is the end of day #8. Out of 41 experiments, performed at voltages ranging from 300 V to 350 V, only 6 resulted in the COP smaller than 1.1 (generating "excess heat" at the rates less than about 50 W). The mean COP and the standard deviation turned out to be consistent with a bell-shaped curve whose mean COP and standard deviation are 1.24 and 0.13 , respectively. Note that 1.24 is significantely higher than the COP at 300 V reported by the French team at the ICCF12. The histogram below shows 14 cases in which the COPs are consistent with what was reported for 350 V.

Why I am using the quotation marks for the "excess heat? " Because I am aware of possible interpretational errors, both in Paris and in Boulder. It is better to be prudent than sorry. Additional verifications will follow. So far we are happy that similar experiments produced similar results.

                                  *
                                  *     
                                  *     *
                                  *     *
                                  *     *
                            *     *     *
                            *     *     *
                            *     *     *
                      *     *     *     *
                      *     *     *     *
                      *     *     *     *
                      *     *     *     *     *
     _____*__________ *_____*_____*_____*_____*_____*______
      0.8   0.9   1.0   1.1   1.2   1.3   1.4   1.5   1.6   COP
It will be useful to invent labels for different projects. The French investigation we originally failed to replicate will be referred to as Paris-1 experiment. Our subsequent investigations, described in the unit #270, will be referred as Texas-1 and Colorado-1 experiments. Our French-American project, now in progress in Boulder, will be referred to as Colorado-2 experiment. I believe that project Paris-2 and Texas-2 will start after the end of the courent experiment. So far reality of excess heat in a setup similar to that in Paris-1 seems to be confirmed. But Colorado-2 results did not confirm that the COP at 350 V is between 1.3 and 1.4, as claimed in the Paris-1 experiments. The mean COP from five tests at 350 V was not significantly different from the main COP at 300 V.

My Internet messages (appended on 3/21/06):
In what follows I will show messages that I posted on the restricted Internet discussion list. The operational definition of the COP can be found in the appendix.

Message 1:
1) We have runs with the COP (coefficient of productivity) above 1.20 at voltages between 300 and 350 volts. But there were many traps on our path. Four days ago I was nearly certain that we are again not seeing excess heat. Here is one illustration. Two days ago we ended the first sequence of consistently large COPs (at least 10 runs). The next morning we worked for about 6 hours and every single COP was nearly one. Frustrated we started to think about all possible differences. Then we realized that the diameter of the W cathode was 2.4 mm yesterday and 3 mm this morning. With the 2.4 mm cathode we again had a sequence of consistently large COPs.

The essential thing is to create conditions under which the electric wattage and the electric voltage are not too different. For example, 350 volts and about 1 or 1.5 A. This happens when the cathode looks as a uniform yellow-orange cylinder and the tip of the cathode is nearly as yellow-bright as the sun. These details were supplied to us by Jean Louis Naudin who seem to be ahead of everybody in this game. According to Pierre, he has a setup that performs "on demand; in his laboratory." He turns the switch on excess heat starts coming, like after turning a bathroom heater. We are still very far from such comfortable situation. Perhaps Naudin will share some details about this device here.

2) Here is our protocol, more or less.
a) Our container is an ordinary two-liters beaker. The level of the K2CO3 electrolyte (0.2 M, as we had) is about 1.2 liters. The beaker is open and fresh hot water is added when the level fols to about 1.0 liters, or more often. There is no splashing at all (or very little, less than one gram for 50 grams evaporated. We do have splashing when power is too high at 300 or 350 W. Under such conditions the COPs are always close to 1.0. Under favorable conditions, on the other hands, we hear a steady sizzling roar. The surface of the electrolyte is surpassingly quite. We do not stir the liquid; intensive boiling takes place only between the electrodes, mostly near the very hot cathode. The thermometer, situated near the wall, shows the temperature of between 89 and 91 C.

b) Mizuno told me that the rate at which the voltage is increased progressively should be small. Here is how we arrive to favorable conditions:

aa) Apply 100 V for about 15 minutes.
bb) Apply 150 V for the next ~5 minutes
cc) Apply 200 V for the next ~15 min
dd) Apply 250 V for the next ~15 min
ee) Apply 300 V for about 60 min
ff) Apply 350 V and start measuring the COP (5 min per run)

Transitions from one voltage to another were also slow, typically in one or two minutes. At stage (bb) blue sparks become visible at the tip. At stage (cc) the entire cathode (initially about 20 mm below the ceramic tube) is orange-red. The inner diameter of the ceramic tube is larger than the cathode and we can push the cathode when what is left become ~5 mm or less. Allowing the cathode to become too short leads to large splashy ark-like booms.

We use distilled water but ordinary water is likely to be acceptable. At one point our beaker was broken and we used a plain glass container. The COP was 1.3 at 350 V. At present we are using a plastic container (the lower part of a large bottle in which the windshield wiper fluid is sold. It works very satisfactory.

What else is important? Our non-evaporative losses are between 66 W (in the now broken beaker) and 95 W in a wider glass container. The losses seems to be more or less proportional to the open area of the container. This indicates that what is lost via conductivity through the walls is only a small fraction of the total non-evaporative loss. This is confirmed by the fact that the plastic foam isolation of the wall does not result is a large reduction of the overall non-evaporative losses. The dependence of our non-evaporative losses, Pl, for our wrapped plastic container on the electric wattage, Pe was as follows:
Pe (W)     Pt (W)
=================
  406         53    measured
  688         73    measured
  848         84    measured
  260         30   extrapolated
  200         20   extrapolated
Non-evaporative loses are, very approximately, are about 10% of the electric power. A good indication of the potential success is to see a very small change in current (preferably going down) when the voltage is changed from 200 to 250 Volts. There is a lot of what we do not understand but at least we know what conditions should be avoided. We suspect that occasional white "explosions" (producing strong splashes) are due to ignitions of hydrogen bubbles. It is important to make sure that hydrogen does not accumulate below the anti-splashing shields. Fountains of water were often seen escaping from the ceramic tube containing the tungsten cathode. The inner diameter of the ceramic tube should be 2 to 3 mm larger than the tungsten cathode. That allows the cathode to be cooled by the rising electrolyte. A teflon disk above the ceramic tube, mounted on the tungsten rod, stops the rising fountains.

3) The radius of the anode does not seem to be important; we used Pierre's anode (diameter of about 4.4 cm) and Richard's anode (diameter 8 cm) at 300 V and obtained very similar COPs, about 1.2 or so. Pierre's anode is platinized Ti while Richard's anode is platinized Nb.

4)The rate at which cathodes are consumed are low when favorable conditions (low input power) are found. We do not even start measuring the COP when the current is higher than about 1.8 A (at 300 V). We reduce the voltage, wait a little and hope for the best. Sometimes favorable conditions (described above) are found rapidly but most often one has to make several attempts before finding them. Results are very reproducible under favorable conditions only. Our protocol does not guarantee favorable conditions. But we now know how to recognize unfavorable conditions, more or less. The COP is nearly always one when we see a lot of white light light and a lot of splashing.

5) The length of the cathode, sticking out of the ceramic tube, should be about 2 cm, initially. Richard has a nice way of pushing the cathode rod down manually, when necessary.

6) Unfortunately, frustration has not been eliminated; we still cannot say that a Mizuno-type experiment can perform "on demand." Looking for a reliable protocol should be part of future efforts. We are counting on help from other experimentalists..

7) Data collected under favorable conditions are in agreement with what was reported by Fauvarque et al. We constructed a histogram of the COP distribution for the run performed at 300 and 350 V. So far it has 33 data "bricks." It shows 24 results with the COP between 1.2 and 1.4, 3 results with the COP between 1.4 and 1.4, 10 data points with the COP between 1 and 1.2, and one result with the COP of 0.81 I hope this description will be useful to many.

8) Another detail worth mentioning is that ordinary tungsten cathodes perform as well as those that have 2% of Th added. This is good news for those who might wish to detect nuclear particles. And to those who might think about possible health effects from consuming too much thorium. Tungsten we use is from the welding supply store; it costs about $50 to buy a box with 10 rods. The fact that 2.4 mm rods work while the 3 mm rods do not work is an indication that even thinner rods (or wires) might perform better. The local strength of the electric field (V/cm) increases rapidly when the diameter becomes smaller. A thin wire will perhaps produce good COP plasma at lower voltages.

Message 2:
Here is how a "watmeter" was calibrated yesterday in our ongoing Mizuno type experiment. We essentially repeated a procedure developed in CNAM in Paris. The electrolytic cell (producing typical glow discharge current) was connected in series with a resistor of about 100 ohms and with a constant-voltage power supply. The average current was, for example, 2 A, depending on the total voltage. The UNIGORE 390 instrument (that we wanted to calibrate) was measuring the electric energy (in W*h) received by the resistor. That resistor was immersed in a second beaker containing boiling water. The function of the electrolytic cell, connected in series was to produce the randomly fluctuating current. The amount of water evaporated, for example, in 20 minutes, was measured at the same time. Knowing that amount, and knowing the rate of non-evaporative losses (conduction and convection)we were able to determine thermal energy generated during the same 20 minutes.

The difference between the thermal energy and the electric w*h better turned out to be 0.43%. Knowing the about 1% uncertainty associated with our determination of thermal losses we concluded that the electric energy measure with the instrument can be reliable at the level of about 1.5%. We are lucky that Pierre Clauzon, who developed the method, was with us, in Richard Slaugther's lab. We hope that the ongoing French-American cooperation will soon resolve the controversy that developed last summer. That controversy, by the way, was published in the form of two reports at the ICFF12. It is also described in the unit #270 at my CMNS website:

 http://blake.montclair.edu/~kowalskil/cf/

Note that the calibration was performed under the worst possible conditions; both i(t) and v(t) were variable (while the total voltage on the power supply was constant). Is there any reason to doubt that the uncertainty of less than 2% (determined as above) cannot be trusted? Comments will be appreciated. The small portable Unigore 390 multimeter, by the way, costs about $1000, I was told.

P.S.
In reading my own message (shown below) I see something that should be specified. The thermal calibration of an electric energy measuring device is valid provided this device is not used in thermal measurements. That was not exactly true in our case. Let me elaborate on this. The thermal energy generated consists of two components, heat lost via evaporation and heat lost by conduction and convection. Only a scale is used to determine the first component, which is dominant. But the second component was determined by methods at the same time, one with Unigore 390 and another with other instruments. I do not have the numbers to report at this time. But the results were nearly the same. Pierre just reminded me that the very same instrument was calibrated many times in Paris and results were also very satisfactory. Let me add that current experiments are done with the open vessel. More about this will be written later.  

Message 3:
1) Can it be that tiny droplets (primage in French) are responsible for the illusion of excess heat? During one experiment we placed two paper towels, of known weight, on the table next to our active container for exactly 2 minutes. Comparing the amount of water lost during that time with the tiny increase of the mass of the paper towel we concluded that no more than 2 +/- 1% of water was lost in the form of droplets that were large enough to hit the towels (smaller droplets evaporate in air).

In principle such tests should be perform during each experiment. But this was practically impossible. We plan to perform another test to address the same issue. What else is worth checking in anticipation of future criticism?

2) Here are five more data points for our histogram; all at 300 C.

COP=1.15, 1.19, 123, 1.19 and 1.21

3) We had only 5 points at 350 V (to stay away from possible troubles. The mean value is not significantly different from what one gets at 300 volts. But we do get the COP of unity at low voltages. This is a very strong argument that instruments are reliable.

4) About our tungsten -- in case somebody might be interested. It was manufactured by Osram Sylvania Inc., Towanda, Pennsylvania, 18848, USA.
The label on the box (purchased today for $43) is: 2 7033 275 53 3/32X7

Message 4:

Prompted by comments about tiny droplets, presumably escaping with the steam, we performed an additional experiment. It was our last test together; Pierre is returning to Paris tomorrow morning, Richard is going back to work tomorrow (after using 1/2 of his yearly three weeks of vacation) and I will start writing a unit about this "Colorado Project" for my CF website. As suggested by somebody on this list (sorry for not remembering the author), we took a very fine stainless-steel screen and covered the cell with it. The holes in our screen represented about 60% of the entire surface, the rest were wires. Each rectangular hole was about 0.2 by 0.2 mm. This would allow the steam to escape but would probably stop most of the drops (both small and large).

The COPs, measured at 300 W, while the cell was ~90%-covered by the screen, turned out to be: 1.20, 1.26 and 1.46. These were more or less the same values as for the open cell. If the excess heat, measured with the open cell, was apparent (due to droplets escaping from the cell) then the COPs would be lowered significantly when the screen was used. But this did not happen. We concluded that the excess heat was not due to droplets. A similar comparison was made with the ohmic heater. The COPs turned out to be the same (1.0) with and without the screen. In the case of the ohmic heater ~95% of the surface was covered by the screen.


Message 5 (posted by M.J. from France)
“I have another concern: non-evap losses surely depend not only on the temperature but also on the convection speed, couldn't this speed be higher for some reason when calibrating with the ohmic heater than during the run, leading to overestimating non-evap losses occurring during the run, and therefore overestimating excess heat? “ [That is a very good question. Paris-2 tests, now in progress, will probably provide the answer. The vessel used now is a thermos. The non-evaporative losses will probably be negligible. Will the COPs still be close to 1.24? We will see.]

Message 6:
1) In an earlier message I said the noise one hears from plasma, under favorable conditions, is "like from a machine gun." A more appropriate description is to say that it is like when something cold is placed on a large frying pan with very hot oil, for example, when frying an egg. That sizzling noise is not as regular as from a machine gun.

2) Reading messages posted today I was thinking about a branch of science called plasma physics. Most of us performing Mizuno type experiments are not event amateurs, as far as that discipline is concerned. Very sophisticated methods of plasma diagnostics have been developed in the last 50 years. Experts in plasma physics would have no trouble in the identification of the red-orange-sizzling-machine-gun-like plasma we were lucky to produce in the Colorado-2 tests. They probably have a short name for it. And they would have no trouble with creating a device generating excess heat on demand. The only person on our list familiar with plasma physics, as far as I know, is George Miley. I would like to know what he thinks about our Mizuno type experiments.

3) Yes, I know what many plasma physicists said about CF in the past. And I recall what Gene Mallove wrote about their selfish motivation. But we should be looking into the future, not into the past. It is silly that we are working in the darkness when powerful sources of light are available. How can we benefit from what is known to real plasma physicists? How can we find access to their sophisticated tools? How can we gain support from at least some of them?

Message 7:
Here is a little rough calculation worth sharing. In our typical 5 minutes test (input electric energy = 400 W * 0.0833 h = 33.3 Wh) the amount of excess heat is about 6.5 w*h (if the COP=1.2). This translates into 1.46*1023 eV. The estimated mass of tungsten lost during the same time is about 0.5 grams. That translates into 1.63*1021 atoms. Thus the excess heat per atom of tungsten seems to be close to 100 eV. That is about two orders of magnitude larger than what I associate with common fuels.

What happens to our spent tungsten? Part of it probably melts and then solidifies in the electrolyte. If so then pure tungsten must be found in the cell (in the deposits we see at the bottom, after some waiting). Suppose that 50 % melts and the rest combines with other atoms or ions. Not being a chemist I might be allowed to postulate that all possible chemical reactions are equivalent to one representative exothermic reaction. Suppose that our excess heat is due to that reaction. What would the Q value of the representative reaction be? It would be 178 eV. That is a lot !

The Q value would be two times larger if only 25 % of tungsten was used in chemical reactions while the rest was simply melted. This shows that Q increases when more tungsten is melted. At the other extreme the Q value would be 89 eV if nothing was melted (all consumed in exothermic chemical reactions). Even this number is unusually high. Did I make a wrong assumption, or a numerical mistake, somewhere?

I think it was Pierre (or was it Richard?) who suggested I make this kind of calculation. Fortunately, a possible 100% error (in the estimated loss of mass lost) would still support the conclusion -- the unusually high energy per atom of tungsten. In the future experiments cathodes should be weighted before and after individual tests. This would provide a more accurate result.

The “what if” fantasizing:
Let me assume, very optimistically that Scott Little, and other investigators, will confirm that excess heat is real. What should be done next to convince mainstream scientists that at least one so-called “cold fusion” phenomenon is worth studying? The answer is obvious -- it is to publish a good paper in a serious refereed journal. But editors of these journals are biased against cold fusion. They already know that our research is pesudo-science and do everything possible to block publications. How to overcome the bias of the editors? How to overcome the bias of the referees?

The first thing is to write the article very well; all possible objections should be anticipated and addressed. We should clearly state, in the introduction, that we have no evidence of nuclear origin of excess heat. The error made nearly 17 years ago should not be repeated. It is not difficult to imagine what would happen if Fleischmann and Pons did not speculate about chemically induced nuclear reactions. Instead of discussing theories scientists would focus on validity of experimental data. That would probably lead to subsequent confirmations and open the door for additional research to understand excess heat. Our paper should be strictly experimental and we should be honest about having no idea what causes the observed plasma anomaly. The term cold fusion should not even be mentioned.

We should also not rush toward publishing the paper. We should write and rewrite the draft, discuss is among the CMNS scientists and perform additional experiments, if necessary, to back our claim. Announcing the discovery too early was another big mistake of Flieschmann and Pons. It will probably take a year to check and double-check everything, and to collect supportive evidence from other researchers. We should publish consecutive paper drafts on the restricted discussion list for CMNS researchers. And we will ask them not to refer to our ongoing investigations. I am assuming that everybody on our list will respect our wishes. The paper should not be submitted before we have a setup which demonstrates excess heat in at least 75% experiments. That is why I think it might take another year before we are ready to publish the paper in a reputable journal. A journal should be carefully chosen to avoid editors who are likely to biased.

Appended on 3/22/06:
In a message sent to Richard this morning I wrote: “I think that the ‘Fantasizing’ part, in the unit #271, should be expanded. What will we do if Scott's results turn out to be negative? Please draft a paragraph about such eventuality. Then we will discuss it.”

His reply was: “I think it is very unlikely that Scott’s results will be negative. . . . But if they are then we must first assist him by phone, internet, photos and web video.  If we could set up the web video I really believe we can walk him to positive results. We should also be prepared to send him our anodes and cathodes.  Pierre’s assembly has produced the best results.  Could we get Pierre to build and test a second assembly and have it ready for Scott? It doesn’t have to be identical but should use the same material and have the same size for the anode.  Perhaps Pierre wants to build a new assembly and Scott could use the old one?  I will also build and test a second assembly but my results so far have been smaller than Pierre’s.  If Scott still can not get positive results then one of us has to go to Texas and assist him in running the experiment.  I would like to go but don’t think it will be possible, I only have so much vacation time remaining.  But I would be able to fly down for a three day weekend to assist.       I think that we should discourage Scott from publishing negative results until we all agree that something was wrong with the French-1, Colorado-2, French-2 and Italian experiments.  It would be best if we could discover exactly what was wrong before any publication. We want to avoid giving the field a black eye due to a defective experiment.  It we can fix the experiment that would be ideal if not then we should post the negative results and our conclusion as to why the experiment is flawed.  It might be best to just do this on your website and on the CMNS list as a discussion thread. CMNS members may have insight into fixing the experiment or challenge our conclusion.”

Well, I think that Richard’s strategy, about what to do if the Texas-2 results are negative, is correct. We will try to identify the cause of disagreement. But what if the cause is not found? How much time and effort should we invest before publishing a paper describing the situation? Perhaps it would be better, at some later time, to submit two papers to the same journal, our and his. Let us hope that Scott’s results will also show some excess heat.

Appendix:
Those who are not familiar with our jargon would probably appreciate if I explain what the COP (coefficient of productivity) is. Imagine a bathroom heater consuming electric energy at the constant rate of 1000 W and releasing heat at the constant rate of 1100 W. Such magic heater would have the COP=1.1; it would produce 10% more energy at the output (heat) than it received at the input (electricity). The term "magic" is used because, according to the low of conservation of energy, we expect a zero difference between the energy going in and the energy coming out. But suppose that such heater is invented. Would this be a device demonstrating that sometimes the energy is not conserved? Not at all. It would mean that some hidden source of energy exists and it is responsible for the unexpected excess heat.

The COP, in our setups, is simply the ratio of the electric energy received by the cell and the thermal energy released by it at the same time. Our COPs were calculated as Pt/Pe, where Pt is the average thermal power (rate of evaporative and non-evaporative losses of heat), while Pe is the average electric power (rate at which electric energy was used), both in watts. Note that hydrogen is produced in our experiments. That hydrogen is a fuel and burning it produces additional heat. The rate of hydrogen production is sufficiently law to neglect potentially useful chemical energy of hydrogen. The COPs would be only slightly larger than the Pt/Pe if generation of hydrogen were also taken under consideration.

This website contains other cold fusion items.
Click to see the list of links