Replication of Fauvarque, Clauzon, and Lalleve's Replication of the Mizuno Experiment
Ludwik Kowalski (a), Scott Little (b), and George Luce (b)

(a) Montclair State University, Montclair, New Jersey, USA.
(b) EarthTech International, Inc., Austin, Texas, USA.

12NOV05

INTRODUCTION

At the 7th International Cold Fusion conference in Vancouver in 1998, Ohmori and Mizuno reported¹ on a new electrolysis experiment in which a W cathode becomes incandescent under certain conditions. They made preliminary calorimetric measurements on cells operating in this mode and concluded that the phenomenon produced significantly more heat energy than the electrical energy required to stimulate it. 

During 1998-2000 an effort was made at Earthtech to replicate their experiment².  Despite having the full cooperation of Dr. Mizuno, including his providing several W cathodes to us, all of our experiments failed to show excess power production.

Jean-Louis Naudin has peformed his own version of this experiment and has published extensive and detailed reports³ of his work on his website. Naudin's results indicate substantial excess power production with power output/input ratios typically around 1.6 but sometimes exceeding 2.5.  In July 2003 Earthtech mounted a  campaign⁴ to replicate Naudin's results.  Although we succeeded in reproducing some of his numerical values, we also demonstrated that they were a result of erroneous data manipulation.   When correctly evaluated none of our experimental results showed excess power production.

Recently Fauvarque, J., P. Clauzon, and G. Lalleve reported⁵ positive excess power from their version of this experiment.  Fauvarque et al employed a fundamental approach to the calorimetry in which the heat power produced by the cell is determined by measuring the rate at which water is evaporated from the cell during operation at the boiling point.   With the cooperation of Ludwik Kowalski of Montclair State University, we now endeavor to replicate this work.

APPARATUS

The experimental setup is depicted below:

DC power at 0-400 volts is fed through a Clarke-Hess 2330 Power Analyzer to the anode and cathode in the cell.   A second DC power supply drives an ohmic heater immersed in the electrolyte.   The cell sits atop a balance for realtime weighing.   A reservoir filled with water provides automatic refilling of the cell during a run.

The output of both power supplies, cell temperature, reservoir temperature, room temperature, and balance reading,  are monitored by the data acquisition computer.

Details of the equipment used in this experiment are available upon request from Earthtech.  Of specific interest, however, is the Clarke-Hess power analyzer.  More information on this versatile instrument is available at http://www.clarke-hess.com. 

To avoid the possibility of breakage during operation, the cell was constructed from a commercial Lexan (polycarbonate) vessel.   We fabricated a phenolic cap (shown in the following photographs) for this vessel that provided a secure mounting for the anode, cathode, temperature probe, refill tube, and steam vent pipe.   A baffle inside the cell just under this cap prevented electrolyte from splashing directly out of the cell.   To permit accurate measurement of the weight of the cell during operation, all the wires leading to the cell were carefully festooned from nearby fixed mooring points so that the fraction of their weight bearing upon the balance would be stable.

This photo shows the arrangement of the cell, balance, reservoir, ohmic heater power supply (blue) and data acquisition computer.   Note the wires leading to/from the cell from the fixed terminal blocks on each side.

The thick polyurethane foam block under the cell provides much-needed shock absorption which allows the balance to function properly under the adverse conditions produced by this experiment.  Under the foam block is a space for the magnetic stirrer which was not present when this photo was taken (and not used in these experiments).

In this photo, the ohmic heater is operating.  The typical foamy appearance of the 0.2M K₂CO₃ electrolyte can be seen.

The second photo shows the main Variac for control of the plasma voltage, the Clarke-Hess 2330 Power Analyzer and the 4:1 step-up transformer used to obtain the required high voltage.  The rectifier and filter capacitors are just behind the cell.

The third photo shows the cell with the Lexan vessel removed.  All the cell components are secured to the machined phenolic lid. At the bottom is the ohmic heater in the form of a stainless steel jacketed single-turn loop. Above the heater is the Pt-coated Nb mesh anode which is supported by two 3mm diameter Ti rods.  One of the rods is merely a structural support. The other rod provides support and the electrical connection to the anode.  Note that these rods are sleeved with Teflon heatshrink to limit the anode area to the submerged mesh. In the center of the cylindrical anode is the cathode assembly. A ceramic tube  protects the upper part of the 2.4mm diameter W rod from contact with the electrolyte.  The tip of the W rod (hidden by the anode mesh in this photo), typically extends about 13mm beyond the tube.  All upper parts of the cathode assembly are sleeved with insulating heatshrink tubing to prevent unwanted paths for the electrical current in the cell. To the right of the anode mesh is the thermistor temperature probe.  The thermistor is housed in a sealed glass jacket. Immediately to the left of the thermistor probe is the refill tube which leads up to the water reservoir atop the cell assembly. Just below the phenolic top is a thin Lexan baffle plate which prevents direct splashing out of the cell.  Three small holes on one side of this plate allow steam to rise above the plate and a larger hole on the opposite side of the lid allows the steam to escape from the cell.  It is a simple labyrinth baffle. The set screws that are visible protruding from the edge of the phenolic top serve to secure the top to the Lexan vessel.  Without these screws, the violent action of the high-voltage electrolysis plasma will easily blow the lid off of the cell.

This drawing shows all of the cell components schematically.

This photo shows the cell during electrolysis plasma operation.  Note the intense glow from the cathode.  The W rod is essentially white hot.

MEASUREMENT STRATEGY

Following the work of Favarque et al, we determined the heat output of this experiment by measuring the rate at which water evaporates from the cell.  The cell always operates at ~100°C and thus it takes 2260 joules to evaporate each gram of water from the cell.  In addition to evaporation, the cell also loses some heat via conduction and radiation.  As Favarque et al showed, this loss is essentially constant provided the cell is kept boiling.  Because the boiling produces strong convection currents in the cell, we elected not to use the magnetic stirrer (shown as optional in the first schematic above) in these experiments.

Prior to the electrolysis plasma runs, we used an ohmic heater to boil the electrolyte at various power levels to determine the heat loss value and to demonstrate that it is relatively constant. This graph shows the observed heat loss (watts) plotted against the heater input power (watts).   The average heat loss is 94.4 watts and the points exhibit a standard deviation of +/- 10 watts about this value.

Each run starts with a warmup period using only the ohmic heater to bring the electrolyte to ~100°C.  This is followed by one or more periods of plasma electrolysis and/or additional ohmic heating periods.  Once every second, the data acquisition system records all temperatures, voltages, currents, powers, and weight readings from the balance.  The weight readings are used in real time to calculate dm/dt, the rate of evaporation from the cell.  The computer then plots an "evaporation power" trace (called Pevap) which shows the thermal power being spent in evaporation.

The electrolysis plasma starts at minute 13.  In this case, the cell voltage was 350 volts (cell voltage, called Vplasma is a dark gray trace on a 0-500 volt vertical scale).   As can be seen from the plasma power trace (called Pplasma), the electrical input power is rather erratic during plasma operation.  But from 14-16 minutes the system is sufficiently well-behaved to obtain a good measure of the relationship between Pplasma and Pevap. The averaged data from that period appears as a single point in the results table below.   After 16 minutes, however, the conditions in the cell become so violent that shocks disturb the balance readings.  Note the resulting "glitches" in the Mass trace. Consequently, the Pevap trace becomes very erratic and unreliable.

The current through the cell also produces electrolysis of the water.  The resulting H₂ and O₂ gas is allowed to escape.  This escaping gas represents an energy flow but it is a negligible fraction of the total input power.  Typical cell current is 2-3 amps so the resulting "gas power" escaping from the cell is only 3 - 4 watts.

RESULTS

A total of 6 runs involving plasma electrolysis were conducted.   The graphical record of each of these runs is included in the appendix to this report but the results are summarized in numerical form below.

Each run includes at least one ohmic heating control period which serves as a fresh check on the heat loss value.  The data from these ohmic periods are treated just like the plasma periods and a Coefficient of Performance (heat output power divided by electrical input power) is calculated for them. 

All of the runs were conducted using 0.2M K₂CO₃ except for Run 11.  To reduce the power consumption and the violent shocks, Run 11 was conducted with 0.02M K₂CO₃ electrolyte.  Despite this reduction in concentration, the cell consumed over 300 watts and the W cathode achieved typical incandescence.

The average COP for all these runs was 1.004 with a standard deviation of +/- 0.026.  The lowest COP observed was 0.97 and the highest was 1.05.  The latter observation occurred while using the ohmic heater and therefore cannot be construed as nascent excess heat.

DISCUSSION

Favarque et al report observing COP's up to 1.41 in their experiments.   Clearly we have not replicated their results.   Among the possibilities to explain the differences between our results are errors in input power measurement and errors in heat output power measurement.

Accurate measurement of the electrical input power during plasma operation is complicated by the highly erratic current.  The average current drawn by the cell is 2-3 amps but the waveform peaks frequently exceed 10 amps.  In our initial efforts with this experiment we did not employ a filter capacitor between the Clarke-Hess and the cell.  As a result, these current spikes had to flow through the Clarke-Hess and we often observed the overload light flashing, which indicates a temporary saturation of the analog front end of the Clarke-Hess.   On a less sophisticated instrument such conditions might be overlooked and will lead to underestimation of the average power being delivered to the cell.   After we placed a second filter capacitor between the Clarke-Hess and the cell this problem disappeared.   It should be noted, however, that this capacitor in no way limits the magnitude of the current spikes drawn by the cell.  On the contrary, it provides a low impedance source closely coupled to the cell and therefore only enhances the magnitude of the spikes.

Accurate measurement of the heat output power in this experiment is critically dependent upon ensuring that the only way the cell loses mass is through evaporation of water.  If any liquid water is ejected from the cell, the resulting mass decrease will produce a false indication of heat production within the cell.  Because of the high specific heat of water (2260 joules/gram), a small amount of liquid water being ejected will create a large false positive signal.  For example, 0.1 grams of liquid water (i.e. 1-2 drops) leaving the cell every second will create a false positive heat signal of 226 watts.  That signal would make a nominal 500 watt run look like it was producing 726 watts, for an apparent COP of 1.45.   Most of the difficulty we encountered with this experiment involved preventing liquid water from being ejected from the cell, especially at higher cell voltages where the electrolyte sloshes around rather violently.  Our first cell designs were rather inadequate in this regard.  It was not until we created the highly closed cell described in this report that we succeeded in largely eliminating the liquid ejection problem.

There is always the possibility that our results do not show excess heat because we did not perform the experiment properly.   If anyone reading this report has suggestions for making the cell perform like it did for Favarque et al, please don't hesitate to pass them on to us (little@earthtech.org).

REFERENCES

1  "Strong Excess Energy Evolution, New Element Production, and Electromagnetic Wave and/or Neutron Emission in Light Water Electrolysis with a Tungsten Cathode", T. Ohmori, T. Mizuno, p. 279, Proceedings of ICCF-7, Vancouver, B.C. April 19-24 1998

2   See numerous reports of our efforts on the Mizuno experiments at www.earthtech.org

3   Jean-Louis Naudin's website is http://jnaudin.free.fr

4   http://www.earthtech.org/experiments/Inc-W/2003/replicationjln.htm

5   Fauvarque, J., P. Clauzon, and G. Lalleve, Abnormal excess heat observed during Mizuno-type experiments. 2005, Laboratoire d'Electrochimie Industrielle, Conservatoire National des Arts et Métiers: Paris.  Available at  http://www.lenr-canr.org/acrobat/FauvarqueJabnormalex.pdf

APPENDIX

During each run, measurement data was plotted on the computer screen and also saved to a disk file. Graphical illustrations in this report are taken from the screen images. Data analysis and tabulation was obtained by spreadsheet handling of the disk file data. The "time interval" heading in the above table indicates the period from which the performance data was taken and averaged. Those periods may be located in the appropriate graphs shown below

RUN 6

RUN 7

RUN 8

RUN 9

Note the sudden drop in mass around 10 minutes into this run.  That was caused by excessive foaming of the electrolyte which forced about 30 grams of liquid water out of the cell.  Of course, the sudden drop in mass caused the Pevap trace to shoot off scale temporarily, as if some huge power dissipation had suddenly evaporated those 30 grams of water.  Later at 14.5 minutes during plasma operation another very sudden mass decrease occurs.  This was caused by part of the PVC steam vent pipe falling off.  Again the Pevap trace shoots offscale.

RUN 10

Here we see another positive excursion in the Pevap trace during the ohmic heater part of the run (at about 9.5 minutes).  This was also caused by a visible ejection of liquid water from the cell but it was not nearly as large as in Run 9.

RUN 11

Again we see the effects of liquid water ejection from the cell around minute 5 in this run.