Run 2 (and Run 1) of our 2nd attempt to replicate the Mizuno-Ohmori Incandescent W Excess Heat Effect - 26MAY99

Run 1 was a failure because, despite repeated reminders to myself, I somehow connected the cell with the reverse polarity! I learned that the cell behavior is very different when the wrong polarity is used. The impedance is very high. The cell never drew over 0.5 amps, even at 150 volts. Apparently the anodic conditions at the W electrode produced an insulating oxide layer. However, apparently at odds with this hypothesis is the observation that, after trying to run the cell backwards for a while, the W electrode was very bright and shiny! It had gone into the cell showing a distinct, rather dark thin-film oxide layer. The Pt electrodes were not damaged visibly by this reverse electrolysis.

After correcting the polarity of the cell, I made an attempt to collect the data for the I vs V curve that M-O present in their papers. The results are similar but not identical. I believe the major difference between our curves is due to temperature effects in the cell. I found that electrolyte temperature had a strong influence (negative) on the current in the region below incandescence. The hotter the electrolyte, the lower the current was for a given voltage. This temperature effect is the opposite of that usually observed and may occur only at temperatures near the boiling point.

In this graph, the incandescent effect started at about 110 volts and was fully developed at 150 volts. As I turned the voltage down (purple line), I observed that the actual incandescence (glowing due to blackbody radiation) decreases steadily with decreasing voltage. However, a second source of light...a sheath of fine purplish sparks that cover the W electrode...continues to provide a glowing appearance until the phenomenon shuts off abruptly at about 75 volts.

This photo shows the experimental setup during Run 2. On the far right at the bottom of the stack is the 150V-10A regulated DC electrolysis power supply. Just above that is the Clarke-Hess 2330 Power Analyzer. Immediately to the left in the foreground is the laptop computer which controls everything and records the data. Behind it is the calorimeter system, which consists of a positive-displacement FMI metering pump and a Peltier heater/cooler that is controlled by the system computer to provide constant temperature water to the calorimetry heat exchanger. The cell, which is surrounded by the calorimetry heat exchanger, is enclosed in the wood-styrofoam box on the left. A small viewport in this enclosure, insulated with a multi-pane window, permits viewing of the cathode during the run. On the extreme left is the gas flow measuring apparatus. The gas bubbles into an inverted graduated cylinder and displaces the water therein. The coil of copper tubing on top of the electronics stack provides extra cooling capacity for the relatively high power levels encountered in this experiment.

Results:

 This plot shows cell temperature, voltage, current, inlet water temperature, and room temperature versus time during Run 2. The horizontal scale spans 6 hours and there are tick marks every hour. The vertical scales are as follows:

Cell temperature: 0-100C (10C/div)

Cell voltage: 0-200V (20V/div)

Cell current: 0-5A (0.5A/div)

Inlet water temperature: 35-45C (1C/div)

Room temperature: 25-35C (1C/div)

 Cell power was off for the first 1.75 hours to allow the calorimeter to equilibrate. During this period you can see the cell temperature level off at 40C, which was the control setpoint for the inlet water temperature. At 1.75 hours, about 30V was applied to the cell. This cause a cell current of about 2.7 amps and the cell temperature rapidly rose to about 65C. A 3 hours, the cell voltage was raise to about 140 briefly to initiate cathode incandescence and then lowered to about 115 volts where the cathode was not incandescent but covered with a sheath of fine purple sparks. This cell voltage was maintained until 4.4 hours when the cell power was turned off. At 5 hours, the calorimeter was again equilibrated and the cell temperature was back at 40C.

 This plot shows the measured electrical input power (Pin) and measured thermal output power (Pout) versus time. The time scale is exactly the same as in the first plot. The vertical scale is -20 to +180 watts (20W/div) and P=0 occurs at the first division from the bottom of the plot.

The first thing to note is that, during the "incandescence" period from 3.2 hours to 4.3 hours, Pin and Pout are essentially equal to each other. In other words, there is no sign of a large excess heat signal.

However, during the warm-up period (2.2 hours to 2.8 hours) Pout is noticeably lower than Pin. Much of this difference can be attributed to the caloric value of the escaping H and O during this high-current part of the run.

In this second version of the power plot, the Pout values have been adjusted by adding a quantity equal to 1.48 times the cell current. This should compensate for the caloric value of the escaping H and O.

During the steady part of the warm-up period (from 2.3 hours to 2.8 hours) the average Pin was 77.8 watts and the average Pout was 75.9 watts, an apparent heat recovery of 97.5%. At these power levels, previous experience has shown that the heat recovery for this calorimeter is typically 97-99%.

However, during the more-or-less steady part of the incandescence period (from 3.2 hours to 4.2 hours) the average Pin was 97.9 watts and the average Pout was 97.6 watts, an apparent heat recovery of 99.7%.

Is this difference in apparent recoveries significant and indicative of ~ 2% excess heat!?

It will be difficult or impossible to know for sure but one additional factor could certainly produce an uncertainty in the Pout/Pin ratio during the incandescence period:

This is the voltage observed across the internal current shunt in the Clarke-Hess during the "incandescence" period in Run 2. In this trace, the horizontal time base is 50 nanoseconds/div. The vertical scale is approximately 2 amps/div. Here you can see oscillations that are apparently around 100 MHz (the bandwidth of this oscilloscope).

This is the same signal viewed at 1 microsecond/div. Note that the average current of 1 ampere is now somewhat apparent.

This is the same signal viewed at 1 second/div.

With such a messy current signal, it should not be surprising if the accuracy of the Clarke-Hess power analyzer suffered a little. It is fortunate that the voltage is constant because all that is required is a good average value for the current in order to obtain an accurate measure of the power delivered to the cell. It is important to note that the analog current meter on the front of the DC power supply indicated about the same current (~ 0.95 ampere) as the Clarke-Hess during the "incandescent" period.

Therefore it is certainly possible that our cell was not producing any excess heat during the "incandescence" period. We must also note the possibility that the cell was producing excess heat and that the Clark-Hess erroneously reported the input power...at a value that happened to match the excess heat. However, it does not seem likely that the cell was producing a large excess heat (as reported by M-O) and that the Clarke-Hess error matched that excess precisely.

Cell Details:

This photo shows the cell after Run 2 ended. Bubbles are still clinging to the thick TFE sheath that protects the 1 mm W lead wire.

The W cathode has eroded significantly. It no longer has sharp corners. It is also quite dark in appearance.

The white thing behind the cathode and below it is the tip of the glass-jacketed temperature probe. The white stuff is thermally-conductive epoxy that is inside the glass jacket.

Note that the Pt wires leading to the Pt anode sheets have been covered with TFE tubing that extends down all the way to the Pt sheet.

The electrolyte is 0.1M K₂CO₃ made with distilled water and ACS grade K₂CO₃ from Alfa Aesar.

Gas Evolution Measurements:

During the warm-up period, when the current was 2.72 amps, we measured 0.54 cc/sec of gas coming from the cell. At room temperature and atmospheric pressure, one expects about 0.19 cc/sec per ampere of cell current so the expected gas flow rate for 2.72 amps is 0.517 cc/sec. The observed rate was therefore 1.05 times higher, which is typical because of water vapor that also emerges from the cell.

During the "incandescent" period, when the current was 0.948 amps, we measured 0.204 cc/sec of gas coming from the cell. This is 1.13 times higher than expected. The increase is probably due partly to increased amounts of water vapor (due to the boiling) and partly to hydrogen gas released from the H₂O as the W cathode is oxidized.

Comments may be sent to little@earthtech.org