Like Run 5, this run employed one of our own W cathodes, fresh 0.1M K2CO3 electrolyte made with ordinary distilled water, and the magnetic stirrer. The primary purpose of this run was to explore the "excess gas" phenomenon. The term "excess gas" refers to the gas produced by the cell in excess of the amount predicted by the current flowing through the cell. If the only mechanism for electrons to move through the cell is via the dissociation of water molecules, it can be shown that for each ampere of current flowing, a total flow rate of 0.19 cm³/sec of H₂ and O₂ gas (at 25° C) should emerge from the cell. In previous runs with this experiment we have observed about 2.5 times this flow rate emerging from the cell. What is the composition of the excess gas?

	Assuming that the excess gas was something other than additional H2 and O2, we elected to use an external recombiner to react the H2 and O2 to H2O and thus allow it to be condensed and removed from the gas stream. We then collected the remaining gas in an inverted graduated cylinder initially completely filled with water. From right to left in this photo you can see the calorimeter enclosure, a small Cu vessel that bubbles the gas through water to arrest flames heading back to the cell, a small stainless steel vessel containing Pd-coated alumina catalyst pellets, and the inverted graduated cylinder arrangement. At the bottom center is a temperature meter that monitored a thermocouple strapped to the outside of the recombiner vessel.
Results:
	The color legend and vertical scales for this plot are as follows: Pin (0-200 watts) Pout (0-200 watts) Tcell (0-100° C) Vcell (0-200V) Icell (0-5A) Tin (39-41° C) The horizontal scale covers 3 hours.

The run begins with the usual equilibration period. At 1.0 hours we started applying voltage to the cell. We ran at 7 volts for a few minutes and then raised the voltage to about 33 volts for 10 minutes. During this time we monitored the temperature of the external recombiner and observed the bubbling of gas into the graduated cylinder. After several minutes of brisk bubbling (cell current was about 4 amps) and little or no rise in temperature of the recombiner, it suddenly "kicked in" and, within two more minutes, the bubbling had completely stopped and the recombiner vessel was quite hot (~55° C). With the recombiner working and totally recombining the exit gas stream from the cell, we then raised the voltage to about 145 volts, where it remained for the rest of the electrolysis period. The cathode responded promptly and assumed the usual orange blackbody glow rimmed with numerous tiny white flashes.

For the next hour, excess gas bubbled slowly into the graduated cylinder while the calorimeter system reported that Pout was almost exactly equal to Pin...indicating no sign of excess heat.

Assuming that all of the 0.42 cm³/sec of gas leaving the cell was a stoichiometric mixture of H₂ and O₂ the lost power due to the caloric value of the gas would be 3.3 watts. During the more-or-less steady operation of the cell from hours 1.5 to 2.2, the average Pin was 122.5 watts and the average Pout was 121.2 watts, 1.3 watts lower. According to our gas measurements it should have been 3.3 watts lower. This is possibly indicative of a very small excess heat from this run but it is also quite possibly a result of calorimetric errors, not the least of which could be in the input power measurement.

	After the run was over, we made a clumsy attempt to collect some of the excess gas in the graduated cylinder. A small stainless steel flask was evacuated and then used to extract the excess gas from the graduated cylinder through a tube that was first inserted into the water and then up into the trapped gas pocket. As a result, the sample flask was filled with a mixture of air (from the sampling tube) and the excess gas. We then connected the sample flask to our vacuum system and, using a very fine needle valve, presented a steady stream of it to our residual gas analyzer (see photo), which is a relatively low-resolution quadrupole mass spectrometer.
	This is the spectrum from the collected gas. On the far left, at mass 2, you can see a strong H2 peak. At mass 28 and 32 we have the characteristic peaks of N2 and O2 in about the right proportions for air. The peaks around 18 are from water and the peak at 12 is carbon, which apparently comes from CO2, which is also visible in this spectrum at mass 44.
	This spectrum was taken with the needle valve closed so that none of the sample gas could reach the RGA. The only peaks remaining are the water peaks around mass 18. These peaks are always present in our vacuum system.

 Knowing that the gas sample was contaminated with air leaves us with the tentative conclusion that the excess gas was relatively pure hydrogen. Supporting this conclusion is the fact that, in our first series of experiments with this cell (May 1998), we observed a similar excess gas and determined at that time that the gas was flammable and burned like hydrogen burns.

Next we turned our attention to the spent electrolyte. We decanted the electrolyte from the cell vessel and then carefully washed and dried the dark metallic-looking residue. To our surprise it weighed only 0.005 grams.

	This photo of the dark residue was taken at 400x magnification and is displayed on your screen at about 800x magnification. The blobs are mostly 5-10 microns in size. This material appears to be identical with "swarf", the metallic residue that is produced by electrical discharge machining. Each electrical discharge melts/vaporizes a small amount of the metal away from the cathode and the molten metal solidifies into a nearly spherical blob that falls to the bottom of the cell. Prior to this run we weighed the W cathode (without TFE coating) at 2.2106 grams. After the run, we carefully removed the TFE coating and found that the used cathode weighed 2.1220 grams, a weight loss of 0.0886 grams of W.
Since there was only 0.005 grams of W swarf collected from the cell, where is the missing 0.0836 grams of W? Maybe the answer lies in the 35 cm3 of "excess" hydrogen collected in the graduated cylinder. Suppose there is a soluble W compound formed during the run and that the reaction that forms this compound releases hydrogen gas. The only viable candidate for this W compound appears to be tungstic acid (H2WO4) and it's not all that soluble. The reaction would have to be: W + 4H2O ® H2WO4 + 3H2 Therefore, 1/3 as many moles of W would be consumed as moles of H2 gas produced. For the 35 cm3 of room-temperature H2 gas produced, 0.088 grams of W would have to be consumed. This is very close to the observed loss of 0.0836 grams. If this hypothesis is correct, the 0.0836 grams of W must be in solution in the spent electrolyte. We used 120 grams of electrolyte so the spent electrolyte should be about 0.070% W by weight.
	We analyzed the spent electrolyte with our XRF system. As can be seen from the spectrum, there is an obvious W content. We measured a W La x-ray intensity of 0.0165 (relative to pure W). Using the mass-attenuation coefficients of W and O (the matrix was assumed to be oxygen) for both the W La and the 20 keV exciting radiation, we computed a sensitivity of 21.3 for this matrix. This yields a W concentration of 0.077% in the electrolyte, very close to the expected 0.070%.

Conclusions:

1. This run didn't show any significant excess heat.

2. About 5% of the W lost by the cathode ended up as swarf on the bottom of the cell. The remainder went into solution.

3. During incandescent operation, the cell produced about 2.5 times as much H₂ and O₂ gas as the average cell current would predict. In other words, excess dissociation occurred.

All comments, suggestions, and criticisms of this work are welcome. In particular we are interested in an explanation of the excess dissociation phenomenon. We also seek a gas-sampling strategy that will eliminate the air contamination encountered in this instance.

little@earthtech.org