While a tenfold reduction in impurities—to about one impurity ion for 500 deuterium ions—is a good start, it is not what Lerner calculated would be needed to preserve the filaments, nor what theory and previous experiments indicate can be achieved. For that, a 50-100 fold reduction in impurities is required, or 5-10 times less impurity than that has been achieved so far.
An alternative source might be a very thin layer of tungsten oxide—too thin to be seen or removed during the electrodes’ cleaning. Tungsten oxide dissociates at 1970 C, far below tungsten’s vaporization point of 5500 C, so an oxide layer will be far more fragile. The oxide layer might well give rise to the tungsten in the plasma as well. If this is the case, repeated firing will burn the oxide layer off and impurities will fall.
Seems a plausible explanation. Eager to see next report now.