Once we were able to resume firing with the full pressure of decaborane, we started to solve our second main problem—the breakdown of the decaborane. Breakdown is the process, at the very start of the pulse, when the high voltage strips the electrons off the atoms, converting the gas to a plasma and allowing the current to flow.
With decaborane, we saw from the first shots that the breakdown was difficult. We saw this from two lines of evidence. First, there were large oscillations in the voltage and the rate of change of the current (fig 1 a and b). These were much larger than those for good, high fusion yield shots with deuterium(fig 1c and 1 d.). We could tell that these were associated with the difficulty of breakdown because they were very similar to those with high-pressure deuterium—around 40 torr—and we knew from both theory and observation that deuterium breakdown gets more difficult with higher pressure.
Figure 1. Oscillations early in the pulse indicate difficult breakdown which leads to no pinch (no large spikes late in the pulse). Top row is second shot with decaborane, voltage, left(a) and rate of change of current, right(b). Bottom row is voltage, right(c) and rate of change of current (d) of a fusion-producing shot with deuterium.
Second, we could observe the breakdown with trigger shots. In these shots, we used the 30-kA pulse from the circuit that send trigger pulses to the switches instead of the much large 1.8 MA current from the main bank. We saw that the breakdown was happening directly from the cathode to the anode, below the end of the insulator. That never leads to the thin sheaths of current that form filaments and drive the formation of plasmoids. Instead, the breakdown has to happen along the insulator, which is coated with a discontinuous layer of metal that channels the current (Fig.2).
Figure 2. In a good breakdown as in the deuterium shot on left, current flows along the insulator (bright circle) and leaves the anode tip dark (black circle.) But in the poor breakdown we started with using decaborane, shown on right, the breakdown went from the outer vanes directly to the anode. The different colors reflect the different gases used. These are trigger shots with only 30 kA of current, not the 1.8 MA of the main shot.
After the first couple of shots, we wondered if part of the problem could be that the boron was depositing on the insulator and anode. Since boron is an insulator, it could cover up the metallic spots on the insulator, preventing breakdown there. To test the hypothesis, we went back to firing with deuterium. Sure enough, the first deuterium shot also had high oscillations and did not produce any fusion.
Fortunately, we found that the deuterium shots themselves could clean off the boron coating, leading to deuterium shots that had high fusion yield. This was a big relief because we had wondered for years how we would clean the boron deposits off. It turns out that simple deuterium does the trick.
Subsequent boron shots on Feb. 14 and Feb. 21 showed that the oscillations decreased after deuterium cleaning shots. Indeed, a combination of this cleaning and hydrogen mixes reduced the oscillations by 75% of the way to our minimum goals: the oscillation levels where deuterium produces fusion reactions. As seen in Fig 3, we are making rapid shot by shot progress toward as goal of 25 KV or less oscillations. While extrapolations are risky, we feel confident that we’ll reach fusion conditions soon.
Fig 3. Oscillations are decreasing shot by shot, showing improving breakdown. The upper orange line shows the voltage oscillation amplitude in kV vs shot number. The lower, horizontal orange line is the level where we expect to get some fusion reactions. The upper blue line shows the rate of current change oscillations in kA/ns and the lower blue line is again the level where we expect to get some fusion reactions.