When we applied a positive bias to the electrode, we began injecting ions into the system. As can be seen in panel one of Figure 3, ion injection is far brighter because inrushing electrons ionize the gas around the electrode and produce a feedback sheath effect, which in a Nitrogen plasma, is a bright pink corona around the electrode. This feedback meant that a great deal of current was drawn from the power supply, and often only a few hundred volts could be established before the current limit of our power supply tripped. In the following set of pictures with positive bias, we selected a high voltage, usually Volts, and pulsed the power supply by continuously resetting it. In this way we injected ions of relatively high energy into the magnetic field.
The top panel of figure 4 shows a Nitrogen plasma at 50 mTorr with electrodes pulsed at 600 and 1400 V. The bottom panel differs only in being at 300mTorr. Note that at 50mTorr, the ions can drift completely around the magnet, though the large mass of the N ion means that ions are trapped only very close to the magnet. So a distinct donut or equatorial distribution can be seen. Occasionally bright spots will flare up around equator, traditionally explained as due to the potential difference between the plasma torus and grounded magnet causing a feedback ionization or sputtering hot spot. Prominent in the 50 mTorr but absent in the 300 mTorr figure are hot spots on the top of the magnet (and presumeably, on the bottom as well). These are individual discharges lasting less than a second that are captured in the 40 s exposure. Not as well resolved (due to poor photographic technique) are the thin filaments from these discharges that form complete loops around the magnet. Since the nickel plated magnet is an equipotential, and the discharge is DC, such a filament should only be able to form if there are parallel potentials along the magnetic field lines, symmetric about the equator. This is what is expected for the QNC.
The discharges are more obvious to the naked eye, which seems to have a better transient response. We began to explore the parameters for these circular discharges. Simultaneous injection of electrons did nothing to enhance the effect, and may, in fact, suppress it. Reducing the pressure to 20mTorr also seemed to suppress, or ``blur'' the circular discharges. Pulsing the positive power supplies at higher voltage also had no effect, nor did spinning the magnet up to approximately 1000 rpm. Since the trapped plasma appears very close to the magnet, it seems that the gyroradius size is important because of wall effects. We also observed that the initial discharges obtained soon after closing up the vacuum chamber were more intense. We thought the discharge might have something to do with water, and reasoned that a less massive background gas might show us gyroradius effects, so we tried Helium and obtained the images described in the next section.