Electrode Injection

With the magnet grounded, negative bias applied to the electrodes induced a breakdown of the gas whenever the pressure was below $\sim 1 Torr$. The electrode injected electrons into the trapping region of the plasma, but other than illuminating the magnetic field lines of the trapping dipole field, little other unusual effects were observed as we reported in SS1.

More excitement ensued when we biased the electrode positive, thereby injecting ions into the dipole trap. In SS1, we were able to break down the gas and achieve ion injection only at high positive electrode voltages ($> 600 V$). The resulting DC discharges, however, tripped the low current power supplies within a second or less, preventing us from studying the phenomena in steady state. We found that by adjusting the resistance of the magnet ground connection enabled us to minimize the current draw of the power supply and thereby obtain a quasi-steady state. (See Figure 2).

In a very clean system, the glow discharge is very steady, however when the magnet is outgassing (either unintentionally due to the cyanoacrylate adhesive used to hold the magnet to the alumina rod, or intentionally as when we paint the magnet with silver paint) then the glow discharge is interrupted by lightning-like field-aligned discharges that appear to emanate from the equator of dipole field lines and terminate a millimeter above the magnet. Occasionally these discharge seem to cross the equator and follow a field line from pole to pole [#!Sheldon01!#].

The current carried by the magnet was monitored by using an oscilloscope to find the voltage across the grounding resistor. The voltage pulses corresponding to arcs observed in the chamber, indicated positive flowing current (ions discharging to ground) with 30 $\mu$s ragged pulses, best approximated by a rapid rise and exponential decay. When we disconnected the ground, the magnet floated at approximately 100 V, whereas the steady state injection from the electrode was maintained at 400 V. Averaging the pulse over 30 $\mu$seconds, gave us a current of 10 mA or a charge of 3e-7 Coulombs. Using 100 V as the charging voltage (the floating potential of the plasma under these conditions), we estimated the capacitance at 3 nanofarads. If the current is determined by the RC time constant of the system, then the 300 $\Omega$ resistor we used would give a 1 $\mu$second e-folding time which is approximately the decay time of the pulse.

Using the frequency of scope triggers as well as movies made with the digicam, we found that increasing the voltage on the probe increased the rate of discharges, whereas increasing the pressure decreased the frequency of discharges. In the discussion section, we develop a simple empirical model that seems to capture the physics of the discharges.