Experimental Setup

We converted a 19 inch bell jar evaporator into a plasma discharge system (see Figure 1). A Nd-B-Fe nickel-plated cylindrical magnet, 1cm radius and 1cm thick with surface field approaching 5 kG, was mounted on a alumina rod. Two high voltage tungsten electrodes protruding from alumina holders were mounted on a rotatable feedthrough. A tungsten electrode threaded through a thin alumina rod was mounted on a separate rotatable feedthrough to act as a single-electrode Langmuir probe. The magnet was electrically connected through a fine tungsten wire also in an alumina sheath. High voltage bias could be applied separately to either electrode, langmuir probe or magnet separately. In the first set of experiments when we were producing kilovolt electrons, fragments of a Sylvania Cool-White fluorescent tube were used glass side up as an X-ray detector, phosphor side up as an electron detector. A roughing pump was used to bring the pressure of the system down to a value of $\sim$ 15 mTorr. A needle leak valve was used to establish higher pressures by bleeding in either Nitrogen or Helium gas. Two 8-bit color digital camcorders were used to take the pictures or movies, which also gave us estimates on rates.

Three high-voltage power supplies ($\sim 5 kV$) were available, but with limited current capabilities ($\sim 5-10 mA$). Therefore we have limited our investigation to low current equilibria. An oscilloscope connected to the magnet provided time resolution of the fast transients that we observed, though we were unable to directly correlate the images with the oscilloscope trace in this experiment. The langmuir probe was stepped through a wide voltage range to characterize the temperature and floating potential of the plasma.

In the first experiment, the magnet was tied to ground through a 5 kOhm resistor and a DC glow discharge was produced by applying high voltage to the tungsten electrodes adjacent to the magnet. In a second series of experiments, we removed the plasma generating electrodes, and biassed the magnet directly with a shielded wire. The langmuir probe remained on a rotatable feedthrough, so that we could probe several locations in the resulting DC glow discharge. Several cylindrical magnets of different aspect ratio were super-glued onto an alumina rod to explore the importance of magnetic field topology.

In the third experiment, we placed the magnet on a hollow alumina rod but now connected to a rotating feedthrough. High voltage was applied at a sliding contact 1/3 the way up the rod, and an internal wire brought the voltage up to the magnet. For better visualization, we replaced the 19'' steel bell jar with an 18'' glass bell jar (and 20'' Lucite blast shield). We also employed a 10-bit CCD monochrome camera to resolve the faint plasma glows that were posterized by the 8-bit cameras.