Basic investigations into physical phenomena of gaseous magnetized plasmas become more important as advances are made in magnetic confinement and control of fusion-grade plasmas. The Davis Diverted Torus (DDT), a relatively small research tokamak, or donut shaped fusion vessel designed to confine plasma with strong magnetic fields, operates to better understand the underlying principles which govern plasma generation and control. A major research topic being explored on the DDT is the beat-wave heating of plasmas, in which the plasma is heated by launching two high energy waves of slightly different frequency into the vessel so that they overlap and create in their interference a lower frequency "daughter wave." In order to understand and model this process fully, new diagnostics and computer acquisition routines must be used to handle the large experimental data sets generated during runs in as automated and as expedient a manner as possible.
Linda Wang, a fellow AWU summer student, and I performed various experiments to familiarize ourselves with the equipment and to better characterize the response of our systems to given electronic input signals. Successful at our early calibrations, we began to learn how to initiate a low density, low temperature filament plasma. A screen mesh filament in the vessel is heated by passing a current through the wires, and then electrons are boiled off the element by biasing this hot filament negatively2. The injected electrons cause ionization of the surrounding neutral gas, creating an electricity and magnetically responsive "plasma." Under our operating conditions, and using argon as our gas, we created plasmas with energies on the order of 5 eV, or roughly 55,000 K (approximately 100,000 degrees Fahrenheit). We proceeded to investigate dependence of filament emission on filament heater currents, on heater voltage, and on pressure in the vessel. Interestingly, we found that greater pressures (on the order of 100 microtorr) increased current at lower bias voltages but decreased the current at higher settings, while lower pressure plasmas (around 1 microtorr) tended to result in the opposite effect. These observations tend to have no grounding as far as we know in previous theoretical and experimental work, but may occur as increased densities inhibit higher voltage electron current but aid negatively biased ion current. At very low gas pressure our results obeyed the Child-Langmuir law, (Current) proportional to (Voltage)3/2.
Having completed filament and non-confined plasma studies, as well as investigations of filament behavior with a magnetic containment field initiated, we moved to using probes, variably biased to attract electrons or ions, downstream of the heating element. A top probe, or non-moveable metal rod protruding into the vessel, was used at first to measure current as a function of voltage and time. After completing these stationary experiments, we wired the system into a moveable 3-D probe and began to take measurements with the added variable of location in the stream. We attempted to optimize the filament heater voltage (at 25 V), filament bias voltage (100 V), and probe location (22 cm into the stream from the outside edge of the 29.8 cm diameter of the vessel) for maximum signal out of the probe, while being forced to stay within certain operating limits of our equipment. We had problems in particular with an amplifier saturating at around 2 volts of input, and had to adjust our settings accordingly.
Matrices on the order of 66,300 cells were processed in Macintosh Excel 4.0 and LabVIEW 2.2 to allow for organization of the data into useful form. These newly formed "time slices," or files of voltage and current arranged in rows at given times, allowed us to produce current versus probe bias voltage curves (I-V curves), correcting for background voltages and using signal averaging to help cancel the transient effects and noise (such as sixty cycle hum from the power lines). From these I-V arrays and curves we are able to generate fairly accurate values for the temperature of the plasma, its "floating potential," and its density.
Ultimate goals include automating the acquisition and processing as fully as possible. After the process is further refined to allow for greater mobility of the probe and processing of larger, 2-D time slice images of the plasma current (proportional to density and temperature), we hope to generate "movies" of plasma density variations through time. This technique will allow for accurate modeling of the plasma response to given inputs and operational parameters, and for careful investigation into the propagation of wave effects resulting from beat-wave heating of plasmas.
Acknowledgments
Thanks to R. Horton for all his expertise and assistance, and for guiding our experiments, to R. Evans for writing all of the LabVIEW codes that I would have labored over for months, to L. Wang for putting up with my jokes, and to J. Thomas and J. Wejnert for tolerating my use of all the computers. Special thanks to D. Hwang for hosting me and for daily reminders not to kill myself with the equipment. Thanks also to the entire AWU staff, on and off site, and especially to J. Gogel and R. Yaffe.