"Impurities and impurity transport in the spheromak SPHEX "
G. Cunningham

Summary:

It is still unknown whether the poor performance of spheromaks compared to tokamaks in studies of plasma sustainment and confinement is due to high levels of impurities in the spheromak plasma or if the problem is intrinsic to the process of magnetic relaxation. This paper presents the results of measurements of impurity influx and transport in SPHEX, a spheromak that is created and sustained by current injection from a magnetized Marshall gun. Experimental diagnostics include an Spex 1 meter spectrometer, a VUV monochromator, an Hα detector, a soft x-ray/UV detector which functions as a bolometer, a CO2 interferometer, and extensive magnetics diagnostics.

A simulation code is used to model the impurity transport, ionization, and emission. Electron density and temperature are assumed not to vary with time, and neutrals and ions have diffusion coefficients which are constant in space and time. Rate equations are solved taking into account collisional and radiative ionization and recombination.

A complete spectrum between 60 nm and 800 nm was taken. The dominant impurities are seen to be carbon and oxygen, despite the fact that the electrodes are predominantly steel. Some FeII lines are seen in the 200-300 nm range, but their intensities varies by more than an order of magnitude from shot to shot. The total radiated power measured by the bolometer does not vary between "high-iron" and "low-iron" shots, implying that Fe does not contribute significantly to the radiated power. Calibration of the VUV monochromator is accomplished by noting that various CIII line intensity ratios do not depend strongly on plasma conditions (53.8 nm / 97.7 nm and 117.5 nm / 97.7 nm are used--these vary by less than 30% for various temperature profiles between 15 and 35 eV).

Resonance lines from each of the major impurity ions are fit to simulation results to determine the diffusion coefficient, central electron temperature, and the impurity influx (the electron density and the parabolic form of the electron temperature profile are held constant). The SXR/UV predicted by the model is about one half of the observed signal, implying that either half of the radiated power is due to impurities other than carbon and oxygen, or the VUV spectrometer calibration is in error by a factor of two. This factor changes with plasma conditions, suggesting that the first hypothesis is correct. Power loss from the plasma occurs through a combination of radiation and the flux of hot particles onto the walls.

The experiments were repeated with a layer of titanium coating the inner walls of the flux conserver in an effort to reduce the impurity concentration. The signal from lower ionization states is significantly reduced; for example OIII emission is reduced by a factor of 50, while OVI emission increases. Comparisons to simulations yield a much higher electron temperature than before titanium gettering. At these high temperatures the CV 227.1 nm line becomes visible, leading to a temperature estimate of 53 eV. This estimate is confirmed by a measurement of the OVI 381.1 nm / 103.1 nm line ratio, which yields a best-fit electron temperature of 51.5 eV. This temperature is then used as a fixed parameter in simulations to determine the diffusion coefficient and impurity influx. The results suggest that emission from carbon and oxygen now account for only 15% of the total radiated power.

In conclusion, in the standard, ungettered condition, the impurity concentration in SPHEX is high, around 10%, of which at least 50% is oxygen and carbon. At least 1/2 of the oxygen influx and 2/3 of the carbon influx is recycling from the flux conserver wall rather than coming directly from the Marshall gun. The power loss can be completely accounted for by radiation. Application of titanium to the flux conserver reduces the electron density and impurity recycling, increasing the electron temperature and improving particle confinement. The results presented here suggest that high transport rates are not intrinsic to the spheromak configuration, but rather the impurity influx can be reduced by improved plasma facing materials.

Applications to my research:

This paper presents results of experiments that are very similar to my project at SSX. The finding that oxygen and carbon are the main plasma impurities in SPHEX agrees with our understanding of the SSX impurity concentrations, and the extensive discussion of spectra and individual carbon and oxygen emission lines will be useful to me as we look for additional lines to observe. The computer models used to calculate best-fit values of plasma parameters likewise closely mirror the approach I am using with PrismSPECT. The process of coating the walls of the flux conserver with titanium is somewhat similar to the helium glow discharge cleaning that we use in SSX--both decrease the concentration of impurity ions and increase the electron temperature. However, this paper does not explain why we see stronger CIII and CIV lines after glow discharge cleaning.