Content area
Abstract
A time-of-flight diagnostic has been developed to measure the energy spectrum d(GAMMA)/dEd(OMEGA) of neutral deuterium atoms D(DEGREES) emitted from the PLT tokamak plasma in the energy range from 20 to 1000 eV. The neutral outflux is gated on a 1 (mu)sec time scale by a slotted rotating chopper disc, supported against gravity in vacuum by magnetic levitation, and is detected by secondary electron emission from a Cu-Be plate. Since the energy dependent detection efficiency is known from prior ion beam experiments, the diagnostic is absolutely calibrated, and quantitative particles fluxes can be determined with 250 (mu)sec time resolution.
The time dependence of d(GAMMA)/dEd(OMEGA) has been measured for various classes of discharges. During an entire discharge, d(GAMMA)/dEd(OMEGA) is monotone decreasing with energy at energies above 100 eV. Early in a discharge, emission is dominantly at low energy with the spectral shape characteristic of a single temperature source, with measured ion temperatures T(,i) as low as 18 eV at 5 msec from initiation. After 50 msec the spectral shape reflects development of a T(,i) profile in minor radius. During steady-state the l/e decay constant of d(GAMMA)/dEd(OMEGA) varies from roughly 100 eV at energies E < 200 eV to about 400 eV at higher energies 500 < E < 1000 eV. Consistent with the PLT ordering of neutral mean free path small relative to the minor radius, the plasma density is the dominant quantity determining the spectral shape of d(GAMMA)/dEd(OMEGA). Steady-state particle emission and average D(DEGREES) energy measured for ohmically heated discharges is 1-2 x 10('14) D(DEGREES) cm('-2) sec('-1) and 200 (+OR-) 50 eV, respectively, at the diagnostic position separated toroidally 100(DEGREES) from the limiter. Both low power RF (200 kW) and high power neutral beam (1.4 MW) auxiliary heating are observed to enhance the total neutral outflux by about 3, leaving the average neutral energy essentially unchanged. Response to RF occurs within 1 msec, in contrast to the 50 msec response time for neutral beams. Edge cooling due to intense gas puffing and injection of titanium flakes is reflected in the neutral atom energy spectrum. The power loss away from the limiter due to emission of charge-exchange neutrals in the energy range from 20 to 1000 eV is found to be a small fraction, less than 1%, of the steady-state ohmic heating power.
A common feature of all discharges is a burst of neutral atom outflux at initiation which is typically 25 times more intense than steady-state emission, as well as a similarly intense burst at termination. A model is presented which determines both the magnitude and time evolution of the start-up burst from toroidally symmetric volume processes driven by the torus fill gas. The good agreement with the measured emission provides a check on the absolute calibration of the diagnostic and also yields information on the toroidal localization of limiter-driven recycling. For the steady-state portion of the discharge, a model is presented in which the dominant recycling mechanism is reflection of the parallel ion flow to the limiter as energetic neutral atoms. The model predicts that the neutral atom outflux has a strong toroidal dependence, peaking at the limiter at a level several hundred times greater than the measured emission 100(DEGREES) from the limiter. The model prediction for the neutral outflux at the toroidal position of the diagnostic is in agreement with the measured value.
