Interpretation of electron cyclotron heating results in overdense plasma in Doublet III

The electron cyclotron heating (ECH) experiment on the Doublet III tokamak with inside oblique launch of the extraordinary mode has shown good heating even at very high densities, up to the operational density limit. According to ray tracing calculations for the high density discharges, the ECH waves are reflected at the periphery of the plasma around r/a ∼ 0.9, and bulk plasma heating is not expected theoretically. But effective good heating has been observed in the increases of both the central electron temperature measured by Thomson scattering and the total energy measured magnetically. Furthermore, in the high density discharges, an extremely high density layer (marfe) exists in front of the antenna before the application of RF, and this layer persists for several milliseconds into the RF pulse. Since the marfe reflects the wave and does not allow its penetration to the bulk plasma, heating is not expected within the framework of standard wave propagation theory. A wave focusing phenomenon is introduced to explain the wave tunnelling through the marfe, using a localized collisional heating combined with the constraint of constant pressure along the magnetic flux line. After the marfe has been removed, after several milliseconds from the RF onset, the power deposition for the bulk plasma appears to be at large radius, according to the soft X-ray diode array diagnostics. But whether the deposition is at about r/a ∼ 2/3 or at the very edge of the plasma has not been confirmed. Theoretically, the accessibility of the overdense plasma for the wave can be improved by introducing a modest level of low frequency density fluctuations into the ray tracing calculations. The damping rate is not, however, large in the inner portion of the plasma. The heating is predicted to take place at large radius, r/a > 0.8, either by heating at the gyroresonance layer after multiple wave reflections on the vessel wall or by a parametric decay instability which excites an ion acoustic mode. Thus, contrary to our intuitive expectation that the deposition might take place in the plasma interior, we are led to conclude that the observed good heating is, indeed, due to the edge heating. The gross energy confinement time τE evaluated at r/a = 2/3 is a factor of about 2.3 larger than the global value τE(a), because of the lack of ECH power deposition within that radius. Since more than 85% of total plasma kinetic energy is contained within r/a = 2/3, this edge heating appears to be very effective in raising the bulk plasma energy.