Though mitigating the adverse effects of a thermal quench (TQ) by impurity injection has been successful, the problem of avalanche runaway electron production in the post TQ still remains because the electric field greatly exceeds the critical field on account of the strong radiation cooling predicted by extant coronal radiation models. In our recent study of the trapping, or self-absorption, of spectral neon lines in a post TQ plasma, a remarkable result was found: The radiant power output in ITER could be dramatically reduced, as much as 50 to 100 below the coronal models currently used in many MHD codes. Our study was based on a systematic examination of the “escape factor” concept which we found can be determined in two ways, one of which by a special solution of the Biberman-Holstein integro-differential equation for excited level population in a two-level system, and the other based on a volume-averaged probability that a photon emitted will escape the radiating body (in this case a slab) without being re-absorbed. The effect of self-absorption can be conveniently treated by multiplying the Einstein spontaneous emission coefficient by the escape factor which depends on the line shape; and it decreases from unity as the plasma opacity, increases from optically thin to thick. We proved that in the extreme optically thick limit the spectral intensity at the surface of a body is given by the Planck (black body) spectrum, as is should be. The dramatically reduced cooling rate in the ITER post TQ plasma means that the electric field and the critical field can be made comparable provided a neon density of 10^15/cm^3 can be attained by massive particle injection, which in turn means that the avalanche runaway mechanism can be stanched. These analytical calculations were supported by PrismSpect radiation code calculations.

Transport simulations of boundary plasma are crucial to understand and model power and particle exhaust in fusion reactors, but remain computationally challenging due to a variety of plasma processes taking place in tokamak SOL. In particular, these simulations must include intrinsic and seeded impurities, ExB drifts and high resolution spatial meshes to model plasma detachment in complex geometry, e.g. the SAS divertor in DIII-D. Whereas UEDGE is a very robust time-implicit solver for boundary plasma transport capable of handling ExB drifts in complex geometry, its computational efficiency significantly decreases as the number of plasma species and the resolution of the meshes increase. To overcome this performance bottleneck, the evaluation of the Jacobian matrix and right-hand side of the plasma equations has been parallelized in UEDGE using both shared-memory Open Multi-Processing (OpenMP) and Message Passing Interface (MPI) while conserving the entire legacy of UEDGE. This parallelized version of UEDGE has been routinely used to simulate boundary plasma in DIII-D with closed and open divertor, showing a speed-up factor of 20 on the Cori cluster at NERSC with only 32 processors. Furthermore, the numerical validity of this parallelized version of UEDGE is guaranteed by the capability of an on-the-fly comparison between serial and parallelized routines during any UEDGE simulation. Taking advantage of these improved performances, UEDGE simulations of DIII-D boundary plasma are currently being conducted to analyze power dissipation in closed divertor. In addition, UEDGE simulations of ITER plasma are currently being setup to be compared to recent SOLPS-ITER simulations with ExB drifts.

The role of nonadiabatic electrons in regulating the hydrogenic isotope-mass scaling of gyrokinetic turbulence in tokamak fusion plasmas is assessed in the transition from ion-dominated core transport regimes to electron-dominated edge transport regimes. We propose a new isotope-mass scaling law that describes the electron-to-ion mass-ratio dependence of turbulent ion and electron energy fluxes. The mass-ratio dependence arises from the nonadiabatic response associated with fast electron parallel motion and plays a key role in altering—and in the case of the DIII-D edge, favorably reversing—the naive gyro-Bohm scaling behavior. In the reversed regime hydrogen energy fluxes are larger than deuterium fluxes, which is the opposite of the naive prediction.

Mitigation of large runaway current during plasma disruption is a critical issue for future tokamak devices. Simulation of relativistic runaway electron (RE) loss due to different forms of non-axisymmetric field for DIII-D, COMPASS and ITER plasmas, shows that RMPs from in-vessel coils are not effective for RE beam mitigation in DIII-D, but do produce finite (>10%) RE loss in COMPASS, consistent with experimental observations in above devices. The simulations utilized the MARS-F code coupled with the recently updated REORBIT module. However, an effective 100% loss of the post-disruption, high-current RE beam in DIII-D, does result in DIII-D from a 1 kG perturbation produced by a fast-growing n = 1 resistive kink instability. This RE loss is independent of the particle energy and the initial location of particles in the configuration space. The major reasons for this difference in RE control by RMP between DIII-D and COMPASS are the coil proximity to the RE beam and the effective coil current scaling versus the machine size and the toroidal magnetic field. For an ITER 15 MA scenario with pre-disruption plasma, about 10% RE loss fraction is also predicted, highlighting the role of the plasma response. For ITER, the applied RMP field, with optimal poloidal spectrum that maximizes the edge localized mode control, leads to REs being lost to the lower divertor region of the limiter surface within a narrow poloidal band. These results thus provide a quantitative assessment on the efficiency of various 3-D perturbation methods on the RE mitigation.