Speaker
Description
As the world fusion program forges into the burning plasma era, it is clear that the successful operation of
a burning plasma will require conditions that simultaneously balance the needs of the power exhaust
challenge with a high performance pedestal and minimal core impurity accumulation. The interconnected
nature of impurity transport demands a robust, integrated understanding of impurity sources and transport
that can only be derived from multi-machine, core to edge investigation of impurities. As part of the
FY20 Joint Research Target, 8 datasets obtained from 3 US tokamaks (Alcator C-Mod, DIII-D, and
NSTX/NSTX-U) were used to perform coordinated research to study impurities from generation at
material interfaces to their eventual accumulation in the core. This analysis validated state-of-the-art SOL
modeling that demonstrates near-SOL impurity accumulation, explained the role of impurity ionization
and neoclassical transport in the recovery of high performance pedestals, validated gyrokinetic impurity
transport predictions across 3 devices, and demonstrated the ability of ECH to eliminate core impurity
accumulation, independent of Z and advanced confinement regime.
Collector probe measurements were used to validate state-of-the-art simulations illustrating the role of
convective transport in explaining near-SOL impurity accumulation. Data obtained in LSN discharges
only differing in the B×∇B drift direction were interpreted by 3DLIM, a new 3D Monte Carlo far-SOL
impurity transport code. Probe results were consistent with the long-hypothesized near-SOL impurity accumulation in the inner target facing (ITF) direction, which modeling suggests may only form when the
∇B direction is away from the active divertor, due to the presence of fast parallel SOL flows when the
∇B direction is towards the divertor. Interpretive modeling assuming purely diffusive radial transport (10
m 2 /s) yields reasonable agreement with experimental measurements of W deposition on the outer-target
facing (OTF) sides of the probe (Figure 1), while failing to reproduce the inner-target facing direction. In
contrast, a purely convective radial transport model (125 m/s radially outwards) is shown to
simultaneously agree with both sides of the collector probe’s deposition profiles - indicating a potentially
dominant role of non-diffusive process in the SOL independent of the ∇B direction studied and
representing the most convincing evidence of near SOL accumulation to date. Also on DIII-D, in-situ
spectroscopic measurements of W on were collected on DIII-D that show that near the outer strike point,
net and gross erosion of tungsten are nearly equal for conditions where the W ionization length is large
compared with its gyro-radius (l W >> ρ W ). These results were found to be well reproduced by analytic
models. In contrast, as l W decreases, the rate at which tungsten net erosion also decreases relative to W
gross erosion, is somewhat reproduced by a recently developed model which relates W net erosion to the
ratio of l W and width of the magnetic pre-sheath, indicating the importance of electric field effects in
regulating prompt re-deposition in addition to gyro-motion. As the spectroscopic measurements only
account for re-deposition from W + , it is concluded that re-deposition from charge states higher than W + is
likely also important in regulating the balance between high-Z sourcing, leakage, and re-deposition.
The first 3D edge modeling of boron impurity transport in DIII-D was also performed, and found that
changes in the balance of parallel friction forces in the SOL alter spatial boron distributions as a function
of upstream plasma density. EMC3-EIRENE simulations of impurity powder dropper experiments
indicate that high density operation favors boron fluxes directly to the inner target, while low plasma
density favors a more uniform B distribution. Simulations using a localized source of boron atoms also
indicate that the boron flux to the divertor surface demonstrates significant toroidal asymmetries, as
measured directly in experiments. However, the simulated spatial scales are not compatible with the
striations observed, indicating this effect cannot be entirely ascribed to the source localization and that
uncompensated error fields also play a key role.
Direct measurements of main-ions identified the role of ionization and neoclassical impurity transport in
the pedestal. Using charge exchange, coupled with measured electron and fully-stripped carbon densities,
were able to unveil the dynamics of inter-ELM evolution (Figure 2). Inter-ELM evolution was found to
occur in two phases: (i) rapid initial recovery of main-ion and electron density followed by (ii) slower
impurity and electron density buildup with primarily impurity fueling. A significant fraction of the
electron density recovery can be directly attributed to impurity ionization and influx. The temporal
dynamics of this process were modeled using the STRAHL code, and are consistent with the
establishment of an neoclassically driven impurity pinch early in the ELM cycle, which preferentially
favors inward impurity transport over main-ion fueling in these conditions. By flattening the electron and
main-ion density (by reduced recycling or increased opacity) it is predicted that the inward impurity pinch
can be reduced, eliminated, or even reversed. The findings are largely consistent with results from both
NSTX, NSTX-U and C-Mod, indicating the robust nature of the results.
Four cross machine datasets (Alcator C-Mod, DIII-D, NSTX/U) were used to validate leading transport
models and establish actuators for the regulation of core impurity accumulation in a wide range of plasma
conditions. DIII-D’s unique neutral beam configurations were leveraged to break typical correlations
between rotation and other plasma parameters, enabling a novel investigation into the role of rotodiffusion
in determining impurity peaking. In contrast to previous work on JET and ASDEX, the DIII-D results
indicate a negligible role of rotodiffusion and suggest that a balance of other pinch contributions likely
explain DIII-D impurity peaking. Multi-machine comparisons were able to identify robust conclusions
about the physical origin of core impurity transport. Near-axis transport is in good quantitative agreement
with NEO calculations nearly independent of both the regime and machine studied, whereas transport
outside of mid-radius appears turbulent in nature. A study of three large databases revealed correlations
with turbulent and neoclassical transport drives, including a robust peaking of impurities correlated with
peaked electron density. Linear and nonlinear gyrokinetic modeling (CGYRO) yields generally
reasonable agreement with experiment but points to regions of parameter space where the origin of
measured impurity transport is still not well understood. Using a dataset that captured a wide range of
advanced operating scenarios on DIII-D (high q-min, hybrids, QH-modes, etc.), it was shown that a proxy
for neoclassical peaking/screening (a/L n,e - 0.5 a/L Ti ) was an effective indicator of peaking independent of
both the operational regime and impurity charge, and a modest (~1.0 MW) Electron Cyclotron Heating
was capable of reducing or eliminating peaked profiles. These results increase confidence in our ability to
predict transport and control impurity buildup with central electron heating in future burning plasmas.
| Affiliation | MIT Plasma Science and Fusion Center, Cambridge, Massachusetts, USA |
|---|---|
| Country or International Organization | United States |
