A fluid modeling perspective on the tokamak power scrape-off width using SOLPS-ITER*

SOLPS-ITER, a 2D fluid code, is used to conduct the first fluid modeling study of the physics behind the power scrape-off width ($\lambda_q$). When drift physics are activated in the code, $\lambda_q$ is insensitive to changes in toroidal magnetic field ($B_t$), as predicted by the 0D heuristic drift (HD) model developed by Goldston. Using the HD model, which quantitatively agrees with regression analysis of a multi-tokamak database, $\lambda_q$ in ITER is projected to be 1 mm instead of the previously assumed 4 mm, magnifying the challenge of maintaining the peak divertor target heat flux below the technological limit. These simulations, which use DIII-D H-mode experimental conditions as input, and reproduce the observed high-recycling, attached outer target plasma, allow insights into the scrape-off layer (SOL) physics that set $\lambda_q$. Independence of $\lambda_q$ with respect to $B_t$ suggests that SOLPS-ITER captures basic HD physics: the effect of $B_t$ on the particle dwell time ($\sim$$B_t$) cancels with the effect on drift speed ($\sim$$1/B_t$), fixing the SOL plasma density width, and dictating $\lambda_q$. Scaling with plasma current ($I_p$), however, is much weaker than the roughly $1/I_p$ dependence predicted by the HD model. Simulated net cross-separatrix particle flux due to magnetic drifts exceeds the anomalous particle transport, and a Pfirsch-Schluter-like SOL flow pattern is established. Up-down ion pressure asymmetry enables the net magnetic drift flux. Drifts establish in-out temperature asymmetry, and an associated thermoelectric current carries significant heat flux to the outer target. The density fall-off length in the SOL is similar to the electron temperature fall-off length, as observed experimentally. Finally, opportunities and challenges foreseen in ongoing work to extrapolate SOLPS-ITER and the HD model to ITER and future machines will be discussed.