Non-monotonic energy flux density profiles at the ITER divertor entrance
SOLPS-ITER simulations of the ITER baseline ($Q=10$) plasma boundary, with tungsten (W) targets and drift terms activated, exhibit a radially localised peak in the energy flux density entering the outer divertor, despite constant anomalous transport coefficients across the scrape-off layer (SOL). This peak occurs radially beyond the near-SOL energy flux fall-off length $lambda_Q$ and is associated with a strong parallel current at the same radial location. It is enhanced by radial $Etimes B$ flows near the target and by fast-reflected atoms from the W surface. The maximum target energy load is dictated by the location and magnitude of this localised peak, which is found to propagate to the target without any dissipation, even in the semi-detached conditions of these simulations. Consequently, partial detachment of the target plasma within the first $lambda_Q$ from the separatrix is textit{not} sufficient for tolerable loads radially further out. This is in contrast to simulations with neither drifts nor parallel current, for which partial detachment inside the first $lambda_Q$ textit{is} sufficient for tolerable loads radially further out. Equivalent simulations without parallel current and without drifts exhibit a maximum load of just 5.6 MWm$^{-2}$ at the outer target, compared to 12.7 MWm$^{-2}$ in their presence. Simulations with parallel current and drifts turned on predict a heat flux profile entering the ITER divertor which cannot be usefully described by a single width parameter $lambda_Q$, at least with the standard seeding and fuelling setup. Implications for simple models are discussed.