Parametric scaling of power exhaust in EU-DEMO alternative divertor simulations

Investigations of parametric scaling of power exhaust in the alternative divertor configuration (ADC) SOLPS-ITER simulation database of the EU-DEMO are conducted and compared to predictions based on the Lengyel model. The Lengyel model overpredicts the necessary argon concentrations for LFS divertor detachment by about a factor of 5-10 relative to the SOLPS-ITER simulations. Therefore, while the Lengyel model predicts that plasmas with acceptable divertor heat loads in EU-DEMO would exceed the tolerable upstream impurity concentrations by a large margin, there are several SOLPS-ITER solutions within an acceptable operational space. The SOLPS-ITER simulations indicate that, unlike assumed by the standard Lengyel model, there are significant heat dissipation mechanisms other than argon radiation, such as cross-field transport, that reduce the role of argon radiation by a factor of 2 to 3. Furthermore, the Lengyel model assumes that the radiation front is powered by parallel heat conduction only, which tends to lead to a narrow radiation front as the radiative efficiency increases strongly with reducing thermal conductivity. As a result, the radiative volume and total impurity radiation are suppressed for a given impurity concentration. However, the SOLPS-ITER simulations indicate that other mechanisms, such as cross-field transport, can compete with parallel heat conduction within the radiative front and increase the radiative volume. As a result, the standard Lengyel model provides a very pessimistic estimate for the necessary concentration for impurities with strong radiative capability in low temperatures around 10 eV, such as argon. However, parametric scaling relations are needed for fast scoping of the operational space without having to run a complex, numerical code, such as SOLPS-ITER. In these applications, the Lengyel model might still be useful to provide guidance on the relative scaling of the exhaust characteristic of the various points in the operational space. The aim of this work is to provide evidence to avoid overconfidence on these simple model predictions.