Divertor shaping with neutral baffling as a solution to the tokamak power exhaust challenge

Abstract

Exhausting power from the hot fusion core to the plasma-facing components is one fusion energy's biggest challenges. The MAST Upgrade tokamak uniquely integrates strong containment of neutrals within the exhaust area (divertor) with extreme divertor shaping capability. By systematically altering the divertor shape, this study shows the strongest evidence to date to our knowledge that long-legged divertors with a high magnetic field gradient (total flux expansion) deliver key power exhaust benefits without adversely impacting the hot fusion core. These benefits are already achieved with relatively modest geometry adjustments that are more feasible to integrate in reactor designs. Benefits include reduced target heat loads and improved access to, and stability of, a neutral gas buffer that 'shields' the target and enhances power exhaust (detachment). Analysis and model comparisons shows these benefits are obtained by combining multiple shaping aspects: long-legged divertors have expanded plasma-neutral interaction volume that drive reductions in particle and power loads, while total flux expansion enhances detachment access and stability. Containing the neutrals in the exhaust area with physical structures further augments these shaping benefits. These results demonstrate strategic variation in the divertor geometry and magnetic topology is a potential solution to one of fusion's power exhaust challenge.

Keywords: Magnetically confined plasmas; Nuclear energy.

Fig. 1. Overview of MAST-U plasma shapes and divertor processes.

a Overview of the magnetic geometry for the Super-X Divertor (SXD, blue), Elongated Divertor (ED, green) and Conventional Divertor (CD, green), from the indicated discharges/times together with the fuelling (D₂) location and first wall geometry. b Lower divertor with diagnostic coverage of the Divertor Monitoring Spectrometer (DMS, grey)^,, X-point position and neutral baffle location (shaded cyan). c Schematic illustration of the characteristic processes in a detached MAST-U Super-X divertor featuring the ion source (magenta), Molecular Activated Recombination (MAR) ion sink (orange) and Electron Ion Recombination (EIR) ion sink (cyan). d Measured 1D profile of the schematic divertor processes in (c), obtained from line-integrated spectroscopic inferences (# 46860 at 45% Greenwald fraction) (ions m⁻² s⁻¹) as function of poloidal distance from the X-point to the target, indicated with a grey arrow in both c and d. The detachment (or ionisation) front position is indicated with a dotted magenta line in both c and d. The experimental results in (d) are derived from a probabilistic sample obtained from a Bayesian spectroscopic analysis, showing the median (solid lines) and the 68% equal-tailed confidence interval (shaded region). See ‘Methods’ section for more information about the analysis and uncertainty propagation.

Fig. 2. Improved divertor performance without adverse impact core for long-legged divertors.

Comparison of divertor (a, b, c) and core (d, e, f, g, h) performance as function of core Greenwald fraction (f_GW in %) for the CD (red), ED (green) and SXD (blue). Divertor parameters: a Integrated ion target flux (symbols) (with polynomial fits (solid, shaded line)), b detachment (ionisation) front position as poloidal distance to the target, c estimated perpendicular target heat load on a logarithmic scale, combining Langmuir probe and spectroscopy measurements (see Methods)^,. The results in (b and c) are derived from a probabilistic sample obtained from a Bayesian spectroscopic analysis, showing the median and the 68% equal-tailed confidence interval (shaded region). See ‘Methods’ section for more information about the analysis and uncertainty propagation. Core parameters: d–g core electron temperatures and densities at two different core Greenwald fractions (corresponding to vertical dotted lines in (a, b, c, h)) indicated by blue crosses (SXD), green dots (ED) and red plusses (CD), h P_SOL (solid lines) deduced from the following contributors: NBI absorption (TRANSP, dashed lines); Ohmic heating (EFIT, not shown); changes to stored energy (EFIT, not shown) and core radiative losses (bolometry, dotted lines).

Fig. 3. Power and particle balance shows additional volume long-legged divertor drives power and particle losses.

Particle (a–c) and power (d–f) balance comparisons between different divertor shapes as function of core Greenwald fraction. a–c Particle balance showing the ion target flux (lower outer divertor) - black, total ionisation source - magenta, Molecular Activated Recombination (MAR - orange) and Electron-Ion Recombination (EIR - cyan) ion sinks (both ion sinks are integrated over the divertor chamber) for the Super-X Divertor (SXD) (a), Elongated Divertor (ED) (b) and Conventional Divertor (CD) (c). d–f Power balance showing hydrogenic power losses $P_{\mathrm{loss}}^{\mathrm{hydro}}$ (orange, integrated over the divertor chamber), target power deposition P_target (black, obtained from spectrocopically inferred temperatures and Langmuir probe particle fluxes) and estimated power flow into the divertor chamber (magenta, $P_{\mathrm{div}}\approx P_{\mathrm{loss}}^{\mathrm{hydro}}+P_{\mathrm{target}}$) assuming that the divertor chamber power losses are dominantly hydrogenic, in agreement with imaging bolometry measurements. Under the assumption that the lower and upper divertors are similar (consistent with Langmuir probe results), $P_{\mathrm{div}}$, $P_{\mathrm{loss}}^{\mathrm{hydro}}$ and P_target have been multiplied by two to obtain integrated values of the upper and lower outer divertors. The results are derived from a probabilistic sample obtained from a Bayesian spectroscopic analysis, showing the median and the 68% equal-tailed confidence interval (shaded region). See Methods section for more information about the analysis and uncertainty propagation.

Fig. 4. Power flow and 1D ion sources/sinks show similar divertor conditions at same poloidal distance to X-point.

Spectroscopically inferred line-integrated ion sources (magenta) and sinks (Molecular Activated Recombination (MAR) - orange; Electron-Ion Recombination - cyan) (part. m⁻² s⁻¹) for the Super-X (SXD) (a), Elongated (ED) (b) and Conventional (CD) (c) Divertors at f_GW = 35% as function of poloidal distance to the X-point. The red (CD), green (ED) and blue (SXD) vertical coloured dotted lines indicate their respective strike point positions, indicated by their magnetic geometry (d–f). The 1D ion source/sink profiles (a, b, c) are extended downstream of their respective strike-points due to convolution of the radial-extent of the SOL/far-SOL with the spectroscopic lines-of-sight, where the plasma is colder than at the separatrix. g Power flow (W) towards the divertor targets as function of poloidal distance to the X-point from the divertor entrance to the target for the CD (red), ED (green) and SXD (blue) at f_GW = 35%, with vertical dotted lines indicating their respective strike points. The part where the divertor leg is detached is shaded in grey in grey. The power flow is inferred by subtracting from $P_{\mathrm{div}}$ the cumulative sum of the hydrogenic power losses from upstream to the target. The results are derived from a probabilistic sample obtained from a Bayesian spectroscopic analysis, showing the median and the 68% equal-tailed confidence interval (shaded region). See Methods section for more information about the analysis and uncertainty propagation.

Fig. 5. Increasing the divertor leg length does not alter the ionisation region after detachment, in agreement with simulations.

Synthetic D₂ Fulcher emission from SOLPS-ITER simulations for the Super-X (SXD) (a, blue), Elongated (ED) (b, green) and Conventional Divertors (CD (c, red), overlaid with 5 eV contours (dashed lines) and the separatrix (solid line). d–f Experimentally measured D₂ Fulcher band emission (595-605 nm) for a strike point scan with magnetic equilibrium shown, moving from CD to SXD at constant density and power, obtained through inverting Multi-Wavelength-Imaging (MWI) imaging data for # 46895. g–i 1D ion sources and sinks (ionisation - magenta, Molecular Activated Recombination (MAR) ion sink - magenta, Electron-Ion Recombination (EIR) ion sink - cyan), obtained from spectroscopic analysis integrated along the spectroscopic lines of sight (Fig. 1b) (part. m⁻² s⁻¹), compared against synthetic diagnostic results from SOLPS-ITER simulations (dotted lines). For the SXD (g) two SOLPS-ITER simulation results are shown: one with default rates and one with corrected molecular charge exchange (${\mathrm{D}}_{\mathrm{2}}+{\mathrm{D}}^{+}\to {{\mathrm{D}}_{\mathrm{2}}}^{+}+\mathrm{D}$) rates (‘Sim. Corr. Rate’), obtained from, which increases MAR. To guide the eye, a shaded magenta vertical line has been added at a radius of 0.95 m and a black arrow has been added at the strike point location (a–i). The experimental results in (g, h, i) are derived from a probabilistic sample obtained from a Bayesian spectroscopic analysis, showing the median (solid lines) and the 68% equal-tailed confidence interval (shaded region). See Methods section for more information about the analysis and uncertainty propagation.

Fig. 6. Overview of simulated and experimental ion sources and D₂ Fulcher emission for different divertor shapes.

a–c 2D ionisation source from SOLPS-ITER simulations (shown in Fig. 5) with horizontal lines at z = −1.6 m (pink) and z = 1.07 m (magenta), demarking the edge of the divertor spectroscopy view and X-point, respectively. The fraction of the ion source downstream these limits compared to the total ion source (outer leg only) are noted. d–f Synthetic diagnostic for the D₂ Fulcher emissivity (arbitrary units) obtained from SOLPS-ITER simulations. g–i Measured D₂ Fulcher emissivity (595-605 nm) obtained from combined divertor imaging and X-point imaging inversions. The indicated time and discharges used are shown and are obtained from repeat discharges for the same core density as used in Fig. 5. A horizontal line at the height of the X-point location is added (magenta). Only emissivities obtained at the same r, z corresponding to the simulation grids are shown. An inversion artefact is present near r = 0.85 m, z = −1.6 m, where there is a gap in coverage between the X-point and divertor imaging systems. Data are shown for the Super-X (SXD, blue, a,d,g), Elongated (ED, green, b,e,h) and Conventional (CD, red, c,f,i) Divertors.