Session GO08: MFE: DIII-D Tokamak

DIII-D: Next Steps on the Reactor Path

The DIII-D program is pursuing a mission to confront the challenges of burning plasmas in ITER and to provide a cost-effective path to reducing physics and engineering risks of a Fusion Pilot Plant. Over the next three years, planned experiments and upgrades at DIII-D target a dissipative power handling solution integrated with a high-performance fusion core. This includes the development and validation of new current drive tools for reactor regimes. An increase in plasma shaping and toroidal field will maximize plasma pressure and density, allowing high opacity/low collisionality regimes to be investigated for core-edge integration. A new modular divertor approach will enable DIII-D to explore optimum configurations for a reactor, benefiting also from a dedicated material interaction test station. Low torque heating and profile control from increased gyrotron power of 6 MW together with plasma control upgrades, will facilitate studies of burning plasma physics and performance limits. Longer term, controlling and mitigating plasma transients using improved magnetic 3D coils and passive runaway destabilizing coils are planned. A balanced approach is now adopted for the DIII-D program to deliver operational reliability, efficiency, and upgrades, thereby laying the foundation for future facility operation at “full potential” for the US fusion research program and collaborators.   

Overview of Recent DIII-D Experimental Results

Recent DIII-D experiments leveraged novel actuators, diagnostics and modeling tools to advance the physics basis for ITER and fusion pilot plants. Pressure broadening for AE control from toroidally steerable off-axis NBI and injection of high harmonic FW from a helicon antenna were demonstrated. Feedback control of density and radiated power enabled super H-mode plasmas with 100% radiative fraction. Enhanced divertor dissipation was obtained by injection of low-Z powders. Fast-ion phase space measurements of unprecedented resolution discovered inward transport by AE. Ultrafast active spectroscopy revealed a weak role of T_e/T_i on long-wavelength T_i fluctuations. Fast local magnetic field measurements over an ELM cycle showed complex pedestal current density redistribution. Frequency bands characteristic of H-mode pedestal fluctuations were reproduced with a reduced model for microtearing modes. Advanced polarimetry observed internal n=3 magnetic fluctuations preceding the n=1,2 tearing modes (TM) that limit high-qmin scenarios. By integrating a TM onset model into scenario modeling, TM stable solutions for non-inductive, β_N>4 operation were identified. The magnetic pre-sheath was characterized with micro-engineered divertor targets, validating theoretical models applicable to ITER and pilot plants.   

Extending the operational space of the high bootstrap current fraction scenario on DIII-D towards ITER steady-state

Recent DIII-D experiments have extended high β_p plasmas to higher performance, meeting the required normalized parameters for ITER Q=5 steady state operation. Plasmas with H₉₈ ≥1.5 and β_N ≥4 have been achieved at q_min≥2 and β_T≥3%, and sustained for about a current diffusion time. The normalized fusion performance G₉₈=H₉₈β_N/q² in DIII-D high β_p experiments reaches the predicted value for ITER from high β_p Q=5 modeling. Strong internal transport barrier leads to a high confinement core with bootstrap current fraction ≥ 80% and line-averaged density at the Greenwald limit. No core impurity accumulation has been observed even with a density ITB. In addition, high β_p plasmas with true single null ITER plasma shape have been obtained for the first time on DIII-D. Achieved normalized parameters (q₉₅~8.0, β_N >2.7 and H₉₈~1.4) are relevant to the ITER high β_p target. With ITER-like shape, excellent core-edge integration with simultaneously sustained ELM suppression, complete divertor detachment (DoD >10), and high confinement core (β_N>2.5) has been demonstrated with neon seeding. The ITB compensates the pedestal reduction from divertor detachment and promotes core-edge integration. These results confirm the high β_p scenario as a highly promising candidate for ITER steady-state operation.   

Deconvolving the roles of E×B shear and pedestal structure in the energy confinement quality of Super H-mode experiments

Recent experiments on DIII-D have validated previous analysis results suggesting that high plasma toroidal rotation, not high pedestal pressure, plays the essential role in achieving very high energy confinement quality (H_98y2>1.5) in super H-mode plasmas. Quasi-linear gyrofluid and nonlinear gyrokinetic transport modeling showed that the contribution from rotation in the core E×B shear is by far the main mechanism responsible for H_98y2  well in excess of standard H-mode. H_98y2 ~ 1.2 was predicted without E×B shear in high pedestal super H-mode plasmas. The new experiments confirm H_98y2~1.2 in high pedestal discharges with low E×B shear obtained by reducing the injected neutral beam torque at constant power. Conversely, high rotation shear discharges maintain very high confinement quality (H_98y2 >2), despite ~40% lower pedestal obtained by reducing the plasma triangularity. These results are consistent with previous simulations of the ITER Baseline Scenario using a super H-mode pedestal [1], showing that the predicted energy confinement quality is not above standard H-mode even at the highest pedestal pressure.

[1] W.M. Solomon, et al., Phys. Plasmas 23 (2016) 056105   

Tungsten in the ITER Baseline Scenario in DIII-D

The impact of Tungsten (W) in burning plasmas can be studied in present machines using W-equivalent radiators, i.e. gases with the same radiative loss rate index (Lz) at low temperature, as W at the higher reactor-relevant Te. Using Kr and Xe, whose Lz index at 1-4 keV matches the Lz of W at 10-15 keV, we studied the ITER Baseline Scenario with the radiated fraction (frad) from W predicted for ITER (<=30%) and a range of torque=0-4 Nm. The scenario remains robustly stable up to frad~30%, where the lost power can overtake the input power and transport losses, decreasing the core Te. Lz being a function of Te in this range, a non-linear process ensues, that reduces the radiated power and can trigger a predator-prey behaviour, increasing the temperature again. A Lotka-Volterra model, modified to include two-region diffusion (core and pedestal) and noise, can account for variable signs in the better known non-linear terms of the equations, representing the experimental conditions of dLz/dT>0 and <0 as in the IBS Te range. The limit of this cycle is reached when Te is low enough to significantly reduce the core current density, raise qmin and eliminate the sawteeth: without the ST flushing out impurities, accumulation increases and the scenario experiences a radiative collapse   

Neoclassical Tearing Mode Seeding by Nonlinear Three-Wave Interactions in Tokamaks

We report seed magnetic island formation by nonlinear three-wave coupling [1]. In this experiment a 2,1 seed island forms due to 4,3 and 3,2 island interaction with a 1,1 sawtooth precursor in the DIII-D ITER baseline scenario. The 2,1 island grows linearly as expected by neoclassical theory, and terminates the discharges. STRIDE shows these plasmas are classically stable and a number of ELMs and sawtooth crashes occur without seeding. Therefore, the 2,1 island is not formed by a classical current driven instability, but an NTM seed island is formed by frequency matching and nonlinearly interacting islands that satisfy the mode number resonance condition. These results are general and relevant for future reactors, as 70% of the analyzed DIII-D ITER baseline scenario plasmas are characterized by frequency matching of resonant tearing modes at the time of 2,1 seeding. As 2,1 seeds are produced in classically stable plasmas, tearing free operation may not be possible by maintaining a classically stable current profile in future reactors. This predicts new challenges for the development of stable plasma scenarios, calling for high differential rotation at q=2, active control and avoidance of high m,n islands as much as possible.   

Directing TGLF saturation rule development with machine learning tools

The trapped gyro-Landau-fluid (TGLF) code solves a reduced set of gyrokinetic equations. To accurately model the nonlinear saturation of turbulence, the TGLF code employs so-called saturation rules SAT0, SAT1 and the recently developed SAT2. Using automated workflows within OMFIT, we have built a curated dataset of 25000 time and space slices of DIII-D plasmas that has been used to successfully validate these saturation rules. We find that SAT1 performs well when the heat flux is largely electrostatic, i.e. with an electromagnetic component satisfying Q_EM / Q_ES < 0.01. To help direct future saturation rule development, we have applied machine learning tools to our curated dataset. First, we applied model optimization tools to the free parameters in SAT0 and SAT1, which have been previously calibrated by the first-principles gyrokinetic code GYRO. We found that free parameters governing the importance of E × B shear could merit further attention during a possible recalibration. Second, we multiplied the wavenumber (k) spectrum of SAT1 with a hypothetical correction factor of the form a / k^c × exp( - b / k), where a, b, and c are the outputs of a neural network. After training this neural network to reproduce the experimentally inferred fluxes, we find that the amplitude of the correction factor is correlated with plasma parameters that are typically associated with trapped electron mode (TEM) turbulence. Therefore, future saturation model development could focus on TEM saturation.   

Effects of divertor closure and X-point height on L-H transition

Divertor configuration and X-point height have a strong effect on the L-H transition power threshold (P_LH) on multiple tokamaks. In a dedicated experiment on DIII-D, the L-H transition was investigated with both Small Angle Slot (SAS) divertor and closed divertor. The P_LH are found to be below 2MW across a wide density range from 1.5-4x10¹³cm^-3, which is about 20-50% lower than the previously observed P_LH in an open divertor. At a density of ~4x10¹³cm^-3 P_LH with closed divertor is larger than the P_LH with SAS divertor, but at densities below 3x10¹³cm^-3 it is similar. It is also found that P_LH is reduced by nearly 30% when the divertor leg length (DLL) is reduced by ~5 cm, which indicates stronger geometry and neutral effects on P_LH than previously observed in an open divertor. Long wave-length density fluctuations and flow dynamics are measured by both BES and DBS diagnostics. Higher turbulence amplitude is observed with longer DLL. Turbulence characteristics and flow dynamics will be presented. The results will improve our understanding of the underlying physics of divertor configuration effects on H-mode access.

    

Direct Identification of Microtearing Modes in the H-mode Pedestal via Current Profile Perturbations

We present a direct, time-dependent experimental identification of low-n microtearing modes (MTMs) in the H-mode pedestal. In a series of dedicated discharges on DIII-D, we decouple the resonant location and the instability drive of pedestal-localized MTMs by exposing the plasma to large, fast vertical oscillations which induce substantial edge current perturbations. These transient events restructure the edge q profile and modify the inter-ELM evolution of the pedestal ω_*,e profile, changing the dynamic frequency evolution of magnetic fluctuations from up-chirping to down-chirping and switching the mode numbers of pedestal instabilities. Clear and distinct changes in the mode response are both qualitatively and quantitatively explained by analytic profile-based predictions, providing a compelling experimental validation of the MTM model. Additional evidence for the MTM identification is given by measurements of propagation direction, gradient saturation and transport fingerprints. Since MTMs primarily drive electron heat flux through the pedestal region, these robust experimental results suggest that, in order to predict electron temperature profile effects, reduced pedestal models should include electromagnetic terms and accurate calculations of MTM stability.   

Poloidal Asymmetries Of The Main Chamber Particle Source In DIII-D

Recent LLAMA (Lyman-α) measurements of DIII-D indicate a one order of magnitude larger ionization rate at the high field side (HFS) main chamber in comparison to the low field side (LFS) when operating with B×∇B directed towards the divertor. In contrast, the ionization at the LFS main chamber is higher by a factor of two with B×∇B drift out of the divertor, while it is lower by a factor of five on the HFS, which eliminates the HFS/LFS imbalance of the main chamber particle source. The LLAMA diagnostic measures the deuterium Lyman-α brightness across the LFS and HFS plasma boundary well below the midplane at the same vertical location. These measurements are of similar order of magnitude as synthetic Lyman-α brightnesses generated by high fidelity EDGE2D-EIRENE simulations with the full set of scrape-off layer (SOL) drifts. However, the model underpredicts the strong HFS/LFS variation when the SOL drift direction is altered and it does not reproduce the steep brightness drop radially from HFS SOL to the separatrix. The presented results highlight the importance of SOL drifts for the HFS particle source above the x-point, which is key for core-edge integrated plasma predictions and edge particle transport studies.   

Low Frequency Instabilities and Alfven Eigenmodes in Mixed Species DIII-D Plasmas

The stability of beam-ion driven instabilities in reversed-shear DIII-D plasmas is clarified. Beta-induced Alfven eigenmodes (AE) are driven by resonant co-passing high-energy deuterium ions¹ but not by hydrogen beam ions. Similarly, deuterium beams drive reversed-shear and toroidal AEs more strongly than hydrogen beams. In contrast, the instability that was previously called a ‘BAAE’ is not driven by fast ions at all.² Theory and experiment indicate that the latter modes are a low-frequency reactive instability of predominately Alfvenic polarization that occurs in plasmas with high electron temperature but modest beta; they are more unstable in hydrogen than deuterium. Simulations reproduce many features of the data but, owing to the approximation of isotropic Maxwellian fast ions, not properties that depend upon beam anisotropy.

¹Nucl. Fusion 61 (2021) 066031.

²Nucl. Fusion 61 (2021) 016029.   

Main Ion Isotope dependence of impurity transport in ion and electron-heating dominated H-mode plasmas in the DIII-D tokamak

The impurity confinement time τ_imp is up to 3x shorter in hydrogen (H) plasmas than in dimensionally similar deuterium (D) plasmas. The transport of calcium is investigated with the laser blow-off system in ELM-y H-mode discharges. Adding ECH to NBI heated D discharge reduces τ_imp 3x while τ_imp remains unaffected by ECH in H plasmas. The lack of τ_imp variation in H is a consequence of the 10x faster impurity diffusion compared to an NBI-heated D reference, which only slightly increases with ECH. The normalized impurity density gradient R/L_n is nearly identical for the H and D plasmas, rising from -2 in NBI heated plasmas to 2 in NBI+ECH heated plasmas. Turbulent transport is modeled by the quasilinear CGYRO model, which correctly reproduces the increase in impurity diffusion caused by a change of heating mix. However, the diffusion in H compared to the D case increases only by a factor 2-3, while in the experiment by a factor of 10. A sensitivity scan by CGYRO indicates impurity transport is mainly affected by indirect effects associated with the isotope change, such as profile gradients, electron-ion heat exchange, and beta stabilization but negligible direct effects of isotope.    

Quantifying impurity transport in runaway electron plateaus in DIII-D

The impurity ion diffusion coefficient in post-disruption runaway electron (RE) plateaus is found to be about 10x higher than classical ion-ion diffusion. The impurity level of RE plateaus has a key role in all aspects of RE plateau dynamics, and is crucial for predicting damage resulting from RE plateau-wall strikes. DIII-D experiments have pursued both stationary and dynamic measurement of RE plateau impurity radial transport. In the stationary experiments, profiles of electron density and radiated power are matched with a 1D radial impurity transport model. This steady-state 1D analysis arrives at impurity ion diffusion coefficients of order 1 - 10 m²/s. Dynamic experiments use small (1 Torr-L), carbon granules fired into the RE plateau, with subsequent toroidal and radial transport of  ions analyzed spectroscopically. The dynamic experiments, performed in a small number of conditions, find impurity diffusion coefficients of order 5 m²/s, consistent with the steady-state estimates. Future experiments and implications for ITER will be discussed.   

Divertor impurity densities at detachment onset in DIII-D discharges

Intrinsic (carbon) and seeded (nitrogen) impurity densities are measured in the DIII-D divertor to assess the impurity concentration c_z needed to achieve divertor detachment and benchmark recently proposed detachment scalings and measurement techniques. Impurity densities are derived from calibrated 2D visible emissivity, line-integrated VUV brightness and local T_e, n_e. Charge state abundancies and emissivity localization are inferred from a collisional radiative model and from experimentally constrained fluid simulations with the UEDGE code to account for effects of transport and charge exchange recombination. Carbon concentrations up to a few % are measured from emissivity of singly and doubly ionized carbon ions in attached conditions and reduced to < 1% in detached regimes at the radiation front due to the reduction in sputtering and the radiation front movement away from the divertor plate. Measurements of nitrogen density from singly ionized nitrogen visible emissivity are used to examine proposed scalings of c_z at detachment on separatrix density and scrape off layer power over the range of heating powers 3-10 MW, while varying the fraction of radiated power due to nitrogen.   

Simulated trends in tungsten erosion and scrape-off layer transport for the new V-shaped, tungsten-coated Small Angle Slot divertor in DIII-D

Narrowing the slot of the DIII-D Small Angle Slot (SAS) divertor to a V-shaped vertex via changes to the inboard baffle is predicted to lower gross erosion of the outboard tungsten (W) coating by a factor of 50 due to increased neutral confinement when the ion grad-B drift is towards the X-point. SOLPS-ITER edge plasma and DIVIMP impurity tracking simulations show that moving the outer strike point (OSP) more than a few cm away from the slot vertex or operating below the detachment threshold will likely negate any benefit of the V-shape, yielding erosion rates that match or exceed that calculated from the original SAS shape. However, moving the OSP outward 4 cm from the slot vertex is also predicted to reduce the fraction of eroded W into the upstream scrape-off layer by a factor of 2. A new, sheath-based prompt redeposition model estimated an average W redeposition fraction in SAS-VW of 0.74 when the OSP is at the slot vertex, improving the accuracy of net erosion estimates in regions of high electron temperature. Predictive modeling of W sourcing and transport in SAS-VW provides useful guidance for future DIII-D experiments, as well as valuable insight into the feasibility of high-Z, closed divertors for reactor-relevant scenarios.