Session CO4: DIII-D Tokamak

Overview of recent DIII-D experimental results

Recent DIII-D experiments have contributed to the ITER physics basis and to physics understanding for extrapolation to future devices. Resonant Magnetic Perturbation ELM suppression was extended to lower shaping as a guide to joint experiments showing first ELM suppression in ASDEX-U. The physics model of suppression was validated in low torque ITER baseline plasmas. Toroidal variation of density gradients and turbulence was documented during RMP. Plasma rotation was predicted from measured scaling of intrinsic torque and momentum transport. Runaway electron plateau dissipation using SPI was demonstrated including new understanding of synchrotron and collisional damping effects. Good coupling of high frequency RF from a helicon antenna during H-mode was observed. Upstream pedestal density at divertor detachment decreased with increasing divertor closure. Bifurcations to detachment in H-mode with increasing density were reproduced with UEDGE including drifts. Sources, SOL transport and core accumulation of tungsten from toroidally continuous divertor target tiles were identified.   

Overview of the DIII-D Divertor Tungsten Rings Campaign

Experiments have recently been carried out with toroidal arrays of W-coated metal inserts at two distinct locations in the lower divertor region. The purpose of the experiments is to determine the high-Z divertor erosion and migration, and its effect on core contamination in high performance, ELM-y H-mode, tokamak discharges in a mixed-material, i.e. C and W, environment. The experiments focused on characterizing the sputtering source from each location, the SOL transport of W, and the subsequent impact on core performance. A wide range of ELM-y conditions was studied, including ELM controlled and ELM-free regimes, to determine the importance of the divertor strike point position relative to W sources in these various regimes. The W penetration efficiency was characterized by using a far-SOL collector probe related to core W density. Correlations between source strength (as measured by W-I spectroscopy) relative to the distance of the strikepoint to each W array, the divertor target magnetic flux expansion, and ELM frequency was seen. These experiments aid in understanding the impact of high-Z divertor source location on core performance in future mixed-material fusion devices, e.g. ITER.   

Active control of high-Z material erosion in the DIII-D divertor

Dedicated DIII-D experiments coupled with modeling were carried out to understand and control high-Z material erosion in a mixed materials environment. The strong correlation of erosion with external biasing voltage suggested by modeling has been demonstrated in recent experiments. High carbon impurity concentration in the background plasma is also found to reduce the net erosion rate of high-Z materials due to increased carbon deposition in the surface. Both modeling and DIII-D experiments show that local methane injection can create a carbon coating on the metal surface. The carbon deposition provides information on radial transport due to E$\times$B drift and cross field diffusion as well as background carbon concentration. Local deuterium gas injection upstream of the W sample reduces the W net erosion rate by a factor of 2 and increases the W re-deposition ratio, mainly due to local plasma perturbation.   

Investigations of detachment bifurcations in DIII-D with UEDGE

UEDGE simulations with cross-field drifts indicate that changes in the divertor electric potential can be largely responsible for the experimentally observed bifurcated onset of detached conditions in DIII-D. Partially detached divertor conditions are presently favored for ITER for optimal plasma performance with tolerable divertor heat loads. Divertor Thomson scattering measurements in DIII-D show that in high confinement mode plasmas with the B$\times$$\bigtriangledown$B-drift towards the active X-point, the low field side (LFS) divertor plate undergoes a rapid transition from well attached to fully detached conditions within a few percent increase in plasma density. UEDGE simulations indicate that at the onset of LFS detachment, the poloidal E$\times$B drift in the private flux region is reduced from about 15$\%$ of the LFS recycling flux to less than 1$\%$, effectively eliminating a strong particle sink in the LFS divertor. This increases the LFS divertor particle content further leading to fully detached conditions.   

Effect of divertor closure and impurities on detachment onset in DIII-D

Heat flux control in future devices requires a detached divertor with upstream parameters compatible with core performance, e.g., at a lower upstream density than presently achievable. Comparison between matched H-mode discharges in the upper and lower divertors of DIII-D demonstrates onset of detachment at a reduced pedestal density for the more-closed geometry of the upper divertor. The upper divertor also produces a lower pedestal density with a less-steep profile than the lower divertor for matched discharges with no additional fueling, presumably due to a reduction in ionization source for the upper divertor. Recent experiments further compare the upper and lower divertors with the addition of impurities injected into the private flux region. These experiments measure the interplay between increased closure and radiating impurities and the effect on divertor detachment, as well as the ability of the more-closed divertor geometry to prevent the accumulation of impurities in the core.   

Improved Confinement in Highly Powered Advanced Tokamak Scenarios on DIII-D

DIII-D has recently demonstrated improved energy confinement by injecting neutral gas into high performance Advanced Tokamak (AT) plasmas during high power operation. Representative parameters are: q$_{95}$ = 6, $P_{IN}$ up to 15 MW, $H_{98}$ = 1.4–1.8, and $\beta_N$ = 2.8–4.2. Unlike in lower and moderate powered AT plasmas, $\tau_E$ and $\beta_N$ increased (and $\nu_{ELM}$ decreased) as density was increased by deuterium gas puffing. We discuss how the interplay between pedestal density and temperature with fueling can lead to higher ballooning stability and a peeling/kink current limit that increasers as the pressure gradient increases. Comparison of neon, nitrogen, and argon as “seed” impurities in high $P_{IN}$ ATs in terms of their effects on core dilution, $\tau_E$, and heat flux ($q_{\bot}$) reduction favors argon. In general, the puff-and-pump radiating divertor was not as effective in reducing $q_{\bot}$ while maintaining density control at highest $P_{IN}$ than it was at lower $P_{IN}$.   

Measurements of helicon antenna coupling in DIII-D

In preparation for 1-MW-level helicon wave experiments, good coupling efficiency of a low-power 12-element phased-array antenna at 476 MHz has been found in several plasma regimes in the DIII-D tokamak. The antenna, a traveling-wave structure of the 'comb-line' type, is designed to excite helicons (fast waves in the LHRF) at a nominal $n_{||}$ of 3. This structure is a low-power prototype (operating at up to 0.4 kW) of a 30-element structure intended for operation at the 1 MW level, which is in the design stage. The dependence of wave coupling on the antenna/plasma distance was found to fit well with a simple model with one adjustable parameter in stationary regimes. In ELMing H-mode discharges that are calculated to have complete first-pass absorption of the coupled waves, strong coupling is found even between ELMs, which supports the design of the high-power antenna. To facilitate quantitative modeling, SOL density profiles were measured with a profile reflectometer, and the density adjacent to the antenna was measured with a fixed Langmuir probe. Future experiments using the high-power antenna will permit measurement of the non-inductive current drive efficiency using helicon waves in high-beta discharges.   

Accessibility and Reproducibility of Stable High-$q_{min}$ Steady-State Scenarios by $q$-profile+$\beta_N$ Model Predictive Control

The capability of combined $q$-profile and $\beta_N$ control to enable access to and repeatability of steady-state scenarios for $q_{min}>$1.4 discharges has been assessed in DIII-D experiments. To steer the plasma to the desired state, model predictive control (MPC) of both the $q$-profile and $\beta_N$ numerically solves successive optimization problems in real time over a receding time horizon by exploiting efficient quadratic programming techniques. A key advantage of this control approach is that it allows for explicit incorporation of state/input constraints to prevent the controller from driving the plasma outside of stability/performance limits and obtain, as closely as possible, steady state conditions. The enabler of this feedback-control approach is a control-oriented model capturing the dominant physics of the $q$-profile and $\beta_N$ responses to the available actuators. Experiments suggest that control-oriented model-based scenario planning in combination with MPC can play a crucial role in exploring stability limits of scenarios of interest.   

Magnetic flux conversion in the DIII-D high-beta hybrid scenario

In DIII-D hybrid scenario discharges, the rate of poloidal magnetic energy consumption is more than the rate of energy flow from the poloidal field coils. This is evidence that there is a conversion of toroidal flux to poloidal flux, which may account for a process known as flux pumping that leads to anomalous broadening of the current profile. The hybrid is a promising high confinement scenario for ITER. The broader current profile aids discharge sustainment by raising the minimum safety factor above unity thereby avoiding sawtooth-triggered 2/1 tearing modes that spoil energy confinement. During long ($\sim$1.5s) stationary intervals with constant stored magnetic energy, a significant flux state deficit rate \textgreater10 mV was observed. This anomalous consumption of poloidal flux only occurred in discharges with $\beta_N\textgreater 2.5$ and when a relatively benign 3/2 tearing mode was present. This suggests the tearing mode plays a critical role in flux conversion. Studies have shown that 3D core displacements can lead to flux conversion, suggesting that the 3/2 tearing mode and its 2/2 side band produce helical perturbations in the core velocity and magnetic field capable of producing a dynamo EMF that drives the observed current redistribution.   

Synchrotron and collisional damping effects on runaway electron distributions

Validated models of runaway electron (RE) dissipation are required to confidently approve safe ITER ${Q=10}$ operation. DIII-D experiments using quiescent REs are exploring the importance of synchrotron and collisional damping terms to RE dissipation. New time and energy-resolved measurements of RE bremsstrahlung hard X-ray (HXR) emission reveal stark differences between high and low energy REs as damping terms are varied. Previously reported anomalously high RE dissipation only applies to low energy REs. At high energy (where synchrotron effects are strongest) low synchrotron damping cases reach higher peak RE energy despite weaker particle confinement. Low-energy RE decay is observed concurrently with high-energy RE growth. RE dissipation models predict bump-on-tail distributions whose properties depend on the damping terms. Measured HXR spectra are very broad, as expected for bump-on-tail distributions.   

The External Kink Mode in Diverted and Limited Tokamaks

The disruptive instability in diverted tokamaks when the safety factor q at the 95\% flux surface, $q_{95}$, is below 2.0 is shown to be a resistive kink. The mode is a counterpart to the ideal mode that explains the corresponding disruption in limited cross sections when $q_a$, the safety factor at the plasma boundary, lies just below a rational value m/n. Experimentally, the 2/1 kink instability is unstable for ${q_a<{2}$. However, for diverted plasmas, $q_a$ is formally infinite and the ideal theory would predict stability. Yet, the disruptive limit occurs in practice when $q_{95}$, reaches 2. It is shown from numerical calculations in L-mode equilibria that a resistive kink mode is destabilized by the rapidly increasing resistivity at the plasma edge when ${q_{95}<{2}$, but ${q_a>>{2}$. The resistive kink behaves much like the ideal kink but the growth rates scale with a fractional power of the resistivity near the q = 2 surface; the exponent transitions smoothly between fractional values up to the ideal scaling. The model also explains an observed discrepancy in the limiter case where the onset actually occurs when $q_a$ is slightly above 2.0.   

Confinement Optimization by Controlling q-shear in DIII-D Steady-State Discharges

Recent experiment in DIII-D to study effects of varying shear of safety factor (q) on transport and fluctuations showed that negative shear (NS) discharges formed an internal transport barrier inside $\rho$ = 0.5, resulting in substantial improvement in ion confinement in the core, and reduced amplitudes of broad turbulent (low- to mid-k) spectra compared with those for the positive shear (PS) discharges. Previous DIII-D experiments under similar conditions but with higher on-axis NBI powers and reversed-B$_T$ direction ($\bigtriangledown$B drift toward upper single null divertor rather than lower single null divertor), showed longer confinement improvement periods with broad NS q-profile. Measured transport characteristics are analyzed in both experiments and compared with predictions of theory-based transport models using the integrated plasma simulation frameworks (OMFIT/IPS). Based on such validated transport models, predictions of prolonged confinement improvement with better sustained ITB using higher powers of off-axis NBI and ECCD available in DIII-D will be discussed.   

Observation of toroidal variation of density gradients and turbulence in DIII-D with 3D fields during ELM suppression

Significant 3D variation in broadband density fluctuations is observed using beam emission spectroscopy and Doppler backscattering near the boundary of weakly 3D plasmas in DIII-D when non-axisymmetric fields are applied to suppress ELMs. The increase in fluctuations is concomitant with an increase in the density gradient measured using profile reflectometry, suggesting that this toroidally localized density gradient could be a mechanism for turbulence destabilization in localized flux tubes. Although changes to magnetic surface topology are shown to be too small to affect turbulence stability directly, two-fluid M3D-C1 simulations find that there is a significant 3D variation of density within flux surfaces in the pedestal. These modeled local density changes modify the local pressure- and density- gradient scale lengths, and measured turbulence is shown to increase on flux tubes with larger gradients.   

Direct measurements and comparisons between deuterium and impurity rotation and density profiles in the H-mode steep gradient region on DIII-D

The recently commissioned edge deuterium charge exchange recombination (CER) spectroscopy diagnostic on DIII-D is providing direct measurements of the deuterium rotation, temperature, and density in H-mode pedestals. The deuterium temperature and temperature scale length can be $\sim$50$\%$ lower than the carbon measurement in the gradient region of the pedestal, indicating that the ion pedestal pressure can deviate significantly from that inferred from carbon CER. In addition, deuterium exhibits a larger toroidal rotation in the co-Ip direction near the separatrix compared with the carbon. These differences are qualitatively consistent with theory-based models that identify thermal ion orbit loss across the separatrix as a source of intrinsic angular momentum. The first direct measurements of the deuterium density pedestal profile show an inward shift of the impurity pedestal compared with the main ions, validating neoclassical predictions from the XGC0 code.   

Enhanced Pedestal Pressure and Increased Pedestal Width in Quiescent H-mode Plasmas in DIII-D

A rapid transition in the pedestal pressure height and width along with increased pedestal turbulence is seen in double null quiescent H mode (QH-mode) discharges in DIII D run without ELMs at low toroidal rotation [1]. The electron pedestal pressure and confinement increase $\sim$60$\%$. Experiments in 2016 have investigated the wide-pedestal parameter space further, demonstrating $\beta_N$ up to 2.3. Power scans run at essentially constant density have found no upper power limit for the wide-pedestal QH mode at NBI input powers up to 5.5 MW. Previous shots exhibiting the return of ELMs at powers as low as 4 MW also had a density which increased with input power. The combined data set indicates that the limit on wide-pedestal operation is a density limit qualitatively similar to that seen in standard QH-mode. Over a power range of 3.9 to 5.5 MW, energy confinement time is independent of input power; the ITER-98y2 confinement factor reaches 1.6 at the highest input power. Comparing with previous single null, standard QH-mode shots at net zero NBI torque [2], confinement is 25$\%$ better at the same density and $\beta_N$.