Session BO4: DIII-D Tokamak

Overview of Recent DIII-D Experimental Results

Recent DIII-D experiments contributed to the ITER physics basis and to physics understanding for extrapolation to future devices. A predict-first analysis showed how shape can enhance access to RMP ELM suppression. 3D equilibrium changes from ELM control RMPs, were linked to density pumpout. Ion velocity imaging in the SOL showed 3D C$_{\mathrm{2}}^{\mathrm{+\thinspace }}$flow perturbations near RMP induced n$=$1 islands. Correlation ECE reveals a 40{\%} increase in Te turbulence during QH-mode and 70{\%} during RMP ELM suppression vs. ELMing H-mode. A long-lived predator-prey oscillation replaces edge MHD in recent low-torque QH-mode plasmas. Spatio-temporally resolved runaway electron measurements validate the importance of synchrotron and collisional damping on RE dissipation. A new small angle slot divertor achieves strong plasma cooling and facilitates detachment access. Fast ion confinement was improved in high q\textunderscore min scenarios using variable beam energy optimization. First reproducible, stable ITER baseline scenarios were established. Studies have validated a model for edge momentum transport that predicts the pedestal main-ion intrinsic velocity value and direction.   

High confinement in negative triangularity discharges in DIII-D

Discharges with negative triangularity ($\delta )$ shape have been created in DIII-D with H-mode-like confinement (H$_{\mathrm{98y2}} \quad =$ 1.2) and high normalized beta (\textit{{\ss}}$_{\mathrm{N}} \quad =$ 2.6) with L-mode-like edge pressure profiles and no ELMs. These inner-wall-limited plasmas with $\delta =$\textit{-0.4 }had the same global performance as a positive triangularity ($\delta =$\textit{0.4)} ELMing H-mode discharge with the same $I_{\mathrm{p}}$, elongation and area. Preliminary fluctuation data shows negative $\delta $ plasmas have lower turbulence levels, typically reduced by 20{\%}, in the outer region of the plasma, 0.7 \textless $r/a $\textless 1.0, compared to equivalent positive $\delta $ discharges. Correspondingly, transport analysis indicates reduced ion and electron diffusivities for negative $\delta $ compared to the positive $\delta $ cases. Also, the positive triangularity discharges had 30-50{\%} lower neutron rates as the identically heated negative triangularity ones, due primarily to impurity retention and deuterium dilution. These results show that negative triangularity is a viable candidate for reactor scenarios with its high confinement, ELM-mitigated characteristics plus a more economical and effective option for divertor placement.   

Shafranov shift bifurcation of turbulent transport in the high $\beta_p$ scenario on DIII-D

The Shafranov shift stabilization of turbulence creates a bifurcation in transport leading to formation of a large radius internal transport barrier (ITB) in the high $\beta_p$ scenario on DIII-D. The high $\beta_p$ scenario exhibits high confinement at high $\beta_N$ and high bootstrap fraction in the absence of rapid rotation or negative central shear. Spontaneous formation of an ITB at fixed $\beta_N$ is examined. The energy confinement improves following formation of the ITB. The improvement is associated with a decrease in the minimum mid-radius characteristic turbulence parameter associated with the Shafranov shift: $\alpha-s$, where $\alpha=q^2R d\beta/d\rho$ is a measure of the Shafranov shift, and s is the magnetic shear. After ITB formation, $\alpha-s>0$ within region of ITB and $\alpha-s<0$ outside the ITB. Before ITB formation, $\alpha-s<0$ throughout the entire core. TGLF transport simulations show a bifurcation of the transport depending on the electron pressure gradient scale length. Before ITB formation, the experimental scale length is on the high-transport side of bifurcation. After ITB formation, experimental scale length is on the low-transport side of the bifurcation in the region of the ITB.   

Plasma Response to n=3 Magnetic Perturbations in Noninductive Hybrid Plasmas in the DIII-D Tokamak

3D magnetic perturbations (MPs) are effective in suppressing Type-I and Grassy ELMs in DIII-D noninductive Hybrid plasmas over a wide range of q95 (5.2-7.5) and beam torque (6 -0.2 Nm) with minimal confinement degradation (\beta$_{N}$\approx3.2, H$_{98}$\approx1.2)$. Recent experiments elucidate the role of the plasma response to n= 3 MPs that is responsible for the effectiveness of ELM suppression in this regime. Scans of the n= 3 applied spectrum were performed using the new ASIPP Super Supplies and by comparing the plasma response to even/odd parity and single row I-coil configurations. Even parity is poor at driving plasma response and for ELM suppression, consistent with model predictions. All other coil configurations showed strong amplification by the plasma, $\approx4x$ larger than for the $\beta$_{N}$\approx1.8$ ITER inductive scenario, consistent with predictions from linear MHD modeling. These results reveal the beneficial role of high beta and elevated q$_{95}$ for the suppression of ELMs by MPs in Advanced Tokamak scenarios.   

Stability of helical cores in high performance tokamak discharges

The threshold for spontaneous growth of m/n $=$ 1/1 helical cores in tokamaks is determined using VMEC. This stability model is based on a DIII-D hybrid discharge with helical core, and it predicts ITER (15MA scenario) to operate far in the helical core unstable regime. Helical cores can only exist in tokamak discharges with monotonic but low, or reversed q-shear and q$_{\mathrm{min}}\approx $ 1 in the core. The helical core is a saturated internal kink mode; its stability limit is proportional to (dp/d$\rho )$/B$_{\mathrm{t}}^{\mathrm{2}}$ around q $=$ 1. Below the stability limit, applied 3D fields can drive a helical core to finite size, as in DIII-D. Above it, a random 3D kick excites a large helical core. In the DIII-D hybrid discharge the helical core contributes to flux pumping, but it is unclear, if helical cores are experimentally detrimental, once they grow in size. Helical cores occur frequently in C-Mod due to impurities; modeling shows, they become unstable in these discharges due to a reversed shear q-profile, which lowers the stability boundary.   

Test of the Eich model for ELM energy densities in DIII-D

A collisionality scan on DIII-D reveals that peak parallel ELM energy densities during Type-I ELMs are within 0.5 -- 2 times of a new model (T. Eich, NME 2017, in publication). In contrast to the model our analysis shows pedestal pressure dependence of ELM energy density. We find proximity to the L-H threshold as important scaling factor beyond the Eich model. ELMs with large energy densities were observed when barely above the LH-threshold. Linear stability analysis with ELITE shows that lower n peeling-ballooning modes with deeper eigenfunctions result in higher divertor heat loads. Measurements with fast soft X-ray and infrared thermography facilitate tracking the distribution of conducted and radiated ELM energy. As ITER will operate close to the LH-threshold our studies emphasize the importance of considering the full pre-ELM phase and not only static profiles for determining the heat load.   

Advances in Understanding and Control of Plasma Rotation on DIII-D

Momentum transport experiments on DIII-D have advanced our understanding of the origin of core and edge rotation by showing that (1) core rotation in low-torque electron-heated ITER-like plasmas displays hollowing driven by turbulence in the absence of MHD, (2) intrinsic rotation in torque-free electron-heated plasmas follows the favorable rho* and nu* scalings as previously found in intrinsic torque experiments using NBI, (3) the edge plasma rotation can be controlled through shaping of triangularity and X-point radius, and (4) rotation and density profiles have separate dependencies on the applied 3D field spectra. These advances inform strategies to avoid low torque disruptions by tailoring turbulent modes that minimize rotation hollowing, and provide confidence in dimensionless scaling of intrinsic torque and rotation to ITER. The triangularity and X-point position provide important new actuators on the rotation beyond neutral beam injection that are available for any diverted tokamak including ITER. The separate spectral dependencies of the momentum and density explain how quiescent braking as well as edge isolated ELM control are possible even in machines with limited toroidal harmonic EFC coils.   

Role of turbulence in determining particle transport in DIII-D

Recent advances in DIII-D identify how changes in turbulence and ExB shear affect particle transport in H-mode plasmas. Using a combination of co- and counter- injected neutral beams to vary applied torque, the ExB shear is systematically scanned at fixed power and fueling. When the ExB shear is reduced below the linear gyro-kinetic growth rates inside the pedestal top (ρ=0.6-0.8), the particle confinement is strongly reduced by an increase in outward diffusion. Furthermore, a slow modulation in ECH power from 1 to 3 MW shows that the density reduction (“pump-out”) originates in the same region. Time-dependent analysis finds that the pump-out begins with a strong increase in density fluctuations measured by DBS at ρ=0.78, where the initial density reduction is largest, along with an increase in the linear growth rate of the Ion Temperature Gradient (ITG) mode. Turbulence modeling by TGLF shows that the plasma eventually transitions from an ITG mode to a Trapped Electron Mode (TEM) regime during high power ECH, but the TEM is not the initial cause of density pump out. For stationary density profiles, the frequency of the dominant unstable mode (i.e., ITG or TEM) correlates with the local density gradient, as predicated by theoretical simulations.   

Local Access Conditions for ELM-free Pedestals in DIII-D Quiescent H-mode Plasmas

Quiescent H-mode (QH-mode) has been identified as an attractive stationary operational regime in tokamaks due to its lack of edge localized modes (ELMs), along with good particle and impurity control due to the presence of MHD or edge turbulence. Local edge access conditions such as a critical edge rotational shear for the transition from a QH-mode to a typical ELMy H-mode in DIII-D are explored. The experimentally determined critical shearing rates are compared to a theoretical model, which demonstrates a linear relationship with $\sqrt{(T_i+T_e)/(k_y L_p \Delta x)}$, where T is temperature, $k_y$ the poloidal wave number, $L_p$ the pressure gradient scale length, and $\Delta x$ the radial width of the mode. Linear BOUT++ stability calculations are performed to calculate edge turbulence characteristics included in the model, such as mode structure, dominant toroidal model number, and growth rates, and are shown to compare well with experimental observations from pedestal profiles, BES, and magnetics.   

A unique predator-prey system of coupled turbulence, drive, and sheared ExB flow in the pedestal of high performance DIII-D plasmas

A unique, long-lived predator-prey oscillation regime (3-12 energy confinement times) is observed to replace coherent edge harmonic oscillations in recent low-torque quiescent high confinement (QH) mode plasmas. The physics of this system has been revealed through simultaneous measurements of local density turbulence \~{n}, ExB velocity V, and ExB shear V' at eight pedestal locations using Doppler backscattering. ExB velocity being poloidally and toroidally symmetric is found to be driven by pressure gradient and not by \~{n}. The phase space of V' and \~{n} exhibits the characteristics of a predator-prey cycle with V' (predator) lagging \~{n} (prey). It is the time-lag in the evolution of V at different pedestal locations which has been found to dictate V' evolution. \~{n} increases while V' decreases and when V' increases, \~{n} is suppressed. Observations of oscillations in edge transport relevant parameters indicate a potentially significant contribution of this mechanism to pedestal transport regulation.   

Understanding the influence of current profile and RF heating on impurity transport in advanced tokamak scenarios

Recent DIII-D experiments show that the advanced tokamak hybrid scenario is compatible with a tungsten (W) divertor. The hybrid scenario, with on-axis electron cyclotron current drive (ECCD) and q$_{min}$~1, shows no degradation in $\beta_n$ or energy confinement compared to an all-carbon divertor. In contrast, a high q$_{min}$ scenario (q$_{min}\geq1.5$) with off-axis ECCD experiences on-axis W accumulation throughout the discharge. With the application of on-axis ECCD in the hybrid scenario, the velocity to diffusion ratio (V/D) calculated with STRAHL reverses sign in the core, producing an off-axis peak in the W density profile. These results indicate that plasmas with broader current density profiles and off-axis ECCD are more susceptible to high-Z impurity accumulation. Results from the hybrid plasmas show the feasibility of a steady-state scenario with a W divertor, thus improving the physics basis of Q=5 steady-state operation on ITER. *Supported by US DOE under DE-AC52-07NA27344 and DE-FC02-04ER54698.   

Characterizing Tungsten Sourcing and SOL Transport during the Metal Rings Campaign

The Metal Rings Campaign on DIII-D utilized two isotopically and poloidally distinct toroidal arrays of tungsten coated inserts in the lower divertor to study W divertor erosion near the outer strike point (OSP) and divertor entrance and subsequent migration in a mixed-material (C-W) environment. In AT hybrid discharges (P$_{\mathrm{AUX}}=$14 MW, H$_{\mathrm{98}}=$1.6, $\beta_{\mathrm{N}}=$3.7) with rapid ELMs (f$_{\mathrm{ELM}}$\textasciitilde 200 Hz, $\delta $W/W\textasciitilde 0.7{\%}) W impurities are seen to reach the midplane predominantly from the OSP region rather than the divertor entrance (far-SOL). Conversely, in scenarios with less frequent larger ELMs (f$_{\mathrm{ELM}}$\textasciitilde 60 Hz, $\delta $W/W\textasciitilde 3.6{\%}) , the W impurities are found to transport equally from the OSP and entrance region. ELM-resolved spectroscopic measurements of W sourcing indicate that large ELMs can source W at many times the inter ELM rate. The peak W erosion rate can shift radially outwards consistent with the ELM energy flux, thereby shifting the balance between strikepoint and far-SOL sources. Changes in the peak erosion locations between forward and reversed Bt discharges are consistent with ExB ion drift effects. Evidence for a near-SOL impurity buildup between the divertors driven by the parallel grad-Ti force is also seen.   

Impact of target material on D and D$_{\mathrm{2}}$ recycling in DIII-D ELMy H-mode discharges

DIII-D operation with W divertor inserts shows molecular recycling flux (measured by Fulcher-a spectroscopy) is reduced between ELMs in comparison with a C divertor where the flux is dominated by D$_{\mathrm{2}}$ molecules ($\ge $90{\%}). This effect is partly explained by the higher reflection probability of atomic D on W. During ELMs, the molecular fraction drops by factor \textgreater 2 on both C and W targets. To study the effect of higher ion impact energy (E$_{\mathrm{imp}})$ on transient D re-emission during ELMs we have applied fast electrostatic bias to a DiMES probe equipped with a W and C sample set. A 50{\%} increase of E$_{\mathrm{imp\thinspace }}$from \textasciitilde 150 eV due to biasing led to transient increase of atomic D re-emission flux on both targets. Similar increase of the D$_{\mathrm{2\thinspace }}$flux was only seen on C. Thus, the ratios of atomic and molecular fluxes on C varied in a similar way to those measured during ELMs. This variation in molecular recycling fraction with material has implications for the dynamics of density pedestal recovery between ELMs, the overall global particle balance of the system, and possibly the overall detachment onset conditions transiently due to the ELM particle influx.   

Effects of divertor geometry on H-mode pedestal structure near divertor detachment

Divertor geometry is found to significantly affect the shape of H-mode pedestal profiles as a function of density up to divertor detachment. In the open divertor, i.e. with the strike point on a flat target plate, the pedestal width is reduced during the detachment state. In contrast, for the closed divertor, i.e. with a baffle-dome divertor or small-angle-slot divertor, the pedestal is significantly wider during the detachment state. In addition, near divertor detachment, the open diverted plasma exhibits a more aligned density and temperature pedestal, while in the closed divertor the detachment results in a greater relative shift between the density and temperature pedestal. Moreover, enhanced fluctuations are excited with divertor detachment in both divertor geometries. The fluctuations appear to be stronger in the open divertor than that in the closed divertor, opposite to previous results with additional transport broadening the pedestal. *Work supported by US DOE under DE-FC02-04ER54698, DE-NA-0003525.   

Modelling the detachment dependence on strike point location in the small angle slot divertor (SAS) with SOLPS

The new Small Angle Slot (SAS) divertor in DIII-D is characterized by a shallow-angle target enclosed by a slot structure about the strike point (SP). SOLPS modelling results of SAS have demonstrated divertor closure's utility in widening the range of acceptable densities for adequate heat handling. An extensive database of runs has been built to study the detachment dependence on SP location in SAS. Density scans show that lower T$_{\mathrm{e\thinspace }}$at lower upstream density occur when the SP is at the critical location in the slot. The cooling front spreads across the entire target at higher densities, in agreement with experimental Langmuir probe measurements. A localized increase of the atomic and molecular density takes place near the SP, which reduces the target incident power density and facilitates detachment at lower upstream density. Systematic scans of variables such as power, transport, and viscosity have been carried out to assess the detachment sensitivity. Therein, a positive role of the viscosity is found.