Improved Particle Confinement with Resonant Magnetic Perturbations in DIII-D Tokamak H-Mode Plasmas

Experiments on the DIII-D tokamak have identified a novel regime in which applied resonant magnetic perturbations (RMPs) increase the particle confinement and overall performance. This Letter details a robust range of counter-current rotation over which RMPs cause this density pump-in effect for high confinement (H mode) plasmas. The pump in is shown to be caused by a reduction of the turbulent transport and to be correlated with a change in the sign of the induced neoclassical transport. This novel reversal of the RMP induced transport has the potential to significantly improve reactor relevant, three-dimensional magnetic confinement scenarios.

Figure 1

With constant neutral beam injection torque (a), the application of $n=2$ RMPs (b) in a discharge rotating (c) opposite $I_p$ (orange) causes a sharp rise in electron density (d) without any correspondingly sudden change in plasma composition (c). The effect is not present in co-$I_p$ rotation (blue).

Figure 2

Measured electron density (a), ion density (b), temperature (c), and rotation frequency (d) pedestal profiles before (blue) and after (orange) the application of RMPs in counter-$I_p$ rotating shot 182 639. The densities rise while the temperatures and rotations are unaffected by the RMPs.

Figure 3

The change in line integrated density (a) and normalized pressure (b). The observed improvements (blue) and fractional changes (green) behave similarly, peaking at moderate negative rotation and reversing when the rotation becomes positive with respect to the plasma current. Upper single null (shaded triangle) and lower single null (inverted shaded triangle) data is shown.

Figure 4

ELM characteristics for counter-$I_p$ rotating shot 182639 without (blue) and with (orange) RMPs. Black error bars show the mean and standard deviation of each dataset. The RMPs increases the size of the ELM density (a) and energy (b) crashes but increase the ELM period (c), resulting in no change to the average ELM particle flux (d). Thus, the observed pump-in (separation of on or off points in the horizontal axis) does not come from a change in ELM particle flux.

Figure 5

Magnetic perturbation induced, neoclassical ion transport across flux surfaces. Axis (a) shows the flux at the pedestal top (${\psi }_N=0.85$) changes for negative rotations, but the linear estimate of pedestal density change does not. Axis (b) shows the neoclassical ion flux profiles, and (c) shows corresponding linear estimates of the extreme case density profiles, which are relatively small deviations from the original (black) profile.

Figure 6

DBS density fluctuation measurements in discharge 182 639. The fluctuations decrease across the entire pedestal when coils are applied (a), and have a causal influence on the observed density rise (b). The corresponding deuterium density model [(a), dashed line] is comparable to the experimental rise in the profiles [(a), solid lines].