The influence of rotation on the βN threshold for the 2∕1 neoclassical tearing mode in DIII-D

Utilizing a capability to vary neutral beam torque injection in the DIII-D [J. L. Luxon, Nucl. Fusion 42, 614 (2002)] tokamak, m∕n=2∕1 neoclassical tearing mode onset thresholds are found to fall by about one unit in βN, from ∼3 to ∼2, in ITER-like sawtoothing high-energy confinement modes of plasma operation [R. Aymar, Plasma Phys. Control. Fusion 42, B385 (2000)] as “co-injected” torque and rotation are reduced. However, increasing levels of torque and rotation in the counter-direction do not lead to corresponding rises in βN thresholds. More encouragingly, error field sensitivity is not found to increase in low rotation plasmas, as might be expected theoretically. These results pose an interesting physics problem, as well as raising concern for future devices such as ITER. Further analyses have explored possible physics origins of the behavior. They suggest many of the usual effects expected to lead to a rotation dependence (mode coupling, wall drag, ion polarization currents) are not significant, with instead models that depend on the size and sign of rotation shear playing a role. Onset behavior suggests the mode is close to being intrinsically (classically) unstable when it appears, and a conceptual explanation is offered for a mechanism by which rotation shear feeds into the onset process through changes to the classical tearing stability index, Δ′. Further parameter extensions and studies are desirable to fully resolve the underlying physics of this interesting process.