Skip to main content
SHARE
Publication

Experimental evidence of edge intrinsic momentum source driven by kinetic ion loss and edge radial electric fields in tokamak...

Publication Type
Journal
Journal Name
Physics of Plasmas
Publication Date
Page Number
092506
Volume
23
Issue
9

Bulk ion toroidal velocity profiles, VD+||, peaking at 40–60 km/s are observed with Mach probes in a narrow edge region of DIII-D discharges without external momentum input. This intrinsic rotation can be well reproduced by a first principle, collisionless kinetic loss model of thermal ion loss that predicts the existence of a loss-cone distribution in velocity space resulting in a co-Ip directed velocity. We consider two kinetic models, one of which includes turbulence-enhanced momentum transport, as well as the Pfirsch-Schluter (P-S) fluid mechanism. We measure a fine structure of the boundary radial electric field, Er, insofar ignored, featuring large (10–20 kV/m) positive peaks in the scrape off layer (SOL) at, or slightly inside, the last closed flux surface of these low power L- and H-mode discharges in DIII-D. The Er structure significantly affects the ion-loss model, extended to account for a non-uniform electric field. We also find that VD+|| is reduced when the magnetic topology is changed from lower single null to upper single null. The kinetic ion loss model containing turbulence-enhanced momentum transport can explain the reduction, as we find that the potential fluctuations decay with radius, while we need to invoke a topology-enhanced collisionality on the simpler kinetic model. The P-S mechanism fails to reproduce the damping. We show a clear correlation between the near core VC6+|| velocity and the peak edge VD+|| in discharges with no external torque, further supporting the hypothesis that ion loss is the source for intrinsic torque in the present tokamaks. However, we also show that when external torque is injected in the core, it can complete with, and eventually overwhelm, the edge source, thus determining the near SOL flows. Finally, we show some additional evidence that the ion/electron distribution in the SOL is non-Maxwellian.