Abstract
Neoclassical magnetic islands are observed to limit the achievable beta in COMPASS-D low collisionality single null divertor tokamak plasmas with ITER-like geometry (R0=0.56 m, B0=1.2 T, Ip=120-180 kA, k=1.6, epsilon =0.3). The limiting beta is typically well below that expected from ideal instabilities with maximum beta N in the range of 1.6 to 2.1. The plasma is heated with up to 1.8 MW of 60 GHz electron cyclotron resonance heating (ECRH) at the second harmonic with X mode polarization. The time history of the measured island width is compared with the predictions of neoclassical tearing mode theory, with good agreement between theory and experiment. The measured islands have a threshold width below which the mode will not grow. The density scaling of the point of onset of the measured instabilities is compared with two theories that predict a threshold island width for the onset of neoclassical tearing modes. Applied resonant helical error fields are used to induce islands in collisionality regimes wherein the neoclassical islands do not occur naturally, allowing the study of the behaviour of neoclassical tearing modes in this regime. The critical beta for the onset of neoclassical tearing modes is seen to be ~3 times higher in the naturally stable region. This observation is compared with the predictions of both threshold theories. A simple expression for the q scaling of the maximum achievable beta N in the presence of neoclassical tearing modes is derived on the basis of the assumption of a maximum allowable island width. The predicted q scaling of this beta limit is compared with data from a q scan, and the results are in good agreement