The effects of changing beam and plasma species on the edge transport barrier are investigated for ELM-free hot ion H mode discharges from the recent DT experiments on JET. The measured pressure at the top of the pedestal is higher for mixed deuterium and tritium and pure tritium plasmas over and above the level measured in pure deuterium plasmas at the same heating power. The pedestal pressure increases with beam tritium concentration for mixed deuterium-tritium beam injection into deuterium plasmas where the measured edge tritium concentration remains low. Alpha heating plays a significant role in the core of such plasmas, and the possible impact on the edge is discussed together with possible direct isotopic effects. Heuristic models for the transport barrier width are proposed, and used to explore a wider range of edge measurements including full power DD and DT pulses. This analysis supports the plasma current and mass dependence for a barrier width set by the orbit loss of either thermal or fast ions, though it does not unambiguously distinguish between them. The fast ion hypothesis could well account for some of the JET observations, though more theoretical work and direct experimental measurement would be required to confirm this. An ad hoc model for the power loss through the separatrix, Ploss ∝ nedge2 Zeff,edgeIp-1, is proposed based on neoclassical theory, a ballooning limit to the edge gradient and a barrier width set by the poloidal ion gyroradius. Such a model is compared with experimental data from JET. In particular, the model ascribes the systematic difference in loss power between the Mark I and Mark II divertors to the change in the measured Zeff. This change in Zeff is consistent with the observed change in impurity production, which is described in some detail, together with a possible explanation provided by the temperature dependence of chemical sputtering.
Discharge optimization and the control of edge stability