High temporal and spatial resolution measurements in the boundary of the DIII-D tokamak show that edge-localized modes ͑ELMs͒ are produced in the low field side, are poloidally localized and are composed of fast bursts ͑ϳ20 to 40 s long͒ of hot, dense plasma on a background of less dense, colder plasma ͑ϳ5 ϫ 10 18 m −3 , 50 eV͒ possibly created by the bursts themselves. The ELMs travel radially in the scrape-off layer ͑SOL͒, starting at the separatrix at ϳ450 m / s, and slow down to ϳ150 m / s near the wall, convecting particles and energy to the SOL and walls. The temperature and density in the ELM plasma initially correspond to those at the top of the density pedestal but quickly decay with radius in the SOL. The temperature decay length ͑ϳ1.2 to 1.5 cm͒ is much shorter than the density decay length ͑ϳ3 to 8 cm͒, and the latter decreases with increasing pedestal ͑and SOL͒ density. The local particle and energy flux ͑assuming T i = T e ͒ at the midplane wall during the bursts are 10% to 50% ͑ϳ1 to 2ϫ 10 21 m −2 s −1 ͒ and 1% to 2% ͑ϳ20 to 30 kW/ m 2 ͒, respectively, of the LCFS fluxes, indicating that particles are transported radially much more efficiently than heat. Evidence is presented suggesting toroidal rotation of the ELM plasma in the SOL. The ELM plasma density and temperature increase linearly with discharge/pedestal density up to a Greenwald fraction of ϳ0.6, and then decrease resulting in more benign ͑grassier͒ ELMs.

Pedestal studies in DIII-D find that the width of the region of steep gradient in the H-mode density is comparable with the neutral penetration length, as computed from a simple analytic model. This model has analytic solutions for the edge plasma and neutral density profiles, which are obtained from the coupled particle continuity equations for electrons and deuterium atoms. In its range of validity (edge temperature between 40 and 500 eV), the analytic model quantitatively predicts the observed decrease in the width as the pedestal density increases and the observed strong increase in the gradient of the density as the pedestal density increases. The model successfully predicts that L-mode and H-mode profiles with the same pedestal density have gradients that differ by less than a factor of 2. The width of the density barrier, measured from the edge of the electron temperature barrier, is the lower limit for the observed width of the temperature barrier. These results support the hypothesis that particle fuelling is an important part of the physics that determines the structure of the H-mode transport barrier.

Densities up to 40 percent above the Greenwald limit are reproducibly achieved in high confinement (H ITER89p = 2) ELMing H-mode discharges. Simultaneous gas fueling and divertor pumping were used to obtain these results. Confinement of these discharges, similar to moderate density H-mode, is characterized by a stiff temperature profile, and therefore sensitive to the density profile. A particle transport model is presented that explains the roles of divertor pumping and geometry for access to high densities. Energy loss per ELM at high density is a factor of five lower than predictions of an earlier scaling, based on data from lower density discharges.

This paper summarizes results from a two-dimensional (2D) physics analysis of the transition to and stable operation of the partially detached divertor (PDD) regime induced by deuterium injection in DIII-D. The analysis [1] shows that PDD operation is characterized by a radiation zone near the X-point at T e ∼ 8-15 eV which reduces the energy flux into the divertor and thereby also reduces the target plate heat flux, an ionization zone below the X-point which provides a deuterium ion source to fuel parallel flow down the outer divertor leg, an ion-neutral interaction zone in the outer leg which removes momentum and energy from the flow and finally a volume recombination zone above the target plate which reduces the particle flux to the low levels measured on the plates and thereby also contributes to reduction in target plate heat flux.

No abstract