Several seemingly unrelated effects in Alcator C-Mod Ohmic L-mode plasmas are shown to be closely connected: non-local heat transport, core toroidal rotation reversals, energy confinement saturation and up/down impurity density asymmetries. These phenomena all abruptly transform at a critical value of the collisionality. At low densities in the linear Ohmic confinement regime, with collisionality ν * ≤ 0.35 (evaluated inside of the q=3/2 surface), heat transport exhibits non-local behavior, core toroidal rotation is directed co-current, edge impurity density profiles are up/down symmetric and a turbulent feature in core density fluctuations with k θ up to 15 cm −1 (k θ ρ s ∼ 1) is present. At high density/collisionality with saturated Ohmic confinement, electron thermal transport is diffusive, core rotation is in the counter-current direction, edge impurity density profiles are up/down asymmetric and the high k θ turbulent feature is absent. The rotation reversal stagnation point (just inside of the q=3/2 surface) coincides with the non-local electron temperature profile inversion radius. All of these observations can be unified in a model with trapped electron mode prevalence at low collisionality and ion temperature gradient mode domination at high collisionality.

A set of key properties for an ideal dissipation scheme in gyrokinetic simulations is proposed, and implementation of a model collision operator satisfying these properties is described. This operator is based on the exact linearized test-particle collision operator, with approximations to the field-particle terms that preserve conservation laws and an H-theorem. It includes energy diffusion, pitch-angle scattering, and finite Larmor radius effects corresponding to classical ͑real-space͒ diffusion. The numerical implementation in the continuum gyrokinetic code GS2 ͓Kotschenreuther et al., Comput. Phys. Comm. 88, 128 ͑1995͔͒ is fully implicit and guarantees exact satisfaction of conservation properties. Numerical results are presented showing that the correct physics is captured over the entire range of collisionalities, from the collisionless to the strongly collisional regimes, without recourse to artificial dissipation.

Nonlinear gyrokinetic simulations of trapped electron mode (TEM) turbulence, within an internal particle transport barrier, are performed and compared with experimental data. The results provide a mechanism for transport barrier control with on-axis radio frequency heating, as demonstrated in Alcator C-Mod experiments [S. J. Wukitch et al., Phys. Plasmas 9(5) 2149 (2002)]. Off-axis heating produces an internal particle and energy transport barrier after the transition to enhanced Dα high confinement mode. The barrier foot reaches the half-radius, with a peak density 2.5 times the edge density. While the density profile peaks, the temperature profile remains relatively unaffected. The peaking and concomitant impurity accumulation are controlled by applying modest central heating power late in the discharge. Gyrokinetic turbulence simulations of the barrier formation phase, using the GS2 code [W. Dorland et al., Phys. Rev. Lett. 85, 5579 (2000)] show that toroidal ion temperature gradient driven modes are suppressed inside the barrier foot, but continue to dominate in the outer half-radius. As the density gradient steepens further, trapped electron modes are driven unstable. The onset of TEM turbulence produces an outflow that strongly increases with the density gradient, upon exceeding a new nonlinear critical density gradient, which significantly exceeds the linear critical density gradient. The TEM turbulent outflow ultimately balances the inward Ware pinch, leading to steady state. Moreover, the simulated turbulent particle diffusivity matches that inferred from particle balance using measured density profile data and the calculated Ware pinch. This turbulent diffusivity exhibits a strong unfavorable temperature dependence that allows control with central heating.

After many years of fusion research, the conditions needed for a D–T fusion reactor have been approached on the Tokamak Fusion Test Reactor (TFTR) [Fusion Technol. 21, 1324 (1992)]. For the first time the unique phenomena present in a D–T plasma are now being studied in a laboratory plasma. The first magnetic fusion experiments to study plasmas using nearly equal concentrations of deuterium and tritium have been carried out on TFTR. At present the maximum fusion power of 10.7 MW, using 39.5 MW of neutral-beam heating, in a supershot discharge and 6.7 MW in a high-βp discharge following a current rampdown. The fusion power density in a core of the plasma is ≊2.8 MW m−3, exceeding that expected in the International Thermonuclear Experimental Reactor (ITER) [Plasma Physics and Controlled Nuclear Fusion Research (International Atomic Energy Agency, Vienna, 1991), Vol. 3, p. 239] at 1500 MW total fusion power. The energy confinement time, τE, is observed to increase in D–T, relative to D plasmas, by 20% and the ni(0) Ti(0) τE product by 55%. The improvement in thermal confinement is caused primarily by a decrease in ion heat conductivity in both supershot and limiter-H-mode discharges. Extensive lithium pellet injection increased the confinement time to 0.27 s and enabled higher current operation in both supershot and high-βp discharges. Ion cyclotron range of frequencies (ICRF) heating of a D–T plasma, using the second harmonic of tritium, has been demonstrated. First measurements of the confined alpha particles have been performed and found to be in good agreement with TRANSP [Nucl. Fusion 34, 1247 (1994)] simulations. Initial measurements of the alpha ash profile have been compared with simulations using particle transport coefficients from He gas puffing experiments. The loss of alpha particles to a detector at the bottom of the vessel is well described by the first-orbit loss mechanism. No loss due to alpha-particle-driven instabilities has yet been observed. D–T experiments on TFTR will continue to explore the assumptions of the ITER design and to examine some of the physics issues associated with an advanced tokamak reactor.

Ohmic energy confinement saturation is found to be closely related to core toroidal rotation reversals in Alcator C-Mod tokamak plasmas. Rotation reversals occur at a critical density, depending on the plasma current and toroidal magnetic field, which coincides with the density separating the linear Ohmic confinement regime from the saturated Ohmic confinement regime. The rotation is directed co-current at low density and abruptly changes direction to counter-current when the energy confinement saturates as the density is increased. Since there is a bifurcation in the direction of the rotation at this critical density, toroidal rotation reversal is a very sensitive indicator in the determination of the regime change. The reversal and confinement saturation results can be unified since these processes occur at a particular value of the collisionality.

In transport barriers, particularly H-mode edge pedestals, radial scale lengths can become comparable to the ion orbit width, causing neoclassical physics to become radially nonlocal. In this work, the resulting changes to neoclassical flow and current are examined both analytically and numerically. Steep density gradients are considered, with scale lengths comparable to the poloidal ion gyroradius, together with strong radial electric fields sufficient to electrostatically confine the ions. Attention is restricted to relatively weak ion temperature gradients (but permitting arbitrary electron temperature gradients), since in this limit a δf (small departures from a Maxwellian distribution) rather than full-f approach is justified. This assumption is in fact consistent with measured inter-ELM H-Mode edge pedestal density and ion temperature profiles in many present experiments, and is expected to be increasingly valid in future lower collisionality experiments. In the numerical analysis, the distribution function and Rosenbluth potentials are solved for simultaneously, allowing use of the exact field term in the linearized Fokker-Planck-Landau collision operator.In the pedestal, the parallel and poloidal flows are found to deviate strongly from the best available conventional neoclassical prediction, with large poloidal variation of a different form than in the local theory. These predicted effects may be observable experimentally. In the local limit, the Sauter bootstrap current formulae appear accurate at low collisionality, but they can overestimate the bootstrap current near the plateau regime. In the pedestal ordering, ion contributions to the bootstrap and Pfirsch-Schlüter currents are also modified. a) landrema@mit.edu

Nonlinear gyrokinetic simulations of impurity transport are compared to experimental impurity transport for the first time. The GYRO code (Candy and Waltz 2003 J. Comput. Phys. 186 545) was used to perform global, nonlinear gyrokinetic simulations of impurity transport for a standard Alcator C-Mod, L-mode discharge. The laser blow-off technique was combined with soft x-ray measurements of a single charge state of calcium to provide time-evolving profiles of this non-intrinsic, non-recycling impurity over a radial range of 0.0 ⩽ r/a ⩽ 0.6. Experimental transport coefficient profiles and their uncertainties were extracted from the measurements using the impurity transport code STRAHL and rigorous Monte Carlo error analysis. To best assess the agreement of gyrokinetic simulations with the experimental profiles, the sensitivity of the GYRO predicted impurity transport to a wide range of turbulence-relevant plasma parameters was investigated. A direct comparison of nonlinear gyrokinetic simulation and experiment is presented with an in depth discussion of error sources and a new data analysis methodology.

Trapped electron mode (TEM) turbulence exhibits a rich variety of collisional and zonal flow physics. This work explores the parametric variation of zonal flows and underlying mechanisms through a series of linear and nonlinear gyrokinetic simulations, using both particle-in-cell and continuum methods. A new stability diagram for electron modes is presented, identifying a critical boundary at ηe = 1, separating long and short wavelength TEMs. A novel parity test is used to separate TEMs from electron temperature gradient driven modes. A nonlinear scan of ηe reveals fine scale structure for ηe 1, consistent with linear expectation. For ηe < 1, zonal flows are the dominant saturation mechanism, and TEM transport is insensitive to ηe. For ηe > 1, zonal flows are weak, and TEM transport falls inversely with a power law in ηe. The role of zonal flows appears to be connected to linear stability properties. Particle and continuum methods are compared in detail over a range of ηe = d ln Te/d ln ne values from zero to five. Linear growth rate spectra, transport fluxes, fluctuation wavelength spectra, zonal flow shearing spectra, and correlation lengths and times are in close agreement. In addition to identifying the critical parameter ηe for TEM zonal flows, this paper takes a challenging step in code verification, directly comparing very different methods of simulating simultaneous kinetic electron and ion dynamics in TEM turbulence.