Table of contents

The feasibility of D-³He reactor plasma conditions in a tokamak of the NET/INTOR class is investigated. It is found that, depending on the energy confinement scaling law, energy breakeven can be achieved in NET without significant modification of its design. Significant improvement in Q (ratio of fusion power to injected power) can be achieved by removing the tritium producing blanket and replacing the inboard neutron shield by a thinner shield optimized for the neutron spectrum in D-³He; this allows the plasma major radius and aspect ratio to be reduced and higher beta and Q-values (up to about 3) to be achieved. The implications of D-³He operation for neutron shielding, the heat loads on the first wall and the divertor as well as plasma refuelling are considered.

The absolute densities of the dominant light impurities in the plasma centre have been measured in JET during neutral beam heating using charge exchange spectroscopy. The behaviour of carbon, oxygen and helium (in helium discharges) is investigated for limiter and X-point discharges. When the plasma is in the X-point configuration, the central carbon concentration is significantly lower than in the material limiter configuration (by a factor of two to three). After the transition from the L-mode to the H-mode, a strong impurity buildup is observed in the plasma centre which significantly reduces the central dilution factor n_D/n_e from initial values near 0.8 to about 0.6 at the collapse of the H-mode. Nonetheless, during the H-mode, the product of deuteron density and ion temperature n_DT_i (both derived by the charge exchange spectroscopy diagnostic) shows a clear enhancement with respect to the L-mode at the same input power. Helium discharges as compared to deuterium discharges have a higher carbon concentration (consistent with physical sputtering) and a much reduced oxygen concentration (n_o/n_e < 0.005, consistent with chemical release of oxygen). The transient behaviour of oxygen and carbon following deuterium pellet injection is discussed using a comparable non-pellet reference discharge. An increase of the oxygen content in the plasma centre is observed some 600 ms after injection of a deuterium pellet, while the central carbon density appears only transiently diluted by the sudden addition of pure deuterium. No evidence of a cleaning action through sawtooth relaxations is detected in the case of low-Z impurities.

The resistive ballooning mode is studied in a separatrix geometry for parameters consistent with the H-mode. Besides the 'usual' ideal and resistive ballooning modes accessing a second stable regime at large plasma pressure gradients (α), there is another class of resistive ballooning modes, which do not stabilize as a increases. The properties of these modes are discussed in detail. Large pressure gradients associated with the H-mode, near the separatrix, can destabilize this class of rapidly growing resistive ballooning modes. It is found that since these modes are stabilized by increasing plasma pressure, the possibility of a second stable window also exists for these resistive ballooning modes.

Exact analytical global solutions of the MHD wave equations are given for a toroidal vessel containing a homogeneous plasma magnetized by a toroidal field B_φ(r). The boundary conditions on the torus are satisfied exactly by a collocation method. The resonant eigenfrequency of the container is calculated for arbitrary meridional crosssection of the torus and arbitrary aspect ratio as a function of magnetic field and of temperature. Plots of the electric field lines are given.

It is planned to use lower hybrid current drive (LHCD) in the Tokamak de Varennes to support continuous operation for the entire duration (30 s) of the toroidal field. In the nominal Ohmic equilibrium, lower hybrid slow wave accessibility is poor (n_{1 access} ≃ 3), which severely limits the number of resonant electrons and LHCD efficiency. Careful modelling is therefore needed to estimate the minimum RF power requirements and the recharge possibilities while maintaining a constant total plasma current. The Bonoli-Englade code has been modified with respect to evolution of narrow Brambilla spectra, RF current diffusion and proportional-integral-derivative feedback control of the plasma current. It appears that 200 kA of current could be maintained in a steady state with about 1 MW of LH power at 3.7 GHz and with n₀ = 4 × 10¹⁹ m⁻³. With the same power, in an alternating Ohmic/RF scenario, the Ohmic transformer could be recharged to 2 V·s in about four seconds at a slightly reduced density.

The local carbon diffusion coefficient has been measured in the S-l spheromak by detecting the radial spread of injected carbon impurity. The radial impurity density profile is determined by the balance of ionization and diffusion. Using the measured local electron temperature and density, the ionization rate is determined from which the particle diffusion coefficient is inferred. The results found in this work are consistent with Bohm diffusion, and the absolute magnitude of the local particle transport coefficient D_⊥ is found to be (4-6) × D_Bohm. The measured particle transport rates are correlated with the plasma pressure gradients, suggesting a pressure driven transport mechanism.

Application of ion cyclotron resonance heating to the central region of the plasma can produce sawtooth-free periods lasting several confinement times and central electron temperatures in excess of 10 keV. The resulting electron transport is studied locally by examining the relationship between total heat flux and temperature and density gradients. The scatter of data is greatly reduced from that found in a previous study of neutral beam heated discharges; this reduction is due to the absence of sawteeth and the inclusion of radiation profiles, and it improves the accuracy of determining heat diffusivities. The spatial profile of the heat diffusivity is found to be virtually independent of plasma current in the central region and strongly dependent upon current in the outer region. When this result is combined with results obtained from radiofrequency modulation experiments, heat pulse propagation measurements and transient plasma responses from pellet injection, a coherent picture of heat diffusivity profiles emerges. Together, these results confirm that the global confinement time has a dependence upon plasma current which is weaker than linear.

The confinement requirements of a D-T tokamak reactor are fully met by extrapolating JET results according to well proven scaling laws. The size of the reactor is determined by technical constraints on blanket thickness, wall loading and magnetic field strength. The energy confinement time is assumed to continue to follow either the empirical L-mode (Goldston) or the offset-linear (Rebut-Lallia) relations.

The Balmer line intensities of atomic hydrogen from the WT-III tokamak plasma have been observed and interpreted in terms of the collisional-radiative model. The atomic and molecular hydrogen densities are deduced together with the electron density. Going from the plasma centre to the edge region along the minor radius, the atomic hydrogen density increases slightly and molecular hydrogen is found only in the periphery. In the edge region, the atomic hydrogen density reaches 1 × 10¹⁶ m⁻³ and the molecular hydrogen density 4 × 10^l7 m⁻³. The electron density distribution is consistent with that determined by infrared laser interferometry.

A transport code simulation of the plasma evolution in the CASTOR tokamak has been performed, which is in agreement with the available experimental data. A method of determining the empirical model for the electron thermal diffusivity is presented. The transport coefficient is consistent with the expectations of the electrostatic drift wave theory.

RF power is widely used as an auxiliary heating method in fusion devices. This paper reviews the relevant theoretical considerations for the ion cyclotron, lower hybrid and electron cyclotron ranges of frequency, and presents the history, the state of the art, and the plans and prospects for antennas and transmission lines for RF heating. Reactor-relevant concerns are discussed, and the information needed to develop realistic antenna designs for a reactor environment is assessed.

This Symposium is the third one treating the subject of aneutronic power (the Symposium on Clean Fusion, held in April 1976 in Washington, DC, and the Symposium on the Feasibility of Aneutronic Power, held in September 1987 in Princeton, NJ, were on the same subject). These meetings were organized by B.C. Maglich and G.H. Miley. The Scientific Committee included H.L. Berk, B. Coppi, L. Lidsky, R. Kulsrud and others. Attendance was by invitation only. The number of participants was approximately 50 in 1987 and approximately 200 in 1989. The proceedings of the 1987 meeting were published in Nucl. Instrum. Methods, Phys. Res., A 271 (1988); the proceedings of the present conference will appear in Fusion Technology.