Table of contents

This year marks the 20th anniversary of the Journal NUCLEAR FUSION which first appeared in September 1960. On a general scientific scale, this point in history may be too recent for a celebration, but for those engaged in the pursuit and preservation of knowledge in such a relatively young field as controlled nuclear fusion research—twenty years is a landmark. Over these two decades, the Journal has developed from a slim quarterly to a mature international monthly. It serves as an important vehicle for the exchange and dissemination of information and thus enhances co-operation among all nations, developed and developing, dedicated to the task of obtaining useful power from the fusion of light nuclei.

In some ways ideas on fusion reactors have not changed during the past twenty years—and in other ways the total development of conceptual reactor designs has occurred within that time. Now, as then, most practical reactor designs are based on the reaction between deuterium and tritium at temperatures in excess of 10⁸ degrees, requiring a breeding blanket containing lithium to regenerate the tritium fuel and substantial shielding to reduce radiation levels beyond the reactor. On the other hand, the form and structure of a reactor was only vaguely perceived twenty years ago, whereas conceptual designs have now been developed in great detail and described at great length. The associated technology has also developed rapidly in recent years, and is now an essential and significant part of any major fusion programme.

Traps in which the plasma confinement region is bounded along the lines of force of the magnetic field are customarily called open traps. The bestknown and most common type of open trap is the so-called 'magnetic-mirror machine' or 'mirror trap', suggested in the early fifties independently by G.I. Budker in the USSR and R.F. Post in the USA. Plasma confinement in the mirror trap is due to the reflection of particles from the regions of strong magnetic field ('mirrors') at the ends of the device. These reflect only particles with sufficiently large magnetic moment. Particles with small magnetic moment freely fly out of the device so that a conical 'hole' without particles, the loss cone, is formed in the phase space.

Pinch research is named after the pinch effect which in turn was first christened and demonstrated experimentally in 1904 by passing a current through a liquid metal. The effect describes the tendency of a current-carrying medium to collapse radially, because of the interaction of the current with the self-magnetic field.

In the history of controlled thermonuclear fusion, there have been no ideas comparable in beauty and conceptual significance with that of the stellarator. The idea behind this device is the general concept of toroidal closed magnetic configurations which is the basis of the theory of magnetic plasma confinement in toroidal systems. It has engendered such important concepts as the magnetic surface, rotational transform and shear of magnetic field lines, the average magnetic well, the separatrix and island structure of magnetic surfaces, stochastization of magnetic field lines, the divertor, etc., and it has become an integral and fundamental part of the physics of high-temperature plasmas.

Inertial confinement fusion research (ICF) is directed towards demonstrating the scientific feasibility of very rapidly heating and compressing small pellets of suitable fuel until conditions exist where thermonuclear fusion can occur and useful amounts of power can be produced. Such pellets may be heated up to the required temperatures either by means of energy drivers or by magnetic compression (imploding liners). Magnetic-confinement fusion (MCF) research has now reached the level of scientific-feasibility experiments after having stayed in purgatory for about twenty years (to quote a well-known dictum by Artsimovich). Based on the results of these efforts the tokamak devices JT-60, TFTR, JET and T-l 5M are expected to become operative in the mid 1980s. By comparison, remarkable progress in ICF has been attained through only ten years' efforts; in this field, too, the level of scientific breakeven does not seem to be too far away.

The name tokamak (derived from the Russian for 'toroidal-chamber-magnetic') is applied to axially symmetric toroidal systems in which the plasma is confined by a strong toroidal magnetic field B_t , produced by an external toroidal solenoid, together with a weaker poloidal field B_p, produced mainly by a toroidal current Ip flowing in the plasma itself. The combination of the two fields produces nested toroidal magnetic surfaces composed of helical field lines. Equilibrium of the plasma is produced by the poloidal field, whilst the toroidal field serves to suppress the main magnetohydrodynamic instabilities, provided the 'safety factor' q = aB_t/RB_p is sufficiently large, where a and R are the plasma minor and major radii, respectively. The toroidal plasma current also provides 'Ohmic' heating of the plasma—the only heating mechanism operative in almost all early tokamak experiments. Particle orbits in tokamaks fall into two classes: there are 'passing' particles that travel completely around the torus, following helical field lines quite closely; and there are magnetically 'trapped' particles that travel back and forth along those parts of field lines that lie on the outer-major-radius side of the torus, undergoing reflection in the region of higher toroidal field on the inner side. In both cases, conservation of canonical toroidal angular momentum, which follows from the axisymmetry of the ideal tokamak configuration, implies that a particle's excursion away from a magnetic surface cannot exceed its gyroradius evaluated with the poloidal field B_p. In this sense, the ideal tokamak configuration confines 'all' particles; in a tokamak reactor, the 3.5-MeV alpha-particles will be confined provided the plasma current exceeds about 3 MA.

A numerical code based on a Δ'-analysis is applied to calculate the saturated amplitude of tearing modes dependent on the current density profile. The only stellarator effect.included is the additional, shearless external rotational transform ι̵₀ in the safety factor profile q(r). In this way, the stellarator field shifts the resonant q = 2 surface toward the outside, where the current density gradient is smaller, and stabilizes the (2, 1) mode as observed experimentally. Also the measured dependence of the (2, 1) mode amplitude on electron density and plasma current can be absolutely predicted by the calculations. – In addition to stabilizing the (2, 1) tearing mode, the current disruption is suppressed in Ohmically heated W VII-A discharges for ι̵₀ ≳ 0.15. The experimental findings, together with the calculated island widths, are compared with the predictions of a theoretical model proposed by several authors to explain the current disruption.

The evolution of magnetic islands is studied by a variational method on the assumption that it consists of a sequence of equilibria. The characteristic time of the evolution is then a resistive time. The sequence may, however, reach a configuration where the angle of the flux lines at the X-point vanishes. This behaviour is plausible in the case of q = 1 islands, in contrast to the case of q > 1. The subsequent evolution must assign a certain role to inertia. It is shown that this role cannot consist of a rapid displacement of the separatrix preserving its topology, but must be due to the onset of small-grain kinetic and magnetic turbulence extending from the separatrix in a large domain.

Data for the calculation of lower-hybrid wave ray trajectories in a tokamak plasma are quoted. It is demonstrated that under actual conditions the refractive index parallel to the total magnetic field varies considerably as the wave is propagated along the trajectory. This implies that there is a considerable change in the conditions of accessibility of the lower-hybrid resonance in a tokamak, as compared with a one-dimensional inhomogeneous plasma model. The accessibility of the resonance area is a strong function of the current distribution in the plasma. When the distribution is highly peaked the central area of the column becomes inaccessible to waves with any initial retardation. When the current profile is flatter, the conditions for accessibility are made easier to fulfil.

Consideration is given to some practical features associated with beta optimization in neutral-beam-heated tokamaks. Previous ideal MHD stability calculations of beta limits have employed ad hoc pressure and current density profiles. A coupled 1 – 1/2-D transport-ballooning eigenmode model was used to match observable experimental parameters and to simulate those not directly measured (such as the beam deposition and heating profiles), including neutral gas and impurity effects. With this model it was found that: (1) beta values exceeding the ideal MHD limits have been obtained, and (2) strong effects of direct beam contributions to the pressure, or indirect effects of strong local beam heating, can outweigh the relatively remote vacuum field shaping in well-penetrated discharges. Since beam-induced currents can produce significant changes in transport by changing the location and width of islands produced by saturated resistive tearing modes, these have been calculated self-consistently.

A simulation study of physical processes involved in ion cyclotron resonance heating and associated radio frequency plugging of a mirror confined plasma was carried out utilizing 1–2/2 dimensional magnetostatic particle code. Three heating regimes are discerned: weak, moderate and strong, characterized primarily by the pump wave amplitude. The pump waves induced other modes to various degrees, particularly so in the moderate and strong heating regimes. The heating increases ion magnetic moments; this property is applied to radio frequency plugging of a mirror. Modest improvements in confinement were observed in some moderate heating regimes. At the same time, pump-induced wave excitation enhances the velocity space diffusion of the ions, which tends to negate the increased magnetic moments. Induced excitation of Alfvén-ion cyclotron waves becomes intense when the heating results in a strongly anisotropic ion distribution, giving rise to the onset of the Alfvén-ion cyclotron instability. When the anisotropy becomes significant, the instability becomes rich in hydrodynamic characters and this is found to be very detrimental to confinement. A hydrodynamic model of this instability is constructed and examined in detail by simulation.

It is shown that drift instabilities in tokamaks may result in the high-energy ion loss only when the spectrum of excited waves is sufficiently wide. Otherwise the turbulent flux of such ions is fictitious and not associated with the loss of particles.

The energy of the electromagnetic noises occurring in the plasma during development of thermonuclear drift instabilities is evaluated. The conditions under which turbulent loss of α-particles takes place are derived. It is shown that when these instabilities develop in a toroidal device with a monotonically decreasing velocity distribution function of α-particles, these conditions are not fulfilled and there is no turbulent diffusion of alpha particles.

The kinetic Alfvén wave – the Alfvén wave with the perpendicular wavelength comparable to the ion gyroradius – is not only capable of heating a plasma but also of producing a current in the direction of the ambient magnetic field. To maintain the current density at the level of the Ohmic heating current in a tokamak, an oscillating magnetic field of 10⁻³ B₀ with a frequency of a few MHz, where B₀ is the magnitude of the toroidal magnetic field, is required.

The effect of toroidal magnetic field ripple on energetic ion confinement in Alcator A is theoretically investigated. A heuristic Monte Carlo particle simulation shows that for sufficiently large v∥/v the ion loss rate due to ripple trapping is independent of the toroidal fraction of the torus contained in the magnetic well. The resulting power loss is estimated to be large enough to limit the efficiency of RF heating on Alcator A.

The effect is considered of magnetic fluctuations on electron heating of tokamak plasmas by lower hybrid waves. For high power experiments, magnetic fluctuations can enhance both the electron heating rate and the anomalous loss rate. Each of these effects can reduce the efficiency of bulk electron heating; the outward shift in power deposition can increase losses even with the same transport rate, and the RF enhanced anomalous transport can provide a significantly increased loss rate.

A positive or negative current pulse induced by a pulsed toroidal electric field much higher than the Dreicer field increases the bulk-ion temperature of the plasma centre two to three times, without macroscopic plasma destruction. The decay time of the raised ion temperature agrees well with the prediction from neoclassical transport theory. The magnitude of the positive current pulse is limited by violent current disruption, and that of the negative current by a lack of MHD equilibrium which is due to a marked reduction of the total plasma current. The relevant current-driven instabilities in the turbulent heating of a tokamak plasma, skin heating and inward transfer of the energy deposition in the skin layer are briefly discussed.

It is attractive to produce a fat and hot plasma in a reversed-field configuration (RFC) by a small magnetic field since adiabatic compression – one of the best supplementary heating methods in RFC – is very effective. In the large-diameter theta pinch, a fat and hot (600–200 eV) RFC plasma could be produced in a small confining field (0.3 T). The heating mechanism was examined. It was found that the delayed pinch effect and the Joule heating due to anomalous plasma resistivity enable the production of such a plasma.

Because of the pressure and magnetic-field gradients and curvature of the magnetic field lines in a bundle divertor of a tokamak device, the plasma may be unstable to local interchange modes. Turbulent transport could be quite large and lead to a thick scrape-off layer which is as large as the radius of curvature of the diverted flux bundle. Such turbulence would be beneficial for lowering the energy and particle fluxes on the collector in a bundle divertor. The effect of a bundle divertor on the β-limit resulting from the ballooning modes of instability in the central plasma is also estimated. The critical β is reduced by less than one percent.

A numerical study shows that in a cylindrical tokamak the internal kink mode (m = 1) develops non-linearly into a helical equilibrium state that possesses a singular current sheet. In the large-aspect-ratio limit, the neighbouring equilibria obtained agree well with the asymptotic analytic theory of Rosenbluth et al.