Table of contents

Measurements of neutrons and hard X-rays are made with a pair of plastic scintillators during injection of deuterium pellets into deuterium TFTR plasmas. Three cases are investigated. During Ohmic heating in plasmas with few runaway electrons, the neutron emission does not increase when a pellet is injected, indicating that strong acceleration of the pellet ions does not occur. In Ohmic plasmas with low but detectable levels of runaway electrons, an X-ray burst is observed on a detector near the pellet injector as the pellet ablates, while a detector displaced 126° toroidally from the injector does not measure a synchronous burst. Reduced pellet penetration correlates with the presence of X-ray emission, suggesting that the origin of the burst is bremsstrahlung from runaway electrons that strike the solid pellet. In deuterium beam heated discharges, an increase in the d-d neutron emission is observed when the pellet ablates. In this case, the increase is due to fusion reactions between beam ions and the high density neutral and plasma cloud produced by ablation of the pellet; this localized density perturbation equilibrates in about 700 μs. Analysis of the data indicates that the density propagates without forming a sharp shock front with a rapid initial propagation velocity (≳ 2 × 10⁷ cm·s⁻¹) that subsequently decreases to around 3 × 10⁶ cm·s⁻¹. Modelling suggests that the electron heat flux into the pellet cloud is much less than the classical Spitzer value.

Series of ohmically heated deuterium and hydrogen discharges in ASDEX, with scans of plasma current and density, and neutral-beam-heated discharges were used to investigate the particle transport by means of computer simulations. It was found that the anomalously high inward drift velocity can be well described by the thermoelectric-type expression v_in(r) = 0.5 D(r) T_e(r)⁻¹ ∂T_e/∂r, where D is the anomalous diffusion coefficient. This relation is similar to the classical result derived from the thermal friction force. The parameter dependences of the inward drift velocity are determined from the scans and explained by the thermoelectric inward flux. An empirical scaling law for the electron heat diffusivity in ohmically heated plasmas is presented.

High aspect ratio pellet targets are ablatively imploded with twelve green beams of GEKKO XII. Multiplexing of shock waves with a quasi-tailored pulse is performed successfully in DT fuel compression. The highest neutron yield achieved is 1.25 × 10¹² per shot. The development of new diagnostics such as neutron activation and secondary fusion reaction diagnostics is also presented.

The heating of torsatron devices by electron cyclotron waves has been investigated using ray tracing techniques. The geometry of the magnetic field and mod-B contours in torsatrons is very different from that found in tokamaks or in most conventional stellarators. The sensitivities of electron cyclotron heating to antenna position, antenna orientation, wave polarization and magnetic field strength are therefore correspondingly different. These sensitivities are identified and quantitative calculations of wave propagation, absorption and radial heating profiles are carried out to develop a systematic picture of heating in these devices. A simple model is introduced which statistically treats the effect of power reflected from the walls. Specific ray tracing calculations were performed using the RAYS geometrical optics code for the Advanced Toroidal Facility torsatron now under construction at Oak Ridge National Laboratory. Significant power absorption (up to 100% of incident wave power) can be obtained with an optimally placed and oriented beam. This is the case when the resonance is at the saddle point in the magnetic field. These results can be achieved through either ordinary mode waves at the fundamental resonance (high or low field launch) or extraordinary mode waves at the second harmonic resonance (low field launch). These results are quite sensitive to antenna placement, antenna orientation and the magnetic field level. Depending on experimental constraints (such as high field or inside accessibility), other considerations can come into play; for example, there are theoretical advantages to launching the waves horizontally in the equatorial plane. The qualitative results from the wall reflection model indicate that polarization control and maximization of the first-pass absorption are more important when the central plasma is resonant near the fundamental resonance thanwhen it is resonant at the second harmonic. Finally, a comparison of the calculations with data from the Heliotron-E torsatron is made and found to be in qualitative agreement.

Energy confinement is studied in lower hybrid current driven (LHCD) plasmas in Alcator C in the density range _e = (1–8) × 10¹³ cm⁻³. In LHCD plasmas, the stored energy in the electron tail, W^tail, can be a significant fraction of the total stored energy, W^tot, especially at lower densities. At sufficiently low densities, the confinement time of the high energy electrons is expected to become shorter than their collisional slowing down time, and direct energy losses from the electron tail can become important in the overall power balance. The global energy confinement time, defined as is found to be comparable to or exceed that in ohmically heated (OH) plasmas at low densities _e ⪅ 3 × 10¹³ cm⁻³, where a steady state current can be maintained with relatively low RF power. However, at higher densities where substantially more RF power is needed (relative to the Ohmic power required to maintain a similar plasma), a deterioration of relative to Ohmic confinement, similar in magnitude to that predicted by the neutral beam heated L-mode scaling, is observed. Theoretical modelling with the aid of a ray tracing Fokker-Planck transport code suggests that the deteriorated confinement in this high density, high power regime may be attributed to an enhanced bulk electron thermal diffusivity. In a combined OH-LHCD plasma, a value of greater than the Ohmic value is obtained as long as the applied RF power does not significantly exceed the Ohmic power.

Simple models for the dissipative trapped electron, circulating electron and toroidal ion temperature gradient drift modes, together with the effect of low-m tearing mode islands, are incorporated into a standard tokamak transport code to study anomalous energy transport in tokamaks, primarily in the Doublet HI tokamak. In simulations of Doublet III in the Ohmic regime, low density neo-Alcator scaling is obtained. At high density the ion temperature gradient mode contributes strongly to the saturation of the energy confinement time. In the beam heating regime, the model predicts that the energy confinement time has weak density and magnetic field dependence, degradation with power and favourable current scaling. The generality of the drift wave model is further tested by simulating smaller tokamaks (ISX-A, Alcator A, Alcator C) as well as the largest tokamak, JET.

Large poloidal asymmetries in density, electron temperature, radial density e-folding length and floating potential have been measured in the plasma existing between the limiter radius and the wall of the Alcator C tokamak. Typically, variations in density by factors of about 4–20 and variations in radial density e-folding length by factors of about 3–8 are recorded in discharges which are bounded by poloidally symmetric ring limiters. These poloidal asymmetries show that pressure is a function of poloidal angle on open magnetic flux surfaces in this region of the plasma. Observations of toroidally symmetric MARFE phenomena further imply that density and perhaps pressure are also a function of poloidal angle on closed flux surfaces existing just inside the limiter radius. The magnitude of these poloidal asymmetries and their dependence on poloidal angle persists independent of machine parameters (central plasma density, plasma current, toroidal field, MARFE versus non-MARFE discharges). Analysis of the data indicates that these asymmetries are caused by poloidal variations in perpendicular particle and heat transport in both the main plasma and the scrape-off layer. A number of possible asymmetric perpendicular transport processes in the scrape-off layer plasma are examined, including diffusion and × plasma convection.

An MHD formulation is used in the study of a wall-stabilized high beta mirror plasma with isotropic pressure. The stabilizing wall extends axially only over a part of the distance between the mirror midplane and the throat. In a model of this arrangement, a wall is used which is near the plasma surface in the bad curvature region and distant from the plasma surface in the good curvature region. A variational method is used to solve the equations for both regions. For the connection of the regions of close and distant plasma-wall proximity, a jump condition is used. The variational calculation is performed with a simple trial function (the choice of the trial function is substantiated with an exact numerical solution). The results show that (a) the removal of the conducting wall in the good curvature region does not significantly degrade plasma stability, (b) the acceptable ratio of the radius of the conducting wall to the plasma radius is about 1.1, and (c) for cases with a low mirror ratio, more conducting wall is needed for stability than for cases with a high mirror ratio. This is in agreement with the physical mechanism of wall stabilization.

A modulation of microturbulence is observed in TEXTOR during low mode number MHD activity using CO₂ laser collective scattering. This is accompanied by a strong enhancement of density fluctuations near ka_s ≈ 3 at the end of soft landing discharges and a displacement of the frequency spectrum towards lower frequencies. The increase is most significant for rapid rampdown of the plasma current accompanied by strong MHD activity but also occurs when the latter is not detected. The evolution of microturbulence is also studied during major plasma disruptions. It was found that disruptions without MHD precursor oscillations are characterized by a rapid increase in the density fluctuations starting ~100 ms before plasma disruption.

It is found that the energy dependent diffusion theory provides excellent accuracy in the modelling of transport of neutral atoms in fusion plasmas. Two reasons in particular explain the good accuracy. First, while the plasma is optically thick for low energy neutrals, it is optically thin for high energy neutrals and the diffusion theory with Marshak boundary conditions gives accurate results for an optically thin medium, even for small values of c, the ratio of the scattering cross-section to the total cross-section. Second, the effective value of c at low energy is very close to 1 because of the downscattering via collisions of high energy neutrals. The first reason is proven computationally and theoretically by solving the transport equation in a power series in c and solving the diffusion equation with 'general' Marshak boundary conditions. The second reason is established numerically by comparing the results from a one-dimensional, general geometry, multigroup diffusion theory code, written for this purpose, with the results obtained using the transport code ANISN. Earlier studies comparing one-speed diffusion and transport theory indicated that the diffusion theory would be inaccurate. A detailed analysis shows that this conclusion is limited to a very specific case. Surprisingly, for a very wide range of conditions and when energy dependence is included, the diffusion theory is highly accurate.

Measurements of the gradient of the d-d fusion rate profile in an Ohmic PLT plasma are used to deduce the gradient of the ion temperature and thus the local ion thermal diffusivity through an energy balance analysis. The inferred ion diffusivity is consistent with neoclassical theory.

A non-linear analysis of parametric instabilities in lower hybrid heating of tokamak plasmas is presented. Harmonic generation of the low frequency mode and its effect on the evolution of the parametric decay of a lower hybrid pump to another lower hybrid wave and a low frequency mode is investigated, and numerical calculations of the relevant wave system are carried out. It is found that the coupling of the low frequency wave to its harmonic may effectively suppress the spatial growth of the primary instability. This may be particularly true at the plasma edge of the tokamak, where electron and ion temperatures are low and the primary instability is strong at high pump intensity. The spectral dependence of the harmonic generation strength is compared with that of the primary instability, and the effect on pump depletion with relevant tokamak parameters is discussed.

Using an approximation of the stability diagram for the high n ballooning modes calculated for an axisymmetric configuration with circular cross-section, the authors show that a flat current density profile permits access to the second stability domain and leads to β values higher than those obtained with usual peaked current density profiles.

Local cavity resonances of fast wave propagation in a hydrogen-deuterium tokamak plasma have been investigated experimentally. The results demonstrate short damping lengths of the fast waves on the high field side of the hybrid layer, and the existence of a local cavity resonance on the low field side, even for a low hydrogen minority concentration (5%) in a deuterium plasma. The strong wave damping disappears when the hybrid layer is moved towards the edge of the plasma.

Current drive by radiofrequency waves is considered for a preionization plasma. The ratio J/P of the driven current density and the absorbed power density is computed, taking into account both Coulomb and electronneutral collisions. A considerable modification of the fully ionized plasma case is found.

In order to achieve a ratio of total fusion output to injected power (Q) of more than one, with minimum tritium consumption in reacting plasma experiments, tritium pellet injections into deuterium plasmas heated by neutral deuterium beams with E_beam > 100 keV are promising. In the computation presented, a tokamak transport code is combined with a pellet ablation code based on a neutral shielding model including fast ion effects in the ablation process. Q > 1 is realized during half a second by injecting a pellet with a radius of 2 mm and a velocity of 5 km·s⁻¹ into a plasma of T_i(0) = 25-30 keV, T_e(0) = 9.5 keV and n_e(0) = 8.5 × 10¹³ cm⁻³, i.e. a 'hot ion mode' plasma. The tokamak parameters are: a (minor radius) = 0.9 m, R₀ (major radius) = 3.0 m, B_t (toroidal magnetic field) = 4.5 T, and I_p = 2.7 MA. The anomalous electron transport is assumed to follow Alcator scaling and the ion transport is assumed to be faster than neoclassical. The total neutral beam power is 17 MW. The total energy confinement time is approximately 0.5 s and the average electron density is around 6.5 × 10¹³ cm⁻³ during Q > 1. For larger pellet radii, smaller injection velocities appear feasible.

Very high fusion reactor net efficiencies (>65%) are shown to be feasible through a new mode of advanced fuel tandem mirror operation. Inducing nonadiabaticity for fusion products not needed to sustain the plasma against losses enhances their end loss rate, leading to a narrow end loss energy spectrum and resultant very high efficiency direct electrostatic conversion. Magnetic field gradients required for non-adiabaticity are estimated, an approximate power balance is calculated, and a reference case is generated. Operation of D-³He reactors in this mode is particularly interesting, because of the recent discovery of a substantial ³He resource on the lunar surface.

The Second International Conference on Fusion Reactor Materials (ICFRM-2) was held on April 13-17, 1986, in Chicago, Illinois, USA. The objective of this conference was to provide an international forum for the exchange of new information and the furtherance of understanding of the science and technology of materials proposed for use in fusion reactors. The international nature of this conference can be seen in the attendance which was composed of 57% from the United States, 24% from the European Community, and 18% from Japan. This conference represented a merging of the three US topical meetings on fusion reactor materials (Miami, January 1979; Seattle, August 1981; and Albuquerque, September 1983) and the First International Conference on Fusion Reactor Materials (Tokyo, December 1984). Conference sponsors were the Argonne National Laboratory and the US Department of Energy, with minor sponsorship from the American Nuclear Society and the Nuclear Metallurgy Committee of the AIME and ASM. The technical programme contained a total of 254 technical papers, of which 59 were oral and the balance posterboard presentations. The conference proceedings will be published in a special supplement to the Journal of Nuclear Materials (Volumes 141, 142, and 143).

The meeting took place at the Eibsee Hotel, near Garmisch-Partenkirchen in the Federal Republic of Germany, May 13-16th 1986. It was a relatively small meeting held in a beautiful location; there were approximately 80 delegates, most of them from Europe but with some from Australia, Japan, USA and elsewhere. The programme was a little unusual in that it brought together computational physics specialists working on astrophysical, magnetic confinement and inertial confinement plasma physics problems. Despite the apparent diversity of interests, the conference was lively and fruitful. Different emphasis of attack to similar numerical and computational problems faced in all three areas gave insight not usual in more narrowly defined topic meetings.