Search

All major JET systems have been fully commissioned for D-T and the DTE1 series of experiments has started with the D-T fuel mixture and operating conditions foreseen for ITER. In the area of ITER physics, significant results have been produced in both D-D and D-T. In D-D, the LH threshold power database has been extended, the bounds on edge-electron temperature and density in ELMy H-modes have been defined and the advantages of Types I and III ELMy discharges have been compared. In D-T plasmas, the isotope effect on H-mode threshold power and transport has been determined so that a more accurate assessment can be made of the ignition margin and heating requirements for ITER. Trace tritium experiments have provided first particle transport measurements and an assessment of the ITER reference ion-cyclotron resonance-frequency heating scenarios has been started, In the area of fusion performance, record D-D neutron yields have been obtained by controlling the plasma and current profiles in hot ion ELM-free H-modes and optimized shear modes. In D-T, internal transport barriers have been readily established in optimized shear discharges and Alfven eigenmodes have been observed.

Impurity fuelling has been studied in TEXTOR by gas puffing under a wide range of experimental conditions. First, neon has been injected into discharges with a range of densities during both Ohmic and neutral injection heating. Then, helium, neon and argon have been injected into similar discharges and the increase in radiated power and Zeff studied as a function of the number of impurity atoms injected. Similar experiments using carbon monoxide and methane yielded markedly different results from those of the rare gases. The fraction of injected atoms entering the confined plasma has been estimated and found to vary widely, from less than 1% for carbon to about 100% for helium. Detailed Monte Carlo calculations performed using the LIM code to simulate the impurity behaviour have shown that good agreement with experiment can be obtained for all species except oxygen, where charge exchange processes are thought to be important

During the JET Preliminary Tritium Experiment (PTE), an estimated 2 × 1012 (1.1 × 1021 atoms) of tritium were injected into the JET vacuum vessel. A series of experiments was performed whose purpose was to deplete the torus of tritium, to compare the effectiveness of different methods of tritium removal and to obtain a quantitative understanding of the processes involved. The effectiveness of the cleaning procedures was such that the normal tokamak programme was resumed one week after the PTE and the routing of exhaust gases to the atmosphere after two weeks. The release of tritium from the vessel was found to scale with the deuterium release from the vessel, suggesting that dilution and mixing of the hydrogen isotopes in the vessel walls is important. High density, disruptive tokamak discharges were found to be the most successful plasma pulses for tritium removal. Purges with deuterium gas were also effective and have the advantage of operational simplicity. Helium discharges, on the other hand, resulted in low tritium release from the vessel walls. It was demonstrated that the tritium release rate could be predicted using data from hydrogen to deuterium changeover experiments. Using the superior quality of data available from the tritium cleanup experiment, the physical mechanisms necessary to describe the hydrogenic uptake and release from the JET torus were identified. Tile release of tritium is reproduced using a model that incorporates implantation into a thin surface layer as well as diffusion of tritium into and out of the bulk material

The START (Small Tight Aspect Ratio Tokamak) experiment became operational in January 1991 at AEA Fusion Culham Laboratory. It is the only tokamak producing hot plasmas at aspect ratios as low as A=R/a approximately 1.3, and is providing valuable insight into the physical processes (such as toroidicity and trapping) involved at low aspect ratio. Three topics are discussed: equilibrium properties, including evidence of neo-classical and bootstrap effects; energy confinement studies; and MHD properties, including possible explanations for the apparent absence of the major disruption at low A.

Low‐aspect‐ratio tokamaks offer both the economic advantage of smaller size and a number of physics advantages which are not available at conventional aspect ratio. The Small Tight Aspect Ratio Tokamak (START) [Fusion Technology 1990, edited by B. E. Keen, M. Huguet, and R. Hemsworth (North‐Holland, Amsterdam, 1991), Vol. 1, p. 353] was conceived as a first substantial test of tokamak plasma behavior at low aspect ratio. It has achieved plasma currents up to 200 kA, peak densities of ∼2×1020 m−3 and central electron temperatures of ∼500 eV at an aspect ratio of 1.3–1.5. Central beta values of ∼13% have been measured and the volume‐averaged beta 〈β〉 can approach the Troyon limit. Plasmas are naturally elongated (κ≲2.0) and are vertically stable without feedback control. Major disruptions have not been observed at low aspect ratios (A≤2.0).

The radial distributions of the low charge states of Be near the JET beryllium belt limiters have been measured experimentally. When interpreted with an impurity production/transport code, these measurements elucidate the plasma behaviour near the limiter, and the source and sink action of the limiter on the impurity ions. We have considered here 3 MA ohmically heated, limiter discharges with ¯n e =1-4×10 19 m -3 . A model which satisfactorily reproduces the measured profiles required the following assumptions: (a) an n e ( r ) profile just inside the LCFS (0 to ≈10 cm radially inwards) which extrapolated smoothly from the SOL (probe) measurements; (b) a T c ( r ) outer edge profile which extrapolated smoothly from the core (ECE) measurements; (c) a cross-field impurity diffusion coefficient D ⊥ ≈0.5 m 2 /s in the outer edge; (d) no strong inward or outward pinch action in the outer edge; (e) impurity production by physical sputtering, rather than by evaporation; (f) sputtering due to parallel plasma flow onto the limiters; no significant direct perpendicular flow onto the limiters, as has been observed for larger limiters.

The first tritium fuelled discharges in JET were followed by a series of cleanup discharges aimed at removing tritium from the torus. Measurements of tritium in the plasma and the exhaust were used to assess the efficiency of the various cleanup techniques. A preliminary estimate of the in-vessel tritium inventory is presented.

Impurity gases have been introduced into TEXTOR at well defined rates to study the fuelling process and the effect they have on the overall energy balance. It is found that carbon and oxygen, introduced as CO and CH 4 gases have radically different behaviour. Oxygen has a high recycling coefficient ∼-0.99 whereas carbon has a low recycling coefficient ∼-0.3. The radial position of the gas source was varied from the wall to the limiter radius but this made little difference to the effect on the global plasma parameters. For very high impurity influxes of both impurities drastic changes in the edge plasma were observed and disruptions could be induced.

Automatic exposure control of an asynchronous CCD camera has been demonstrated on TEXTOR, JET and COMPASS and in a single field, random event mode, on TVD and START. The system has been successful in correctly exposing and timing variable intensities requiring a range of exposures from 0.2–300 ms in a single discharge.

Isotope exchange experiments have been performed in JET using hydrogen, deuterium, and, in the recent preliminary tritium experiment (PTE), tritium. The rate of change-over from one isotope to another involves two quite different time constants. We have modelled this behaviour using a multireservoir model which splits the accessible hydrogenic particles into two groups, each having a different rate of exchange of particles with the plasma. By applying this model to the sequence of discharges during and after the PTE, we can determine the parameters in the model. The resulting fit also gives a good representation of hydrogen/deuterium change-over experiments, indicating that the tritium behaves in the same manner as other hydrogen isotopes, at least as far as recycling is concerned. Discrepancies between the model and the actual measurements of tritium recovery after the PTE lead us to conclude that isotope exchange processes resulting from collisions of molecules with the vessel walls play a significant role in spreading tritium around the machine.

Measurements have been made of the tritium recovery from the JET pumping system both after high power neutral beam injection where the tritium is produced from D(d, p)T reactions and after tritium is injected into the torus in the first tritium experiment (FTE). The tritium has been detected using a liquid scintillation counter which has allowed measurements to be made over a wide range of concentrations. The recovery rate has been found to follow a power law R=At−n over the period from 10 to 3000 s after each discharge. The value of n varies from 0.8 when tritium is first introduced into the plasma to ∼0.5 many discharges later. This effect is attributed to a change in the depth distribution of the tritium in the surface.

Low aspect ratio tokamaks may provide a low cost route to developing magnetic fusion energy and may offer simpler and more economic fusion reactors. The first test of this concept with hot plasmas has been conducted with the Small Tight Aspect Ratio Tokamak (START) experiment now in operation at Culham Laboratory. The novel method of plasma formation is described and observations of plasmas at aspect ratios as low as 1.3 are presented, indicating high temperature discharges with good global stability properties and typical tokamak characteristics. This represents a new and significant extension of the parameter range of the conventional tokamak

The energy distributions of tritons and helium ions escaping to the wall and limiters in JET have been measured. The tritons are produced from the D-D reaction during NBI heating and are born with an energy of 1 MeV, while the 3 He ions have been accelerated by ICRH. The tritium is collected in carbon samples placed in a probe which discriminates energy on the basis of larmor radius and the ratio of perpendicular to parallel velocity. The helium ion energy is deduced by measuring the depth distribution in a nickel collector. The average energy of the escaping tritons is not greater than 6 keV and is probably thermal. The energy distribution of the helium is very broad, most ions having energies less than 30 keV but some having energies greater than 1 MeV.

In this work the results of gas balance measurements during the beryllium phases of JET are presented and compared with similar data from the carbon phase. Gas balance calculations are carried out by measuring the gas release after a discharge and comparing it to the total fuelling. These measurements give direct information on the in-vessel retention of hydrogen isotopes. After the introduction of beryllium the fuelling requirements of a typical discharge increased by a factor of ∼- 2 and the measured gas release increased substantially. The average deuterium recovery after a non-disruptive shot is ∼- 35–50% of the input. This percentage increases to ∼- 80% for long term measurements. The deuterium inventory in the walls is similar for C and Be. Some tritium release measurements are presented. The fraction of tritium recovered, both following the discharge and on the long term, is about half of the measured fraction of deuterium. The analysis of D and T release data leads to the observation that surface effects dominate the gas release after the discharge.

START continues to provide key results on the 'spherical tokamak' concept. This paper outlines the engineering solutions adopted to overcome problems of space restrictions in a very low aspect ratio tokamak, whilst providing high performance at low cost.

Time-resolved arcs have been observed in the DITE tokamak using three independent techniques, the observation of the current flowing to a probe, the spectral emission from the probe material, and the arc tracks on a rotating disc. The results obtained with the different methods are in good agreement and show that arcing typically occurs in DITE during the first 20 ms of the discharge. This short period of arcing is consistent with measurements of the metal impurity flux at the wall, which show a similar time dependence. Observations of the arc current as a function of radial position show that arcing is negligible 3 cm beyond the radius of the fixed limiter, in good agreement with the observed arc tracks on projecting surfaces of the torus. The erosion of a probe was estimated from the charge flowing in the arcs and is consistent with previous measurements of the metal impurities deposited at the torus wall.

Measurements have been made of the plasma parameters in the DITE bundle divertor using an array of heat flux probes which can be scanned horizontally and vertically through the diverted flux bundle. Horizontal profiles of the saturation ion current, the electron temperature and the deposited power have been obtained for both ohmic and neutral beam heated discharges. Complete contours of these parameters have been obtained for the beam heated discharges. Maximum current densities of 7 A cm−2, temperatures of 20–30 eV and deposited power of 25 MW m−2 have been observed. In addition surface collector probes exposed for 163 discharges have been analysed and the main impurity shown to be stainless steel. The impurities were found to have penetrated ~100 nm into the carbon collector and this is consistent with diffusion.

The incidence of arcing in DITE as a function of the plasma parameters has been investigated using a time-resolved arc detector. The arcs are detected by observing the current and the optical emission from a concentric probe at the same minor radius as the fixed limiter. The measured currents are in good agreement with the estimated currents previously obtained from the track dimensions on probes. The arcing signals were correlated with pulses in the loop voltage, MHD activity, and rational values of the safety factor q (a) . Arcing during the initial phase of the discharge was reduced by a factor 5 by reducing the plasma current rise-time from 11 to 3 ms. Simultaneous measurements of the Ti flux and TiII and TiXVIII spectral lines did not, however, show a significant reduction in impurities for most of the discharge.

Erosion of limiters, walls and divertor plates is likely to be a serious concern in the next generation of fusion devices. The main mechanisms of erosion are reviewed and the spatial distribution of erosion and deposition are discussed. Measurements of erosion and redeposition have been made at the JET limiters under different conditions and net erosion in the range 10–100 μum detected after typically 3000 10 s discharges. Modelling of the distribution is in quite good agreement with the experimental data and net to gross erosion ratios of 0.2 to 0.3 are calculated. The problem of extrapolating to reactor conditions is discussed. While there are serious difficulties in carrying out this extrapolation, the higher edge densities envisaged should lead to lower ratios of net to gross erosion than in present experiments.

Results from the DITE tokamak experiment with the Mk II bundle divertor and neutral beam heating are described. Particle and energy balances have been investigated and both show substantial fractions of the inputs flowing to the divertor. Observations of asymmetries in power deposition in the torus are presented. Measurements of the plasma parameters in the divertor scrape-off layer and in the diverted plasma are reported. Flow into the divertor is adequately described by convective processes.

The transport of impurity ions originating at the limiter in a tokamak is critically dependent on the location of the ion in the boundary plasma. In the confined plasma, just inboard of the limiter, impurity ions will disperse freely into the discharge whilst in the scrape-off layer the pre-sheath plasma flow and the associated ambipolar electric field may tend to sweep impurities back to the limiter surface. In this paper we have studied, both by experiment and by theory, the transport of carbon impurity ions in the vicinity of the limiter. By comparing experimental measurements of the spatial distributions of impurities around the limiter with that predicted from a Monte Carlo computer code it appears that the parallel dispersal on closed field lines in the confined plasma is consistent with classical transport processes and that in the scrape-off layer the dispersal is indeed impeded by the pre-sheath plasma flow.

An experiment to test beryllium as a limiter material has been performed in the ISX-B tokamak. The effect of the plasma on the limiter and the effect of the limiter on the plasma were studied in detail. Heat and particle fluxes to the limiter were measured, and limiter damage by melting was documented as a function of power flux. Strong melting and evaporation of the limiter caused beryllium gettering of the vacuum vessel. Postmortem analysis of the limiter was performed to document the amount of retained hydrogen and the erosion and impurity deposition on the limiter. The effect of the limiter on the plasma performance was studied in terms of parameter space, impurity content, and confinement for the ungettered and gettered cases. Operational experience with beryllium in a fusion experiment is discussed.

The DITE bundle divertor MkII has been operated with high power neutral beam injection, up to 2.1 MW, and with gaseous fuelling in the divertor. Discharges with good confinement have been used to study the properties of the plasma exhaust channel in the presence of a high density of neutral molecules in the divertor, up to ~ 1 ×1021 m−3. An inner layer of the channel carries a large part of the heat exhaust by collisionless convection at a power density ~ 100 MW m−2 and particle flux ~ 1024 m−2 s−1. Ionisation in the divertor has little effect on this layer. A collisional, low power density outer layer is strongly affected by the neutral particles in the divertor, with plasma flow out of the divertor at high neutral pressure, and ionisation losses supplied by electron heat transport from the torus. This layer has not previously been demonstrated in the plasma exhaust channel of a fusion research device. Calculations using the DEGAS code predict blocking of the divertor ducts by the exhaust plasma: retention of neutral particles in the divertor is seen experimentally. The amount of ionisation in the divertor is however lower than predicted, and it is shown that collisional interaction between neutral particles from the diverted plasma and the molecular background could be responsible for this reduction.

Unipolar arcing has been observed in the DITE tokamak. These arcs occur on the fixed limiter, on probes inserted in the plasma and on parts of the torus structure. The amount of material removed by arcing is consistent with the concentration of metal in the plasma. A measurement of erosion of the divertor target by the arcing has been made by the thin layer activation technique. A 25 μm layer at the surface of a titanium sample is activated using the 48 Ti (p, n) 48 V reaction and the removal of this layer is monitored by measuring the reduction in γ-rays emitted. Another form of surface erosion observed is the presence of craters and well defined cylindrical holes. Large numbers of these holes are present near regions which have been locally melted by runaway electron beams.

Measurements have been made of the number of metal atoms deposited on collectors during tokamak discharges in DITE. The amount of metal is shown to be larger than that expected due to sputtering by hydrogen ions and charge-exchange neutrals. It is proposed that arcing is responsible for this metal deposition and direct observation of arc tracks on limiters and probes is described.

High speed cine photography is a useful diagnostic aid for studying plasma behaviour and plasma surface interactions. Several workers have filmed discharges in tokamaks including ASDEX, DITE, DIVA, ISX, JFT2, TFR and PLT. These films are discussed and examples given of the observed phenomena which include plasma limiter interactions, diverted discharges, disruptions, magnetic islands and moving glowing objects often known as ‘UFOs’. Examples of plasma structures in ASDEX and DITE not previously published are also given. The paper also reports experiments in DITE to determine the origin of UFOs.

Wall long term samples (LTS) which were exposed during the 1984 and 1985 operational periods of JET have been investigated by means of surface analysis techniques. The results indicate that deposition of wall material (Fe, Cr, Ni) and its distribution on the vessel wall were strongly affected by wall erosion on bellows protection plates during unstable discharge phases. Erosion sources and deposition sinks are spatially correlated indicating that most of the eroded material is deposited close to its source.

Physical measurements of the JET graphite limiter tiles, made before and after exposure to ~ 3000 tokamak discharges, have shown that there is erosion of the surface nearest the plasma and redeposition of the carbon radially further out. The spatial distribution of the erosion/redeposition agrees with a previously proposed analytical model. The sputter yield of the redeposited layer has been measured and found to be similar to bulk material. Metals from the antennae screens are injected into JET discharges during ICRF heating. The spatial redistribution of these metal impurities on the limiter has also been studied. The observed alternate layers of metal and redeposited carbon explains the decay of the metal concentration in the plasma in ohmic discharges following ICRF heated discharges.

Detached discharges are described in which the radiated power loss is from within a maximum radius up to 50 mm less than the limiter radius. Outside this radiating zone the plasma density and temperature profiles are flat with values of approximately 5 × 1018 m−3 and 7 eV respectively. Typical conditions for detached discharges in DITE are a plasma current of 100 kA ,a safety factor of ≈ 7 (toroidal field 2.5 T), and a line average density of 3 × 1019 m−3 in deuterium. Detachment can also be observed in similar discharges in helium at a density of 6 × 1019 m−3. These values are 70–100% of the critical density for disruption.

The latest analysis of results from the Bundle Divertor and prospects for application to a tokamak reactor are presented. The upgraded version of the machine, DITE IB is described briefly. Initial results of a programme to study the causes of disruption at high density in ohmic and beam-heated discharges are also described.

The authors discuss the impurity control limiter (ICL) which has an inverted geometry. The ICL shape is designed to direct the impurities towards the wall. They present the results from a two-dimensional neutral particle code which maps the ionisation of carbon physically sputtered by deuterons from a carbon limiter. This ionisation source is coupled to a one-dimensional impurity transport code which calculates the implied central impurity density. The results demonstrate that the ICL achieves impurity control in two ways. Firstly, many of the sputtered impurities directed towards the wall are not ionised and return to the wall as neutrals. Secondly, much of the ionisation which does occur is located in the scrape-off layer. They conclude that a reduction in central impurity density of a factor of 10 is possible in a Tokamak such as DITE provided that the limiter is the main source of impurities.

An experiment has been performed on the Impurity Study Experiment (ISX‐B) tokamak to test beryllium as a limiter material. Beryllium is an attractive candidate for a limiter and has been proposed for use in the Joint European Torus (JET) experiment. A temperature‐controlled, segmented, beryllium top‐rail limiter was located inside the plasma radius described by the existing titanium limiters. An extended set of diagnostics was added for measurement of scrapeoff and limiter parameters. These included visible and infrared monitoring systems, probes, and surface analysis experiments. Tokamak experiments included parameter surveys of both Ohmically heated and neutral‐beam‐heated plasmas and an extended fluence test of the limiter. The most significant effect of operation with a beryllium limiter was the reduction in low‐Z impurities caused by gettering action of beryllium deposited on the liner walls. The experiment required the design and implementation of contamination control apparatus and work procedures ...

A series of experiments has been performed on the DITE tokamak in which helium and methane have been puffed from the tip of an instrumented probe limiter immersed in the edge of a helium plasma. The spatial distributions of He I, He II, C I, C II and C III have been obtained with a CCD camera using interference filters. The intensity distributions for He I, C I and C II are very localised while in the cases of He II and C III, streaming along field lines has been observed. The toroidal distributions are in approximate agreement with a parallel transport model that includes ionisation, heating and diffusion.

This paper presents a review of hydrogen and helium recycling phenomena in tokamaks with limiters and walls largely made out of carbon (graphite, a-C:H layers). The key points of interest are the plasma fuelling efficiency, the wall pumping phenomena as observed in JET, TFTR, TEXTOR and other machines under various fuelling schemes (gas, neutral beams, pellets), the release of hydrogen/helium from material surfaces during and after plasma discharges and the long term retention (total particle inventory) of hydrogen in graphite or carbonised structures in tokamaks. The effect of a combined hydrogen/helium plasma on recycling is also discussed. It is shown that only part of the above phenomena can be understood in terms of processes between hydrogen/helium and carbon as known from simulation experiments (ion beams, gas discharge facilities) and that others (in particular the JET wall pumping phenomenon) have still to be explained. Possible mechanisms are outlined and discussed by means of global models.

Characteristics of the operation of the DITE tokamak with the MkII bundle divertor and neutral beam heating are described. In these experiments the divertor was pumped by a liquid helium cooled cryopump, and fuelling was by gas input into the tokamak vacuum vessel. Generally a fuelling rate of 1–2 × 1021 atoms/s is required to sustain a constant density with the divertor operating, and about 50% of the gas input flows to the divertor. Global power balances have been carried out using instrumentation distributed around the tokamak and in the divertor. It is found that the wall loading exhibits substantial toroidal and poloidal asymmetries, with a high loading near the gas feed when the divertor is operating. The percentage power exhaust to the divertor is highest, up to 70%, in ohmic discharges with the gas feed turned off and the density falling. With 1.4 MW of neutral beam heating, about 24% of the power is seen at the divertor plates. Typically, between 70% and 80% of the input power can be counted for. Measurements in the scrape-off plasma show that the density is about 1× 1019 m−3 at 30 to 40 mm outside the separatrix, when the line average density in the tokamak is 2−2.5 × 1019 m−3. However there is less than 10% reionisation of recycled neutrals in the divertor, as determined from Hα measurements. In addition, the strong coupling observed between the particle and power flow to the divertor suggests that the exhaust is convective, and this may be a consequence of the strong mirror field characteristic of the bundle divertor magnetic geometry.

In the JET tokamak concentrations of tritium have been observed m graphite many orders of magnitude above the background level. By examining limiters, wall tiles and small graphite wall samples, a spatial distribution for tritium was obtained. The distribution shows a marked asymmetry with a maximum at the top of the torus. The depth of penetration of the tritium into the graphite has also been measured and this, together with the spatial distribution, enables estimates of the tritium inventory to be made.

A series of measurements of the plasma parameters in the boundary layer have been made for discharges using a top rail limiter and neutral beam heating. Results have been obtained using heat flux/Langmuir probes, an EXB gridded energy analyser and with an infra red camera. Profiles of the saturation ion current, the electron temperature and the power deposited on the limiter have been taken. A large difference in the e-folding length between the ion and electron drift sides of the probe was observed and this is attributed to the difference in their respective connection lengths. The e-folding distances have been measured over a wide range of plasma parameters including plasma current, density and beam heating power. No significant change in the cross field diffusion coefficient was observed during these variations, the value being typically ( 1.5 ± 0.5 ) m2 s−1.

An experiment has been undertaken to see if chemical sputtering of carbon can be directly observed under tokamak operating conditions. The test material consists of a well instrumented graphite probe limiter which can be heated up to 900°C under steady state conditions and up to >1200° C with additional plasma heating during a discharge. The carbon released from the limiter is monitored spectroscopically by using a visible spectrometer and a CCD camera with interference filters. Measurements have been made from 20°C to 700°C in steady state conditions and up to 1200°C with plasma heating. Within experimental error, there is no change in the carbon production rate over the whole temperature range. The inferred chemical sputtering yields are low, in agreement with beam simulations that indicate reduced yields at high incident current densities, low energies and in the presence of metal surface contaminants.

A proposal has been prepared for three facilities for diagnosing the plasma surface interactions and the plasma edge parameters in JET. The principal system proposed will be for plasma edge diagnostics. A carriage driven by linear motors and containing up to five probes will be transferable into the torus from a surface analysis station. After exposure to a discharge the probes can be returned to the analysis station and analysed before the next discharge. The second system will be for short-term specimens. It will consist of a magazine of samples which can be exposed in the torus for one or more well-defined discharges. These samples will be introduced and removed from the torus via a “vacuum suitcase”. The third system, termed “long-term” samples, consists simply of regions on the inner wall of the torus where samples can be fixed at different toroidal and poloidal locations. These samples will be removed in stages during the lifetime of the machine and examined to see the long-term integrated effects on surface damage, mechanical properties, and hydrogen isotope implantation.

Through a series of controlled experiments on the DITE tokamak the main impurity production mechanisms at graphite limiters are investigated. Three principal processes can release carbon impurities: deuterium chemical sputtering, physical sputtering and oxygen chemical erosion. By varying, in turn, the limiter temperature, the bombarding particle energies and the local oxygen flux the rate of each erosion process is changed independently. From these experiments it is found that oxygen erosion dominates. Hence, carbon impurities leave the graphite surface primarily in the form of CO gas. The observed radial transport of intrinsic carbon and oxygen originating from the limiter is confirmed to be similar to that resulting from the localized puffing of CO. Metal impurities produced at the graphite limiter are physically sputtered, predominantly by carbon and oxygen impurity ions. The observed transport of chromium from the limiter has been modelled with a physical sputtering transport code.

The fluxes of impurities from limiters in tokamaks can largely be explained by sputtering. However the contribution of this flux to the impurity concentration in the discharge centre is very dependent on the transport of particles in the boundary layer. In order to improve our understanding of this transport, the neutral and low ionization states in the boundary layer have been studied. The measurements were made using a CCD camera with filters, for well identified and spectrally isolated lines of the different impurity species. By looking tangentially and vertically into the torus, the radial distribution at a poloidal limiter and the toroidal distribution at a probe limiter, were observed. These profiles have been unfolded using models of the spatial distribution. At the same time as the radiation is measured, a moving Langmuir probe is inserted up to 30 mm into the plasma. This produces radial profiles of the density and temperature in a period of 200 ms, which are used in conjunction with a 1-D model of the impurity radiation profile for comparison with the experimental results. The results show that the radiating layer of each of the species covers a radial extent of at least 50 mm with the lower ionization states being closest to the limiter, as expected. The radiation within the layer is peaked and in good agreement with the predictions of the model.

Experimental results from DITE are reviewed, with particular emphasis on the study of plasma-wall interactions. The energy and particle fluxes to the walls and limiters, and the return flux of particles into the discharge, have been measured by several methods. The effect of the bundle divertor on these fluxes, and the resulting changes in the plasma parameters are discussed.