G. Naylor

Design of a new Nd:YAG Thomson scattering system for MAST

A new infrared Thomson scattering system has been designed for the MAST tokamak. The system will measure at 120 spatial points with approximately 10 mm resolution across the plasma. Eight 30 Hz 1.6 J Nd:YAG lasers will be combined to produce a sampling rate of 240 Hz. The lasers will follow separate parallel beam paths to the MAST vessel. Scattered light will be collected at approximately f/6 over scattering angles ranging from 80 degrees to 120 degrees. The laser energy and lens size, relative to an existing 1.2 J f/12 system, greatly increases the number of scattered photons collected per unit length of laser beam. This is the third generation of this polychromator to be built and a number of modifications have been made to facilitate mass production and to improve performance. Detected scattered signals will be digitized at a rate of 1 GS/s by 8 bit analog to digital converters (ADCs.) Data may be read out from the ADCs between laser pulses to allow for real-time analysis.

A 130 point Nd:YAG Thomson scattering diagnostic on MAST.

A Thomson scattering diagnostic designed to measure both edge and core physics has been implemented on MAST. The system uses eight Nd:YAG lasers, each with a repetition rate of 30 Hz. The relative and absolute timing of the lasers may be set arbitrarily to produce fast bursts of measurements to suit the time evolution of the physics being studied. The scattered light is collected at F/6 by a 100 kg six element lens system with an aperture stop of 290 mm. The collected light is then transferred to 130 polychromators by 130 independent fiber bundles. The data acquisition and processing are based on a distributed computer system of dual core processors embedded in 26 chassis. Each chassis is standalone and performs data acquisition and processing for five polychromators. This system allows data to be available quickly after the MAST shot and has potential for real-time operations.

L–H transition and pedestal studies on MAST

On MAST studies of the profile evolution of the electron temperature (Te), electron density (ne), radial electric field (Er) as well as novel measurements of the ion temperature (Ti) and toroidal current density (j) in the pedestal region allow further insight into the processes forming and defining the pedestal such as the H-mode access conditions and MHD stability. This includes studies of fast evolution of Te, ne and Er with Δt = 0.2u2009ms time resolution and the evolution of pe and j through an edge-localized mode (ELM) cycle. Measurements of the H-mode power threshold, PL−H revealed that about 40% more power is required to access H-mode in 4He than in D and that a change in the Z-position of the X-point can change PL−H significantly in single and double null configurations. The profile measurements in the L-mode phase prior to H-mode suggest that neither the gradient nor the value of the mean Te or Er at the plasma edge play a major role in triggering the L–H transition. After the transitions, first the fluctuations are suppressed, then the Er shear layer and the ne pedestal develops followed by the Te pedestal. In the banana regime at low collisionality (ν) ∇Ti ≈ 0 leading to Ti > Te in the pedestal region with Ti ~ 0.3u2009keV close to the separatrix. A clear correlation of ∇Ti with ν is observed. The measured j (using the motional Stark effect) Te and ne are in broad agreement with the common peeling–ballooning stability picture for ELMs and neoclassical calculations of the bootstrap current. The j and ∇pe evolution Δt ≈ 2u2009ms as well as profiles in discharges with counter current neutral beam injection raise questions with respect to this edge stability picture.

Collisionality and safety factor scalings of H-mode energy transport in the MAST spherical tokamak

A factor of 4 dimensionless collisionality scan of H-mode plasmas in MAST shows that the thermal energy confinement time scales as . Local heat transport is dominated by electrons and is consistent with the global scaling. The neutron rate is in good agreement with the ν* dependence of τE,th. The gyrokinetic code GYRO indicates that micro-tearing turbulence might explain such a trend. A factor of 1.4 dimensionless safety factor scan shows that the energy confinement time scales as . These two scalings are consistent with the dependence of energy confinement time on plasma current and magnetic field. Weaker qeng and stronger ν* dependences compared with the IPB98y2 scaling could be favourable for an ST-CTF device, in that it would allow operation at lower plasma current.

New physics capabilities from the upgraded Thomson scattering diagnostic on MAST

The newly upgraded MAST Thomson scattering (TS) system provides excellent spatial resolution (~1?cm) at over 130 radial locations across a full plasma diameter, and utilizes eight individual Nd:?:YAG laser systems which can be fired sequentially, providing electron temperature and density profiles approximately every 4?ms throughout a plasma discharge. By operating the system in burst mode, whereby the laser separation can be adjusted to within a few microseconds of each other, it is possible to obtain detailed profiles of transient and periodic phenomena such as sawteeth crashes, massive gas injection for disruption mitigation and the temperature perturbations associated with neoclassical tearing mode (NTM) islands. Following Fitzpatrick et al (1995 Phys. Plasmas 2 825), we consider a simplified model in which finite parallel diffusive heat transport can provide a threshold for NTM island growth and demonstrate that the TS derived electron temperature profiles around an island can be used to obtain both the island width and the critical island width below which temperature gradients are maintained across the island, potentially removing the bootstrap current drive for the NTM. Initial results from high beta, neutral beam injection heated discharges on MAST show that the measured island width inferred from the TS data is in good agreement with magnetic estimates of the island width (considering both a cylindrical approximation and using a full field line tracing estimate). The temporal behaviour of the island width obtained from the magnetic diagnostics indicates that for the scenarios considered to date, finite parallel diffusion is likely to play an important role in NTM threshold physics in MAST.

Synchronization of Thomson scattering measurements on MAST using an FPGA based ''Smart'' trigger unit

The MAST Thomson scattering diagnostic has recently been upgraded to make electron density and temperature measurements at 130 points across the 1.5 m diameter of the plasma. The new system is able to take 240 measurements per second using eight Nd:YAG lasers, each running at 30 Hz. The exact firing time of these lasers is adjusted with 100 ns precision using a field programmable gate array based trigger unit. Trigger pulses are produced to fire the lamps of all lasers and the Q switches with the appropriate delay depending on the warm-up status. The lasers may be fired in rapid bursts so as to achieve a high temporal resolution over eight points separated down to the microsecond level. This trigger unit receives optical trigger events and signals from external sources, allowing the trigger sequences to be resynchronized to the start of the plasma pulse and further events during the shot such as the entry of a fuelling pellet or randomly occurring plasma events. This resynchronization of the laser firing sequence allows accurate and reproducible measurements of fast plasma phenomena.

Absolute calibration of LIDAR Thomson scattering systems by rotational Raman scattering

Absolute calibration of LIDAR Thomson scattering systems on large fusion devices may be achieved using rotational Raman scattering. The choice of calibrating gas molecule presents different options and design trade-offs and is likely to be strongly dependent on the laser wavelength selected. Raman scattering of hydrogenic molecules produces a very broad spectrum, however, with far fewer scattered photons than scattering from nitrogen or oxygen at the same gas pressure. Lower laser wavelengths have the advantage that the Raman cross section increases, sigma(Raman) proportional to 1/lambda(0)(4), but the disadvantage that the spectral width of the scattered spectrum decreases, Deltalambda(Raman) proportional to lambda(0)(2). This narrower spectrum makes measurement closer to the laser wavelength necessary. The design of the calibration technique presents a number of challenges. Some of these challenges are generic to all Thomson scattering systems. These include detecting a sufficient number of photoelectrons and designing filters that measure close to the laser wavelength while simultaneously achieving adequate blocking of the laser wavelength. An issue specific to LIDAR systems arises since the collection optics operates over a wide range of depth of field. This wide depth of field has the effect of changing the angle of light incident on the optical interference filter with plasma major radius. The angular distribution then determines the effective spectral transmission function of the interference filter and hence impacts on the accuracy of the absolute calibration. One method that can be used to increase absolute calibration accuracy is collecting both Stokes and anti-Stokes lines with optical filter transmission bands specifically designed to reduce systematic uncertainty.

Microstability analysis of pellet fuelled discharges in MAST

Reactor grade plasmas are likely to be fuelled by pellet injection. This technique transiently perturbs the profiles, driving the density profile hollow and flattening the edge temperature profile. After the pellet perturbation, the density and temperature profiles relax towards their quasi-steady-state shape. Microinstabilities influence plasma confinement and will play a role in determining the evolution of the profiles in pellet fuelled plasmas. In this paper we present the microstability analysis of pellet fuelled H-mode MAST plasmas. Taking advantage of the unique capabilities of the MAST Thomson scattering system and the possibility of synchronizing the eight lasers with the pellet injection, we were able to measure the evolution of the post-pellet electron density and temperature profiles with high temporal and spatial resolution. These profiles, together with ion temperature profiles measured using a charge exchange diagnostic, were used to produce equilibria suitable for microstability analysis of the equilibrium changes induced by pellet injection. This analysis, carried out using the local gyrokinetic code GS2, reveals that the microstability properties are extremely sensitive to the rapid and large transient excursions of the density and temperature profiles, which also change collisionality and βe significantly in the region most strongly affected by the pellet ablation.

Laser system for high resolution Thomson scattering diagnostics on the COMPASS tokamak.

A new Thomson scattering diagnostic has been designed and is currently being installed on the COMPASS tokamak in IPP Prague in the Czech Republic. The requirements for this system are very stringent with approximately 3 mm spatial resolution at the plasma edge. A critical part of this diagnostic is the laser source. To achieve the specified parameters, a multilaser solution is utilized. Two 30 Hz 1.5 J Nd:YAG laser systems, used at the fundamental wavelength of 1064 nm, are located outside the tokamak area at a distance of 20 m from the tokamak. The design of the laser beam transport path is presented. The approach leading to a final choice of optimal focusing optics is given. As well as the beam path to the tokamak, a test path of the same optical length was built. Performance tests of the laser system carried out using the test path are described.

H-mode access by pellet fuelling in the MAST tokamak

Access into H-mode is studied in the MAST tokamak when plasma is fuelled by cryogenic pellets. It is shown that pellet fuelling from the high-field side allows access to H-mode in plasmas heated by neutral beams. Simple and two-stage L–H transitions are identified. The results of comparison of the L–H transitions with pellet injection with transitions where plasmas are fuelled solely by gas puffing depend on the gas puff geometry: fuelling by high-field side gas leads to an L–H transition at the density comparable to the transition with pellet injection. In contrast low-field gas can completely prevent the L–H transition.