Density Pump-out Research Articles

Analysis of stochastic magnetic fields formed by the application of resonant magnetic perturbations on MAST and comparison with experiment

In MAST experiments with applied resonant magnetic perturbations (RMPs), clear reduction in line-averaged density has been observed in a wide range of L-mode plasmas when there is an alignment between the perturbation and the equilibrium magnetic field that maximizes the size of the resonant components of the applied magnetic field, as well as in a few H-mode plasmas but with a much stronger sensitivity to this alignment. This density pump-out is the result of increased particle transport, which is thought to be caused by the formation of a stochastic magnetic field in the plasma edge. This paper presents an analysis of the magnetic field structures formed by the application of n = 3 RMPs on MAST, including various parameters characterizing the degree of stochasticity in the plasma edge. Values for these parameters are calculated and compared with the amount of density pump-out observed in MAST experiments. It is found that density pump-out is fairly well correlated with some of the parameters calculated using vacuum modelling, but none of them provides a single threshold value for pump-out that applies to both L- and H-mode plasmas. Plasma response modelling provides a robust criterion for density pump-out that applies both to L- and H-mode plasmas.

ELM pacing using modulated non-axisymmetric magnetic fields on DIII-D

Edge localized mode (ELM) pacing using modulated n = 3 non-axisymmetric fields has been demonstrated on DIII-D over a wide range of conditions, including significant variations in temperature, βN, density and shape. At low collisionality, the pacing results in a clear reduction in the ELM size and peak heat flux to the divertor, up to a factor of 5–6 for short time windows, although only a factor of two for sustained periods with the present hardware capability. At higher collisionality, although similar increases in the ELM frequency have been demonstrated, no meaningful reduction in the heat flux is typically observed as a direct result of the pacing. However, it appears that the ELM size may be reduced indirectly via changes in the L–H power threshold as a result of density pumpout associated with the application of non-axisymmetric fields. At this stage, it remains unclear whether the failure to reduce the ELM size with modulated fields is a limitation associated with high collisionality, high density, or relative proximity to the L–H power threshold.

Comparison of resonant magnetic perturbation-induced particle transport changes in H-mode (DIII-D) and L-mode (MAST)

Recent experiments show the impact of resonant magnetic perturbations (RMPs) on the density (Schmitz et al 2008 Plasma Phys. Control. Fusion 50 124029, Evans et al 2008 Nucl. Fusion 48 024002, Kirk et al 2008 Nucl. Fusion 50 024002, Liang et al 2007 Phys. Rev. Lett. 98 265004), leading to a so-called density pump-out. Previous comparisons between DIII-D and TEXTOR have focused on the similarities of the deformation of the separatrix and the creation of striations at the intersection of the main chamber wall (Schmitz et al 2008 Plasma Phys. Control. Fusion 50 124029; Schmitz et al 2009 Phys. Rev. Lett. 103 165005). In this paper, we compare the difference in magnitude of the experimentally observed density pump-out in L-mode with H-mode in two diverted tokamaks: MAST and DIII-D. In order to address the differences in magnetic field from the coils, plasma shape and q95 between the two devices, we compute a weighted magnetic diffusion coefficient with a vacuum field line tracing code. This allows us to compare the changes in density pump-out with the weighted magnetic diffusion coefficient, using a simple particle diffusion model. We find that the density pump-out is vastly different in the two confinement regimes, suggesting different particle transport mechanisms. Since one main difference in transport characteristics between L- and H-mode is turbulence, we compare turbulent particle characteristics. We find that in L-mode (MAST) the fluctuations and E × B shear increase at the plasma edge, whereas in H-mode (DIII-D) the fluctuations decrease at the plasma edge. Deeper inside the core, the E × B shear remains similar in L-mode (MAST), whereas a large decrease that quickly saturates with RMP strength is observed in H-mode (DIII-D). These results suggest that the RMP-induced particle transport at the plasma edge in L-mode (MAST) is the result from increases in turbulent particle transport, whereas the results in H-mode (DIII-D) suggest a decrease in turbulent particle transport.

Magnetic perturbation experiments on MAST L- and H-mode plasmas using internal coils

Experiments have been performed on MAST using internal (n = 3) resonant magnetic perturbation (RMP) coils. The application of the RMPs to L-mode discharges has shown a clear density pump-out when the field line pitch angle at the low-field side of the plasma is sufficiently well aligned with the applied field. The application of the RMPs before the L–H transition increases the power required to achieve H-mode by at least 30%. In type I ELMing H-mode discharges, at a particular value of q95, the ELM frequency can be increased by a factor of 5 by the application of the RMPs. This effect on the ELMs and the L-mode density pump-out is not correlated with the width of the region for which the Chirikov parameter, calculated using the vacuum field, is greater than 1 but may be correlated with the size of the resonant component of the applied field in the pedestal region or with the location of the peak plasma displacement when the plasma response is taken into account.

ELM suppression by resonant magnetic perturbation in high-performance, stationary plasmas

The method of resonant magnetic perturbation (RMP) has been shown to suppress edge-localized modes (ELMs) in the high-performance, stationary (or ‘hybrid’) scenario in the DIII-D tokamak. Calculations of stability to peeling–ballooning (P–B) modes are shown to be consistent with the observed suppression of type-I ELMs, while the ELM suppression, dependence on edge safety factor and density pump-out are similar for hybrids and standard H-mode discharges. However, other small ELMs can appear when the edge safety factor is outside the resonance window or when the H-mode pedestal is perturbed, which are not related to P–B stability. The role of the edge bootstrap current in determining stochastic heat transport during RMP is discussed.

Active control of type-I edge localized modes with n = 1 and n = 2 fields on JET

Recent experiments on JET have shown that type-I edge localized modes (ELMs) can be controlled by the application of static low n = 1 external magnetic perturbation fields produced by four external error field correction coils (EFCC) mounted far away from the plasma between the transformer limbs. When an n = 1 field with an amplitude of a few mT at the plasma edge (the normalized poloidal flux, Ψ, is larger than 0.95) is applied during the stationary phase of a type-I ELMy H-mode plasma, the ELM frequency rises from ∼30 Hz up to ∼120 Hz. The energy loss per ELM normalized to the total stored energy, ΔWELM/W, decreased from 7% to values below the resolution limit of the diamagnetic measurement (<2%). Transport analysis using the TRANSP code shows up to 20% reduction in the thermal energy confinement time because of density pump-out, but when normalized to the IPB98(y, 2) scaling the confinement time shows almost no reduction. Stability analysis of controlled ELMs suggests that the operational point with n = 1 perturbation field moves from the intermediate-n peeling–ballooning boundary to the low-n peeling boundary, and the radial width of the most unstable mode reduced from ∼3% down to ∼1% of the normalized minor radius. The first results of ELM control with n = 2 fields on JET demonstrate that the frequency of ELMs can be increased by a factor of 3.5 with the present capability of the EFCC power supply. During the application of the n = 1, 2 fields, a reduction in the absolute ELM size (ΔWELM) and ELM peak heat fluxes on the divertor target by roughly the same factor as the increase in the ELM frequency has been observed. The reduction in heat flux is mainly due to the drop in particle flux rather than a change in the electron temperature. Similar plasma braking effects have been observed with n = 1 and n = 2 external fields when the same EFCC coil current was applied. Compensation of the density pump-out effect has been achieved by means of gas fuelling in low triangularity plasmas. An optimized fuelling rate to compensate the density pump-out effect has been identified. When the n = 1 field is applied in plasmas with reduced toroidal rotation and density due to increased TF ripple of 8%, both the magnitude of the toroidal braking and density pump-out are found to become smaller; however, the increase in the ELM frequency with the n = 1 field is still observed. Active ELM control by externally applied fields may offer an attractive method for next-generation tokamaks, e.g. ITER.

Overview of ELM Control by Low n Magnetic Perturbations on JET

A series of ELM control experiments have been performed on JET aiming at a better understanding of the plasma response to applied magnetic perturbations. The dynamics of the edge profiles with applied n = 1 fields have been studied in the type-I ELM H-mode plasmas. Typically, the pedestal density decreased by about 20 % when the n = 1 perturbation field was applied (So called pump-out effect). However, there is no influence of the n = 1 fields on the recovery rate of the pedestal pressure, but the ELM crash occurs earlier and at a lower level of the pedestal pressure and pressure gradient. The compensation of the density pump-out has been demonstrated using either gas fuelling or pellets injection in low triangularity H-mode plasmas. Strong toroidal rotation braking by more than 60 % has been observed, and found to be independent on the safety factor. The calculated Neoclassical Toroidal Viscosity (NTV) torque profile in the ν regime including the boundary effect is in agreement with the observed torque profile induced by the n = 1 fields on JET. No complete ELM suppression was observed by application of either n = 1 or n = 2 fields with a Chirikov parameter above one at √ Ψ ≥ 0.925, which is one of the important criteria for the design of ITER ELM control coils.

Edge localized mode control experiments on MAST using resonant magnetic perturbations from in-vessel coils

Edge localized mode (ELM) control has been investigated on MAST with a set of in-vessel coils producing n = 3 perturbations. In spite of a broad enough region where the Chirikov parameter is above 1, according to the criterion employed for the design of the ITER ELM control coils, no effect was seen on type I ELMs. The behaviour on MAST is different from that in DIII-D in this respect. The coils do have an effect, however, on Ohmic L-mode plasmas and plasmas just above the L–H transition. In the former, they induce a density pump-out, modify the turbulence characteristics and edge radial electric field and produce a strike point splitting, while in the latter they can trigger ELMs in ELM-free plasmas.

Main Scientific Achievements of the Tore Supra Ergodic Divertor Experiments

The main scientific outputs from the Tore Supra ergodic divertor experiments are recalled as well as their theoretical framework. Major references are given for further consideration and in-depth information. The surprising results, such as density pump-out or unaffected electron core temperature, are reported. These have also been observed in the recent edge-localized mode (ELM) control experiments. The understanding of their physics, stemming from experimental and theoretical work, is recalled, although further assessments are still required; this can provide an interesting starting point for the ongoing effort in the ITER framework of the control of ELMs and plasma-wall interaction.

The influence of resonant magnetic perturbations on edge transport in limiter H-mode plasmas in TEXTOR

In this contribution, we report on experimental results on edge transport in limiter H-mode plasmas in TEXTOR under the influence of the Dynamic Ergodic Divertor (DED). These plasmas are characterized by a pedestal structure mainly visible in the electron density, resulting in increased electron pressure gradients of up to 30 kPa/m over a pedestal width of 25 mm at high pedestal collisionalities ( ν e ∗ = 1 - 10 ) , and with high frequency ELMs in the range of 300–1500 Hz. Under the influence of DED the pedestal pressure is gradually reduced and completely collapses to L-mode when the laminar zone extends all the way across the pedestal width. Toroidal plasma rotation is maintained at H-mode levels by the torque introduced by DED in the stochastic region. The perturbed magnetic topology has been optimized to access conditions with a density pump-out which are strongly governed by wall pumping capabilities in TEXTOR.

Experimental comparison of recycling and pumping changes during resonant magnetic perturbation experiments at low and high collisionality in DIII-D

Resonant magnetic perturbations (RMPs) have been shown to successfully suppress edge localized modes (ELMs) in the DIII-D tokamak. A previous study of target plate conditions highlighted differences in RMP discharges between low and high electron collisionality, ν e ∗ , operation in DIII-D. This paper reports on a systematic study of the electron density pump-out associated with the turn-on of the RMP over a wide range of operating conditions in DIII-D, including shapes and collisionalities similar to those anticipated in ITER. It is shown that the pump-out magnitude, Δ n e, has an upper envelope that is inversely proportional to the pedestal ν e ∗ . The particle decay times, which are calculated based on global D 2 particle balances, show an increase as the pedestal ν e ∗ is increased. Both results are suggestive that the underlying physics mechanism is an increase in edge particle transport and/or that wall depletion is playing a role in the pump-out magnitude.

Fluid modeling of an ELMing H-mode and a RMP H-mode

In this paper, we investigate the role of wall conditions during the density pump-out observed in Resonant Magnetic Perturbation (RMP) experiments. We use a 2D fluid code, coupled to a neutral Monte-Carlo code (SOLPS5) to model a reference ELMing H-mode. Next we modify either the recycling at the target plates or the pumping efficiency at the pump entrance. This causes a change in upstream profiles, that alone cannot account for the shape change observed in the RMP H-mode. Therefore, the radial transport must be altered to match the RMP H-mode. By comparing the transport model for the ELMing H-mode and the RMP H-mode, we show that inside the separatrix an increase in transport is necessary to model the effects of the applied perturbed field in RMP H-mode.

Preliminary Results of Sawtooth Behaviour and Electron Heat Transport During ECRH Experiment on HL-2A

An electron cyclotron resonance heating (ECRH) system with a maximumoutput power of 1 MW, a frequency of 68 GHz and a duration of1s was operated on HL-2A tokamak. The temperature profile controland the sawtooth behaviour during ECRH experiments are investigated.`Density pumpout' during on-axis ECRH is analysed. Features ofconfinement and electron heat transport during ECRH are studied.

Start-up assist by magnetized plasma flow injection in TPE-RX reversed-field pinch

A reversed-field pinch (RFP) start-up assisted by a magnetized plasma flow injection was demonstrated for the first time on a TPE-RX machine. This sequence of experiments aimed to establish a new method of ionization, gas-fill and helicity injection in the start-up phase of an RFP. In this start-up method, magnetized and well-ionized plasma is formed by a magnetized coaxial plasma gun and injected into the torus chamber as an initial pre-ionized plasma for RFP formation. In the initial experiments, attenuated density pump-out and comparatively slow decay of the toroidal flux and plasma current were observed as evidence of its being an effective start-up method.

Global modelling of tank gas density and effects on plasma density control in MAST

A zero-dimensional, or global, description of fuelling and exhaust processes is introduced to help assess techniques for reducing the density of gas, and its impact on plasma-density control, within the large vacuum tank of the MAST spherical tokamak. Model calculations for sample MAST discharges reveal their typical ion confinement time (⩽50 ms) and contrasting fuelling efficiencies for outboard or inboard puffing. In steady state, the ratio between tank molecular and plasma densities is fixed irrespective of any sinks when gas is puffed from the vessel wall, but for fuelling at the plasma edge with moderate ionization efficiency, unsaturated boronization of MAST surfaces can yield ≈5 × lower gas density and excellent plasma pump-out. A ‘two-chamber’ adaptation of the global model with zones connected by variable conductances is also defined to address a possible in-vessel, cryopumped divertor. Exact limiting solutions reveal that for a constant plasma, steady-state gas density in the main torus generally rises for stronger divertor pumping when puffing into the tank, while it only decreases with pumping for fuelling at non-zero efficiency directly into the divertor plasma or 100% efficiency at any location. Full ‘two-chamber’ calculations for MAST indicate that the tank gas would be lowered principally through high divertor closure and fuelling efficiency, which combined with unsaturated boronization lead to > 10 × less steady gas density. Divertor cryopumping would be useful chiefly for plasma density control in long pulses or sustained exhaust in the absence of wall sinks. Global modelling offers a way similarly to examine fuelling behaviour in any tokamak device.

Understanding of the density profile shape, electron heat transport and internal transport barriers observed in ASDEX Upgrade

In this paper several transport phenomena are described and explained through a (quasi) linear description of the micro-instabilities. This paper deals with the following phenomena: density peaking, electron heat transport, density pump-out, reactor density profiles and the stabilization of the ion temperature gradient (ITG) mode in transport barriers. Density peaking is observed to increase with decreasing collisionality—a phenomenon that can be explained by the influence of collisions on the trapped electron response. The density pump-out due to central electron heating is observed to occur only if the dominant instability is the trapped electron mode (TEM). The proposed explanation involves the thermo-diffusive contribution to the particle flux which is outwards for the TEM while it is inwards for the ITG. The insight in the density profile behaviour can be used to predict moderately peaked density profiles in a reactor. Collisionality is also found to influence the electron heat flux in discharges with dominant electron heating even at relatively small values of the density. The dominant instability under dominant electron heating is found to be the TEM, and the dependence of the electron heat flux on the electron temperature gradient is reasonably described by the quasi-linear results. At higher density (and collisionality) the growth rate of the TEM is reduced and a transition to a dominant ITG is found. This transition is reflected in the speed of the electron heat pulse propagation. Finally, it has been found that the ITG under experimentally relevant conditions is not stabilized by a uniform radial electric field.

Overview of TCV results

The TCV tokamak (R = 0.88 m, a<0.25 m, BT<1.54 T) is equippedwith six 0.5 MW gyrotron sources operating at 82.7 GHz for secondharmonic X mode ECH. By distributing the ECCD current sources overthe discharge cross section, fully driven stationary plasmas withIp = 210 kA, ne0 = 2 × 1019m-3, Te0 ≈ 4 keV, were obtained for the fulldischarge duration of 2 s. Highly peaked electron temperature profileswith Te0 up to 12 keV were obtained in central counter-current drivescenarios with off-axis ECH. Absorption measurements using a 118 GHzgyrotron have demonstrated the importance of suprathermal electronsfor third harmonic absorption. A coupled heat-particle transportphenomenon known as `density pumpout', which leads to the expulsion ofparticles from the plasma core, has been linked to the presence of m = 1modes, suggesting that it is due to the existence of locally trappedparticles associated with the loss of axisymmetry. Highly elongateddischarges have been developed with ohmic heating (κ<2.8) and off-axisECH. The latter exhibit considerably improved vertical stability dueto current profile broadening. A `gateway' for ELMy H modes has beendiscovered, which allows stationary ohmic ELMy H mode operation over a wide range of elongation, triangularity and density. Divertordetachment experiments suggest the existence of recombination pathwaysother than three body or radiative processes.

Particle transport with high power central ECH and ECCD in TCV

A coupled heat and particle transport phenomenon, leading to particle depletion from the plasma core is observed in a variety of plasma conditions with centrally deposited electron cyclotron heating and electron cyclotron current drive in TCV. This phenomenon, which causes inverted sawteeth of the central density in sawtoothing discharges and leads to stationary hollow profiles in the absence of sawteeth, has been linked to the presence of m/n = 1/1 MHD modes. In particular, this phenomenon, known as `density pumpout' can be suppressed by stabilizing the mode by means of operation at high triangularity. The correlation of pumpout with the loss of axisymmetry suggests that neoclassical transport processes involving locally trapped particles near the helically displaced magnetic axis, previously believed to be important only in stellarators, may account for the phenomenon in tokamaks as well.

Analysis of RF sheath interactions in TFTR

New theoretical and experimental tools are applied to the analysis of ICRF antenna-edge plasma interactions in the TFTR tokamak. A new numerical method for computing the three dimensional (3-D) rf sheath voltage distribution is used, and the quantitative predictions of rf sheath theory are compared with measurements of the edge density profile obtained by microwave reflectometry and with titanium impurity concentration data. It is shown that the local density depletion at the antenna is consistent with density pump-out by strong E × B convection into the Faraday screen (FS). Modelling of the FS impurity influx shows that the calculated titanium impurity concentration based on this direct influx agrees with the measured concentration for π phasing. It is also shown that screening of impurity neutrals by ionization in the SOL is a large effect and increases with rf power. At high power over many shots, a fraction of the metal impurities migrates around the machine and is deposited on the limiters, providing a secondary source of titanium. The data show that the central titanium concentration is strongly dependent on antenna phasing. Possible explanations for this phasing dependence are discussed.

Generation of poloidal electric field with electron cyclotron resonance heating and its effects on particle confinement in Heliotron-E

The potential formation near the edge, X-points, and on the divertor footprints are studied using Langmuir probe arrays with respect to the density pump-out due to ECRH. It is found that a poloidal electric field is generated in the interior region, and this is considered to cause the enhanced particle flux across the LCFS. A poloidal structure of the edge potential distribution seems to be an m = 2 type and is discussed based on the orbit topology of warm electrons in the l = 2 heliotron configuration.