Hiroshi Naitou

Gyrokinetic simulation of internal kink modes

Internal disruption in a tokamak has been simulated using a three‐dimensional magneto‐inductive gyrokinetic particle code. The code operates in both the standard gyrokinetic mode (total‐f code) and the fully nonlinear characteristic mode (δf code). The latter is a quiet low noise algorithm. The computational model represents a straight tokamak with periodic boundary conditions in the toroidal direction and a square cross section with perfectly conducting walls in the poloidal direction. The linear mode structure of an unstable m=1 (poloidal) and n=1 (toroidal) kinetic internal kink mode is clearly observed, especially in the δf code. The width of the current layer around the x‐point, where magnetic reconnection occurs, is found to be close to the collisionless electron skin depth, indicating the importance of electron inertia. Both codes give very similar nonlinear results, in which full reconnection in the Alfven time scale is observed along with the electrostatic potential structures created during this...

Kinetic effects on the convective plasma diffusion and the heat transport

Cross field heat and particle transport due to the thermally excited convective cell mode is investigated both by theoretical work and by particle simulations. It is shown that there is a parameter range where the heat diffusion coefficient D H is considerably less than the particle diffusion coefficient D P . This results from the fact that high energy particles in a velocity distribution diffuse much more slowly than the low energy particles because the fluctuating electric fields are averaged over their finite Larmor radii. The magnetic field scaling of D H / D P is also discussed. The results imply that ions in a thermonuclear reactor should be less affected by turbulence than are electrons as appears to be the case.

Relationship between extraction of H− ions optimized by plasma grid potential and plasma parameters in a bucket source

Although optimizing the magnetic filter position and the plasma grid potential is one of the most effective factors to enhance H− yield, details concerning their roles are not now clarified well. In this article, spatially resolved measurements of the electron energy distribution function, plasma fluctuations, and plasma parameters are presented. On the basis of these experimental results, we will discuss the roles of both the magnetic filter and the plasma grid biasing voltage Vb on enhancement of H− production and extraction of H− ions.

Nonlinear simulation of tearing mode based on 4-field RMHD model

Simulations of dynamics of tearing modes based on a 4-field reduced magneto hydro dynamics model are performed, laying an emphasis on interaction with microscopic and transport processes. The simulation results show the importance of turbulent fluctuations for the onset of tearing mode. The helical current perturbation associated with a magnetic island is studied. The perturbed bootstrap current has a structure in the magnetic island (including a phase difference), indicating the need to improve assumptions in the theory of neoclassical tearing mode.

Spatial control of electron energy distribution function in a magnetically filtered multicusp plasma source

Using a movable magnetic filter, a technique to control plasma parameters spatially has been investigated in a multicusp plasma source. At any filter position, plasma parameters change steeply across the magnetic filter. Mainly, a multicusp plasma source is divided into two parts, i.e., a source plasma region with energetic electrons and a diffused plasma region. In the latter region, the high-energy component of electrons is eliminated thoroughly by the localized magnetic field of the filter.

Control of reactive plasmas in a multicusp plasma source equipped with a movable magnetic filter

With the use of both a movable magnetic filter and a plasma grid, plasma parameters (H2‐CH4 or Ar‐CH4 plasmas) are spatially well controlled. At any filter position, plasma parameters change steeply across the magnetic filter. Then, a plasma source is divided into the two parts, i.e., the source plasma region (high density plasma with energetic electrons) and the diffused plasma region (low density and low‐temperature plasma without energetic electrons). Plasma parameters in the diffused plasma are well controlled by changing the plasma grid potential. The role of the magnetic filter (i.e., preferential reflection of energetic electrons) is well clarified by computer simulation. The relation between plasma parameters and some species of neutral radicals is also briefly discussed.

Observation of the limit cycle in asymmetric plasma divided by a magnetic filter

An asymmetric plasma divided by a magnetic filter is numerically simulated by the one-dimensional particle-in-cell code VSIM1D [Koga et al., J. Phys. Soc. Jpn. 68, 1578 (1999)]. Depending on the asymmetry, the system behavior is static or dynamic. In the static state, the potentials of the main plasma and the subplasma are given by the sheath potentials, φM∼3TMe/e and φS∼3TSe/e, respectively, with e being an electron charge and TMe and TSe being electron temperatures (TMe>TSe). In the dynamic state, while φM∼3TMe/e, φS oscillates periodically between φS,min∼3TSe/e and φS,max∼3TMe/e. The ions accelerated by the time varying potential gap get into the subplasma and excite the laminar shock waves. The period of the limit cycle is determined by the transit time of the shock wave structure.

Beam instability excited by the magnetic filter

By the one-dimensional electrostatic particle simulation, the ion beam instability is observed in the plasma divided by the magnetic filter (MF). The strength of the MF is selected to influence only electron dynamics; ions move freely across the MF. There are grounded walls at the left and right ends of the system. Particles hitting the walls are absorbed there. The high temperature and high density plasma (main plasma) faces the low temperature and low density plasma (subplasma) across the MF located at the center of the system. The averaged space potential of the main plasma is higher than that of the subplasma. Due to the potential gap at the MF, ions in the main plasma are accelerated into the subplasma. Depending on the extent of the asymmetry of the system, steady or the periodic (dynamic) state manifests. For the periodic state, high density clumps get into the subplasma and excite the strong ion beam instability. The new clump comes into the subplasma when the old clump reaches the wall.

Particle simulation of the potential formation across the magnetic filter

The potential formation across the magnetic filter (MF) is studied by the two‐and‐one‐ half‐dimensional (two configuration and three velocity spaces) electrostatic particle simulation. The strength of the MF is selected to influence only electron dynamics; ions move freely across the MF. The MF divides the plasma into the main plasma and the diffused plasma. The density and temperature of the main plasma is controlled, while there is no artificial treatment for the diffused plasma. It is found that the (collisionless) anomalous diffusion of electrons plays an important role in determining the potential structure in which the potential of the main plasma is higher than that of the diffused plasma. It is also found that by changing the potential of the wall adjacent to the diffused plasma, the plasma potential of the diffused plasma can be controlled without affecting the main plasma.

Particle simulation on extraction of positive and negative ions from a volume source

A two dimensional electrostatic particle simulation was done to study the extraction of positive (negative) ions from a volume plasma source. The simulation model is a rectangular system which consists of an extraction grid (left wall), a plasma grid, and a grounded wall (right wall). Upper and lower boundaries are connected by the periodic boundary condition. Full dynamics of charged particles are followed. Positive (negative) ions are extracted from the plasma region through a slit in the plasma grid to the extraction grid. Electrons are reflected by the magnetic filter and confined in the region to the right of the magnetic filter. Simulation results are compared with the Child–Langumuir law where the extracted ion current is proportional to the three‐halves power of the potential of the extraction grid. In the case of the positive ion extraction, simulation results agree quite well with the Child–Langumuir law. Whereas, in the case of the negative ion extraction, simulation results agree with the law ...