LOW-FREQUENCY PERCOLATION SCALING FOR PARTICLE DIFFUSION IN ELECTROSTATIC TURBULENCE

An important point for turbulent transport consists in determining the scaling law for the diffusion coefficient D due to electrostatic turb ulence as a function of the control parameter A approximate to E/omega B proportional to the ratio of the rms electric field to the magnetic held strength times an average frequency omega. It is well known that for weak amplitudes or large frequencies, the reduced diffusion coeff icient D approximate to D/omega approximate to A(gamma) has a quasilin earlike (or gyro-Bohm-like) scaling (gamma=2), while for large amplitu des or small frequencies it has been traditionally believed that the s caling is Bohm-like (gamma=1). Only recently a percolation critical ex ponent (gamma=7/10) has been predicted by Isichenko. The aim of this w ork consists of testing this prediction for a given realistic model. T his problem is studied here by direct simulation of particle trajector ies. Guiding center diffusion in a spectrum of electrostatic turbulenc e is computed for test particles in a model spectrum, by means of a ne w parallelized code RADIGUET 2 described here. The spectrum involves o nly one frequency omega but a large number of randomly phased electros tatic plane waves, propagating isotropically in the plane perpendicula r to the confining strong magnetic field. This ensures chaotic traject ories. This set of waves represents standing waves. Their amplitudes d epend on wavelength in order to reproduce the k(-3) domain of the obse rved spectrum in tokamaks. The results indicate a continuous transitio n for large amplitudes toward a value of gamma=0.704+/-0.030 which is compatible with the Isichenko percolation prediction.