Measurement of the Magnetic and Temperature Dependence of the Electron-Electron Anisotropic Temperature Relaxation Rate

Using a pure electron plasma, the magnetic field and temperature dependences of the electron-electron anisotropic temperature relaxation rate are measured. The anisotropy is characterized by T_{parallel } not= T_{| }, the temperatures associated with the degrees of freedom parallel and perpendicular to the applied magnetic field. The relaxation rate is measured for large magnetic fields (30 kG to 60 kG) and for an unusual plasma temperature range (25 K to 10^4 K). In this parameter regime, the ratio of the gyroradius, r _{rm c}, to the classical distance of closest approach, b, plays an important role in determining the relaxation rate. At high temperatures the plasma is in the regime r_{rm c}/b > 1. (At all temperatures lambda_{rm D} gg r_{rm c} where lambda_{rm D} is the Debye length.) In this regime the measured rates are compared to a Fokker-Planck prediction as modified by a strong magnetic field approximation due to Montgomery, Joyce and Turner. The rates are also compared to a Fokker-Planck prediction without this approximation. The modified prediction yields much better agreement to the measured rates. For temperatures where r_{ rm c}/b ~ 1, the measured rate peaks. As the temperature is lowered below this point, the plasma enters a regime where r_{ rm c}/b < 1. Here, as r_{rm c}/b goes from about 1 to about 0.03 the measured rate normalized by T^{3/2} drops by a factor of 10^4. (Whereas, for r_{rm c}/b gg 1 the normalized rate is essentially independent of temperature.) This rapid decrease is consistent with a theoretical prediction of O'Neil and Hjorth, who argue that the collisional dynamics is constrained by a many electron adiabatic invariant in the regime r_ {rm c}/b << 1. The plasma is contained axially by a potential well and radially by the magnetic field. To measure the relaxation rate the shape of the potential well is modulated sinusoidally which similarly modulates the plasma length parallel to the magnetic field. If the modulation frequency is neither slow nor rapid compared to the relaxation rate the compression is irreversible and the plasma is heated. Maximum heating per cycle is predicted to occur when 2 pi f = 3nu, where nu is the relaxation rate and f is the modulating frequency. The relaxation rate is thus determined by measuring which modulating frequency produces the most heating per cycle.