Monte Carlo Simulation of an Electron Swarm in an Argon-like Gas

Abstract

Aust. J. Phys., 1987,40,61-4 A Monte Carlo simulation (MCS) of an electron swarm in an argon-like gas (of atomic mass a.m.u.), with Golden's cross section with a very deep minimum, is presented in order to examine the breakdown of the validity of the conventional two-term expression for the transverse diffusion coefficient Dr indicated by Milloy and Watts. It is found that the MCS value of Dr approaches the two-term value much slower than other swarm parameters such as the drift velocity Wand the longitudinal diffusion coefficient ~. The discrepancy between the MCS and two-term values of Dr revealed by Milloy and Watts is attributable to the evaluation of Dr at the long transient (correlation) stage in the MCS. It is concluded that the two-term expressions for Dr and ~ (and other swarm parameters) are sufficiently accurate even for Golden's cross section with a very deep minimum.Because the conventional two-term solution of the Boltzmann equation is very useful for the analysis of electron swarms in gases, the validity of the two-term solution has been thoroughly examined in its various aspects. By making comparisons with electron swarm parameters obtained by the Monte Carlo simulation (MCS), Milloy and Watts (1977) indicated the breakdown of the validity of the two-term expression for the transverse diffusion coefficient Dr in an argon-like gas (of atomic mass 1 a.m.u.) with Golden's (1966) momentum-transfer cross section cr m·(E) with a very deep minimum. It seems curious, however, that the MCS and two-term values of the drift velocity W, the mean energy <E), and even the overall energy distribution fo(E) are in agreement but Dr values disagree by up to 40% when the ratio of electric field strength to atomic number density (E/N) is about 2 Td «E) -1 eV). (The longitudinal diffusion coefficient DL was not considered.)In this paper, a MCS of an electron swarm in the argon-like gas is presented in order to resolve this inconsistency. The Monte Carlo method is a straightforward extension of the Monte Carlo technique (Koura 1983(Koura , 1984 for the simulation of electron thermalisation in gases. At time t = 0, we assume that all electrons of number no were isotropically released from the origin (x = y = z = 0) with the same energy < E)O. The scattering of electrons was taken to be isotropic and cr m (E) was obtained from equations (2) little influence on cr m(E) at E < 2 eV.] The electric field strength E was chosen as E/N = 2 Td so that the two-term value of <E) was about 1 eV, where the MCS

Modelling electroluminescence in liquid argon

“…These larger values seem to be caused by trajectories that have not fully reached an uncorrelated state after the artificial termination of the simulation imposed after 2 × 10 9 collisions, corresponding to a simulated drift time of about 11.4 µs. Similar challenges to diffusion constant simulations have been pointed out previously in [33] and have been traced back to steep changes in cross sections at low energies. Any experiment able to measure diffusion will actually measure the spread of the electron cloud rather than each single electron trajectory.…”