Monte Carlo determination of electron transport properties in gallium arsenide

Royal Radar Establishment, Malvern, Worcestershire, England
Journal of Physics and Chemistry of Solids 09/1970; DOI: 10.1016/0022-3697(70)90001-6

ABSTRACT A Monte Carlo technique has been used to calculate the electron distribution functions in the (000) and (100) valleys of gallium arsenide. This method avoids having to make any of the conventional approximations used to solve the Boltzmann transport equation, but instead evaluates the distribution function exactly once the scattering rates have been specified. Polar, acoustic and relevant intervalley scattering processes have been included, together with the non-parabolicity and wavevector dependence of the cell-periodic part of the Bloch functions in the (000) valley. The structure found in the distribution function in the (000) valley is interpreted in terms of the energy dependence of the scattering processes, particular reference being made to the prediction of a population inversion for fields in excess of about 10 . The mobility, mean energy, and electron population in each valley and the mean velocity are calculated as functions of the electric field strength, and comparison is made with previous theoretical results and the experimental data.

  • [Show abstract] [Hide abstract]
    ABSTRACT: An explicit proof of a simple time-step propagation scheme is given in the framework of basic probability theory. It can be used in Monte Carlo simulations solving the Boltzmann transport equation. If the stochastically selected first scattering event occurs before a given time t1, the particle is propagated as usual until the end of the corresponding free-flight time; otherwise, however, the propagation can be stopped in this scheme at t1 and a new random number can be generated to decide whether the first scattering event occurs before the next specified time t2.
    Mathematics and Computers in Simulation 03/2003; 62(s 3–6):323–326. · 0.84 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A homogenous (bulk) silicon semiconductor is studied by using the Direct Simulation Monte Carlo (DSMC). Two DSMC algoritms are considered, the self scattering technique (SST) and the constant time technique (CTT). First, the results obtained by CTT are shown to converge (with vanishing time step) to the results obtained by SST. The truncation error of CTT turns out to be of first order with respect to the time step. Second, the efficiency of both algorithms is compared. It is found that SST is more efficient if a high precision (relative error less then three percent) of the results is needed.
    COMPEL International Journal of Computations and Mathematics in Electrical 11/2005; 24(4). · 0.28 Impact Factor
  • La Rivista del Nuovo Cimento 18(6):1-33. · 3.27 Impact Factor