Abstract
This paper presents an accurate Monte Carlo calculation of the high‐field drift velocity of carriers in the p‐ and n‐channel inversion layers of silicon. The results are then compared with the results obtained by Hess and Sah by assuming a Maxwellian distribution of carriers. The models used in both these calculations are, however, the same as used by Hess and Sah. For the (100) surface of the n‐channel device, the electric subband expected to have the largest population at high fields is considered. For the p channel, only the lowest subband is taken and scalar effective mass is assumed. Lattice scattering mechanism alone is assumed to exist. It is found that the Monte Carlo calculation differs from the Maxwellian calculation. Also, the agreement between the former and the experiment of Fang and Fowler and of Sato et al. is not satisfactory. Possible origin of this discrepancy with the experiment is discussed.