It is well known that during epitaxial growth of thin films of almost
all II-VI semiconductors, the growth rates show a pronounced temperatu
re dependence which is due to desorption of one or both components fro
m the growing surface. The measured desorption rate appears to be ther
mally activated with a strikingly small value: a few tenths of an eV.
The explanation generally put forward is that the desorption of a weak
-binding state acts as a ''precursor'' to chemisorption. According to
this point of view, the small measured activation energy is a real ene
rgy corresponding to a well-defined microscopic process. We argue that
no weak-binding precursor state is needed for reproducing the experim
ental growth rate of CdTe. Using Burton, Cabrera and Frank's theory an
d by performing Monte Carlo simulations of a one-particle model for de
position, diffusion, aggregation and desorption, we have found that th
e macroscopic desorption rate appears to be thermally activated over a
large range of temperatures. This rate is a combination of all the mi
croscopic energies - diffusion barrier and desorption barrier - and it
can take values of a few tenths of an electronvolt, even though all m
icroscopic energies are much larger. A very simplified model of CdTe g
rowth is thus proposed and tested against experimental measurements of
growth rates for various temperatures and deposition fluxes (C) 1998
Elsevier Science B.V. All rights reserved.