We study the sensitivity of LEED to stacking faults in the early stage
s of epitaxial growth. By theoretical simulation of stacking faults in
different locations, in particular for Cu/Cu(111), Co/Cu(111), and p(
4 x 4)Pb/Cu/Cu(111), we show that quantitative LEED can detect stackin
g faults down to three or four layers deep into the substrate. This is
also true in the presence of a surfactant layer. The modification of
the intensity spectra by stacking faults is large enough to be identif
ied by visual comparison with defect-free surfaces. In the cases where
different stacking sequences coexist on the surface, a minimum covera
ge of 30%-50% imperfect stacking sequences is needed in order to be cl
early detected in the LEED I-V curves. Therefore, our results provide
guidance to conduct experiments where early detection of stacking faul
ts is important.