Reconstructed domains on a stepped W(100) surface

A model calculation is carried out for the scattering of low energy electron diffraction from reconstructed (√2 × √2)R45° domains on a stepped W(100) surface. The existence of monotonically increasing steps causes the integral order beams to split and the separation of the splitting oscillates with the incident electron energy. We show that due to the existence of an antiphase relationship among the randomly nucleated reconstructed domains, the half-order beams neither split nor oscillate with the incident electron energy. This nonsplitting of half order beams is in agreement with the observation by Debe and King (DK). We also show that the measured intensity profile of a half order beam is equal to the signal intensity profile from the individual finite size domains convoluted with the instrument response function. This gives a simple way to evaluate the reconstructed domain size quantitatively. From the angular distribution of the half-order beam intensity we deduce that the reconstructed domains are somewhat round in shape, instead of the “long strips” proposed by DK. Also, the long range inhibition (20 Å) of the reconstruction near the step edge suggested by DK does not necessarily follow from our analysis. As a matter of fact, there is evidence showing that the inhibition (if it exists at all) can be short range in nature. Our suggestion is in agreement with the observation of the reconstructed W(100) surface by Melmed, Tung, Graham and Smith using FIM technique.