Abstract
Electron diffraction patterns of 45 A thick two-dimensional crystalline arrays of a cell membrane protein, bacteriorhodopsin, have been recorded at two electron voltages, namely 20 and 120 kV. Significant intensity differences are observed for Friedel mates at 20 kV, but deviations from Friedel symmetry are quite small at 120 kV. It does not seem likely that the measured Friedel differences can be accounted for by complex atomic structure factors, by curvature of the Ewald sphere, or by effects that might occur as a result of inelastic scattering (absorption). It is therefore concluded that dynamical diffraction within the single molecular layer of these crystals is responsible for the observed Friedel differences. The results are useful in estimating the maximum specimen thickness for which the kinematic approximation may be safely used in electron crystallography of biological macromolecules at the usual electron voltage of 100 kV, or even at higher voltages. The results show that the Friedel differences are independent of resolution and this opens up the possibility that dynamical effects occurring at lower voltages might be used to phase higher-voltage kinematic diffraction intensities.