Abstract

Electron-density maps are calculated by Fourier syntheses with coefficients based on structure factors. Diffraction experiments provide intensities up to a limited resolution; as a consequence, the Fourier syntheses always show series-termination errors. The worse the resolution, the less accurate is the Fourier representation of the electron density. In general, each atomic peak is shifted from the correct position, shows a deformed (with respect to the true distribution of the electrons in the atomic domain) profile, and is surrounded by a series of negative and positive ripples of gradually decreasing amplitude. An algorithm is described which is able to reduce the resolution bias by relocating the peaks in more correct positions and by modifying the peak profile to better fit the real atomic electron densities. Some experimental tests are performed showing the usefulness of the procedure.