phase error is much stronger. In reality, however, the standard deviation of the (random) error will be smaller than 0.08, but the data will be affected by systematic errors as well.The results reported in the present paper show that it is possible to estimate triplet invariants from artificial two-wavelength data (which may be corrupted by random errors) and use the invariants in a multi-solution procedure to obtain structure-factor phases without the need to solve the heavy-atom structure.References FORTIER, S., WEEKS, C. M. & HAUPTMAN, H. (1984). Acta Cryst.A40, 544-548. GIACOVAZZO, C. (1983). Acta Cryst. A39, 585-592. HAUPTMAN, H. (1982a). Acta Cryst. A38, 289-294.
HAUPTMAN, H. (1982b
AbstractThe dependence of the X-ray transmission coefficient of a thin-film coating as a function of incidence has been measured in the grazing-angle range. The method is based on the use of a substrate Bragg reflection to redirect the incident or transmitted beams. It allows grazing incidence from both outside and inside the substrate to be performed. The geometry of the experiment is described. The results are interpreted by means of dynamical theory combined with an optical formalism for stratified systems. Experimental results and applications are compared with reflectivity data.