A method for the accurate comparison of lattice parameters

A new method of comparison of lattice parameters is described. The method uses a standard double-crystal diffractometer, fitted with a specimen rotation stage, which compensates for tilt errors on the specimen setting. Some improvement is possible through the use of a monochromatizing system with a well defined wavelength. The use of a single reference standard enables the instrument zero to be accurately determined; the use of a second reference or a different reflection from the first allows the wavelength to be accurately determined. The errors are carefully assessed and it is shown that some errors important in other methods are self-cancelling through the use of the rotation stage, which need not itself be an ultraprecision component. Peak location to a confidence level of 10 '' permits absolute traceable lattice-parameter determination to a few tens of parts in 10(6), with much greater relative sensitivity. The method is verified using both a tangent-arm-driven and direct-drive double-axis diffractometer at room temperature on samples of semi-insulating gallium arsenide. Comparison of the lattice parameter of a sample using this method and the conventional Bond technique is satisfactory.