Improvement of absolute accuracy for a multiple bounce reflectometer through a detailed effort to reduce systematic errors.

We have constructed a multiple-bounce reflectometer similar to the one designed by Kelsall. Systematic errors present the reflectivities measured on our multiple-bounce reflectometer have been significantly reduced for high-reflectivity mirrors. This reduction came about through careful study of all components in our system. The diffraction effects of the apertures in the reflectometer were studied with the aid of a computer program that calculated the radial intensity distribution due to the Fresnel diffraction from each aperture. The systematic errors arising from the attempt to measure a concave spherical mirror were studied and minimized. A second computer program calculated the systematic error introduced by pathlength changes as the number of bounces in the reflectometer increased. The replacement of the sample thermopile detector with a thin-film bolometer along with other equipment changes has improved the absolute accuracy to 0.001 for the measured reflectivities with a demonstrated precision of 0.0003. Previous measurements had been about 0.008 low compared to measurements on the V-W pass reflectometer at China Lake, California.