Abstract

Collecting mirror reflectometers such as the Coblentz hemisphere and the hemiellipsoid are useful for measuring diffuse reflectance and transmittance. The three primary sources of systematic error in these instruments are: optical aberrations (and associated overfilling of the detector), nonuniform angular response of the detector, and multiple reflections between the sample and detector. In previous work, Snail¹ derived an approximate upper bound for the magnification of a hemiellipsoid and suggested using an inverted nonimaging² compound parabolic concentrator to eliminate the angular response of the detector. This paper considers the effect of multiple reflections between the sample and detector of a hemiellipsoidal mirror reflectometer. For a point sample and point detector, multiple reflections can be included in the throughput equation analytically. The result is a power series. For finite sized detectors and samples, a ray tracing approach must be used. In addition to reporting on such a ray tracing study, techniques for correcting for and/or minimizing multiple-reflection effects are discussed.

© 1987 Optical Society of America