Submillimeter Wavelength Scattering From Random Rough Surfaces

We describe bistatic scattering measurements on eight reference targets constructed from Al 2 O 3 grit of various sizes embedded in an absorptive epoxy matrix. These samples’ surface topographies were measured using focus-variation microscopy, and their autocorrelation lengths and root-mean-square (RMS) roughness levels were extracted. Bistatic scattering measurements were then performed in the principal plane over the 325–500 GHz range and at 650 GHz, using either a vector network analyzer or a narrow band source and detector, respectively, in both s- and p -polarizations, and at incidence angles of 25°, 45°, and 65°. The samples’ RMS roughness levels cover the range of 36–280  μ m, corresponding to $0.040 . Their autocorrelation lengths cover the range of $0.086 , generally shorter than the samples in most previous studies. The measurements of bidirectional reflectance distribution function (BRDF) include regimes of both diffuse scattering and specular reflectance. The measurements were compared to two ab initio rough-surface scattering theories: the modified integral equation method (IEM-B) and the Generalized Harvey–Shack (GHS) model. There are no adjustable free parameters in these comparisons, but other important shortcomings exist and are described for both theories. Although there are several individual cases where either the IEM or the GHS theory (or both) provide an excellent match to measurement, their overall agreement with measurement across the entire dataset is poor. In addition, the diffuse BRDF in each bistatic scan has been fitted to a Lambertian (constant) dependence on scattering angle, and a purely empirical model has been developed for the dependence of the Lambertian scattering on frequency, roughness, polarization, and incidence angle. Across the entire dataset, this empirical model provides the best match to measurement, and can be used for reliable phenomenology studies of submillimeter imaging or wireless telecommunication.