Size-dependent Measurements of the Scattering Properties of Planetary Regolith Analogs: a Challenge to Theory

Introduction: The nature of the scattering of light is thought to be well understood when the medium is made up of independent scatterers that are much larger than the wavelength of that light. This is not the case when the size of the scattering objects is similar to or smaller than the wavelength or the scatterers are not independent. In an attempt to examine the applicability of independent particle scattering models, to planetary regoliths, a dataset of experimental results were compared with theoretical predictions. Method: Light scattering by samples of three well-characterized powders (aluminum oxide, calcium car-bonate, and iron oxide) were analyzed. Particle sizes of the samples range from smaller than to larger than the wavelength. Scanning electron microscope (SEM) analyses were performed to try and determine the actual particle size distribution of the sample. However, due to the difficulty of obtaining precise results with small particles, the SEM results are questionable, and the manufacturer's labeled sizes are assumed to be accurate. The samples were analyzed using both the long arm (0.05° < phase angle (g) < 2.5°) and short arm (5° < g < 140°) goniometers in the Goniometer Laboratory at JPL. Instrument descriptions and data for a variety of samples have been previously reported [1-5]. A comparison with predictions of scattering models was made by first calculating scattering parameters from the goniometer data. These were determined by fitting the radiative transfer-based, improved model of [6] to the curves. The parameters fitted are the single scattering albedo (w), single particle scattering function (p(g)) and the coherent backscatter angular width parameter hc. The single particle scattering function was represented either as an expansion of Legendre polynomials (2 nd or 3 rd order), or as a double Henyey-Greenstein function (symmetric or asymmetric), depending on which representation produced the best fit to the data. Where the opposition effect was wide enough to affect the short arm goniometer data, both the shadow-hiding and coherent backscatter opposition surges were then fit to this dataset. When this was not possible, the data from the long arm goniometer were fit using a simplified set of equations that contain only terms for the coherent backscatter opposition surge and the background intensity. The resulting fit parameters were then combined [6] to calculate fundamental scattering variables: transport mean free path (L), average scattering angle , scattering coefficient (S), and extinction coefficient (E). A Mie code [7] was used to …