Abstract
In small-angle scattering theory and data modeling, it is generally assumed that each scattered ray - photon or neutron - is only scattered once on its path through the sample. This assumption greatly simplifies the interpretation of the data and is valid in many cases. However, it breaks down under conditions of high scattering power, increasing with sample concentration, scattering contrast, sample path length and ray wavelength. For samples with a significant scattering power, disregarding multiple scattering effects can lead to erroneous conclusions on the structure of the investigated sample. In this paper, the impact of multiple scattering effects on different types of scattering pattern are determined, and methods for assessing and addressing them are discussed, including the general implementation of multiple scattering effects in structural model fits. The modification of scattering patterns by multiple scattering is determined for the sphere scattering function and the Gaussian function, as well as for different Sabine-type functions, including the Debye-Andersen-Brumberger (DAB) model and the Lorentzian scattering function. The calculations are performed using the semi-analytical convolution method developed by Schelten & Schmatz [J. Appl. Cryst. (1980 ▸). 13, 385-390], facilitated by analytical expressions for intermediate functions, and checked with Monte Carlo simulations. The results show how a difference in the shape of the scattering function plotted versus momentum transfer q results in different multiple scattering effects at low q, where information on the particle mass and radius of gyration is contained.