testing characterization algorithms

[Maxim Yurkin]

Dear colleagues,

Let me bring your attention to the paper: Doicu A., Efremenko D.S., and Yurkin M.A. An optimization tool for inverse problems of multilayered spherical particles, J. Quant. Spectrosc. Radiat. Transfer 362, 110034 (2026). (open access). 

The main result of this paper is a flexible optimization framework designed to be robust with respect to all kind of experimental noise, including model uncertainties, i.e. when the real particle is not exactly described by the used model (like concentric multi-layered sphere in this paper). The latter errors are rarely rigorously considered in the literature, but they may make the whole analysis of uncertainties invalid. The paper is not that easy to understand (full of formulae), but that is the best source of information and ready-to-use tool if you ever delve into these issues.

However, with respect to the DDA, it highlights an interesting niche application. Here we use the DDA (and ADDA, in particular) to easily answer "what if?" questions by simulating specific deviations from an ideal concentric multilayered sphere. While there may exist faster methods to solve such light-scattering problem (at least for simple deviations), the DDA probably provides the fastest turnaround time for a small number of numerical experiments. We have used the simplest procedure by utilizing the built-in shape (coated sphere with non-concentric core), but you can easily generalize it to any other shape. The only limit is your imagination, although you then also need to generate the shape file for ADDA.

Maxim.