Reproducible validation and replication studies in nanoscale physics

Credibility building activities in computational research include verification and validation, reproducibility replication, uncertainty quantification. Though orthogonal to each other, they are related. This paper presents validation replication studies electromagnetic excitations on nanoscale structures, where the quantity of interest is wavelength at which resonance peaks occur. The study uses open-source software PyGBe: a boundary element solver with trecode acceleration GPU capability. We replicate result by Rockstuhl et al. (2005, doi:10/dsxw9d) two-dimensional method silicon carbide particles, despite differences our method. second case from Ellis (2016, doi:10/f83zcb) looks aspect ratio effects high-order modes localized surface phonon-polariton nanostructures. results partially replicate: wavenumber position some mode match, but for other differ. With virtually no information about original simulations, explaining discrepancies not possible. A comparison experiments that measured polarized reflectance nano pillars provides case. dominant two more do remain minor modes. Results this were produced strict practices, we share packages all, including input files, execution scripts, secondary data, post-processing code plotting figures (deposited Zenodo). In view many challenges faced, propose reproducible practices make feasible.