2023MNRAS.519.2646B

Polluted white dwarfs serve as astrophysical mass spectrometers - their photospheric abundances are used to infer the composition of planetary objects that accrete onto them. We show that due to asymmetries in the accretion process, the composition of the material falling onto a star may vary with time during the accretion of a single planetary body. Consequently, the instantaneous photospheric abundances of white dwarfs do not necessarily reflect the bulk composition of their pollutants, especially when their diffusion time-scales are short. In particular, we predict that when an asteroid with an iron core tidally disrupts around a white dwarf, a larger share of its mantle is ejected, and that the core/mantle fraction of the accreting material varies with time during the event. Crucially, this implies that the core fraction of differentiated pollutants cannot be determined for white dwarfs with short diffusion time-scales, which sample only brief episodes of longer accretion processes. The observed population of polluted white dwarfs backs up the proposed theory. More white dwarfs have accreted material with high Fe/Ca than low Fe/Ca relative to stellar abundance ratios, indicating the ejection of mantle material. Additionally, we find tentative evidence that the accretion rate of iron decreases more rapidly than that of magnesium or calcium, hinting at variability of the accreted composition. Further corroboration of the proposed theory will come from the upcoming analysis of large samples of young white dwarfs.