Nanoparticles of various materials were proposed as carriers of nuclides in targeted alpha particle therapy to at least partially eliminate the nuclear recoil effect causing the unwanted release of radioactive progeny originating in nuclear decay series of so-called in vivo generators. Here, we report on the study of 211Pb and 211Bi recoils release from the 223Ra surface-labelled TiO2 nanoparticles in the concentration range of 0.01–1 mg/mL using two phase separation methods different in their kinetics in order to test the ability of progeny resorption. We have found significant differences between the centrifugation and the dialysis used for labelled NPs separation as well as that the release of 211Pb and 211Bi from the nanoparticles also depends on the NPs dispersion concentration. These findings support our previously proposed recoils-retaining mechanism of the progeny by their resorption on the NPs surface. At the 24 h time-point, the highest overall released progeny fractions were observed using centrifugation (4.0% and 13.5% for 211Pb and 211Bi, respectively) at 0.01 mg/mL TiO2 concentration. The lowest overall released fractions at the 24 h time-point (1.5% and 2.5% for 211Pb and 211Bi respectively) were observed using dialysis at 1 mg/mL TiO2 concentration. Our findings also indicate that the in vitro stability tests of such radionuclide systems designed to retain recoil-progeny may end up with biased results and particular care needs to be given to in vitro stability test experimental setup to mimic in vivo dynamic conditions. On the other hand, controlled and well-defined progeny release may enhance the alpha-emitter radiation therapy of some tumours.