FPT3 was the last of the five in-pile integral experiments in the Phébus FP programme, whose overall purpose was to investigate fuel rod degradation and behaviour of fission products (FPs) released via the primary coolant circuit into the containment building. The results contribute to validation of models and computer codes used to calculate the source term for a severe accident with core meltdown in light water reactors. Unlike the previous tests, FPT3 used B4C as absorber material in the pre-irradiated (24.5 GWd/tU) fuel bundle, while featuring a steam-poor period as in FPT2, which used Ag/In/Cd absorber. The main FPT3 containment results are summarised: the source term of FPs, fuel and structural materials from the experimental circuit into the containment; the composition, morphology and deposition processes of aerosols in the containment atmosphere; the specific behaviour of the radiologically significant FP iodine; and finally the performance of passive autocatalytic recombiner (PAR) coupons exposed to the containment atmosphere just after the transient. The major elements contributing to the aerosol mass in the containment are the volatile FPs Cs and Mo, the control rod material B, the cladding material Sn, and the instrumentation materials Re and W (specific to Phébus tests). The fractional compositions, leaving aside the control rod materials, were very similar in FPT2 and FPT3. After reactor shutdown, homogenisation of the aerosol size in the containment led to only one aerosol population, similar to the previous tests. Long-term aerosol deposition in the containment was dominated by gravitational settling and diffusiophoresis, but significant deposits were also measured on the vertical wall, consisting of multi-component aerosols, again comparable with FPT2. A significant result of FPT3 was that iodine is mainly in gaseous form in the atmosphere up to containment isolation; the rest in aerosol form. Another important result was the fast decrease of the iodine concentration in the atmosphere, for total iodine (gas and aerosol), mainly due to deposition on the painted condensers; the depletion of the airborne aerosols Cs and Te was about a factor of 3 slower. As in FPT2, gaseous iodine was mainly in molecular form. Concerning the PAR coupons, which were introduced in the vessel when the overall aerosol, FP and hydrogen releases in the containment atmosphere had ended, analysis of their surface temperature evolution indicated that, despite the very low oxygen concentration in the FPT3 containment, they worked quite well and were not poisoned by CO or FPs during their 30 min exposure time. © 2012 Elsevier B.V.