Hmm...assuming atmospheric pressure inside the little glass tube (atom density ~2.7e19 /cm3) and the ~5300 barn (5.3e-25 cm2) cross section, thats a interaction factor of ~1.4e-5 per cm. For a 1mm square vial, to convert the helium to tritium, you would need a fluence of 1.9e24/cm2 (to convert 100% of the 2.7e17 atoms in each 1x1x10mm vial)...that a bit high (and by "a bit" I mean 60 years in HIFR). Unless the capsules only need a few % of the helium-3 converted back to tritium, I don't see any way you could recharge them in a reasonable timeframe with neutron irradiation with any sort of conventional source. Are my figures just way off? Even mounting them end-on to the flux, and a need for only 1% to be back converted, you would still need a flux of ~2e21...which is nearly a month in a high flux system like HIFR, and a year or more in a conventional system. Since He3 doesn't bond to anything, I cant see anyway of making this work other than aggregating the He3 from tons of them, liquifying it, and irradiating that (and separating out the tritium produced).
Am I missing something? I dont' see how this would work.