[NukeMan]

Most modern weapons use a neutron pulse tube which accelerate tiny amounts of Tritium or Deuterium into each other, producing a fusion reaction (yup, fusion) which then yields neutrons. The electronic pulse can be exactly controlled (probably to nansecond or smaller accuracy) to intitiate the chain reaction. For the Polonium-Beryllium initiators used in early weapons, the limiting factor is the half-life of Polonium-210 which is ~138 days. You do not need to replace after one half life, but a good rule of thumb is that after 5 half-lives the isotope needs replenishing. That would be about 700 days or ~2 years. Perhaps more frequently than that, as I don't know offhand the strength required. Polonium-210 is an alpha emitter, not neutron emitter. You don't want neutrons emitted until the actual time of initiation. Po-210 reacts with Be-9 to produce neutrons, but both isotopes are kept apart by a foil to avoid premature initiation.

On the other hand, Pu-239 (and Pu-240 which is unavoidably in the isotope mix) produces boatloads of neutrons via spontaneus fission, and any competent terrorist should be able to detect them to see if he is getting the real goods. U-235 is not nearly so bad (about 2 neutrons per second per kilogram of U-235) but still detectable, if you think to look for it.

[steve-b]

You do not need to replace after one half life, but a good rule of thumb is that after 5 half-lives the isotope needs replenishing.

That depends on the nature of the decay product. For instance, tritium needs to be replaced much sooner than a single half-life (12 yr), because its decay product (He3) soaks up neutrons like a sponge and muffles the chain reaction.