>5.1V zener
Way too low for Vdd=5V. The leakage current at 5V will cause problems so
close to the zener voltage.

>1Meg, high Vdd, damage, when
The real problem is, that the decoupling capacitor that is obligatory in
parallel with the chip Vdd will build up enough energy to damage almost
anything as the chip fails in thyristor mode (voltage raises until it
trips, then you have full current through the chip). You can compute how
many Joules from 24V to 0.6V (assuming it is not fried already) given
W=C*U^2/2, the stored energy (actually only half of it goes into the chip,
but that's another story).

The worst part is, that this almost never happens during normal operation,
only when sleeping. Then, you nicely debug your thing, it works, you have
it sleep, you make 20 of them, you test them, and 10 never wake up
again... Go figure.

>one stuck, works only with high voltage
Various things are known to go wrong with PICs that were exposed to er...
unhealthy conditions. They mostly boil down to charges injected into the
bulk at unexpected locations. Or so I am told. Anyway, it happened to me
too, even after 'overprogramming' an EPROM type. The cure was, to put the
chip in a normal toaster oven at 80 degrees C overnight (NOT a microwawe
please, and do this at your own risk of course). Then it was erased for 2
hours under a mercury vapor lamp (500W) (DON'T do this without eye
protection and ozone evacuation etc.). This made it as good as new. Both
procedures are based on known ways to accelerate charge evacuation from
semiconductor bulks...

Peter (plp)