What battery chemistry would be most suitable for an application where the cells are expected to sit a very long time, say 100 years, disconnected, with minimal self discharge?
Assume the storage temperature would be something on the order of outdoor extremes experienced in inhabited places on Earth, say -50°C to 50°C or so.
Assume the operating temperature range would be narrower, say 0°C to 30°C or so.
It need not be a chemistry currently available in any commercial products. Assume environmental concerns can be overcome.
Assume the current demands are relatively low, for a small electronic data processing load and that physical space can be found to make a battery of sufficient capacity.
Assume that a hard switch (not a semiconductor) will disconnect the battery from the load.
It doesn't have to be a voltaic cell necessarily, just a self-contained means to store or generate electric energy. It must not rely on external sources of energy like sun, wind, water, or physical effort (like cranking a dynamo) which might not be available when the system is finally turned on.
To clarify the intent, this would be for a "time capsule" type use where it's expected to be sealed into construction, but must be able to power up in the future with no expectation of available external power.