Random physics analogy of the day:
First, semiconductor PN junctions. They store charge under forward bias [current]. Under reverse bias, that charge gets sucked out the terminals, and then the voltage reverses.
Electrochemical cells. They store charge under forward bias. Under reverse bias, that charge gets sucked out the terminals, until the reactants are spent.
Both exhibit logarithmic V(Q) charge curves (well, give or take the variation in chemistry available in cells).
Both exhibit self-discharge. The cell discharges through leakage and slow reactions. The junction discharges through recombination (...leakage and slow reactions).
Suppose we shove charge carriers into a junction without using forward bias. Say by shining a light on the semiconductor (or bombarding it with nuclear radiation). Now voltage magically appears at the terminals, and although we can dynamically deplete the charge in the same way, there's a constant flow of it, so that even at zero DC voltage, it delivers a current.
Welllll, suppose we delivered a constant supply of reactants to the cell. It would be a... fuel... cell.
So there you have it folks, a solar cell is obviously a photonic fuel cell. And likewise, a PN junction is a very, very teensy, high speed battery.
And likelikewise, as I'm sure anyone who works with fuel cells must know, an MPPT can be used with them, just as they can with solar cells (though I would guess, since fuel is a controlled variable, it's not nearly as useful in that domain!).
It's left as An Exercise For The Student to calculate how many 2N3904s are needed to replace a AA cell... ;-)
(This would be a Sunday morning musing, except that it's a day late. So there. :^) )
Tim