storage - What's the highest theoretical energy density for a chemical battery?

This is more a physics/chemistry/nanotech question, but what's the theoretical best energy density you could get out of a chemical battery (or fuel cell), if you could arrange atoms in any manner you wanted? I'm thinking of the nanotech batteries described in Diamond Age. How does it compare to current technologies?

This is specifically about chemical batteries, which could be built atom-by-atom in the charged state, not nuclear, antimatter, CAM, or other more exotic technologies.

Answer

I don't know the actual answer to this question, but I know a least upper bound to the answer, and a means of figuring out the real answer.

Battery scientists have a metric called maximum theoretical specific energy; you can read about the definition in Advanced Batteries by Robert Huggins. Right now, the most energy dense batteries you can buy are lithium ion, which are in the 100-200 Wh/kg range. I don't know what the best battery is, but later in the book, Huggins shows calculations that indicate that Li/CuCl₂ cells have an MTSE of 1166.4 Wh/kg. (5x the capacity of current batteries!)

We know that the highest MTSE is at least 1166.4 Wh/kg; you could use his method to calculate the same value for other chemistries, but the search space is pretty large.

I've also seen references on the internet to Li/O₂ and Al/O₂ batteries with MTSE of 2815 and 5200 Wh/kg, respectively. Not sure how credible those references are. Later references, like this 2008 article in the Journal of the Electrochemical Society, suggest that the MTSE for a Li/O₂ cell is around 1400 Wh/kg.