I think this is going nowhere, but even if you only get 80% of the manufacturing energy back, that might be competitive with other forms of long term energy transport and storage.
For storage you'd have to have the current leakage down to next to nothing for that to be effective. Think about it: 10^9th devices in one meter, that's one micrometer thickness per device. At that surface area and thickness you'd have to have an extremely high grade insulator between the layers or the leakage current (essentially a parasitic resistor of 10^9 meter area with a thickness of a fraction of a micrometer electrically in parallel with the device) will kill any storage application.
Another thing I'm missing very much in this discussion is the equivalent of Betz for this device: it only works if the air has a way to exchange with the surface which would greatly increase the required volume. Without air exchange (and thus humidity exchange) it would eventually just stop working. Unless Brownian motion alone is enough to make it work but I find that hard to imagine. Come to think of it: moving the air through a sandwich of layers that thin would require considerable power!
The point is it doesn't leak because it doesn't actually store energy. It takes 100k joules to make a box that will generate 80k joules over its lifetime. But you can make it here, where and when energy is plentiful, and put it there, where energy is scarce. That's effectively energy storage, if not technically.
If that works out to the same physical size then at 22 Wh that would be a complete waste of space and money. It would likely also cost a very large multiple of that to transport it to the location where energy is scarce. A nice high quality battery with plenty of reserve power would likely be much cheaper, easier to interface to and easier to transport. There are batteries with a 20 year shelf life.
