An intriguing article on the origin of capacitance scaling limitations showed up in EE Times recently. In what seems almost an obvious and definitely late-arriving finding, a research team at UCSB claim that capacitance dielectric scaling fails at small thicknesses due to depolarizing of the interface layers. This happens at a scaling where it becomes a significant fraction of the dielectric layer itself. Although ostensibly applicable to microelectronics, the significance may find its way into the manufacture of compact energy storage caps. The caveat: we need lots of expensive gold and platinum electrode material whose high electron mobility counteracts the depolarizing effects.
(standard disclaimer: don't trust too much of the stuff written in trade journals like EE Times, especially in regards to explaining the fundamental physics)