Idle speculation can teach physics, vacuum energy edition

To start the new year, a silly anecdote ending in real science.

Back when I was in grad school, around 25 years ago, I was goofing around chatting with one of my fellow group members, explaining about my brilliant (ahem) vacuum energy extraction machine.  See, I had read this paper by Robert L. Forward, which proposed an interesting idea, that one could use the Casimir effect to somehow "extract energy from the vacuum" - see here (pdf).  

Fig from here.

(For those not in the know: the Casimir effect is an attractive (usually) interaction between conductors that grows rapidly at very small separations.  The non-exotic explanation for the force is that it is a relativistic generalization of the van der Waals force.  The exotic explanation for the force is that conducting boundaries interact with zero-point fluctuations of the electromagnetic field, so that "empty" space outside the region of the conductors has higher energy density.   As explained in the wiki link and my previous post on the topic, the non-exotic explanation seemingly covers everything without needing exotic physics.)

Anyway, my (not serious) idea was, conceptually, to make a parallel plate structure where one plate is gold (e.g.) and the other is one of the high temperature superconductors.  Those systems are rather lousy conductors in the normal state.  So, the idea was, cool the system just below the superconducting transition.  The one plate becomes superconducting, leading ideally to dramatically increased Casimir attraction between the plates.  Let the plates get closer, doing work on some external load.  Then warm the plates just slightly, so that the superconductivity goes away.  The attraction should lessen, and the plate would spring back, doing less work of the opposite sign.  It's not obvious that the energy required to switch the superconductivity is larger than the energy one could extract from running such a cycle.   Of course, there has to be a catch (as Forward himself points out in the linked pdf above).  In our conversation, I realized that the interactions between the plates would very likely modify the superconducting transition, probably in just the way needed to avoid extracting net energy through this process.  

Fast forward to last week, when I randomly came upon this article.  Researchers actually did an experiment using nanomechanical resonators to try to measure the impact of the Casimir interactions on the superconducting transition in (ultrasmooth, quench-condensed) lead films.  They were not able to resolve anything (like a change in the transition temperature) in this first attempt, but it shows that techniques now exist to probe such tiny effects, and that idly throwing ideas around can sometimes stumble upon real physics.