New paper out:

M. Polettini and M. Esposito, *A marginal observer’s effective thermodynamics*, arXiv:1703.05715

The story is quite simple. Suppose an observer can only measure a (time-integrated) current *J* flowing through a system, that fluctuates (say, the total charge through a resistance with Johnson-Nyquist noise, in some lapse of time). And that he can tune a parameter *X* that is the thermodynamic force conjugate to the current (e.g., the voltage drop). Suppose that he has no idea whatsoever about what happens in the rest of the circuit. Still, he can make a simple experiment. First, he should calibrate the system by tuning *X* to that value *X** where the current vanishes (on average). He can then restore it to its value *X*. Then our claim is that, in a sufficiently long time, the Integral Fluctuation Relation (“Jarzynski identity”) will be observed

*< exp (X-X*) J > = 1*

This comes with a lot of consequences. You can find some in the presentation I gave today at the APS spring meeting in New Orleans. Here the slides

## 2 thoughts on “A marginal observer’s effective thermodynamics”