Is Thibado’s Graphene Brownian Capacitor Charger Perpetual Motion of the Second Kind? In Fluctuation-induced current from freestanding graphene (peer-reviewed version on Phys. Rev. E, note: behind a paywall) Thiabado, et al, report the extraction of work from brownian motion.  The experimental set up involves graphene in close but insulated contact with an electrode that charges a battery and a storage capacitor until a switch shunts the potential through a resistor.

This seems to be a Feynman-Smoluchowski ratchet hence perpetual motion of the second kind.  If so, where is the flaw in the experiment?
 A: Thibado "published" about this idea three years ago  https://researchfrontiers.uark.edu/good-vibrations/

https://youtu.be/wrleMqm3HiU
He now added complications (diodes etc) but that won't help.
The system with mechanical noise makes it a bit more complicated but not really different from a resistor with thermal noise. It is like trying to get energy from the thermal voltage of a resistor. Cannot be done.
Edit: there is an obvious source of energy in the schematic of their new preprint: that battery. I suspect that the bias voltage is the source of the power that they detect. (But I have not analyzed this in detail.)
A: The interpretation is almost certainly wrong.  A graphene film cannot "ripple" due to a static, spatially uniform temperature.  It can only mechanically deform in response to a changing temperature or a temperature gradient.  If all the calculations are done properly, the graphene film device will be determined to be a heat engine whose performance is consistent with the usual laws of thermodynamics.  I can't state that from direct evidence (i.e., an experiment that measures temperature and entropy changes associated with this device), but the indirect evidence is overwhelming.
