A patented time crystal:

https://www.google.com/patents/US9212828
"An Epicatalytic Thermal Diode (ETD) includes one or more ETD cells. Each cell comprises first and second surfaces with a cavity between them, which contains a gas that is epicatalytically active with respect to the pair of surfaces. The surfaces chemically interact with the gas such that the gas dissociates at a faster rate proximate to the first surface than it does proximate to the second surface. Thus, a steady-state temperature differential between the first surface and the second surface is created and maintained.. In various applications, multiple ETD cells are connected in series and/or parallel."

Note that, apart from the steady-state temperature gradient, there are steady-state concentration gradients and respective ETERNAL flows of reactants between the two surfaces.

That catalysts can shift chemical equilibrium was my first heretical idea, about 25 years ago. I believed my argument was convincing and enthusiastically submitted a short paper to Nature - they rejected it without reading it (let alone giving it to referees). My efforts to publish continued, mainly in The Journal of Physical Chemistry, and I was also active on Internet forums. The main result was this:

http://bip.cnrs-mrs.fr/bip10/valevfaq.htm
Athel Cornish-Bowden 1998: "Reading Mr Valev's postings to the BTK-MCA and other news groups and trying to answer all the nonsense contained in them incurs the risk of being so time-consuming that it takes over one's professional time completely, leaving none for more profitable activities. On the other hand, not answering them incurs the even greater risk that some readers of the news group may think that his points are unanswerable and that thermodynamics, kinetics, catalysis etc. rest on as fragile a foundation as he pretends. [...] Can a catalyst shift the position of an equilibrium? No. Absolutely not if it is a true catalyst present at very low concentrations. If it is present at a concentration comparable with that of one or more of the reactants then it may appear to shift the position of equilibrium by mass action effects. However, when it does this it is acting as a reactant, not as a catalyst. Mr Valev's claims to have shown otherwise..."

In my first article submitted to Nature the argument was essentially like this:

Consider the dissociation-association reaction

A - B + C

which is in equilibrium. Let us assume that the forward reaction

A - B + C

is exothermic while the reverse

B + C - A

is endothermic. We add a catalyst, e.g. a macroscopic catalytic surface, and it starts splitting A efficiently but is unable to get together B and C and join them into A. In other words, the catalytic surface accelerates the forward reaction but fails to accelerate the reverse. This looks realistic - the probability that a B molecule and a C molecule will hit the catalytic center simultaneously, so that the center can combine them, could be vanishingly small. Yet, if this is so, the second law is obviously violated - even at equilibrium, there will be local temperature and concentration gradients at the catalytic surface that can in principle be harnessed to do work.

Pentcho Valev