http://canadianjuniorreport.com/grap...ectric-current
"Physicists have built a graphene battery that harvests energy from the thermal movement of ions in solution. (...) When an ion smashes into the graphene strip, the collision generates enough energy to kick a delocalised electron out of the graphene. The electron then has two options: it can either leave the graphene strip and combine with the copper ion or it can travel through the graphene strip and into the circuit. It turns out that the mobility of electrons is much higher in graphene than it is through the solution, so the electron naturally chooses the route through the circuit. It is this that lights up the LED. "The released electrons prefer to travel across the graphene surface...instead of going into the electrolyte solution. That is how the voltage was produced by our device," say Zihan and co. So the energy generated by this device comes from ambient heat. These guys say there were able to increase the current by heating the solution and also by accelerating the copper ions with ultrasound. They even claim to have kept their graphene battery running for 20 days on nothing but ambient heat. But there's an important question mark. One alternative hypothesis is that some kind of chemical reaction is generating the current, just as in an ordinary battery. However, Zihan and co say they ruled this out with a couple of control experiments."

So if the current is not generated by a chemical reaction, this is a violation of the second law of thermodynamics par excellence. Is it so unnatural that some particular ions easily kick a delocalised electron out of the graphene? Perhaps not - it seems that, for graphene, the process is easy in general:

http://physicsworld.com/cws/article/...infrared-light
"For example, it [the graphene] has an ideal "internal quantum efficiency" because almost every photon absorbed by graphene generates an electron-hole pair that could, in principle, be converted into electric current."

Pentcho Valev