Harnessing molecular power: Electricity generation on the nanoscale

https://phys.org/news/2023-10-harnessing-molecular-power-electricity-generation.html

"natural motion of molecules in liquid...zinc oxide strands (nano whiskers) are piezoelectric... implantable medical devices, can scale to kW-scale generators... working to improve energy density of the device by testing different liquids, high-performing piezoelectric materials, and new device architectures and by enlarging the device"

Simon Derricutt commented:

"yep, this works, but the efficiency of conversion is only somewhere around 1.5e-12 when you look at the actual power available. 1.5pW per cm² or 15nW per m². Thus you'd need a pretty huge device (66.7 million m²) to get a kW, and that's quite a bit of Gold needed. The ZnO is of course pretty cheap material, but you still need to grow the nanorods so the end result isn't cheap. Sure, you can stack these devices provided you have enough heat-transfer between the layers, so you don't need 66.7 square kilometres of ground area, but it's still going to be a lot of mass.

"Back in May 2013, in my first article on R-G, I put forward a better design to do this, that might be as much as 0.1% efficient and so produce around 1mW per cm². Might be more efficient than that, of course. Needed a chip-fab to make it, so cost per cm² around equivalent to normal chips, but the design/developments costs also equivalent to chips and so somewhat expensive and way beyond what I could afford.

"The main thing from this paper is that it's a valid violation of 2LoT, that anyone can replicate if they have enough skill and money. Either 2LoT is totally inviolable, or it isn't, and this shows that you can get power from Brownian motion where standard theory (backed by Feynman) says it's totally impossible to do so. Thus the main value here is in proving it's not actually impossible, though I expect most mainstream scientists will still think it's such a small output as to be negligible and so insist that 2LoT stands.

"Thus the value here isn't the actual output, but instead the proof. It's big enough to be measurable from a 2x2cm device, and that itself shows that 2LoT isn't absolute. There are better ways, though. If you go back to Robert Murray-Smith's replications and extensions of the Lovell Monotherm, he produced enough power to light an LED (microwatts rather than the picowatts shown here), 6 orders of magnitude better.... Can we get a few orders of magnitude more by designing it better? I think we can, and I'm working on that. Since that's going to need manufacture, and the others need to make a profit, I'm not going to detail design parameters here, but I figure it's pretty obvious once you've seen the basics.

"For this device, the description of operation isn't quite accurate. They've shown the characteristics of the Schottky diode when the contact between the Gold and ZnO is already there, but in fact that Schottky barrier is only produced when the vibration of the ZnO and Gold nanorods combine to cause such a contact. Then, the difference in work-function creates a Schottky barrier by dragging the carriers out of the depletion-zone that that contact produces. May not be the surface charges on the piezo-electric ZnO, but just the formation of the Schottky barrier and the extraction of the carriers while that depletion-zone is being created. Since the work-function of ZnO is around 4.9eV, and that of Gold is 5.0eV, the barrier won't be deep and the amount of charge transferred won't be much, and this is why they're only seeing a few mV output. They'd probably do far better using SnO2 (about 4.6eV) as the nanorods, or to dope the ZnO with around 6% Ta2O5 (work-function 4.69eV). Maybe use Indium-Tin Oxide (around 4.2eV). Maybe replace the Gold (5eV) with Platinum (5.3eV). Getting the design better can improve things. You need to look at all the details and then optimise each bit. It's still largely random, though, so getting more than a couple of orders of magnitude improvement could be tricky." http://disq.us/p/2wbsvrx