This 60% UV is just ONE of the configs...

The nanoparticles improve efficiency by 60% in the ultraviolet spectrum. The visible light spectrum is only nominally affected.

It's still pretty cool, though.

This whole series of "only 60% of the UV part" threads is missing the rest of the article. That was just for ONE size of naonparticle, suitable for converting light to the upper end (blue) of the visible range. They ran the tests for another size, suitable for converting to the lower end (red), and got a higher conversion result, including some improvement from converting visible light, as expected.

Solar cells completely miss photons below the bandgap energy and only peel off the bandgap energy from those above it. They have a bandgap in the infrared so they get most photons, but only take that first 0.6 electron-volt chunk of their energy and lose the rest as heat. That's great if you have an infrared photon at 0.603 eV, not so hot for visible light photons at 1.8-3.1 eV, and pretty crummy for UV photons at 3.1 to 12 or so eV.

Films of nanoparticles have an interesting property: They absorb photons of various wavelengths and emit photons of particular wavelengths related to their size. But they don't do that in the solar-cell style of chopping the right-sized hunk off a more energetic photon and throwing the rest away. Instead they are able to combine energy from multiple lower-energy photons to generate one of the desired energy, chop several desired energy photons out of a high-energy one (and keep the leftover shavings to combine with others to make more desired-energy photons), and trade energy among their neighboring particles.

So it was expected that a film of nanoparticles on a solar cell would grab the energy from photons all over the spectrum, convert it to the energy characteristic of the nanoparticle size, and re-emit that. The improvement from efficiently salami-slicing and stacking photons should be better than losses from such things as emitting the photon in the wrong direction, giving a big boost to the cell.

And to some extent that was happening: Feed UV photons to nanoparticles that chunk 'em into something in the 3 eV range and you get more out of the UV hitting the cell than you would without the film - without appreciably affecting the output from the visible light. You're averaging about 1 2/3 IR photons worth of energy, instead of 1, for each incoming photon. Feed it to nanoparticles that chunk it up finer, down to 2 eV or so, and you get more out of your UV and also start improving on even visible light.

That's a good sign for doing what you really wanted to do: Use nanoparticles that emit just a tiny squidge above the solar-cell's bandgap, chunking all the photons into the right size for the cell and wasting very little of their energy. (But maybe still losing a bunch by emitting them in the wrong direction. That might be improved by putting the nanoparticles at the bottom of wells in the cell rather than on a flat surface.)

But the experiment produced a surprise: The VOLTAGE went up! WTF?

That means one of two things:
  a) The nanoparticles affected the bandgap.
  b) The nanoparticles coupled directly into the cell's "circuitry" in some non-obvious way.

b) might lead to something even better: Nanoparticles that capture the photons, chunk and stack them into some desired size (voltage), and deliver them directly to the wiring. That could get virtually ALL the incoming energy into your wires.

A solar cell with efficiencies in the .90s could be a whole heck of a lot better than even the experimenters were originally chasing. So it's no wonder they published now, with only two sizes of particles tested.

Hot DAMN!

On looking at it further it seems they're not getting the "stacking" effect but are getting the salami-slicing (which was the original point) plus this new direct-coupling thing.  That's just fine:  With normal solar cells' small bandgap the salami-slicing alone might produce a sufficient improvement to push the power/cost ratio ahead of grid power for most sunny locations.

It will be interesting to see whether anything comes of this in the next couple years.

If it turns out that nanoparticles tuned to the bandgap and applied by this process capture 60% of the extra photons (as the blue-tuned ones did for UV photons) they would produce a 60% improvement on deep red, 120% on red, 180% on green and 240% on blue.  Even with glass cutting off the bulk of the UV that would about triple a panel's output - just by adding a cheap coating to the cells before panel assembly.

If they are also getting "stacking" they're also losing more photons to re-radiation in wrong directions, so the improvement would be less - like in the "cheap
coating only doubles output" range (if they stay with a flat surface).  Oh, boo hoo!  B-)