On the other hand, we can get a basic idea of the theoretical maximum power efficiency of both a silicon solar cell and photosynthesis using some relatively simple arguments.

If you take a hypothetical crystalline silicon solar cell with well-engineered antireflection coatings or structure, then it will absorb pretty much all the sunlight with a wavelength shorter than about a micron. The internal quantum efficiency of crystalline silicon solar cells is often really close to 100%, so essentially all of the absorbed photons are turned into electrons, each of which carries about 1.1 electron-volts of energy. For comparison, most visible-light photons have somewhere between 1.5 and 3 electron-volts of energy, so there are two sources of energy loss here. First, all of the photons with longer wavelengths than about a micron are not absorbed. Second, most of the absorbed photons are converted to electrons that have lower energy than the original photon. Put everything together with the actual solar spectrum, and you come up with 31% being the theoretical efficiency limit of a crystalline silicon solar cell (if memory serves).

Photosynthesis is more complicated, since as Ronald points out below, a lot of light is simply not absorbed. Also, in bright sunlight the internal quantum efficiency drops (the 100% quantum efficiency number is only true under low illumination). Finally, the electrons that photosynthesis produces are rather lower in energy than those of a silicon solar cell. Strictly speaking, photosynthesis doesn’t produce electrons so much as produce chemical reactions (turning water into oxygen, for example), so it’s even harder to really gauge energy efficiency. If you do something like the article suggests and take the electrons straight from the photosynthetic centers, then the energy of the electrons will strongly depend on exactly where in the photosynthetic chain the electrons were taken from. Putting everything together, I would hazard a guess that natural photosynthesis alone is 5-10% energy efficient, based on the above facts and based on the fact that whole plants are limited to 3% or 4% total energy efficiency (in terms of converting sunlight into plant mass).

In terms of theoretical limits, if you were to take Photosystem II and directly collect its electrons at their most energetic point (which would be really hard), I’d guess around 15% total efficiency to be the limit. The bandgap is about 1.8 electron-volts at the highest-energy point which corresponds to a theoretical Shockley-Queisser efficiency limit of about 28%, but I’m guessing only about half the incident photons with more energy than 1.8 electron volts will be absorbed.

There are many ways to evaluate what might be judged as “best.” Considering the environmental impacts of mining and processing solar cell components, photosynthesis seems to be “best” in terms of being “greenest” in more ways than one.

If you’re doing ultrafast research, I suspect you must be way ahead of most everyone else in the field (of green).

At the end of the day, energy efficiency is almost always more important than quantum efficiency, and by that metric commercial solar cells have a clear advantage.

For background: I do ultrafast research on photosynthesis, motivated in part because of its high quantum efficiency and the potential implications for solar power.