the idea of POF or power over fiber is to couple a moderate amount of solid state laser light, say a few hundred milliWatts or even several Watts (or potentially more, but it gets more challenging), into an optical fiber, using a different wavelength than that that used for communications.
Polymer optical fiber is flexible and can have a much larger core diameter than single mode fiber used for long distance. People use it to hook their stereo equipment together some times.
Converting a few Watts of narrow-wavelength IR light to electrical power can be much more efficient than conversion of full-spectrum solar power; a III-V laser and a III-V photodiode run in photovoltaic mode can be 80 or 90% efficient since the photon energy is just a little bit higher than the bandgap.
Answers and comments to the question Are cubesats deployed with fully discharged batteries? Even those on Sherpa? make it clear that launch delays may lead to situations where batteries can run down. These can be months or even a year or more. Batteries can die and real time clocks can drift.
Question: Has there been any research or reports of testing or just going ahead and using flexible, polymer multimode optical fiber to keep smallsat or cubesat batteries (or any other spaceflight-critical battery) trickle-charged or just plain charged through the use of optical fiber? Or to communicate with them, for example to update real-time clocks or change instructions? Or both?
Why might this be a good thing?
As mentioned in this answer cubesats are often not provided with individual electrical docking ports, especially using connectors qualified for the vibrations of spaceflight, and that would disconnect easily and without risk of failure at satellite deployment. A big advantage of optical fiber is that it requires no physical connection whatsoever! The last centimeter or so could be absolutely free space. The numerical aperture or NA of polymer optical fiber can be around 0.12 for example (there are a range of NA available). At a distance of 1 centimeter from the end of a 1mm core optical fiber, the spot would be less than 4 millimeters in diameter. A 6mm photovoltaic cell on the surface of the cube, or behind a tiny window, or even a plastic lens and a smaller III-V PV cell is all it takes to receive the power and data over the 1 centimeter gap, allowing for a few millimeters of misalignment.
Every cubesat I've seen is completely disconnected from the rocket electrically, unlike most satellites. They typically reside inside of a pod of some kind, such as the P-POD. This means there is no method to charge the satellites. In addition, as such a low priority user, they are often loaded into the satellite weeks, or occasionally months ahead of time (For classified launches). Bottom line, it seems unlikely anytime soon that they would be charged at launch. The cubesats that have been launched thus far seem to be able to compensate for this.