Physicists trap light in a bottle
To get light into an optical cavity, it has to fit. That is, when the photon travels a complete circuit of the sphere, it must travel a whole number of half-wavelengths. The spheres are tiny, so the color difference between two wavelengths that fit is huge. This becomes a problem because we can’t precisely control the size of the spheres during manufacture, and nature chooses which colors of light atoms will interact with—the two rarely end up matching.
The usual solution is to make a bajillion spheres and find one that is close to right. Then you can heat the sphere so that it expands until you get to exactly the right color. It would be much better to just have a resonator that could adapt itself to any color of light.
Inject photons from the outside? Dope the glass with smidgeon of lanthanide ion. Pump the ion into fluorescence. The Stokes-shifted output is your internal source. Was that so hard? Make the balls out of Plexiglas. Now you can put the lanthanide ion into an organic antenna ligand, kick it up to a high coordination number anion overall, and have hugely efficient giant Stokes-shifted fluorescence. UV input, red or near-IR output.
Engineers make things, scientists make stuff. If you need different stuff, things won’t save you.