[pdonis]

> Isn't there a fixed amount of energy per mass that it takes to convert liquid water into vapor?

Yes.

> Why does it matter that the energy comes from light?

The paper is drawing a distinction between light and "heat", which in the context of these experiments basically involves how you deliver the energy: do you do it by heating up the whole mass of water, or do you do it by shining light at it and having the light interact with individual water molecules?

In a practical sense, this would be expected to potentially increase the efficiency of evaporation, since bulk heating of water involves significant losses--much of the energy you expend doesn't go into the water. If you can find particular wavelengths of light that interact strongly with the water and cause evaporation, you can greatly decrease the amount of input energy that gets lost in the process.