Well, yes, I agree, physically the sun heats the solid surface and only after its power transfers into a temperature wave, one can see heat conduction inside of solid and initiated convective flow outside. So, probably, it will be accurate enough to model a thin part of a solid volume as a heat source of defined power. I only cannot find any instruction how to do it correctly. Does it have to be a separate cell zone? Refinements? Or just a separate solid with the same material assigned and one side of it will be an interface?