Thermoelectric has terrible efficiency, tho. Also, graphene isn’t inexpensive, and actual macroscopic conductivity is much worse than aluminum or copper.
Efficiency may not be much of an issue if you're spinning up a kilometers-wide, microns-thick graphene sheet at a Lagrange point or something.
At the solar flux levels at the point of that probe (light intensity decays by distance squared), efficiency is probably mostly irrelevant. What's the challenge though is how you irradiate all the heat away to provide a large enough temperature delta.
It’d be neat to see a divergent lens made of carbon metamaterial at the aft of the craft. (... or convergent depending on the shape of the craft)