- absorption and radiation are correlated (a black surface would radiate more than a white surface when heated to 6000K (assuming they survive the temperature unchanged))
- a solid polished surface reflects better than a powder surface (thats imho a flaw in the article)
- basically you want a surface that is white in solar and black in atmospheric
Exactly. The trick here is that you're exploiting overlaps in emission and absorption spectra in the atmosphere.
One of my high school professors touched on this when talking about heatsink design and I thought I remembered it wrong, glad to see it come up here.