[ramchip]

> Thus, as long as total energy is conserved, there's no problem.

Where did the other laws go?

> The second issue is that solar heating is due to radiation, not conduction/convection, so there's no heat exchange as such.

I don't understand how there could be "solar heating" but no "heat exchange".

Unless I misunderstood the papers I cited, the efficiency equations show that there is a limit (efficiency drops to 0% with Ts=Ta), but I'll still try to address the "intuition" part of your post.

Let's say we are in a black room with no windows, everything being at room temperature. I use lenses to concentrate the infrared light from the walls onto a black cube. Would the cube heat up from the concentration of photons?

[anthonyb]

> I don't understand how there could be "solar heating" but no "heat exchange".

Hmm, I thought that "heat exchange" referred to direct transfer of heat energy, perhaps I'm misremembering. http://en.wikipedia.org/wiki/Heat_transfer#Radiation doesn't seem to differentiate, but bear in mind that there's no way to focus molecular vibration with a mirror.

And if more photons hit the cube with your lens than without it, then yes, the cube will heat up.

[ramchip]

Then focus the infrared light on a thermoelectric generator instead of a cube. The generator powers a widget. The dissipated heat from the generator and the friction on the widget goes back to the walls and is later reemitted as infrared.

That's a perpetual motion machine.

[anthonyb]

...until it hits a new equilibrium, at which point you're back to square one.

btw, from what I've been reading, it looks like you're right (ie. you can't heat anything directly to > the sun's temperature) but I can't find any explanation which makes any intuitive sense at all, just lots of stuff like http://en.allexperts.com/q/Physics-1358/Black-body-radiation... (note the hand waving) or else a bunch of hard core entropy equations.