Re: what would it take to melt an aluminum can w/ only sunlight

This is a tricky one.  It all depends on the amount of insulation you 
use.  To have any chance of this working, you will need to crush the can 
down as much as possible first to minimise heat loss.  The sun puts out a 
nominal energy flux of around 1400 J/m2/s at 1 a.u. (the distance to the 
earth).  I believe you would need between 10 and 20 kJ to melt an empty 
coke can.  This kind of excludes the use of a lens system unless you 
really go to town with many lenses all focussed down on the can.  The 
alternative would be to use parabolically or hemispherically convex 
mirrors.  For the same power, the area of mirrors required is much larger, 
so a logistical problem may be encountered here.

You may find more optimistic news at
foundry 
or
aluminium_cycle

But my feeling on this is that you will have to move the can a lot closer 
to the sun to get this to work.

Alternatively you could capture the sun's energy as stored electrical 
energy via solar cells and use the derived electrical current to melt the 
can in an inductive or electric arc furnace.  But I expect you will come 
up against power constraints again here.

Sorry I can't be of more help.  Good luck if you do go for it.