Physicists at MIT have created the first perfect mirror. If this is true, could enough mirrors be directed at a space ship like Archimedes "Burning Mirror" would at a boat from automated solar farm in orbit work to power spacecraft: How far can usable light be projected from multiple points in orbit to intersect at a single point in space?
The limiting factor for focusing light in this way is actually not the mirror's material properties but the quality of the grinding of the mirror. To focus light on a point, the mirror must be nearly a parabola (it would need to be a parabola if the sun was infinitely far away... but practically speaking, it's far enough away to treat it as infinitely far).
Parabolas are very difficult to grind. Typically we grind a section of a sphere (which is geometrically easy to construct) and then do clever tricks to grind that out into a parabola. It's hard. When I got to tour the lab which made the Hubble Space Telescope primary mirror, they said that the last round of polishing to get the mirror as parabolic as possible was done with a human finger and some talcum powder. You can imagine how long it must have taken to finish the 2.4m diameter mirror!
Modern telescope mirrors have been playing with active optics, where the mirror has a set of actuators behind it which bend the mirror ever so carefully into the right shape. These can do so adaptively, so if your mirror gets hot on one side and cold on the other, these actuators can mitigate some of the effect of the warping due to uneven heating.
Of course, one thing worth thinking about in this scenario is that no matter how many mirrors you have focused on a point, they cannot heat a target up any hotter than the surface of the sun. This is a thermodynamic issue. If the object were hotter, it would actually radiate so much energy that it would actually heat the sun with the mirrors! For instance, there's no way to heat an object up to the temperature of the core of the sun, where fusion occurs.
So the limit for mirrors is once you have enough solar-multiplication to counteract the radiative heat losses of the object into space. As it turns out, its actually pretty hard to cool down an object in space. Modern satellites have to spend a lot of care dealing with the sun's energy. So you actually don't need a very large mirror in order to heat a point up to rather close to the temperature of the surface of the sun.