The Finest Mirrors Ever Made, And Their 10-Minute Mission

Check out the smoothness claim in this article "If one could see the surface at an atomic level, it would show no greater valleys or peaks than two atoms in either direction."

Any thoughts on how this is achieved: 10⁶ grit sand paper?

If I recall, I read in Reader Digest, similar boasting the same type of claims on the Hubble mirror before launch.

And we know how that turned out. I hope they gave it an eye test before launch.

Hi BDThompson, This is achieved by polishing the surface with cerium oxide, and what happens is that the tops of the small hills are melted and run into the valley's, thus creating a near perfect finish, and before you ask how I know this, I cut gemstones for a living.!

The accuracy of polishing they've done isn't wildly greater than is typical for telescope mirrors. There are two parameters used for defining a telescope surface. One is the overall optical figure. The primary mirror optical surface may be a section of a sphere, or a paraboloid, for example. There are a number of concave shapes used in different telescope designs, including some fairly complex aspherical surfaces. So one parameter measures how much the finished surface deviates from the theoretical 'perfect' surface they are trying to obtain. This deviation from perfection is measured in fractions of a wavelength of light, and a patient amateur telescope maker can achieve 1/20th of a wavelength without any exotic equipment. Special tools can help achieve even higher levels of accuracy - if the tools are made correctly. In the case of the Hubble, one tool was off by about a millimeter and caused the optician to polish the surface to the wrong focal point. Had a simple test known as a 'foucalt test' been done - a test used by every amateur telescope mirror maker - the flaw in the Hubble mirror would have been detected.

The other parameter measured is surface roughness, sometimes referred to as 'orange peel'. The type of polishing compound used somewhat determines this surface roughness, but it's also dependent on the patience of the optician and the precision he uses in preparing the working surfaces. Traditionally 'jeweler's rouge', which is a fine iron oxide powder, can yield a very smooth surface free of 'orange peel'. It may be that high-tech optical companies have an better material these days. (Cerium Oxide yields a fast polish, but tends to leave an 'orange peel' surface. For telescope mirrors, sometimes cerium oxide is used first, then followed with 'jeweler's rouge'.)