How many megapixels would a camera need to resolve individual atoms?

Won't work - you'll run out of light befor you'll be able to see atoms.

As you may be aware, light has wave characterisitics, and (unless you do som _very_ fancy tricks), you won't be able to see anything smaller than the wavelength of light (and even with those fancy tricks, you'll gain maybe a factor of 10). As for the numbers: wavelength of visible light is (roughly) in the 350...700 nm range, while atoms are in the 100 pm (=0.1 nm) range (and below).

Next few problems: how will you deal with the atoms in the lens(es) and the air (inside the camera, inside the lens, between lens and object)?

Oh, and finally, the number of (mega)pixels is irrelvant for your question - at least unless you state the sensor size and focal length of the lens (and, if you want to be exact, the diameter of the circle of confusion).

pixel density alone would not work. You'd need a lens to focus the image of the individual atoms. And if the lens can do that, then pixel density becomes a secondary issue - but the pixel density and the lens need to work together. Electron microscopes are used as the "lens" - and you need to slow the fast moving electrons so they are visible.

http://newsroom.ucla.edu/portal/ucla/new-microscop...