Bad results make nice photos

I am making thin films of transparent metal hydrides, using a method called reactive sputtering. The samples I prepare react strongly with oxygen after the preparation, which I really do not want them to do. To avoid the sample to take up oxygen from air, I therefore cover them with metal before exposing them to air. But it it is not always successful. Some times the metal capping is too thin, and the oxygen goes through the film, creating tiny cracks and craters. The images above are microscope images of the cracks and craters, taken with an optical microscope trough the transparent films.

The colours are created by light interference in the film, the same effect that creates rainbow-like colours when a thin cap of oil is lying on top of water or wet asphalt, which you probably all have seen.