In many laboratory sciences, issues of cleanliness and purity are everpresent concerns. In materials science, keeping things (instruments, materials, people) clean structures the knowledge-making process. Using the work of mary douglas, i examine various contaminants, impurities and defects that are relevant to materials scientists. Importantly, though, definitions of what constitutes `dirt' are multiple, overlapping and, often, formally contradictory; this means that impurities are as much positive resources as threatening pollutants. In materials science labs, where many kinds of actors and forms of life intersect, pollution may be used to rein in confusion and ambiguity. This paper traces various manifestations of laboratory dirt, then examines how (un)cleanliness enables certain moves in the materials science game.