Crystal dislocations were invented to explain why pure metal crystals were soft compared with homogeneous shear strengths calculated from atomic theory. However, it was later found that pure semiconductor crystals had hardnesses at room temperature comparable with calculated homogeneous shear strengths. Furthermore, it was known that pure metal–metalloid crystals were very hard, although they conduct electricity like metals. The differences cannot be explained by means of a classical mechanical model. However, they could be explained by considering the chemical bonding in differing solid crystals. In particular, hardness depends on the degree of localization of the valence (bonding) electrons; a very old idea. What was new was that it was possible to provide quantitative theories by using the results of quantum chemistry, and relatively simple analysis. Selected sets of crystal types must be treated, of course, just as selected sets of molecules were treated in theoretical chemistry. Otherwise the rationalization became unmanageable.

Chemical Theory of Dislocation Mobility. J.J.Gilman: Materials Science and Engineering A, 2005, 409[1-2], 7-12