It was noted that widely varying experimental determinations of diffusion parameters had led to models, of solid-state diffusion, which included various mechanisms. By using Si as an example, it was shown that published experimental data obeyed a compensation rule with a characteristic temperature which was close to the temperatures at which measurements were made. This was not consistent with accepted diffusion models, and could only be explained by experimental error. It was proposed that the compensation rule could be used as a test for experimental error, and it was concluded that there was no evidence for more than one self-diffusion mechanism in Si.

The Role of Experimental Error in Arrhenius Plots D.J.Dunstan: Solid State Communications, 1998, 107[4], 159-63