It was noted that the average Gibbs energy per defect was equal to about -kT for any type of lattice defect which was in thermodynamic equilibrium. This result was independent of the defect type, the material, and the environment, but required that a sufficient number of defects exist so that the Gibbs energy changed slowly with the number of defects at equilibrium, while the defect concentration was still so small that they did not interact.

R.A.Johnson: Physical Review B, 1994, 50[18], 13799-800