Comment on "Analytical and numerical verification of the Nernst theorem for metals"

Recently, Høye, Brevik, Ellingsen, and Aarseth [Phys. Rev. E 75, 051127 (2007)] claimed that the use of the Drude dielectric function leads to zero Casimir entropy at zero temperature in accordance with Nernst's theorem. We demonstrate that their proof is not applicable to metals with perfect crystal lattices having no impurities. Thus there is no contradiction with previous results in the literature proving that the Drude dielectric function violates the Nernst theorem for the Casimir entropy in the case of perfect crystal lattices. We also indicate that the approximation of temperature independent relaxation frequency used in the paper leads to incorrect values of numerical coefficients in the obtained asymptotic expressions for metals with impurities.