Critical analysis of a variational method used to describe the molecular electron transport

In a recent paper [I.\ B^aldea and H.\ K"oppel, \prb {\bf 78}, 115315 (2008)], we showed that a variational approach [P.\ Delaney and J.\ C.\ Greer, \prl {\bf 93}, 036805 (2004)] proposed to compute the electron transport through molecules, which is based on boundary constraints of the Wigner function, is unable to correctly describe the zero-bias conductance of the simplest uncorrelated and correlated systems. In the present paper, we extend our previous analysis of the linear response limit of that approach, by considering, instead of the Wigner function, general constraints. We demonstrate that, if, as usual in transport theories, the quasi-particle distributions in electrodes are constrained, this method yields the completely unphysical result that the zero-bias conductance vanishes. Therefore, we conclude that the variational approach itself is defective.