Academia.eduAcademia.edu

Path integral formulation of mean-field perturbation theory

1977, Annals of Physics - ANN PHYS N Y

Abstract

We develop a convenient functional integration method for performing mean-field approximations in quantum field theories. This method is illustrated by applying it to a self-interacting 4" scalar field theory and a PJ, four-Fermion field theory. To solve the 4" theory we introduce an auxiliary field x and rewrite the Lagrangian so that the interaction term has the form ~4~. The vacuum generating functional is then expressed as a path integral over the fields x and 4. Since the x field is introduced to make the action no more than quadratic in #, we do the 4 integral exactly. Then we use Laplace's method to expand the remaining x integral in an asymptotic series about the mean field x0 . We show that there is a simple diagrammatic interpretation of this expansion in terms of the mean-field propagator for the elementary field 4 and the mean-field bound-state propagator for the composite field x. The 4 and x propagators appear in these diagrams with the same topological structure that would have been obtained by expanding in the same manner a ~4% field theory in which x and 4 are both elementary fields. We therefore argue that by renormalizing these theories so that the mean-field propagators are equivalent, the two theories are described by the same renormalized Green's functions containing the same three parameters, pa, m*, and g. The quartic theory is completely specified by the renormalized masses $' and mp of the x and 4 fields. These two masses determine the coupling constant g = g($, ma). The cubic theory depends on pa and me and a third parameter g, , g = g($, m2, g,), where g,, is the bare coupling constant. We indicate that g($, m2, gO) < g($, mz) with equality obtained only in the limit g,, + co. When g, + co the wavefunction renormalization constant for the x field in the cubic theory vanishes, and the cubic theory becomes identical to the quartic theory. Our approach guarantees that all quartic theories have the same * Sloan Foundation Research Fellow.