Physical Review Research (Jan 2022)

Single-boson exchange representation of the functional renormalization group for strongly interacting many-electron systems

  • Pietro M. Bonetti,
  • Alessandro Toschi,
  • Cornelia Hille,
  • Sabine Andergassen,
  • Demetrio Vilardi

DOI
https://doi.org/10.1103/PhysRevResearch.4.013034
Journal volume & issue
Vol. 4, no. 1
p. 013034

Abstract

Read online Read online

We present a reformulation of the functional renormalization group (fRG) for many-electron systems, which relies on the recently introduced single-boson exchange (SBE) representation of the parquet equations [F. Krien, A. Valli, and M. Capone, Phys. Rev. B 100, 155149 (2019)10.1103/PhysRevB.100.155149]. The latter exploits a diagrammatic decomposition, which classifies the contributions to the full scattering amplitude in terms of their reducibility with respect to cutting one interaction line, naturally distinguishing the processes mediated by the exchange of a single boson in the different channels. We apply this idea to the fRG by splitting the one-loop fRG flow equations for the vertex function into SBE contributions and a residual four-point fermionic vertex. Similarly as in the case of parquet solvers, recasting the fRG algorithm in the SBE representation offers both computational and interpretative advantages: The SBE decomposition not only significantly reduces the numerical effort of treating the high-frequency asymptotics of the flowing vertices, but it also allows for a clear physical identification of the collective degrees of freedom at play. We illustrate the advantages of an SBE formulation of fRG-based schemes, by computing through the merger of dynamical mean-field theory and fRG the susceptibilities and the Yukawa couplings of the two-dimensional Hubbard model from weak to strong coupling, for which we also present an intuitive physical explanation of the results. The SBE formulation of the one-loop flow equations paves a promising route for future multiboson and multiloop extensions of fRG-based algorithms.