Establishing the Gauge Invariant Linear Response of Fermionic Superfluids with Pair Fluctuations: A Diagrammatic approach

TL;DR

This paper develops a gauge invariant linear response theory for ultracold fermionic superfluids with pair fluctuations, ensuring consistency with Ward identities and including pseudogap effects within a diagrammatic framework.

Contribution

It introduces a simplified diagrammatic method to construct gauge invariant vertices in fermionic superfluids with pair fluctuations, maintaining Ward identities in the superfluid phase.

Findings

Gauge invariance achieved with a full external EM vertex.

Constructed a WI-maintaining vertex including pseudogap effects.

Applicable to G0G t-matrix approach for pair fluctuations.

Abstract

We present a manifestly gauge invariant linear response theory for ultra-cold Fermi gases undergoing BCS-Bose-Einstein Condensation (BEC) crossover with pair fluctuation effect included, especially in the superfluid phase, by introducing an effective external electromagnetic (EM) field. For pure BCS-type superfluids, the gauge invariance of the linear response theory can be maintained by constructing a full external EM vertex by including the fluctuation of the order parameters in the same way as the the self-energy effect is included in the quasi-particle, therefore the Ward identity (WI) is satisfied. While for the Fermionic superfluids with pairing fluctuation effect included in the quasi-particle self-energy, the construction of a gauge invariant vertex is non-trivial, since in the broken symmetry phase the effect of Nambu-Goldstone modes (collective modes) intertwines with that of the pairing fluctuation. In this paper, we find that under a suitable diagrammatic representation, the construction of such vertex is greatly simplified, which allow us to build a WI-maintaining vertex with pseudogap effects included in the superfluid phase. We focus on the $G_0G$ $t$-matrix approach to the pair fluctuations, although our formalism should also works equally well for the $G_0G_0$ $t$-matrix theory.