Inversion in a four-terminal superconducting device on the quartet line: I. Two-dimensional metal and the quartet beam splitter

Abstract: In connection with the recent Harvard group experiment on graphene-based four-terminal Josephson junctions containing a grounded loop, we consider biasing at opposite voltages on the quartet line and establish lowest-order perturbation theory in the tunnel amplitudes between a two-dimensional (2D) metal and four superconducting leads in the dirty limit. We present in addition general nonperturbative and nonadiabatic results. The critical current on the quartet line $I_c(\Phi/\Phi_0)$ depends on the reduced flux $\Phi/\Phi_0$ via interference between the three-terminal quartets (3TQ) and the nonstandard four-terminal split quartets (4TSQ). The 4TSQ result from synchronizing two Josephson junctions by exchange of two quasiparticles "surfing" on the 2D quantum wake, and this mechanism is already operational at equilibrium. Perturbation theory in the tunnel amplitudes shows that the 3TQ are $\pi$-shifted but the 4TSQ are $0$-shifted if the contacts have linear dimension which is large compared to the elastic mean free path. We establish the gate voltage dependence of the quartet critical current oscillations $I_c(\Phi/\Phi_0)$. It is argued that "Observation of $I_c(0)\ne I_c(1/2)$" implies "Evidence for the four-terminal 4TSQ" for finite bias voltage on the quartet line and arbitrary interface transparencies. This statement relies on physically-motivated approximations leading to the Ambegaokar-Baratoff-type formula for the quartet critical current-flux relation. It is concluded that the recent experiment mentioned above finds evidence for the four-terminal 4TSQ.