Towards Arbitrary Control of Lattice Interactions in Nonequilibrium Condensates

TL;DR

This paper introduces a method to engineer reconfigurable networks of nonequilibrium condensates with precise control over pairwise interactions, enabling advanced applications like analog Hamiltonian optimization and reservoir computing.

Contribution

It presents a novel approach to control individual interactions in nonequilibrium lattice condensates using dissipative channels and spatially tailored pump profiles.

Findings

Demonstrated control of sign and strength of interactions in a 2D polariton condensate lattice.

Showed how dissipative barriers block unwanted interactions.

Proposed potential applications in analog optimization and reservoir computing.

Abstract

There is a growing interest in investigating new states of matter using out-of-equilibrium lattice spin models in two dimensions. However, a control of pairwise interactions in such systems has been elusive as due to their nonequilibrium nature they maintain nontrivial particle fluxes even at the steady state. Here we suggest how to overcome this problem and formulate a method for engineering reconfigurable networks of nonequilibrium condensates with control of individual pairwise interactions. Representing spin by condensate phase, the effective two spin interactions are created with nonresonant pumping, are directed with dissipative channels, and are further controlled with dissipative gates. The dissipative barriers are used to block unwanted interactions between condensates. Together, spatial anisotropy of dissipation and pump profiles allow an effective control of sign and intensity of the coupling strength between any two neighboring sites independent of the rest of the spins, which we demonstrate with a two dimensional square lattice of polariton condensates. Experimental realisation of such fully-controllable networks offers great potential for an efficient analog Hamiltonian optimizer and for reservoir computing.