Variational Quantum Simulation of Valence-Bond Solids

TL;DR

This paper presents a hybrid quantum-classical variational algorithm for simulating ground-state phase diagrams of frustrated quantum spin models, successfully identifying various magnetic and valence-bond phases in the thermodynamic limit.

Contribution

It introduces a novel cluster-Gutzwiller ansatz with a parameterized quantum circuit using XY gates for efficient valence-bond generation, enabling simulation of complex quantum phases.

Findings

Successfully mapped the phase diagram of the J1-J2 Heisenberg model.

Identified Neel, columnar anti-ferromagnetic, and valence-bond solid phases.

Demonstrated convergence guided by long-range order onset.

Abstract

We introduce a hybrid quantum-classical variational algorithm to simulate ground-state phase diagrams of frustrated quantum spin models in the thermodynamic limit. The method is based on a cluster-Gutzwiller ansatz where the wave function of the cluster is provided by a parameterized quantum circuit whose key ingredient is a two-qubit real XY gate allowing to efficiently generate valence-bonds on nearest-neighbor qubits. Additional tunable single-qubit Z- and two-qubit ZZ-rotation gates allow the description of magnetically ordered and paramagnetic phases while restricting the variational optimization to the U(1) subspace. We benchmark the method against the J1-J2 Heisenberg model on the square lattice and uncover its phase diagram, which hosts long-range ordered Neel and columnar anti-ferromagnetic phases, as well as an intermediate valence-bond solid phase characterized by a periodic pattern of 2x2 strongly-correlated plaquettes. Our results show that the convergence of the algorithm is guided by the onset of long-range order, opening a promising route to synthetically realize frustrated quantum magnets and their quantum phase transition to paramagnetic valence-bond solids with currently developed superconducting circuit devices.