Digital quantum simulation of strong correlation effects with iterative quantum phase estimation over the variational quantum eigensolver algorithm: $\mathrm{H_4}$ on a circle as a case study

TL;DR

This paper demonstrates a hybrid quantum-classical approach combining variational quantum eigensolver and iterative quantum phase estimation to accurately compute ground state energies of the H4 molecule, emphasizing static correlation effects.

Contribution

It introduces a minimally parametrized unitary coupled cluster ansatz and highlights the importance of initial state preparation for reducing noise in quantum phase estimation.

Findings

Initial state quality significantly affects phase estimation accuracy.

The proposed ansatz effectively captures static correlation with fewer parameters.

Proper initial state preparation reduces sampling noise impact.

Abstract

The iterative quantum phase estimation algorithm, applied to calculating the ground state energies of quantum chemical systems, is theoretically appealing in its wide scope of being able to handle both weakly and strongly correlated regimes. However, the goodness of the initial state that is sent as an input to the algorithm could strongly decide the quality of the results obtained. In this work, we generate the initial state by using the classical-quantum hybrid variational quantum eigensolver algorithm with unitary coupled cluster ansatz. We apply the procedure to obtain the ground state energies of the H4 molecule on a circle, as the system exhibits an interplay of dynamic as well as static correlation effects at different geometries. Furthermore, we argue on the importance of static correlation in construction of the reference determinant, and propose a minimally parametrized unitary coupled cluster ansatz, which drastically reduces number of variational parameters while incorporating the static correlation effects in the wavefunction. We demonstrate that a carefully and appropriately prepared initial state can greatly reduce the effects of noise due to sampling in the estimation of the desired eigenphase.