Ab-initio low-energy dynamics of superfluid and solid 4He

TL;DR

This paper presents a novel method using Genetic Algorithms to extract real-time dynamics from noisy imaginary-time correlation functions in 4He systems, revealing detailed spectral features of superfluid and solid helium.

Contribution

It introduces a new approach combining QMC data with evolutionary algorithms to accurately reconstruct spectral functions of helium systems.

Findings

Recovered sharp quasi-particle excitations in superfluid 4He

Identified multiphonon branches in spectral functions

Discovered a roton-like feature in vacancy-wave excitations

Abstract

We have extracted information about real time dynamics of 4He systems from noisy imaginary time correlation functions f(tau) computed via Quantum Monte Carlo (QMC): production and falsification of model spectral functions s(omega) are obtained via a survival-to-compatibility with f(tau) evolutionary process, based on Genetic Algorithms. Statistical uncertainty in f(tau) is promoted to be an asset via a sampling of equivalent f(tau) within the noise, which give rise to independent evolutionary processes. In the case of pure superfluid 4He we have recovered from exact QMC simulations sharp quasi-particle excitations with spectral functions displaying also the multiphonon branch. As further applications, we have studied the impuriton branch of one 3He atom in liquid 4He and the vacancy-wave excitations in hcp solid 4He finding a novel roton like feature.