Magnetic Properties of YBa_2Cu_3O_{7-\delta} in a self-consistent approach: Comparison with Quantum-Monte-Carlo Simulations and Experiments

TL;DR

This study uses a modified fluctuation-exchange approximation to analyze magnetic properties of YBa2Cu3O7−δ, comparing results with Quantum-Monte Carlo simulations and NMR experiments, highlighting the importance of temperature renormalization for accurate modeling.

Contribution

It introduces a temperature renormalization in the fluctuation-exchange approximation to better match QMC and experimental results for the Hubbard model of YBCO.

Findings

Good semi-quantitative agreement for U/t=4.5 at high temperatures.

Larger U/t (~8) yields less satisfactory agreement with experiments.

Temperature renormalization improves the comparison between methods.

Abstract

We analyze single-particle electronic and two-particle magnetic properties of the Hubbard model in the underdoped and optimally-doped regime of \YBCO by means of a modified version of the fluctuation-exchange approximation, which only includes particle-hole fluctuations. Comparison of our results with Quantum-Monte Carlo (QMC) calculations at relatively high temperatures ($T\sim 1000 K$) suggests to introduce a temperature renormalization in order to improve the agreement between the two methods at intermediate and large values of the interaction $U$. We evaluate the temperature dependence of the spin-lattice relaxation time $T_1$ and of the spin-echo decay time $T_{2G}$ and compare it with the results of NMR measurements on an underdoped and an optimally doped \YBCO sample. For $U/t=4.5$ it is possible to consistently adjust the parameters of the Hubbard model in order to have a good {\it semi-quantitative} description of this temperature dependence for temperatures larger than the spin gap as obtained from NMR measurements. We also discuss the case $U/t\sim 8$, which is more appropriate to describe magnetic and single-particle properties close to half-filling. However, for this larger value of $U/t$ the agreement with QMC as well as with experiments at finite doping is less satisfactory.