Table of Contents
Fetching ...

Finite-temperature signatures of underlying superconductivity in the electron-doped Hubbard model

Wen O. Wang, Thomas P. Devereaux

Abstract

We perform numerically exact determinant quantum Monte Carlo simulations of the Hubbard model and analyze pairing tendencies by evaluating correlation functions at the imaginary-time midpoint ($τ=β/2$), which suppresses high-frequency weight and emphasizes low-energy physics. Using this diagnostic, we identify clear finite-temperature signatures of underlying $d$-wave superconductivity for electron doping, while finding no clear indication upon cooling for hole doping. Our analysis enables direct comparison with ground-state DMRG, revealing consistent real-space pairing patterns. These results provide a practical route to bridge the gap between finite-temperature and ground-state numerically exact simulations of the Hubbard model despite the fermion sign problem.

Finite-temperature signatures of underlying superconductivity in the electron-doped Hubbard model

Abstract

We perform numerically exact determinant quantum Monte Carlo simulations of the Hubbard model and analyze pairing tendencies by evaluating correlation functions at the imaginary-time midpoint (), which suppresses high-frequency weight and emphasizes low-energy physics. Using this diagnostic, we identify clear finite-temperature signatures of underlying -wave superconductivity for electron doping, while finding no clear indication upon cooling for hole doping. Our analysis enables direct comparison with ground-state DMRG, revealing consistent real-space pairing patterns. These results provide a practical route to bridge the gap between finite-temperature and ground-state numerically exact simulations of the Hubbard model despite the fermion sign problem.
Paper Structure (9 sections, 24 equations, 14 figures)

This paper contains 9 sections, 24 equations, 14 figures.

Figures (14)

  • Figure 1: (A,B) Unequal-time $d$-wave pair-field susceptibility, $\chi_d \equiv \beta\langle \mathcal{T}_\tau\Delta_d(\tau=\beta/2)\Delta^\dagger_d\rangle/N_s$ for hole doping ($x > 0$; A) and electron doping ($x < 0$; B). The inset of (A) highlights half filling ($x=0$); the inset of (B) zooms in on the electron-doped regime $|x| \geq 15\%$. (C, D) Superconducting pairing vertex from the unequal-time susceptibility and its uncorrelated bubble, $\Gamma_d \bar{\chi_d} \equiv \bar{\chi_d}/ \chi_d -1$ (with $\bar{\chi_d}$ built from single-particle Green’s functions), for hole doping (C) and electron doping (D). Insets in (C, D) show the doping dependence of $\Gamma_d \bar{\chi_d}$ at fixed temperatures. Parameters: $U/t=8$, $t'/t=-0.25$; cluster $8\times 8$.
  • Figure 2: Real-space pattern of the unequal-time pair-pair correlator $\beta S_{\alpha,\alpha'}(\mathbf{r},\tau=\beta/2) \equiv \beta \sum_i \langle \mathcal{T}_\tau \Delta_{\alpha}(\tau=\beta/2, \mathbf{r}_i + \mathbf{r}) \Delta_{\alpha'}^\dagger(0, {\mathbf{r}_i}) \rangle/{N_s}$ (see Supplementary Materials Formalism Section), at (A, D) hole doping $x=0.03$, (B) half filling $x=0$, and (C, E) electron doping $x=-0.03$. We take the vertical bond connecting $(0,0)$ and $(0,1)$ as the reference. Parameters: $U/t=6$, $t'/t=-0.25$; Temperatures and sizes: $T/t=0.2$ on $16\times 4$ (A-C); $T/t=0.222$ on $8\times 8$ (D, E).
  • Figure 3: Superconducting pairing vertex $\Gamma_d \bar{\chi_d}$ for various interaction strengths $U$ at electron doping $x=-1\%$ (A) and $x=-7\%$ (B). Parameters: $t'/t=-0.25$; cluster $8\times 8$.
  • Figure 4: Negative temperature slope of the pair susceptibility, $-\partial \chi_d/\partial T$, estimated from an Akima spline fit to $\chi_d(T)$. Parameters: $U/t=6$, $t'/t=-0.25$; cluster $8\times 8$.
  • Figure S1: Unequal-time $d$-wave pair-field susceptibility and vertex (definitions as in main-text Fig. \ref{['fig:Pd_and_vertexU8']}), now for $U/t=6$ (other parameters unchanged). Trends are qualitatively consistent with the $U/t=8$ case in main-text Fig. \ref{['fig:Pd_and_vertexU8']}.
  • ...and 9 more figures