Strong Pairing Originated from an Emergent $\mathbb{Z}_2$ Berry Phase in La$_3$Ni$_2$O$_7$

Jiaxin Zhang, Hao-Kai Zhang, Yi‐Zhuang You, Zheng-Yu Weng

Abstract

The recent discovery of high-temperature superconductivity in La$_3$Ni$_2$O$_7$ offers a fresh platform for exploring unconventional pairing mechanisms. Starting with the basic argument that the electrons in $d_{z^2}$ orbitals nearly form local moments, we examine the effect of the Hubbard interaction $U$ on the binding strength of Cooper pairs based on a single-orbital bilayer model with intralayer hopping $t_{\|}$ and interlayer super-exchange $J_{\perp}$. By extensive density matrix renormalization group calculations, we observe a remarkable enhancement in binding energy as much as $10$-$20$ times larger with $U/t_\|$ increasing from $0$ to $12$ at $J_{\perp}/t_\|\sim 1$. We demonstrate that such a substantial enhancement stems from a kinetic-energy-driven mechanism. Specifically, a $\mathbb{Z}_2$ Berry phase will emerge at large $U$ due to the Hilbert space restriction (Mottness), which strongly suppresses the mobility of single particle propagation as compared to $U=0$. However, the kinetic energy of the electrons (holes) can be greatly restored by forming an interlayer spin-singlet pairing, which naturally results in a superconducting state even for relatively small $J_\perp$. An effective hard-core bosonic model is further proposed to estimate the superconducting transition temperature at the mean-field level.

Introduction

The discovery of superconductivity in pressurized $\text{La}_3\text{Ni}_2\text{O}_7$ (LNO) at temperatures up to 80K has sparked intense debate regarding its underlying pairing mechanism. A critical question is the role of the on-site Coulomb repulsion ($U$), or "Mottness." While some argue that the high doping levels in LNO might render $U$ irrelevant, this research demonstrates that strong electron-electron repulsion is not merely a hindrance but a primary catalyst for high-$T_c$ superconductivity in this system.

Methods and Model

The researchers utilized a single-orbital bilayer model focusing on $d_{x^2-y^2}$ orbitals, with the assumption that $d_{z^2}$ orbitals provide the necessary interlayer magnetic coupling ($J_\perp$). To solve this model, they employed the Density Matrix Renormalization Group (DMRG) method, a high-precision numerical technique for finding ground states in low-dimensional quantum systems. Simulations were performed on bilayer lattices of various sizes (up to $32 \times 2 \times 2$) to investigate the dependence of binding energy on $U$ and $J_\perp$.

Results: The "Mottness" Boost

A central finding of the study is the dramatic enhancement of Cooper pair binding energy in the presence of strong correlations. When the Hubbard interaction $U$ is increased from 0 (the weak-correlation limit) to 12 (the strong-correlation limit), the binding energy $E_b$ increases by a factor of 10 to 20 ($R_b$). This peak enhancement occurs near the experimentally relevant regime where $J_\perp/t_\parallel \sim 1$.

Enhancement of binding energy $R_b$ and $E_b$ as a function of interlayer coupling $J_\perp$ for different system widths $L_y$.

The Phase-String Mechanism

To explain this enhancement, the authors identify an emergent $\mathbb{Z}_2$ Berry phase, also known as a "phase-string." In the large-$U$ limit, the movement of a single hole is restricted by the requirement that it cannot occupy a site already holding an electron. This restriction introduces a non-trivial sign change every time a hole hops past a spin, leading to destructive interference that severely limits single-particle mobility.

However, when two holes from different layers form an interlayer spin-singlet pair, these individual phase frustrations cancel out. This allows the pair to move through the lattice much more freely than a single hole could. Consequently, the pairing is driven by the system's drive to lower its kinetic energy, which is otherwise suppressed by Mottness.

Illustration of the phase-string mechanism (a-c) and the resulting superconducting transition temperature $T_c$ dome (e) peaking at filling $\nu=1/4$.

Discussion and Implications

This kinetic-energy-driven mechanism provides a robust explanation for several experimental observations in LNO. The effective bosonic model derived from this theory predicts a $T_c$ dome that peaks at a filling of $\nu = 1/4$, matching the experimental electronic structure. Furthermore, the theory accounts for the "strange metal" behavior (linear-$T$ resistivity) observed above $T_c$ as a consequence of holes scattering off the random phase fluctuations of broken pairs. While future work is needed to fully address the multi-orbital nature of LNO and extend these findings to a full 2D limit, this research establishes a fundamental link between Mottness and high-temperature pairing.

Summary updated

Abstract

The recent discovery of high-temperature superconductivity in La$_3$Ni$_2$O$_7$ offers a fresh platform for exploring unconventional pairing mechanisms. Starting with the basic argument that the electrons in $d_{z^2}$ orbitals nearly form local moments, we examine the effect of the Hubbard interaction $U$ on the binding strength of Cooper pairs based on a single-orbital bilayer model with intralayer hopping $t_{\|}$ and interlayer super-exchange $J_{\perp}$. By extensive density matrix renormalization group calculations, we observe a remarkable enhancement in binding energy as much as $10$-$20$ times larger with $U/t_\|$ increasing from $0$ to $12$ at $J_{\perp}/t_\|\sim 1$. We demonstrate that such a substantial enhancement stems from a kinetic-energy-driven mechanism. Specifically, a $\mathbb{Z}_2$ Berry phase will emerge at large $U$ due to the Hilbert space restriction (Mottness), which strongly suppresses the mobility of single particle propagation as compared to $U=0$. However, the kinetic energy of the electrons (holes) can be greatly restored by forming an interlayer spin-singlet pairing, which naturally results in a superconducting state even for relatively small $J_\perp$. An effective hard-core bosonic model is further proposed to estimate the superconducting transition temperature at the mean-field level.
BibTeX
@inproceedings{zhang2024strong,
  author = {Jiaxin Zhang and Hao-Kai Zhang and Yi‐Zhuang You and Zheng-Yu Weng},
  title = {{Strong Pairing Originated from an Emergent $\textbackslash\{\}mathbb\{Z\}\_2$ Berry Phase in La$\_3$Ni$\_2$O$\_7$}},
  booktitle = {Physical},
  year = {2024},
  doi = {10.1103/physrevlett.133.126501},
}
Figure