Superconductivity in moiré transition metal dichalcogenide bilayers: comparison of two distinct theoretical approaches

This paper investigates superconductivity in twisted WSe$_2$ bilayers by comparing two theoretical frameworks—a conventional negative $U$-Hubbard model yielding an isotropic $s$-wave gap and a strongly correlated $t$-$J$-$U$ model allowing for unconventional symmetries—to evaluate their consistency with recent experimental observations.

The Gist

Imagine a microscopic dance floor made of two ultra-thin sheets of material (twisted layers of a substance called WSe₂). When you twist these sheets slightly against each other, they create a giant, repeating pattern called a "moiré superlattice." On this dance floor, electrons (the dancers) can move around. Sometimes, instead of just dancing individually, they pair up and move in perfect sync, creating a state called superconductivity, where electricity flows with zero resistance.

The goal of this paper is to figure out why and how these electrons pair up in this specific material. The authors tried two different "rulebooks" (theoretical models) to explain the dance, and they compared which one fits the real-world observations better.

Here is a breakdown of their two approaches using simple analogies:

Approach 1: The "Magnetized Floor" (Negative U-Hubbard Model)

Think of this approach as a scenario where the dance floor itself has a special property that encourages partners to form immediately.

• The Rule: In this model, the electrons are like people who are naturally drawn to each other because of a "negative repulsion" (an attractive force). It's as if the floor is sticky for pairs.
• The Result: The electrons pair up in a very simple, uniform way (called s-wave). Imagine everyone on the dance floor holding hands in a perfect circle, moving in the same direction.
• The Problem: When the authors ran the numbers, this model predicted that superconductivity could happen almost anywhere on the dance floor, as long as the crowd density was just right. However, real experiments show that superconductivity only happens in a very specific spot: right when the dance floor is exactly half-full. This model was too "lenient" and didn't match the strict conditions seen in the lab.

Approach 2: The "Tug-of-War" (t-J-U Model)

This second approach is more complex and realistic. It treats the electrons as if they are playing a high-stakes game of tug-of-war.

• The Rules: Here, electrons naturally hate being on top of each other (strong repulsion), but they also want to move around (kinetic energy). To get along, they have to compromise. They pair up not because the floor is sticky, but because they are forced to cooperate to avoid crashing into each other.
• The Renormalization (The "Heavy Backpack"): The authors used a method called the "Gutzwiller approximation" to account for how much the electrons are pushing against each other. Imagine the electrons are wearing heavy backpacks. When they are in a crowded room (high repulsion), the backpacks get heavier, changing how they move.
• The Result: This model predicts a much more exotic dance. The electrons pair up in a twisted, complex pattern (a mix of d-wave and p-wave symmetries).
• Why it fits better: This model correctly predicted that superconductivity would be unstable if the dance floor was too crowded or too empty. It only became stable right at the "half-full" mark, exactly where the real experiments show it happens. The "heavy backpack" effect (correlations) actually helps stabilize the pair, but only in that specific sweet spot.

The Final Verdict

The authors compared their two rulebooks against real experimental data:

1. The Simple Model (Approach 1) was like a map that said, "You can find treasure anywhere." It was too broad and didn't match the reality that treasure is only found in one specific spot.
2. The Complex Model (Approach 2) was like a detailed map that said, "The treasure is only here, at the intersection of the half-full line and the Van Hove singularity."

The Conclusion:
The paper concludes that the "Complex Model" (t-J-U) is the better description. It suggests that in these twisted material sheets, superconductivity isn't just a simple attraction; it's a delicate balance of strong repulsion and movement. The electrons only pair up successfully when the "crowd density" is just right (half-filling) and the "backpacks" (correlation effects) help stabilize them. This explains why the superconducting state appears as a small, specific "dome" in experiments, rather than spreading out everywhere.

In short: The electrons aren't just falling in love; they are navigating a crowded, high-pressure environment where they can only hold hands if the conditions are perfect.

Technical Summary

Technical Summary: Superconductivity in Moiré Transition Metal Dichalcogenide Bilayers

Problem Statement
Recent experiments have identified a superconducting (SC) phase in twisted bilayer WSe$_2$ (tWSe$_2$), a moiré transition metal dichalcogenide system. This state emerges near half-filling, adjacent to a correlated insulating phase, and exhibits a dome-like stability as a function of electronic concentration. While the system offers unique tunability via twist angle, displacement fields, and gate voltages, the fundamental pairing mechanism and the complete theoretical description of the SC phase remain unestablished. Existing literature suggests Coulomb-driven pairing with anisotropic gaps and mixed singlet-triplet character, though phonon-mediated mechanisms and pure spin-singlet channels have also been proposed. This paper aims to confront two distinct theoretical frameworks applied to tWSe$_2$ to determine which better aligns with available experimental observations regarding the stability and symmetry of the superconducting state.

Methodology
The authors investigate the superconducting state of a single moiré band in twisted WSe$_2$, incorporating Ising-type spin-orbit coupling via spin- and direction-dependent complex hoppings. Two complementary theoretical approaches are employed:

1. Negative $U$-Hubbard Model: This approach utilizes a conventional attractive onsite interaction ($U < 0$) combined with the Hartree-Fock approximation. It assumes a momentum-independent pairing potential, leading to an isotropic $s$-wave gap. This framework treats strong electron-electron repulsion as not directly affecting the paired state, representing a relatively standard pairing scenario.
2. $t$-$J$-$U$ Model with Gutzwiller Approximation: This approach employs a model where pairing is induced by intersite kinetic exchange ($J > 0$) in the presence of substantial onsite Coulomb repulsion ($U > 0$). The interaction term includes a Dzyaloshinskii-Moriya-type term arising from the Ising spin-orbit coupling. To account for strong correlation effects, the authors apply the Gutzwiller variational wave function method. This allows for the analysis of correlation-induced renormalization of Hamiltonian terms, leading to unconventional gap symmetries.

Key Results

• Negative $U$-Hubbard Model:

  • The model predicts a stable isotropic $s$-wave superconducting state.
  • The stability of the SC state is primarily driven by the density of states (DOS). Consequently, the superconducting gap exhibits a dome-like behavior that tracks the Van Hove singularity line in the $(n, D)$ phase diagram (where $n$ is band filling and $D$ is the displacement field).
  • Discrepancy: The model predicts SC stability even away from half-filling ($n=1$) provided the displacement field is tuned to place the Van Hove singularity at the Fermi level. This contradicts experimental data, which shows the SC state is stable only in the vicinity of half-filling.

• $t$-$J$-$U$ Model:

  • The pairing mechanism is intersite, resulting in a $k$-dependent gap with a mixed $d+id$ (singlet) and $p-ip$ (triplet) symmetry, characterized by non-trivial topology (Chern numbers $C = \pm 2, \pm 4$).
  • Renormalization Effects: Strong onsite Coulomb repulsion induces significant renormalization of the Hamiltonian terms. The kinetic exchange term responsible for pairing is enhanced ($\lambda_s^4 > 1$) in the moderately correlated regime ($U/W \lesssim 1$), favoring SC stabilization near half-filling.
  • Suppression at Half-Filling: In the strongly correlated regime ($U/W > 1$), the SC gap is suppressed at exact half-filling due to the emergence of a Mott insulating state ($q^2 \approx 0$).
  • Agreement with Experiment: When the model is extended to include nearest-neighbor Coulomb repulsion (as detailed in Ref. [33]), the SC state is suppressed across most of the phase diagram. The superconducting phase survives only at the intersection of the Van Hove singularity line and half-filling. This specific condition allows correlation-induced renormalization and high DOS effects to act simultaneously, reproducing the experimentally observed stability of the SC state exclusively near half-filling.

Significance and Conclusions
The paper concludes that while the negative $U$-Hubbard model offers a straightforward description of isotropic pairing, it fails to capture the specific experimental constraint of SC stability being confined to the half-filled regime. In contrast, the $t$-$J$-$U$ model, particularly when incorporating Gutzwiller renormalization and intersite Coulomb repulsion, provides a more consistent theoretical picture. It suggests that the superconductivity in twisted WSe$_2$ is an unconventional state driven by kinetic exchange and strong correlations, where the interplay between the Van Hove singularity and correlation-induced renormalization stabilizes the paired phase only in a narrow region near half-filling. This work highlights the necessity of accounting for strong correlation effects and specific gap symmetries to accurately describe the superconducting properties of moiré TMD bilayers.