Orbital-Selective $d$-wave Superconductivity in the Two-Band $t$-$J$ Model: Possible Applications to La$_3$Ni$_2$O$_7$

Using variational Monte Carlo simulations on a two-band $t$-$J$ model motivated by La$_3$Ni$_2$O$_7$, the study reveals that a robust orbital-selective $d$-wave superconducting state emerges exclusively from the itinerant orbital, while the quasi-localized orbital suppresses superconductivity by forming competing local bound states, suggesting that enhancing $T_c$ requires minimizing the involvement of localized $d_{z^2}$ orbitals.

The Gist

The Big Picture: A Dance Floor with Two Types of Dancers

Imagine a crowded dance floor representing a superconductor (a material that conducts electricity with zero resistance). In this specific material, La₃Ni₂O₇ (a type of nickel-based superconductor), the "dancers" are electrons.

For decades, scientists studied a simpler version of this dance floor (the "cuprates") where everyone danced in a very specific, coordinated way called d-wave pairing. This coordination is what allows the electricity to flow without friction.

This new paper asks a big question: What happens if we add a second, different type of dancer to the floor?

In this material, there are two types of electron "orbitals" (dance spots):

1. The "Itinerant" Dancer (Orbital-0): These electrons are energetic, fast, and love to move around the whole floor. They are the ones who usually lead the coordinated dance.
2. The "Quasi-Localized" Dancer (Orbital-1): These electrons are shy, slow, and prefer to stay in one spot. They don't move much.

The Discovery: The Shy Dancer Ruins the Party

The researchers used a powerful computer simulation (called Variational Monte Carlo) to watch how these two types of dancers interact. They found a surprising and somewhat counter-intuitive result:

The shy, slow dancer (Orbital-1) actually hurts the superconductivity.

Here is the metaphor for what is happening:

• The Ideal Dance: When only the fast dancers are present, they link up in pairs and move in perfect unison across the floor. This is the superconducting state.
• The Disruption: When the slow, shy dancers are introduced, they don't join the group dance. Instead, they act like sticky spots or potholes on the dance floor.
• The "Energy Defect": The fast dancers get distracted. Instead of linking up with other fast dancers to form a smooth wave, they get "stuck" trying to pair up with the slow, stationary dancers.
  • Think of it like a fast runner trying to hold hands with a person standing still. The runner gets stuck, the rhythm breaks, and the smooth flow of the race is ruined.
  • The paper calls these stuck pairs "local bound states" or "energy defects." They act like anchors that drag down the performance of the whole group.

The Key Finding: "Orbital-Selective" Superconductivity

The most important discovery is that the superconductivity is orbital-selective.

• The fast dancers (Orbital-0) can still superconduct, but only if they ignore the slow dancers.
• The slow dancers (Orbital-1) never join the superconducting dance; they just sit there and block the path.
• The more slow dancers you add, the worse the superconductivity gets. The paper shows that as the number of "shy" electrons increases, the ability of the material to conduct electricity without resistance drops steadily.

Why Does This Matter? (The La₃Ni₂O₇ Connection)

This material, La₃Ni₂O₇, is a hot topic because it might be a "high-temperature" superconductor (one that works at temperatures we can actually use, not just near absolute zero).

The researchers realized that in this specific material, the "shy" electrons (derived from the $d_{z^2}$ orbital) are naturally present and are causing trouble. They are acting as the potholes on the dance floor.

The Solution:
If you want to make this material a better superconductor (raise its critical temperature, $T_c$), you shouldn't try to make the shy dancers dance. Instead, you should try to make them leave the dance floor.

The paper suggests that if scientists can use physical tricks—like squeezing the material (pressure), changing the chemicals slightly (substitution), or stretching it (strain)—to push the energy of the "shy" electrons up, they will stop participating. This would leave only the "fast" dancers, allowing them to dance in perfect harmony and creating a much stronger superconductor.

Summary in One Sentence

This paper explains that in the new nickel-based superconductor, a second type of electron acts like a "traffic jam" that stops the main superconducting current from flowing, and the best way to fix it is to find a way to remove that traffic jam entirely.

Technical Summary

1. Problem Statement

The paper addresses a fundamental gap in the theoretical understanding of high-temperature superconductivity (SC) in multi-orbital systems.

• Context: The single-band $t$-$J$ model has been the canonical framework for cuprates, successfully explaining robust $d$-wave SC via Variational Monte Carlo (VMC) simulations. However, the recently discovered high-$T_c$ superconductivity in bilayer nickelates (specifically La$_3$Ni$_2$O$_7$) involves a complex multi-orbital electronic structure ($d_{x^2-y^2}$ and $d_{z^2}$ orbitals) that cannot be captured by a single-band model.
• The Question: How does the inclusion of a second active orbital (specifically a quasi-localized one) impact the well-established $d$-wave superconducting state? Does it enhance pairing through new channels, or does it compete with and suppress the existing order?
• Motivation: La$_3$Ni$_2$O$_7$ exhibits a tunable multi-orbital environment where the interplay between itinerant and localized orbitals is crucial, yet the specific mechanism governing SC in this two-band regime remains unresolved.

2. Methodology

The authors employ a two-band $t$-$J$ model on a square lattice, solved using Variational Monte Carlo (VMC) simulations.

• Model Hamiltonian:
  • Orbitals: Two orbitals are defined:
    • Orbital-0: Itinerant, $d_{x^2-y^2}$-like, with large isotropic hopping $t_{00}$.
    • Orbital-1: Quasi-localized, $d_{z^2}$-like, with small hopping $t_{11}$.
  • Hopping: Includes intra-orbital ($t_{00}, t_{11}$) and inter-orbital ($t_{01}$) terms. Crucially, due to orbital symmetry, inter-orbital hopping $t_{01}$ changes sign between $x$ and $y$ directions ($t_{01}^x = -t_{01}^y$).
  • Interactions: Derived from the large-$U$ limit of a Kugel-Khomskii model. The exchange term ($H_{ex}$) includes spin-spin interactions and density-density interactions mediated by superexchange.
  • Constraints: The Gutzwiller projector ($P_G$) forbids double occupancy, enforcing the Mott insulating character at quarter-filling.
• Simulation Technique:
  • Wavefunction: A Gutzwiller-projected mean-field wavefunction $|\psi\rangle = P_G |\psi_{MF}\rangle$, where $|\psi_{MF}\rangle$ is the ground state of a Bogoliubov-de Gennes (BdG) Hamiltonian.
  • Optimization: Variational parameters (hopping amplitudes and pairing gaps) are optimized to minimize the total energy $E = \langle \psi | H | \psi \rangle / \langle \psi | \psi \rangle$ using the stochastic-reconfiguration method.
  • Parameters: Simulations were performed on $20 \times 20$ lattices with doping $\delta = 0.16$ (quarter-filling corresponds to $n=1$, so $n=0.84$). The study varies $t_{01}$, $t_{11}$, and an onsite energy splitting parameter $\mu_z$.

3. Key Contributions and Theoretical Framework

• Energy Hierarchy Analysis: The authors derive an analytical understanding of the superexchange energy scales. They identify a competition between:
  1. Intra-orbital AFM superexchange: Drives $d$-wave pairing in the itinerant orbital-0 (energy scale $\sim t_{00}^2/U$).
  2. Inter-orbital density-density attraction: A spin-independent term ($\sim t_{00}^2/U$) that binds an electron in orbital-1 to an electron in orbital-0.
• The "Defect" Mechanism: The paper proposes that the quasi-localized orbital-1 does not act as a partner for superconductivity. Instead, the inter-orbital attraction forms local, spin-inert bound states between orbital-0 and orbital-1 electrons. These bound states act as "energy defects" that sequester the itinerant electrons, disrupting the phase coherence required for long-range $d$-wave superconductivity.

4. Key Results

• Orbital-Selective $d$-wave SC:
  • The superconducting order parameter is exclusively orbital-selective. Robust $d$-wave pairing ($\Delta_{00}$) emerges solely from the itinerant orbital-0.
  • Pairing amplitudes in other channels ($\Delta_{11}$ and $\Delta_{01}$) are negligible (vanishingly small, $\sim 0.005$), effectively acting as a pseudogap rather than a coherent condensate.
• Suppression by Orbital-1 Occupancy:
  • Increasing the inter-orbital hopping $t_{01}$ or the intra-orbital hopping $t_{11}$ leads to a transfer of electrons from orbital-0 to orbital-1.
  • As the occupancy of orbital-1 ($\langle n_1 \rangle$) increases, the SC order parameter $\Delta_{00}$ is monotonically suppressed.
  • This confirms that the presence of the second orbital is detrimental to SC stability in this regime.
• Application to La$_3$Ni$_2$O$_7$:
  • The model is mapped to La$_3$Ni$_2$O$_7$, where the active bands correspond to interlayer antisymmetric molecular orbitals ($|x, -\rangle$ and $|z, -\rangle$).
  • The authors introduce an onsite energy splitting $\mu_z$ to mimic structural or chemical perturbations.
  • Result: Increasing $\mu_z$ (which lowers the energy of the less-itinerant $|z, -\rangle$ orbital) increases its occupancy and suppresses the SC order parameter.
  • Prediction: To enhance $T_c$ in La$_3$Ni$_2$O$_7$, one must suppress the participation of the localized $d_{z^2}$-derived orbital. This can be achieved by tuning the system to reduce the energy splitting between the orbitals (making them more degenerate) or by structural/chemical means that raise the energy of the localized orbital.

5. Significance

• Theoretical Advancement: The paper moves beyond the single-band paradigm, establishing that in multi-orbital Mott systems, the presence of a quasi-localized orbital can fundamentally compete with and degrade superconductivity, rather than simply adding new pairing channels.
• Mechanism for Nickelates: It provides a microscopic explanation for the SC in La$_3$Ni$_2$O$_7$, identifying the "orbital-selective" nature of the pairing and the detrimental role of the $d_{z^2}$ orbital.
• Experimental Guidance: The work offers a concrete, testable strategy for material engineering: maximizing $T_c$ requires minimizing the occupancy of the localized orbital. This suggests that future experiments should focus on strain engineering, chemical substitution, or interface engineering to tune the orbital energy levels and suppress the localized character.
• General Principle: The findings suggest a general principle for multi-orbital correlated systems: if a localized orbital forms strong inter-orbital bound states with itinerant carriers, it acts as a source of decoherence, suppressing the global superconducting order.