Superconductivity and competing orders in honeycomb $t$-$J$ model: interplay of lattice geometry and next-nearest-neighbor hopping

Using large-scale DMRG simulations and slave-boson mean-field theory, this study reveals that next-nearest-neighbor hopping in the doped honeycomb $t$-$J$ model induces a robust $d$-wave superconducting phase coexisting with armchair stripes, while highlighting a significant dependence of competing charge and superconducting orders on cylinder boundary geometry.

The Gist

Imagine you are trying to organize a chaotic dance party in a very specific room. The room is shaped like a honeycomb (think of a beehive), and the dancers are electrons.

In this paper, scientists are trying to figure out how these electrons behave when the music changes speed and the room's shape is tweaked. Specifically, they are looking for a special kind of dance called Superconductivity, where the dancers move in perfect, frictionless unison, allowing electricity to flow with zero resistance.

Here is the story of their discovery, broken down into simple concepts:

1. The Setting: The Honeycomb Dance Floor

Usually, scientists study these electron dances on square grids (like graph paper). But recently, we've discovered materials (like twisted layers of graphene) that look more like honeycombs.

• The Problem: When you add a few extra dancers (doping the system) to this honeycomb floor, they start fighting. Some want to form lines (stripes), while others want to dance in pairs (superconductivity). It's a tug-of-war.
• The New Variable: The researchers introduced a new rule: dancers can now hop not just to their immediate neighbors, but also to the next nearest neighbor. They call this "next-nearest-neighbor hopping" ($t'$). Think of this as giving the dancers a slightly longer stride or a new way to move across the floor.

2. The Experiment: The "Cylinder" Test

To study this, the scientists used a super-powerful computer simulation (called DMRG). However, simulating an infinite 2D floor is too hard for computers. So, they rolled the honeycomb floor into long, narrow tubes (cylinders) to make the math manageable.

Here is the twist: The shape of the tube matters.

• Tube Type A (YC4-0): Imagine a tube where the floor tiles are aligned one way. On this tube, when the researchers increased the "longer stride" ($t'$), something magical happened. The electrons formed a robust superconducting state.
  • They found a "sweet spot" (around $t' \approx 0.4$) where the superconductivity was strongest. It was like finding the perfect tempo where the whole dance floor synchronized perfectly.
  • Even though some dancers still tried to form lines (stripes), the superconducting dance won out.
• Tube Type B (XC8-0): Imagine a tube where the tiles are rotated 90 degrees. On this tube, the same rules applied, but the result was totally different! The electrons refused to superconduct. Instead, they locked themselves into rigid, long lines (stripes) and stopped dancing freely.

The Lesson: The boundary conditions (how you wrap the floor into a tube) acted like a referee, forcing the electrons to choose different strategies. This showed that the "winner" of the competition depends heavily on the geometry of the space.

3. The Theory: The "Mean Field" Crystal Ball

Because the tubes are narrow, the scientists worried: "Is this just an illusion caused by the narrow tube? What happens on a giant, infinite floor?"

To answer this, they used a different method called Slave-Boson Mean-Field Theory (SBMFT). Think of this as a crystal ball that predicts what happens in the real, infinite world.

• The Prediction: The crystal ball confirmed that on a real, infinite honeycomb floor, the "Stripe" pattern (specifically the "armchair" stripes) is very stable.
• The Surprise: However, when the "longer stride" ($t'$) gets very large, the stripes dissolve, and the electrons switch to a uniform, frictionless superconducting dance. This confirmed that the superconductivity seen in the narrow tubes wasn't a fluke; it's a real, stable phase that could exist in the real world.

4. The Big Picture: Why Should We Care?

This paper is a roadmap for finding new superconductors.

• The "Sweet Spot": They found that you don't need exotic quantum states to get superconductivity. You just need the right amount of "next-nearest-neighbor hopping" ($t'$).
• The Connection to Real Life: Since we can now create these honeycomb structures in the lab (using twisted layers of materials like graphene or transition metal dichalcogenides), this research tells experimentalists exactly what to look for. If they tune the material to have that specific "stride length" ($t' \approx 0.4$), they might finally unlock high-temperature superconductivity in these honeycomb materials.

Summary Analogy

Imagine a crowded room where people are trying to either form a conga line (Stripes) or dance in perfect pairs (Superconductivity).

• The scientists found that if you change the music tempo (the hopping parameter $t'$), the room can switch from a conga line to a perfect partner dance.
• They also discovered that if you arrange the furniture (the cylinder geometry) differently, the people might be forced to pick one dance over the other, even if the music is the same.
• By combining the furniture arrangement experiments with a crystal ball prediction, they proved that there is a specific music tempo where the whole room can dance in perfect, frictionless harmony.

The Bottom Line: By tweaking how electrons hop in honeycomb materials, we can likely engineer a new, robust form of superconductivity, potentially leading to better electronics and energy transmission in the future.

Technical Summary

1. Problem Statement

The paper addresses the fundamental challenge of understanding the ground states of doped Mott insulators on honeycomb lattices, a system of high relevance due to its connection to twisted bilayer graphene and other moiré materials. While the undoped honeycomb Hubbard model is known to host Antiferromagnetic (AFM), Quantum Spin Liquid (QSL), and Valence Bond Solid (VBS) phases, the nature of the ground state upon hole doping remains controversial.

Specifically, the authors investigate the role of next-nearest-neighbor (NNN) hopping ($t'$) and the corresponding superexchange coupling ($J' \propto (t')^2$) in the extended $t-J$ model. Key open questions include:

• Does $t'$ enhance or suppress superconductivity (SC) in doped honeycomb systems?
• How does $t'$ mediate the competition between SC and Charge Density Wave (CDW)/stripe orders?
• What is the true 2D thermodynamic limit of these phases, given that finite-size simulations often suffer from geometry-dependent artifacts?

2. Methodology

The authors employ a synergistic approach combining two complementary numerical and theoretical techniques:

• Large-Scale Density Matrix Renormalization Group (DMRG):

  • Model: The $t-t'-J-J'$ model on honeycomb lattices with $t=2$, $J=1$, and $J' = (t'/t)^2 J$.
  • Geometry: Simulations are performed on three distinct cylinder geometries to probe boundary effects:
    • YC4-0: 4 zigzag chains, periodic along $\vec{e}_y$, with bonds lying along $\vec{e}_y$.
    • XC8-0: 8 sites along $\vec{e}_x$ (armchair orientation), periodic along $\vec{e}_x$.
    • YC4-4: 4 zigzag chains with a shift in periodic boundary conditions.
  • Parameters: Doping concentrations $\delta = 1/12$ and $\delta = 1/8$, with $t'$ varying from 0 to 1.1.
  • Technical Details: Up to $m=24,000$ block states are kept, with finite-truncation error extrapolation to ensure convergence.

• Slave-Boson Mean-Field Theory (SBMFT):

  • Used to extrapolate results to the 2D thermodynamic limit, overcoming the width limitations of DMRG.
  • Electrons are decomposed into fermionic spinons ($f$) and bosonic holons ($b$).
  • The Hamiltonian is decoupled in particle-hole and particle-particle channels.
  • The method evaluates the stability of various candidate CDW patterns (a-stripe, z-stripe, bidirectional CDW) and uniform SC states by comparing converged ground-state energies.

3. Key Contributions and Results

A. Robust Superconductivity on YC4-0 Cylinders

On YC4-0 cylinders, the ground state exhibits a robust quasi-long-range nematic $d$-wave superconducting (SC) phase coexisting with sub-leading armchair-oriented stripes (a-stripe).

• SC Correlations: The SC pair-pair correlation function decays as a power law $\Phi(r) \sim r^{-K_{sc}}$, with Luttinger exponent $K_{sc} < 1$, indicating dominant SC order.
• Optimal $t'$: The SC strength exhibits a non-monotonic, "valley-shaped" dependence on $t'$. The SC is maximized (optimal $K_{sc} \approx 0.80$) at $t'_{op} \approx 0.4$.
  • This optimal point corresponds to a $J' \approx 0.04J$, which lies deep within the undoped AFM regime, suggesting that SC is driven by spin fluctuations within the doped AFM state rather than quantum criticality near QSL phases.
• Symmetry: The relative magnitudes of correlations ($\Phi_{aa} \sim 4\Phi_{bb} \sim 4\Phi_{cc}$) and sign changes between bond types confirm nematic $d$-wave symmetry.

B. Geometry-Dependent Competing Orders

The study reveals a striking sensitivity to cylinder boundary conditions, highlighting the challenge of extrapolating finite-width results:

• XC8-0 Cylinders: For $t' > 0.5$, the system stabilizes a long-range zigzag stripe (z-stripe) phase with no superconductivity. The SC correlations decay exponentially ($\xi_{sc} \sim 0.4$).
• YC4-4 Cylinders: At $t'=0$, a coexisting SC and z-stripe phase is found (consistent with prior work), but increasing $t'$ drives the system into phase separation.
• Implication: The "true" ground state is highly sensitive to how the lattice is cut, necessitating the use of SBMFT to resolve the 2D limit.

C. 2D Thermodynamic Limit (SBMFT Results)

The SBMFT calculations clarify the stability of phases in the infinite 2D limit:

• $\delta = 1/12$: The a-stripe pattern is the global ground state across the entire $t'$ range, dominating over z-stripes and bidirectional CDW.
• $\delta = 1/8$: A phase transition occurs. For small $t'$, the a-stripe is dominant. However, for $t' > 0.5$, the system transitions to a uniform nematic $d$-wave SC phase with no charge order.
• Consistency: The SBMFT prediction of SC dominance at large $t'$ for $\delta=1/8$ aligns with the DMRG observation on YC4-0 cylinders where CDW correlations are significantly suppressed while SC remains dominant.

4. Significance and Conclusion

• Mechanism of SC: The work provides strong evidence that NNN hopping ($t'$) can significantly enhance superconductivity in doped honeycomb lattices. The optimal SC occurs at $t' \approx 0.4$, driven by AFM spin fluctuations rather than proximity to a QSL phase.
• Role of Geometry: The paper demonstrates that finite-size cylinder geometry acts as a "tuning parameter" that can stabilize distinct competing orders (SC vs. Stripes). This underscores the necessity of combining DMRG with mean-field theory to accurately map 2D phase diagrams.
• Experimental Relevance: Given the tunability of $t'$ and $J'$ in moiré materials (e.g., twisted bilayer graphene, twisted TMDs), these results suggest that unconventional nematic $d$-wave superconductivity is a robust candidate for the ground state in doped honeycomb systems, particularly in regimes where $t' \sim 0.4t$.

In summary, the paper resolves the competition between SC and CDW in the honeycomb $t-J$ model, identifying a robust, $t'$-enhanced nematic $d$-wave superconducting phase that persists in the 2D limit, particularly at higher doping levels.