From Ferrimagnetic Insulator to superconducting Luther-Emery Liquid: A DMRG Study of the Two-Leg Lieb Lattice

Using density matrix renormalization group (DMRG) simulations motivated by recent ultracold atom experiments, this study reveals that the two-leg Lieb ladder Hubbard model transitions from a ferrimagnetic Mott insulator at half-filling to a Luttinger liquid at lower fillings, with a distinct superconducting Luther-Emery phase exhibiting dominant $s_{xy}$-wave pairing emerging in a narrow window near the critical filling of $n_c \approx 2/3$.

The Gist

Imagine a bustling city made of tiny, energetic people called electrons. In most cities (materials), these people move freely, creating electricity like a flowing river. But sometimes, if the city gets too crowded or the people start hating each other too much, they stop moving and freeze in place, turning the city into an insulator.

This paper is a story about a very specific, unusual city layout called the Lieb Lattice. Think of it not as a perfect grid like a chessboard, but as a city where every fourth building is missing. This creates a unique pattern of streets and alleys that changes how the residents behave.

The authors, Alexander Nikolaenko and Subir Sachdev, used a powerful computer simulation technique called DMRG (which is like a super-smart detective that can track millions of interactions at once) to figure out what happens in this city under different conditions.

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

1. The Perfectly Crowded City (Half-Filling)

When the city is exactly half-full (one person for every two spots), the residents behave in a very specific way.

• The Rule: A famous mathematician named Lieb predicted that in this specific city layout, the residents would naturally form a Ferrimagnetic state.
• The Analogy: Imagine the residents are split into two groups: "Team Red" and "Team Blue." In a normal city, they might fight evenly. But here, because of the missing buildings, there are more "Blue" spots than "Red" spots. The Blues dominate, but they still argue with their Red neighbors. The result is a city that is mostly "Blue" (magnetic) but has a chaotic, fighting vibe locally. It's a solid, frozen block of magnetic order.

2. Letting Some People Out (Doping)

The researchers then started removing people from the city (lowering the "filling" or density).

• The Surprise: As long as the city is still more than two-thirds full, the "Blue" dominance remains. The city stays magnetic, even though it's less crowded.
• The Transition: Once the city drops below that two-thirds mark, the magnetic order breaks down. The residents stop fighting and start flowing freely again, turning into a Luttinger Liquid.
• The Analogy: Think of this like a traffic jam clearing up. The cars (electrons) are no longer stuck in a gridlock; they are flowing in two distinct lanes: one lane for "Charge" (moving forward) and one lane for "Spin" (spinning in place).

3. The Magic Window: The Superconducting Phase

This is the most exciting part of the story. Right before the magnetic order completely disappears (in a narrow window just above the two-thirds mark), something magical happens.

• The Discovery: The researchers found a Superconducting phase.
• The Analogy: Imagine the residents suddenly deciding to dance in perfect pairs. Instead of moving individually, they link arms and glide across the floor without any friction. This is superconductivity—electricity flowing with zero resistance.
• The "Luther-Emery Liquid": The authors call this state a "Luther-Emery Liquid." Think of it as a dance floor where the "Spin" lane has frozen solid (the dancers are locked in pairs), but the "Charge" lane is still flowing freely. This specific combination allows the pairs to glide effortlessly.

4. The Dance Style (sxy-wave)

The researchers didn't just find that they were dancing; they figured out how they were dancing.

• The Geometry: In this city, the residents live on different types of streets. The strongest dancing partnerships form between specific types of neighbors (the "px" and "py" sites).
• The Shape: The pattern of their dance isn't a simple circle or a straight line. It's a complex, four-leaf clover shape. The authors call this sxy-wave symmetry.
• Why it matters: Usually, when magnets are involved, scientists expect a different dance style (p-wave). Finding this specific "clover" dance suggests that the unique geometry of the city (the missing buildings) is the secret ingredient that makes this superconductivity possible.

Why Should We Care?

This isn't just a math puzzle.

1. Real Experiments: Scientists are currently building these exact "cities" using ultracold atoms in laser traps (optical lattices). This paper gives them a roadmap of what to expect.
2. Superconductors: We want to build better superconductors (materials that conduct electricity perfectly) for things like lossless power grids and super-fast computers. Understanding how to make electrons pair up in these weird, flat-band cities could help us design new materials that superconduct at higher temperatures.
3. The "Flat" Mystery: The city has "flat bands," which means the residents have nowhere to go but to interact with each other. This paper shows that when you force electrons to interact this strongly, you get exotic new states of matter that we are only just beginning to understand.

In a nutshell: The authors simulated a weirdly shaped city of electrons. They found that when you take just the right amount of people out, the city transforms from a frozen, fighting magnetic block into a friction-free superconductor where electrons dance in perfect, clover-shaped pairs. It's a new chapter in understanding how to make electricity flow without losing a single drop of energy.

Technical Summary

1. Problem Statement

The paper investigates the ground-state phase diagram of the Hubbard model on a two-leg Lieb ladder (a quasi-one-dimensional strip of the 2D Lieb lattice). This geometry is motivated by:

• Cuprate Physics: The Lieb lattice mimics the CuO$_2$ planes in high-temperature superconductors (Cu sites on a square lattice, O sites on bonds).
• Ultracold Atom Experiments: Recent experiments (Lebrat et al.) using fermionic $^6$Li atoms in optical Lieb lattices have observed signatures of ferrimagnetism at half-filling and anomalous compressibility away from half-filling.
• Theoretical Gap: While the half-filled case is rigorously understood (Lieb's theorem predicts a ferrimagnetic Mott insulator), the behavior away from half-filling, particularly the interplay between flat bands, strong interactions, and potential superconductivity, remains unclear. Previous Hartree-Fock studies suggested magnetic ordering but failed to capture strong correlation effects near quantum critical points.

2. Methodology

The authors employ Density Matrix Renormalization Group (DMRG) simulations, a state-of-the-art tensor network method for 1D and quasi-1D systems, to overcome the limitations of mean-field theories.

• Model: The Hamiltonian includes nearest-neighbor hopping ($t$) and on-site repulsion ($U$) on a two-leg ladder ($L_y=2$) with $L_x$ unit cells (total sites $N = 3L_x L_y$).
• Parameters:
  • Interaction strength: Primarily studied at strong coupling $U=16$ (in units of $t$).
  • Filling fraction ($n$): Varied from $n=1$ (half-filling) down to $n \approx 0.3$.
• Simulation Types:
  • Finite DMRG: Used for systems up to $L_x=30$ with open boundary conditions (OBC) to calculate total spin and energy gaps. Bond dimensions ($\chi$) up to 4000.
  • Infinite DMRG (iDMRG): Used for periodic boundary conditions (PBC) on unit cells up to $L_x=10$ to extract correlation lengths and central charges. Bond dimensions up to $\chi=8000$.
• Analysis Tools:
  • Calculation of total spin $S^2$ and charge/spin gaps ($\Delta_c, \Delta_s$).
  • Extraction of central charge ($c$) via entanglement entropy scaling.
  • Analysis of correlation lengths ($\xi$) for charge, spin, density, and pairing operators to identify dominant instabilities.
  • Real-space analysis of pairing correlation functions to determine pairing symmetry.

3. Key Contributions and Results

A. Phase Diagram at Strong Coupling ($U=16$)

The study maps out a rich phase diagram as a function of filling $n$:

1. Ferrimagnetic Mott Insulator ($n=1$):

   • Consistent with Lieb's theorem, the system exhibits a total spin $S^2 > 0$ and a charge gap. The ground state is a ferrimagnet with antiferromagnetic local ordering but a net magnetic moment.

2. Ferromagnetic Metallic Phase ($2/3 < n < 1$):

   • Upon doping away from half-filling, the system remains metallic with a finite total spin ($S^2 \neq 0$) down to a critical filling $n_c \approx 2/3$.
   • This phase is linked to the flat bands of the non-interacting spectrum, which enhance magnetic instabilities.

3. Superconducting Luther-Emery Liquid ($0.55 \lesssim n < 2/3$):

   • Discovery: In a narrow window just below the ferromagnetic transition ($n_c = 2/3$), the system enters a phase with a finite spin gap ($\Delta_s \approx 0.03$) but a gapless charge mode.
   • Nature: This is identified as a Luther-Emery liquid.
   • Instability: Analysis of correlation lengths shows that the pairing correlation length ($\xi_p$) diverges and significantly exceeds the density correlation length ($\xi_{ns}$). Specifically, $\xi_p / \xi_c \approx 4.1$ and $\xi_p / \xi_{ns} \approx 2.1$, indicating that superconductivity is the leading instability.
   • Symmetry: Real-space pairing correlations are strongest on the bonds connecting $p_x$ and $p_y$ sites (the "h" and "u" bonds in the unit cell). The correlations satisfy $P_{hh} = P_{uu} = P_{hu}$, suggesting $s_{xy}$-wave pairing symmetry. Spin-triplet correlations decay much faster, ruling out $p$-wave pairing.

4. Paramagnetic Luttinger Liquid ($0 < n < 0.55$):

   • For lower fillings, the system becomes a standard Luttinger liquid with one charge and one spin mode ($C_1S_1$, central charge $c=2$).
   • Commensurate Fillings:
     • At $n=1/2$ and $n=1/6$: The charge mode becomes gapped ($C_0S_1$), forming a Mott insulator with gapless spin excitations.
     • At $n=1/3$: Both charge and spin modes are gapped ($C_0S_0$), resulting in a fully gapped Mott insulator. This is associated with a quadratic band touching point specific to the 2-leg ladder geometry.

B. Critical Behavior

• The spin gap in the superconducting region is intrinsically linked to the strong-interaction regime near the ferromagnetic quantum critical point (QCP).
• As $U$ decreases, the spin gap closes rapidly at $U \lesssim 6$, indicating the superconducting phase is a strong-coupling phenomenon.
• The central charge $c$ exceeds 2 near the critical point $n_c=2/3$, attributed to slow convergence of entanglement entropy due to nearly flat bands.

4. Significance

• Experimental Relevance: The results provide a theoretical framework for interpreting recent ultracold atom experiments on optical Lieb lattices, specifically explaining the doubling of compressibility and the nature of the magnetic-to-metallic transition.
• Mechanism of Superconductivity: The paper identifies a novel route to superconductivity driven by proximity to a ferrimagnetic quantum critical point in a flat-band system. Unlike standard antiferromagnetic QCPs which often favor $d$-wave, or ferromagnetic QCPs favoring $p$-wave, the unique geometry of the Lieb lattice leads to $s_{xy}$-wave pairing.
• Methodological Validation: It demonstrates the necessity of non-perturbative methods (DMRG) over Hartree-Fock for capturing the subtle interplay of flat bands, strong correlations, and competing orders in low-dimensional systems.
• Future Directions: The work suggests that the full 2D Lieb lattice may host similar exotic phases, motivating future studies using Neural Quantum States or larger-scale DMRG to explore the 2D limit.

In summary, the paper establishes that the doped Hubbard model on a Lieb ladder hosts a robust superconducting Luther-Emery liquid with $s_{xy}$-wave symmetry in a narrow doping range near a ferrimagnetic transition, offering a new paradigm for flat-band superconductivity.