🔬 mesoscale physics

Coexistence of insulating phases in confined fermionic chains with a Wannier-Stark potential

Using density matrix renormalization-group simulations, this study reveals that repulsively interacting fermions in a confined chain with a Wannier-Stark potential exhibit a ground state characterized by a staircase of coexisting insulating phases, including charge density waves, band insulators, and Mott insulators, separated by incompressible incommensurate regions.

Original authors: N. Aucar Boidi, K. Hallberg, A. Aharony, O. Entin-Wohlman

Published 2026-08-04

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: N. Aucar Boidi, K. Hallberg, A. Aharony, O. Entin-Wohlman

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a world where tiny particles, like electrons, don't just bounce around randomly but dance to the rhythm of their neighbors. This is the realm of quantum physics, specifically the study of "many-body systems," where the behavior of a crowd of particles is far more complex than the sum of its parts. To understand this, scientists often use a simplified mental model called the "Hubbard model." Think of it like a game of musical chairs on a grid: particles (the players) can hop from one chair (a lattice site) to the next, but if two players try to sit in the same chair, they get annoyed and push each other away (repulsion). Sometimes, they also get annoyed if they sit in chairs right next to each other.

Now, imagine you tilt the entire game board. Instead of a flat floor, the chairs are arranged on a ramp. This creates a "Wannier-Stark potential," a fancy term for a linear slope that makes it harder for particles to hop uphill and easier to slide downhill. Why does this matter? Because in the real world, materials can act like perfect insulators (blocking electricity) or conductors (letting it flow), and sometimes they do both in strange, layered ways. By studying how these particles arrange themselves on a tilted, finite track, scientists hope to understand how to control these phases, which is crucial for building better electronics or simulating new materials using cold atoms in a lab.

In this study, the authors N. Aucar Boidi, K. Hallberg, Amnon Aharony, and Ora Entin-Wohlman decided to play with this tilted game board using a powerful computer simulation technique called the density matrix renormalization group (DMRG). They set up a finite chain of 41 sites (an odd number to keep things symmetrical) and filled it with fermions (a type of particle like an electron) that repel each other. They applied a linear potential, effectively creating a ramp where the energy of each site changes steadily from one end of the chain to the other.

What they found was a fascinating "staircase" of phases. Instead of the particles smoothly sliding from a crowded, full state on one side to an empty state on the other, they got stuck in distinct, locked-in patterns. It's as if the particles formed a series of terraced gardens along the slope. On the far left, where the potential is low, the sites are completely full (two particles per site). On the far right, where the potential is high, the sites are completely empty. But in the middle, instead of a smooth transition, the system organizes itself into specific, repeating blocks.

Depending on how strongly the particles repel their neighbors (a parameter they call $V$ ), these blocks take on different shapes. Sometimes, they form a "Mott insulator" where every single site holds exactly one particle. Other times, they form "charge density waves" (CDW), where particles arrange themselves in a pattern like "full-empty-full-empty" (202020...) or "full-half-full-half" (212121...). These are the "plateaus" the authors talk about—regions where the local density of particles stays roughly constant, like a flat step on a staircase, even though the overall slope of the chain is changing.

Crucially, the paper suggests that between these flat, locked-in steps, there are "domain walls." These are regions where the density changes more smoothly, but they aren't just messy transitions. The authors' simulations show that even in these in-between zones, the particles are still stuck in an insulating state, unable to flow freely. They call these "incompressible incommensurate-filling phases" (IIF). It's a bit like a traffic jam where the cars are moving slowly but are still packed so tightly they can't change lanes. The authors found that these IIF regions have tiny energy gaps, meaning they are still insulators, just with a different kind of order than the neat steps next to them.

The team also discovered that they could control this entire landscape. By changing the steepness of the slope (the parameter $\mu_0$ ) or the strength of the repulsion ( $V$ ), they could make the "Mott" region grow or shrink, or make the "charge density wave" patterns take over the whole chain. For instance, when the repulsion between neighbors was strong enough ( $V=6$ ), the central Mott phase disappeared entirely, replaced by a half-filled CDW pattern that stretched across the chain. They also noted that if they tilted the slope too steeply, the interesting middle phases would vanish, leaving only a fully filled side and a fully empty side.

The authors suggest that these findings could be observed in experiments with cold atoms trapped in optical lattices, where scientists can actually create these linear potentials and tweak the interactions. While the paper doesn't claim to have built a new device, it suggests that by tuning the slope and the interactions, experimentalists could create a "menu" of different insulating phases coexisting on a single chip. The results are based on high-precision numerical simulations, which the authors trust to be accurate, but they emphasize that these are theoretical predictions waiting for experimental verification. The key takeaway is that a simple linear slope can force a quantum system to organize itself into a complex, multi-layered structure, revealing a rich coexistence of different insulating states that wouldn't exist on a flat surface.

Technical Summary: Coexistence of Insulating Phases in Confined Fermionic Chains with a Wannier-Stark Potential

Problem Statement
The paper investigates the ground-state properties of a one-dimensional (1D) fermionic system confined by a linearly varying potential (Wannier-Stark potential) and subject to repulsive interactions. While the interplay of strong interactions and quantum statistics in 1D systems is known to generate diverse phases—such as Mott insulators, charge density waves (CDW), and metallic states—the specific behavior of these phases in finite, confined geometries with spatially varying potentials remains less explored. Previous studies on uniform extended Hubbard models have identified various insulating and metallic phases, and experiments with cold atoms have observed some of these. However, the structure of phases near the "edges" of confined samples, particularly when confinement is achieved via site-energy variation rather than hard-wall boundaries, requires further theoretical clarification. The authors aim to generalize the extended Hubbard model to a finite 1D chain under a linear potential to understand how different insulating phases coexist and transition within a single system.

Methodology
The authors study a generalized 1D extended Hubbard Hamiltonian defined on a finite chain of $L$ sites. The Hamiltonian includes:

Nearest-neighbor (nn) hopping ( $t$ ).
On-site repulsive interaction ( $U$ ).
Nearest-neighbor repulsive interaction ( $V$ ).
A site-dependent chemical potential $\mu_i$ representing the linear Wannier-Stark potential, defined as $\mu_i = \mu_0[i/i_c - 1]$ , where $i_c$ is the center of the chain and $\mu_0$ controls the slope.

To solve this many-body problem, the authors employ the Density Matrix Renormalization Group (DMRG) technique, a highly accurate numerical method for correlated 1D systems. Calculations were performed with approximately $m=500$ states and 4–6 finite-size sweeps, achieving an energy precision of $\sim 10^{-10}$ . The study focuses on the local average site density $\langle n_i \rangle$ , nn density-density correlations $\langle n_i n_{i+1} \rangle$ , nn spin-spin correlations $\langle s^z_i s^z_{i+1} \rangle$ , and the local density of states (LDOS).

Key Results
The numerical results reveal a "staircase" of (quasi) plateaus in the local site density $\langle n_i \rangle$ along the chain, decreasing from doubly-filled ( $n=2$ ) to empty ( $n=0$ ) as the potential increases. The system exhibits a simultaneous coexistence of distinct insulating phases separated by incompressible regions with incommensurate fillings.

Phase Coexistence:
- Band Insulators (BI): At the extremes of the chain (large negative or positive potential), the system forms completely filled ( $n=2$ ) or empty ( $n=0$ ) regions.
- Mott Insulators (MI): For intermediate potentials and specific interaction strengths (e.g., $V=0, U=10$ ), a central region with one particle per site ( $n=1$ ) emerges, characterized by antiferromagnetic spin correlations.
- Charge Density Waves (CDW): As the nearest-neighbor interaction $V$ $V$ increases, CDW phases appear.
  - For $V=3$ , "incipient" or doped CDW phases with local mean fillings around $\langle n_i \rangle \approx 1.5$ and $\approx 0.5$ coexist with the Mott phase. These exhibit a structure of $2121\dots$ or $1010\dots$ superimposed on a linear background.
  - For $V=6$ (where $V > U/2$ ), the Mott phase is replaced by a half-filled CDW ( $2020\dots$ ) in the center, coexisting with doped CDW phases at the edges. This specific coexistence of two different CDW types within a single chain is a novel finding.
Incompressible Incommensurate-Filling (IIF) Phases:
The transitions between these commensurate phases are not smooth metallic regions but are separated by "domain walls" with incommensurate fillings. Crucially, the LDOS calculations show that these intermediate regions possess small but finite energy gaps at the Fermi energy. This indicates that these incommensurate regions are incompressible (insulating), contrary to the compressible regions often found in mean-field approximations of 2D systems. The authors term these "incompressible incommensurate-filling phases" (IIF).
Dependence on Parameters:
- Potential Slope ( $\mu_0$ ): As the slope $\mu_0$ decreases, the central Mott or CDW regions expand, while the band insulating regions at the edges shrink. For very steep potentials, only the band insulator phases remain.
- Interaction Strength ( $V$ ): Increasing $V$ drives transitions from Mott to CDW phases and alters the width of the IIF regions.
- Chemical Potential ( $\mu$ ): The sequence of phases is robust against changes in the global chemical potential, though the entire profile shifts spatially along the chain.

Significance and Claims
The paper claims to demonstrate that a linear potential in a finite 1D system induces a spatial coexistence of various insulating phases (Band, Mott, and CDW) that have previously been studied individually in uniform systems. A primary contribution is the identification of incompressible incommensurate-filling phases acting as domain walls between these locked-in commensurate structures. The authors suggest that these results, particularly the coexistence of different CDW orders and the nature of the IIF regions, could be observed experimentally in cold-atom arrays subjected to longitudinal electric fields. The work highlights that the specific sequence and size of these phases can be tuned experimentally by varying the slope of the potential and the range of repulsive interactions. The authors modestly note that further research is required to determine if these incompressible walls are related to Stark many-body localization.

More like this