Lattice studies of entanglement entropy in $O(N)$ models at finite densities

This paper presents a method for calculating entanglement entropy in $O(N)$ models at finite density by integrating the replica trick with worm algorithm Monte Carlo simulations, showcasing initial results for the 3D nonlinear $O(4)$ model.

The Gist

Imagine you have a giant, complex puzzle made of billions of tiny, dancing pieces. In the world of quantum physics, these pieces are particles, and they are constantly interacting with each other. One of the most fascinating things about these particles is entanglement.

Think of entanglement like a magical, invisible thread connecting two particles. Even if you pull them apart to opposite sides of the room, they still "know" what the other is doing. If you spin one, the other spins instantly. This isn't just a party trick; it's a fundamental rule of how the universe works.

This paper is about a team of scientists (Aatu, Niko, and Tobias) who are trying to measure just how strong these invisible threads are in a specific type of quantum system, especially when the system is "crowded" (at a high density of particles).

Here is a breakdown of their journey, explained simply:

1. The Problem: The "Sign Problem"

Usually, when scientists try to simulate these quantum systems on a computer, they run into a massive wall called the "Sign Problem."

• The Analogy: Imagine trying to count the number of people in a room, but half the people are wearing "positive" shirts and the other half are wearing "negative" shirts. When you add them up, the positives and negatives cancel each other out, leaving you with zero or a confusing mess. You can't tell how many people are actually there.
• The Reality: In quantum physics, when you add "density" (more particles), the math gets so complicated that the computer can't figure out which configurations are real and which are just mathematical noise.

2. The Solution: The "Worm" Algorithm

To get around this, the scientists didn't look at the particles directly. Instead, they looked at the flow of energy and charge between them.

• The Analogy: Instead of trying to count every single person in the room, imagine you are a worm crawling through the cracks in the floor. You don't care about the people; you just care about the tunnels they leave behind.
• How it works: They reformulated the math so the computer could "walk" through the system like a worm. This worm leaves a trail, updating the connections between particles step-by-step. This bypasses the "Sign Problem" entirely, allowing them to simulate the crowded system without the math breaking down.

3. The Goal: Measuring the "Entanglement Entropy"

They wanted to measure Entanglement Entropy (EE).

• The Analogy: Imagine you have a long loaf of bread (the whole system). You want to know how "connected" the left half of the loaf is to the right half. If you cut the loaf in half, how much information is lost? If the two halves are totally independent, the connection is weak. If they are deeply entangled, the connection is strong.
• The Challenge: To measure this, you usually have to cut the loaf in different places and compare the results. But in a computer simulation, "cutting" the loaf is a huge, messy operation that requires restarting the whole simulation every time. It's like trying to measure the length of a river by building a new dam every time you want to take a measurement.

4. The Innovation: The "Boundary Deformation"

The team invented a clever trick to measure this without rebuilding the whole simulation.

• The Analogy: Instead of cutting the loaf in a new place, imagine you have a flexible ruler. You gently nudge the ruler one inch to the left, then another inch, then another. You measure the tiny changes as you slide the ruler along.
• The Trick: They developed a method to "nudge" the boundary between the two halves of their system one tiny step at a time. As they slide this boundary, they watch how the "worm" (the flow of particles) reacts. By measuring these tiny nudges, they can calculate the total entanglement without ever having to do a massive, impossible calculation.

5. The Results: What They Found

They tested this on a 3D model (the $O(4)$ model) which is like a simplified version of how certain forces in nature work.

• What they saw: As they increased the "density" (crowdedness) of the system, the entanglement changed in a very specific way.
  • At first, as they added more particles, the entanglement grew.
  • But once the system hit a "critical point" (like water turning into steam), the behavior changed, and the entanglement started to drop.
• The "Aha!" Moment: They also checked their math by comparing two different ways of calculating the same thing. It was like checking your bank account balance by looking at your receipts and your bank statement. Both matched perfectly! This proved their "worm" method and their "nudging" technique were working correctly.

Why Does This Matter?

This isn't just about abstract math. Understanding entanglement in crowded systems helps us understand:

• Superconductors: Materials that conduct electricity with zero resistance.
• Neutron Stars: The incredibly dense cores of dead stars.
• Quantum Computers: How to build machines that use these quantum rules to solve problems.

In a nutshell: These scientists built a new kind of "microscope" (the worm algorithm) and a new way to "measure" (the boundary deformation) to peek inside the quantum world. They successfully measured how tangled particles get when they are packed tightly together, proving their new tools work and opening the door to understanding some of the universe's most extreme environments.

Technical Summary

Here is a detailed technical summary of the paper "Lattice studies of entanglement entropy in $O(N)$ models at finite densities."

1. Problem Statement

The paper addresses the challenge of computing Entanglement Entropy (EE) in interacting Quantum Field Theories (QFTs) at finite chemical potential ($\mu$).

• The Sign Problem: Standard Monte Carlo (MC) simulations of QFTs at finite density suffer from a "sign problem" because the action becomes complex, rendering importance sampling of the original field configurations ($\phi$) impossible.
• Computational Difficulty of EE: EE is defined via the logarithm of the reduced density matrix ($S_{EE} = -\text{tr}(\rho_A \log \rho_A)$), which is notoriously difficult to compute directly. While the replica trick allows the calculation of Rényi entropies ($H_r$) via partition functions, calculating the derivative of EE with respect to the entangling region size ($\ell$) involves comparing partition functions with different topologies ($Z(\ell, r)$ vs. $Z(\ell+1, r)$).
• Sampling Overlap: In lattice simulations, increasing the entangling region size $\ell$ by one unit is a non-local change. This results in minimal overlap between the configuration distributions of $Z(\ell, r)$ and $Z(\ell+1, r)$, causing severe statistical inefficiency in standard importance sampling.

2. Methodology

The authors propose a novel framework combining dual variable reformulation, the worm algorithm, and a modified boundary deformation method.

A. Dual Reformulation and Worm Algorithm

To bypass the sign problem in $O(N)$ models at finite $\mu$:

• The theory is reformulated in terms of integer-valued dual variables (fluxes $k$, neutral pairs $l$, and monomers $p, q, n$).
• The partition function is expressed as a sum over these integer variables subject to local constraints (charge conservation and evenness constraints).
• Worm Algorithm: Instead of local Metropolis updates, which are inefficient due to constraints, the authors employ a worm algorithm. This involves inserting a source-sink pair (a "worm") that moves through the lattice, updating flux variables ($k$ and $\chi^{(i)}$) along its path to satisfy constraints dynamically. This allows for efficient sampling of the constrained dual system.

B. Entanglement Entropy via Boundary Deformation

To calculate the derivative $\partial_\ell S_{EE}$ without the overlap problem:

• Replica Trick: The authors utilize the relation $\text{tr}(\rho_A^r) = Z(\ell, r) / Z^r$. They focus on the second Rényi entropy ($r=2$) to approximate $\partial_\ell S_{EE}$.
• Boundary Deformation: Instead of jumping directly from $\ell$ to $\ell+1$, the boundary between regions $A$ and $B$ is deformed piece-by-piece across individual spatial sites. This connects $Z(\ell, 2)$ and $Z(\ell+1, 2)$ via a sequence of local updates.
• Handling Defects: Changing boundary conditions on a spatial site swaps the endpoints of temporal links. In the dual formulation, this can violate local constraints (creating "defects"). The authors developed two specific procedures to handle this:
  1. Plaquette Worms: Before changing the boundary, constrained worm updates are performed on temporal plaquettes to adjust flux variables ($k$ and $\chi$) so that the boundary change does not create defects. These are reversed after the boundary change to maintain detailed balance.
  2. Defect-Anti-defect Annihilation: Alternatively, if defects are created, they appear as pairs. A worm update is used to move the defect to its anti-defect partner to annihilate them, ensuring the system remains valid.

3. Key Contributions

• Algorithmic Innovation: The first successful application of the boundary deformation method for EE in $O(N)$ models at finite density, specifically adapted to work with the dual worm algorithm.
• Defect Resolution: The development of specific "plaquette worm" and "defect-annihilation" protocols to maintain local charge conservation and evenness constraints during non-trivial boundary deformations.
• Validation via Thermodynamic Relations: The authors derived a theoretical identity linking the mixed derivative of the entropy to the spatial derivative of the charge density:
  $$\frac{\partial^2 H_2}{\partial \mu \partial \ell} = -4 N_t N_s^{d-1} \frac{\partial n(\ell, 2)}{\partial \ell}$$
  They use this relation to validate their algorithm.

4. Results

The authors presented initial results for the nonlinear $O(4)$ model in 3 dimensions ($d=3$) on lattices of size $2 \times N_t \times N_x \times N_s$.

• Dependence on $\ell$: The derivative $\partial_\ell H_2$ decreases rapidly as the entangling width $\ell$ increases, eventually reaching a plateau for $\ell \geq 5$.
• Dependence on $\mu$ and $N_t$:
  • $\partial_\ell H_2$ decreases as the temporal lattice size $N_t$ increases (approaching the continuum limit).
  • The behavior with respect to chemical potential $\mu$ is non-monotonic: $\partial_\ell H_2$ increases with $\mu$ up to a critical value, after which it decreases.
• Algorithm Validation: The authors computed both sides of the derived identity (Eq. 12) numerically. The plots of $\partial_\mu \partial_\ell H_2$ and $-4 N_t N_s^{d-1} \partial_\ell n$ showed excellent agreement, confirming that the algorithm correctly samples the partition function and respects thermodynamic relations.

5. Significance and Outlook

• Phase Transitions: The ability to compute EE at finite density opens new avenues for studying phase transitions in QFTs, where entanglement measures can serve as order parameters.
• Future Directions:
  • Computing critical exponents using EE.
  • Comparing results at large $N$ with $N \to \infty$ analytical calculations.
  • Simulating canonical ensembles to study symmetry-resolved entanglement.
  • Investigating the connection between large-$\ell$ EE and thermal entropy density.
• Impact: This work provides a robust computational tool for exploring quantum many-body systems at finite density, a regime previously inaccessible to standard lattice methods due to the sign problem.