⚛️ high-energy theory

Universal Statistics of Measurement-Induced Entanglement in Tomonaga-Luttinger liquids

This paper employs conformal field theory and a replica trick to derive closed-form expressions for the statistics of measurement-induced entanglement in one-dimensional Tomonaga-Luttinger liquids, revealing distinctive critical behavior, fat-tailed bimodal distributions, and an equivalence between microscopic Born averaging and conformal boundary condition averaging at low energies.

Original authors: Kabir Khanna, Romain Vasseur

Published 2026-06-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Kabir Khanna, Romain Vasseur

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

The Big Picture: Measuring a Quantum Soup

Imagine you have a giant, bubbling pot of "quantum soup" (a one-dimensional quantum system). This soup is in a special, critical state where everything is deeply connected to everything else, like a massive, invisible web of relationships. In physics, we call this a Tomonaga-Luttinger liquid (TLL).

Usually, when you measure a quantum system, you break these connections. It's like poking a hole in a balloon; the structure collapses. However, this paper explores a very specific and counter-intuitive scenario: What happens if you measure most of the soup, but leave two small islands untouched?

The researchers wanted to know: Does the act of measuring the middle part destroy the connection between the two islands, or does it somehow create a new, strong connection between them?

The Problem: The "Average" Lie

In quantum mechanics, every time you measure something, you get a random result (like flipping a coin). If you measure a huge system, you get a massive string of random results (a "bit string").

The Old Way: Scientists often looked at the "average" result. They would say, "On average, the connection between the islands is X."
The Reality: The paper argues that this average is misleading. Because the results are random, the "average" hides the true story. Sometimes, a specific random outcome creates a huge amount of connection (entanglement). Other times, it creates almost none.

The authors wanted to know the full story: What is the probability of getting a strong connection? A weak one? Is the distribution of these connections a smooth hill, or does it have wild spikes?

The Method: The "Replica Trick" and the "Boundary" Analogy

To solve this, the authors used a mathematical tool called the replica trick. Imagine you have a single quantum system, but you make $N$ identical copies of it (replicas) to do your math. By looking at how these copies interact, you can figure out the statistics of the randomness without having to simulate every single random outcome (which would take forever).

They made a brilliant discovery to simplify the math:
Instead of thinking about the messy, random results of measuring individual atoms on a lattice (like a grid of Lego bricks), they realized that at low energies, all these random outcomes behave exactly like setting different "boundary conditions" on a smooth, continuous field.

The Analogy:
Imagine the quantum system is a drumhead.

The Lattice View: You are pinching the drumhead at thousands of random points with random forces.
The Paper's View: You don't need to track every pinch. You can just imagine the edge of the drumhead is held down at different heights.
The Result: The authors proved that calculating the "average" over all the random pinches is mathematically identical to calculating the average over all possible ways to hold down the edge of the drum. This turned a nightmare of randomness into a clean, solvable geometry problem.

The Key Findings

1. The Distribution is "Fat-Tailed" and Bimodal

When they calculated the full distribution of entanglement (how likely you are to get a certain amount of connection), they found it wasn't a normal bell curve.

Bimodal: The graph had two distinct peaks. One peak was at very low entanglement, and the other was at very high entanglement.
Fat Tails: The "tails" of the distribution were heavy. This means that while rare, outcomes with massive entanglement happen much more often than you would expect in a normal system.

2. The "Quantum Wire" Effect

In the limit where the two unmeasured islands are as far apart as possible, the researchers found something surprising.
Even though the islands are far apart, there is a tiny, non-zero probability that the measurement in the middle will "snap" the system into a state where the two islands are perfectly entangled (like a Bell pair or an EPR pair).

The Metaphor: It's as if the critical state acts like a quantum wire. Even if you cut the wire in the middle and measure the cut, there is a small chance the measurement "re-wires" the ends so they are instantly connected again.
The probability of this happening is small, but it is finite. This is what drives the "fat tails" in the distribution.

3. The "Typical" vs. The "Average"

The paper highlights a massive difference between the average entanglement and the typical entanglement.

Average: Because of those rare, high-entanglement events (the "fat tails"), the mathematical average is pulled up high.
Typical: If you actually performed the experiment once, you would likely get a result near the lower peak, not the high average.
Takeaway: The "average" is a bad predictor of what you will actually see in a single experiment. The system is dominated by rare, dramatic events.

4. Disorder vs. Measurement

They also compared this to "disorder" (random impurities in the material). They found that if you treat the measurement outcomes as simple random noise (ignoring the quantum probabilities, or "Born weights"), the results are completely different. This proves that the unique physics of this system comes specifically from the quantum nature of the measurement probabilities, not just random noise.

Summary in a Sentence

The paper shows that when you measure a critical quantum system, the randomness of the results doesn't just create noise; it creates a wild statistical landscape where rare, "miraculous" events can suddenly link distant parts of the system together, a phenomenon that simple averages completely miss.

Technical Summary: Universal Statistics of Measurement-Induced Entanglement in Tomonaga-Luttinger Liquids

Problem Statement
The paper investigates the statistics of Measurement-Induced Entanglement (MIE) in one-dimensional quantum critical systems described by Tomonaga-Luttinger liquids (TLLs). MIE is defined as the entanglement entropy of unmeasured regions, averaged over all possible projective measurement outcomes on the remainder of the system, weighted by their Born probabilities. While previous works have established that MIE exhibits conformal invariance and universal scaling in the limit of maximal separation between unmeasured regions, a rigorous analytical understanding of the full statistical distribution of MIE—beyond just the mean—has been elusive. Theoretical challenges arise because computing Born-averaged observables requires averaging over an exponentially large space of measurement records, a task that typically necessitates the replica trick and is difficult to perform exactly for non-linear functionals of the density matrix. Furthermore, it remains unclear whether the "typical" behavior of MIE (dominated by rare events) differs significantly from the average, and how the full probability distribution behaves in the critical regime.

Methodology
The authors employ a combination of Conformal Field Theory (CFT) and the replica trick to derive exact analytical expressions for the cumulants and the full distribution of MIE. The methodology proceeds as follows:

Setup and Field Theory Mapping: The system is modeled as a spin-1/2 XXZ chain (or its continuum limit, a compact free-boson CFT). Projective measurements of the charge density ( $\sigma^z$ ) on a disjoint region $B$ are modeled in the continuum as imposing inhomogeneous Dirichlet boundary conditions on the bosonic field $\phi(x)$ .
Replica Trick for Cumulants: To compute the $l$ -th cumulant of the MIE, the authors introduce a generalized replica partition function $Z_A(k_1, k_2) = (\text{tr}\rho_{m,A}^n)^{k_1} (\text{tr}\rho_m)^{k_2}$ . This allows them to express the cumulants as derivatives of the logarithm of this partition function in the replica limit $k \to 0$ .
Conformal Mapping and Quantum-Classical Split: The geometry of the $n$ -sheeted Riemann surface (arising from the replica construction) is mapped to a finite cylinder. The partition function is decomposed into a "quantum" fluctuation part (independent of measurement outcomes) and a "classical" winding part.
Averaging over Boundary Conditions: A crucial step involves summing over all possible measurement outcomes (random boundary profiles). The authors demonstrate that, at low energies, this Born-average over microscopic lattice outcomes is equivalent to an average over conformal boundary conditions (specifically, uniform Dirichlet conditions labeled by a phase $\delta\phi$ ), weighted by the partition function of the system with those boundary conditions.
Analytic Continuation: The resulting winding contribution to the partition function is expressed as a sum over winding numbers. The authors perform an analytic continuation of this sum to non-integer replica indices to extract the cumulants and the full probability distribution.

Key Contributions and Results

Equivalence of Averaging Schemes: The paper establishes a rigorous connection showing that Born-averaging over microscopic measurement outcomes in TLLs is equivalent to averaging over conformal boundary conditions weighted by their respective partition functions. This allows the MIE and its cumulants to be expressed as averages of "forced" MIE (MIE conditioned on a fixed outcome) over a specific probability distribution $p(\delta\phi)$ .
Universal Scaling of Cumulants: In the limit of maximal separation between unmeasured regions (cross-ratio $\zeta \to 0$ $ζ \to 0$ ), the authors derive closed-form scaling expressions for the cumulants $\kappa_l$ $κ_{l}$ .
- For large Rényi indices ( $n > 1/2l$ ), all cumulants exhibit a universal scaling behavior: $\kappa_l \sim \zeta^{g/2} / \sqrt{\log(1/\zeta)}$ , where $g$ is the Luttinger parameter. This scaling is identical across all cumulants for sufficiently large $n$ .
- For smaller $n$ , the scaling changes, exhibiting different power laws dependent on $n$ and $l$ .
- Notably, the scaling includes a $\sqrt{\log(1/\zeta)}$ factor, breaking the usual power-law scaling typical in CFTs.
Full Distribution and Bimodality: The authors derive the full probability distribution $P(S_m)$ $P (S_{m})$ of the post-measurement entanglement. They find that the distribution is generically bimodal with heavy tails at both extremes ( $S_m \to 0$ $S_{m} \to 0$ and $S_m \to \log 2$ $S_{m} \to lo g 2$ ).
- The right tail (near $S_m = \log 2$ ) corresponds to rare measurement outcomes that generate a maximally entangled Bell pair (EPR pair) between the separated regions.
- The probability weight of this high-entanglement tail vanishes as $\zeta \to 0$ (scaling as $\zeta^{g/2}/\sqrt{\log(1/\zeta)}$ ), yet remains finite. This indicates that the critical state acts as a "quantum wire" for specific, rare outcomes.
- The leading scaling of the cumulants in the $\zeta \to 0$ limit is dominated by this rare, high-entanglement tail, explaining why the mean MIE differs significantly from the "typical" MIE (which scales as $\zeta^{2g}$ ).
Disorder Induced Entanglement (DIE): As a corollary, the authors analyze entanglement induced by uncorrelated quenched disorder (uniform averaging over outcomes rather than Born-weighted). They find that DIE scales as $1/\log(1/\zeta)$ , independent of the Rényi index and Luttinger parameter, highlighting a stark contrast with the Born-averaged MIE.
Numerical Verification: The analytical predictions are corroborated by numerical calculations using Matrix Product States (MPS) for the XX chain and Density Matrix Renormalization Group (DMRG) for the interacting XXZ chain. The numerical results show excellent agreement with the theoretical predictions for the second and third cumulants and the full distribution.

Significance
The paper provides one of the first complete analytical descriptions of the statistical properties of measurement-induced entanglement in quantum critical systems. By moving beyond simple averages to the full distribution and higher cumulants, the work reveals that the "typical" behavior of MIE is not representative of the average, which is instead dominated by rare events. The finding that Born-averaging is equivalent to averaging over conformal boundary conditions simplifies the theoretical treatment of measurement-induced phenomena in CFTs. Furthermore, the identification of a heavy-tailed distribution with a vanishing but finite probability of generating maximal entanglement offers a deeper understanding of how measurements reshape the entanglement structure of critical states, suggesting their utility as quantum wires for specific outcomes. The results are benchmarked against lattice models, confirming the universality of these effects in the low-energy limit.