Measurement-induced phase transition in interacting bosons from most likely quantum trajectory

This paper proposes a new theoretical method based on the most likely quantum trajectory to describe monitored interacting bosonic systems, demonstrating its exactness for Gaussian theories and its ability to reveal an entanglement phase transition from area-law to logarithmic-law scaling in the interacting Sine-Gordon model.

The Gist

The Big Picture: The "Most Likely" Path Through a Storm

Imagine you are trying to predict the path of a leaf floating down a river. But this isn't a normal river; it's a chaotic, stormy one where invisible gusts of wind (representing quantum measurements) keep hitting the leaf, pushing it in random directions.

In the quantum world, when you measure a system, you don't just get a single result; you get a "trajectory"—a specific history of how the system evolved based on that specific sequence of random measurement results. Because the measurements are random, there are billions of possible paths the system could take.

The Problem: To understand the system, physicists usually have to calculate the average of all these billions of paths. It's like trying to predict the weather by simulating every single raindrop in a storm. It's computationally impossible for complex systems.

The Solution: The authors of this paper propose a clever shortcut. Instead of tracking every single possible path, they ask: "What is the single most likely path the leaf would take?"

They developed a method to find this "Most Likely Trajectory." They proved that for simple systems, this single path is actually a perfect representative of the whole storm. For complex, interacting systems, it's an approximation, but a very powerful one that turns a chaotic, random problem into a smooth, predictable one.

---

The Core Concepts

1. The Quantum Trajectory (The Leaf's Journey)

Think of a quantum system as a dancer.

• Unitary Dynamics: The dancer follows a choreographed routine (the laws of physics).
• Measurement: Every few seconds, a spotlight hits the dancer, and a random gust of wind pushes them slightly off-balance.
• The Trajectory: The specific path the dancer takes is a "quantum trajectory." Because the wind is random, the dancer could stumble left, right, or spin.

2. The "Most Likely" Trajectory (The Ideal Dancer)

Usually, to understand the dance, you'd need to watch 1,000 dancers and average their movements. This paper says: "Let's just watch the one dancer who never stumbles."

They mathematically define a "Most Likely Trajectory" where the random wind gusts (fluctuations) are effectively zeroed out. Instead of a chaotic, wobbly path, the system follows a smooth, deterministic path.

• Why it works: In simple systems (like free bosons), this "ideal dancer" moves exactly the same way the average of all 1,000 real dancers would.
• The Magic: This turns a messy, random equation into a clean, solvable one.

---

The Experiment: The Sine-Gordon Model (The Trampoline Park)

To test their method, the authors applied it to a complex system called the Sine-Gordon model. Let's imagine this as a giant trampoline park with many trampolines connected to each other.

• The Setup:

  • The Potential (The Wells): The trampolines have little pits or "wells" in them (caused by the Sine-Gordon interaction). If a ball (the particle) rolls into a well, it gets stuck there. This represents localization (things staying put).
  • The Measurements (The Shakers): Every few seconds, someone shakes the whole trampoline park (momentum measurement). This shaking tends to knock the ball out of the pits. This represents delocalization (things spreading out).

• The Battle:

  • If the pits are deep and the shaking is weak, the ball stays stuck in the pit. The system is "Massive" and "Localized."
  • If the shaking is strong, the ball gets knocked out of the pits and bounces freely across the whole park. The system becomes "Massless" and "Delocalized."

• The Phase Transition:
  The authors used their "Most Likely Trajectory" method to find the exact tipping point where the ball switches from being stuck to bouncing freely. They found a Measurement-Induced Phase Transition (MIPT).

  • Before the transition: The system is quiet, and the "entanglement" (how connected the trampolines are) grows very slowly (Area Law).
  • After the transition: The system is chaotic, and the entanglement grows much faster (Logarithmic Law).

---

Why This Matters (The Takeaway)

1. Solving the Unsolvable: Traditionally, studying how quantum systems behave under constant measurement is a nightmare because of the randomness. This method cuts through the noise by focusing on the "statistical hero"—the most probable path.
2. A New Tool for Complex Systems: They proved this works perfectly for simple systems and used it to discover new physics in complex, interacting systems (like the trampoline park) that were previously too hard to analyze.
3. The "Area" vs. "Log" Law: They showed that by just changing how often you measure the system, you can fundamentally change how the quantum information is shared across the system. It's like turning a quiet library (low entanglement) into a bustling marketplace (high entanglement) just by changing the volume of the background noise.

Summary Analogy

Imagine trying to predict the traffic flow in a city during a massive festival.

• Old Way: Simulate every single car, every red light, and every pedestrian jaywalking. (Too hard, too slow).
• New Way (This Paper): Identify the "Most Likely Route" that the average driver would take if they ignored the rare, crazy detours.
• Result: You get a surprisingly accurate map of the traffic flow without needing a supercomputer to track every single car. And, you discover that if you add enough police cars (measurements) at the intersections, the whole traffic pattern changes from a gridlock to a free-flowing highway.

This paper provides the mathematical blueprint for that "Most Likely Route" in the quantum world.

Technical Summary

1. Problem Statement

The paper addresses the challenge of characterizing Measurement-Induced Phase Transitions (MIPTs) in interacting bosonic many-body systems.

• Context: MIPTs arise from the competition between unitary dynamics (which builds quantum correlations) and quantum measurements (which suppress them).
• The Challenge: Describing MIPTs requires analyzing non-linear functionals of the quantum state (e.g., entanglement entropy) conditioned on specific measurement outcomes (quantum trajectories).
  • Theoretical Difficulty: The stochastic nature of measurements leads to an infinite, unclosed set of coupled equations for interacting systems. Standard approaches like the Replica Trick or full numerical simulations of stochastic trajectories are computationally prohibitive for interacting bosons.
  • Experimental Difficulty: Fully characterizing the state for individual trajectories (post-selection) is exponentially hard.
• Gap: While free bosonic systems have been studied, interacting bosonic systems under continuous monitoring lack robust analytical tools to predict steady-state properties and phase transitions.

2. Methodology: The Most-Likely Trajectory Approach

The authors propose a novel theoretical framework based on the Most-Likely Trajectory (MLT).

• Core Concept: Instead of averaging over the entire ensemble of stochastic trajectories, the method identifies the single trajectory that dominates the joint probability distribution of measurement readouts.
• Derivation:
  • The joint probability distribution $P[r; t]$ of a trajectory $r(t)$ is expressed via a path integral formalism.
  • The authors apply a Saddle-Point approximation to the action $S[r; t] = \log P[r; t]$.
  • The dominant trajectory $r^*(t)$ satisfies $\partial S / \partial r = 0$, which implies $r^*(t) = \langle \hat{R} \rangle_{\rho^*}$, where $\hat{R}$ is the measured operator and $\rho^*$ is the state conditioned on the trajectory.
• Resulting Dynamics:
  • This approach yields a deterministic, non-linear master equation for the density matrix $\rho^*(t)$.
  • Crucially, it sets the stochastic noise term ($dW$) to zero ($dW=0$), effectively filtering out fluctuations while retaining the back-action of the measurement on the state.
  • For Gaussian theories (free bosons), this approximation is exact because the path integral is quadratic, making the saddle point the exact solution.

3. Application to Interacting Systems: SCTDHA

To apply the MLT method to interacting bosons, the authors combine it with the Self-Consistent Time-Dependent Harmonic Approximation (SCTDHA).

• The Model: The monitored Sine-Gordon model, consisting of a free boson chain with a cosine potential ($\cos(\alpha \hat{x}_j)$) and continuous weak momentum measurements.
• The Approximation:
  • Interactions prevent the closure of equations of motion at the level of second moments (variances).
  • SCTDHA maps the interacting Hamiltonian onto a time-dependent quadratic Hamiltonian by expanding the interaction term around the instantaneous mean field and variance.
  • This creates a "dressed" harmonic oscillator with time-dependent effective mass and drive, allowing the use of Wick's theorem to close the equations.
• Synergy:
  • In the standard stochastic approach, the SCTDHA parameters (mass, drive) would become stochastic variables dependent on the trajectory, leading to stochastic correlations that are hard to solve.
  • By using the MLT approach, the state $\rho^*$ is deterministic. Consequently, the SCTDHA parameters become deterministic functions of time. This allows the system of equations to be solved analytically or via simple numerical integration without simulating an ensemble of trajectories.

4. Key Results

A. Benchmark: Free Bosons CFT

• The method was first tested on the Free Bosons Conformal Field Theory (CFT) (the non-interacting limit).
• Result: The MLT approach exactly reproduces the trajectory-averaged results found in previous literature (Ref. [49]).
• Validation: It correctly predicts the absence of a phase transition in the free case, showing logarithmic negativity scaling as $\log N$ (log-law) for all measurement strengths, and power-law decaying correlations.

B. Interacting Bosons: The Sine-Gordon Model

The authors applied the combined MLT + SCTDHA method to the interacting Sine-Gordon model.

• Phase Transition Discovery: They identified a Measurement-Induced Phase Transition driven by the competition between the Sine-Gordon potential (localizing particles in potential wells) and momentum measurements (delocalizing particles).
• Phase Diagram:
  • Massive/Localized Phase (Low $\gamma/J$): The potential dominates. The system exhibits an Area Law scaling for entanglement (logarithmic negativity is constant or scales with boundary size) and exponentially decaying correlations. The effective mass in the SCTDHA is non-zero.
  • Massless/Delocalized Phase (High $\gamma/J$): Strong measurements overcome the potential, preventing localization. The system exhibits Logarithmic Law scaling for entanglement ($\log N$) and power-law decaying correlations. The effective mass vanishes ($h=0$).
• Critical Line: The transition occurs at a critical measurement strength $\gamma_c$ dependent on the interaction parameter $\alpha$. As $\alpha \to \infty$ (free limit), the massive phase shrinks, consistent with the free boson results.

C. Perturbation Theory Validation

• To validate the SCTDHA approximation, the authors performed a perturbative analysis for strong interactions ($J \gg 1$).
• They calculated the relevance of the cosine operator using the steady-state correlations of the free theory.
• Result: The critical line derived from perturbation theory qualitatively matches the critical line found via the MLT+SCTDHA method, confirming the reliability of the approximation in capturing the critical behavior.

5. Significance and Conclusions

• Theoretical Breakthrough: The paper provides the first analytical framework capable of describing MIPTs in interacting bosonic systems without resorting to full stochastic simulations or the replica trick.
• Efficiency: By reducing the problem to a single deterministic trajectory, the method drastically reduces computational complexity, enabling the study of system sizes and dynamics previously inaccessible.
• Physical Insight: The work reveals that continuous weak measurements can drive a delocalization transition in interacting bosons, changing the entanglement structure from area-law to log-law.
• Future Directions: The authors suggest the method could be extended to fermionic systems, spin systems (via bosonization), and other non-Hermitian master equations (e.g., no-click measurements).

In summary, the paper establishes the Most-Likely Trajectory as a powerful, exact tool for Gaussian systems and a highly effective, controlled approximation for interacting systems, successfully predicting a new class of measurement-induced critical phenomena in bosonic matter.