Arrow of Time as an indicator of Measurement-Induced Phase Transitions

1. Problem Statement

Measurement-Induced Phase Transitions (MIPTs) occur in monitored quantum many-body systems due to the competition between unitary dynamics (scrambling information) and quantum measurements (collapsing the state).

• Current State of the Art: MIPTs are traditionally diagnosed using entanglement-based measures (e.g., entanglement entropy), which transition from a volume law (weak monitoring) to an area law (strong monitoring).
• The Gap: These entanglement measures are non-local and require post-selection over exponentially many quantum trajectories to compute average properties. Furthermore, MIPTs lack a conventional thermodynamic order parameter.
• The Question: Can the Arrow of Time (AoT)—a thermodynamic quantity quantifying the irreversibility of a process—serve as a novel, local diagnostic for MIPTs? Specifically, does the AoT exhibit critical behavior at the transition point?

2. Methodology

The authors develop a thermodynamic perspective on MIPTs by analyzing the AoT across different regimes:

• Definition of AoT: The AoT ($Q$) for a specific quantum trajectory is defined as the logarithmic ratio of the probability of the forward trajectory ($p_m$) to the probability of the time-reversed backward trajectory ($\tilde{p}_m$):
  $$Q_m = \log \frac{p_m}{\tilde{p}_m}$$
  Crucially, to ensure finite and well-defined AoT, the authors utilize invertible (non-projective) measurements. Projective measurements are fully irreversible (making the backward probability zero), whereas invertible measurements allow for a finite backward probability.

• Three Analytical Frameworks:

  1. No-Click Limit (Single Trajectory): The authors analyze the "no-click" trajectory (where no detection events occur) in a monitored Ising chain. This dynamics is governed by an effective non-Hermitian Hamiltonian. They relate the AoT to the steady-state expectation value of the local measurement operator.
  2. Continuous Measurements (Stochastic Approach): They generalize the analysis to systems with continuous monitoring described by Stochastic Schrödinger Equations (SSE). They derive expressions for the mean and variance of the AoT, linking them to non-linear correlation functions of the local measured observable ($\langle \sigma_z^2 \rangle - \langle \sigma_z \rangle^2$).
  3. Random Quantum Circuits (Exact Solvability): To rigorously prove critical behavior, they study random quantum circuits with Haar-random unitary gates interrupted by invertible measurements.
     • They map the problem to a classical statistical mechanics model using the replica trick ($n \to 1$ limit).
     • They replace projective measurements with invertible Kraus operators parameterized by $\alpha$, allowing for exact analytic solutions in the large local Hilbert space dimension ($q \to \infty$) limit.

3. Key Contributions

• Local vs. Non-Local Diagnostic: The paper establishes that while entanglement entropy is a non-local quantity, the AoT is associated with a local operator (specifically, the expectation value of the measured observable).
• Non-Linearity: The average AoT is shown to be a non-linear functional of the averaged density matrix. It depends on the variance of local observables, distinguishing it from linear response functions.
• Exact Mapping to Percolation: By introducing invertible measurements in random circuits, the authors successfully map the calculation of the average AoT to a bond percolation problem on a 2D square lattice. This allows for the exact determination of critical exponents.
• Thermodynamic Nature: The AoT is identified as a bulk thermodynamic quantity (related to the free energy of the theory) rather than a boundary or surface phenomenon (like entanglement entropy).

4. Key Results

• Single Qubit & No-Click Dynamics:

  • In a single qubit under continuous monitoring, the AoT is proportional to the excited-state population.
  • In the monitored Ising chain (no-click limit), the AoT per spin exhibits a non-analytic behavior (a kink in the derivative) at the critical measurement rate ($g_c = 1$), coinciding with the known entanglement transition.

• Continuous Measurements (Many-Body):

  • The average AoT is proportional to the equal-time, equal-position connected correlation function: $Q \propto \langle \sigma_z^2 \rangle - \langle \sigma_z \rangle^2$.
  • While numerical results for the mean AoT show a smooth crossover, the variance and higher derivatives of the AoT are predicted to reveal the critical behavior, as they probe correlations at distinct space-time points.

• Random Circuits & Critical Exponents:

  • In the limit of large local dimension ($q \to \infty$), the average AoT maps exactly to the average cluster size in a bond percolation model.
  • The transition occurs at a critical measurement strength $\alpha_c = \pi/4$ (corresponding to bond probability $p_c = 1/2$).
  • The AoT exhibits non-analytic behavior characterized by the specific heat critical exponent $\alpha = -2/3$ (in 2D percolation universality class).
  • Specifically, the singular part of the AoT scales as $|\alpha - \alpha_c|^{8/3}$. The third derivative of the AoT with respect to the measurement strength diverges at the transition.

5. Significance and Implications

• New Diagnostic Tool: The AoT provides a robust, alternative indicator for MIPTs that does not rely on entanglement entropy. It is particularly useful because it is a local observable, making it potentially more accessible in experiments where global entanglement is hard to measure.
• Thermodynamic Perspective: The work reframes MIPTs not just as information-theoretic transitions (scrambling vs. localization) but as thermodynamic phase transitions. The AoT probes the bulk free energy, whereas entanglement probes boundary conditions.
• Experimental Feasibility:
  • Unlike entanglement-based probes which suffer from severe post-selection overhead (requiring the preparation of identical states many times to estimate averages), the AoT can be extracted directly from the entropy of measurement outcomes (Shannon entropy of the trajectory record).
  • The paper suggests that while the mean AoT might show a crossover, the critical behavior is visible in the derivatives of the AoT with respect to the measurement rate, offering a concrete experimental signature.
• Broader Applications: The framework extends to various non-projective measurement phenomena, including the stabilization of topological states, non-unitary quantum computing, and entropy production in mesoscopic systems (e.g., Kondo effect in quantum dots).

In summary, this paper successfully bridges stochastic thermodynamics and quantum information theory, demonstrating that the irreversibility of quantum measurements (the Arrow of Time) undergoes a sharp, critical transition at the same point as the entanglement transition, governed by the universal laws of percolation.