Experimental Observation of Dynamical Phase Transitions in a Dephased Photonic Quantum Walk

This paper experimentally demonstrates both first- and second-order dynamical phase transitions in a dephased photonic quantum walk on a three-node graph, revealing how tunable gauge flux and dephasing control the crossover between detailed-balance and non-detailed-balance regimes while linking Liouvillian spectral topology to relaxation criticality.

The Gist

Imagine a tiny, invisible traveler (a single photon) hopping around a playground with just three swings (nodes). This traveler doesn't just hop randomly; it follows strict rules of quantum mechanics, which means it can be in multiple places at once, like a ghost walking through walls. However, in the real world, things get messy. The environment "eavesdrops" on the traveler, causing it to lose its spooky quantum magic and start acting more like a normal, classical ball bouncing around.

This paper is about watching that traveler and discovering that the way it settles down into a calm, steady state isn't always smooth. Sometimes, it changes its behavior abruptly, like a light switch flipping. Other times, it changes gradually, like a dimmer switch. The scientists found they could control which type of change happens by tweaking two "knobs" on their machine.

Here is a breakdown of their discovery using simple analogies:

The Setup: The Quantum Playground

Think of the experiment as a high-tech game of "musical chairs" played with light.

• The Traveler: A single photon.
• The Playground: A triangle of three spots (nodes).
• The Rules: The photon hops between spots based on a set of quantum instructions.
• The Noise (Dephasing): Imagine someone whispering secrets to the photon, telling it exactly where it is. The more they whisper (higher "dephasing"), the more the photon forgets its quantum superpowers and acts like a normal ball. The scientists could turn this whispering up or down at will.

The Two Types of "Settling Down"

When the game starts, the photon is in a chaotic state. Eventually, it settles into a pattern where it visits all three spots equally. The paper shows that the journey to that calm state can happen in two very different ways, depending on the "knobs" the scientists turned.

1. The "Light Switch" Change (First-Order Transition)

The Scenario: The scientists turned off the "synthetic gauge flux" (a special magnetic-like field they created) and turned up the noise (dephasing).
What Happened: As they adjusted the speed of the photon's hops, the way the system settled down suddenly flipped.
The Analogy: Imagine a group of people trying to find their seats in a theater. At one speed, the people in the front row sit down instantly, while the back row takes a long time. Suddenly, you tweak the speed, and poof—now the back row sits instantly, and the front row takes a long time. It's a sudden, jarring switch. The paper calls this a First-Order Dynamical Phase Transition. It's like a light switch: it's either "on" or "off," with no in-between.

2. The "Dimmer Switch" Change (Second-Order Transition)

The Scenario: The scientists turned on the "synthetic gauge flux" (breaking the symmetry) and kept the noise high.
What Happened: Instead of a sudden flip, the system started to oscillate. The photon didn't just settle; it wobbled back and forth, getting quieter and quieter until it stopped.
The Analogy: Imagine pushing a child on a swing. If you push at the right rhythm, they go higher and higher. If you push at the wrong rhythm, they wobble and slow down. Here, the system started to "wobble" (oscillate) as it settled. As the scientists adjusted the speed, this wobbling behavior grew smoothly and continuously. There was no sudden jump; it was a smooth slide from "no wobble" to "lots of wobble." This is a Second-Order Dynamical Phase Transition. It's like a dimmer switch: you can turn the light up or down smoothly.

The Special "Sweet Spot" (The Exceptional Point)

The most exciting part of the discovery is a specific point where the two types of behavior meet.

• The Analogy: Think of two cars driving on parallel roads. In the "Light Switch" scenario, they just cross paths and keep going. But in the "Dimmer Switch" scenario, at a specific moment, the two cars merge into a single lane, drive together for a split second, and then split apart again.
• The Science: The scientists found a point called an Exceptional Point (EP). At this exact moment, the two different ways the system relaxes (the "modes") merge into one. This is a rare and special state where the rules of the game change fundamentally. They proved that this merging only happens when they broke the symmetry (turned on the gauge flux).

Why Does This Matter?

The paper claims that by using light and a simple three-node setup, they successfully demonstrated that:

1. Open systems (systems that interact with their environment) can have sharp, dramatic changes in how they relax, not just closed, perfect systems.
2. You can control whether the change is sudden (like a switch) or smooth (like a dimmer) just by adjusting a magnetic-like field and the amount of "noise."
3. They tracked this behavior all the way from a very noisy, classical world to a quieter, more quantum world, finding that these special transitions still exist even when the system is mostly quantum, as long as there is a little bit of noise.

In short, they built a tiny, controllable playground for light to show that the way things calm down can be engineered to be either a sudden crash or a smooth glide, depending on how you set the rules.

Technical Summary

Technical Summary: Experimental Observation of Dynamical Phase Transitions in a Dephased Photonic Quantum Walk

Problem Statement
Dynamical phase transitions (DPTs) are well-explored in closed quantum systems, often characterized by non-analyticities in the Loschmidt echo. However, realistic physical systems are inherently open, and extending DPT concepts to open quantum systems coupled to an environment remains challenging. In open systems, the mechanisms governing dynamical critical behavior are poorly understood, particularly regarding how relaxation toward a stationary state is governed by the spectral properties of the dynamics generator. While theoretical works suggest that breaking time-reversal symmetry (TRS) and detailed balance can lead to second-order DPTs associated with exceptional points (EPs), realizing these phenomena in a controlled quantum platform has been difficult. Furthermore, distinguishing between first-order transitions (driven by eigenvalue crossings) and second-order transitions (driven by eigenvalue/eigenvector coalescence) in the presence of decoherence requires a system where both coherent evolution and dissipation can be precisely tuned.

Methodology
The authors experimentally investigate DPTs using a discrete-time open quantum walk (OQW) on a three-node graph implemented in a photonic platform.

• System Architecture: The experiment utilizes a phase-stable single-photon interferometric network. The qutrit basis is encoded in the polarization ($H, V$) and spatial ($U, D$) modes of single photons, representing nodes $|0\rangle, |1\rangle, |2\rangle$.
• Dynamics Model: The system evolves via a discrete-time map $\rho(s+1) = e^{-L}\rho(s)$, where $-L$ is the effective Floquet-Liouvillian superoperator. The evolution consists of a coherent unitary step $U = e^{-iH\tau}$ followed by a dephasing channel.
  • The Hamiltonian $H$ includes hopping amplitudes $J_0, J_1, J_2$ and a synthetic gauge flux $\phi$ threading the triangular loop.
  • Dephasing is introduced via local pure dephasing in the node basis, controlled by a parameter $q \in [0, 1]$. $q=1$ corresponds to the fully dephased (classical Markov) limit, while $0 < q < 1$ represents the partially dephased (quantum) regime.
• Control Parameters: The study varies the coherent step strength $\beta \equiv J_0\tau$ and the synthetic gauge flux $\phi$.
  • $\phi = 0$ preserves TRS and detailed balance.
  • $\phi \neq 0$ breaks TRS and detailed balance.
• Measurement: The researchers reconstruct the quantum channel (transition matrix $Q$ for $q=1$ and the full superoperator $S$ for $q<1$) using quantum process tomography. They extract the eigenvalues (Floquet exponents $\lambda_k$) and eigenvectors (relaxation modes) to analyze the spectral topology and relaxation dynamics.

Key Contributions and Results
The paper reports the first experimental observation of both first-order (FOPT) and second-order (SOPT) dynamical phase transitions in a photonic OQW.

1. First-Order DPT under Time-Reversal Symmetry ($\phi = 0$):

   • In the fully dephased limit ($q=1$), the system obeys detailed balance, resulting in a real spectrum.
   • As $\beta$ is tuned, the authors observe a diabolic crossing of the real Liouvillian eigenvalues ($\text{Re}[\lambda_1] = \text{Re}[\lambda_2]$) at a critical point $\beta_c \approx 0.82$.
   • This crossing causes an abrupt rearrangement of the relaxation modes: the "slow" relaxation branch switches identity, leading to a discontinuous change in the relaxation pattern (e.g., which node relaxes last).
   • An order parameter $m(\beta)$, measuring the weight of the slow mode on a specific node, exhibits a jump $\Delta m \approx 0.66$, confirming the FOPT nature.

2. Second-Order DPT under Broken Time-Reversal Symmetry ($\phi \neq 0$):

   • When TRS is broken ($\phi = \pi/3$), detailed balance is violated, and the spectrum becomes complex.
   • The system exhibits a second-order transition at an exceptional point (EP) where both eigenvalues and eigenvectors coalesce.
   • Below $\beta_c$, relaxation is monotonic with real eigenvalues. Above $\beta_c$, the non-stationary eigenvalues form a complex-conjugate pair, leading to damped oscillatory relaxation.
   • The transition is continuous, characterized by the square-root scaling of the eigenvalue splitting ($\Delta \lambda \propto \sqrt{|\beta - \beta_c|}$) and the eigenvector overlap parameter $g$ approaching unity at the EP.

3. Quantum Regime Validation ($0 < q < 1$):

   • By progressively reducing the dephasing strength ($q=0.8, 0.5$), the authors track the crossover toward the quantum-coherent regime.
   • The spectral signatures of both FOPT (diabolic crossing) and SOPT (EP coalescence) persist down to a finite dephasing threshold.
   • As coherence increases ($q \to 0$), the critical point $\beta_c$ shifts monotonically, and for very weak dephasing ($q=0.1$), the spectral signatures become smeared, indicating the suppression of the transition in the near-unitary limit.
   • Crucially, the authors note that in the continuous-time (Lindbladian) limit, TRS is preserved, and SOPTs driven by EPs cannot arise; thus, the observed SOPTs are intrinsic to the stroboscopic (Floquet) dynamics.

Significance
The authors claim that their results provide a direct experimental link between Liouvillian spectral topology and relaxation criticality in open quantum systems.

• Fundamental Physics: The work demonstrates that sharp dynamical critical behavior can occur in finite open systems, governed by the spectral structure of the Floquet-Liouvillian superoperator rather than the thermodynamic limit. It experimentally validates that breaking detailed balance via synthetic gauge flux enables second-order transitions driven by non-Hermitian degeneracies (EPs), which are forbidden under TRS.
• Platform Utility: The photonic OQW serves as a controllable platform for engineering dissipative dynamics. The ability to tune the order of the transition (FOPT vs. SOPT) via the gauge phase offers a handle for spectrum-guided relaxation engineering.
• Future Directions: The authors suggest this platform enables the exploration of non-Hermitian spectral topology, including bulk-boundary correspondence and dissipative topological phases. They also highlight potential applications in relaxation engineering, such as testing non-Hermitian speedup phenomena (e.g., Mpemba-like effects) and fast state preparation, though these are presented as potential extensions rather than realized outcomes in this specific study.