Experimental violation of a Bell-like inequality for causal order

This paper reports the first experimental violation of a Bell-like inequality for causal order involving four parties with simulated spacelike separation, achieving a 5.7-sigma result that certifies indefinite causal order under conditions that exclude bidirectional signaling, though the certification remains device-dependent.

The Gist

The Big Idea: A Game Where "First" and "Second" Don't Exist

Imagine you are playing a game with two friends, Alice 1 and Alice 2. Usually, in our daily lives, things happen in a strict order: you put on your left shoe, then your right shoe. Or you send a text, and then the other person receives it. This is called a definite causal order.

However, quantum mechanics (the physics of the very small) suggests that sometimes, two things can happen in a "superposition" of orders. It's as if Alice 1 and Alice 2 are both doing their tasks at the exact same time, and it's impossible to say who went first. This is called indefinite causal order.

For a long time, scientists could only theorize about this. They had a mathematical rule (an inequality) that said: "If the world works in a normal, definite order, the results of this game must add up to less than a certain number." If the results went over that number, it would prove that the order of events was truly indefinite.

The problem? Building a machine to test this is incredibly hard. It requires perfect timing, perfect light, and a setup where one person is so far away that they can't possibly send a signal to the others in time to cheat.

What this paper did:
A team of researchers built a complex machine using light (photons) to play this game. They successfully broke the mathematical rule by a significant margin, proving that in their experiment, the events did not happen in a fixed "first, then second" order.

---

The Characters and the Setup

To understand the experiment, let's meet the four players:

1. Alice 1 and Alice 2: They are the "doers." They are inside a special machine called a Quantum Switch. Their job is to perform operations on a photon (a particle of light).
2. Bob: He is the "remote observer." He is located 3 kilometers away from the switch.
3. Charlie: He is the "judge." He is near the switch and checks the final result.

The Goal:
Bob and Charlie want to see if Alice 1 and Alice 2 are acting in a fixed order (Alice 1 then Alice 2, OR Alice 2 then Alice 1) or in a fuzzy, indefinite order (both at once).

The Analogy: The "Magic" Train Station

Imagine a train station with two tracks (Track A and Track B) and a magical switch that controls which track a train takes.

• The Control: In this experiment, the "switch" is a photon's polarization (the direction its light waves are vibrating).
• The Train: The "train" is another photon carrying information, encoded in time (arriving early or arriving late).

How the Quantum Switch works:

• If the control photon is vibrating Horizontally, the train goes down Track A: It passes Alice 1 first, then Alice 2.
• If the control photon is vibrating Vertically, the train goes down Track B: It passes Alice 2 first, then Alice 1.

The Magic Trick:
The researchers prepared the control photon in a special state where it is vibrating both horizontally and vertically at the same time. This means the train is effectively traveling down both tracks simultaneously. The photon interacts with Alice 1 and Alice 2 in a superposition of "Alice 1 first" and "Alice 2 first."

The Challenge: The "3-Kilometer" Test

To prove this isn't just a trick where Alice 1 whispers to Alice 2 to coordinate their moves, they had to ensure spacelike separation.

Think of it like this: If Alice 1 and Alice 2 are in the same room, they could easily talk to each other. But if Alice 1 is in New York and Alice 2 is in London, and they have to make a decision in the blink of an eye, they can't possibly communicate fast enough (since nothing travels faster than light).

• The Setup: The researchers put Bob 3 kilometers away. They used long fiber optic cables to simulate this distance.
• The Speed: They had to perform the operations on the light particles incredibly fast (in nanoseconds).
• The Result: Because the operations were so fast and the distance was so great, it was physically impossible for Alice 1 to send a signal to Alice 2 (or vice versa) to coordinate their answers before the measurement was made.

The "Cheating" Loophole (Why it's not 100% perfect yet)

The paper is very honest about a small "loophole."

In a perfect world, Alice 1 and Alice 2 would be in two completely separate rooms that are sealed off from each other. In this experiment, they are in the same lab, and the light travels between them.

• The Loophole: Because the light stays in the lab for a tiny fraction of a second, it is theoretically possible (though highly unlikely in this specific setup) that the two Alices could be "talking" to each other via the light beam itself, rather than the order of events being truly indefinite.
• The Fix: The researchers argue that based on how their machine is built, this "talking" shouldn't happen. However, to be 100% sure (device-independent), they would need to put Alice 1 and Alice 2 in completely separate, sealed locations. They haven't done that yet, but they have shown that with current technology, they are very close.

The Result: Breaking the Rule

The researchers ran the experiment thousands of times. They measured the correlations between the choices made by Alice 1, Alice 2, Bob, and Charlie.

• The Rule: If the world has a definite order, the score should be 1.75 or lower.
• The Result: Their score was 1.807.

This might not sound like a huge difference, but in the world of quantum physics, this is a massive victory. It was 5.7 standard deviations away from the limit. In simple terms, the odds of this happening by random chance are less than one in a million.

Summary

This paper is a major step forward because:

1. It proved the concept: They showed that you can experimentally violate a rule that assumes events happen in a fixed order.
2. It used real distance: They used 3 kilometers of fiber optic cable to ensure the players were far enough apart to prevent easy cheating.
3. It was fast: They synchronized complex electronics to operate at speeds where light couldn't travel between the players in time to coordinate.

They haven't built a time machine, but they have proven that at the quantum level, the universe doesn't always agree on who went first. The "order" of events can be as fuzzy and undefined as the particles themselves.

Technical Summary

Technical Summary: Experimental Violation of a Bell-like Inequality for Causal Order

Problem Statement
Quantum mechanics permits scenarios where physical processes occur in an indefinite causal order, a phenomenon prototypically realized by the "quantum switch." While theoretical frameworks, such as causal witnesses and process tomography, exist to detect this indefiniteness, they often require full characterization of the quantum devices (device-dependent). A more rigorous approach involves violating inequalities on observed correlations, analogous to Bell inequalities for nonlocality. Specifically, van der Lugt, Barrett, and Chiribella (VBC) proposed a Bell-like inequality (the VBC inequality) that can be maximally violated by a quantum switch, potentially allowing for device-independent certification of indefinite causal order. However, experimental realization has been hindered by the stringent requirements: high-fidelity quantum switches, high-quality entanglement, and, crucially, spacelike separation between the operations inside the switch and an external party. Previous attempts lacked the necessary speed and coordination to simulate this separation effectively.

Methodology
The authors report an experimental violation of the VBC inequality using a photonic setup involving four parties: Alice 1, Alice 2, Bob, and Charlie.

• Quantum Switch Architecture: The experiment utilizes a photonic quantum switch where a control qubit (encoded in photon polarization) coherently controls the order of operations performed on a target qubit (encoded in time-bin). The target qubit undergoes a "measure-and-reprepare" protocol at the local stations of Alice 1 and Alice 2.
• Entanglement and Spacelike Separation: The control qubit is maximally entangled with an idler photon shared with Bob. To simulate spacelike separation between Bob and the quantum switch operations, the idler photon is transmitted through a 3-kilometer fiber spool, while the signal photon traverses the switch. This distance, combined with the operational speed, ensures that Bob's measurement events are spacelike separated from the operations of Alice 1 and Alice 2 within the switch.
• High-Speed Operations: The experiment operates at a repetition rate of 0.1 MHz. Key components include:
  • Time-bin Encoding: Used for the target qubit to allow rapid manipulation via electro-optic (EO) switches within asymmetric Mach-Zehnder interferometers (AMZIs).
  • Delayed Readout: To prevent the measurement outcomes of Alice 1 and Alice 2 from revealing the causal order (which would collapse the superposition of orders), the authors employ a delayed-readout strategy. Measurement outcomes ($a_i$) and reparation settings ($x_i$) are encoded in the temporal domain using an ancillary time-bin qubit and trigger signals, allowing them to be decoded by a time-to-digital converter (TDC) only after the control qubit has been measured.
  • Synchronization: A 10-MHz master oscillator synchronizes the acousto-optic modulator (AOM), EO switches, random number generators (RNGs), and TDCs to ensure precise timing.
  • Phase Stabilization: Active temperature stabilization maintains fluctuations below 0.05°C, ensuring high interference visibility (0.98) in the long fiber loops required for the time-bin qubit.

Key Contributions

1. First Experimental Violation of the VBC Inequality: The study achieves the first experimental demonstration of a violation of the VBC inequality, a Bell-like inequality specifically designed for causal order involving a spacelike separated party.
2. Novel Photonic Implementation: The authors implement a quantum switch where a polarization-encoded control qubit coherently controls operations on a time-bin target qubit. This combination allows for fast switching via EO elements, a necessity for the VBC scenario.
3. Technical Advances in Readout and Stabilization: The paper introduces a delayed-readout technique for time-bin qubits that avoids the need for multiple spatial modes (unlike previous polarization-based methods) and implements active temperature control to stabilize kilometer-scale interferometers against phase drift.
4. Simulation of Spacelike Separation: By integrating 3-km fiber spools with high-speed operations, the setup effectively simulates the spacelike separation required to close the "locality" aspect of the causal order test, a prerequisite for the VBC inequality derivation.

Results

• Inequality Violation: The experimental value for the left-hand side of the VBC inequality was measured to be 1.807 ± 0.010. This exceeds the theoretical bound of 1.75 (derived from the assumptions of Definite Causal Order, Relativistic Causality, and Free Interventions) by 5.7 standard deviations (SDs).
• Component Analysis: The measured terms were $0.490 \pm 0.004$, $0.492 \pm 0.004$, and $0.825 \pm 0.009$, closely matching theoretical predictions.
• No-Signaling Verification: The authors verified that the observed correlations satisfy the no-signaling constraints imposed by the causal structure. Statistical analysis showed that probability differences regarding signaling between parties fell within 3 SDs of zero, confirming the validity of the causal assumptions.
• Error Sources: The deviation from the theoretical maximum (1.8536) is attributed primarily to bit-flip noise from EO switch cross-talk (isolation 20–23 dB) and residual phase noise due to temperature fluctuations.

Significance and Limitations
The paper claims that these results provide a statistically significant certification of indefinite causal order under conditions that guarantee spacelike separation with current state-of-the-art devices. This advances the field from theoretical proposals and device-dependent witnesses to a correlation-based test that approaches device-independent certification.

However, the authors explicitly state that the demonstration is not fully device-independent or loophole-free.

• Bidirectional Signaling Loophole: The primary limitation is that the laboratories of Alice 1 and Alice 2 accept input photons for an extended period. In principle, this allows for bidirectional signaling between the two Alices (e.g., Alice 1 affecting Alice 2's outcome) within the classical spacetime of the experiment. The violation of the inequality could, in theory, be explained by such signaling rather than indefinite causal order.
• Device Dependence: The exclusion of this signaling loophole relies on knowledge of the experimental setup (specifically, that a single photon mediates the interaction and cannot travel both ways simultaneously in a classical spacetime). Therefore, the certification is not "device-independent" in the strict sense used for Bell nonlocality.
• Future Outlook: The authors note that a fully loophole-free, device-independent certification would require future experiments probing the quantum nature of spacetime or scenarios where bidirectional signaling is fundamentally ruled out.

In summary, this work demonstrates a major experimental step toward certifying indefinite causal order by violating a Bell-like inequality, combining high-speed photonic operations with long-distance fiber delays, while acknowledging the inherent limitations of implementing such tests within classical spacetime.