⚛️ quantum physics

Direct and Efficient Detection of Quantum Superposition

This paper presents a resource-efficient scheme that uses local measurements and an XOR game to directly verify quantum superposition with exponentially high confidence, successfully demonstrating 99% certainty in a single-photon experiment without requiring the recombination of superposed states.

Original authors: Daniel Kun, Teodor Strömberg, Michele Spagnolo, Borivoje Dakić, Lee A. Rozema, Philip Walther

Published 2026-04-16

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel Kun, Teodor Strömberg, Michele Spagnolo, Borivoje Dakić, Lee A. Rozema, Philip Walther

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

Imagine you have a magical coin. In the real world, a coin is either Heads or Tails. But in the quantum world, a particle can be both Heads and Tails at the exact same time. This is called superposition.

For a long time, scientists have had a hard time proving this "magic" is real without breaking the spell. Usually, to prove a coin is in a superposition, you have to flip it, catch it, and then smash the two possibilities together to see an interference pattern (like ripples in a pond). It's like trying to prove a ghost is in the room by making it walk through a wall and watching the wall shake. It's indirect, complicated, and sometimes people argue, "Maybe it's just a trick of the light."

This new paper introduces a clever, direct way to catch the ghost in the act without ever making it walk through a wall. They do this using a game.

The Game: "Guess the Secret Code"

Imagine two friends, Alice and Bob, are in separate rooms. They can talk to each other, but they can't see what the other is doing.

The Referee: A third person, the Referee, has a special "Test Particle" (let's call it a Quantum Token). He secretly decides to put a "Heads" mark on the left path or a "Tails" mark on the right path. He might do both, or neither. He then sends the token to either Alice or Bob, but he keeps the secret code (the combination of marks) hidden.
The Challenge: The Referee asks Alice and Bob to guess the XOR of his secret code.
- What is XOR? It's a simple logic rule: "Are the two marks different?" If one is Heads and one is Tails, the answer is "Yes" (1). If they are the same (Heads/Heads or Tails/Tails), the answer is "No" (0).
The Rules:
- If the token is a normal, classical object (like a real coin), it can only travel down one path. It carries information about only one mark. Alice and Bob can only guess. Their best chance of winning is 50/50.
- If the token is a quantum superposition, it travels down both paths at once. It carries information about both marks simultaneously.

The Secret Weapon: The "Ghost" Helper

Here is the genius part. Alice and Bob aren't just guessing blindly. They share a second "Helper Token" that is also in a superposition, split between their two rooms.

When the Referee sends the Test Token, Alice and Bob mix it with their Helper Token on a special splitter.

If the Test Token was classical: The mixing does nothing special. They still have to guess.
If the Test Token was quantum: Because the token was in two places at once, it interacts with the Helper Token in a way that creates a "secret handshake" between Alice and Bob. When they look at their detectors, the results are perfectly correlated (or anti-correlated) based on the Referee's secret code.

By looking at these correlations, they can deduce the secret code with a success rate of 75% (or even higher in ideal conditions), far beating the 50% limit of classical guessing.

The Analogy: The Invisible Ink

Think of it like this:

Classical Particle: A letter written in invisible ink that is either on the left page or the right page. You can only read one page. You can't know what's on the other.
Quantum Superposition: A letter written in invisible ink that is simultaneously on both pages.
The Experiment: Instead of trying to read the letter directly (which would fade the ink), Alice and Bob use a "magic scanner" (the second photon) that reacts to the presence of the ink on both pages. The scanner beeps differently depending on whether the ink was on one page or both.

Why This Matters

Direct Proof: They didn't have to smash the paths together to see interference. They proved the particle was in two places at once just by playing a game and looking at the results.
Efficiency: The paper shows that you don't need millions of trials to be sure. With just 37 copies of the particle, they could be 99% confident that the particle was truly in a superposition. It's like flipping a coin 37 times and getting 37 heads; you'd be pretty sure it's not a normal coin.
No Complex Gear: Previous methods required complex, delicate equipment (like homodyne detectors). This method uses simple detectors and just counts "clicks," making it much easier to build and use.

The Bottom Line

The researchers built a "non-local interferometer" (a fancy way of saying a machine that checks two places at once without the particle ever meeting itself). By turning the verification of quantum superposition into a simple logic game, they proved that the particle was indeed in two places at once, efficiently and directly.

It's like proving a magician is using a double-sided card not by looking at the card, but by playing a game where the magician's score is impossible to achieve with a single-sided card.

1. Problem Statement

The principle of quantum superposition is fundamental to quantum theory, yet its direct verification remains challenging.

Indirect Nature of Current Methods: Standard verification methods (e.g., Young's double-slit, delayed-choice experiments) rely on re-interfering the superposed paths of a single particle. This is an indirect inference; the particle is never observed in the superposition state itself, only in the resulting interference pattern.
Classical Explanations: Recent theoretical work suggests that some interference-based demonstrations can be explained classically, and Bell inequality violations using single photons often require complex homodyne measurements or are limited by particle-number superselection rules when applied to massive particles.
The Goal: The authors aim to develop a method to directly verify that a single particle exists in a coherent spatial superposition without re-interfering its paths, using only local measurements and a resource-efficient protocol.

2. Methodology

The authors adapt a theoretical framework based on XOR games (specifically proposed by Del Santo and Dakić) and implement it using a nonlocal interferometer.

Theoretical Framework: The XOR Game

Setup: A Referee challenges two separated players, Alice and Bob, to guess the XOR sum ( $x \oplus y$ ) of two bits ( $x, y$ ) encoded by the Referee on two spatial paths.
Classical Limit: If the test particle is a classical object (localized in one path), it carries information about only one bit. The optimal winning probability for classical strategies is $P_{win} = 0.5$ (random guessing).
Quantum Advantage: If the particle is in a coherent superposition of both paths, Alice and Bob can share an ancillary photon (also in a superposition) and perform joint measurements. This allows them to achieve $P_{win} > 0.5$ .
Specific Protocol: The Referee applies $\pi$ $π$ -phase shifts ($0$ or $\pi$ $π$ ) to the paths. Alice and Bob interfere their local test photon with their shared ancilla photon on local beamsplitters.
- If both photons arrive at the same lab (50% probability), no phase information is gained, and they guess randomly.
- If one photon arrives at each lab (50% probability), the detection events (correlated vs. anti-correlated) depend non-locally on the sum of the Referee's phases ( $\phi_x + \phi_y$ ).
- By outputting the detector index, they can win with a theoretical probability of $P_{win} = 0.75$ for pure quantum states.

Experimental Implementation

Source: A Type-II Spontaneous Parametric Down-Conversion (SPDC) source generates indistinguishable photon pairs. One acts as the Test Photon (T), the other as the Ancilla/Measurement Photon (M).
Nonlocal Interferometer:
- Both photons are prepared in spatial superpositions ( $|1,0\rangle + |0,1\rangle$ ).
- The Referee (central station) controls phase shifts on the Test Photon's paths using Piezoelectric Transducers (PZTs).
- Alice and Bob receive one mode of the Test Photon and one mode of the Ancilla Photon.
- They locally interfere these modes on 50:50 beamsplitters and record coincidence counts between their detectors.
Decoherence Control: To test the transition from quantum to classical, the authors introduce controlled phase noise (Gaussian distribution) to the Referee's phase setting, effectively tuning the purity ( $P$ ) of the test photon's state from 1 (pure) to ~0.54 (mixed).

3. Key Contributions

Direct Verification without Re-interference: The experiment demonstrates that superposition can be verified by measuring correlations between separated parts of a delocalized state and a shared resource, without ever recombining the superposed paths of the test particle itself.
Resource Efficiency: The protocol converts the verification task into a statistical game. The confidence that the particle is superposed approaches unity exponentially fast with the number of measured copies.
Discrete Variable Approach: Unlike previous single-photon Bell tests that required complex homodyne detection (continuous variables), this method uses simple single-photon detectors and discrete variable logic (XOR games).
Robustness to Imperfections: The authors derived a theoretical model accounting for experimental imperfections (beamsplitter imbalance, finite Hong-Ou-Mandel visibility) and validated it against experimental data without free parameters.

4. Results

Winning Probability: In the pure state regime, the experiment achieved an average winning probability of $P_{win} = 0.716 \pm 0.007$ .
- This is significantly above the classical limit of $0.50$.
- It closely matches the theoretical maximum of $0.7162$ predicted by their model (Eq. 6), which accounts for the measured HOM visibility ( $V = 94 \pm 2\%$ ) and beamsplitter ratios.
Purity Transition: As the purity of the test photon was reduced (via phase noise), the winning probability decreased linearly, following the predicted curve. This confirmed that the XOR game sensitivity scales directly with the coherence of the superposition.
Confidence Estimation:
- The authors calculated the statistical confidence ( $C = 1 - p\text{-value}$ ) that the particle is in a superposition.
- Key Finding: Only 37 copies (game instances) were required to reach a 99% confidence level that the particle was in a superposition.
- The residual confidence error decays exponentially with the number of copies ( $N$ ).

5. Significance

Foundational Impact: This work provides a "direct" demonstration of the superposition principle, addressing long-standing concerns about whether superposition is merely a mathematical artifact or a physical reality observable without self-interference.
Practical Utility: The method offers a highly efficient tool for quantum resource verification. It can be used to certify the quality of superposition states in quantum networks or quantum computing architectures with minimal overhead.
Scalability: The exponential convergence of confidence suggests this protocol is highly scalable for verifying quantum resources in larger systems or for massive particles (where particle-number superselection rules might otherwise hinder Bell tests).
Technological Advancement: By utilizing nonlocal interferometry and XOR games, the authors bridge the gap between game-theoretic quantum protocols and practical experimental verification, avoiding the complexity of homodyne detection.

In summary, Kun et al. successfully demonstrated a direct, efficient, and statistically robust method to verify quantum superposition using local measurements and an XOR game, proving that a single photon can be certified as delocalized with high confidence using only 37 experimental runs.