Theory-agnostic nonclassicality certification in an integrated photonic circuit

The quest to understand the quantum world often begins with a simple, stubborn question: how do we know something is truly quantum? For decades, scientists have relied on the mathematical framework of quantum mechanics to answer this, assuming that the equations describing atoms and light are the final word on reality. But what if those equations are just a temporary map, a useful tool that might one day be replaced by something deeper? To build technologies that are truly secure and powerful, researchers need a way to certify that a device is behaving in a genuinely nonclassical way without assuming that quantum theory is the only truth. This requires a method that looks only at the raw data—the inputs fed into a machine and the outputs it produces—without peeking at the internal machinery or assuming a specific physical theory. The goal is to find a signature of "contextuality," a property where the result of a measurement depends on the context in which it is taken, a behavior that has no parallel in the classical world of everyday objects.

In a recent study, a team of researchers took this theoretical challenge and brought it into the physical world using a tiny chip made of glass and light. They built an integrated photonic circuit, a microscopic maze of channels etched into a chip, designed to guide single photons. By injecting a single particle of light into one end and measuring where it emerged from the other, they created a controlled experiment. The researchers did not assume the chip was behaving according to quantum laws. Instead, they treated the device as a black box, recording the statistical patterns of where the photons landed under different settings. Using a technique called theory-agnostic tomography, they reconstructed the effective "shape" of the experiment's possibilities directly from these numbers, asking what kind of underlying reality could produce such data. They then tested this reconstructed reality to see if it could be explained by any classical, noncontextual model. The result was a definitive certification: the data could not be explained by any classical theory, proving the device was exhibiting genuine nonclassical behavior without ever invoking the standard equations of quantum mechanics.

The experiment took place on a specialized chip that acts as a three-way intersection for light. The researchers sent single photons into the first port of this circuit. Inside, the path of the light was manipulated by changing the relative phases of the waves, a process controlled by applying small electrical voltages to the chip. These voltages acted as the "knobs" for the experiment, defining how the photon was prepared and how it was measured. The team varied these voltages across a wide range, from zero up to about 55 millivolts for the preparation stage and up to 52 millivolts for the measurement stage. For every specific combination of these voltages, they counted how many photons arrived at each of the three possible output ports. Over thousands of trials, they built a massive table of probabilities, mapping out exactly how the light behaved under every condition they tested. This data was the only thing they trusted; they did not start by assuming the photons were following quantum rules.

To make sense of this mountain of numbers, the researchers used a method that builds a physical model from the ground up, using only the data itself. They asked a fundamental question: what is the simplest, most accurate description of the system that fits these numbers without overcomplicating things? In the language of the study, they looked for the "rank" of the data, which corresponds to the number of independent dimensions needed to describe the system's behavior. If the system were purely classical, a very simple, low-dimensional description would suffice. If it were quantum, it would require a more complex, higher-dimensional space. By analyzing the data with a computer algorithm that tested different levels of complexity, they found that the best fit required a seven-dimensional space. This was a crucial finding because a standard three-level quantum system, which is what the chip was designed to mimic, would typically require a nine-dimensional space to be fully described. The fact that the data naturally settled on a seven-dimensional description meant the experiment was capturing the essential physics without needing the full, unverified machinery of quantum theory.

With this reconstructed model in hand, the team then performed the final test for nonclassicality. They checked whether the seven-dimensional model they had built could be embedded into a classical "simplex," a geometric shape that represents all possible classical states. If the data could fit inside this classical shape, the experiment would be explainable by ordinary physics. If it could not, the system was exhibiting contextuality. The researchers ran a linear programming test, a rigorous mathematical check, to see if their model could be squeezed into this classical framework. The answer was a clear no. The model refused to fit. The data showed a robustness of 0.200, meaning the nonclassical nature of the system was strong enough to withstand a significant amount of simulated noise before it would collapse into a classical explanation. This provided solid, theory-independent evidence that the photonic circuit was doing something that classical physics simply cannot do.

To ensure their findings were not an artifact of their specific method, the researchers also built a separate, standard quantum model of the chip based on known physics. They ran the same analysis on this theoretical model and found it produced the same seven-dimensional structure and a similar level of nonclassicality. This consistency acted as a powerful cross-check, confirming that the theory-agnostic method was working correctly. In contrast, when they applied the same test to a purely classical simulation of the device, the result was zero robustness, correctly identifying it as classical. This comparison ruled out the possibility that the method was falsely flagging classical data as quantum. The study successfully demonstrated that integrated photonic circuits, which are a leading platform for future quantum computers, can be certified as nonclassical without assuming quantum theory is the final truth.

This work marks a significant step forward in connecting the foundations of physics with practical technology. By proving that a scalable, chip-based system can be certified as nonclassical using only its input-output statistics, the researchers have shown that we do not need to trust the internal workings of a device to know it is quantum. This approach is vital for the future of quantum technologies, where security and reliability depend on being able to verify that a device is truly operating beyond classical limits, even if the underlying theory of the universe changes. The experiment confirms that the strange, counterintuitive behaviors of the quantum world are not just mathematical abstractions but real, measurable features that can be harnessed and certified on a chip, paving the way for more robust and trustworthy quantum devices.