Experimental test of quantum causal influences

TL;DR

This paper experimentally demonstrates quantum nonclassicality using causal influence tests in an instrumental scenario, revealing quantum effects beyond Bell inequality violations by employing interventions in a photonic setup.

Contribution

It introduces and experimentally verifies a new quantum nonclassicality witness based on causal influence in the instrumental scenario, distinct from Bell tests.

Findings

Quantum correlations violate classical bounds with interventions.

The experimental setup faithfully implements the instrumental causal structure.

Quantum bounds for causal influence are validated as reliable tools.

Abstract

Since Bell's theorem, it is known that the concept of local realism fails to explain quantum phenomena. Indeed, the violation of a Bell inequality has become a synonym of the incompatibility of quantum theory with our classical notion of cause and effect. As recently discovered, however, the instrumental scenario -- a tool of central importance in causal inference -- allows for signatures of nonclassicality that do not hinge on this paradigm. If, instead of relying on observational data only, we can also intervene in our experimental setup, quantum correlations can violate classical bounds on the causal influence even in scenarios where no violation of a Bell inequality is ever possible. That is, through interventions, we can witness the quantum behaviour of a system that would look classical otherwise. Using a photonic setup -- faithfully implementing the instrumental causal structure and allowing to switch between the observational and interventional modes in a run to run basis -- we experimentally observe this new witness of nonclassicality for the first time. In parallel, we also test quantum bounds for the causal influence, showing that they provide a reliable tool for quantum causal modelling.