Experimental evaluation of non-classical correlations between measurement outcomes and target observable in a quantum measurement

TL;DR

This paper experimentally investigates the role of non-classical correlations in quantum measurements by analyzing a sequence of photon polarization measurements, revealing significant deviations from classical expectations at the transition between weak and strong measurements.

Contribution

It provides an experimental analysis of non-classical correlations affecting measurement outcomes, highlighting their impact during the transition from weak to strong measurements.

Findings

Non-classical correlations influence measurement evaluations.

Deviations from classical eigenvalue ranges occur at measurement transition.

Quantum effects become prominent when varying measurement resolution.

Abstract

In general, it is difficult to evaluate measurement errors when the initial and final conditions of the measurement make it impossible to identify the correct value of the target observable. Ozawa proposed a solution based on the operator algebra of observables which has recently been used in experiments investigating the error-disturbance trade-off of quantum measurements. Importantly, this solution makes surprisingly detailed statements about the relations between measurement outcomes and the unknown target observable. In the present paper, we investigate this relation by performing a sequence of two measurements on the polarization of a photon, so that the first measurement commutes with the target observable and the second measurement is sensitive to a complementary observable. While the initial measurement can be evaluated using classical statistics, the second measurement introduces the effects of quantum correlations between the non-commuting physical properties. By varying the resolution of the initial measurement, we can change the relative contribution of the non-classical correlations and identify their role in the evaluation of the quantum measurement. It is shown that the most striking deviation from classical expectations is obtained at the transition between weak and strong measurements, where the competition between different statistical effects results in measurement values well outside the range of possible eigenvalues.