Quantum Correlations of Exciton-Polaritons

Abstract

n this work we study the quantum nature of a hybrid light–matter system, where we use the interactions of the matter component to introduce a sizable nonlinearity to photons. It is a longstanding goal to achieve interactions that overcome the losses in the system, resulting in the photon blockade effect that leads to nonclassical signatures measurable in photon correlation experiments. To implement such a system we combine an optical cavity with excitons in an InGaAs quantum well where the strong light–matter coupling leads to new hybridised eigenstates called polaritons. To further enhance the existing interactions between these polaritons we introduce lateral confinement of the optical mode and a second quantum well. This allows for the coupling to indirect excitons, where the spatial separation of electrons and holes introduces dipolar interactions to the system. In order to measure correlations over a wide range of polariton compositions, we use an open cavity design allowing for in situ tuning of the cavity length. This design naturally results in sizable fluctuations of the cavity length, directly translating to the energy of the optical resonance. We therefore developed a measurement procedure based on postselecting photon arrival times based on their countrate, allowing us to overcome these fluctuations and resolve the correlations of the polaritons. The interactions between purely direct exciton–polaritons allowed for the first observation of nonclassical polariton correlations by continuous wave excitation, with a value g(2)(0) = 0.90(1) at an cavity content of 28 %, marking the current record for comparable systems. Correlation measurements at lower exciton fractions show the presence of a new regime, where a small antibunching persists independent of the detuning between the excitation laser and the polariton mode. This behaviour cannot be explained by polariton–polariton interactions. We attribute them to the presence of the biexciton and propose a “dissipative blockade” mechanism, where the selective coupling of the biexciton to the doubly excited polariton state lowers the probability of the two-photon emission. The good agreement of the data with our numerical simulations strongly supports this mechanism as origin of the detuning independent antibunching. To our knowledge this would mark the first observation of nonclassical correlations originating from a dissipative blockade mechanism, and could potentially offer a novel approach to the creation of single-photon states. At the same time, the polariton interaction strengths extracted from the simulation depend much stronger on the exciton content as expected from the commonly used quadratic scaling law. And while we do not have an explanation for this behaviour, it suggests that a microscopic theory might be necessary to capture the details in the scaling of the interactions of these composite particles. At the same time careful measurements of the polariton properties show that as a result of the increased indirect exciton content, the exciton oscillator strength is substantially reduced. This leads to a broadened linewidth and a lower transmission which limits the available parameter space. As a result, the expected interaction enhancement could not be observed in this system.