Master's defense: Kasper H Nielsen

Title: Towards programmable nonlinear photonic quantum simulation

Abstract: Quantum computers have the potential to change what is possible in the field of simulating quantum systems and other tasks that utilize the extra computational complexity of quantum states. While a universal fault tolerant quantum computer is not available yet, we can still benefit from the extra computational complexity of quantum states in noisy intermediate-scale quantum (NISQ) devices with experiments and technology available today. In this thesis we investigate the possibility of simulating molecular quantum dynamics with non-linear photonic circuits consisting of linear optical interferometers and quantum dot non-linearities. In particular, we develop tools to simulate and design such programmable non-linear photonic circuits to implement quantum simulation of molecular dynamics, and report preliminary experimental efforts towards their implementation with time-bin photonic interferometers and quantum dots.

The main results form the thesis are the classical simulation of two mode quantum dot non-linear optical circuits which is programmed to simulate vibrations in a water molecule and the experimental realisation of multi-photon interference in a multimode time-bin interferometer. The classical simulation of a quantum dot non-linear photonic circuit shows promising results for the approach, with the possibility for scaling up to more modes and photons.

The main experimental effort of this these was to build and characterise a multimode time-bin interferometer, which integrates naturally with the quantum dot single photon source. Finally we perform 2-photon experiments with the interferometer for demonstration of the bosonic suppression law and entanglement generation between two single photons. Fidelities of 88 ± 3% and 89 ± 1%, respectively, are reported. The experiments are also simulated where partial indistinguishability of the photons and experimental errors are accounted for. These simulations yield a fidelity between simulation and experiment of 99±3% and 98±1%, respectively, demonstrating a good understanding of the noise processes in the setup.