Time-bin encoding for optical quantum computing
File(s)
Author(s)
Sempere Llagostera, Santiago
Type
Thesis
Abstract
Scalability has been a longstanding issue in implementing large-scale photonic experiments for optical quantum computing. Traditional encodings based on the polarisation or spatial degrees of freedom become extremely resource-demanding when the number of modes becomes large, as the need for many nonclassical sources of light and the number of beam splitters required become unfeasible. Alternatively, time-bin encoding paves the way to overcome some of these limitations, as it only requires a single quantum light source and can be scaled to many temporal modes through judicious choice of pulse sequence and delays. Such an apparatus constitutes an important step toward large-scale experiments with low resource consumption.
This work focuses on the time-bin encoding implementation. First, we assess its feasibility by thoroughly investigating its performance through numerical simulations under realistic conditions. We identify the critical components of the architecture and find that it can achieve performances comparable to state-of-the-art devices. Moreover, we consider two implementation approaches, in fibre and free space, and enumerate their strengths and weaknesses. Subsequently, we delve into the lab to explore these schemes and the key components involved therein. For the fibre case, we report the first implementation of time-bin encoded Gaussian boson sampling and use the samples obtained from the device to search for dense subgraphs of sizes three and four in a 10-node graph.
Finally, we complement the study of the time-bin encoding with two side projects that contribute to the broad spectrum of enabling techniques for quantum information science. First, we demonstrate the ability to perform photon-number resolving measurements with a commercial superconducting nanowire single-photon detector system and apply it to improve the statistics of a heralded single-photon source. Second, we demonstrate that by employing a phase-tunable coherent state, we can fully characterise a multimode Gaussian state through solely the low-order photon statistics.
This work focuses on the time-bin encoding implementation. First, we assess its feasibility by thoroughly investigating its performance through numerical simulations under realistic conditions. We identify the critical components of the architecture and find that it can achieve performances comparable to state-of-the-art devices. Moreover, we consider two implementation approaches, in fibre and free space, and enumerate their strengths and weaknesses. Subsequently, we delve into the lab to explore these schemes and the key components involved therein. For the fibre case, we report the first implementation of time-bin encoded Gaussian boson sampling and use the samples obtained from the device to search for dense subgraphs of sizes three and four in a 10-node graph.
Finally, we complement the study of the time-bin encoding with two side projects that contribute to the broad spectrum of enabling techniques for quantum information science. First, we demonstrate the ability to perform photon-number resolving measurements with a commercial superconducting nanowire single-photon detector system and apply it to improve the statistics of a heralded single-photon source. Second, we demonstrate that by employing a phase-tunable coherent state, we can fully characterise a multimode Gaussian state through solely the low-order photon statistics.
Version
Open Access
Date Issued
2022-11
Date Awarded
2023-10
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Kolthammer, William
Walmsley, Ian
Sponsor
Engineering and Physical Sciences Research Council
Grant Number
P510257
Publisher Department
Physics
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)