Nonperiodic silicon photonic devices for nonlinear photon generation and multimode design for optical computing

Zhetao Jia

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2024-200
December 1, 2024

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-200.pdf

Integrated photonics has experienced significant growth in both academic research and industry startups. Chip-scale photonic devices are transforming key areas including communication, computing, sensing, navigation, and augmented/virtual reality (AR/VR). However, challenges in current platforms hinder the practical application of these technologies. Specifically, weak materials nonlinearities have limited direct on-chip photon generation for both classical and quantum light sources, and also existing photonic devices' bandwidth for communication/computing is not meeting current demands.

The goal of this thesis is to provide possible solutions to the two issues in silicon photonics. First, I will discuss using topological structures to enhance the transport and nonlinear frequency generation in silicon photonics. We propose a non-periodic photonic system with a structural disorder that confines light near the system boundary, enhancing the nonlinear photon generation rate compared to periodic systems. Next, I will introduce a systematic inverse-design method, aiming to boost the efficiency of on-chip nonlinear photon generation, along with a physical interpretation of these results. Specifically, I will present our experimental demonstration of a compact, robust, and efficient entangled photon pair source based on spontaneous four-wave mixing, achieving a generation rate of 1.1MHz and a coincidence to accidental ratio of 162. This method can also be generalized for other nonlinear optical processes. In the second part of the thesis, I will show the possibility of expanding the optical computing bandwidth with a mode-division multiplexing (MDM) strategy, offering a new degree of freedom in optical computing with the micro-ring resonator platform. I will outline an MDM strategy suitable for small-scale neural networks and a multi-dimensional architectural approach for large-scale optical computing applications. I will present the experimental results of our device for matrix multiplication fabricated at foundry.

Advisor: Boubacar Kanté

\"Edit"; ?>

BibTeX citation:

@phdthesis{Jia:EECS-2024-200,
    Author = {Jia, Zhetao},
    Title = {Nonperiodic silicon photonic devices for nonlinear photon generation and multimode design for optical computing},
    School = {EECS Department, University of California, Berkeley},
    Year = {2024},
    Month = {Dec},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-200.html},
    Number = {UCB/EECS-2024-200},
    Abstract = {Integrated photonics has experienced significant growth in both academic research and industry startups. Chip-scale photonic devices are transforming key areas including communication, computing, sensing, navigation, and augmented/virtual reality (AR/VR). However, challenges in current platforms hinder the practical application of these technologies. Specifically, weak materials nonlinearities have limited direct on-chip photon generation for both classical and quantum light sources, and also existing photonic devices' bandwidth for communication/computing is not meeting current demands.


The goal of this thesis is to provide possible solutions to the two issues in silicon photonics. First, I will discuss using topological structures to enhance the transport and nonlinear frequency generation in silicon photonics. We propose a non-periodic photonic system with a structural disorder that confines light near the system boundary, enhancing the nonlinear photon generation rate compared to periodic systems. Next, I will introduce a systematic inverse-design method, aiming to boost the efficiency of on-chip nonlinear photon generation, along with a physical interpretation of these results. Specifically, I will present our experimental demonstration of a compact, robust, and efficient entangled photon pair source based on spontaneous four-wave mixing, achieving a generation rate of 1.1MHz and a coincidence to accidental ratio of 162. This method can also be generalized for other nonlinear optical processes. In the second part of the thesis, I will show the possibility of expanding the optical computing bandwidth with a mode-division multiplexing (MDM) strategy, offering a new degree of freedom in optical computing with the micro-ring resonator platform. I will outline an MDM strategy suitable for small-scale neural networks and a multi-dimensional architectural approach for large-scale optical computing applications. I will present the experimental results of our device for matrix multiplication fabricated at foundry.}
}

EndNote citation:

%0 Thesis
%A Jia, Zhetao
%T Nonperiodic silicon photonic devices for nonlinear photon generation and multimode design for optical computing
%I EECS Department, University of California, Berkeley
%D 2024
%8 December 1
%@ UCB/EECS-2024-200
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-200.html
%F Jia:EECS-2024-200