Projects
Neuromorphic photonics is an emerging technology that combines principles from photonics and neuroscience to create advanced computing systems inspired by the human brain. It involves the design and development of photonic systems that mimic the behaviour and functionality of biological neural networks.
Neuromorphic photonics offers unprecedented computational power, energy efficiency, and parallel processing capabilities, surpassing the limitations of traditional computing methods. This technology has the potential to revolutionize areas like machine learning, complex data analysis, all-optical computing, sensing, and communications.
The George Green Institute for Electromagnetics Research group focuses on innovative photonic artificial neuron architectures. These architectures leverage the unique properties of light and chaos physics to achieve efficient and parallel computation. Additionally, we explore applications of neuromorphic photonics in sensing by decoding complex spectral sensory systems, capitalising on light's interaction with artificial neurons for fast and accurate computations. Our research also delves into novel architectures based on non-linear Stimulated Brillouin scattering phenomena, enabling high-speed and energy-efficient information processing without active optical components.
On-chip artificial neuron based on photonic reservoir computing. [DOI: 10.1049/pbcs077g_ch12 and 10.1117/12.2548716]
Application of photonic neuron to perform complex statistical analysis for sensor’s signal processing. Photonic neuron can learn based on discrete training dataset and make an inference on unknown data. [Optical Materials Express 12 (5), 1767-1783, 2022]
Application of photonic neuron to perform complex signal processing to improve the quality of transmitted information by reducing distortion and noise caused by the communication channel, ensuring clearer and more reliable communication. Image shows for a specific 4-bits communication system. [Optics Express 31 (13), 22061-22074, 2023]
Photonic Crystal Surface Emitting Lasers
A photonic crystal is a pattern of holes in a transparent material that can strongly affect the way in which light travels through the crystal. This control over the flow of light can be used to enhance the performance of lasers. The resulting devices are called “Photonic Crystal Surface Emitting Lasers” (PCSELS)..
The photonic crystal allows these devices to operate at a much higher power and faster switching speed than other semiconductor lasers [1, 2]. High speed operation may earn these lasers a home in data centres as optical interconnects, while narrow beams and high power make them suitable as LiDAR sources [3].
We work on the simulation and theory of the lasers. Among our methods is the Unstructured Transmission Line Method (UTLM) for simulating complicated three-dimensional optical geometries [4] that offers a lot of freedom to include the full 3D complexity of the problem in the simulation
UTLM simulation of a photonic crystal laser. Various materials are layered from left to right in slabs. One layer contains a single air hole, triangular but with rounded corners. Thousands of such holes are arranged in a grid to realise the laser, only one square of the grid is shown. (a) Shows the electric field along a plane, while (b) shows the mod-squared electric field in a series of planes orthogonal to that of (a). These electric fields are associated with the light of the laser.
Parity and Time (PT) symmetric structures
The recently discovered Parity and Time (PT) symmetric structures allow a judicious balancing act of both loss and gain to yield distinctive features and unique properties.
The discovery of such structures has been regarded as one of the top 10 discoveries in Physics over the last 10 years (J. Cham, Nature Phys, 11, 799, 2015) since they offer a novel approach to the on-chip control and manipulation of light. In collaboration with colleagues in the School of Mathematical Sciences we have studied the distinctive features of PT symmetric Bragg gratings, coupled resonators and chain resonators and unravelled some of their great potential. Our numerical and semi-analytical studies of one- and two-dimensional PT symmetric structures with idealised material properties was extended to include the impact of realistic material properties via a time-domain Transmission-Line Modelling (TLM) method.
S. Phang, et. al. Sci. Rep. 6 (20499), 2016
Study of wave propagation in large-core asymmetric step index fibres (SIF)
In further work with colleagues in the School of Mathematical Sciences we have developed a ray dynamic approach for the study of wave propagation in large-core asymmetric step index fibres (SIF), especially those made from chalcogenide glasses (Also see Mid-Infrared Photonics). An operator is developed that propagates the intensity distribution of light along the fibre via a discrete mapping of pixelated phase space. This provides an efficient means by which to determine for the power accumulated in the fibre core following an arbitrary excitation, vital in the design of both fibre-based amplifiers and fibre-based luminescent light sources for spectroscopy.
Numerical studies of the impact of plasmonic effects
Our wave-modelling work also includes numerical studies of the impact of plasmonic effects, associated with silver nanowires, on the behaviour of the lasing frequencies and thresholds of the grating modes in a binary grating formed from silver wires and quantum wires.