Were you always fascinated by multidisciplinary research and had the ambition during your PhD research to go beyond your own specialized field to build a bridge to other research areas? Would you like to do exactly that and connect integrated photonics with synthetic organic chemistry in order to tackle the challenges in neuromorphic computation? Then you should read on and consider applying to this Postdoctoral Researcher position.Background
The current computing paradigm based on the Van-Neuman architecture is reaching its limits. The bottleneck lies in the serial link between processing and memory units and with the adoption of deep learning for neural networks in various scientific and industrial fields, parallel computing hardware based on neuromorphic architectures promise to provide better performance at higher energy efficiency. Within neuromorphic hardware, synaptic weighting, or in general, a linear matrix transformation consumes the most energy.
Photonic implementations of such linear matrix transformations can exploit various properties of light to achieve a high degree of parallelism. Many photonic matrix multiplication devices are based on free space optics which is difficult to scale due to their large size. A handful of concepts exist that use networks of optical interferometers, ring resonators, amplifiers or phase change materials that are suitable to be miniaturized and integrated on a photonic chip.Project Description
In this project, we will use multiple waveguide modes in an integrated photonic device to perform matrix multiplication. For that, local refractive index control at high resolution is needed. The candidate will investigate how to achieve that with islands of liquid crystals that can be either electrically or optically tuned. The candidate will address several challenges connected to that:
- What types of liquid crystals have the most suitable electrical and optical properties.
- How to best deposit islands of liquid crystal on InP photonic integrated circuits.
- How to best contact and control a matrix of liquid crystal islands.
The project creates a bridge between material sciences and engineering by targeting a specific concept envisioned for neuromorphic computing. Your research will enable a whole new class of reconfigurable photonic devices with micrometer control, which could lead to highly efficient parallel optical processing engines.
This project is part of the Eindhoven Hendrik Casimir Institute's
(EHCI) collaborative projects framework within the focus area Neuromorphic Computing
. EHCI aims to make the roadmap for exponential technology happen, by developing new technologies, such as new types of processors, highly-efficient communication transceivers, secure communication technology and ubiquitous sensors.