CDT Compound Semiconductor Physics (PhD Progression 1+3)

点击翻译按钮获取中文摘要

Vertical cavity surface emitting lasers (VCSELs) have widespread use in applications such as datacoms, atomic clocks, LiDAR, face/gesture recognition and augmented reality. In the field of neuromorphic computing, optical pulses or 'spiking' is used to imitate the firing of neurons. If designed in the correct way, VCSELs can also be used to target photosensitive proteins (opsins), causing activation of biological neurons used in optogenetic therapy. Combining this with a VCSEL designed for spiking in neuromorphic computing could provide an effective route to neural-protheses (brain-computer interfacing), an exciting new area of neuroscience and biomedical engineering. Individually, VCSELs are relatively low power emitters (10s mW), but in arrays or coupled configurations this power can be scaled, albeit with appropriate thermal management. Germanium (Ge) substrates have a higher thermal conductivity than gallium arsenide and allow for thinner substrates further aiding heat extraction, an area which can be explored in the project. Depending on the configuration, arrays of VCSELs can be sparsely or densely packed together and this can cause significant variation in device architecture due to the impact on process parameters, such as 'aspect-ratio dependant etching'. Additionally, the need for gain and saturable absorber sections in short-pulse VCSELs requires careful control of etching parameters to meet the required depths for intra-cavity electrical contacts. This project will focus on optimising process parameters to achieve addressable arrays of VCSELs for spiking or high pulse-powers and testing will be used to inform improvements in fabrication. Furthermore, routes to separate out dense arrays using plasma-die separation techniques will be explored, in the case of Ge substrates this can become critical due to cleaving difficulties.