Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.578586
Title: Nanophotonic filters for digital imaging
Author: Walls, Kirsty
Awarding Body: University of Glasgow
Current Institution: University of Glasgow
Date of Award: 2013
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
There has been an increasing demand for low cost, portable CMOS image sensors because of increased integration, and new applications in the automotive, mobile communication and medical industries, amongst others. Colour reproduction remains imperfect in conventional digital image sensors, due to the limitations of the dye-based filters. Further improvement is required if the full potential of digital imaging is to be realised. In alternative systems, where accurate colour reproduction is a priority, existing equipment is too bulky for anything but specialist use. In this work both these issues are addressed by exploiting nanophotonic techniques to create enhanced trichromatic filters, and multispectral filters, all of which can be fabricated on-chip, i.e. integrated into a conventional digital image sensor, to create compact, low cost, mass produceable imaging systems with accurate colour reproduction. The trichromatic filters are based on plasmonic structures. They exploit the excitation of surface plasmon resonances in arrays of subwavelength holes in metal films to filter light. The currently-known analytical expressions are inadequate for optimising all relevant parameters of a plasmonic structure. In order to obtain arbitrary filter characteristics, an automated design procedure was developed that integrated a genetic algorithm and 3D finite-difference time-domain tool. The optimisation procedure's efficacy is demonstrated by designing a set of plasmonic filters that replicate the CIE (1931) colour matching functions, which themselves mimic the human eye's daytime colour response. The best designs were fabricated and demonstrated a least-mean-square error, in comparison to the desired colour matching functions, of 6.37*10^3, 2.34*10^3 and 11.10*10^3 for the red, green, and blue filters respectively. Notably the spectrum for the red filter contained a double peak, as present in the corresponding colour matching function. Such dual peak behaviour cannot be achieved using a single current dye-based filter. The filters retain the same layer thickness for all structures so they can be defined in a single lithography step. A new approach to enable the fabrication of a multispectral filter array on a CMOS imager is also presented. This combines a Fabry-Perot filter with effective medium theory (EMT) to enable the fabrication of multiple filters in a single cavity length via lithographic tuning of the filter passband. Two approaches are proposed; air-filled nanostructures and dielectric backfilled nanostructures. The air-filled approach is demonstrated experimentally producing three filters with FWHM of 60 - 64 nm. Using the backfilled design, and incorporating a highindex cavity material, a set of twenty three narrowband filters, with a FWHM of 22 - 46nm is demonstrated. A virtual image reproduction process was developed to quantify the image reproduction performance of both the plasmonic and Fabry-Perot filter sets. A typical rgb dye-based filter set used in conventional imagers achieves a mean colour error of 2.711, whereas the experimental data from the plasmonic filters achieves an error of 2.222 which demonstrated a slight improvement in colour reproduction. The multispectral filter set developed in this work performed even better, with 4 filters giving an error of 0.906, 10 filters an error of 0.072 and continued improvement in the colour error reaching 0.047 for 23 filters. All the filter sets proposed are fully compatible with the CMOS process so as to enable direct integration onto CMOS image sensors in industrial foundries in future. The performance of the presented filters also suggest new compact applications in art reproduction, agricultural monitoring and medical imaging.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.578586  DOI: Not available
Keywords: TK Electrical engineering. Electronics Nuclear engineering
Share: