This thesis is split into three parts.
Part I: An Automatic Self-optimising Continuous-flow Reactor for Electrosynthesis
Assembly-line and/or continuous, steady-state strategies are widely spreading in the manufacturing industry. Electrosynthesis in an automatic continuous flow manner is drawing more and more attention to both pharmaceutical industry and research. This part describes the develop and use of an automated self-optimising continuous-flow electrochemistry reactor system. The main focus of this work is to develop an bi-language (MATLAB and LabVIEW), server-client structure automation software framework to conduct automation, control and monitoring of flow electrochemistry processes, which enables quick system setup, reconfiguration and high flexibility. Stable noisy optimisation by branch and fit (SNOBFIT) and simplex algorithm (modified simplex and super-modified simplex method) were developed and tested on simulators, which were then applied to the synthesis action of methoxylation of N-formylpyrrolidine and electro-oxidation of 3-bromobenzyl alcohol. Searching of the optimum operation condition of a reaction in an automatic manner is a major step forward to establish convenient and straightforward use of organic electrosynthesis in routine laboratory synthesis or industrial applications.
Part II: Applying FTIR Imaging to Address Challenges in Plastic Recycling
Plastic pollution is ubiquitous throughout the earth, and reusing/recycling of plastic has the potential to reduce the global abundance and weight of waste plastics. The main focus of this work is investigating plastic sample using Fourier transform infrared spectroscopy (FTIR), single point and imaging, for recycling purpose. An quantitative calibration of talcum concentration in talcum reinforced virgin polypropylene sample with IR peak ratio/integration was conducted, and the application of the result to analyse the talcum disperse in polypropylene matrix was reported. Micron scale FTIR imaging was conducted on the film sample. Pseudo-colour image visualising the distribution of talcum in the polypropylene matrix indicated a highly uneven distribution, a result of the reprocessing method.
FTIR imaging was applied to investigate the composite structures of 'real-world' composites sample for recycled industry plastics, including: virgin polypropylene with short milled recycled carbon fibre, virgin polypropylene, maleic anhydride grafted polypropylene with carbon fibre, acrylonitrile butadiene styrene with calcium carbonate and virgin polypropylene with poly(ethylene terephthalate) on the micron scale. Imaging technique in FTIR spectroscopy not only provide micron level spatial information of the composition but also direct solution of improving the inter-facial interactions between compositions, thus improving the physical/chemical performance of the plastic products. Those pilot studies provide insights into applying FTIR imaging for plastic sample analysis.
Part III: FTIR spectroscopy for Breast Cancer Prognosis
Breast cancer is a major cause of deaths for females worldwide. Cancer prognosis provides a patient's likely outcome based on their current standing, which can help to decide the treatment for the patient. The current golden standard prognosis index, Nottingham Prognosis Index, is a time-consuming, un-objective process to which limited confidence can be assigned because of inherent operator variability. Applying Fourier transform infrared spectroscopy (FTIR) imaging to breast cancer tissue offers a non-destructive, label free tool for cellular and extracellular breast cancer tissue studying. In this work, we evaluate the prognostic ability of FTIR spectroscopy for identifying different grades of breast cancer. Different combinations of data pre-processing, feature extraction and unsupervised learning methodologies are explored. Spectrum quality control methods are applied to correct or minimise spectral problems, including high noise level, baseline offset and outlier. A multi-stage data analysis algorithm developed can provide statistical control over the breast cancer classification process and produce a precise cancer prognosis.
|Date of Award||8 Jul 2020|
- Univerisity of Nottingham
|Supervisor||Michael George (Supervisor) & Peter Summers (Supervisor)|
- Data mining
- Breast Cancer
- Software Engineering
- Automation and Optimisation