Mahmood, Irfhan
(2025)
Exploring self-optimisation methods for hydrogenation reactions in supercritical CO2.
PhD thesis, University of Nottingham.
Abstract
The work described in this Thesis investigates the use of self-optimisation methods to improve the efficiency of hydrogenation reactions in supercritical CO2 (scCO2) in continuous flow. Chapter 1 introduces the concepts of flow chemistry, emphasising its advantages. An overview of various optimisation methods used in flow chemistry, with a focus on recent advancements in self-optimisation methods is outlined. There is a brief description of supercritical fluids and discussion of previous work related to hydrogenation reactions in scCO2. The Chapter concludes with the Aims of the Thesis.
Chapter 2 details the development of an automated high temperature and high pressure reactor designed for the self-optimisation of hydrogenation reactions in supercritical CO2. The chapter begins by outlining the reactor’s components, which were based on previous iterations of similar reactors used by the research group. This reactor design incorporates various optimisation algorithms and utilises an on-line Gas Chromatography (GC) system for monitoring the reactions. The reactor was commissioned with a test reaction of hydrogenation of isophorone with a palladium based catalyst (up to 800 mg) in scCO2. The hydrogenation of isophorone was successful with the formation of a single product 3,3,5-trimethylcyclohexanone (TMCH), with the optimum yield of >99% achieved.
In Chapter 3, the reactor was used to investigate the hydrogenation of isophorone for self-optimisation, where reactant was pumped into the system, and the reaction conditions were optimised automatically for the product (TMCH) yield by varying different parameters such as temperature, pressure, CO2 and isophorone flow rates, and H2:isophorone ratios. Three different optimisation algorithms, Super Modified Simplex (SMSIM), Stable Noisy Optimisation by Branch and FIT (SNOBFIT) and Bayesian, were used, all producing comparable results for the hydrogenation of isophorone. When comparing the performance of these algorithms at different catalyst loadings (400, 200, and 100 mg of 2% Pd/type31 SiO2-Al2O3), the Bayesian algorithm outperformed the others by requiring the fewest experiments to achieve the optimal yield of TMCH.
In Chapter 4, a new accelerated self-optimisation method was developed using kinetic modelling, applied to the same model reaction of hydrogenation of isophorone. Kinetic models were employed to predict the optimal conditions, which then guided the optimisation algorithms more efficiently. This approach reduced the number of required experiments to obtain the optimum by more than 50% compared to the previous self-optimisation method, resulting in significant savings in time, materials, and resources, while also minimising waste production. Further improvements in efficiency were achieved through the integration of in-line ReactIR, which offers a much faster sampling rate compared to on-line GC, contributing to a more sustainable self-optimisation process for the hydrogenation of isophorone. In this case, the SMSIM algorithm performed best, requiring the least number of experiments to reach the optimum when guided by the kinetic model.
In Chapter 5, the substrate scope for the self-optimisation of hydrogenation reactions was expanded to include x,β-unsaturated carbonyl compounds, specifically x-methyl-trans-cinnamaldehyde. The initial strategy focused on selecting a suitable catalyst capable of selectively hydrogenating the starting material into the desired products, 2-methyl-3-phenylpropanal and 2-methyl-3-phenyl-2-propen-1-ol, by precisely optimising reaction parameters. Initial Design of Experiment (DoE) studies were carried out using 5% Ru/Al2O3 to determine the conditions required for forming each product. These DoE results then informed the Bayesian algorithm for self-optimisation. The optimal conditions for high yields of each product differed, with milder conditions required for selective hydrogenation of the aldehyde functional group in x-methyl-trans-cinnamaldehyde compared to the alkene functional group. Through self-optimisation, the system successfully optimised the production of the desired products using a single catalyst. To improve on the yield for the hydrogenation of x-methyl-trans-cinnamaldehyde towards the 2-methyl-3-phenyl-2-propen-1-ol product, catalytic transfer hydrogenation (CTH) was explored using ethanol as both the solvent and hydrogen donor, in the presence of acidic alumina as the catalyst. This approach achieved 100% selectivity and a 97% yield of the desired product, with no by-products formed.
Chapter 6 of this Thesis outlines details of the experimental work conducted. Chapter 7 summarises the Thesis, evaluates the success of the techniques and approach outlined in the aims of Chapter 1, and offers suggestions for future research based on the findings.
| Item Type: |
Thesis (University of Nottingham only)
(PhD)
|
| Supervisors: |
George, Michael W.
Poliakoff, Martyn
Hirst, Jonathan D.
Graham, Richard S. |
| Keywords: |
alkylation, heterogeneous catalysis, reaction optimization, supercritical fluids, sustainable chemistry |
| Subjects: |
Q Science > QD Chemistry > QD 71 Analytical chemistry |
| Faculties/Schools: |
UK Campuses > Faculty of Science > School of Chemistry |
| Item ID: |
80616 |
| Depositing User: |
Mahmood, Irfhan
|
| Date Deposited: |
30 Jul 2025 04:40 |
| Last Modified: |
30 Jul 2025 04:40 |
| URI: |
https://eprints.nottingham.ac.uk/id/eprint/80616 |
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[thumbnail of Corrections]](https://eprints.nottingham.ac.uk/style/images/fileicons/application_pdf.png "Corrections")