Search

The continuous catalytic transfer hydrogenation of benzonitrile to benzylamine is demonstrated using a palladium on carbon catalyst with triethylammonium formate as reducing agent. Solvent choice was critical in overcoming rapid catalyst deactivation. A 15-fold increase in catalyst productivity was observed in flow compared to batch, which was achieved using an ethanol–water solvent in combination with intermittent catalyst regeneration by washing with water.

This forms part of our Humans of AI3SD Series.In this Humans of AI4SD interview he discusses data-driven approaches to mechanistic investigation, communicating across disciplines, the lack of progress with AI in synthetic chemistry, and his advice to early career researchers.

Predicting an optimal set of conditions for a given reaction is a challenging task. This becomes even more challenging when also trying to predict the performance of a reaction, such as the final yield of product or how long it will take. Existing approaches to this problem use machine learning algorithms that are fed a large volume of single timepoint yield data (the amount of product generated after a set time). Using this data is problematic; if a reaction only reached 50% would it reach 99% if run for longer? How much longer? Was there a slow catalysts activation period followed by a rapid reaction? Or was a catalyst poisoned, stopping the reaction midway? These are factors described by the reaction kinetics. We propose that using reaction kinetic data will lead to better predictive models. However, collecting kinetic data is significantly more costly and time consuming than collecting single timepoint yields. To minimise these issues, we will be using a cutting-edge machine learning technique called active learning. Rather than performing every possible reaction combination to build our model, active learning first picks the most important set of conditions. The user collects this data, and theactive learning process is updated and begun again, resulting in a highly cost-efficient procedure minimising experimental requirements.

The difficulty in accessing and safely utilising non-stabilised diazo species has in the past limited the application of this class of compounds. Here we explore further the use of oxadiazolines, non-stabilised diazo precursors which are bench stable, in direct, non-catalytic, aldehyde C-H functionalisation reactions under UV photolysis in flow and free from additives. Commercially available aldehydes are coupled to afford unsymmetrical aryl-alkyl and alkyl-alkyl ketones while mild conditions and lack of transition metal catalysts allow for exceptional functional group tolerance. Examples are given on small scale and in a larger scale continuous production.

High power, phase-pure Nb-doped LiFePO4 (LFP) nanoparticles are synthesised using a pilot-scale continuous hydrothermal flow synthesis process (production rate of 6 kg per day) in the range 0.01–2.00 at% Nb with respect to total transition metal content. EDS analysis suggests that Nb is homogeneously distributed throughout the structure. The addition of fructose as a reagent in the hydrothermal flow process, followed by a post synthesis heat-treatment, affords a continuous graphitic carbon coating on the particle surfaces. Electrochemical testing reveals that cycling performance improves with increasing dopant concentration, up to a maximum of 1.0 at% Nb, for which point a specific capacity of 110 mAh g−1 is obtained at 10 C (6 min for the charge or discharge). This is an excellent result for a high power cathode LFP based material, particularly when considering the synthesis was performed on a large pilot-scale apparatus.

With a resurgence of interest at the turn of the millennium, organocatalysis has become an established field in chemical research. Proline, in particular, has played a central role in the profusion of research. As the archetypal organocatalyst, proline mediated transformations have acted as a point of reference against which the performance of new developments in the field are benchmarked and compared. The field as a whole has benefited from a timely merging of complementary experimental, kinetic, and computational studies, through which a comprehensive picture of reaction mechanisms has been drawn. The research documented in this thesis attempts to build on this successful multi-faceted approach to elucidate further mechanistic insights into organocatalysis. The proline mediated conjugate addition of aldehydes to nitroalkenes has been investigated via a ‘kinetics first’ approach, through the use of Reaction Progress Kinetic Analysis (RPKA) methodology. It was discovered that proline is irreversibly removed from the catalytic cycle via initiation of a polymerisation reaction of the nitrostyrene substrate. Additionally, it was found that this off-cycle pathway was not limited to proline alone, but also occurs with a number of secondary amine compounds, including other organocatalysts. The synthesis of two constrained bicyclic proline analogues was carried out in order to experimentally verify their computationally predicted behaviour in a benchmark aldol reaction. Excellent agreement between the experimentally determined and computationally predicted product enantioselectivity was found. Kinetic investigation of these bicyclic proline analogues revealed deactivation of the catalyst to be interfering with the catalytic cycle and so critically impairing their activity as compared to proline. The catalytic capability of these catalysts was explored, revealing universally poor behaviour in all but the reaction for which they were designed. An intriguing case of a lack of selectivity in the HPESW reaction was explored computationally and the decrease in stability of the anti transition state relative to the syn was determined as the cause. A brief exploration of the application of modern DFT functionals, M062X and ωB97XD, in a synthetically relevant setting was conducted and their extremely poor performance noted, compared to the excellent and highly accurate performance of the B3LYP functional. With the recent description of a new stereochemical and mechanistic regime for organocatalysts lacking a directing group, termed Curtin-Hammett kinetics, a kinetic investigation of the diphenylprolinol silyl ether mediated α-selenation reaction was undertaken to generalise this regime across a further carbonyl α-functionalisation reaction. A kinetic profile for the reaction was observed through the use of the RPKA methodology. The shape, an initial spike followed by zero order kinetics, closely resembles that of the conjugate addition and α-chlorination reactions. Further exploration revealed product selectivity to be a function of both the reaction solvent, correlated to dielectric constants, and the electronic properties of the aryl rings of the catalyst. An in-depth computational investigation of the conformational preferences of two important species on the catalytic cycle, the starting material and product selenoenamines, was undertaken in order to probe the underlying cause of the solvent-dependent reversal on selectivity. In agreement with experimental findings in the literature, the starting material enamine was found to have a high energetic preference for the anti E enamine in a number of solvents, dependent on the both the conformation of the pyrrolidine ring and the exocyclic CPH2OTMS group for each enamine C-N rotamer. This implies a solvent-dependent change in starting material enamine conformational preference is unlikely to explain the effect of solvent on product selectivity. Study of the product selenoenamine in various solvents found the relative thermodynamic stabilities of the E and Z enamines to change with solvent, with the Z enamine more stable in the gas-phase and non-polar solvents but little-to-no difference in stability between the E and Z enamines in polar solvents. This suggests that the experimental observation of the E enamine only in toluene is not thermodynamically preferred but is rather the result of an inherent kinetic preference. The above evidence supports the proposed hypothesis that solvent dependent product selectivity is the result of the erosion of kinetic preferences due to equilibration of relative intermediates in the catalytic cycle, as opposed to a change of enamine conformation in the stereogenic carbon-carbon bond forming step. Open Access

Controlling polymer composition starting from mixtures of monomers is an important, but rarely achieved, target. Here a single switchable catalyst for both ring-opening polymerization (ROP) of lactones and ring-opening copolymerization (ROCOP) of epoxides, anhydrides, and CO2 is investigated, using both experimental and theoretical methods. Different combinations of four model monomers - ϵ-caprolactone, cyclohexene oxide, phthalic anhydride, and carbon dioxide - are investigated using a single dizinc catalyst. The catalyst switches between the distinct polymerization cycles and shows high monomer selectivity, resulting in block sequence control and predictable compositions (esters and carbonates) in the polymer chain. The understanding gained of the orthogonal reactivity of monomers, specifically controlled by the nature of the metal-chain end group, opens the way to engineer polymer block sequences.

A continuous flow method for the direct conversion of alcohols to amines via a hydrogen borrowing approach is reported. The method utilises a low loading (0.5%) of a commercial catalyst system ([Ru(p-cymene)Cl2]2 and DPEPhos), reagent grade solvent and is selective for primary alcohols. Successful methylation of amines using methanol and the direct dimethylamination of alcohols using commercial dimethylamine solution are reported. The synthesis of two pharmaceutical agents Piribedil (5) and Buspirone (25) were accomplished in good yields employing these new methods.