Novel catalysts for the hydroxymethylation of allyl alcohol : a convenient synthetic route to 1, 4-butanediol

Abstract

Hydroxymethylation catalysis provides a valuable strategy for the high volume production of alcohols from α-alkenes. Generally this involves a hydroformylation-hydrogenation sequence, but the capacity to optimise selectivity for each transformation is limited. Condensation reactions between aldehyde products and alcohol products frustrate process economics. By an alternative scheme, all relevant bond-forming reactions occur in a single mechanism. This thesis describes several approaches to catalyst development and the application of derived systems for the hydroxymethylation of allyl alcohol. A review of auto-tandem hydroxymethylation and domino hydroxymethylation is presented in Chapter 1.

In Chapter 2 the synthesis of bis-(diethylphosphine) ligands based on a modular series of chiral alicyclic scaffolds is described. High pressure NMR studies have shown that the catalytically active complex [RhH(CO)₂(L-L)] adopts preferentially ea geometry, with [Rh(CO)(L-L)(μ-CO)]₂ as the primary competing species. Catalyst performance can be correlated with the flexibility of the chelating ring; this favoured a high monomer/dimer ratio which enhances activity, but could not rigidify the configuration of the diethylphosphine groups which inhibits linear selectivity. Deuterium labelling studies were suggestive of a domino hydroxymethylation scheme. From the rhodium-hydroxyalkyl-hydride-carbonyl cation, a reductive elimination furnishes the diol derivatives and a β-hydride abstraction furnishes the hydroxyaldehyde derivatives. Up to 53 mol% selectivity to 1, 4-butanediol was attained. The catalysts could be recycled via biphasic separation, however poisoning by methacrolein caused a decline of activity upon reuse of the solution.

An investigation of enhanced specific activity via the meta-effect is the subject of Chapter 3. The effect of systematic meta-substitution in triphenylphosphine upon physicochemical properties was investigated by IR spectroscopy and electrochemistry, both of which showed no significant structural impact on the uncoordinated triarylphosphine. Variable temperature ¹H NMR studies however revealed a change in the solution dynamics of the corresponding Vaska complex. The activation barrier to phosphorus-(ipso)carbon rotation increases as a function of meta-substitution, with rotation of substituted aryl rings past each other being more strained. This should create a well-defined coordination sphere around rhodium, and is proposed to account for the high linear selectivity observed in the hydroformylation of allylic alcohols with [RhH(CO){(3, 5-Me₂Ph)P}₃]. Linear-selectivity reached 96 mol%. Catalyst recycling was executed via biphasic separation, retaining on over twelve cycles an average of ~ 94 % efficiency. The kinetics of allyl alcohol hydroformylation with [RhH(CO){(3, 5-Me₂Ph)P}₃] was found to be well represented by Equation 11 (Section 3.6)

A detailed analysis of how substrate-specific the influence of the meta-effect remains to be performed.

In Chapter 4 domino hydroxymethylation by multi-component L-L/PEt3/Rh systems is described. The regioselective performance of a diphosphine rhodium catalyst in hydroformylation was translated for hydroxymethylation upon introduction of triethylphosphine at a L-L/PEt3 molar ratio ≥ 1. The highest observed selectivity to 1, 4-butanediol was 66 mol%. Competitive activity of triethylphosphine-modified rhodium species presumably accounts for the reduced linear selectivity observed when L-L/PEt3 molar ratio < 1. Despite aggravated catalyst decomposition at higher triethylphosphine concentrations, heterogeneous hydrogenation does not appear to take place. Deuterium labelling studies also discount a sequential homogeneous hydrogenation. There is evidence for the activation of a tris-phosphine-modified rhodium-acyl-carbonyl complex, but such a species could not be isolated from complexation reactions with a variety of precursors. It would be of interest to determine alternative promotors and to establish whether it is preferential to employ a high concentration of mildly acidic species or a low concentration of highly acidic species.

The self-assembly of DNA base pair analogues 2-N-pivaloylaminopyridyl phosphine and isoquinolyl phosphine, each modified with diphenylphosphine, diethylphosphine, dicyclohexylphosphine and bis(3, 5-dimethylphenyl)phosphine, is described in Chapter 5. In the presence of a rhodium precursor, exclusive formation of the heteroleptic complex was observed. Although the intramolecular hydrogen-bonding network is sensitive to temperature and free hydroxyl functionalities, highly regioselective catalysts were generally afforded under the appropriate operating conditions. Only the catalyst based on the bis(dicyclohexylphosphine)-heterodimer performed poorly, presumably due to the formation of mono-phosphine complexes. High chemoselectivity was correlated with the heterodimer acidity constant, however this is rendered non-linear by a trans influence when electronic distinction between the platforms is high. Overall, complexes based on the assembly of a dicyclohexylphosphine platform and a bis(3, 5-dimethylphenyl)phosphine platform were found to be optimal; up to 73 mol% selectivity to 1, 4-butanediol was reached.

It has been demonstrated in this thesis that in order to effect linear-selective domino hydroxymethylation of allyl alcohol, two distinct transition state structures must be optimised. High regioselectivity demands an asymmetric rhodium-hydride-dicarbonyl complex, which can be generated by an asymmetric chelate or by rigidifying the configuration of the substituents on phosphorus. Interestingly, chelation geometry in this transition state has little impact on this parameter. It has been shown that domino hydroxymethylation is activated by an electron-rich rhodium-acyl-dicarbonyl. The state of electron density on rhodium can be controlled by the substitution pattern on the phosphorus donors, but can also be changed by the inclusion of a suitable promoter. The chelation geometry in this transition state is more significant; placing the acyl functionality trans to a phosphorus donor concentrates the electronic effect in the rhodium-alkyldiol-hydride-carbonyl cation to such an extent as to impede hydride migration and reductive elimination of the diol, favouring β-hydride abstraction and reductive elimination of the hydroxyaldehyde.