Decomposition of Phosphine-Stabilized Metathesis Catalysts by Lewis Donors

Title: Decomposition of Phosphine-Stabilized Metathesis Catalysts by Lewis Donors
Authors: McClennan, William
Date: 2016
Embargo: 2018-11-02
Abstract: Olefin metathesis has had tremendous impact on synthetic approaches to the formation of new carbon-carbon bonds. Heterogeneous metathesis catalysts have been used in industry for decades, to effect redistribution of olefin chain lengths in petrochemicals processing. Only recently, however, has olefin metathesis emerged in pharmaceutical and specialty chemical manufacturing. The nearly 20-year gap between the discovery of easily-handled ruthenium catalysts and industrial implementation in these sectors is a result of many factors. One key contributor is the limited understanding of decomposition mechanisms that limit the reliability of olefin metathesis. Poor catalyst selectivities and yields remain a challenge for industrial uptake of olefin metathesis. Much academic effort focuses on designing methods and new catalysts for catalyst separation and reuse. Exploration of catalyst decomposition pathways has seen much less study in comparison. The limited recognition of decomposition as a problem in academia is undoubtedly due to the tendency to use high catalyst loadings, which mask catalyst decomposition problems. The catalyst loadings commonly reported in academia need to be decreased by a factor of 100–1000 for industrial viability in many processes. This thesis explores the decomposition of metathesis catalysts bearing a phosphine ligand. Such Grubbs-type catalysts, particularly second-generation versions containing an N-heterocyclic carbene ligand, are the most widely used metathesis catalysts in current use. This study follows up on an earlier discovery from the Fogg group, which showed that pyridine and amine donors drastically accelerate decomposition of the Grubbs catalysts. For the first-generation Grubbs catalyst, decomposition of the resting-state methylidene complex RuCl2(PCy3)2(=CH2) Ru-6 was shown to proceed via nucleophilic attack of PCy3 on the Ru=CH2 bond, forming a σ-alkyl complex that was intercepted and characterized crystallographically. Further reaction led to liberation of the methyl phosphonium salt [MePCy3]Cl 2. Under the same conditions, the second-generation methylidene complex Ru-4a decomposes rapidly without any detectable σ-alkyl intermediate. In this thesis, decomposition of the important second-generation catalysts is shown to proceed via the methylidene-abstraction pathway. These studies centered on the methylidene complex RuCl2(NHC)(PCy3)(=CH2), where the NHC ligand is H2IMes (Ru- 4a) or IMes (Ru-4b). The short lifetime of the σ-alkyl complex was tentatively attributed to the ease of activation of a C–H bond on the NHC ligand (specifically, those on the mesityl o-methyl group). This intermediate could not be observed for Ru-4a, but could be observed for the IMes system Ru-4b. This is suggested to reflect the slower rotation of the H2IMes ligand about the Ru–NHC bond, which promotes C–H activation. Also examined is the ability of other Lewis donors to trigger this methylidene abstraction pathway during catalysis. Lewis donors are shown to greatly accelerate decomposition of a wide range of phosphine-stabilized metathesis catalysts. Remarkably, even weak donors such as water, ethers, alcohols, and nitriles (i.e. functionalities that are widespread among contaminants, functional groups, and “green solvents”) are shown to promote this pathway. Phosphine-stabilized catalysts are generally regarded as more robust towards such donor functionalities, and the detrimental impact of such weak donors has gone widely unrecognized. These findings have profound implications for catalyst choice and use. Finally, a deuterium-labelling study was undertaken, in which the mesityl substituents on the NHC ligand were fully deuterated, to confirm that the proton in the phosphonium salt 2 indeed originates in the mesityl methyl group. Of note, the energies of the PCy3 dissociation and C–H activation steps were found to be closely similar. This was unexpected, given prior evidence that PCy3 loss is rate-determining in the Ru-2 precatalysts, and the much stronger Ru–PCy3 bonding known to be present in the methylidene complexes. This finding highlights the need to impede C-H activation for phosphine-stabilized metathesis catalysts. More broadly, it underscores the risks inherent in the use of phosphine-stabilized catalysts, or indeed of catalysts that are stabilized by a donor group that can function as both a Lewis base, and a nucleophile.
CollectionThèses, 2011 - // Theses, 2011 -