Grease is the word…

Many years ago, I heard an anecdote regarding a fortuitous discovery in the Sharpless lab. A graduate student had been awarded their PhD and the group were celebrating in the lab with bottles of red wine (I know- but they were different times). Someone in the group-who had obviously consumed a sufficient quantity of the red elixir- decided it would be a good idea to run a metal-catalysed epoxidation reaction using red wine as solvent. The major research focus of the team at this time was stereoselective olefin oxidation. The storey goes that the reaction was somewhat successful and more significantly the epoxide product was produced in very good enantiomeric excess. The hypothesis was that tartrates in the red wine acted as a ligand for the titanium and the inherent stereochemistry in the ligand backbone controlled the epoxidation facial selectivity. In another version of this tale the wine-infused cork from a spent red wine bottle was rammed into the neck of a reaction flask and the tartrate salts leached into a stirring reaction mixture. As I’m sure your aware, tartrate ligands became the vital component for what became known as the Sharpless asymmetric epoxidation reaction.1

I confess I don’t really believe either of these stories- but wouldn’t it be great if there was some truth in it?

Why am I going on about this? Well maybe we can run chemical reactions using other non-conventual materials as solvents. And if these materials are cheap, non-toxic and easily obtained we could be on to a winner- certainly from a green chemistry perspective.2 Okay- you’ve waited long enough. Someone’s published a Suzuki Reaction run in vegetable oils, fish oils, butter, and beeswax (as discrete reactions of course).3

First some statistics. In 2020, the production of organic solvents surpassed 28 million metric tons, placing a significant burden on environmental health and safety. In the pharma industry solvents can exceed 80% of the total mass of materials used to make the API. Generally, this is something industry is aware of and many companies have invested heavily in trying to reduce the solvent burden for product manufacture, implementing recovery and recycling technologies.However, using mixtures of solvents in the same process can cause problems in separation and recovery. Mixed solvents may end up on the road to hell- incineration.

Use of lipid solvents for extraction of bioactive compounds from organic plant-based materials is well known. Many drug formulations rely on a lipid-based recipe to maintain drug substance solubility, and the use of micellular-type systems as replacements for polar aprotic solvents has been developed by the Lipshultz group and implement on large scale by Novartis.5

Figure 1: triglyceride consisting of glycerol, palmitic acid and a-linolenic acid

The new paper I hinted at above by Gevorgyan and his team at UiT Arctic University of Norway describes the use of lipid-based procedures for homogeneous transition-metal catalysed cross-coupling.3 A first thought might be that a very greasy oil might not be the best solvent for a polar aryl heterocycle or some such substrate, however the team claim the polarity can be adjusted by addition of phospholipids or other amphiphiles derived from vegetable oils (Figure 1).

Initial studies were carried out using a standard Suzuki cross coupling reaction (Figure 2, Pd(PPh3)4, KF, 80°C, 6hrs). As with optimising any reaction a robust analytical method is required from the outset. This proved challenging due to the physical properties of the oil, however an NMR screening technique using trimethoxybenzene as an internal standard enabled rapid screening of a wide range of oils, most of which seemed to perform as well as the “standard solvent” controls (MeTHF, acetal, dioxane, toluene, DMF).

Figure 2: “solvent” screen for a standard Suzuki reaction (% yield by NMR with 1,3,5-trimethoxybenzene as internal standard)

 

The optimised reaction performed well using activated substrates (Figure 2). For unactivated aryl halides a DavePhos/Pd2dba3 precatalyst system (using rapeseed oil as solvent) restored a high yield (e.g. 4-florophenyl boronic acid / 1-bromo-3,5 trifluomethylbenzene: Pd tetrakis 42% yield, Pd2dba3 / DavePhos 94% yield).

Rapeseed oil became the focus of further studies- readily available and edible (or should I say non-toxic and non-hazardous). Oil from various sources and brands were compared head-to-head and showed consistent yields suggesting that differences in production methods have little influence on the performance of rapeseed oils from diverse sources as solvents for the cross-coupling reaction. Waste rapeseed oil also appeared to work.

The paper showed a wide scope of examples- mostly simple aryl halides and boronic acids. “Difficult” substrates were not described- a step too far perhaps?

Heck-type couplings proved more difficult- most likely because vegetable oils and related lipids possess a number of internal double bonds which can potentially undergo heck-type coupling reactions giving side products. That said, unsubstituted styrenes, acrylates and terminal olefins could be coupled with aryl halides using tBu3PHBF4/Pd2dba3 as precatalyst combined with Bu4NOAc as base. Terminal olefin substrates resulted in formation of isomeric internal oilefins. Unactivated aryl halide substrates required an increase in catalyst loading- 2mol% v’s 0.5 mol% for activated substrates (Figure 3).

Figure 3: Heck couplings in rapeseed oil

Surprisingly, saponification of the ester-based oils was not a problem unless water was presence or strong inorganic bases. Carbonates, fluoride, acetates and amines under anhydrous conditions were all tolerated.

Non-polar products could be isolated from oils using column chromatography. In some ways use of hydrocarbon solvents such as heptane as an eluent negates the use of a sustainable oil-based solvent replacement as reaction solvent. To address this the team investigated the use of natural terpenes as eluents. -pinene, 3-carene, (R)-(+)-limonene, g-terpinene and sabinene were successful as eluents for column chromatography.

The third approach involved saponification of the oil followed by extraction of the product. This was problematic in terms of functional group stability (esters, nitriles etc) and would potentially be a significant problem on any kind of scale.

There we have it- a green approach to cross-coupling chemistry with an unusual solvent. Some of it I think is a bit of a stretch- the isolation stage in particular and using a terpene as a chromatography eluent (how easy is it to remove traces from your product?). When your chemistry has hit a wall and the chips are down, try (fry?) using oil as solvent- it might just help….

 

References:

  1. The first practical method for asymmetric epoxidation, K. B. Sharpless et al, J. Am. Chem. Soc. 1981, 102, 5974-5976.
  2. Green and sustainable solvents in process chemistry, J. Hallett et al, Chem. Rev. 2018, 118, 747-800.
  3. Lipids as versatile solvents for chemical synthesis, A. Gevorgyan et al, Green Chem. 2021, 23, 7219-7227.
  4. Systems level roadmap for solvent recovery and reuse in industries, K. Yenki et al, iScience 2021, 24, 103114.
  5. For example: a general and practical alternative to polar aprotic solvents, F. Gallou et al, Org. Process Res. Dev. 2016, 20, 1388-1391.