Safer solutions

Ionic liquids could provide a safer alternative to organic solvents in manufacturing processes, as Dr Sarah Houlton discovered

The vast majority of processes involved in the synthesis of pharmaceutical ingredients and intermediates use some form of organic solvent.However, organic solvents are often volatile, creating potential emission problems, and flammable, making process safety an issue. Toxicity is a cause for concern, and pharma ingredients have to be free from all traces. Waste disposal has to be considered, too.

One alternative attracting a great deal of attention in the past year or two is ionic liquids. These have no vapour pressure and are not flammable, so the emissions and process safety problems of organic solvents are eliminated. Despite the lack of toxicity data, ionic liquids have great potential as solvents for organic reactions, and in addition to a vast amount of interest from the academic community, several companies have been working on their application in organic synthesis.

An ionic liquid is a salt that is molten at low temperature. Common salts like sodium chloride are solid at room temperature because the small, regularly sized positive and negative charged ions pack together very efficiently, so a large amount of energy is needed to break them up. Ionic liquids contain ions that fit together much less well, so their melting points are significantly lower because less energy is needed to break up the lattice.

The properties of the ionic liquid, such as melting point, viscosity and density, depend on the precise nature of the ions. More complex systems, containing several different ion species, have been created in the past year or two, which mean the exact properties of an ionic liquid can be tuned.

Another feature is that ionic liquids can affect the course of the chemical reaction, depending on the properties of the ions they contain and how they interact with the substrates, reagents and catalysts that are involved. All these factors mean that they have great potential as media for carrying out chemical reactions in a cleaner, more efficient manner than conventional organic solvents.

ugly cations

As Professor Ken Seddon of The Queens University, Belfast, explained at a recent Royal Society of Chemistry conference on the commercialisation of ionic liquids, the archetypal cation that has been used to make ionic liquids is methyl imidazole, typically combined with PF₆^- or BF₄^-. 'But I don't think these will be the most used in the future,' he claimed.

The properties of the ionic liquid are very dependent on the chain length of the organic species in the ions, with the melting point dropping as it gets more unsymmetrical, before rising again as van der Waals forces get larger as the chains get longer. 'We need an "ugly" unsymmetrical cation,' he said. 'While it's not essential, it does make the liquid more fluid.'

Ionic liquids have a number of properties that make them potentially useful for carrying out chemical reactions. As well as the liquid range, they are excellent solvents for organic, inorganic and polymeric materials. While some are water sensitive and must be used in a dry box, many are hydrophobic and air stable. Acidic compositions are superacids, with a pKa of around -20, giving interesting possibilities for reaction mediation. They are frequently thermally stable up to 200°C.

Furthermore many are now commercially available, some on a large scale. 'I don't think we have yet found an insoluble organic,' Seddon claimed. 'Inorganics invariably dissolve, too, but not those with huge lattice structures. This does mean that the steel, glass, PTFE etc of reaction vessels doesn't dissolve.'

Perhaps the main reason for the recent interest in ionic liquids is their potential for 'green' chemistry. In addition to the above advantages, they can be highly solvating and thus low reaction volumes are needed. They can act as catalysts as well as solvents, and give highly selective reactions. As so many ionic liquids are possible - current estimates are that there could be as many as 10¹⁸ ternary mixtures. 'This puts the number into the combinatorial range,' Seddon explained. 'There are advantages and disadvantages because there are so many. But it presents a new paradigm, too - you can use the solvent to design the chemistry.'

A huge challenge facing chemists trying to establish reactions in ionic liquids is working out precisely which anion and cation to use. 'The anion controls the chemistry, while the cation controls the liquid's properties,' Seddon explained. 'There are around a million (10⁶) simple ionic liquids, 10¹² secondary ones, and 10¹⁸ ternary systems.

'Despite this, maybe half the publications that have appeared use R-methylimidazole hexafluorophosphate - the Antichrist of ionic liquids,' he claimed. 'Although it is air stable, it rapidly absorbs water from the atmosphere, and is hydrolytically unstable, generating five moles of HF for every mole of ionic liquid, which is not good. It will never be used in industry. So why do academics use it? Because it's cheap - and everyone else does! One chemical reason is that PF₆^- is octahedral, and very highly non-hydrogen bonding, so it is good for studies. The same is true for tetrafluoroborates and they hydrolyse 50 times faster.'

Chris Adams, of Llangollen-based XeF₆, pointed out that there has to be a very strong economic driver for an existing process to be re-engineered using an ionic liquid.

He quoted figures from BP that savings had to be in excess of 10% for a high volume process, but for a fine or speciality chemicals process, the economic realities are much more stark. 'Experience shows that a new process has to demonstrate 30-50% cost savings for implementation to succeed if there is no product improvement.'

potential issues

He also cited several potential issues that could affect commercial success. For example, if less than 100% conversion and selectivity are achieved, this is likely to add to process complexity. Intellectual property could be a problem too - how can protection be secured completely when there is an infinite number of alternative ionic liquid compositions that could also do the job?

The first commercial process to use ionic liquids was introduced last year by BASF. The biphasic acid scavenging utilising ionic liquids (BASIL) process is used for the production of an alkoxyphenylphosphine. In the original process, triethylamine was used to scavenge the acid that was formed in the course of the reaction, but this made the reaction mixture difficult to handle. Replacing triethylamine with methylimidazole results in the formation of an ionic liquid, which separates out of the reaction, making the process much easier to run. It is now carried out at a multi-tonne scale, proving that handling large quantities of ionic liquids is possible.

BASF is also pursuing ionic liquids in other processes. Another under development is nucleophilic HCl. 'This is normally done with phosgene, because of the side reactions that take place with HCl,' research scientist Matthias Maase explained. 'But in an ionic liquid, you can get a very selective reaction with HCl.' BASF has gained 98% selectivity, and the process is at an advanced stage of development and is expected to be introduced on a large scale in the second half of 2004.

A third example, using ionic liquids to break azeotropes, is also at an advanced stage and he hopes it will be available in the near future. The company is launching its range of commercial ionic liquids at Chemspec.

Alan Curzons, director of new technologies in GlaxoSmithKline's corporate environment, health and safety group, explained that solvent directives can limit what can be used in processes. Pharma manufacturing is highly regulated by government agencies, particularly regarding process changes, and the use of recovered or recycled solvents. Route and process changes after approval are costly - so there have to be very good reasons for any alterations. Early route definition is necessary, but the high attrition rate in pharma means that early stage effort in developing processes is expensive.

He described the metrics principles that GSK uses for green chemistry. 'These have to be simple.' he said. 'We score for a number of different categories, including waste, impact, health, safety and lifecycle.' Waste involves looking at issues such as recovery and recycling, plus disposal issues.

The main issue for impact is bioaccumulation, while he claimed that health is not merely exposure. 'Ideally, plants should run in a "shirt-sleeves" environment, but ionic liquids are often eye irritants,' he said.

safety issues

Safety issues include flammability and compatibility with other chemicals, while lifecycle means looking at cradle to grave impacts for manufacture.

Green issues are not the only ones that must be considered when selecting an ionic liquid. Drug actives are frequently polar, so isolating the product and removing traces of ionic liquids have to be considered. If organic solvents are needed to extract the product from the ionic liquid, this could negate the perceived benefits of using the ionic liquid in the first place.

Significant loss of ionic liquid into the organic phase is possible and this is generally not feasible with solvents with low polarity anyway. Distillation is rarely an option; direct crystallisation is much preferable.

'The ICH guidelines are very strict to stop solvents contaminating drugs,' Curzons said. 'This is a particular issue if an ionic liquid is to be used in the final stage of manufacturing the active.'

He added that GSK has seen ionic liquid residues in products by using mass spec. 'I would like to see life cycle analysis done,' he said. 'And how do we dispose of them if they don't biodegrade? We are unlikely to be able to reuse them in the pilot plant. The pharmaceutical industry has to demonstrate the quality and stability of its products.

'This is required for regulatory approval, but it takes time and is expensive. I would like to see telescoping - the same solvent used over several steps, which is cheaper to manage. At the moment, ionic liquids are expensive and there are thousands to choose from; organic solvents are cheap, and there are only maybe 10 choices.'

So what would it take for ionic liquids to be used routinely in the pharmaceutical industry? The availability of a limited number of ionic liquids with broad applicability to a significant number of different processes, with established environment, health, safety and operational data, would be a distinct advantage.

One of the main suppliers of ionic liquids is Merck KGaA. 'We don't think there will be a one size fits all ionic liquid,' said the company's Will Pitner. 'I believe there will be ionic liquids tailored for specific reactions.'

The key issue facing the commercialisation of ionic liquids is the lack of toxicity information. If quick, relatively cheap environmental screening were possible, then ionic liquids are more likely to be used, particularly in the pharma sector. Professor Bernd Jarstorff at the University of Bremen in Germany is running an extensive programme looking at just this issue, working with Merck. Initial results look promising,¹ as they indicate they have low toxicity, ecotoxicity and biodegradability. His results also suggest that long, aromatic side-chains are best avoided, while the ecotoxicity is on a parallel with toluene and the oral toxicity in rats is around 1400mg/kg.

Purification of ionic liquids is also an issue, Pitner claimed. 'They can't be recrystallised, distilled or sublimed,' he said. 'The majority contain halides, water and organic starting materials, especially unreacted base. These can all wreak havoc on chemical processes. They can also give undesired reactions, and poison catalysts. Halides are usually less of a problem if you are not using a catalyst, for example.' The company offers 250 different ionic liquids, and others are available on request by custom synthesis.

South African company SASOL has been investigating the potential of ionic liquids in metathesis. This metal catalysed reaction of two olefins, figure 1, is an increasingly popular synthetic technique, in which alkylidene fragments are exchanged between two olefins.

Ideally, said the company's Mike Green, the reaction should be homogeneous. The reaction is very selective, with little or no isomerisation. Operating conditions are mild, from ambient temperature up to about 80°C, and high activity and long life can be obtained from the catalysts, in this case ruthenium, which is cheaper than either rhodium or palladium.

Ionic liquids offer potential advantages by making catalyst recovery easier, product separation simpler, and presenting the possibilities of enhanced catalyst activity and better selectivities.

Green gave the example of the self-metathesis reaction of 1-octene to give ethylene and 7-tetradecene. 'Cross-metathesis is a big potential problem here,' he said. 'There is usually no need to use a solvent as the catalyst is soluble in the neat olefin, but as the temperature is raised, we get advantages with ionic liquids over the olefin alone. While conversion is not complete, it is better and faster. The selectivity is also slightly better, with fewer secondary metathesis reactions.

'This is significant if you want a successful industrial process. Clearly, the ionic liquid is interacting with the catalyst in some way. And we've seen some very promising preliminary results. There are no doubts whatsoever that these would justify a process.'

Novartis has also been investigating the possibilities of ionic liquids. Ernst Kuesters explained that one area of investigation - methylation - has great potential as ionic liquids could remove the need for toxic reagents. The company has been looking at dimethyl carbonates as methylating reagents, and the results, particularly when combined with microwave irradiation, are very promising. It has also found that, with an ionic liquid and microwaves, benzylation using dibenzyl carbonate, slow reactions can be accelerated from days to minutes.

While industrial applications of ionic liquids are beginning to appear, they still have the perception of being academic curiosities. Queen's University's QUILL has published a map for industrial innovation, in conjunction with several industrial partners, which sets out the basics about the properties of ionic liquids and their manufacture, and explains the strategic drivers that are needed for them to become accepted in different sectors.

'The estimated timescale from the lab to implementation is three to four years,' said BP's Martin Atkins. 'There are major benefits to be gained from working across the disciplines and boundaries of science and technology. Catalysis and reaction engineering will build the bridges.'