Chemical companies are coming under increasing pressure to clean up their activities by finding alternative cleaner syntheses rather than by dealing with the after-effects.
Chemical companies are coming under increasing pressure to clean up their activities by finding alternative cleaner syntheses rather than by dealing with the after-effects, says Tim Lester.Cleaner processes could avoid costly waste treatments, Courtesy of Cleanaway.
When I was appointed technical adviser for cleaner synthesis on the Science and Engineering Council’s (now EPSRC’s) Clean Technology Programme four years ago, one of the first things I did was to carry out some literature searches. Using the keyword ’clean’ yielded almost nothing, but I suspect that the position would be very different today. The terms clean technology and clean (or cleaner) synthesis are heard much more frequently nowadays at gatherings of industrial chemists. Journal publishers have also seized on the opportunity - the Journal of Chemical Technology and Biotechnology now boldly states on its front cover that its coverage encompasses clean technology, while the Journal of Cleaner Production was launched in 1993. Clean technology forestalls pollution by circumventing waste production and minimising the use of energy ? avoiding the problem in the first place rather than treating the effluents. So where does cleaner synthesis fit in?
Cleaner synthesis involves making changes to the chemistry, biology, or engineering of the original process. It is just one of several waste minimisation options1 and may not always be the most appropriate; another strategy might be to change the product to one that serves the same purpose, but is cleaner to manufacture. Alternatively, carefully implementing ’good practice’ can lead to substantial reductions in discharges to water and land, as demonstrated by a recent project to clean up the rivers Aire and Calder.
For those involved in manufacturing chemicals, cleaner synthesis is particularly satisfying - making a product by a novel route that is both commercially advantageous and less of a burden on the environment. Many sub-disciplines of chemistry and process engineering can contribute to the technology at various levels of sophistication - from leading edge science to the innovative application of existing knowledge.
Chemical manufacturing processes vary enormously in their waste production characteristics. Commodity chemicals made from petroleum feedstocks are often produced by using catalytic processes with very high yields and selectivities. Byproducts can frequently be used in other processes or, as a last resort, as fuels to help run large integrated petrochemical plants. Nevertheless, even the petrochemical and refining industries acknowledge certain challenges, notably finding a selective route to ethene synthesis; currently only ca 30 per cent of the naphtha cracked product is ethene. It would also be useful to identify an alternative acid catalyst to replace HF or H2SO4 used for alkylating various gasoline components.
At the other end of the scale lies the synthesis of complicated biologically active molecules in the pharmaceutical and agrochemical industries, often requiring multistep procedures giving overall yields as low as 10 per cent.
Roger Sheldon has categorised sectors of the chemical industry by quantity of byproduct per kilogram of product (Table 1). Many of the byproducts that arise in these various industry sectors do so by stoichiometric reactions, side reactions or further reactions.
Table 1. Sectors of the chemical industry by quantity of byproduct per kg of product
|Industry sector||Product tonnage||kg by products/
kg of product
|Oil refining||106 - 108||ca 0.1|
|Bulk chemicals||104 - 106||<15|
|Fine chemicals||102 - 104||550|
|Pharmaceuticals||101 - 103||25100+|
A + B ____> C + D
Unless by some fortunate chance both C and D are useful products, stoichiometric reactions of this type produce a molecule of a waste for each molecule of product. Acids and alkalis are often reagents, with salts as the unwanted byproducts.
Other reactions that fit this type of scheme are AlCl3-catalysed Friedel?Crafts alkylations and acylations, which generate large quantities of aluminium wastes. Despite considerable advances towards replacing AlCl3 with other catalysts, such as zeolites and Envirocats, worldwide demand for AlCl3 is ca 75 000 t pa - much of it used to catalyse Friedel?Crafts reactions.
Researchers are also developing cleaner alternatives to oxidising agents based on chromium and manganese, which can give toxic metal byproducts that are subject to regulatory pressures.
We can take a quantitative look at the waste produced by different reactions using the concept of atom utilisation, calculated by dividing the molecular weight of the desired product by the sum of the molecular weights of all ’products’.
The manufacture of t-butylamine illustrates the concept. The conventional route is via the Ritter reaction:
HCN + H3O H3O
(CH3)2C = CH2 ___> (CH3)3CNHCHO ___> (CH3)3CNH2 + HCOOH
Based on these reactions the atom utilisation is 61 per cent, but the reaction suffers from salt formation during work up. However, BASF has developed a cleaner synthesis using zeolite catalysis, which demonstrates 100 per cent atom utilisation:
(CH3)2C=CH2 + NH3 ___> (CH)3CNH2
Stoichiometric reactions are common in work-up procedures to neutralise an acidic or basic medium that was used in a previous reaction step. Product work up and isolation can pose as great a constraint on clean synthesis as the synthetic reaction itself and careful reappraisal of a longstanding work up procedure can sometimes realise significant improvements.
A + B ___> C (desired reaction)
A + B ___> D (side reaction)
Aromatic substitutions are a well known example of this type of reaction, with their competition between ortho, meta and para sites. Nitrating systems that maximise the production of para-nitrotoluene are commercially attractive because the para-isomer costs about twice as much as ortho-nitrotoluene. Keith Smith at the University of Swansea recently reported5 producing ca 80 per cent of para-nitrotoluene using nitric acid and acetic anhydride to generate acetyl nitrate in the presence of zeolite beta.
Attempts to make particular enantiomers are also subject to side reactions as a consequence of imperfect stereoselectivity.
A + B ___> C ___> D
In these reactions C is the desired product, but the reaction is difficult to stop cleanly at this stage. During the manufacture of ethene oxide by oxidising ethene, ca 20 per cent of the starting material undergoes further oxidation to carbon dioxide and water. Another example is the production of methylene chloride by chlorinating methyl chloride, which also results in the over-chlorinated byproducts chloroform and carbon tetrachloride.
Fate of solvents
Besides avoiding the above three reaction types wherever possible, cleaner synthesis also involves finding the most appropriate solvents. Water is frequently thought of as the only solvent that can be readily discharged, but finding chemistry that works in water is not helpful unless the reactants and/ or products (plus byproducts) are harmless or can be thoroughly removed. Supercritical carbon dioxide6 is one of very few other environmentally friendly solvents, but as yet its use is limited to a few specialist applications - extracting flavours, or caffeine from coffee beans, for example.
We can sometimes avoid conventional solvents by using an excess of one of the reagents, which circumvents the need to separate and contain another component. For example, polypropene is now prepared in propene, replacing solvents such as kerosene. In many cases, however, effective containment and recycling of solvents may be the best option.
The manufacture of lactic acid illustrates several different aspects of waste production. Two routes ? chemical synthesis and fermentation - satisfy most of the worldwide demand for the acid, ca 40 000 t pa.7 In the chemical synthesis route, the first step involves reacting ethanal with HCN under base-catalysed conditions at atmospheric pressure to give lactonitrile:
CH3CHO + HCN ____>CH3CHOHCN (1)
After purifying the product by distillation, the next step is to hydrolyse the lactonitrile using concentrated sulphuric (or hydrochloric) acid - a typical stoichiometric reaction resulting in the low value byproduct ammonium sulphate (or chloride):
CH3HCOHCN + 2H2O + ?H2SO4 ___> CH3CHOHCOOH + ?(NH4)2SO4 (2)
The crude lactic acid is then esterified to give methyl lactate:
CH2CHOHCOOH + CH3OH ___> CH3CHOHCOOCH3 + H2O
which is purified and hydrolysed under acid conditions to give lactic acid and methanol, for recycling back into the system.
The lactic acid molecule is ’built’ in steps (1) and (2), with atom utilisation efficiencies of 100 per cent and ca 60 per cent respectively, but producing product of a suitably high quality also involves two distillations besides the esterification/hydrolysis steps and their associated energy requirements and wastes.
The second route to lactic acid manufacture relies on fermenting carbohydrates, such as molasses or corn syrup. The resulting acid is neutralised by adding calcium carbonate, to yield calcium lactate. Fermentation takes four to six days and, to keep the calcium lactate in solution so that it can be separated by filtration, its concentration must be below ca 10 per cent, with inevitable consequences for plant cost and capacity. Work up proceeds by carbon treatment, evaporation and acidification with sulphuric acid, to produce lactic acid plus stoichiometric quantities of calcium sulphate. The resulting lactic acid is only technical grade. If higher quality product is needed, we need to include esterification, distillation and hydrolysis steps as for the previous route. Both processes therefore fall well short of the ideals of clean synthesis for a number of reasons.
Is it cleaner?
Choosing the cleanest process from various different routes to make the same product is not always straightforward. One measure that we can look at is the E factor - the amount of waste produced for a given amount of product, taking into account yields and solvent losses as well as atom utilisation.3 However, wastes vary in composition as well as quantity, and consequently may need to be discharged to different media ? for example to landfill rather than to water. Finding ways to quantify the environmental damage done by different wastes is not always straightforward.
It is also important, when evaluating the options for cleaner synthesis, to include all the relevant factors, such as the implications of making the necessary reagents or the energy consumed during product separation. For example, ozone looks like an attractive oxidising agent because it does not leave any toxic residues behind unlike metal-based reagents, but we need to weigh this up against the fact that 90 per cent of the electrical energy used in the electric discharge ozone generator results in heat rather than ozone.
What started out as a simple matter of judging the cleanest process may begin to look like a lifecycle analysis. At least one pharmaceutical company has found that an innovative piece of chemistry required so much extra solvent compared to the existing reaction path that the perceived improvements turned out to be a mirage.
Moreover, while clean synthesis addresses environmental concerns, we must not overlook the implications of such changes for health and safety. In a recent review of inherently safer chemistry, Abe Mittelman and Daniel Lin8 point out that substituting HF for aluminium trichloride in Friedel?Crafts reactions may not be as good for employees as it is for the environment. Companies need to assess the consequences of any changes to processes fully before implementing them.
Manufacturers have been changing and refining their processes in ways that produce less waste since the early days of the chemical industry. Many changes have been driven by economics - by reduced costs for raw materials and energy, or improvements to produce purer products that command higher prices. Tighter regulations for waste disposal, leading to increased treatment or off-site disposal costs, have added to economic pressures - sometimes tipping the balance in favour of less polluting processes. For example in the chlor-alkali industry, membrane cells, which avoid the use of mercury, are now preferred to the traditional mercury cathode cells.
Today, the literature contains many examples of industrially applied cleaner syntheses.3,4,9-13 A good example is the Hoechst-Celanese route to ibuprofen from isobutylbenzene, which involves three steps - two of them catalytic - and has an atom utilisation of 100 per cent, a marked improvement on the previous six-step synthesis.
Companies are continuing to conduct in-house studies on ways to improve their manufacturing processes, but are also collaborating with each other, for example through the Cefic (European Chemical Industry Council) Sustech R&D programme, which is subdivided into: bio-Sustech; catalyst design and application; process intensification; safety and environmental management; process modelling, simulation and control; contaminated land issues; separation technologies; particulate solids processing; and recovery, recycling and reuse.
At the supra-national level the UN, through its environmental programme UNEP, has identified a number of cleaner production industry sector groups, among them several sectors using chemical processes - the textile and pulp and paper industries for example. By organising conferences, newsletters and other forms of networking, UNEP aims to encourage better exchange of information between different sectors. A recent example concerns the transfer of ’know-how’ about clean production in the metal finishing industry between Australia and Hong Kong. When implemented, the resulting recommendations - based on good housekeeping, minimising water use and the recovery of excess metals - should reduce discharges to Hong Kong Harbour.14
Industrial development work on clean processes is underpinned by government funded research programmes in the UK, US and other countries. In the UK, the Cleaner Synthesis programme was launched in 1992 and in the US the National Science Foundation is running a Benign Chemical Synthesis and Processing programme with similar aims.15 A Gordon research conference on Environmentally benign organic synthesis was held in July 1996, while in Italy a number of universities have established an Inter-university Consortium on Chemistry for the Environment. The EPSRC’s financial commitment to cleaner synthesis through its funding of research in universities reached ?15m, with the award of a further ?1.5m for nine projects in June 1996. Topics that are attracting most support include various aspects of catalysis, supercritical fluids, chemistry in water, radical reactions that circumvent undesirable initiators, such as tin hydrides, electrochemical synthesis and some novel reactor concepts.
The EPSRC’s Clean Technology Programme also has a related research target on Waste minimisation through recycling, re-use and recovery in industry and there is a parallel Link programme on this subject.
The past few years have seen much more attention being given to cleaner ways of making chemicals and there is a substantial commitment to research in both industry and universities. Industry is introducing cleaner processes, though confidentiality issues can delay their disclosure. However, to quote a speaker at a recent Royal Society meeting: ’clean technology has to be cost effective, unlike some environmental projects’. Clean synthesis is about a ’win-win’ approach - a ’win’ for the environment and a ’win’ for the manufacturer.
Source: Chemistry in Britain
Dr Tim Lester is EPSRC technical adviser for cleaner synthesis and can be contacted. He would be pleased to hear from any readers wanting to discuss industry’s needs or ideas for research projects to explore cleaner synthesis.
The next Gordon research conference on Environmentally benign organic synthesis will be held in Oxford during August 1997. For details contact Dr Stephen DeVito.
- B. Crittendezi and S. Kolaczkowski, Waste minimisation guide. London: Institution of Chemical Engineers, 1994.
- Waste minimisation: a route to profit and cleaner production. An interim report of the Aire and Calder project. Centre for Exploitation of Science and Technology, 1994.
- R. A. Sheldon, Chem.Tech., 1994, 24, 38.
- K. G. Malle in Waste minimisation: a chemist’s approach, K. Martin and T. W. Bastock (eds), p 35. Cambridge: RSC, 1994.
- K. Smith, Chem. Commun., 1996, 469.
- M. Poliakoff and S. Howdle, Chem. Br., 1995, 31, 118.
- Encyclopedia of chemical technology Kirk-Othmer, 4th edn, vol 13, p 1048. New York: Wiley, 1995.
- A. Mittelman and D. Lin, Chem. Ind., September 1995, 694.
- Chemistry of waste minimisation, J. H. Clark (ed). Oxford: Blackie Academic, 1995.
- J. F. Hayes and M. B. Mitchell, Chem. Br., 1993, 29, 1037.
- M. J. Braithwaite and C. L. Ketterman, Chem. Br., 1993, 29, 1042.
- I. G. Laing in Waste minimisation: a chemist’s approach, K. Martin and T. W. Bastock (eds), p 93. Cambridge: RSC, 1994.
- G. Steffan, Optimisation of classical processes and combination with modern reactions for the synthesis of fine chemicals. Chemspec ’95, Essen.
- Australia Centre for Cleaner Production, The clean advantage, p 1. February 1996.
- P. T. Anastas and C. A. Farris (eds), Benign by design - alternative synthetic design for pollution prevention, ACS symposium series 577. Washington DC: ACS, 1994.
- K. Fischer and S. Hunig, J. Org. Chem., 1987, 52, 564.
- Design Expert, distributed by QD Consulting near Cambridge
1. Better by design
In October 1995 the antibiotic project team at Glaxo Wellcome’s R&D division in Stevenage was engaged in trying to improve an existing route for preparing a promising antibiotic that is currently in clinical trials. One step of the synthesis that posed a particular challenge was the desilylation of the trinem (1) to give the corresponding hydroxyester (2), (see Scheme).
The original route involved deprotecting the trinem using excess tetrabutylammonium bromide, potassium fluoride and acetic acid. In the laboratory these conditions gave us a yield of 75 per cent of the hydroxyester after product isolation. On scaling up the reaction in a pilot plant to produce about 250 kg of hydroxyester, the yield fell to ca 68 per cent. This was because of the instability of the product during the protracted work up, which involved numerous extractions to remove the acetic acid. Of greater concern, however, was the considerable amount of waste generated at this scale (ca 150 kg of waste per kg of product). The waste included 5 kg of quaternary amine byproducts per kg of product, which give off considerable amounts of nitrogen oxide gases during disposal by combustion.
Initially, our literature search for alternative desilylation procedures looked very promising. However, although many methods are available, conditions for deprotecting silylesters are often very harsh and completely degraded our starting compound and/or the required product. Eventually, however, we tried out a milder method described by Seigfried H?nig and coworkers at W?rzburg University, using triethylamine trihydrogen fluoride in tetrahydrofuran.16 Switching to this method gave us just 50 per cent of the unisolated product (still in solution), but we later increased solution yields to over 90 per cent simply by changing the solvent to dimethylformamide or N-methylpyrrolidone. Because the new method no longer uses acetic acid, we did not have to carry out extensive extractions and we were able to isolate the product in 80 per cent yield, irrespective of the scale. This reduced the amount of waste to 60 kg per kg of product, including just 0.8 kg per kg of amine byproducts.
It was an impressive result, but as usual there was a catch. Under the new reaction conditions the lactone impurity (3) gradually forms, and is difficult to remove using the simpler isolation conditions. The amount of impurity that forms is critically dependent on when we stop the reaction - too early and conversion is incomplete, or too late and the level of impurity becomes unacceptable. Traditionally, chemists have used a ’one factor at a time’ approach to optimisation - in other words assessing the effect of a particular factor by keeping all other conditions constant. But while this can be a quick method of improving a single response such as yield, it does not give us an overall picture of what is going on. If there are many conflicting goals, such as simultaneously maximising yield and purity while minimising material and waste, then it becomes very difficult to apply this approach efficiently.
Instead, we chose to use an experimental design software package called DesignExpert17 to generate a model of the reaction based on the statistical analysis of a series of 20 experiments in which we altered all of the factors simultaneously. From this we determined the optimum reaction conditions for various desired outcomes - maintaining high product yields, controlling quality, maximising throughput, and minimising wastes. After applying these conditions experimentally, we were reassured to find that the actual responses closely matched the expected outcomes predicted by the computer.
One drawback of this approach may be having to carry out 20 experiments; fortunately automation can help to remove the drudgery . but that’s another story.
Martin Owen is automation project leader in the process research and development division of Glaxo Wellcome
2. Catalysts for change
In September 1995 Contract Chemicals’ Knowsley site became one of the first fine chemical companies in the UK to be accredited to the new standard for environmental management systems, BS7750. This recognises the company’s commitment to uphold the principles of responsible care, which aim to improve environmental performance. As part of the company’s continuous improvement programme, Contract Chemicals is developing environmentally benign chemistry for synthesising a range of organic intermediates for the pharmaceutical, agrochemical and related industries.The explosive addition of AlCl3 to water, cf the inert behaviour of Envirocat EPZG
One area that my colleagues and I in the research department have recently been investigating, in conjunction with researchers at the University of York, is the development of a range of supported-reagent catalysts called Envirocats, which demonstrate significant advantages over traditional catalysts in terms of processing efficiency, health, safety, and the environment. Important applications for Envirocats include replacing conventional catalysts such as para-toluenesulphonic acid and methane sulphonic acid in a number of high temperature esterifications. Using the solid Envirocat alternative - a strong Br?nsted acid - avoids the problem of coloured final products and reduces processing times without necessitating any plant modifications. Envirocats also have the advantage of producing better quality esters for certain applications, using fewer aqueous washes, and do not require solvent or carbon treatment to purify the products.
We have also applied another Envirocat to synthesise the pharmaceutical intermediate para-chlorobenzophenone (see Scheme), which conventionally involves Friedel?Crafts acylation using stoichiometric quantities of the toxic and potentially hazardous catalyst AlCl3. Unlike AlCl3 , the Envirocat doesn’t complex with the final products, eliminating the need for an aqueous quench and the problem of acidic effluent. Besides reducing HCl emissions by as much as 75 per cent, using the alternative non-toxic Envirocat reduces the weight of catalyst by 10-fold while yielding 70 per cent of the desired para-isomer, with only minimal amounts of the ortho-isomer byproduct.
Keith Martin is a technical service specialist at Contract Chemicals.
Contact and Further Information
Dr Tim Lester
Deft Technology and Design, 11 Nightingale Road, Hampton, Middlesex TW12 3HU
Dr Stephen DeVito
US Environmental Protection Agency
Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY
Penrhyn Road, Knowsley Industrial Park South, Prescot, Merseyside L34 9HY