Reducing the risks

Dynamit Nobel now makes pharmaceutical ingredients rather than dynamite, but these can be just as explosive. Manufacturing Chemist reports on how these materials can be handled safely

It's dark in the bunker, and cold. Tunnelled into a small, wooded mound, the walls are lined with blackened planks and studded with shards of twisted metal. The air smells acrid.
'We place the sample in here,' says the technician, indicating a clamp set three feet above the concrete floor. Surrounded on three sides by a thick steel shield, the clamp is at the focus of four powerful gas jets. The technician indicates the metal fragments buried in the ceiling. 'Sometimes, the tests are quite energetic.'
It's not what you'd expect for a chemical plant, but this is no ordinary facility. Dynamit Nobel's Dynamic Synthesis works, located at Schlebusch, just outside Leverkusen in Germany, has been a centre for explosives expertise for over a century, although these days it uses and processes explosive materials for the pharmaceutical industry rather than making dynamite.
Even so, a major part of the activities at the site involves determining how explosive the materials are. The bunker is an integral part of this process, housing the Kroenen Test, which shows whether a substance will explode when heated. Other tests involve under strictly controlled conditions scraping the materials across ceramic plates, showering them with sparks, and bashing them with falling weights.
Dynamit Nobel's diversification into the pharma business is not as big a leap as it sounds. It stems from its expertise in making nitroglycerine originally as an explosive for the mining industry, but since the 1970s made to treat angina and other heart disorders. Making and handling explosive materials, with their very specific hazards, is highly specialised, so the company has used its knowledge to carve out a niche. Where drug companies need to use explosive materials to synthesise their APIs, Dynamit Nobel is one of the few ports of call.
<b>additional toxicity</b>
Explosive reagents are often called for in pharmaceutical manufacture. For example, many drugs contain heterocyclic rings in their molecular structure similar to benzene rings, but with one or more carbon atoms substituted for nitrogen or other elements. As Juergen Haase, head of r&d at Schlebusch, explains, making heterocycles often requires the use of azides, which contain three multiple-bonded nitrogen atoms, or hydrazines, carcinogenic compounds that contain two double-bonded nitrogens. 'Sodium azide, NaN₃, is comparable to sodium cyanide in toxicity,' Haase says. 'In addition, in almost all azide reactions you have to expect the generation of HN₃, which is extremely sensitive to shock and friction, and detonates with a speed comparable to TNT.'
If that weren't bad enough, use of NaN₃ usually leads to the generation of N₂O, better known as laughing gas, which is also toxic and highly explosive in its own right. And this is one of a group of molecules that are in daily use at Schlebusch. Others include carbon disulphide and diborane compounds.
So what kind of facilities are needed to handle such aggressive substances? The main concern, as with all chemical plants, has to be safety, but this shows itself in ways quite different from most complexes. For a start, standing in the middle of Schlebusch, you'd be forgiven for thinking you weren't on a chemicals complex. In place of the usual towering columns and steel pipework are a widely scattered collection of unassuming buildings, surrounded by trees. But the forestry isn't just for show: in the event of an explosion, the trees would help absorb and contain the blast.
<b>process controls</b>
The setting doesn't compromise the hygienic standard of the plants as befits a supplier of chemicals to the pharmaceutical industry, all the equipment complies with cGMP standards. However, the facilities themselves are also designed to cope with the possibility of explosions. Johannes Schlupp, head of cGMP production, explains that the design philosophy behind the plants has changed little since 1912.
'The need is to direct any blast that might occur, so you need some thick walls that will withstand the pressure, and some blowdown walls that will fall down.' Control rooms are separated from the processing area by strong walls, so operators are always safe.
The philosophy might be the same, but the practical aspect has, of course, changed over the years. The most recent building on the site - the active pharmaceutical ingredients facility was built in 1999 - is designed so that the most risky operation, drying, occurs on the ground floor and has extra reinforcement. 'Drying is much more dangerous, especially with tetrazoles,' says Schlupp. 'Once the liquid solution starts to evaporate, the thermocouples are no longer immersed and they are not reliable.'
Another instantly noticeable and surprising design feature is the widespread use of glass within the plants. This is a safety feature, and is due to the formation of shock-sensitive sublimates during the recrystallisation of acidic substances. 'If the condenser and receiving vessels are made of glass, you can see if these sublimates are forming,' Schlupp says.
'If it happens, we rinse the equipment from the top down, to redissolve the sublimates and make the plant safe. If the equipment were made from steel, we wouldn't be able to see the sublimate; it would build up to dangerous levels, and it's a more difficult material to clean.'
Despite these features, however, the general design of reactors at Schlebusch is very similar to other fine chemical production plants. But according to Stefan Loebbecke of the Fraunhofer Institute of Chemical Technology, new developments in microreactors could present some interesting alternatives.
Microreactors, where the reactions take place in channels a fraction of a millimetre across, are particularly suitable for reactions involving or producing explosive materials, Loebbecke says. The mixing characteristics are very good, so hot spots do not form. Retention times are small, reducing the opportunity for explosive conditions to develop. Moreover, the small hold-up characteristics of microreactors allow process conditions to be controlled precisely, with unwanted reactions and side-products suppressed, enhancing the selectivity and yields of the reactions.
Microreactors consist of a series of channels around 700µm in diameter, etched into small plates and chips of silicon and similar materials, looking very similar to electronic components. A typical microreactor would measure around 15x25mm. The channels are arranged to maximise the mixing of the materials. For example, the channel might be interrupted by a series of shapes like a capital G, which splits the flow and forces it to recombine.
Silicon is an ideal material for microreactors, Loebbecke says, but not because of its semiconducting properties - it is transparent to infra-red, he explains, which helps the researchers monitor reactions. Loebbecke uses IR to scan the reaction mixture at the entrance point of the microreactor, at the exit, and at the points where the 'G' shapes force the streams to recombine.
Among Loebbecke's experiments is the synthesis of nitroglycerine in silicon microreactors. 'Apart from the microreactor itself, the set-up is unspectacular, it's just junk engineering,' he says. The reaction is controlled thermally, by immersing the whole microreactor inside a water bath, and the reagents are introduced into the system using simple syringe pumps.
<b>production concerns</b>
Normally, microreactors would use peristaltic pumps for reagent dosing, Loebbecke says, but when explosive materials are being used, it's important to avoid any pulsation.
Nitroglycerine is made by nitrating glycerin with a mixture of concentrated sulphuric and nitric acids, and oleum. The product is highly sensitive to shock, friction and elevated temperature; conventionally, it is kept below 30°C to avoid the risk of explosion.
However, the microreactor method does not require the use of oleum, and allows safe handling of the reaction mixture at temperatures as high as 45°C, says Loebbecke, and higher temperatures mean faster reactions. Moreover, if a microreactor becomes too hot and an explosion does occur, the damage caused is as small as the equipment. Generally, a corner of the microreactor might be blown out or the wafer would break into pieces. The volume of the reactants is too small to cause any serious damage.
Although the volume of individual microreactors is tiny, an array of them could provide industrial-scale capacity. 'Several hundred kilogrammes, up to several tonnes, is possible,' Loebbecke says. Safeguards would need to be in place to prevent fouling and clogging of the system purging cycles or ultrasonic cleaning techniques could be used for this, Loebbecke suggests.
The technique is not limited to nitrations: many reactions could be tackled, as long as they occur in the liquid and gas phase. 'The reactions themselves are carried out continuously, but we could use batch-phase,' Loebbecke says. 'We could also use different solvents. It's not just copying and miniaturisation, it's a chance to work on the chemistry itself.'