Fred Starr recalls the Flixborough disaster
As the new year begins, I am reminded that the 40th anniversary of the Flixborough disaster, the biggest explosion at a UK chemical plant, is approaching.
The blast, at Nypro UK’s north Lincolnshire plant on 1 June 1974, was the equivalent of up to 15 tonnes of TNT, which totally demolished the plant and killed 28 people.
However, Flixborough was alleged to have been responsible for an even greater disaster, the BSE crisis, (better known as mad cow disease), which hit UK farming in the early nineties. The Flixborough plant produced caprolactam, which involved oxidising cyclohexane at temperature and pressure in six interconnected large-diameter reactors. Caprolactam was an intermediary in the manufacture of nylon, but it was also used in the rendering of beef carcasses, whereby the fat was separated from the flesh and bones. The waste products were then crushed to make a type of gruel that was fed back to cows. What the digestive system of a ruminant, evolved to eat grass, made of this form of cow cannibalism is a puzzle to me as a mere metallurgist.
The brains and the spinal cord counted as waste that, when infected with the mad cow prion, would – if nothing were done to break the cycle – result in British cow herds becoming more and more infected. The suggestion was that the caprolactam treatment in some way, eliminated the prions, since the incidence of BSE in cows began to rise a few years after Flixborough. The total destruction of the plant resulted in a shortage of caprolactam, and this form of rendering was stopped.
The devastation at Flixborough was the culmination of a chain of consequences, starting with a massive stress corrosion crack in number five of the six reactors. Each of these were linked by 70cm diameter pipes, through which the hot, pressurised cyclohexane drained from one reactor to another. One of the valves on the fifth reactor was in bad shape and had to be kept cool by spraying it with water. The nitrate in the water, plus the heat of the reactor vessel, resulted in stress cracking of the vessel wall.
The plant was shut down, the offending vessel removed and a narrower pipe (50cm rather than 70cm) was built to connect reactors four and six. Two months later, the new section of pipe gave way and tens of tonnes of boiling cyclohexane surged across the plant. A vast explosion was inevitable.
Although the 50cm pipe was badly designed, it did not fail immediately. One hypothesis was that a localised fire had occurred elsewhere on the site. This caused a nearby stainless steel pipe to fail by creep or zinc embrittlement, releasing its own torrent of boiling cyclohexane, whose explosion blasted away the pipe. In the first stages of the fire it was supposed that zinc, in molten form, had dripped down from galvanised walkways onto the stainless pipe.
The implications of the Flixborough disaster reverberated throughout the chemical engineering industry. Because of the amount of stainless pipework on our own reforming plants, I did some tests. The results were amazing. A rod of Type 316 stainless, which even at 800°C was quite hard to bend, snapped like a carrot when smeared with a piece of zinc.
The irony of Flixborough, with respect to the mad cow tragedy, is that the disaster can be traced back to the farmers themselves. It was the spraying of the reactor using water from the mains supply that had caused the stress corrosion. But the nitrates in the water had come from the farmland run-off heavily fertilised with ammonium nitrate, not to mention the highly nitrogenous waste produced by cattle.
The mistakes leading up to Flixborough have probably been forgotten, as those working in the chemical plant sector at the time have since retired. So doubtless as you read this, somewhere in the world, a hapless plant attendant has been told to spray water on a valve, oblivious to the dangers of stress cracking. Awareness of zinc embrittlement has gone the same way, although there was a well justified, post-disaster row about its significance. How the argument arose, and its pros and cons, will be reserved for a future recollection.