Cleaning Process Leads to Jet Flame. Hickson and Welch Ltd 1992

The Incident

On 21st September 1992 at about 1.20 pm at Hickson & Welch Ltd in West Yorkshire, a jet of flames exploded from a manway access opening on the side of a batch still.  

Two people were killed and three people suffered burn injuries when the flame cut through a nearby office building.   

The flame also reached a four storey office block and as a result windows were shattered and rooms set on fire.  63 employees managed to escape, but one female employee, although rescued by fire crews, later died of smoke inhalation. 

The Cause

A process vessel known as 60 still base used to distil batches of an organic liquid was being raked out to remove a build-up of semi-solid residue rich in dinitrotoluenes (DNT) and nitrocresols.  The vessel had never been cleaned out (for 30 years) and the bottom seam battery was covered in sludge.  Heat had been applied for about three hours through an internal steam coil before the raking out process began.   This began a self-heating runaway reaction in the residue leading ,with disastrous consequences, to deflagration and a jet flame. 

A formal investigation under the Health and Safety at Work etc. Act 1974 was carried out by the Health and Safety Executive (HSE) .  A special report was made which concentrated on the cause of the incident and the precautions that should have been taken but doesn’t mention the satisfactory arrangements that were found during the investigation. However, several important lessons of general relevance to the chemical industry were revealed by the investigation.

Manufacturing was suspended on the MNT plant by the company immediately following the accident and a technical committee was set up to find alternative methods to remove DNT and nitrocresols from the manufacturing system.  The resulting options were discussed with the HSE.  The system was modified to eliminate batch distillation of whizzer oil in 60 still base.  A new system was devised before MNT production re-commenced which involved distillation followed by fractional crystallisation.

Conclusion 

It was concluded that the accident occurred following the attempt to remove sludge in the still base that had accumulated over a long period of time.  The material  which was heated to make it soft was assumed to be making the plant less efficient.  They knew that the materials processed in the vessel were highly energetic, but didn’t attempt to monitor the accumulation of residue, even though in October 1988 a significant process change had been introduced.  Hickson and Welch management believed that the level of residue in the still base ebbed and flowed with each distillation.  However, on 21st September 1992 the removal of sludge from 60 still base was authorised without attempting to identify the hazards, risks or even the material.  The residue contained organic nitro compounds and it is known that they can undergo exothermic decomposition leading to thermal runaway when temperatures are raised.

The investigation revealed several important lessons, of which, where the batch distillation of highly energetic materials (such as Mononitrotoluenes or other organic nitro compounds) is carried out still residues should be analysed, monitored and removed at regular intervals to prevent possible build-up of unstable impurities.