Process Step Omitted Causes Boil up
During a new preparation of a sulphonyl chloride, the reaction mixture (in dichloromethane) was added to an ice/water slurry to destroy the unwanted phosphorus oxychloride. After some mixture had been added, a delayed exotherm ensued causing the DCM to boil rapidly, only moderated by rapid addition of ice by the operators (in full BA)
The process had been carried out many times before in the laboratory, however the laboratory had always stripped out the phosphorus oxychloride to a "brick", which was subsequently dissolved in dichloromethane and quenched into ice/water to kill the last traces of phosphorus oxychloride and phosphorus pentachloride. This step was omitted in the plant process, which was written based on a preparation of a different sulphonyl chloride, where the phosphorus oxychloride (but less used) was not removed. Thus there was "some" justification for not removing the phosphorus oxychloride. In addition this was a "rush-job".
However when the batch was quenched the phosphorus oxychloride / dichloromethane mixture failed to react with the ice-cold water until it started to go, this then caused a warming, with subsequent fast reaction of the (already) added mixture.
Subsequent quenching of the batch was done in water with some hydrochloric acid added (which is known to assist the hydrolysis of phosphorus oxychloride), and at 15°C. The final quality of the product was, however, only 90%
- Plant processes must be based on a sound laboratory process, not a "blend" of one or two more processes, especially where highly reactive materials are used.
- Although time is often of the essence, it is essential that short cuts are not taken that could prove detrimental to the safety of the process.