An epoxide intermediate polymerized leading to a drum rupture. The intermediate was thought to be stable.
It was shown that the polymerization was base catalysed. Normally the compound was purified by a silica treatment prior to drumming up. In this particular case less silica was available to carry out the purification. The process change was not perceived as significant.
- The use of less silica resulted in not all the base being removed.
- Thus the intermediate was drummed up with traces of base present, ideal conditions for the subsequent polymerization.
Once again we see a minor "insignificant" process change being the main cause of an incident. All process changes, however minor must be properly assessed, starting with laboratory trials.