UK Chemical Reaction Hazards Forum
 |  Previous |  Home |  Next |  SiteMap |
ICRA Label
Safe Surf
 
Miltronic DHTML menu
Creative Commons License

Sodium Hydride / DMF process stopped

Scenario

A manufacturer had contracted out the R&D for a new product.  The contractor had spent some time optimizing the process and the isolation of the product, scaling up as they went, up to a 10 litre scale.  Before the manufacturer would fund his three cGMP validation batches, it was decided to have one large batch made by a sub-contractor (under non cGMP) to prove that the chemistry worked and the isolation of the product was successful.

During the discussions with the chosen sub-contractor, the sub-contractor's Safety Scientist on seeing the process, raised the point that the process used sodium hydride in dimethylformamide and that the mixture was warmed.  He then pointed out that they could not possibly run this process in the pilot plant.  This was a surprise for the contractor (and the manufacturer) as they had seen many processes in the literature using this mixture.

The Safety Scientist insisted it was unsafe to run and left to look up some firm data to show the contractor.  He showed them the papers published in 19821, & 2, the data in Bretherick3 plus unpublished data he had determined at another employment4.   The contractor left and the project was now almost certainly doomed to be terminated.

Lessons Learnt

Conclusions

It is plain that the use of dimethylformamide (DMF) and sodium hydride continues on, mainly due to chemists following old publication procedures.

The use of sodium hydride and dimethylformamide often proceeds without event, especially on a small scale.  It is only when the reaction is scaled up does one encounter potential problems with runaway reactions

The runaway reaction observed between sodium hydride and dimethylformamide is not very reproducible at plant scale.  Some batches have successfully been made without incident.  This may well be due to the quality of the sodium hydride (traces of free sodium) and the dimethylformamide (water content has been suggested).

Nevertheless, at least two pilot plant vessels at different locations have been damaged by this mixture, so one must not look favourably on implementing a plant process based on mixtures of sodium hydride and these protic solvents.

References
Back to Top