Monday, November 10, 2014

University of Minnesota chemistry has limited the scale of azide reactions?

Also in this week's C&EN, a really worthwhile article from Jyllian Kemsley talking to William Tolman, the chair of the chemistry department, on their recent TMS-azide explosion. While the whole article is worth reading, I think this section detailing Professor Tolman's decisions was very interesting (emphasis mine):
“Overall, there was clearly a lack of proper hazard assessment,” Tolman continues. “They didn’t stop and say, ‘This is a really dangerous procedure. Should we be doing this at all, or should we be taking extra precautions?’ ” The lab became complacent after doing the reaction several times without incident, he believes. And warnings included with literature protocols were “pretty lame,” he says. 
Tolman notes that no lab in his department has the proper equipment to allow the reaction to be performed safely at the 200-g scale. He has now set a limit of 5 g for any procedure involving azide. 
Tolman announced that limit and other follow-up actions in a department-wide meeting in July at which he discussed the incident investigation findings. What he calls a “lively” 45-minute discussion ensued. Such meetings are important for leaders to publicly acknowledge the importance of significant incidents and show support for discussing safety concerns, Tolman says. “The emphasis was not on assessing blame,” he adds, “but rather on what we should all do to improve risk assessment.” 
In addition to limiting the scale of azide reactions, Tolman ordered lab groups in his department to assess their standard operating procedures and update them if necessary. The goal was to get everyone to stop and think about whether they’re doing anything that is potentially hazardous, whether they have a procedure for that activity, and whether that procedure is correct, he says. That lab self-assessment was completed in August, and the department’s safety committee is now working out how to do a peer review of the methods.
The remainder of the article discusses the "safe operation cards" that label what reaction is being run in a hood and what the potential hazards might be. There's also a renewed focus on safety meetings, hazard assessment and getting students to think about what potential risks might lie behind their experiments. As with all procedural changes, an accurate assessment of whether or not this will do any good will come 5 to 10 years from now.

I would like to know a little more about the reasoning behind the 5-gram azide limit. (UPDATE: Prof. Tolman explains more.) Also, what are general limits around the academic chemistry community? (The Sharpless laboratory at Scripps is probably the most concentrated collection of organic azides in academic chemistry -- what are their internal procedures?) I tend to think hard limits on reaction scale are less than useful, but hey, maybe they have some DSC data to back this up.

I cannot help but contrast the University of Minnesota's response to this event to what happened almost 6 years ago at UCLA with the Sheri Sangji case. Between now and June (5 months?), they've published a safety letter to the chemistry community on what happened and how to avoid it. You'd think that the millions of dollars coughed up by UCLA would have produced a similar document, but I don't think that's happened. (There's an argument to be made that the main difference between that incident and this one has been the presence/absence of lawyers, too.) 


  1. My place of business (industry, government contractor) imposes scale limits on high explosives. We work with energetic materials (think stuff that goes BOOM intentionally). The logic is somewhat morbid, but: if 2 grams of RDX detonates in a flask you are holding it won't kill you. You might lose some fingers, or suffer some lacerations from shattered glass. But ultimately you'll walk away with your life. I suspect similar reasoning is behind the 5 grams rule being imposed at UMinn.

  2. I would se the limit higher, after making quite a few azides on 50+ gram scale, based on these considerations: What is the product carbon/nitrogen ratio? (3-hydroxy-1-propyl azide is probably dangerous, the analogous 5-hydroxypentyl azide is probably not). Are you using aqueous/protic reaction conditions? Can HN3 formation in the reaction flask happen? Are you distilling neat azide material at the end - and on what temperature?

    I think one should not make one simple rule and then use it instead of knowledge and common sense, but I think azide scale-up experiments run by grad students do require extra supervision

  3. When there's an incident like this I see a lot of warnings go around about the specific chemical involved, which is good, but these warnings often get extrapolated out to every chemical in that same class, which isn't so helpful. I'd much rather see a detailed analysis and more specific guidelines about when certain chemicals and/or conditions really do carry a higher risk that necessitates additional controls, and when they don't. I hear broad statements like azides are dangerous, and it is true that many of them are dangerous, but the actual risk profile varies dramatically from compound to compound.

    Maybe the information I want is available and I'm just too lazy to find it, but, if it is, I feel like it should be highlighted in cases like this. Or maybe there are too many variables and we're talking about low probability but very dangerous events, so its hard to provide useful guidance? Or maybe there's too much liability involved in saying something is less dangerous, and then having that thing cause an incident?

    1. Liability is a big factor here. Also, a good process safety review takes time and involves trained specialists from several areas. These people are often viewed as overhead and show stoppers and trimmed/cut/laid off at every opportunity.

      Some years ago I heard about a small scale reaction review at Ely Lilly. The entire review was on paper/computer and was turned around in 1-2 days. The number of accidents and near-accidents was reduced to a small fraction of the original number.

      As for the usefulness of broad statements of hazard.... When I was in college I could pick up a bottle of acetonitrile from an open shelf any time. However, methyl cyanide was locked up in a armor plated enclosure and getting it required three signatures and a witness. It is all in the name :).

  4. I'd also like to know the reasoning for why they needed 200 g of TMSN3. That seems excessive for an academic setting. During my grad school years, most of my reactions were on milligram scale. I did a few reactions on the tens of grams scale, but these were mostly aqueous type reactions that had simple work up procedures.