Wednesday, June 18, 2014

Twitter: TMS-azide explosion at University of Minnesota

Yesterday, University of Minnesota chemistry professor Chris Cramer tweeted about an explosion within UMN's laboratories. Initial reports indicate that TMS-azide (MSDS here) was the culprit. The graduate student involved was both burned and hurt from shards of glass and appears to be recovering.

As you can see, the damaged hood is pretty wrecked. A hearty "glad you're okay" and "best wishes" to the grad student who was injured. Stay safe out there, folks.

[It sounds like a possible culprit may have been hydrazoic acid. If so, we have a pretty good idea now why process groups work so hard to avoid an accumulation of it in their reactors.]

36 comments:

  1. I presume the postdoc likely attempted the Orgsyn procedure on double the original scale, without making sure the diglyme and NaN3 was dry and TMSCl free of HCl... Its those little things that can kill you. I bought and used TMS-azide couple times, for Curtius and also to make tetramethylguanidinium azide, but never prepared it myself. TMS-azide definitely likes to solvolyze as soon as it see a hydroxy group - i.e. isopropyl alcohol. I think people need to be instructed how to properly handle the stuff, any spill in the lab should prompt evacuation because people get pretty sick from breathing the vapors.

    ReplyDelete
    Replies
    1. Reportedly, this student used azides all the time, so I'm guessing they were fully trained.

      Delete
    2. If you've been to graduate school, then you know that's a terrible assumption.

      Delete
  2. so if hydrazoic acid is explosive and TMS azide isn't, when the hydrazoic acid impurity detonates does it set off the TMS azide as well? i'm wondering if it was really a "200 g" explosion or if the unhydrolyzed azide did not contribute so much to the explosion

    ReplyDelete
  3. It's possible TMSN3 is a relatively stable explosive, so that only when subjected to lots of force does it detonate/explode. The MSDS doesn't note that, though - it (not MSDS, but European version - http://www.apolloscientific.co.uk/downloads/msds/OR4280_msds.pdf) does say that its vapors may form explosive mixtures with air, though, so that if HN3 forms and explodes it might atomize the TMSN3 and generate an explosive mixture with air which can then go off (because the trigger is present).

    ReplyDelete
  4. Thanks for this. Please keep posting these incidents. As a biochemist, I am clueless about the volatility of organic (or inorganic) compounds.

    Could someone explain to me why sodium azide is not volatile but hydrazoic acid is volatile?

    ReplyDelete
  5. comment died - ?

    1) TMSN3 isn't noted as explosive but can form explosive mixtures with air, so if the hydrazoic acid is present, it can detonate/explode spontaneously (because it's highly sensitive) and atomize the TMSN3 which would then also explode (like a fuel-air bomb).

    2) HN3 has a low molecular weight, and it can probably hydrogen bond a little bit - there aren't many intermolecular forces to hold it together, so it has a low bp (body temp). NaN3 is ionic, so charge holds the ions in place, and since the charge is all over the place (a Na+ ion sticks to multiple nearby N3- ions, and vice versa) there are lots of forces to hold the ions together and not let them get away (volatilize). I think that's a rough explanation.

    ReplyDelete
  6. I meant volatile in terms of explosive. Why is NaN3 not explosive but HN3 is explosive.

    ReplyDelete
  7. For the same reasons given in 2) above.

    ReplyDelete
  8. I don't know. Guessing - I don't think there's as easy a way to get NaN3 to N2 and something else - with HN3, it can spit out N2 and the remaining HN probably has reasonable pathways (with O2?) to get to small molecules, probably H2O and N2; the secondary reactions probably generate a lot of energy, as well. With NaN3, you have to overcome the lattice energy holding the salt together (ion attractions), and then find some good reaction to get rid of NaN (or the barrier to spitting out N2 would be too high). With heavier, softer transition metals, they have the option of redox, because the energy to remove or add electrons to them isn't so much, but Na needs lots of energy to accept an electron, and way too much energy to give one up. If there isn't a good pathway to volatile products, then it probably is less likely to explode. Again, though, I don't really know.

    ReplyDelete
  9. NaN3 is explosive. It's detonated in airbags to inflate them during car accidents, for example.

    ReplyDelete
  10. An explanation of the events that led to the picture being temporarily removed:

    1. I received a request from the grad student to take it down.
    2. I thought about it (consulted some trusted people) and agreed to the request.
    3. The picture was taken down.
    4. The request to remove the picture was retracted by the grad student.

    It'll be back up in about 4 hours.

    ReplyDelete
  11. I don't know how it's initiated, though, and how much energy it takes. HN3 is spontaneously explosive, and I couldn't find anything on NaN3's shock sensitivity, just its toxicity, ability to generate HN3 and other more explosive metal azides. Since it's an azide, it probably still wants to be N2, and other gorp, though.

    I was assuming that insensitive = not explosive. Perhaps it would be better to say that NaN3 doesn't easily explode (high kinetic barrier), while HN3 explodes really easily (really low kinetic barrier), as do lots of other azides.

    ReplyDelete
  12. Solid azides are shock-sensitive. If you put a few milligrams of NaN3 on the bench top and hit it with a hammer, it would create a little spark and a pop sound. Explosions in the chemistry lab are part of the business. This is why it's important to have good safety habits and treat unknown compounds like they can kill you. I remember one time my unknown compound detonated in the flask I was holding it in my hand. It required surgery to remove the bigger shards glass. Glass still comes out of that hand from time to time. Luckily my face was protected by the hood sash.

    ReplyDelete
  13. sodium azide does not detonate in airbags. It deflagrates. If it detonated you would be dead.

    ReplyDelete
  14. I work with azoic acid. The way to generate it is adding slowly sulphuric acid to a mixture of toluene, water and sodium azide. The toluene solution thus formed can be later titrated and used. It is excellent for Mitsunobu reactions. Obviously it is a highly toxic and explosive and must be handled carefully. Azides per se are tricky business, but they are very good protective groups and are found ubiquitously in the chemistry literature. As chemists we must be prepared to handle any type of reactions, dangerous or not, however we must fully understand what we are doing. Accidents as the one described are a good example of what happens when someone does something that he does not fully understand. It always comes back to the same old story, document you're self before running a new reaction and be sure you understand well what are you doing. If you still have doubts, ask.

    ReplyDelete
    Replies
    1. If you want to make organic azides by a Mitsunobu procedure, try Rollin's method (Synthesis, 1990, 130) which uses the safe, stable, crystalline and easy-to-make bis-pyridine complex of zinc azide. No H+ at all! Anyone using a hazardous procedure should first consider all alternatives.

      Delete
  15. So, I ask the crowd, how would you quench/dispose TMS-azide? I have been reading about several reactions recently where a slight excess (1.2-1.5 equiv.) of TMS-azide is used but it's not apparent how the authors deal with the extra azide.

    Is something like this reasonable?

    http://www.ehs.neu.edu/hazardous_waste/fact_sheets/sodium_azide/

    ReplyDelete
    Replies
    1. I actually used that procedure for a couple of years with no incidents during deactivation; this was for sodium azide and some transition-metal azides. You just have to make sure you're using enough nitrite in the first place.

      Delete
    2. The MSDS provides hazards, first aide, toxicology, recommended PPE, phys. props etc. Also keep in mind that there are a few other texts which are very useful when preparing to hande materials such as azides: “Bretherick’s Handbook of Reactive Chemical Hazards” and “Destruction of Hazardous Chemical in the Laboratory.” Always use the literature as much as possible to raise awareness and be safe.

      Delete
    3. All good choices. I am a huge fan of "Prudent Practices in the Laboratory", since it is free and searchable. It also contains safe destruction procedures:

      http://www.nap.edu/openbook.php?record_id=4911

      Delete
  16. "4. The request to remove the picture was retracted by the grad student. "
    Kudos to the grad student for wanting others to learn from this incident so soon after it happened. That takes a huge amount of emotional strength. I know, because I did not have such strength in a less severe situation. I hope you heal quickly.

    ReplyDelete
    Replies
    1. The student who was injured in the azide reaction and the student who posted it to Twitter are two different people, both at the University of Minnesota.

      Delete
  17. Why is anyone making this in the first place? $3 /g from Aldrich

    ReplyDelete
  18. Because TMSCl is $0.60/g and NaN3 is $0.50/g. When there is no money, grad students are forced to do what's cheap.

    ReplyDelete
    Replies
    1. I went to grad school so I understand this mentality, but damn. That hood isn't going to be cheap to repair.

      Delete
  19. If you wanted to make a 10g batch of TMSazide, would take 10g of TMSCl and 6 g of NaN3, for a cost of $9 plus labor, solvents, waste, lost time doing something better, etc. Thats about an even trade if you ask me

    ReplyDelete
    Replies
    1. Solvents are cheap. Labor, and waste is free (to the PI).

      Delete
    2. Not really. Just came back from a conference and one of my friends who is a PI mentioned that 1h of grad student labor costs ~$50-70, counting tuition and overhead. So, that this should happen is stupid. Inefficient, too.

      Delete
  20. What scale was this on? I wonder if the injuries are worse than they say because that hood got wrecked.

    ReplyDelete
    Replies
    1. The original tweet said 200 grams.

      Delete
    2. 200 g preparation of TMSN3 is borderline insane in an academic setting. Definitely some negligence on part of the PI as well as the chemist performing the task.

      Delete
    3. All this when it only costs $600 to buy 200g .....

      Delete
  21. TMS-azide is often touted as "safe". However, for safety analysis it should be considered the same as hydrazoic acid as there are routes beside hydrolysis by which TMS-azide can decompose to hydrazoic acid. Likely another such route was discovered here. All azide reactions should be handled with care especially when hydrazoic acid can be formed. I'm not a fan of this reagent and other chemistry should be used where possible. Otherwise use in dilute solution.
    Tom
    archibaldmail@gmail.com

    ReplyDelete
  22. These guys substituted PEG for diglyme which is specified in the orgsyn prep. PEG stands for polyethylene glycol. As in alcohol. It's a short-chain oligomer of ethylene oxide terminated with hydroxyl groups. It is different in that way from diglyme, which is diethylene glycol dimethyl ether. Now it's obvious why hydrazoic acid was formed in this accident, whereas the orgsyn prep is safe.

    ReplyDelete