Skip to main content

Chemical Biology

4-Azidophenylalanine: A Warning

There’s a reagent used in chemical biology and protein labeling that should be getting a bit more attention than it does. Not because it’s useful – that’s already known – but because it can explode.

Here’s the paper (from UC-Irvine and Amgen), and the compound is the 4-azido derivative of phenylalanine. Interestingly, the authors were not trying to investigate its hazards, but were after an easier, safer, and cheaper synthetic route. They succeeded, by the way, along with providing some valuable reference work about how to determine its chiral purity and also (re)establishing a poorly characterized coupling reagent (Boc-Oxyma) as a useful species. Now, some people will look at this target compound and say “Well, yeah, sure, azide”, but organoazides are not all explosive. The lower the molecular weight, the more likely you are to run into trouble, but there are smaller azides than this one that still aren’t a problem to handle. I’ll say right off that I don’t know of any incidents with azido-Phe, which is a good thing, but the data in this paper suggest that such an incident is certainly possible. More on that in a moment.

One of the concerns about the synthetic routes to this compound, I should note, is the presence of an azide ion / copper combination. It is a very solid general rule to avoid the formation of any transition-metal or heavy-metal azide species; they should all be considered explosion hazards (“Things I Won’t Work With” reference, for emphasis). Want some more backing for that statement?

. . .The sensitivity of dry cupric azide to friction is so great that it is prone to explosion during its removal from filter paper. Copper(I) azide (cuprous azide) is similarly perilous, and explosions have been reported for this complex while completely submerged in water and after being lightly touched with a feather in open air. An intelligence bulletin from The National Institute for Occupational Safety and Health warned in 1976 that several explosions had occurred in the United States and Canada, apparently resulting from the accumulation of heavy-metal azides in hospital plumbing. These events were likely due to the routine disposal of NaN3-laced waste from automatic blood cell counting machines directly into hospital drains. Since many of these incidents occurred while attempting to clear a blocked pipe, it seems unlikely that any metal azides present were dry.

Said the manuscript, dryly. So no, I don’t think we’ll be wanting any copper azide species if that can be helped (the covalent azides used in the copper-catalyzed “click” reaction are a different thing altogether). Since this new synthesis does still involve a copper-catalyzed azide formation, the authors carefully investigated the material in each step for explosion hazards. The good news is that none of the intermediates (nor the precipitates formed) seem to be hazardous. The bad news is that the final product is.

This was determined through differential scanning calorimetry (DSC), a technique where a (small!) sample is heated to increasing temperatures and monitored for excess heat being given off (exothermic decomposition), and its partner technique, accelerating-rate calorimetry (ARC). What these show is that the N-Boc derivative of azido-Phe is possibly a cause for concern, but the unprotected species is absolutely a problem. Its calorimetric behavior is similar to that of a high explosive, and it appears as if it could be shock-sensitive as well. I would say that this problem is unappreciated in the chemical literature, and it’s a good thing that no one discovered this the hard way. Perhaps a salt of the compound has better behavior and would be a safer reagent and article of commerce?

26 comments on “4-Azidophenylalanine: A Warning”

  1. Boomtown says:

    In grad school when I worked with organic azides, I was told the rule of thumb was that if your carbon to nitrogen & oxygen ratio was < 2, it had a good chance of being explosive. In this case it is 9 to 7, so it should have been assumed to be explosive unless proven otherwise.

  2. Synthon says:

    My chemistry teacher worked on munitions in WW2. As a young employee he was introduced to copper azide’s properties. A few crystals were placed on a heavy quarter inch steel plate which was then heated underneath with a Bunsen. They all stood back, there was a loud crack and the plate fell apart in two pieces.
    Surprised about azidoPhe though. In practice it’s not going to be a problem except now transport regulations will probably make it difficult to get hold of.

  3. Anon says:

    Just curious if Fmoc derivative of above amino acid would be fine? We are increasing the MW (more stable?) by putting Fmoc and can be removed only under basic (piperidine etc.) conditions.

    1. Mark Richardson says:

      We didn’t test the Fmoc derivative, but the Boc protected compound was not overly concerning as an explosion risk. If you are performing solid-phase synthesis then you are probably only handling mgs of the stuff in solution, so I think you’ll be okay.

      1. MagickChicken says:

        I read that as “mugs of the stuff in solution” and my eyes bugged a little.

  4. cato says:

    Could have tried the old hammer test–throw a couple mgs on a paper towel, bang it with a hammer, and if you see black residue, maybe think of another route….

    1. Derek Lowe says:

      That’s certainly a lot cheaper than buying an official drop-hammer test rig. . .

    2. Mark Thorson says:

      My dad was a high school science teacher, and he warned me about some incident in which a student set off a whole roll of caps with a hammer. The hammer was blown back into the kid’s face. I suppose the lesson is if you hit a potential explosive with a hammer, have your elbow at a roughly 90 degree bend so that the hammer goes flying elsewhere.

      1. Wheels17 says:

        I blew the hammer head off the hammer body with stacked caps as a kid. Most of the energy was absorbed in the fracture, so the head didn’t develop a lot of velocity. It was fun being a kid back then. The big bang cannon was a great lesson in stoichiometry. If one scoop was loud, shouldn’t two scoops be louder?? No, it was quieter, but the flames out the front were longer….

        1. Kazoochemist says:

          Geeeez…. I must have done the “multiple rolls of “caps” “ experiment a zillion times when I was a kid. We hit them with hammers, we dropped heavy weights on them, we carefully cut out the little “dot of goodness” to concentrate the minimal force present in a roll of “caps”.

          Can you actually purchase “caps” today?

        2. fajensen says:

          I blew a tiny, sharp sliver of steel off a hammer as a kid. This little piece embedded itself into the bone of my thumb, right between the nail and the joint. I had to use needle pliers and some effort to pull it out again (why I know it hit the bone).

          If that had hit an eye instead, that would not have been good. I became aware that even the dumbest, of dumb & simple objects can be dangerous in unforeseeable ways.

          I think it wasn’t the caps that did it, rather the energy came from the mechanical impact from smashing the hammer head into a steel plate the caps were placed on.

  5. Uncle Al says:

    How to determine its optical purity.
    … The world has improved – including absolute configuration determination
    … Got a mess? Get chiral answers anyway.

  6. Chrispy says:

    Years ago a colleague was generating a bunch of azide/acid fluoride amino acids to use as N-protected, C-activated building blocks. On a whim he dripped the tiniest drop of a solution of the stuff (not sure which amino acid) on a hot hotplate. BANG! The resulting explosion was so loud that it shook the windows and drew people from all over the building, and it shut down this line of research. This was too bad because they were really excellent building blocks, with very little steric hindrance around the hot carbonyl carbon.

  7. milkshake says:

    Arylazides are known to be more sensitive than alkyl azides. I distilled on a large scale (100g) neat N3CH2C(Me)2CH2OH: that neopentyl azide was made from the corresponding chloro compound at 140C/2days in DMSO with NaI. It was fun.

  8. Quintus says:

    To quote from the abstract” A reliable, scalable, cost-effective….” Sure is cost effective and reliable as it can always be guaranteed to destroy equipment, buildings etc and does so reliably; but not scalable. Why be surprised about this?

    1. Mark Richardson says:

      Hi Quintus. We didn’t destroy anything. The sequence is scalable in the sense that it worked well on milligram scales as well as on gram scales. These reports are mostly to inform University laboratories; I am not certain how this compound is produced industrially. Most labs will be happy to acquire a gram or two to support their projects, and so taking out a building is unlikely. In either case, the Cu (I) catalyzed azidation we investigated is a much less hazardous alternative to azidodediazoniation or diazotransfer. The trick to staying safe is to stop large-scale production at the penultimate N-Boc protected intermediate and only bring through small amounts of the potentially explosive final product as needed. As with most risks in chemistry, the hazards can be effectively managed by limiting the scale, and should abate the risk of damaging equipment or, worse, injuring the chemist operating it. I hope this clears things up a bit.

      1. Quintus says:

        You need to investigate the physical properties of this material, especially if you may be shipping it around the place. If it is explosive, try find someone willing to transport it!
        So hazards can be managed by limiting the scale, well I’m surprised that we have chemical manufacturing because potentially every process has its associated hazards. And, after a significant number of years spent in process research, I came across more than a few hazards, especially in the azide area of expertise, where we produced lots of 10KMol batches of a particularly nasty one. So please get Amgen to have a thorough look at this compound.

        1. Mark Richardson says:

          If you contact me by e-mail, I would be happy to send you a link where you can download the full article free of charge (I have 50 or so of these to hand out). My correspondence information is featured with the abstract. I assure you that we have no intent to manufacture or distribute 4-AzPhe, but we did characterize it thoroughly. The thermal decomposition profile is detailed in our report.

        2. userdfx says:

          Quintus, how do you treat/dispose aqueous azide waste on scale?

          1. Quintus says:

            Diazotise it with conc. HCl, NaNO2

  9. Scott says:

    I still think you chemists occasionally make up some explosives just to get Accounting to pay for a new, much better [whatever]…

    “Reason for need: Previous unit destroyed in experimental accident, replacement is now [much better unit]”

    1. Mark Richardson says:

      You caught me. I told accounting that my 1998 Mercedes C230 was destroyed during the DSC and ARC experiments, but they refused to buy me a new one. Rude!

  10. PotStirrer says:

    I’ve always wondered about the carbon to nitrogen ratio rule of thumb for azide safety. Does a nitrogen atom in your molecule that isn’t part of the azide count in the calculation? If so, why? My personal experience is that synthesizing a pyridyl version of benzyl azide resulted in a pretty hefty explosion in our, at the time, brand new lab. So, apparently the answer is yes, but I still don’t know why. We did successfully make multigram quantities of t-butyl azide though.

    1. milkshake says:

      it is the strong electron withdrawing effect of pyridine ring most likely the culprit, not the extra nitrogen. As you know pyridine is more electron-deficient ring than nitrophenyl. The stablity of azides follows the rule: alkyl azides > aryl azides >> acyl azides. I was making 4-azido-substituted pyridine once (I could not use ammonia exchange because of a solvolysis of labile difluoromethyl group nearby) and I made sure to decompose the crude azido to aminopyridine in situ, by Staudinger with PMe3

  11. Derek, thanks for this post and highlighting this potential explosion danger. This article has caught much attention over the past few weeks, with lots of discussion. So we have made an entry of this chemistry into the Pistoia Alliance Chemical Safety Library (which data is now pushed to PubChem as well).

Comments are closed.