Post Tagged with: "Azides"

Explosive Solutions

mercury azides

Instead of starting at the beginning of a paper I want to kick off this commentary with a statement from near the end:

Caution! Covalent azides are potentially hazardous and can decompose explosively under various conditions! Especially Hg2(N3)2, α– and β-Hg(N3)2, and [Hg2N]N3 in this work are extremely friction/shock-sensitive and can explode violently upon the slightest provocation. Appropriate safety precautions (safety shields, face shields, leather gloves, protective clothing) should be taken when dealing with large quantities. Hg compounds are highly toxic! Experimental details can be found in the Supporting Information.”

This wonderful statement appears in a recent publication by Professors Schultz and Villinger at the University of Rostock in Germany. They discuss the preparation of mercury azides and the azide of something called Millon’s Base. This compound was new to me and it turns out to be nitridodimercury hydroxide, [Hg2N]OH.2H2O, which Millon1 discovered by the reaction of mercurous oxide and ammonia in the mid 1800s. In a classic example of understatement the authors’ state that as is the case with most transition metal nitrogen compounds the extremely low energy barriers to explosive decomposition result in difficulties in the isolation and manipulation of said species! Curtius, of rearrangement fame, was apparently the first person to isolate mercury azide Hg2(N3)2 from the reaction of hydrogen azide and mercury2. I guess this was after the discovery of his famous rearrangement.

Structural data for this compound is available from x-ray and revealed two modifications, called α and β. Due to its lability the β modification has not been fully characterised. Schultz etal have now rectified this situation and also report the preparation of the azide salt, [Hg2N]N3 of Millon’s base. They prepared α & β-Hg(N3)2, the latter compound by slow diffusion of aqueous sodium azide into a solution of mercury (II) nitrate separated by a layer of aqueous sodium nitrate. In this synthesis one wonders how any yield was obtained because when the needles of β-Hg(N3)2 begin to form in the lower mercury(II) nitrate layer spontaneous explosions occur during crystal growth. If you want large crystals of either modification, usually obtained by slow crystallisation, I would not recommend it as apparently large crystals seem to explode when you look at them the wrong way, even in solution they detonate. Explosive solutions would be a great name for a company! Anyway, in spite of these difficulties an X-ray structure along with a melting point was obtained.

Turning now to the synthesis of the azide of Millon’s base the authors note that the normal method always produced a mixture of the two modifications. Pure α-[Hg2N]N3 was obtained by treatment of α-[Hg2N]Br with concentrated aqueous sodium azide for 300 days, so you need patience when dealing with these compounds, not only because they are explosive but they suffer from long reaction times. However starting with β-[Hg2N]NO3 the reaction was faster, only taking 4 days for the exchange with azide but produced a mixture of modifications. However, they did manage to obtain both modifications.

Elemental analysis could not be carried out due to their explosive nature and both modifications are sensitive towards heat, shock and especially friction. The bigger the crystal the more sensitive it is. However, slow heating in a DSC instrument showed that they are stable up to 283°C for the β form and 313°C for the α. Rapid heating in a closed vessel caused violent heavy detonation accompanied by a bright blue flash.

The paper has some fascinating x-ray pictures of all the molecules discussed and allowed determination of the N-Hg bond lengths. Together with the chemistry and the dangers involved in this chemistry, a great piece of work has evolved into a wonderful very readable paper. Congratulations to all who participated.



1      E. Millon, J. Prakt. Chem. 1839, 16, 58.

2      T. Curtius, Chem. Ber. 1890, 23, 3023

Life in Chemical Development, Part 2.

In the first part of this little series I recounted my experience with two steps of a four-step sequence, now I would like to move on to the last two steps: The preparation of a benzyl chloride and it’s conversion to a benzyl azide.

If you remember I had to convert 7098 kg of the benzyl alcohol ultimately to the azide. According to the plan:

Now benzyl halides are well known for their lachrymatory properties and this one made me cry just thinking about it. All that was required was to walk past the building, where it was being produced, to burst into tears and I had to run 46 batches (1.02 kMol) to make this stuff plus 9 for use tests. In fact we made the chloride then almost immediately concerted it to the azide.

As part of the safety checks in the pre-reaction control of the equipment the conductance of the enamelled stainless steel reactor was checked to make sure there were no cracks in the enamelling, it was deemed to be ok so we carried on. The alcohol was placed in a 630 L reactor and 312.8 kg of 37% hydrochloric acid was pumped in. The solution was heated slowly to an internal temperature of 90-93°C ( to avoid loosing too much HCl) and held there for 5 hours. During the reaction a two-phase system formed and we all cried. The product was on the bottom and it was separated from the acid after cooling to 40°- 45°C because the compound solidified at 37°C. It was then filtered and the pH adjusted to 9-12 with 30% NaOH solution and stored at 40°C as a two-phase system with water with minimal stirring and constant pH adjustment maintaining the 9-12 range. In the meantime we got things going for the conversion to the benzyl azide, more about that later.

When we examined the filter from the very last reaction we observed bits of blue glass. I hear you say “not again”. I don’t seen to have much luck with enamelled reactors. Well this time we were really lucky, and I mean really. Have a look at these two pictures.

The hole was a hairline crack in the enamel. Now this did not show up in the conductivity tests as it was right up at the top of the reactor where the stirrer joined together with the motor and could not be reached with the equipment we had, a pathetic excuse really. Maybe we should have used, you know that beer that reaches places that other beers can’t.  Remember under the enamel is stainless steel and we were using almost boiling 37% hydrochloric acid. So the acid seeped through during the course of the 46 batches and started munching away at the steel. The metal was so thin that if you pinched it between thumb and forefinger you could move the bottom part back and forward. I would say that one more reaction and the stirrer would have broken off at 100 rpm making God only knows what kind of mess. Furthermore it is well known that the presence of iron (rust) benzyl halides decompose exothermically at quite low temperatures. I can’t remember the exact temperatures but it moves the decomposition point (where the exotherm begins in DSC measurements) down about 50 or so degrees and increases the size of the exotherm markedly. So I guess we were lucky on two points, we stopped just in time and we were using steel with a very low iron content. After I saw this and realised the implications I my knees started knocking together and I staggered across the road to a pub and had a few stiff drinks and went home where I continued the treatment.

Back to the chemistry: Working with azides is particularly dangerous because of potential explosion and health hazards. Sodium azide is a very nasty compound. It is a CNS depressant and breathing the dust causes almost immediate breathing problems amongst others, see this page for more information, azides. Furthermore it also contains traces of hydrogen azide, which has similar biological behaviour to sodium azide but has the pleasant habit of being shock sensitive and hence explosive. The stirrer episode was bad enough; and we were using 70 kg of sodium azide per batch, my poor knees (never mind the liver). Even at pH 9 or above one can still detect HN3 in the gas phase. For the reaction we had an extensive gas washing system with 4 washers filled with 30% sodium hydroxide solution through which the exhaust went. At the end of this chain we periodically monitored for the presence of HN3 using ferric chloride spot tests, which are very sensitive for this compound. I’m happy to say that at the end of the chain we never detected any HN3. The reactor was specially made out of high quality tantalum steel, where the heavy metal content was minimised so we hopefully avoided the formation of heavy metal azides, I do not know if tantalum azide exists (perhaps someone who reads this may know) and heated glass tubing was employed for the transfers.

We threw the following into a 630 L reactor; 200 kg water, 2.6 mL of 30%NaOH solution, 700g tetra-n-butylammonium bromide, 70 kg sodium azide and a pH electrode. After heating this mixture to 90-95°C internal temperature and added the alkaline mixture of the benzyl chloride to it within 60 minutes. The pH drifted during this reaction and it was constantly monitored and kept between 9 and 12. The reaction is exothermic and the temperature control was also monitored closely during the 2 hour stirring at 90-95°C.

We then cooled to room temperature and filtered the lower organic phase (this time no glass was observed!) and removed the aqueous layer. This time everything went ok and from 55 batches we obtained a total of 9284.64 kg with an average purity of 94% and an average yield of 97.9%. All of the batches were released for the next step by QA. At last I was almost finished, I still had to dispose of all the azide containing waste from all the gas washers and all the water layers and reactor cleaning! This was really funny. We disposed of it by treatment of the waste with 37% hydrochloric acid and sodium nitrite, generating nitrogen, laughing gas and various other oxides or nitrogen that were washed out by the exhaust treatment. This was another foaming reaction, but by this time I was immune to foaming, didn’t worry me anymore. The aqueous phases went down to the water treatment plant.

There it was finished at last, with enough material for my colleague to play with. There is still more to tell about this chemistry but that will be part 3.

I hope you enjoyed my ramblings and look forward to many comments!