Wednesday, February 15, 2017

A report from Bristol on that TATP incident

Via Jyllian Kemsley's The Safety Zone blog, a writeup of that TATP incident from senior people at the University of Bristol. It was written by the dean of the Faculty of Science (Prof. Timothy Gallagher) and the head of the school of chemistry (Prof. Nicholas Norman). I'm going to excerpt this, but you should go over there and read the whole thing: 
On 3 February 2017, a graduate student in the School of Chemistry at the University of Bristol was carrying out a literature procedure to oxidise an aldehyde to the carboxylic acid using aqueous acidified chlorite. The experiment was carried out on a 5 mmol scale (just under 1g of aldehyde) and risk assessments identifying all hazards had been undertaken and signed off by both student and supervisor. The reaction solvent was acetone (50 mL). 
Part of the procedure involved adding a quantity of 30% hydrogen peroxide (H2O2) solution to remove some of the by-products of the reaction, whose presence was (apparently) associated with a yellow colour (possibly including chlorine dioxide). The literature indicated that H2O2 be added until this yellow colour had disappeared, which should have required about 1 mL of peroxide solution. 
The student, focusing on the yellow colour, which did not completely disappear, continued to add hydrogen peroxide solution until about 50 mL had been added. During workup to remove the solvent, the student realised that the solvent volume was not decreasing and that the liquid was becoming viscous, and so likely contained far more “product” than was expected. GCMS analysis indicated the presence of triacetone triperoxide (TATP), and it was estimated that this could amount to 30-40g if all the excess H2O2 had reacted with the acetone solvent. 
At that point, the graduate student immediately alerted the supervisor, who escalated this to the Head of School. A series of decisions were made and actions taken that resulted in the disposal of the suspected TATP by means of a controlled explosion carried out by the emergency services. 
Nobody was injured and no damage was done in the lab. Although the TATP presented an explosion hazard, the risk of explosion was considered minor due to all material remaining in solution; TATP is far more sensitive to detonation as a solid. Immediate disposal was warranted, however, due to the risk of precipitation/crystallisation of a solid material.
The authors have some statements that I will summarize (errors made are mine)
  1. The student was overfocused on removing the yellow color in the reaction, even as the risk of H2O2 + acetone had been identified.
  2. The role of the acetone was overlooked. 
  3. When the student recognized how they had made an error, they reported it immediately to their supervisor. "This was highly responsible – the most important thing done – and shows the value of investing in developing and fostering a culture in which colleagues recognise errors and misjudgements, and they are supported to report near misses."
This seems like a reasonable response and a fairly quick one, all to the credit of the University of Bristol.

I'd sure like to know what the "workup to remove the solvent" was. Did the student put this material on a rotary evaporator?

This incident reminds me of a favorite list of "what if" questions from McConville's "Pilot Plant Real Book", including the question "Consider the possibility of operator error - over charging or undercharging raw materials, adding materials in the wrong order, omitting a component, overheating, holding for too long at reaction temperature, opening or closing the wrong valve, etc."

Finally, it would be great if there was some kind of central repository of chemical incident information. This report would certainly be a good candidate for inclusion in it. I wonder if the chair of the department of chemistry at UCLA has such a report to file? 


  1. I assumed rotovapping - I thought they said that they were trying to pull off solvent but the mixture didn't seen to decrease in volume, but only to get more viscous; that was when they realized that something was wrong.

    This incident seems like one of those moments right up there with this: I guess everyone's glad, though, that he didn't just leave it there and go to lunch.

  2. again, the reaction of the student, boss, the university was completely overblown and unnecessary, given the amounts of materials, but I guess everyone wanted to handle this by the book. Diluting the react mix with alcohol and pouring it down the drain with running water would be appropriate. I would not do this with 10 liters of potentially hazardous mixture, but 100mL of mix is kind of a joke - adding half a liter of methanol would be enough to render it innocuous.

    The yellow color in chlorite oxidation is most likely ClO2 from a combination of acid and chlorite. Additives like sulfamic acid are good at removing Cl2 and ClO(-) but they do not do anything for ClO2. But chlorine dioxide is volatile, on gram scale rotovaping it is perfectly reasonable, only on kilo scale I would worry about hazards of ClO2. Also, acetonitrile is far better solvent for oxidations than acetone, I have used aqueous acetonitrile for chlorite oxidations and it worked very nicely.

  3. A useful repository for information: I think no longer updated but has some good knowledge for process chemists.

  4. "Finally, it would be great if there was some kind of central repository of chemical incident information."

    Seems like this would be a valid thing for the ACS to do as part of its mission. Maybe some of the C-suite execs could lead the charge.