We have two comments about the recent letter by Alan R. Katritzky and Mirna El Khatib concerning the organic azide workup explosion at the University of Florida, Gainesville (C&EN, Jan. 9, page 4). We commend the authors for pointing readers to the recent book “Organic Azides, Syntheses and Applications” by S. Bräse and K. Banert (John Wiley & Sons, 2009). The first chapter provides a good summary of hazard testing and evaluation, including an excellent list of general precautions and rules of thumb for azides and, by analogy, other energetic materials.
Dow Chemical has a reactive chemicals program that has been refined over the past 45 years and has been highly effective in reducing highly reactive chemical incidents. On the basis of our experience, we suggest additional resources and approaches for the hazard evaluation of reactive chemicals.
First, strategies for an overall reactive chemical hazard assessment based on scale and energy release potential have been described in “Selection of the Proper Calorimetric Test Strategy in Reactive Chemicals Hazard Evaluation” by David J. Frurip (Org. Process Res. Dev., DOI: 10.1021/op800121x). Many of the described tests are covered by ASTM standard testing procedures (see “The Role of ASTM E27 Methods in Hazard Assessment,” Parts I & II in Process Safety Progress, DOI: 10.1002/prs.10046 and 10.1002/prs.10058). These papers describe how many chemical companies define safe operating limits using a sensible and balanced formalized method. Several external contract laboratories have experience in testing and post-testing evaluation for this reactive chemicals hazard evaluation method. These include Fauske & Associates, Chilworth, HEL, ioMosaic, and Fike, among others. Using these labs can be advantageous and avoids the pitfalls of “self-testing.” For example, differential scanning calorimetry (DSC) results can be impacted greatly by sample encapsulation procedures and material of construction (Thermochim. Acta, DOI: 10.1016/0040-6031(80)87111-5 and 10.1016/0040-6031(88)87443-4).
Second, this event demonstrates the advantage of a “management of change” (MOC) policy, which assists successful and safe start-up of new procedures and chemistries. A MOC policy is a formal, structured process to evaluate any change (in this case, a workup in acid) by knowledgeable experts for its consequences. Completion of the MOC process is required before any new chemistries or procedures can be implemented.
This MOC policy is used currently with great success at Dow and by a large number of chemical companies worldwide. At the R&D scale, this MOC policy may sound burdensome, but the effort and review are scaled to address the perceived risk associated with the reviewed activity, and the policy is appropriately and efficiently applied. This sometimes consists of simply requiring a conversation with at least one independent and knowledgeable colleague.
By David J. Frurip and David B. Gorman
Katritzky and Khatib reported an explosion when handling benzotriazole-1-sulfonyl azide despite DSC data showing the compound to be thermally stable below 95 °C. However, when it comes to exploring potential hazards of azides, one needs to consider more than just its thermal stability on melting.
The properties of thermal stability, shock sensitivity, and explosivity are related, but not in a direct manner. For example, picric acid melts at 122 °C without decomposition but is shock-sensitive and explosive. Some compounds are shock-sensitive but not explosive, and others are explosive but only mildly shock-sensitive.
The DSC data given really say nothing about the potential of this azide to be either shock-sensitive or explosive. Most organic azides will thermally decompose, usually at temperatures below 200 °C. If the DSC experiment was carried out through the complete decomposition of the azide, the decomposition exotherm would in general yield the temperature at which decomposition is detected, the total energy released, and an estimate of the rate at which the energy is released.
Total decomposition energies above 1–2 kJ/g should be of greater concern. A broad decomposition exotherm in the DSC indicates that decomposition occurs slowly. However, a sharp exotherm indicates rapid decomposition and is indicative of explosive potential. A DSC experiment run to high enough temperatures to look for decomposition should be the first step in exploring potential hazards of any compound of unknown stability. Should high decomposition energies or rapid decomposition rates be found, or if the compound belongs to a class of materials like azides, diazo compounds, etc., that are known to be hazardous, then shock sensitivity should also be examined. This can be done with a standard drop-weight shock sensitivity test [ASTM E680-79(2011)e1], but the crude, old-fashioned hammer and anvil test can also be tried if the appropriate equipment or testing service is not readily available.
In this latter test, a small amount of sample (1–10 mg) is placed on an anvil and hit sharply with a hammer. If the sample is stable, it should appear unchanged. However, if the sample has darkened or charred, or if a loud report is heard beyond the normal bang of the hammer hitting the anvil, then the compound should be considered shock-sensitive. Shock sensitivity is a good indication that a compound might be explosive, but it is not proof.
Shock sensitivity simply means that a mechanical force can impart enough energy to a compound to make it decompose. The rate of decomposition is what is important. A good explosive is one with low shock sensitivity so it can be handled safely, but with high explosive power to do the intended job. Unless further testing is done, any compound showing shock sensitivity and high energy release should not be handled in other than milligram quantities except by someone who is an expert in the area of explosives.
All azides and diazo compounds should be handled as if they were shock-sensitive and explosive unless there are clear data to the contrary. When using sodium azide in a reaction, one should destroy any residual azide ion or hydrazoic acid with nitrous acid, followed by destroying any excess nitrous acid with urea during the workup so there is no hydrazoic acid or azide ion present when the product is isolated. When a particular azide structure is desired but found to be shock-sensitive, putting large inert substituents such as phenyl, phenoxy, or alkyl groups on the compound can result in a less sensitive material.
Hydrazoic acid is a water-soluble liquid that boils at 37 °C and can spontaneously explode when isolated. Whether hydrazoic acid or the acid workup has anything to do with the explosion, or whether the sulfonyl azide product in question is inherently shock-sensitive and explosive, is not clear from the information given. However, the properties of hydrazoic acid make it seem unlikely that the acid would end up mixed with the solid product in the bottom of a flask after an aqueous workup. Further testing of the benzotriazole-1-sulfonyl azide is required to definitively answer that question.
By Gary R. Buske