By Dr. Vyto Babrauskas
On August 4, 2020, much of the port of Beirut, Lebanon, blew up due to a massive ammonium nitrate (AN) explosion. On April 17, 2013, a portion of the town of West, Texas, was leveled when an ammonium nitrate storage facility exploded. Based on initial information concerning the Beirut disaster, we shall examine here what lessons could have been learned from Texas that might have averted the Beirut disaster.
Ammonium Nitrate Disasters and Safety Measures
In the early 20th century, disasters used to regularly happen in the manufacturing of ammonium nitrate. But in more recent times, manufacturers learned to be safety conscious and there have been very few accidents in connection with the actual manufacturing process. Where disasters continue to recur with regularity is when the material is either in transport or in storage. A study I did a few years ago demonstrated that 100 percent of the AN accidents in storage or in transport have been due to a single cause: an uncontrolled fire (1). There have been no other causes.
But if this is the case, it should not be hard to see that the solution needed is simple and robust. No accidents will occur if we make an uncontrolled fire to be an impossibility. This is not hard; in fact, it is fairly easy. One needs to do only two things:
- Only store the material in a fully non-combustible building.
- Make sure there are no combustible goods nearby which could ignite and burn.
It is not a challenging engineering problem to accomplish these things. We have known how to build fully non-combustible structures since the 1880s (2). And it again requires no advanced technical knowledge to make sure there is nothing combustible nearby that could threaten the storage with its flames, whilst burning.
Other safety precautions can be marginally helpful(3), but some are genuinely misguided. Foremost of these is the concept of setback distances, intended as a safety zone surrounding an explosion. Looked at very superficially, this might seem to be a good idea, but a closer examination reveals the opposite. Consider first if the management of combustibles is inadequate, there remains a risk of explosion. Obviously, if an explosion is immanent, the surrounding area needs to be cleared of all individuals. This means first-responders need to be sent into an area where they may be blown up. This is not consistent with protecting the lives of the first responders themselves. On the other hand, if combustibles have been successfully managed so that there is no risk of explosion, then explosion-safety setback distances clearly serve no purpose.
The Sad History of Disasters
Unfortunately, entirely preventable disasters keep recurring. Over the last century, there have been about 70 major fires involving AN in storage or in transport. Of these, 30 percent resulted in explosions, while 15 percent resulted in loss of life(4). Since enough data are available, these statistics can be taken as a direct indication of expected consequences of what will happen if a serious fire threatens stored AN. I consider that any activity which entails a 15 percent risk of involving fatalities to be deemed unacceptably risky, but evidently more education is needed to bring this message across.
In the U.S., the most recent AN disaster occurred on April 17, 2013 when a warehouse holding about 30 tons of AN in the town of West, Texas, caught fire and resulted in an explosion where 15 persons were killed and some 260 injured. A sizable fraction of the town was blown up, including two schools which were badly enough damaged that they needed to be demolished (5). The disaster was especially poignant since most of the individuals killed were either firefighters or individuals directly responding to assist the local volunteer fire department in their efforts. Subsequent to that, I published nine articles on the fire/explosion hazards of AN and how to prevent such disasters in the future. Evidently none of this information reached the officials in Beirut, Lebanon, since it hardly seems plausible that they would have known of a 15-percent probability of fatalities and simply accepted the risk. And, of course, it is deeply tragic that the Beirut tragedy was an order of magnitude worse than the one in Texas.
Texas vs. Beirut: Some Similarities and Some Differences
The two incidents shared a crucial similarity: In both cases, it was clear that uncontrolled fire was the proximate cause of the explosion. In both incidents, videos clearly show flaming from a rapidly worsening fire just prior to a detonation. The combustible materials in Texas were a wooden building plus diverse combustible agricultural commodities. In the Beirut incident, photographs indicate that building itself was not combustible. The early reports have not elucidated all the conditions, but it appears that piles of fireworks were burning. It is not clear what else was burning, but it is clear that the explosion was preceded by—and triggered by—a massive fire.
The AN in Texas was a fertilizer product, intended for sale to local farmers. The AN product in Beirut was confiscated from a vessel and was intended to be shipping to a mining operation in Mozambique, possibly as an ingredient for making ammonium nitrate/fuel oil (ANFO). This means the physical specifications of the prills were slightly different, but this would not affect the propensity for the material to explode. The biggest difference, of course, is that some 30 tons sufficed to level a sizable portion of West, Texas, whereas in Beirut some 2,750 tons were involved. There was also a substantial difference in casualties. In Texas, 15 persons died in the incident. In Beirut, the official tally is not yet available but initial estimates suggest about 150. Since there were nearly two orders of magnitude more material in Beirut, it is fortunate indeed that there were not two orders of magnitude more of fatalities.
Ammonium Nitrate Properties
Since AN is not in widespread general usage, it is useful to understand some of its basic properties and traits. Ammonium nitrate (NH4NO3) is a chemical which is not combustible. Therefore, reports saying that stored AN “ignited” are incorrect, since only fuels can ignite and AN is an oxidizer, not a fuel (6). AN can, however, explode (7), and the explosion can be massive, if the quantity is large. Some details now need to be considered. Since AN is an oxidizer, it can react with various fuels, producing heat in the process. But this is not how explosions proceed. The typical AN explosion is due to excessively rapid decomposition. In addition to being an oxidizer, the AN molecule is unstable and it can decompose, that is, come apart into fragments. Decomposition can be provoked by various means, but by far the most common is heat; this is termed thermal decomposition (8). Next one can ask: Where does the heat come from? Again, in theory, there are many possibilities. But in practice, what usually happens is that a large, uncontrolled fire occurs nearby, and this proceeds to heat the AN.
If the burning material is very close, then an additional complication arises. With normal fires, heat rises upwards. Then, if there is a ceiling above, the hot gas plume gets stopped by the ceiling and turned sideways. This is how fires spread in a wooden structure, for example. But AN has a special property. If a pile of AN is being heated by an adjacent fire, the material will get heated up, and will eventually melt at 169.6ºC (337.3ºF). Molten AN flows like a liquid. Like any liquid, it will flow downwards and establish a flow path governed by gravity. If this “river” of flowing AN then encounters some combustible, such as wood or cardboard, the fuel will ignite hypergolically. Hypergolic ignition is what happens with certain very reactive chemicals when ignition occurs upon contact of that chemical with some other material, without needing any external ignition source. Thus, if there is a fire involving wood, cardboard, or other fuels next to a pile of AN, fire spread will be substantively accelerated. Instead of just spreading up-and-across, fire will also spread along a second path, that of the flowing AN “river.” The result can be a much faster fire development than might have been expected for the particular fuel that is burning.
Ammonium nitrate is primarily used in two industries: as an ingredient for explosives-making and as a fertilizer in agriculture. For both applications, the substance is typically sold in the form of “prills,” which are small spherules of about 2 mm diameter. Chemically, the material is identical for both industries, though there are some physical differences, primarily in density and porosity, of the grades sold for explosives-making and for agriculture. But either grade is susceptible to explosion, including its most severe form, detonation. Detonation occurs when a reaction wave moves at a velocity greater than the local speed of sound in the unreacted material.
What might be surprising to individuals who do not work with AN is that it is not classified as an explosive, even though it can explode and these explosions can produce massive loss of life. The U.S. Federal government has a narrow definition (9) of what constitutes an ”explosive,” and AN is not classified as an “explosive” but, rather an “oxidizer.” This practice is the same worldwide, according to guidelines set forth by the UN (10), which specify that AN is classed as an oxidizer and not an explosive, unless the product contains combustible material in excess of a certain limit value. This is for two reasons: (1) AN is a less sensitive explosive than TNT or RDX. This means it requires significantly more energy to be put into the material to initiate an explosion. However, once an explosion is proceeding, this limitation will not offer much solace to the victims. (2) Classifying AN as an explosive is strongly resisted by the agricultural industry, since this would make it more difficult for them to use AN as a fertilizer. This, however, is also a poor reason, since a very similar product, calcium ammonium nitrate (CAN) is available for fertilizer use which shows no significant risk of explosion (11).
In his review, Médard (7) concluded that not a single case could be found of the accidental explosion of pure, unheated AN. In laboratory studies, on the other hand, it is easy to explode AN by mixing in large amounts of contaminants (8). But such contamination just does not tend to happen in reality. (It is sometimes claimed that the large explosion of AN which occurred in 2001 in Toulouse, France, was due to contamination. However, Guiochon (12), the world’s leading expert on decomposition of AN, found instead that the actual cause was due to a buried, unexploded bomb which detonated beneath the factory). What is inarguable is that the governing cause of AN explosions has always been uncontrolled fire.
Since the Beirut explosion, I have received queries about toxicity hazards from AN explosion incidents. AN explosions are caused by very rapid decomposition of the material. The chemical reactions involved in decomposition are complex and numerous8. In principle, the decomposition of ammonium nitrate produces some reaction products which can be toxic. These include not only nitric acid and ammonia, but also various oxides of nitrogen. In large concentrations, any of these would be toxic and could even be lethal, but the blast effect is such that these substances are dispersed very widely and do not accumulate in any particular location. So while there will be trace readings that can be obtained, there is not likely to be any surviving person who would be exposed to injurious concentration levels of the toxic chemicals. In the West (TX) disaster, for example, numerous individuals who were not close enough to the warehouse to be killed were injured and many of them hospitalized. But the injuries were primarily blunt trauma (from falling objects) or from flying glass. There were not any cases identified of injuries due to chemical poisoning. One might also ask about the dangers of being exposed to unreacted prills of AN. Here, again, the danger is very modest. AN can be an irritant, especially to mucous membranes, and would be toxic if ingested in sufficient quantities (13), but the latter seems highly improbable. Thus, while there are numerous extremely serious concerns about AN explosions, toxicity is generally not one of the primary concerns.
Ammonium nitrate is a chemical which can be highly dangerous, but only under certain circumstances. As with any explosible substance, it can detonated by an existing explosion reaction. In practice, however, the most likely danger is from fire. If an uncontrolled fire impinges on stored ammonium nitrate, the material can explode. The safety takeaway, which needs to be disseminated as widely as possible, is this:
Do not dispatch ammonium nitrate to any storage facility which cannot safely handle the material. The primary requirements for the storage facility is that the structure be non-combustible, and that no combustible goods or combustible structures be located close to the storage facility, since such a condition may result an uncontrollable exposure fire.
1. Babrauskas, V., “Explosions of Ammonium Nitrate Fertilizer in Storage or Transportation Are Preventable Accidents”, J. Hazardous Materials 304, 2016, 134-149.
2. Freitag, J. K., The Fireproofing of Steel Buildings, Wiley, New York, 1899.
3. Babrauskas, V., “The West, Texas Ammonium Nitrate Explosion: A Failure of Regulation,” J. Fire Sciences 35, 2017, 396-414.
4. Babrauskas, V., “Will Firefighters Be Any Safer Under the New Hazardous Materials Code?”, Fire Engineering 168:11, November 2015, 66-70.
5. West Fertilizer Company Fire and Explosion (15 Fatalities, More Than 260 Injured), Report 2013-02-I-TX, U.S. Chemical Safety and Hazard Investigation Board, Washington, 2016.
6. Babrauskas, V., Ignition Handbook, Fire Science Publishers/Society of Fire Protection Engineers, Issaquah, Washington, 2003.
7. Médard, L. A., Accidental Explosions, 2 vols., Ellis Horwood, Chichester, England, 1989.
8. Babrauskas, V., and Leggett, D., “Thermal Decomposition of Ammonium Nitrate,” Fire & Materials 44, 2020, 250-268.
9. Code of Federal Regulations, 49 CFR 172.101 Hazardous Materials Table.
10. Recommendations on the Transport of Dangerous Goods. Model Regulations. Vol. 1, United Nations, 2019.
11. Babrauskas, V., “The Ammonium Nitrate Explosion at West, Texas: A Disaster That Could Have Been Avoided,” Fire & Materials 42, 2018, 164-172 .
12. Guiochon, G., “On the Catastrophic Explosion of the AZF Plant in Toulouse, September 21, 2001,” 8th Global Congress on Process Safety, Houston, Texas, AIChE, 2012, 737-759.
13. Wexler, P., ed., Encyclopedia of Toxicology, 3rd ed., Academic Press, London, 2014.
VYTO BABRAUSKAS, Ph.D., earned degrees in physics and structural engineering and a Ph.D. in fire safety. As a researcher at the National Institute of Standards and Technology, he developed devices to measure the heat release rate of products and developed a computer program for modeling the development of room fires. He founded a consulting firm in 1993 and provides fire safety science expertise to fire investigation and litigation. His Ignition Handbook is widely used in the fire services, and he has two books forthcoming: Electrical Fires and Explosions and Smoldering Fires. He is currently based in New York City, where he is affiliated with the John Jay College of Criminal Justice.