Some substances are referred to as explosive because under certain conditions they are liable to undergo a rapid chemical reaction which produces large quantities of hot gas. There is not a precise definition of the term explosive, but the three most important characteristics of explosives are the speed of the transformation to gas, the quantity of gas produced and the temperature of the gas. Unfortunately there are two distinct types of explosion which are referred to as deflagration and detonation and these tend to be confused in the literature on account of the fact that the effects of deflagration and detonations are not always a clear guide to the nature of the explosion. The destructive effects of a deflagration can occasionally be comparable to those of a detonation, but in general the characteristics of the two phenomena are rather different.
During a deflagration, the chemical reaction zone travels from one particle of the substance to another by thermal conduction and convection; therefore, the physical state of the substance has a marked effect on its explosive power. The velocity of the deflagration front in a solid in the open air is quite different from that in a solid contained in a closed vessel, because the velocity of reaction front increases with the pressure exerted on the solid by the combustion gases. Thus, in a closed container the deflagration front accelerates as pressure increases, but rarely exceeds 1m/s.
The strength of the container governs the maximum pressure rise, and hence the maximum speed and duration of the deflagration, which in turn governs the maximum air pressure generated by the explosion. Containment of explosives should therefore be avoided wherever possible. If this is not possible containers should be open or of low strength and should not be stacked high because this increases the confinement of the central and lower layers.
The characteristics of condensed explosive fall into two main groups the first of which contains chemical information such as composition and heat of formation at constant volume. The second group includes the physical properties of the substance and the conditions under which it is stored (exploded). The first group, together with the quantity, determines the consequences of an explosion should one occur, while the second relates to the probability of an explosion and hence the safety of the storage arrangements.
Since an explosion is the rapid conversion of a solid into a gas at high temperature, the basic parameters governing explosive yield are the quantity of gas produced and the heat released by the reaction, which determines the maximum temperature reached.
The quantity of gas produced is usually expressed as a volume at 0oC and 1 atmosphere pressure per 1kg of the explosive. It is denoted by Vo and is calculated from the chemical equation representing the explosive reaction, treating the gases as ideal gases.
The heat of the explosion is the quantity of heat released during the decomposition. It is determined by the chemical equation representing the explosion reaction in which the explosive products on the right hand side are in a physical state appropriate to the high pressure and temperature conditions of the explosion. Any water formed is in the vapour state. The heat of an explosion, measured per unit mass of explosive, is designated Qv and is closely approximated by the heat of reaction; however, there is not a unique way of determining Qv and different thermodynamic relations yield different results.
There is little point in striving for high accuracy in a calculation of explosive yield because not only is the final thermodynamic state of the products uncertain, but the physical state of the containment, storage conditions etc. can have a marked effect on the energy released.
Calculation and experiment have determined the energy of explosion of most common explosives, but in practice not all of the energy which could theoretically be released is actually converted into a blast wave. There are numerous reasons for this, but the main one is that only a fraction of the mass of the explosive actually explodes - the rest is dispersed. The ratio of actual energy released to that theoretically available is usually referred to as the explosion efficiency.
The energy released by an explosion is therefore the product of the mass of the explosive, the energy of explosion of 1kg of the substance and the explosion efficiency. The specific explosion energy is usually measured in terms of the energy of detonation of TNT and is referred to as explosive power: -
Es = Energy of decomposition of 1kg of substances.
Etnt = Energy of detonation of 1kg of TNT.
Since the consequences of explosions are documented in terms of the mass of TNT, the consequences of explosions of other substances are most conveniently determined by calculating an equivalent mass of TNT. This is defined as:-
TNT Equivalent = M x (explosive power) x (efficiency).
Once the TNT equivalent of a substance has been established, the effect of an explosion of M kg of that substance is easily determined by reference to documented effects of an explosion of an equivalent quantity of TNT.
An explosion by its very nature releases a large quantity of hot gas in a very short period of time. This generates a pressure front which moves away from the source. As the explosion progresses on a microsecond time scale, the temperature and pressure disturbances move out with increasing speed and catch up earlier ones, resulting in the formation of a steeply rising pressure front or shock wave, which moves out with constant velocity.
Over the years data have been collected on the peak overpressure versus distance for various types of explosion. Although these exhibit considerable scatter they have been plotted in terms of equivalent mass of TNT. (R. Merrifield: Simplified calculations of blast induced injuries and damage, HSE Specialist Inspector's Report Number 37). Some typical results are shown below:
|Scaled distance (m.kg-1/3)||Peak overpressure (kPa)|
The consequences of blast waves are often tabulated in terms of the effect of different levels of overpressure on people and buildings. Examples can be seen in the list below:
|Serious level of death||500 mbar|
|Dangerous level (1% lethality)||140 mbar|
|Windows usually shattered (all sizes)||35-70 mbar|
|Frame distortion of steel framed buildings||140 - 170 mbar|
|Rupture of oil storage tanks||210 - 280 mbar|
|Steel framed buildings pulled from foundations||210 mbar|
|Rail cars overturned||490 mbar|
|Complete destruction of all non-reinforced buildings||700 mbar|
Damage to humans caused by explosions may be classified as primary or secondary. Primary damage includes being blown to pieces, being hurled through the air into a hard object or structure and suffering death via rupture of body organs. Inhalation of hot toxic and dusty gases, missile impact, collapse of buildings and fire may cause secondary damage. Experience has shown that roofing tiles, bricks and broken glass account for most missile injuries.
Sodium Chlorate is the most common explosive material found in mixed warehouses, it forms colourless hygroscopic crystals which are soluble in water (solubility is 790 g/l) and toxic. It melts at 248 oC and decomposes at 300 oC releasing oxygen. Sodium chlorate is strongly oxidising and mixtures containing combustible material are extremely sensitive to impact and flame. Fine powders and surfaces impregnated with sodium chlorate are likely to catch fire and/or explode spontaneously.
Sodium chlorate is normally transported in 25 or 50 kg steel drums but occasionally it is packed in paper bags lined with aluminium. It is classed as an oxidising agent and not an explosive, although its explosive and unpredictable behaviour is well known. There have been several incidents of sodium chlorate explosions and research, carried out by HSE has shown that, while drums of pure sodium chlorate engulfed in a small fire in the open are unlikely to explode, a small degree of confinement, such as that offered by a three sided roofed enclosure, is sufficient to cause a fire to produce an explosion.
The reaction mechanism of high temperature sodium chlorate is complex and not fully understood. The reactions which are generally thought to take part in the explosive decomposition are: -
|2NaC1O3||=||2NaC1 + 3O2|
|4NaC1O3||=||3NaC1O4 + NaC1|
|NaC1O4||=||NaC1 + 2O2|
Other possible reactions include: -
|4NaC1O3||=||2C12 + 5O2 + 2Na2O|
|2NaC1O2 + C12||=||2NaC1 + 2C1O2|
Many substances catalyse the thermal decomposition of sodium chlorate and some form explosive mixtures. Mixtures with organic materials such as sugar, sawdust, oil etc., and with inorganic substances, such as sulphur, finely divided metal acids etc., can be extremely sensitive to shock friction and/or heat and can burn or explode spontaneously. Spillage should be avoided at all cost and if they do occur should be removed with water by an operator wearing protective clothing.
The explosive power of sodium chlorate is usually assumed to be 0.14, indicating that the energy of decomposition is about 650 kJ/kg. The efficiency of sodium chlorate explosions is generally taken to be about 0.25, thus the TNT equivalent of a stack of sodium chlorate is: -
|TNT equivalent||=||M x 0.14 x 0.25 kg|
Ammonium nitrate is a hygroscopic colourless crystalline solid, which is very soluble in water. In the dry state it is non-corrosive, but, when moist, it reacts with various metals forming a variety of compounds, some of which are highly unstable (e.g. copper nitrate tetramine). The main decomposition product when ammonium nitrate is heated above 200oC is N2O, but above 250oC other oxides of nitrogen can be formed.
The most common form of ammonium nitrate is fertiliser. This exists in a variety of forms but these are classified into two groups according to the nitrogen content. All fertilisers with a nitrogen content of more than 28% are assumed to have the same hazard potential, although it is known that low density material and compounds containing potassium are more likely to detonate.
Pure ammonium nitrate is not shock or friction sensitive and cannot be induced to detonate under normal storage conditions; however, the following parameters increase its sensitivity: -
Fertiliser is generally considered to be less hazardous than pure ammonium nitrate unless it is contaminated with substances that make it more sensitive.
There is some confusion and uncertainty in the literature and in safety reports about the explosive power of fertiliser. This can be traced back to the question - can a stack of fertiliser detonate or only deflagrate? Experiments have shown that, to all intents and purpose, it is incapable of doing either unless at least some of the stack is heated above its melting point. Detonation, which is characterised by a supersonic pressure wave moving through the material, can occur only if the dimensions of the explosive are greater than some particular value known as the critical charge diameter. The corresponding diameter for molten ammonium nitrate is only about 10 cm.
Deflagration is not constrained by dimensions and is said to occur when a subsonic combustion generated pressure wave moves through the material. Under certain conditions the energy released and the damage caused by the two process (detonation and deflagration) in a sample of ammonium nitrate can be different, but, in hazard analysis, it is not usual to distinguish between them and to refer only to an explosion.
The consensus of opinion on ammonium nitrate hazards is that, in the event of a large fire at an fertiliser store, a pool of liquid ammonium nitrate will be formed at the side of the stack that is nearest to the fire. If this pool is struck by a high speed missile (e.g. something falling or part of a drum that has exploded) then a local explosion will occur sending a shock wave into the main fertiliser stack that has not melted. A suitable basis for the assessment of an ammonium nitrate explosion would be to use a TNT equivalence of 55% and an efficiency of 25%.
Stacks of ammonium nitrate in the open are assumed to be less likely to explode because the probability of an explosion initiator such as a girder falling into a molten pool is very low.
Organic peroxides are highly reactive, combustible and thermally unstable due to the presence of the unstable -O-O- peroxy link in their molecular structure. Some are low flash point highly flammable liquids, while others are classified as explosive. Most are liquids, but some are produced in the form of a paste. In the pure state nearly all organic peroxides are detonable, but their reactivity is suppressed by dilution or phlegmatisation with liquids such as water or phthaltes. Even so, essentially all of them are capable of self heating and runaway decomposition brought about either by temperature or contamination. The results of a runaway are a violent pressure burst of their container and a sudden release of hot flammable vapour which usually ignites spontaneously. If the container is particularly strong an explosion may occur.
All organic peroxides are characterised by a self-accelerating decomposition temperature (SADT) which tends to decrease with increasing packaging size. Clearly storage should be at a temperature well below the SADT. It is recommended that when the SADT is 20oC or less storage should be 20oC below the SADT. When the SADT is between 20 and 35oC storage should be at 15oC below the SADT, and when the SADT >35oC, the storage temperature can be 10oC below SADT or lower. If the temperature in an organic peroxide store rises above the SADT, due to, for example, a fire in a neighbouring building, a rapid decomposition can take place and give rise to three types of major accident hazard, all of which should be addressed in a safety report. These are: -
The convenient way of determining the consequences of an explosion is to calculate the equivalent amount of TNT. Unfortunately TNT equivalence fractions are not readily found for all peroxides, but the values in the table below can be used in hazard analysis.
Table 9: TNT equivalent of some common organic peroxides
|pure dibenzoyl peroxide||0.09|
|t-butyl-peroxy maleate (pure)||0.14|
|methyl ethyl keytone peroxide 60%||0.26|
|cyclohexanone peroxide 60%||0.13|
|dibenzoyl peroxy benzoate||0.25|
Use of a weighted average TNT equivalent for the contents of a mixed peroxide store will produce a conservative estimate of the blast potential.
Self-accelerating decomposition of a large quantity of organic peroxide produces a rapid rise in temperature and release of highly flammable vapours, which are likely to ignite. The result is a fireball with properties similar to fireballs from flammable hydrocarbon liquids:-
R = fireball radius (m)
t = duration of the fireball (s)
A = substance specific constant (~29)
J = atmospheric transmissivity factor
H = humidity
VF = view factor between fireball and target
x = distance between target and point on the ground under the centre of the fireball
k = (x2 + R2)1/2.
A fire in a peroxide store that does not result in an explosion or a fireball will burn very intensely and the thermal radiation from the flames may produce knock-on effects. Quantification of the hazard is difficult because the store may contain the flames, but this is unlikely. Complete failure of the roof is more probable and then the flames are likely to extend well above the walls of the building.
The extent of the hazard can be determined by assuming that the peroxide store forms a pool fire with a high burning rate. The difficulty with this approach is that burning rate data for most peroxides is sparse and at best uncertain. A figure that is probably conservative for most peroxide stores containing Type 1 compounds (see CS21) is 0.5 kg/s.m2. This is about 10 times the burning rate of hydrocarbon liquids. If the stores contains mainly type 3 organic peroxides, this figure should be reduced to 0.2. The rate of heat release can be obtained from the mass burning rate by assuming an effective heat of combustion of around 30 MJ/kg.
The burning rate can be used to predict the flame height for a pool covering the whole of the floor of a peroxide store:-
L = height of the flames (m)
Q = rate of heat release in kW
There is little or no data in the literature on organic peroxide pool fires hence uncertainty surrounds the emissive power of the flame pillar. In the absence or more reliable data, a figure of 200 kW/m2 over the whole length should be assumed. The duration of the pool fire should be accounted for in consequence calculations because at 0.5 kg/s.m2, the contents of a small store may be consumed quickly.