This article originally appeared in the June 2005 issue of SPRAY.
Because it is still relevant information we are offering it again to our readers.
It has been reviewed by the author and independent sources for accuracy.
Three potent little letters: LEL. Not even an acronym. It is sometimes called LFL. Yet, they hold tremendous significance for the aerosol industry. They are a crucial factor in our flammability test methods, in formula development, in plant safety and, ultimately, in consumer safety. As an estimate, while aerosol-related lawsuits are very rare, probably more than 50% of them involve LELs in some way.
LEL is “quick-speak” for Lower Explosive Limit. Similarly, LFL stands for Lower Flammability Limit. The terms were originally called the “Lower Limit of Flammability,” as used in the watershed book titled Limits of Flammability of Gases & Vapors by H.F. Coward & G.W. Jones of the Bureau of Mines (U.S. Dept. of the Interior). It was first published in 1952 as 155 data-packed pages called Bulletin 503 for 70¢ per copy. In it, the authors pulled together and codified 368 references dating back to at least 1886.
As the name suggests, the LEL is the lowest concentration of gas or vapor which, when mixed with air and touched with an ignition source, will propagate a flame. The term erroneously suggests that an explosion will always follow; however, this is only the case when the containment vessel is damaged due to the pressure build-up. All combustions produce heat and many also produce additional molecules—typically, carbon dioxide and water vapor.
As an example, a 3.0 v.% mixture of ethane in air, uniformly distributed in a pressure-resistant, four liter glass globe, then ignited, will produce a pressure of 4.3 atmospheres, or 61.5 psi-g. The pressure will quickly subside as the gases cool, and the water vapor condenses. Even if this mixture occupied only 1% of a closed vessel, the almost instantaneous pressure increase would be about 0.615 psi-g—or 88.5 psf-g (pounds per square foot—gauge)—discounting any slight cooling effect of unburned gas (oxygen) in the container. This is enough pressure to do very serious damage to cars, rooms or houses. For hydrocarbon gases in air, flame propagation proceeds at the typical rate of about 12.5 feet per second (3.8 m/s) in all available directions from the source. In almost all cases, this means that all the hydrocarbon gas-air mixtures within the flammable range will have combusted within less than a second, often sending a shock wave or pressure front outward at very high speed, which can extend far beyond the range of the fire itself. Tests conducted by Factory Mutual Research Corp. have shown that the instantaneous release of isobutane, next to an ignition source, will result in about 80% of the gas being combusted. The remainder is outside the flammable range of about 1.86 to 8.4 vol.%.
The LEL for a given gas or vapor is not a fixed volume-% figure; it depends upon many factors. The propagation of flame depends upon the transfer of energy from the burned to the unburned gas. In a limit mixture, the amount of energy available for transfer is only just enough to maintain propagation. Anything that reduces the available energy will affect the LEL. The lowest LEL values will result from the use of vessels large enough that cooling from their walls is negligible. Large pressure-resistant globes, of at least two liters capacity, will give lower LEL results than smaller containers.
Closed glass tubes, containing LEL mixtures, will often show vibratory flame front movements. This is due to pressure increases, resulting in combustion at one end, followed by cooling at that end as the flame moves onward. In many cases, the flame will simply go out before reaching the far end. The direction of flame travel, in vertical tubes, is also important. Upward proliferation produces lower LEL results than downward travel. The gentle convection air currents, set up by rising flames, assist in their propagation. Conversely, for flames initiated at the top, the rise of heated air inhibits downward flame travel, and the LEL increases.
Atmospheric changes in air pressure have a small effect on LEL values. They tend to increase LEL and decrease UEL (upper explosive limit) values, but only when the initial pressure is about 8 to 10% lower than the average sea-level pressure of 760mm Hg absolute. If the atmospheric pressure increases, the LEL value will decrease. In some extreme situations, “non-flammable” gases will develop an LEL and UEL. For example, methylene chloride vapors are flammable near the Dead Sea (bordering Israel), or at the bottom of a 1500-foot deep mine shaft. In the same way, mixtures of HFC-134a and air, at about 15psi-g, will develop an LEL at about 12.5 vol.% gas and 87.5 vol.% air. (The UEL will be very slightly higher). Similarly, humidity conditions have almost no effect on the LEL, although the difference between very dry (0% R.H.) and very wet (100% R.H.) air can change the UEL of hydrocarbon and air mixtures by 0.4 to 0.5 vol.%. For aerosols, this effect is academic. Extremely dry LEL mixtures are typically harder to ignite—also academic.
The gas-air temperature is a complicating factor. Raising the temperature to 100°C (212°F) acts to reduce the LEL value by about 9%. At 400°C (752°F), the LEL is reduced by about 32%. This means that large ignition sources can cause the combustion of gas-air mixtures where the gas concentration is below the LEL level for ambient temperatures, especially in quiescent mixtures. Most flammability limit testers use either a 2–4mm induction coil spark or a very small flame as the preferred ignition source. In a few cases, a tuft of guncotton is fired by a spark or hot platinum wire to produce an instantaneous large flame. The use of electrically heated, white-hot platinum wire has been attempted; however, the results are unreliable. Rhodium wire, which melts higher, at 3,560°F (1,960°C), gives better results.
The effect of testing conditions for propane-air mixtures is shown in Table 1.
If propane is added to an “air” mixture containing 11.6 vol.% oxygen or less, combustion is not possible. For the butanes, this figure is 12.1 vol.%. Similarly, if propane is added to air that is under 98mm of absolute pressure (24.8″ Hg. of vacuum), combustion cannot occur. In another experiment, propane gave an LEL of 2.32 vol.% in atmospheric air (0.0 psig). When the air pressure was increased to 12 atmospheres (or 176psig), the LEL only decreased slightly—to 2.01 vol.%.
Many gas houses in the U.S. still have Halon 1301 receptacles near the ceiling to almost instantly extinguish any possible gas-air fire. At least one has carbon dioxide receptacles, for the same purpose, although the extinguishing process is somewhat slower. In the case of propane and isobutane, experiments have shown that there can be no LEL (or flammability) when certain amounts of gas diluents are added (Table 2).
In practice, larger amounts are needed to ensure that all susceptible areas of the gas house are quenched—even though diluent concentrations will vary from locale to locale—and to positively control the quickly expanding ball of potentially explosive fire (typically within about 0.12 seconds for the Halon units). On that basis, the engineering target is about 12.5 vol.% of Halon 1301 as a total room average. The Halon atomized gas-liquid is primarily directed at the most likely flammability sites, and typically suppresses the fire ball by the time it reaches a diameter of about 13 to 18″ (33 to 46mm).
In the case of nitrogen and carbon dioxide, huge amounts of gas would have to be very rapidly released, producing a suffocating atmosphere, and even then the flame suppression would probably be too slow for worker safety.
The fire triangle
Perhaps the most fundamental concept in fire technology is the relationship between fuel (gas), air (oxygen) and an ignition source. These elements can conveniently occupy the corners of what is termed the “fire triangle” or “flammability triangle” (Figure 3). Each is subject to certain parameters—outside of which combustion cannot take place. First, the fuel must be gaseous and within the flammability range; i.e. between the LEL and UEL. The gaseous aspect may sound strange, if one considers a piece of wood or a chunk of coal. However, these objects must first be heated to the point where they produce a flammable vapor at the surface before combustion can proceed. In the case of liquids, the temperature at which this occurs is called the flash point.
The volatile aliphatic hydrocarbons, in liquid form, have flashpoints below room temperature, ranging from propane (-156°F or -104°C) to n-octane (56°F or 13°C). Isobutane, which is the most important for the aerosol industry, flashes at any temperature above approximately -117° F (-83°C). The importance of flash points can be demonstrated using a small quantity of liquid n-butane (-101°F or -74°C). If as little as 1.0 gram of liquid n-butane (boiling point 31°F) is placed in a 1oz. bottle and a flame is passed over the top, an immediate flare is produced, up to about 10″ (254mm) high. Likewise, if that 1.0 gram is poured onto a concrete floor, it will very quickly evaporate, producing about 414mL of invisible gas. The gas is cool, and about 2.1 times as heavy as air. It spreads very rapidly (and surprisingly widely). In doing so, it will produce virtually all possible mixtures with the air.
Just taking the LEL concentration for calculating purposes, it will generate about 22.3 liters (5.9 U.S. gallons) of flammable gas-air mixture. This is sufficient to create a brief fire of about 10 U.S. gallons in volume if ignited. Obviously, such experiments should not be done or else be conducted with extreme precautions.
The second aspect of the classical fire triangle is air—or more specifically, the availability of oxygen. Dry air contains 20.95 vol.% of oxygen, even up to 35 miles above the earth, and this is quite sufficient to support combustions unless it is significantly diluted with non-flammable gases. In practical terms, the air can be considered to be a rather constant aspect of the fire triangle, and one that does not affect the LEL value of any flammable gas.
The final element involves the ignition source. While numerous origins are available, the most common gas house source is a sufficiently energetic spark: electrical, static or mechanical. In contrast, the usual source for consumer incidents is a flame, followed by electrical sparks.
Flames can be found in gas-fired hot water heaters, gas stoves, furnaces and propane barbecues and fish cookers. Sparks arise from electric toasters, on/off wall switches, electric motors under refrigerators or deep-freezers, washing machines, driers, coffee pots and other appliances. Many household fires—due to either aerosols or other products—are initiated at essentially the floor level. This is because the products are often used at floor level, such as certain insecticides, solvent cleaners and so forth. Also, nearly all flammable gases and vapors have densities from about 1.5 to 4.0 times that of air, enabling them to invisibly pour downward to the floor level, after which they can spread prodigiously. In fact, some states have recognized this aspect and have issued regulations that gas-fired hot water heaters must be elevated about 16″ (406mm) above the floor for safety reasons.
The ignition source must confer sufficient energy to the super-LEL gas-air mixture to overcome the reaction threshold and initiate combustion. For example, electrical equipment of the micro-ampere type is considered intrinsically safe in hazardous atmospheres because any sparks would be too “cold” to cause an ignition. While some experts contend that sparks from dropping a steel wrench or similar object onto a gas house floor could cause an ignition, the costly 3% beryllium, 97% copper tools will not. They do, in fact, create tiny sparks on a hard surface, but these are not sufficiently energetic to incite combustion.
From hot-wire experiments, it is thought that temperatures of over about 2,550°F (1,400°C) are required to ignite an LEL gas-air mixture. (However, within the flammable range—and particularly at the stoichiometric composition—lower temperatures are sufficient to cause auto-ignition.) Flames, household 110V sparks and many static sparks are hotter than this and thus are sufficiently energetic to start fires.
Fire engineers (including Fire Marshals), insurance inspectors and other experts generally have a very good grasp of these basic flammability aspects. They are applied to safety designs (ventilation, detection, dousing) and safety precautions (no smoking, no use of cotton or wool clothing, minimal occupancy) of gas houses. Beyond this, however, they are also considered in such matters as product labeling, home construction, home safety devices and so forth.
Until about 1960, the determination of LEL values was a relatively inexact science. This was due to the proliferation of experimental equipment and test methods, the unavailability of highly purified flammable gases, deviations in atmospheric pressure and other factors. Today, for a specific method, repeatability is about ±0.5% and reproducibility is about ±1.0%. The equipment is usually quite complex, and most parts are customized. The fact that different instruments and techniques have been used to obtain LEL data has been a source of concern and some confusion. For example, a major specialty gas supplier reported the LEL of n-butane in his catalogs as 1.6 vol.% in 1970 and as 1.9 vol.% in the year 2000. (The LEL of isobutane was given as 1.8 vol.% in both editions.)
During the 1960s, the LEL of isobutane was reported as 1.6 vol.% (2000 mL globe) and also as 2.2 vol.% (downward flame travel in a tube). The current value is often given as 1.83 vol.%. Some authors have elected to “err on the side of caution,” and have decided to use the lowest experimental figure.
The accompanying graph (Figure 4) provides a plot of LEL values from methane to n-decane homologous series, also showing LEL ranges for those compounds where this data is available. At first glance, the higher hydrocarbons would seem to have lower LELs and therefore be more flammable than the lower ones. This is not really the case for two reasons. First, if we calculate the LEL values on a weight-% basis (almost never done), all the numbers become very similar. Some representative data will illustrate this (Figure 5).
Chemical reactions, such as the burning (oxidation) of hydrocarbons, are always on a weight basis, as are the calories/gram or kiloJoules/gram heat outputs. The second reason is that the vaporization rate of the higher hydrocarbons is much slower, usually allowing ventilation to waft their vapors away soon after they are produced.
The graph does not differentiate between isomers, such as n-butane and isobutane, n-pentane, isopentane and neopentane, the five hexanes, the nine heptanes and so forth. Except for the butanes, there is a scarcity of LEL data on these compounds. For instance, pure neopentane is an extremely rare commodity, quite possibly unobtainable commercially. There is no incentive to determine an LEL value for it. Within experimental error, it is probable that the LEL values for every set of isotopes are either identical or very close together.
Significance of LEL values
The gaseous hydrocarbons, typically propane, n-butane and isobutane, have extremely low LEL values (Figure 6). Consequently, a small amount can produce a large volume of potentially flammable gas-air mixture. An often used example is where about 9.2 grams of liquid n-butane or isobutane (roughly16mL) is poured into a standard 55 U.S. gallon (204 liter) open-top drum. The liquid will almost immediately evaporate, producing about 3.81 liters of pure gas. If this is stirred up, using a canoe paddle or some similar device, to give an LEL (1.87 vol.%) volume of 204 liters, then the entire gas-air content of the drum will become potentially flammable. If ignited, the flame volume will typically be about 300 liters (80 U.S. gallons) with fire tongues easily reaching heights of 5–6 feet.
To complete the example, if twice the amount of butane is added (18.4 grams), then stirred into the air mass in the drum, the composition will closely approach the stoichiometric concentration needed for complete combustion. The consequences of ignition will then be much more dramatic, including a shock wave, faster flame propagation, and easily three times the LEL flame volume.
For propane, n-butane and isobutane, the problem of low LEL values is exacerbated by their gas densities, which are 1.5503, 2.076 and 2.011 times as heavy as air (at 60°F or 15.6°C). Once released, the gases tend to sink slowly in air. However, if they are sprayed or otherwise atomized, dilution with air quickly follows. The density of the gas-air mixtures will vary, but at the LEL of 1.87 vol.% for isobutane, the mixture is only 1.9% heavier than air. The settling rate in tranquil air then becomes miniscule or even debatable.
It is possible to take photographs, and even motion pictures, of hydrocarbon gas-air clouds, presumably with coated plates sensitive to ultraviolet molar extinction coefficients at about 265 millimicrons. In one moving picture example, an aerosol, sprayed upon a large flat surface, produced yellow-tan clouds of isobutane, between 8″ and 12″ (250 to 300mm) high, and having roiling, rounded tops around 6″ to 10″ in diameter—with little or no evidence of gradation into gas-air mixtures. There appeared to be some degree of cohesiveness. The surface characteristics could be likened to those of unstable, drained and severely coalesced foams with macro-bubbles—except these rounded gas structures are vastly bigger.
The aerosol industry also uses pentanes, hexanes and heptanes, often as isomeric mixtures (Shell, et. al.) or as distillation cuts (ExxonMobil, et. al.). For example, n-heptane vapor is 3.45 times as heavy as air and has a flash point of 25°F (-4°C). The vapor is relatively slow to evaporatively generate, but once formed, can progress along floors and down ramps or stairways in a relatively thin layer of supra-LEL content.
In one warehouse episode, a gaseous mixture of hexanes and heptanes drifted across a concrete floor for a distance of almost 80′ (24 meters) before contacting an electrical switchboard that ultimately caused it to ignite. If a floor fire does occur, the flames will travel fanwise and back to the source at the typical rate of about 12.6′ per second (3.85 m/s), making it virtually impossible for people to get out of harm’s way.
Serious plant explosions have occurred as a result of traveling supra-LEL gas-air mixtures—not only of the volatile hydrocarbons, but also of ethanol, diethyl ether, dimethyl ether and acetone. In fact, diethyl ether vapors have an insidious degree of cohesiveness that will allow the vapors emanating from a small, open container to travel over 26′ feet (8 meters) along a laboratory work-bench to an ignition site. This unusual characteristic has caused most toll packers to eschew the filling of diethyl ether (DEE) into cans of engine starter and similar products.
The volatile liquid hydrocarbons, now used rather widely in mold releases, certain adhesives, certain de-dusting type cleaners and other specialty aerosols, are often treated the same as the normally gaseous types; for instance, stored in pressure resistant bulk tanks, particularly in the case of n. and isopentanes. Drums of mixed hexanes (boiling range beginning at 121.5°F or 49.7°C) are known to deform—and in at least two cases, to leak—if stored in the hot sun during summer months. The volatile liquid hydrocarbons are markedly different than the normally gaseous ones in that they can remain as liquids during aerosol productions. Many years ago, during the production of a hair spray that included 15.0% isopentane and 5.5% carbon dioxide, two major safety hazards were encountered. Upon their release from the T-t-V gasser about 1.0 gram of isopentane sprayed out and onto the gas house floor. It formed a pool of liquid, cooled by the evaporation of a portion. The ventilation system was obviously incapable of removing the liquid, which was being continually replenished as production continued. Inevitably it got onto the boots worn by the gas house operator, making him even more nervous.
The cans then traveled into the main building and into the hot water bath. They still contained some liquid isopentane in the well of the valve cup, which sizzled into gas as soon as the can submerged in the bath. It was estimated that about 1.0 lb. (454g) of isopentane vapors were generated in the hot tank every three minutes or so. When a “cup blower” was installed between the gas house and the main establishment, liquid isopentane was blasted out of the cup—solving the hot-tank problem—but the mist landed on the ground below the blower. This could be tolerated fairly well during the hot summer months. However, as winter weather developed, the isopentane evaporation rate decreased precipitously, and some actually seeped through holes or porous areas of snow to remain for long periods as trapped (flammable) gas. Much to the disappointment of the marketers, these formulas had to be changed to eliminate the isopentane ingredient.
The LEL of flammable gases and liquids is one of the most important safety concepts of the aerosol industry. Around the world, many industrial and consumer accidents could have been avoided if LEL aspects were better understood and implemented. Since over 10 billion aerosols per year (worldwide) (Editor’s note: More than 15 billion in 2018) are now made using hydrocarbon gases or volatile liquids, LEL technology affects virtually every aerosol plant on the planet, and it is an important factor to be considered in our efforts to provide continuing consumer protection. SPRAY