Combustion Properties of Solid Fuels PDF

CHAPTER 5 Combustion Properties of Solid Fuels Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept. KEY TERMS cellulosic, p. 125 p...

CHAPTER 5 Combustion Properties of Solid Fuels Courtesy of Jamie Novak, Novak Investigations and St. Paul Fire Dept. KEY TERMS cellulosic, p. 125 point of origin, p. 138 spontaneous ignition, p. 132 char, p. 128 pyrolysis, p. 125 synthetic, p. 143 flame resistant, p. 156 pyrophoric, p. 132 thermoplastic, p. 145 piloted ignition, p. 128 soot, p. 143 thermosetting (resin), p. 125 OBJECTIVES After reading this chapter, you should be able to: Describe different modes by which fuel vapors are generated from a solid fuel. Define pyrolysis. Demonstrate a clear understanding of the combustion properties of wood, paper, plastics, paint, metals, and coal. Differentiate the ignition and combustion processes of thermoplastics from those of wood. Describe the reasons why wood does not have a fixed ignition temperature. Explain the importance of dust particle size and concentration in dust explosions. Explain the role of flame color and smoke production during fire development. 123  For additional review materials, appendices, and suggested readings, visit www.bradybooks.com and follow the MyBradyKit link to register for book-specific resources. Register for MyFireKit by following directions on the MyFireKit student access card provided with this text. If there is no card, go to www.bradybooks.com and follow the MyFireKit link to Buy Access from there. I n addition to the liquids and gases examined in Chapter 4, there is a large group of very common fuels for which accurate and precise combustion data cannot generally be tabulated—solid fuels. Many of the typical combustion properties may have little or no meaning for this group. For this reason, it is not often possible to list such familiar properties as flash point or explosive range for the most common fuels of all: wood, plastic, paper, and the like. Despite this limitation, the properties of these fuels are of the greatest importance, and to some extent they can be described, either by numerical data or in general terms. The ignition and combustion of solid fuels are more complex than that of liquid or gaseous fuels because solid ignition usually depends on pyrolyzing enough of the material into a combustible fuel and having those products mix with the correct amount of air to be ignited. The ignition of any solid fuel is a surface phenomenon, which means that the temper- ature of the surface is the critical factor, not the bulk temperature. The ignition of the typical solid depends on providing a high enough heat flux to bring about the necessary pyrolysis (as introduced in Chapter 2). The criterion of “high enough” depends, as we have seen, on the properties of the potential fuel, that is, its thermal conductivity, thermal capacity, and density, as well as its rate of absorption of heat. Ignition has been defined as the initiation of a self- sustaining combustion.1 Because combustion can be either flaming or smoldering, when we examine ignition processes we have to be aware of the difference between those processes. As Babrauskas has pointed out, some ignition temperature data are the surface temperatures of fuels as they ignite into flames, and others are the temperature of an oven into which the object is placed that results (after some time elapses) in smoldering combustion.2 Reasonably pure cellulose is rare in nature, occurring in substances like cotton or linen fiber. Thus, an undyed cotton dress or linen shirt is fairly pure cellulose. A wooden beam is composed largely of cellulose, but in a different structural arrangement with many addi- tional complex organic substances. The chief reason that any numerical values at all can be attributed to the solid fuels is that when heated, most, if not all of them, undergo heat decomposition or pyrolysis with the production of simpler molecular species that do have definite and known properties. Even so, precise values are not often available, because in some instances the pyrolysis of solid materials has been inadequately studied and because the pyrolysis of a single solid material gives rise to a great many simpler products. The resulting complex mixture varies with the thermal and environmental conditions and has physical and chemical properties unlike those of any pure compound. The properties of solid fuels and pyrolysis, and the mechanism of their combustion, will be examined in this chapter. 124 Chapter 5 Combustion Properties of Solid Fuels Pyrolysis As discussed earlier, the most significant part of all flaming fires is the flame itself, in which the combustion reaction takes place solely between gases. This remains true even though the fuel feeding the flame is a solid: wood, cloth, plastic, paper, or even coal. How then does a solid fuel maintain a gaseous reaction in the flames that burn around it? There are several routes by which this can occur, as shown in Figure 5-1. Some solid fuels can evaporate directly to vapors, by a process called sublimation. Naphthalene, a flammable solid formerly used as mothballs, and methenamine, used for ignition tests, sublime at room temperatures. Other solid fuels, such as candle waxes, melt and then evaporate. Thermoplastics melt, decompose into smaller molecular species, and then evaporate. Another class, such as polyurethane, decomposes when heated to form volatile liquid products, which evaporate. The final category is the very large class of products that when heated, decompose to yield both volatile products and a charred matrix. This cate- gory includes wood, paper, other cellulosic products, and most thermosetting resins. The cellulosic Plant- key to most of these mechanisms is the phenomenon of pyrolysis, which occurs within the based materials based on a natural polymer of solid fuel as a result of being strongly heated. sugars. In chemistry, pyrolysis is defined as the decomposition of a material brought on by heat in the absence of reactions with oxygen. In fires, pyrolysis of fuel also includes ther- thermosetting (resin) mal decomposition that takes place in the presence of oxygen (since oxygen is nearly Polymer that decom- always present in the combustion zone). Oxidation by-products are therefore common in poses or degrades as it is heated rather than real-world combustion products. The phenomenon has been known for a very long time, melts. but its mechanism is only now beginning to be understood, and then to only a limited degree. To understand what occurs during the pyrolysis of fuel, it is necessary to remem- pyrolysis The chemi- ber that all practical fuels are organic in nature; that is, they are complex compounds cal decomposition of substances through the based on carbon. action of heat, in the Nearly all fuels of importance to the fire investigator are vegetable in origin or are absence of oxygen. derived by decomposition, bacterial action, or geologic processes from “living” material, both animal and vegetable. This is true of oil and coal as well as of natural gas. Wood, the most common fuel in ordinary fires, is the direct result of life processes in which very com- plex organic structures are synthesized by natural processes within living cells. To date, no one can write a complete structural formula for wood, although it is well known that its major constituent is cellulose, a very large molecule synthesized from many glucose (sugar) molecules in long chains of undetermined length (see page 27). In addition to cellulose, there are many other compounds in wood: hemicellulose and lignin are the most prevalent (up to about one quarter each), and there are various resins, pitches, oils, and other sub- stances in various quantities. Certain softwoods, such as pine, have large quantities of volatile oils called terpenes and oleoresins (the source of commercial rosin), while most hardwoods contain little or no resin and only low concentrations of some of the terpenes. Organic compounds (including the constituents of wood) when heated are subject to complex degradations to simpler compounds that are more volatile and therefore more Sublimation Melting Evaporation FIGURE 5-1 Different modes by which fuel LIQUID VAPOR vapors are generated SOLID Melting Decomposition + from a solid fuel. Source: evaporation Introduction to Fire Dynamics, Decomposition + Decomposition + 2nd ed., D. Drysdale. 1999. melting evaporation New York: John Wiley & Sons Limited. Reproduced with Decomposition + evaporation permission. Chapter 5 Combustion Properties of Solid Fuels 125  flammable than the original ones. It is these compounds that oxidize in the flame. If a sample is pyrolyzed by heating it electrically in an inert atmosphere, the decomposition products may be separated by gas chromatography, identified by mass spectrometry, and studied by a variety of tests. Such studies have been important in recent years in elucidat- ing the mechanism of heat decomposition. Pyrolysis products from wood include water vapor, methane, methanol, and acrolein, for instance. For the purposes of fire investiga- tion, it is not important to understand the exact mechanism of pyrolysis or all the inter- mediate products that result. However, it is important to understand the basic facts of heat decomposition because they are at the heart of understanding the basic nature of fire fed by solid fuel. In the sense that they do not evaporate or sublime directly into fuel vapors, wood, paint, plastics, and coal do not burn. When they are heated, they decompose into smaller molecules with greater volatility and flammability, and, of course, into carbon, which is combustible directly. This is the process of pyrolysis, and it is fundamental to nearly all fires. It explains how fuel vapors generated by the solid are ignited to support the flames visible above the solid fuel surface. There is a zone just below the surface where pyroly- sis occurs, and beneath (or behind) there is a region where no pyrolysis has yet occurred. The gaseous materials formed by the pyrolysis of the large, nonvolatile molecules of the solid fuel are the materials that burn in the flame. Without them there could be no flame. For instance, recent work by Stauffer has demonstrated that polyethylene pyrolyzes into a homologous series of alkanes, alkenes, and dienes; polystyrene forms aromatic products while burning; and polyvinyl chloride first releases noncombustible hydrogen chloride (HCl) gas and then subsequently yields burnable organic fragments.3 Recent work by DeHaan has demonstrated that animal fats pyrolyze into homologous series of alkenes, aldehydes, and alkanes in ratios different from those produced by polyethylene.4 CROWN FIRES AND FIREBALLS The pyrolysis process explains why fires may become “crown” fires in the burning of a forest. When foliage is heated to the point that it exudes a large quantity of volatile gases and vapors that are flammable, the vapor cloud can burst explosively into flames, and the entire bush or tree becomes almost instantaneously enveloped in fire. This is the phenom- enon that gives rise to the comment that a tree or bush “exploded” into fire. The mate- rial composing it simply was heated to the point that it decomposed by pyrolysis to form a large quantity of volatile, flammable gases that when mixed with air, underwent rapid combustion, or deflagrated. Pyrolysis also explains the mechanism by which fires are set at a distance by radiant heat from a large fire nearby. The radiation is absorbed until the temperature produced starts pyrolytic action, and temperatures rise until the autoignition temperature of the pyrolysis products is reached. At this time, fire erupts all over the fuel exposed to the radiated heat. An additional curious fire phenomenon that is explainable only on the basis of pyrolysis products is the fireball, or “firewhirl.” It may be generated in a very intense fire and can travel through the air for appreciable distances while burning. The “ball” is formed by the emission of quantities of pyrolysis products far in excess of the air locally available to burn them. This means they accumulate in a limited region, burn- ing primarily on the periphery. The strong updraft created by the fire carries the pyrol- ysis products upward into the air, where they continue to burn independently of the original fuel from which they were formed. In wildland fires, these products can drift out of the direct updraft and carry a mass of flaming gases into areas not yet on fire. It is possible for such fireballs to entrap personnel fighting the fire. Fatalities from fire- balls are known. Anyone who has viewed a very intense fire will have observed this phenomenon, often in miniature, as masses of flame that detach from the fire and rise into the air, as shown in Figure 5-2. 126 Chapter 5 Combustion Properties of Solid Fuels  FIGURE 5-2 Fire plumes above very energetic fires can become detached and move away from their source as firewhirls or fireballs. Courtesy of Calvin Bonenberger, Lafayette Hill, PA. NONPYROLYZING FUELS Not all solid fuels have to undergo pyrolysis before they can combust. Reactive metals like sodium, potassium, phosphorus, and magnesium combust in air when oxygen com- bines directly on the exposed surface. The resulting heat vaporizes the fuel and produces hot gases and incandescent oxides (ash), but the fuel does not first pyrolyze into simpler compounds. A few “solid” fuels like asphalt and wax melt and vaporize when heat is applied, and the vapors support the flames. Carbon (as charcoal) can undergo combus- tion without flames (smoldering combustion) as a solid–gas interaction on its myriad sur- faces as oxygen diffuses into it. Combustion Properties of Wood Far more wood is burned as fuel in structural and outdoor fires than any other solid material. Thus, its properties as a fuel, in regard to its behavior during combustion, are generally of greater importance to the fire investigator than those of any other solid com- bustible material. COMPONENTS OF WOOD The term wood is generic and covers a wide variety of materials, natural and man-made, whose chief component is vegetable in origin. The major constituent of wood is cellulose ( 50 percent), while numerous other constituents are present: hemicellulose ( 25 percent) and lignin ( 25 percent), with resins, salts, and water making up variable percentages.5 The chemical makeup of cellulose was discussed in Chapter 2. Wood products, which are materials composed chiefly of cellulose, include all kinds of manufactured boards and panels, formed wood items such as furniture, and a host of paper and cardboard prod- ucts manufactured, for the most part, from wood pulp. Wood is obtained from many varieties of trees, some resinous, some not; some dense, others light. Woods vary greatly in their water content, volatile components, and other Chapter 5 Combustion Properties of Solid Fuels 127  chemical properties. In addition to having diverse origins, wooden materials are also greatly altered by manufacturing processes that result in a variety of prepared sheet, board, and formed materials with all types of treatment and finishes. IGNITION AND COMBUSTION OF WOOD Wood, being a chemically and physically complex cellulosic fuel, introduces a number of variables. Each of the major constituents has different temperatures at which it undergoes pyrolysis: hemicellulose at 200°C to 260°C (400°F to 500°F), cellulose at 240°C to 350°C (460°F to 660°F), and lignin at 280°C to 500°C (536°F to 930°F).6 The thermal conduc- tivity of wood varies with orientation, as does its permeability to air, both being signifi- cantly higher in the direction of the grain than across it. This will affect the ignitability of a wood mass or surface, for some portions will be more easily ignited than others. Heat must penetrate the wood to trigger pyrolysis and charring, as illustrated in Figure 5-3. The production of volatile oils and resins is also faster along the grain than at right angles to it, adding further variation. Wood discolors and chars relatively quickly at temperatures above 200°C to 250°C (400°F to 480°F), as shown in Figure 5-4, but prolonged heating at temperatures above 107°C (225°F) will have the same effect.7 (There are limited data on long-term [days or weeks] exposure to lower temperatures, but it appears that destructive pyrolysis can occur very slowly at temperatures down to 85°C [175°F].) As wood chars, its absorp- tivity of incident heat flux becomes higher due to the darker surface and the lower ther- char Carbonaceous mal inertia (k c) of the char, so its temperature begins to increase faster once charring remains of burned begins. All this means that wood does not have a fixed ignition temperature; it varies organic materials. with the rate and manner in which heat is applied to it. It is clear that on heating, a wood surface has one temperature at which it generates enough volatiles (which are evidenced by the visible smoke) that can be ignited by the application of a pilot piloted ignition (external) flame (piloted ignition), and another much higher temperature at which Ignition aided by the the vapors themselves will ignite without any external pilot flame (non-piloted or spon- presence of a separate external ignition source taneous ignition). Along with all the other variables, the nature of the heating (whether such as a flame or elec- radiant or convective) will influence both piloted and spontaneous ignition tempera- tric arc. tures. Other variables include moisture content and size thickness of the sample, Char base Pyrolysis Char layer zone Pyrolysis zone base Normal wood FIGURE 5-3 Normal combustion of wood resulting in progressive formation of char and pyrolysis zones. Courtesy of the USDA Forest Service, Forest Products Laboratory, Char Madison, WI. depth 128 Chapter 5 Combustion Properties of Solid Fuels  FIGURE 5-4 When Douglas fir and similar softwoods are exposed 1 to a fire temperature of 700°C to 900°C (1500°F 350 ºF to 1900°F), the pyrolysis and accompanying char- 220 ºF ring occur at tempera- ½ Inner char tures of approximately zone 550 ºF 177°C to 288°C (350°F ASTM E119 to 550°F), as shown. The Fire temp. depth of the pyrolysis is dependent on the mois- 1500 ºF to 1900 ºF ture content, density of Char rate app. the wood, and intensity ¼0 in./min. of the approaching fire. Note that the char rate is only approximated even under these idealized lab- oratory test conditions. Courtesy of the USDA Forest Service, Forest Products Laboratory, Madison, WI. orientation (vertical or horizontal), oxygen concentration, duration of heat exposure, and especially whether piloted or nonpiloted (autoignition) is being observed. Due to its complexity, the ignition temperature of “wood” is not a single, clearly defined value for all woods and thermal conditions, as evidenced by the scattering of tem- peratures reported here. It is generally agreed that even with fresh (undecomposed) woods, the piloted ignition temperature will vary with the species of wood, the size and form of the sample itself, the ventilation, and the intensity, manner, and period of heat- ing. The autoignition temperature is similarly variable. It might be thought to be depend- ent on the ignition temperatures of the pyrolysis products formed in the wood, but the major species—methane, methanol, and carbon monoxide—all have autoignition temper- atures on the order of 450°C to 600°C (850°F to 1,100°F), which are much higher than some observed autoignition temperatures for wood. Clearly, other factors not yet identi- fied are involved, perhaps the glowing combustion of surface char. The auto- or self-ignition temperature is closely related to the temperature at which the oxidation of the wood becomes sufficiently exothermic to raise the temperature of the surrounding fuel and thereby become self-sustaining. Various authors have reported autoignition temperatures to be on the order of 230°C to 260°C (440°F to 500°F). Browning reported that ignition temperatures could be as low as 228°C (442°F) for some woods, and temperatures rang- ing from 192°C (378°F) to 393°C (740°F) have been reported elsewhere in the literature, but it is not clear whether they were observing flaming ignition or self-sustaining smol- dering combustion.8 The reported variabilities reflect the effects of different test methods, availability of oxygen, and even different masses and geometries of the sample. Browning’s tests were of small wood chips embedded with thermocouples in a heated chamber, and his criterion for ignition was the oven temperature at which the wood chip temperature exceeded that of the environment. Many of these earlier studies did not dis- criminate between smoldering and flaming combustion, so it is not surprising that reported results vary so much. Because ignition of wood is a surface phenomenon, the measurement of surface temperature at the location of ignition is the desired result. Clearly, the complexity of the ignition of wood makes measurement of relevant tempera- tures very difficult and subject to measurement error. Babrauskas has assembled the results of numerous studies and concluded that values around 250°C to 260°C (480°F) Chapter 5 Combustion Properties of Solid Fuels 129  represent a defensible surface temperature for both piloted or nonpiloted ignition of fresh whole wood in a reasonably short period of time9 [ignition near a minimum flux (immer- sion in oven) being smoldering combustion that may transition to flaming]. Piloted autoignition (surface) temperatures measured in tests where loose specimens were in open air being heated by a radiant heat source were typically 300°C to 400°C (570°F to 750°F). For smoldering ignition, surface temperatures under 300°C are typical. Higher temperatures favor gasification and thereby flaming combustion.10 Ignition Variables Dry wood is more readily ignited than wood with a higher moisture content (for the rea- sons explored in Chapter 3). Continuous or even recurrent exposure to raised tempera- tures will dry wood, with the result that the ignition temperatures may be somewhat lowered. However, the work of McNaughton does not indicate a significantly increased hazard as the wood dries until the temperatures reach 275°C to 280°C (527°F to 536°F), by which point both paint and wood are thoroughly charred.11 One recent study of piloted ignition of Masson pine by radiant flux (in horizontal specimens 10 cm 10 cm) revealed surface ignition temperatures of 301°C to 405°C 12 (574°F to 761°F). The higher the moisture content of the sample (0–30 percent MC), the higher the ignition temperature (301°C at 0 percent MC to 368°C at 30 percent MC, for example). Janssens reported that the ignition temperature is increased by 2°C for every 1 percent moisture increase.13 Hemicellulose has the lowest ignition temperature and lignin the highest. Softwoods, with less hemicellulose and more lignin, have slightly higher autoignition temperatures (349°C–364°C; 660°F–687°F) than do hardwoods (300°C–311°C; 572°F–592°F) under the same test protocol.14 Some woods, like pine, will ignite into flames more readily and burn vigorously, because they contain resinous materials that easily produce volatile flammable vapors when heated. These add greatly to the limited ability of cellulose and lignin alone to sup- port the fire. Ease of ignition of wood is correlated with the content of pitch and other components that readily decompose to generate combustible vapors. Some woods, like pitch pine and slash pine, contain so many combustible components that they may easily be ignited with a match flame and will burn furiously, as if soaked in kerosene. Other than volatile resins found in many species of wood, the main fuel component is cellulose. Because the reader is probably familiar with the low heat output from such pure cellulose fuels as cotton or linen fabrics, it is readily appreciated that the cellulose (with its ¢Hc 16 kJ/g) itself is not an efficient part of wood combustion. As we have seen, cellu- lose, as a derivative of glucose, is already partly oxidized, and therefore it has proportion- ately less fuel (per unit weight) to provide for oxidation in fire. Ease of ignition makes wood (¢Hc 16 kJ/g) or paper suitable for kindling a fire, but the total heat output is less on a weight-for-weight basis than that of fuels with a higher percentage of hydrogen such as coal or petroleum (¢Hc 46 kJ/g). Thus, fires that have only wood for fuel may be somewhat less intense than those that are fed by hydrocarbon liquids, which are often used as accelerants (because they can release more energy per second per unit of surface area, sometimes called energy density). Interestingly, charcoal has a ¢Hc of 34 kJ/g. Decomposed Wood The ignition temperatures of fresh, undecomposed wood as measured in the laboratory may be of less importance to the fire investigator, since fresh, new wood is not always involved in a structure fire. Of more interest is the effect of organic decomposition (decay) and thermal decomposition (pyrolysis) on the ignitability of wood. Angell reported that the ignition temperature of southern pine, measured at 205°C (400°F), when sound, dropped to 150°C (300°F) when the same wood decayed.15 (These values seem far too low for autoignition and may represent values for piloted ignition or sustained smolder- ing, possibly as a result of self-heating.) The minimum radiant heat flux for ignition of wood has been measured at 12 kW/m2 for piloted ignition and 20 to 40 kW/m2 for 130 Chapter 5 Combustion Properties of Solid Fuels autoignition (depending on time of exposure and test protocol). It should be noted that these samples were small (0.5 in. 2 in.) and exposed to convective, short-term heating. Ignition times were recorded as 427 seconds for the sound wood and 105 seconds for the rotted wood (which indicate that the samples were exposed to very high heat fluxes).16 The relationship between applied heat flux and piloted ignition has been studied by Janssens. As we saw in a previous chapter, the higher the heat flux and the lower the ther- mal inertia, the faster ignition will take place. Because different woods have different den- sities and thermal conductivities, the ignition time will depend on the variety of wood involved. Moisture content will affect both, but since most interior wood is in the mois- ture content range of 9 to 12 percent, its effect can be ignored. Regarding time to ignition, Babrauskas’s data offered the empirical correlation # tig 130 0.73>(qex – 11.0) 1.82 # where density of the wood (kg/m3), and qex – is the radiant heat flux to which the (thick) wood is exposed (over the range 20–40 kW/m2).17 The lower the density or the higher the radiant heat flux, the shorter the ignition time. There is a relationship between environmental (immersion) temperature and ignition time for wood. Very old data pub- lished by NFPA show that longleaf pine, for instance, does not produce piloted ignition at 157°C (315°F) even after exposure for 40 minutes, yet piloted ignition can occur after 14.3 minutes at 180°C (360°F), 11.8 minutes at 200°C (390°F), 6 minutes at 250°C (480°F), 2.3 minutes at 300°C (570°F), or 0.5 minute at 400°C (750°F).18 (Very low reported ignition temperatures of 180°C–200°C, however, should be interpreted as com- ing from faulty measurements when only short-term heating has been performed.) “LOW TEMPERATURE” IGNITION OF WOOD Of more frequent interest to fire investigators are the effects of slow, prolonged heating of wood, which dehydrates and then decomposes the wood by pyrolytic action. The normal combustion of wood is a progressive process with zones of char and pyrolysis that progress inward from surfaces exposed to heat, as illustrated in Figure 5-4. At temperatures between 100° (212°F) and about 280°C (540°F) wood loses weight slowly as moisture and volatile oils are released gradually, and a large percentage ( 40 percent) of the wood is turned to char- coal.19 At temperatures above about 180°C (360°F), the pyrolysis of all three major solid constituents (i.e., cellulose, hemicellulose, and lignin) reaches its maximum rate, leaving a smaller percentage (10 to 20 percent by weight) as char. If the heat being accumulated by the char is retained, and there is an adequate supply of oxygen, the temperature of the mass can rise to the point at which combustion can take place (ignition temperatures of about 300°C; 570°F). The retention of heat, of course, depends on the amount of thermal insulation avail- able and the amount of heat that is being lost to convective and conductive processes. If there is too much insulation, the supply of oxygen becomes inadequate to sustain combustion, although smoldering combustion can be sustained even at very low oxygen levels. It is for these reasons that the formation of charcoal by slow heating of wood can play a significant role in the initiation of some accidental smoldering fires20 as depicted in Figure 5-5. Failure of the insulation around metal chimneys or fireboxes can expose wooden building components to such temperatures, while the failure of the integrity of metal or masonry flues may allow flames or sparks to provide piloted ignition. Obviously, direct flame exposure from an opening in a flue or chimney will lead to relatively rapid ignition of exposed wood. There is also a relationship between incident heat flux, piloted ignition temperature, and time to ignition. Yudong and Drysdale demonstrated that the higher the heat flux (between 15 and 32 kW/m2), the lower the observed ignition temperature and the shorter the time to ignition.21 Although whole wood itself may not ignite at “low” temperatures, it is well known that prolonged exposure to heat at temperatures below 120°C (248°F) over an extended Chapter 5 Combustion Properties of Solid Fuels 131 FIGURE 5-5 Charring in ceiling from prolonged contact with electric radiant-heat ceiling panels. Courtesy of Robert Toth, IRIS Investigations. period of time can cause wood to degrade to charcoal by the distillation and pyrolysis process described.22 Shaffer calculated that exposure to temperatures as low as 150°C (300°F) for long periods can decompose finely divided cellulosics to charcoal, which then may ignite.23 This charcoal has been referred to as pyrophoric carbon or pyrophoric char- coal, alluding to the properties of activated (laboratory) charcoal to oxidize with room air pyrophoric Capable even at modestly elevated temperatures. It has been argued that pyrophoric is a misnomer, of igniting on exposure since the U.S. Department of Transportation defines pyrophoric materials as those that can to atmospheric oxygen ignite spontaneously within 5 minutes of exposure to air. (The Chemical Manufacturers at normal temperatures. Association defines pyrophoric as a material that will ignite spontaneously in dry or moist air at a temperature of 54.4°C (130°F) or below.)24 Bowes, in his comprehensive study of self-heating mechanisms, points out that activated charcoal is made by heating a carbona- ceous fuel (often, coconut hulls for laboratory use) under reducing conditions (in the absence of all oxygen). Although this material is capable of self-heating when exposed to air, especially when warm, exposure to air over a long period of time simply permits oxi- dation of the carbon.25 If a wooden surface is exposed to a radiant heat source when air is freely available, the pyrolysis products will distill away, but the activated form of char- coal will not be formed in bulk. The resulting char may not be susceptible to self-heating, and flames could not be supported because the required volatiles have already dissipated. Bowes’s examination of the problem showed, however, that if a heated cylinder such as a flue or pipe runs through a massive wood member with minimal clearance, the conditions will be appropriate for the formation of a charcoal that can self-heat to smoldering com- spontaneous ignition bustion if the temperatures are high enough spontaneous ignition. Bowes offered the A chemical or biological opinion that the temperatures produced by ordinary steam pipes would not be adequate to process that generates create this condition but that superheated steam pipes or flues might. He commented, how- sufficient heat to ignite the reacting materials. ever, that if a suitably hot source (i.e., one with adequately high radiant heat output) was See self-heating. close to a wood surface overlaid with a noncombustible layer of sheet metal, tile, or simi- lar barrier to oxygen, a reducing atmosphere could be produced, as shown in Figure 5-6. The failure of the barrier or some other change that would suddenly bring the heated char- coal into contact with fresh air could then result in flaming combustion. Bowes presented the thermodynamic argument that to accomplish this, the temperature of the source would have to be considerably hotter than steam pipes operating at normal pressures. 132 Chapter 5 Combustion Properties of Solid Fuels  FIGURE 5-6 Fire in a wood floor under the hearth of an improperly installed fireplace. Courtesy of Greg Lampkin, Knox County, TN, Fire Investigation Unit. Heaters and Flues One study of the problem of ignition of wooden structures by heaters and flues has shed further light on this problem. Martin and Margot evaluated a series of fires in houses where a change had been made to high-efficiency wood stoves, sometimes 5 and 10 years prior to the fire.26 They discovered that the affected wood surfaces had been faced with metal or tile, often with an intervening layer of solid insulation. With prolonged heating, even at temperatures below 120°C (250°F), the insulation would allow enough heat to be conducted into the wood to trigger degradation while at the same time excluding air from the surface. The lower the temperature was, of course, the longer the time required. If after a suitable mass of charcoal had been created and maintained at a high temperature by the surrounding mass of wood oxygen was admitted (by way of structural collapse, wood shrinkage, or movement of the responsible heater), the mass of charcoal could sus- tain self-heating combustion to the point of igniting flaming fire in adjoining fuels. Martin and Margot’s analysis of the thermodynamics showed that temperatures had to be main- tained for a very long period of time, that air circulation had to be minimal (to minimize heat losses as well as to exclude fresh air), and that the involved wooden members had to be fairly massive for this process to take place. Degradation to Char Recent work by Cuzzillo and Pagni has explored a number of factors associated with the heating (or “cooking” at temperatures well below normal ignition temperatures) of wood.27 Their findings show that the degradation of wood to char over a long period of time decreases the thermal conductivity and dramatically increases the porosity of wood. Cracking of the char makes it even more accessible to atmospheric oxygen. (Fresh wood is much less permeable across its grain than it is along its grain, and it has higher thermal conductivity along its grain than across it as well.) When wood is heated at low temperatures, charcoal is formed (even in the presence of some air). If enough is produced to be a critical mass for self- heating, and the char is sufficiently porous to permit diffusion of oxygen through the mass, the heated material can be ignited as a self-sustaining, smoldering combustion. If conditions promote a sufficient increase in heat release rate, there can be a transition to flaming com- bustion in adjoining masses of wood that have not been degraded completely. Chapter 5 Combustion Properties of Solid Fuels 133  Cuzzillo’s work showed that self-heating of the char formed by low-temperature “cook- ing” is a likely mechanism leading to runaway self-heating when the conditions are right. It has to be realized that runaway self-heating of such char does not necessarily lead to flam- ing combustion, because so much of the volatile content has been cooked out of the char. But a change in ventilation (such as shrinkage of wood timbers, failure of a metal or tile cov- ering, or even a shift in the direction and intensity of ambient wind) could be the factor to push the combustion process into flames. The processes that prompt transition from smol- dering to flaming are not well understood. Flaming may be triggered by better ventilation that causes a higher heat release rate (above a critical threshold), or the creation of hot spots in the char that cause piloted ignition. The role of the processes involved in long-term, low- temperature ignition of wood will be discussed in more detail in Chapter 6. Areas in which oxygen-deprived “cooking” processes may take place include areas surrounding light fixtures and fireplaces. Flues, vents, and chimneys (see Figure 5-7) nor- mally have adequate clearances built into them because they are recognized sources of heat, but untreated cellulose insulation (wood shavings, sawdust, or paper pulp) can inad- vertently be brought into contact with such heat sources. Pipes carrying steam under pres- sure can reach temperatures of more than 150°C (300°F) and should be well ventilated with adequate clearances where they pass through wooden structural members. In a recent review, it was concluded that any heat source that can create surface temperatures on wood members in excess of 77°C (170°F) must be properly separated or insulated (as UL recommended years ago). In fact, that is a requirement of the City of New York Building Code.28 FIGURE 5-7 Displaced metal chimney sections allowed heat to ignite wood framing of chimney enclosure. Courtesy of Lamont “Monty” McGill, Fire/Explosion Investigator. 134 Chapter 5 Combustion Properties of Solid Fuels  FIGURE 5-8 Massive floor timbers over 80 years old ignited by hot-water pipes. Residents smelled smoke for 2 weeks prior to detecting smoldering fire. Courtesy of Thomas Goodrow, Fire/Explosives Technical Specialist, ATF (retired). The review of low-temperature ignition issues by Babrauskas revealed numerous well-documented cases where cellulosic materials (including sawdust, wood boards, and timber) were “ignited” by sources between 90°C and 200°C (196°F–390°F).29 Charring of wood is known to take place at temperatures as low as 105° to 107°C (221°F–225°F), and that wood charred at such low temperatures exhibits similar exothermic oxidation as char produced at very high temperatures ( 375°C; 700°F). It is also thought that mois- ture and cyclic heating and cooling play roles in making charred wood susceptible to self- heating. There are well-documented fire cases where massive wood timbers in a residence were ignited by prolonged contact with steam heating pipes whose surface temperatures never exceeded 120°C (250°F), as illustrated in Figure 5-8. As with other self-heating materials (as will be seen in Chapter 6), if there is a sufficient reactive mass, adequate oxy- gen, insulation, and time, such self-heating can reach runaway growth achieving smolder- ing that, in many cases, can transition to open flame. In the 2004 MagneTek court decision, several scientific errors crept in and were “val- idated” by the published decision.30 These included a claim that pyrolysis is just a theory that is not suitably supported by scientific data. Nothing could be further from the truth. Pyrolysis as described in this text has been known for over 200 years to be part of the combustion process, and thousands of papers support and prove its mechanisms. Somehow the court became convinced that wood has a fixed ignition temperature of 400°F (195°C) and that any ignition has to be caused by a source with at least that tem- perature. The court failed to note that wood is a very complex fuel that undergoes numer- ous chemical and physical changes as it is heated. The various forms created by these processes have different chemical and physical properties, some of which include an increased propensity to self-heat. Even though the term pyrophoric is misapplied to wood and charcoal (and that is presumably the term the court meant to discredit), the mecha- nisms of self-heating are well known. No, there is no tested and published “formula” that Chapter 5 Combustion Properties of Solid Fuels 135  can accurately predict the relationship between heat (or temperature) exposure and time to ignition for such a complex and multivariate problem as “low-temperature” ignition of wood; however, that does not mean that scientifically defensible hypotheses cannot be offered and tested in fire investigation. Some solar water heaters (which are deliberately designed to capture the energy of the sun’s heat radiation) can produce temperatures above 150°C (300°F) both in the heat-collecting element and in the pressurized transfer medium (if any). When the circulating pumps for these heaters fail, charring of underly- ing wood structures and roof fires have reportedly resulted. Investigators tend to associate the time of discovery with the time of first ignition. This assumption may introduce serious errors to the fire analysis. Due to its slow output of heat and smoke, smoldering may proceed for an extended period without being noticed by passersby. When the combustion transitions to flame, it is almost certain to be discov- ered quickly. Flame Temperatures The actual flame temperatures produced by burning wood vary greatly, with such vari- ables as oxygen content, forced-air draft, resin content, and degree of carbonization hav- ing significant effects. Using a typical net heat of combustion for wood of 18 MJ/kg (7755 Btu/lb) gives a calculated adiabatic (ideal condition) flame temperature of 1,590°C (2,900°F). The actual flame temperature of wood has been measured and reported as being some 500°C less (estimated to be 1,040°C, or 1,900°F).31 It should be noted that the combustion of wood is so complex that a calculation of adiabatic flame temperature is of little relevance. It merely serves to illustrate the contrast between such calculated values and observed flame temperatures. It should also be remembered that measure- ments of turbulent flame temperatures produce time averages of rapidly fluctuating temperatures. Char Rates Once wood is ignited, the rate at which it will char is dependent on the heat flux to which the wood is exposed, its material, and its physical form. The standard exposure for wall and floor materials uses the E119 furnace, whose temperature and heat flux rise from near ambient to a maximum after 60 to 90 minutes. Under these test conditions the char rate for plywood ranged from 1.17 mm/min to 2.53 mm/min (0.05 to 0.1 in./min). For 1/2-in. (12.7-mm) plywood, burn-through times were 10 to 12 minutes. For 3/4-in. plywood (18–19 mm), burn-through times were 7.5 to 17.6 minutes. For boards (with tongue-and- groove edges or no gaps), burn-through times were 10.5 minutes for 19.8-mm pine, 14.17 minutes for 20.6-mm oak, and 24.3 minutes for 38-mm boards.32 The presence of gaps dramatically decreased burn-through times, reducing them by half that of tightly fitted boards of the same thickness. Testing by the authors has shown that post-flashover burn- ing of wood surfaces (such as plywood) can achieve 3 to 4 mm/min (0.12 to 0.16 in./min) char rates, particularly where ventilation is at a maximum (in doorways for example).33 These rates are in accordance with results reported for “jet” fires against softwoods and for higher-temperature fires than the E119 protocol, in which for exposures to mild fluxes ( 20 kW/m2) the char rates are on the order of 0.6 to 1.1 mm/min (0.02 to 0.04 in./min).34 Butler tested the charring rate of one type of wood at a wide range of radiant heat fluxes and found an arithmetic relationship, as shown in Figure 5-9. Note that at a radi- ant heat flux of less than 10 kW/m2, the char rate was zero, while at 20 kW/m2 it was approximately 0.6 mm/min (0.02 in./min). At flame contact heat fluxes of 50 kW/m2 the rate was 1.2 mm/min (0.05 in./min), and at 150 kW/m2 the rate was on the order of 4.2 mm/min (0.17 in./min).35 Babrauskas reported that other researchers reported simi- lar correlation. With a correlation # for mass loss rate, the best mathematical fit for surface char rate was 0.028 qtot (where , char rate, is in mm/min) for radiant flames below 200 kW/m2. 136 Chapter 5 Combustion Properties of Solid Fuels  FIGURE 5-9 Variation of 1 charring rate of wood with radiant heat flux. Note the log/log scale and sharp decrease below 20 kW/m2. From Butler, C. P. “Notes on Charring Rates in Wood.” Fire Research Note Rate of charring (mm/s) 10–1 No. 896, 1971. Reproduced in Drysdale, D. An Introduction to Fire Dynamics. New York: John Wiley & Sons, 1985, p. 182. 1/40"/min 10–2 0.6 mm/min 10–3 10 100 1000 Radiant heat flux (kW/m2) Doors often provide the only separation between compartments and provide delays to fire spread, so an appreciation of fire penetration times is important to timeline recon- structions. The critical feature is usually the tightness of fit, with loose or badly fitting doors providing only 3 to 5 minutes of resistance to flame spread. If the door is known to be a reasonably good fit, a solid-core door 44.5 mm (1-3/4 in.) thick can withstand 14 to 30 minutes of E119 exposure.36 Shoub found “panel” doors [panels 9.5 mm (3/8 in.) thick] burned through in 5 to 6 minutes of E119 exposure (1.6 to 1.9 mm/min). Hollow- core doors with 4.8-mm (3/16-in.) wood panels burned through in 9 to 10 minutes.37 Many hollow-core doors in low-cost housing or offices are coated fiberboard and have been seen to fail within 3 to 5 minutes of fire exposure in room fire tests by these authors. As noted previously, intensities (and exposure times) of real fires in modern rooms can be more severe than the E119 test protocol.38 Role of Adhesive Another factor that may contribute to the influence of manufactured wood products is the combustion properties of the adhesive used. Often, the adhesive is less readily com- bustible than the wood layers; if it is more combustible, it may contribute what is effec- tively a fire accelerant. The resins used in oriented strand board (OSB) have the same ignition and combustion properties as the bulk solid timber, so burn-through times are reportedly similar.39 In the evaluation of any fire involving these materials, special atten- tion to the adhesive involved is essential. Some plastic binders can melt and support a flame like a large candle, leaving little residue after prolonged combustion. CHARCOAL AND COKE Although wood contains oxygen in its structure and could be considered partially oxi- dized, one of its main combustion products, charcoal, contains no significant oxygen and is therefore an entirely carbonaceous fuel. Charcoal is an excellent fuel, having a heat of combustion on the order of 34 MJ/kg. Charcoal is, of course, formed by pyrolyzing and destructively distilling the volatile materials out of the wood, leaving behind the non- volatile constituents, chiefly carbon, as shown in Figures 5-3 and 5-4. Charcoal is corre- spondingly difficult to ignite and, lacking volatile materials, gives little flame. The small Chapter 5 Combustion Properties of Solid Fuels 137  blue flames associated with its burning result from the formation of carbon monoxide (CO) gas, which burns rapidly close to the surface of the charcoal from which it was gen- erated, to form carbon dioxide (CO2). A charcoal fire is largely a smoldering fire, that is, one without flame, but intensely hot. Here the solid is reacting with gaseous oxygen to produce combustion on and slightly below the surface of the fuel. Coke, formed similarly from coal or petroleum, displays burning characteristics much like those of charcoal and for the same reasons. It is not a cellulosic fuel, because coke (and coal) is almost entirely carbon and hydrocarbons. WOOD PRODUCTS Not all wood present in structures is in the form of structural lumber. There is an infinite variety of manufactured wood products in every building. The susceptibility of these products is not the same as for wood alone. These products deserve special mention because they greatly modify the burning characteristics of the wood from which they are derived, and they can add significantly to the fuel load and ease of ignition. As a general rule of heat transfer, the smaller the dimensions, the thinner the fuel, and the more edges presented to a heat source, the faster the fuel package will be ignited. Plywood and Veneer Board Plywood and veneer board are very common. Both are made from very thin sheets of wood (peeled from suitable logs) laminated to one another with layers of adhesive. The nature of the adhesive determines the suitability of the product for interior (dry condi- tions only), exterior (some weather exposure), or marine (extended exposure to water) use. Standard plywood is largely made from fir, with cheaper, imperfect sheets and pieces in the interior layers. Veneer board generally is made with a top layer of hardwood lam- inated to less expensive base layers. The authors have investigated a number of fires in which plywood and veneer board were heavily involved. In no instance was the ignition of the fire traced to such materials, but the rapidity of buildup and spread of the fire was nearly always markedly increased. (See fire tests on vertical wood surfaces in Chapter 3.) In one instance of a fire of electri- cal origin, an extremely thin plywood finish at some distance from the origin had pro- point of origin The duced such severe local burning that the location was for a time suspected as the point of specific location at origin. This illustrates the necessity of checking on all possible types of causation and the which a fire was ignited. behavior of involved fuels before adopting any one of them. Failure to do this has led to numerous errors on the part of some fire investigators. Plywood and veneer board curl as they delaminate as they burn. (See Figure 3-30.) This behavior can contribute to rapid downward flame spread and cause floor-level burns, sometimes on flooring some distance from the wall. (See Chapter 7 for an example.) Particleboard or Chipboard Particleboard or chipboard is made from small chips, sawdust, and waste from wood and paper mills bonded together under great pressure with a suitable adhesive. Particleboard is very dense but cheaper to make than plywood and is widely used in floors, cabinets, and furniture where appearance is not of importance. Particleboard has very limited strength under wet conditions and tends to swell and crumble when exposed to water for any length of time. Both plywood and particleboard are often fin- ished with a thin vinyl covering that can be combustible. A product that is very widely used for exterior sheathing and manufactured structural components is called oriented strand board (OSB) or chipboard. These are large splinters, shards, and thin fragments of wood that are oriented so that their grain is roughly parallel and then glued together under great pressure and heat using formaldehyde resins. The resulting product is stronger than particleboard and cheaper than plywood because it can be made from waste from wood processing. 138 Chapter 5 Combustion Properties of Solid Fuels Burning Behavior of Thin-Cut Woods With regard to the burning characteristics of wood products, it must be remembered that the finer or thinner wood is cut, the better it ignites, and the faster it burns, because the thin fuel is exposed to heat from both sides, and there is no underlying mass into which the excess heat can be conducted. As a result, the thin fuel reaches its surface ignition tem- perature more quickly and ignites. It then burns faster because of the large surface area exposed to heat and air. A laminated or glued-composite board will burn much like a solid-wood board of the same dimensions except for the influence of the adhesive or coat- ing, which can be significant. If the adhesive softens under heat (or if it has been exposed to water or moist conditions), the layers will delaminate and open like the pages of a book. This will make the thin layers much easier to ignite, expose a greatly increased sur- face area to the flames, and cause comparatively fiercer burning. Thus, in this regard, the adhesive may be more important than other considerations. Plywood from the most com- mon sources is not much different in its combustion properties from other wood of the same thickness. Some imported veneer boards are made with unsuitable adhesive and may greatly accelerate a fire involving them. Particleboard is generally more difficult to ignite because of its dense nature. Once ignited, it behaves in the same way as densely packed sawdust and is prone to smoldering combustion that is difficult to extinguish. Prolonged exposure to moisture causes many particleboard-type composite materials to disintegrate, with corresponding changes in their ignition and combustion properties. Exterior Deck Materials Several new products intended for exterior decks are being seen more frequently today. These consist of cellulose (wood flour, wood fiber, or ground paper) blended with a poly- ethylene or polypropylene plastic binder. They can vary from 41 to 65 percent wood fiber, with 31 to 50 percent plastic and ash (possibly added as a filler or opacifying agent).40 They may be solid in cross section, hollow, or channeled (as seen in Figure 5-10). These products have been tested for ignitability by burning brand and for fire performance when ignited by an underdeck burner, or in a cone calorimeter. These composite materials have been found to exhibit increased heat release rate (i.e., a bigger fire) than the equivalent solid wood material.41 They also exhibit a tendency to drip molten plastic as they burn (thus increasing the chances of spreading a fire down onto combustibles below) and to fail structurally rapidly and without warning (increasing risks of injury to firefighters).42 Other Cellulosic Building Materials Other cellulosic building materials, not necessarily made from wood but alike in their combustion properties, include Celotex, Masonite, and similar boards made from com- pressed and bonded fibrous materials of a cellulosic nature. The low-density products (such as Celotex) were formerly used as acoustical tiles to cover ceilings and today are used, in large sheets, as insulation, floor underlay, or for other such purposes (see Figure 3-27). Tests show that some of these materials can support rapid flame spread across their surfaces and are subject to a smoldering fire along their edges or, occasionally, in their interiors. In a large surrounding fire, they burn well and contribute much fuel. Thus, it is common to see that in a room fire, all the acoustical tiles have fallen and burned on the floor. They fall readily because they are generally glued into place, and the heat softens the adhesive. If ceiling insulation or ceiling tiles are made of low-density cellulose (paper fibers), they can contribute significantly to the intensity of the fire, but only rarely are they involved in starting it. Once on fire, they are difficult to extinguish, since they tend to con- tinue smoldering. The high-density products (Masonite) are considerably more ignition resistant. Blown-in ceiling insulation can be shredded or macerated paper with a flame retar- dant added. A similar product is also made for application by being blown in while wet for use in walls, where the dried mass is more resistant to mechanical disturbance and fire penetration. Blown-in insulation can also be mineral fiber, which is noncombustible. Both Chapter 5 Combustion Properties of Solid Fuels 139  types of products are typically gray in color, and careful examination (including labora- tory analysis) is required to identify the product.43 Loose blown-in cellulose insulation is a triple threat when it comes to accidental ignition for the following reasons: Its good insulative properties minimize heat loss from even small heat sources buried in it. Its loose texture means it can be moved or can fall into contact with heat sources. Its fire retardance can be compromised by moisture or mechanical separation, mak- ing it more susceptible to smoldering. Some brands of cellulose insulation have been formulated with fire retardants which can be highly corrosive to metals, although these are not common. Fresh, proper flame-retardant treatment will, however, make the bulk material very resistant to both flaming and smoldering combustion. Samples of any suspected struc- tural, decorative, or insulation materials should be retained for laboratory analysis. Paper Paper is one of the more interesting substances involved in fires, partly because of its fre- quent use in starting fires and partly because of its unusual combustion properties. Everyone who has ignited a fire with paper will realize that newspaper works well, while a picture magazine makes very poor kindling. The reasons are simple but not always understood. The basic ingredient of all paper is cellulose, the same material as that in cotton and a major constituent of wood, both of which are used in the manufacture of paper. Although cellulose is readily combustible, not all the extraneous constituents of paper are. “Slick” papers have a very high content of clay, which is not combustible. Many writing papers also have a high clay content. The differences in the combustion properties of var- ious papers are readily demonstrated by burning a piece of newspaper, which leaves a small and very light ash, followed by burning a piece of heavy, smooth, slick magazine paper, which leaves a very heavy ash. The ash in most instances is the clay, although many specialty papers also contain such constituents as titania (titanium dioxide), barium sul- fate, and other noncombustibles. Because paper is a cellulosic product like wood, its ignition properties are not eas- ily measured. Graf studied the ignition temperatures of a variety of papers.44 With few exceptions, he found the ignition temperature to be between 218°C and 246°C (425°F and 475°F). At 150°C (300°F) most papers appeared to be unchanged during the dura- tion of his tests. Many papers tanned in the range up to 177°C (350°F), went from tan to brown by about 204°C (400°F), and if not ignited, went from brown to black at about 232°C (450°F). Such events are affected by the rate of heating, manner of igni- tion, nature of heat exposure, and amount of ventilation; therefore, a precise ignition temperature cannot be cited. Smith’s data (as reported in the Ignition Handbook) indi- cate that most uncoated papers ignited at surface temperatures of 260°C to 290°C (500°F to 550°F) as a result of a validated test.45 Other reported tests generally agreed. Long-term ignition was reported to occur in paper wrapped around a pipe maintained at 200°C (400°F). Paper ignites easily because, like wood, it has a low thermal inertia [calculated from k c, the product of the fuel’s thermal conductivity (k), its density ( ), and thermal capac- ity (c)], so its surface temperature goes up rapidly, and it is thermally thin, allowing rapid heat saturation of its entire thickness. The minimum ignition heat flux for various paper products has variously been reported from 20 to 35 kW/m2.46 If paper is exposed as a single sheet rather than compressed in a stack, its large surface area promotes rapid burning, and a high, but brief, heat release rate. When paper is stacked, it no longer has much exposed surface, and a pile of stacked papers is very difficult to burn, although it 140 Chapter 5 Combustion Properties of Solid Fuels  FIGURE 5-10 Plastic and cellulose composites for building applications. (a) Plank form for decks. (b) Board form for fences. Courtesy of John D. DeHaan. (a) (b) will char on the exposed surface. Figure 5-11 illustrates the effect of thermal thickness. Ventilation is absent on the inside of such a stack, thus making paper a good insulator against penetration of heat from a surrounding fire. In fact, a small fire may be smoth- ered by paper quite effectively, merely by covering the burning surface with a few layers of it. The air supply is thereby shut off, and the fire is quenched. This is true even of Chapter 5 Combustion Properties of Solid Fuels 141 FIGURE 5-11 Loose paper exposed to radiant heat from a developing room fire demonstrates the effect of “thermal thickness”—scorched where it curled away and unaffected where in con- tact with book beneath. Courtesy of John D. DeHaan. newspaper, which is among the most combustible types of paper in existence. A stack of paper, however, can support smoldering fire within its mass. Single sheets of paper, either flat or crumpled, have proven very difficult, if not impossible, to ignite by contact with a glowing cigarette. Carbon paper, however, can be ignited with a cigarette and will sup- port a smoldering fire. Once burning at a steady rate, cardboard cartons (flats) have been measured to yield a maximum mass flux of 14 g/m2/s compared with that of wood cribs, 11 g/m2/s.47 In the interpretation of a fire, it is not the presence of paper that is significant but its distribution in terms of exposed surface. Sometimes, the presence of a stack of papers is taken as the origin of a general fire. Although it is possible to burn a stack of folded newspapers to completion over an extended time, it is surprisingly difficult to do because of the limited surface area exposed to the air. In contrast, a pile of loosely crumpled sheets of the same paper is not only easy to ignite but quickly raises the temperature of nearby combustibles to their ignition points and thus initiates a fast-growing, very destructive fire. In our experiments, rooms with a small general fuel load (bed, dresser, carpet, and wooden doors) could be quickly brought to a fully involved stage of fire with a large pile of crumpled newspaper ignited with a match. The papers burned nearly to completion, leaving only small, unremarkable flakes of ash. Although fiberglass thermal insulation is not combustible, it is often faced with a kraft paper layer secured with an asphaltic binder. The thin paper, if exposed on walls or ceilings, can support extremely fast fire spread. At the same time, it should be remembered that a flammable liquid poured onto a pile of papers is a common arson set. If the papers are closely packed, the accelerant burns at the surface of the papers, with the absorbent paper acting like a wick. The residue of unburned papers within such a pile is a good place to look for unburned traces of the liq- uid accelerant. Here the primary purpose of the paper is to retain the accelerant, not to contribute largely to the fire by its own combustion. 142 Chapter 5 Combustion Properties of Solid Fuels Plastics Plastics are nearly universal in modern structures and vehicles and are found burned in nearly all fires. As a result, fire investigators need a working knowledge of the kinds of plastics, their applications, and their burning behaviors. GENERAL CHARACTERISTICS There are many kinds of plastics, with a variety of physical and chemical properties. Some, like Teflon (tetrafluoropolyethylene), are not readily combustible and will exhibit heat damage only at extremely high temperatures. Others, such as nitrocellulose, are readily ignitable and violently flammable. Most of the commonly used plastics range between these two extremes. Called synthetics as a class because they are not natural synthetic Material products but are manufactured in a factory (usually from petroleum feedstocks), most that is human-made, usually referring to plastics are readily combustible. Some may burn to completion only in an intense sur- organic polymers. rounding fire, and others will readily burn almost to completion. Many laboratory tests of the ignition and combustion properties of plastics are used to identify them and clas- sify them as to fire hazard. It is important to remember that nearly all plastics are made from long chains of hydrocarbons linked together in various ways. Thus, sufficient heat input will rupture the chemical bonds holding the chains together, and the plastic will pyrolyze to simpler, more volatile compounds. These pyrolysis products may be very toxic, as in the case of styrene monomer and cyanogens; extremely flammable, as in the case of carbon monoxide (CO) and the short-lived radicals CH and CH2; or readily sus- ceptible to further pyrolysis. In addition, some plastics and rubbers create pyrolysis prod- ucts that can have serious consequences in the fire environment. Polyvinyl chloride plastic releases hydrochloric acid (HCl), and Teflon releases hydrogen fluoride (HF) and free fluorine (F2) when it finally decomposes in a fire. Some rubbers generate hydrocyanic acid (HCN) during combustion. For a full description of pyrolysis products from plastics in fires, the reader is referred to other references.48 In addition, thermoplastic polymers will melt and flow, sometimes dripping onto a source of flame to fuel it and sometimes collecting in a large pool of burning plastic, which burns fiercely and resists extinguish- ment. The behaviors of some of the more common plastics when exposed to a small flame in air are listed in Table 5-1. As with laboratory tests with other solid fuels, it must be remembered that the fire behavior of synthetics is largely dependent on their degree of molecular cross-linking, physical form (sheet, rod, fiber, etc.), test conditions, and pres- ence of inorganic fillers and fire retardants, so it is best not to generalize about the fire behavior of a particular type of plastic. Because the chemistry of plastics is similar to that of petroleum products, they tend to produce similar sooty flames when burning in air, and the open-flame temperatures produced by burning plastics are on the same order as those of common petroleum dis- tillates. The flames of thermoplastics can be sustained by the formation of a pool of molten plastic, producing very intense fires. The flame temperatures produced by some plastics are quite high. Polyurethane is one example for which flame temperatures up to 1,300°C (2,400°F) have been measured over polyurethane mattresses during fire experi- ments.49 The reported maximum mass burning fluxes for plastics are shown in Table 5-2 (for horizontal steady-state burning). Multiplied by the heat of combustion for each fuel, these mass fluxes can be used to approximate the HRR of horizontal fuel surfaces of var- ious sizes. Note that these heat release rates are nearly the same as for ignitable liquids [for polypropylene, for example: 24 g/m2 · s 44 kJ/g 1 MW/m2]. The combustion of many polymers contributes largely to the formation of the greasy soot The carbon- or sticky dense soot found at many fire scenes and is responsible for the dense black based solid residue smoke more frequently noted during the early stages of structure fires. The dense smoke created by incomplete many polymers produce is rich in flammable vapors, and the ignition of dense clouds of combustion of carbon- these vapors quickly involves an entire room upon flashover. Only polyethylene, polypropylene, based fuels. Chapter 5 Combustion Properties of Solid Fuels 143  TABLE 5-1 Flaming Burning Characteristics of Small Samples of Common Polymers POLYMER FLAME COLOR ODOR OTHER FEATURES Nylon Blue with yellow tip Resembles burning Tends to self-extinguish; grass clear melt Polytetrafluoroethylene Yellow None Chars very slowly; difficult to burn Polyvinyl chloride Yellow with green Acrid Tends to self-extinguish; base acidic fumes Urea-formaldehyde Pale yellow Fishy formaldehyde Very difficult to ignite Butyl rubber Yellow, smoky Sweet Burns Nitrile rubber Yellow, smoky Sickly sweet Burns Polyurethane Yellow, blue base Acrid Burns readily Silicone rubber Bright yellow-white None Burns with white fumes; white residue Styrene-butadiene rubber Yellow, very smoky Styrene Burns Alkyd plastic Yellow, smoky Pungent Burns readily Polyester (contains styrene) Yellow, very smoky Styrene Burns Polyethylene Yellow, blue base Resembles burning Burns readily; clear when candle wax molten Polymethyl methacrylate Yellow Characteristic Burns Polypropylene Yellow, blue base Resembles burning Burns readily; clear when candle wax molten Polystyrene Yellow, blue base, Styrene Burns readily very smoky Source: K. J. Saunders, The Identification of Plastics and Rubbers (London: Chapman and Hall, 1966). See also Fire Protection Handbook, 19th ed. (Quincy, MA: NFPA, 2003), fig. 8.10.6. and polyoxymethylene (Delrin) generally burn with a clear flame and little soot. Polyethylene and polypropylene also produce a candle-wax smell, since upon heating, they pyrolyze into much the same components as wax, with the resulting similar behavior. Although the autoignition temperatures of plastics are generally on the order of 330°C to 600°C (650°F to 1,100°F), their universal applications in packaging, utensils, TABLE 5-2 Maximum Mass Flux Rates (Steady-State Burning) PLASTIC g/m2 · s Polystyrene (granular) 38 Acrylic plastic (PMHA, granular) 28 Polyethylene (granular) 26 Polypropylene (granular) 24 Rigid polyurethane foam 22–25 Flexible polyurethane foam 21–27 Polyvinyl chloride (granular) 16 Source: From J. Quintiere, Principles of Fire Behavior (Albany, NY: Delmar, 1998), 109. 144 Chapter 5 Combustion Properties of Solid Fuels  TABLE 5-3 Thermal Properties of Polymeric Materials THERMAL BEHAVIOR MELTING POINT (°C) NATURAL POLYMERS ¢Hc (DRYSDALE) (MERCK INDEX)a Wool Chars above 200°C Cellulose 16.1 Chars above 100°C Thermoplastic Polymers Polyethylene (HD) 46.5 Melts 130°C–135°C (85–110)a°C Polypropylene (isotactic) 46.0 Melts 186°C (165)a°C Polymethyl methacrylate 26.2 Melts 160°C (250)°C Polystyrene 41.6 Melts 240°C (180)°C Polyacrylonitrile Melts 317°C Nylon 66 31.9 Melts 250°C–260°C Nylon 6 — (225)°C Polyvinyl chloride (PVC) 19.9 — (Chars 200–300)°C Thermosetting Polymers Polyesters (Dec. 250)°C Polyurethane foam 24.4 Decomposes 200°C–300°C: polyols, isocyanates Phenolic foams 17.9 Chars Polyisocyanurate foams 24.4 Chars b Silicones 26.8 Decomposes/oxidizes to SiO2b Source: From D. Drysdale, An Introduction to Fire Dynamics, 2nd ed. (Chichester, UK: Wiley, 1999), 3. a The melting point of many polymers is determined by the amount of cross-linking between the chemical structures and can vary considerably. b Buch, RR, “Rates of Heat Release and Related Fire Parameters for Silicones”, Fire Safety Journal, 17, 1991, 1–12. furnishings, wall coverings, windows, walls, and even exterior structure panels provide many opportunities for them to come into contact with temperatures of this magnitude (see Table 5-3).50 BEHAVIOR OF PLASTICS When considering plastics as fuels, it is helpful to classify them by their behavior when they are heated. Plastics that undergo reversible melting without appreciable chemical degradation are called thermoplastics. Plastics that do not melt but rather decompose and thermoplastic Polymer leave behind a solid char are called thermosetting plastics. There is a small group of mate- that can undergo reversible melting rials that decompose into volatile materials that are not rigid and are sometimes called without appreciable elastomers; these include polyurethane rubber. The chemical and physical behaviors of chemical change. some of the common polymers are shown in Table 5-3. Minimum autoignition temperatures of common plastics are shown in Table 5-4. Polyvinyl Chloride The behavior of a particular plastic in a general fire is the result of many factors, which all should be kept in mind when evaluating its potential contributions to the fire. Vinyl (polyvinyl chloride—PVC) may burn slowly when tested in solid form in a laboratory but has been observed to burn rapidly and contribute greatly to flame spread when present as a thin film on wall coverings (probably due to the adhesives and plasticizers used). PVC wire insu- lation softens and melts when heated by overheated wiring but chars if exposed to a flame. Chapter 5 Combustion Properties of Solid Fuels 145  TABLE 5-4 Autoignition Temperatures of Common Plastics MINIMUM IGNITION TEMPERATURE PLASTIC °C °F Polyethylene 365* 488 910 Polyisocyanate 525 977 * Polymethyl methacrylate 310 467 872 *