hw

The essay should be a one- to two-page narrative focusing on arguments that support what the authors discuss in the attachments , as well as other research that you conduct related to these concepts. You are to focus on one or two concepts in this unit’s chapters and relate to the fire that now involves several rooms and other apartments. Make sure the points are tangible. The narrative should consist of your evaluation of the textbook and other resources, listing the most compelling evidence or argument.

The questions below may help you to focus your essay on the one or two concepts chosen.

• How does the rate of flame spread over the surface of the cooking oil (liquid) depend on the flash point?
• If the cooking oil is at or above its flash point, is the flame spread rate fast, and will the spilled oil engulf with flames within seconds?
• With the intense heat in the room, is the cooking oil already evaporating sufficiently to reach the lower flammability limit in the vapor phase over the fuel surface?
• Does chemical change occur within solid materials in and around the apartment and within the chemistry of the volatiles during pyrolysis?
• Does the dripping or running of the synthetic fabric that is draped from the ceiling to the floor increase the fire spread rate due to the increase in burning surface?
• Can the heat from the apartment fire cause skin burns and/or weaken the structure to the point of collapse?
• Can a single exposure to the smoke in the apartment have acute effects (e.g., death, incapacitation, impaired mobility, and reduced clarity of thinking) and post-exposure effects (e.g., lung damage) to the downed firefighter?

CHAPTER
11

Smoke and Heat Hazards

OBJECTIVES

After studying this chapter, you should be able to:

•   List the hazards to people and property from a fire.

•   Explain the following types of harm from a fire: acute effects, postexposure effects, and chronic effects.

•   List the most important toxic gases in smoke.

•   Explain the differences between narcotic gases and irritant gases.

•   Explain the concept of fractional effective dose.

•   Explain the underlying principle of Haber’s rule.

•   Explain the concept of limiting hazard and its role in fire protection.

Introduction

Shortly after 7:00 A.M. on November 21, 1980, a small fire broke out in one of the restaurants on the casino floor of the 26-story MGM Grand Hotel in Las Vegas. By the time the fire was under control, around

10

:00 A.M., 85 people had died and more than 700 were injured. Although the fire never extended above the casino level, victims were located as high as the 25th floor. The smoke had spread with minimal dilution up the elevator shafts and other vertical pathways, exposing occupants of the upper floors. Seventy-five of the fatalities were from smoke inhalation.

This outcome is indicative of the profile of causes of fire deaths in the United States. From 1970 to 1985, 41 percent of the fire deaths in structures were caused by smoke inhalation, and another 46 percent were caused by a combination of smoke inhalation and burns. Eighty percent of the fatalities resulted from fires that spread beyond the room of origin, and the majority of the victims of those post-flashover fatal fires were outside the area of origin when the fire began. Nearly all of these remote victims of post-flashover fires died from smoke inhalation

—

either smoke inhalation only (54 percent) or smoke inhalation with burns (36 percent)—as opposed to burns only (2 percent) or other causes (8 percent) [1].

These data are consistent with two key points made in the preceding chapters: (1) post-flashover fires are underventilated, and (2) underventilated fires serve as the source of products of incomplete combustion. This section relates the products of incomplete combustion to the toxic potency of the fire-generated atmosphere.

Hazards of Smoke Exposure

If an occupant of a building is confronted with a fire, multiple levels of harm might result from exposure to the smoke [2]. At the highest level of exposure, the occupant might perish while still in the building. Somewhat lower exposures might have incapacitating effects that hinder escape capability and make it unlikely that a person would survive without assistance. Sub-incapacitating effects include impaired mobility and other physical capabilities, reduced clarity of thinking, and detrimental behavior. Any of these sublethal effects might lead to injury or death due to lengthened immersion in the smoke.

These 
acute effects
 are experienced at the time of the fire and result from smoke exposure in a single fire. Additional effects might be experienced well after leaving the fire scene. For building occupants, 
postexposure effects
 can result from a single smoke exposure; for fire fighters, 
chronic effects
 can result from accumulated damage from multiple exposures.

Both forensic investigations and laboratory studies have enhanced our knowledge of the smoke components that lead to these casualties. The most thorough forensic program was based on autopsies of all fire victims in the State of Maryland over several years in the 1970s [3]. Of the 530 victims, approximately 60 percent had more than half of the hemoglobin in their blood converted to carboxyhemoglobin (COHb) as a consequence of having inhaled CO shortly before dying. CO binds to hemoglobin approximately 200 times more effectively than does O2, so its presence reduces the blood’s capacity for transporting oxygen throughout the body. In experiments in which test animals were exposed to carbon monoxide, conversion of more than about half the hemoglobin to COHb caused incapacitation and usually death. An additional 20 percent of the Maryland fire victims had a somewhat lower COHb level, but also had preexisting cardiovascular disease, suggesting that they were vulnerable to lower exposures of carbon monoxide than healthy people. Thus 80 percent of the victims could have been killed by carbon monoxide; of the remaining 20 percent, 11 percent died from burns, and the cause of death was not established for 9 percent.

Hydrogen cyanide (HCN) can be generated from burning materials that contain nitrogen in the chemical structure. High levels of blood cyanide were found in the blood of a number of victims in the Maryland study, and HCN inhalation was suspected to be a cause or contributor to death in those cases. Autopsies performed in subsequent investigations of fires in Pennsylvania [4] and Puerto Rico [5] found that hydrogen cyanide inhalation was the primary cause of death for some of the 136 fire victims.

There are no standard procedures for quantifying the presence of other inhaled toxic gases in the deceased. Hence, there are no data on these smoke components from these or any other fire victim studies.

Broader knowledge of the smoke components that contribute to fire deaths has been gathered through laboratory studies that have identified the gases generated during the flaming, pyrolysis, or smoldering of materials and products. These results enabled construction of extensive lists of potentially harmful gases, although two important limitations apply when using these lists to assess the toxic potency of a fire environment:

•   The completeness of the lists is unknown, as different chemical analysis techniques were used for different groups of chemicals, and the chemical methods were chosen based on expectations of the nature of the combustion products. As a consequence, an unexpected, highly toxic chemical might not have been found.

•   These studies did not provide a basis for quantifying the toxicological effects of any one of the identified gases or of combinations of those gases. A mixture of components might be much more or less toxic than the summed potency of the individual gases.

This line of research was complemented by bioassays, in which laboratory animals were exposed to single gases, combinations of gases, or smoke from both pyrolysis and combustion, providing extensive quantitative toxicological data [6, 7]. In these test series, the animals were exposed for different time intervals to different concentration levels. The data were reduced to determine the EC50, the concentration (C) that led to an effect (E) in 50 percent of the test subjects. In nearly all the studies, the effect was lethality (LC50) or one of several measures of incapacitation (IC50). Ideally, the animal species were selected because of the similarity of their toxicological response to humans’ response. However, the high cost of such testing using monkeys, baboons, or dogs led to many of the experiments being performed using carefully bred rodents, in order to reduce variability among the test subjects. In any case, the quantitative extrapolation of the data to humans introduced some degree of uncertainty. No new animal smoke exposure studies have been reported for more than two decades.

Based on the various research studies, we can identify some general precepts that are fundamental to fire toxicology [6]:

1.  The toxic gases can be grouped into three classes:

a.  Asphyxiants (or narcotics). Upon inhalation, 
asphyxiant gases
, also called narcotic gases, deprive the body of oxygen, either by interfering with oxygen’s transport to cells or by reducing the cells’ ability to use the oxygen. The two most important asphyxiants are CO and HCN.

b.  Irritants. 
Sensory irritant gases
 (or, simply, irritant gases) that affect the senses (e.g., eye irritation) and upper respiratory tract tend to influence physiology and behavior during exposure. Those that affect the lower pulmonary tract (lungs) tend to have later effects. The most important irritant gases are the halogen acids and some partially oxidized organic gases, such as acrolein and formaldehyde. Some sulfur- and phosphorus-containing compounds may contribute if the combustibles contain these atoms.

c.  Others. In a few instances, the smoke from a small material specimen generated extremely toxic smoke or an unusual toxic effect on laboratory animals. In each case, the offensive combustion product was identified by specialized chemical analysis. There was no commonality of chemical structure of the materials or the toxicants.

2.  Within each of the first two groups, the effects of the gases can be combined.

3.  The asphyxiants’ effects are dose related; that is, the effect results from the concentration of the inhaled gases and the time interval over which they are inhaled. A person moving through a smoky building accumulates a 
fractional effective dose (FED)
 during each time increment, such as 0.25 minute [8]. These values are summed, and when the FED reaches 1, the average person experiences the effect (e.g., death, incapacitation). 
Haber’s rule
 is often used to relate these two components: the time and intensity of exposure. It states that, for a given toxicological effect, the product of the volume fraction and the exposure time is a constant. While this tool gives a fair estimation for individual asphyxiants, it is less accurate for combinations of such gases.

4.  The sensory effects of the irritant gases are concentration related; that is, they occur very quickly. As a person experiences successively higher concentrations of an irritant gas, the 
fractional effective concentration (FEC)
 increases, with the average person experiencing the effect at FEC = 1.

5.  Despite the large number of combustion-generated, potentially toxic chemicals, the deaths of rats exposed to fire smoke can be estimated from the contributions of only a few (N) gases. In the resulting 
N
-gas model
, N has been found experimentally to be no larger than about 7 for a wide variety of materials [7]. The N-gas model is also applied to lethal and incapacitating effects of smoke on people.

6.  The smoke exposure that is lethal is approximately two to three times the incapacitating exposure [9, 10].

There are thresholds of both exposure time and concentration below which no noticeable effect occurs. Experts have compiled the following equations for the incapacitation due to inhalation of gases [11]:

•   Narcotic gases:

where [CO] is the average volume fraction (µL/L) of CO during the time increment, Δ t; [HCN] is the average volume fraction (µL/L) of HCN during the time increment, Δ t; and Δt is the time increment (min). At each time increment, the two terms on the right-hand side of 

Equation 11-1

 are multiplied by a frequency factor, vCO2, to allow for the increased breathing rate (and thus the increased uptake of the narcotic gases) caused by the presence of CO2:

where [CO2] is the average volume percent of CO2 during the time increment.

•   Irritant gases:

where [Z] is the volume fraction of an irritant gas, Z, (µL/L), and FZ is the volume fraction of that gas (µL/L) expected to lead to an FEC value of unity. The best-judgment approximate values for FZ for some irritant gases are provided here:

HCl or HBr: 1000 µL/L
HF: 500 µL/L
NO2 or formaldehyde: 250 µL/L
SO2: 150 µL/L
Acrolein: 30 µL/L

For example, a person exposed to 1000 µL/L of a mixture of HBr in air would immediately experience severe eye irritation and throat constriction that would render the person unable to effect his or her own escape. The same outcome would occur with a mixture of 125 µL/L of NO2 and 250 µL/L of HF.

The following sections provide additional information for specific toxicants.

Toxicity of Prominent Fire Gases

Carbon Monoxide

Carbon monoxide (CO) is a narcotic toxicant common to nearly all fire smoke. It often makes the major contribution to an incapacitating or lethal exposure. As with all smoke, the concentration in the upper layer of the room is particularly important because this gas mixture emanates from the room as the temperature rises and the gas expands. As noted earlier, most fire deaths occur outside the room of fire origin, and people outside the fire room may be exposed to this flow, albeit perhaps in a diluted version.

Table 11-1
 shows the symptoms that develop in a healthy person as the percentage of COHb increases. People who are more susceptible will experience symptoms at lower COHb levels; people who are less susceptible will experience symptoms at higher COHb levels.

Equation 11-1 relates the CO volume fraction in the inhaled air to the exposure time that would lead to incapacitation of the average person. Other equations [6] relate the percent COHb in the blood to the volume fraction of the inhaled air and personal variables, such as a person’s breathing rate, body mass, and activity level. 

Figure 11-1

 shows the exposure times required to incapacitate a 154 lb (70 kg) person performing different levels of activity [6]. CO appears to follow Haber’s rule over most of the volume fraction and time ranges, although asymptotes are visible at the extremes of curves B and C in Figure 11-1. Low volume fractions of CO (less than a few hundred µL/L) apparently would not be incapacitating for common exposure times to fire smoke. At the very high CO volume percentages, exposures of less than two minutes would result in lethal COHb levels.

Table 11-1 Symptoms as the Percentage of Hemoglobin Converted to Carboxyhemoglobin (COHb) Increases in the Human Bloodstream [12]

Percent by Volume of Carboxyhemoglobin		Symptoms
10	Judgment inefficiencies
10–20	Slight headache
20–30	Headache, fatigue, dizziness
30–40	Severe headache, weakness, dizziness, confusion, vision dimness, nausea, vomiting and collapse
40–50	Death for some
50–60	Coma, death for most
60–80	Death for all within a few hours
80–90	Death within an hour
>90	Death within a few minutes

Data from: ISO 27638-11, “Analysis of Blood for Asphyxiant Toxicants – Carbon Monoxide and Hydrogen Cyanide,” International Standards Organization, Geneva, 2011.

Figure 11-1 Relationships between the volume percent of CO and the time to incapacitation by carbon monoxide for a 70 kg (154 lb) human at different levels of activity [6].

Curve A: Person at rest, 40 percent COHb, breathing rate of 8.5 L/min

Curve B: Light work (e.g., walking at 6.4 km/h, 4 mph), 30 percent COHb, breathing rate of 50 L/min

Curve C: Heavy work (e.g., slow running at 8.5 km/h, 5.3 mph, or walking up a 17 percent incline at 5.6 km/h, 3.5 mph)

Reproduced from: Purser, D. A. (2008). Assessment of Hazards to Occupants from Smoke, Toxic Gases, and Heat. In: The SFPE Handbook of Fire Protection Engineering, 4th ed., DiNenno, P. J., et al., eds. Quincy, MA: National Fire Protection Association.

Carbon Dioxide

Carbon dioxide (CO2) is generated in all fires involving organic materials. Because most of the oxygen depleted from the air during a fire is bound into CO2, an estimate of the volume fraction of CO2 in the fire effluent can be obtained by subtracting the volume fraction of oxygen in the smoke from 0.21. For severely underventilated burning or for fuels that contain large amounts of oxygen (i.e., significant numbers of oxygen atoms), this estimate can be adjusted using the balanced chemical reaction equation.

Up to concentrations of 5 percent by volume, CO2 is not toxic [6]. Instead, it mainly increases the respiration rate (

Equation 11-2

) and dilutes the toxicants and oxygen in the air. As the CO2 concentration approaches 10 percent, however, a person would experience dizziness and loss of consciousness.

As described later in this chapter, in a residential fire room, the temperature and heat exposure would reach untenable levels before the CO2 concentration reaches 5 percent. Post flashover, the CO2 concentration in the smoke flowing from the fire room could exceed 10 percent by volume; however, this flow is also untenably hot. By the time that the smoke temperature has cooled below 100 °C, the CO2 concentration will generally be less than 5 percent due to dilution of the effluent by air outside the fire room.

Hydrogen Cyanide

The lethal toxic potency of hydrogen cyanide (HCN) is approximately 20 times that of CO. Fortunately, the maximum measured HCN volume fractions in most smoke are lower than CO volume fractions by at least the same factor. Like CO, HCN affects the amount of oxygen made available to cells in the body and especially the brain. However, its mechanism in doing so differs from that used by CO. HCN does not interfere with the hemoglobin transport of the oxygen. Rather, after entering the bloodstream through the lungs, HCN combines with the enzymes in the cells and inactivates them, so the cells can no longer accept oxygen. 
Table 11-2
 summarizes the effects of acute exposure to HCN.

Equation 11-1 is a best fit to data for the exposure of monkeys to various concentrations of HCN in air. This equation is deemed reasonably appropriate for simulating the susceptibility of people performing light activity, such as a brisk walk. Note that the exponent of 2.36 means that HCN does not follow Haber’s rule [6]. Equation 11-1 indicates that the average person would be incapacitated from a 10-min exposure to 130 µL/L of HCN in air and from a 3-min exposure to 220 µL/L of HCN in air. The data reveal that a threshold effect occurs: exposure to HCN concentrations below 80 µL/L for up to an hour results in only minor hyperventilation.

Hydrogen Chloride and Hydrogen Bromide

The toxic potencies of hydrogen chloride (HCl) and hydrogen bromide (HBr) are approximately equal. Researchers have published far more data for HCl, so the following discussion focuses on these data.

When laboratory animals, or people, are exposed to HCl at very low volume fractions, an immediate irritant reaction occurs in the sensitive areas of the eyes, nose, and throat. At higher volume fractions, the irritation becomes painful, impairs escape, and eventually becomes incapacitating. For high doses (volume fraction multiplied by duration), an insult to the lower respiratory tract can result in long-term harm and even death. This lower respiratory tract damage can be enhanced by HCl adsorbed on small soot particles or dissolved in small liquid droplets in smoke. Upon inhalation, these submicrometer aerosols bypass the body’s defenses in the upper respiratory tract and reach the lungs unabated. The magnitude of this enhancement has not been measured.

Table 11-2 Symptoms as the Concentration of Blood Cyanide Increases in the Human Bloodstream [12]

Symptoms

Blood Cyanide (µg/mL)
0.5–1.0	Conscious, but flushed skin, rapid pulse, and headache
1.0–2.5	Stuporous, but responsive to stimuli, excessively rapid breathing and heart rate
≥ 2.5	Comatose, blue skin tone, gasping for breath, death

Data from: ISO 16312-1, “Guidance for Assessing the Validity of Physical Fire Models for Obtaining Fire Effluent Toxicity Data for Fire Hazard and Risk Assessment. Part 1: Criteria,” International Standards Oragnization, Geneva, 2011.

The data are inconsistent regarding the volume fraction of HCl that results in almost immediate incapacitation of the average person. The consensus among experts is that this level approximates 1000 µL/L [10]. Noticeable impairment of escape is expected at approximately one-fifth this volume fraction [6].

Nitrogen Oxides

Oxidation of the nitrogen in the burning products leads to the formation of nitrogen dioxide (NO2) and nitric oxide (NO). NO2 is the favored combustion product, and its toxic potency is approximately five times that of NO. NO2 is a strong irritant with an estimated incapacitating volume fraction of 250 µL/L [11].

Organic Irritants

Among the small, partially oxidized organic molecules that act as strong irritants (formaldehyde, acetaldehyde, and acrolein), acrolein (CH2CH—CHO) generates notable concern, because the toxicological literature indicates that exposure to volume fractions as low as 10 µL/L could be lethal to people. However, baboons exposed to 500 µL/L of this gas for 5 min were able to perform a task and survived afterward [13]. Equation 11-2 indicates an incapacitating exposure of 30 µL/L of acrolein, but in light of the baboon experiments, this may well be excessively conservative. This discordance may be unimportant, because toxicologically significant concentrations of acrolein have not been quantified in the smoke from burning or smoldering materials.

Other Toxic Species

Many other gases occasionally found in smoke are believed to be harmful. These gases, which rarely appear in significant concentrations except when special combustibles are involved in the fire, include ammonia (NH3), sulfur dioxide (SO2), hydrogen sulfide (H2S), hydrogen fluoride (HF), isocyanates, and phosphorus compounds. A very few compounds whose toxic potency is far higher than that of the compounds discussed here are labeled as “supertoxicants.” Reference [6] offers additional information on these and other gases.

Oxygen Deficiency

Even if only carbon dioxide, water vapor, and nitrogen were present in the air inhaled by a fire victim, the oxygen concentration would be reduced below the normal 21 percent by volume. As the percentage drops toward 14 percent, the effects of oxygen deficiency appear, including lethargy, impairment of coordination, and nausea. Judgment, memory, and work capacity decline as the oxygen concentration drops toward 10 percent. All these effects are heightened when the change in oxygen concentration occurs suddenly—the typical case in a fire. In an environment containing less than 10 percent oxygen, people rapidly lose consciousness and will die if not revived quickly.

As implied in the discussion of CO2, fire fighters and building occupants who are enveloped by the undiluted smoke flow from a flaming fire might be exposed to oxygen concentrations in the lethal range, but this flow is also untenably hot and rich in toxic gases. By the time that dilution has cooled the smoke to a temperature less than 50 °C, the O2 concentration will generally be greater than 15 percent. In some instances, a reduced O2 level is not accompanied by a highly elevated temperature. A person who is asleep in a bed or on an upholstered chair that is smoldering might inhale smoke that is only warm, but contains significantly diminished oxygen compared to normal air.

Inasmuch as two of the principal toxicants in smoke, CO and HCN, act to deprive the brain of oxygen, their effects would be enhanced if the oxygen level in the air fell significantly below 21 percent [4]. These effects can be reversed by immediately administering oxygen to persons rescued from fires (
Figure 11-2
).

Figure 11-2 Fire fighter administering oxygen to a fire victim.

© Patti McConville/Alamy Images.

Smoke Toxic Potency Measurement

There is insufficient time and laboratory capacity to burn all commercial products under the multiple possible fire conditions and quantify the toxic potency of the smoke. Therefore, over the years, a number of test methods had been developed as means to obtain such information more quickly and less expensively [14]. The apparatus in these methods, also known as 
physical fire models
, differ markedly in their combustion conditions, test specimen size and configuration, and products or materials tested. They also differ in their gas sampling and measurement methods and in how (and whether) laboratory animals are exposed to the smoke.

These test methods were intended for use as stand-alone indicators of the acceptability of a material. A low LC50 value or a high concentration of one or more toxic gases would make a material unacceptable for use. However, the results obtained from different test procedures do not always agree. For example, one test procedure showed the LC50 for red oak to be 3.4 times lower than that for polystyrene, while the polystyrene LC50 was deemed 1.7 times lower than the value for red oak by another test procedure [15].

It is now recognized that a fire hazard or risk assessment is the most technically sound basis for specification of a product’s acceptability. In addition to toxic potency values for the potential combustibles, this assessment takes the following information into account:

•   Mass and burning properties of the combustibles

•   Room/building properties, including dimensions and locations of doors and windows

•   Nature of the building (e.g., apartment, hospital)

•   Nature of the occupants, such as age and mobility

•   Locations of the occupants

•   Types of potential fires

•   Outcomes to be avoided (e.g., deaths, serious injuries, loss of building function)

A standard that identifies the principles for describing and evaluating toxic potency measurement methods, including those that do not involve exposing test animals has been established [16]. These principles are classified into seven categories:

1.  The apparatus replicating one or more fire stages: well-ventilated burning, vitiated burning, post-flashover burning, or smoldering

2.  The test specimen being representative of the product being evaluated

3.  If test animals are to be exposed to the smoke, statement of the choice of physiological effect to be measured and documentation of the relationship between the animals’ response and the anticipated human response

4.  Analysis of the combustion or pyrolysis products that are known to be generally produced and those that might be specific to the product being investigated

5.  Measurement of the mass loss of the test specimen

6.  Determination of the accuracy of the toxic species yields and the resulting estimate of smoke toxic potency, properly by comparison with a set of real-scale tests of the same products

7.  Determination and documentation of the within-laboratory repeatability and the inter-laboratory reproducibility of the test results

A companion document applies these principles to a dozen currently cited methods [17]. Most of the apparatus have been used to test materials, rather than products.

Since there is a very large number of different products and materials in our residences, and since a single fire usually involves multiple products, it would be a useful simplification if nearly all combustibles could be represented by a single LC50 or IC50 value. Products whose smoke is extremely toxic could be treated as exceptions. 

Table 11-3

 summarizes the results of a 2004 compilation of published toxic potency data. At that time, these were all the published results for animal exposure tests that generated an EC50 and for which the combustion conditions in the apparatus could be related to a single fire stage [10]. Only animal data were used, as these were the only tests in which it could be assumed that the effects of all the toxicants were included; all the test animals were rats or mice, and their exposure to the smoke was for nominally 30 min. The materials and products included materials/products whose chemical structures were aromatic and nonaromatic, and some that contained atoms other than C, H, and O. The numbers in parentheses indicate the number of materials and products for which data had been reported, followed by the number of specimens whose test results were at least 30 percent above the average value and the number that were at least 30 percent below the average value. (The estimated uncertainty in the EC50 for an individual material was estimated to be ±30 percent.) Examination of these results reveals the following points:

•   The mean EC50 values are not very sensitive to the mode of combustion.

•   The dose that leads to incapacitation is approximately one-third to one-half the lethal dose.

•   For many specimens, their EC50 values lie outside the experimental range of uncertainty. (The individual values ranged from as low as 1 g/m3 to as high as 100 g/m3.) Recalling that low EC50 values represent smoke of higher toxic potency, the data indicate that the smoke toxic potency for some materials can be more than 10 times that of the average material. If a significant mass of one of these materials became involved in a fire, it would be prudent to use caution in assessing the tenability using an average EC50 value.

Some engineering calculations have assumed that CO was the only toxicant and based the survivability time on this premise. The average LC50 values from Table 11-3 are comparable to the value for a post-flashover atmosphere in which the only harmful gases are CO and CO2. Thus, the large numbers of materials with lower LC50 values reflect materials containing other toxicants and/or unusually larger amounts of CO. The CO-only assumption, then, is a practical starting point but has clear limitations.

Table 11-3 Mean LC50 and IC50 Values (g/m3) [10]

Fire Stage	LC50	IC50
Well-ventilated flaming	30.4 (101; 22, 15)	11.2 (51; 18, 26)
Underventilated flaming	25.8 (10; 4, 0)	—
Oxidative pyrolysis	27.8 (87; 18, 19)	11.5 (53; 14, 18)

Data from: ISO 13571, “Life-threatening Components of Fire – Guidelines for the Estimation of Time to Compromised Tenability in Fires,” International Standards Organization, Geneva, (2012).

ISO 13571 provides an alternative to the CO-only hypothesis for engineering calculations in which the fuel chemistry is unknown [11]. In these cases, a “generic” LCt50 value may be used—that is, 900 g/m−3·min for well-ventilated, pre-flashover fires and 450 g/m−3·min for vitiated, post-flashover fires. For incapacitation, the ICt50 values would be 450 g/m3·min for well-ventilated, pre-flashover fires and 220 g/m3·min for vitiated, post-flashover fires. The data in Table 11-3 indicate that this simplification has limitations similar to those for the CO-only assumption.

These values (and all those mentioned previously with a subscript “50”) correspond to conditions that would affect the average animal or person. Of course, some people may be more susceptible or less susceptible to the airborne toxicants. Designing a facility using these average values would put half the occupants at risk in the event of a fire. It is common practice to apply a safety factor in such situations. For smoke hazard calculations, one might use an ICt50 value that is one-third of the generic value; however, a more careful rationale is warranted. The factor of 3 might be too restrictive (i.e., it might rule out some otherwise desirable products) in a building in which the occupants are predominantly young, healthy adults. Conversely, the same factor of 3 might be insufficient for a building in which the occupants are expected to be more susceptible to airborne contaminants, such as a preschool or an assisted-living facility.

Thus far, the discussion has focused on the acute effects of fire smoke—that is, the effects that happen while a person is still in the burning building or shortly upon escape or rescue. Longer term effects, both on people and on the environment, are also possible. One such threat, mentioned earlier, is pulmonary damage from a high exposure to irritant gases.

Some combustibles either are carcinogenic (cancer-causing) compounds themselves or can generate carcinogenic compounds under fire conditions. These compounds can be transported (as part of the smoke) to remote parts of the building and to the outdoors. Indoors, their presence can lead to an extensive decontamination process of the entire building, which could delay reoccupation of the building for years. Outdoors, their effect can be widely spread if they enter the food chain.

As an example, consider the polychlorinated biphenyls (PCBs) formerly used as electrical insulating fluids. 

Figure 11-3

 shows the structure of a PCB, which contains five chlorine atoms per molecule. The number and locations of chlorine atoms in the molecules can vary.

PCBs have been found to be toxic, and their manufacture was banned in 1979. Nevertheless, they may persist in older electronics, lighting fixtures, electric power transformers, and capacitor banks. Of even greater concern than the toxicity of PCBs is their possible thermal transformation, in part, to the much more toxic chemical class known as chlorinated dioxins, an example of which is also shown in Figure 11-3.

Figure 11-3 Structure of a pentachlorobiphenyl and a chlorinated dioxin.

Nonthermal Smoke Damage

Fire smoke is not only dangerous to people; it can also impair the functioning of critical equipment. This type of damage can be caused by a relatively small, localized fire, whose smoke spreads to distant points in the building.

This kind of damage from fire smoke often leads to effects that are chemical in nature. The active chemicals include organic acids (e.g., formic acid, acetic acid) and inorganic acids (e.g., HCl, HNO3) generated from the building materials and contents. HCl can penetrate porous structural concrete and attack its steel reinforcing rods. The subsequent slow corrosion can lead to structural weakening at a later time. Extinguishing agents, which may not be corrosive themselves, can generate corrosive by-products (discussed in the Fire fighting Chemicals chapter). For example, when applied to a fire, the halogenated fire suppressants partially decompose to form HCl, HBr, and/or HF. Some dry-powder agents contain corrosive salts. Water can cause damage in some cases (e.g., to books or documents), although vacuum-drying procedures have been developed to salvage most items. Smoke moving through a building will be absorbed by many textiles and will impart an odor to these materials. Sometimes these odors can be removed by washing or dry cleaning; however, the chemistry associated with residual odors is not well understood.

Perhaps the most important category of nonthermal smoke damage is to electronic equipment, such as computers, telephone switchgear, manufacturing plant control rooms, and the electronic controls on everything from smoke alarms to microwave ovens. The most obvious type of damage is corrosion—the slow oxidation of metal exposed to air—which can be accelerated by a substance in the smoke, often an acid. For example, HCl in smoke will attack most metals to form the metal chloride, which will then promote (catalyze) further attack of the metal. The presence of moisture or high relative humidity (greater than 40 percent) generally is necessary for rapid corrosion to occur. A careful examination of the surface of a metal exposed to air under normal (nonfire) conditions often will show a chloride deposit of as much as 10 mg/m². This amount usually is not harmful. However, after exposure to smoke from a fire involving polyvinyl chloride, surface contamination of as much as thousands of mg/m² has been found, which can lead to significant damage.

Electronic equipment also can be rendered non-operative because of a short caused by conductive soot particles, which can bridge a gap between conductors in the circuit. Smoke aerosol deposits on electric contacts (as in connectors or relays, where protective plastic coatings are not feasible) can cause the contacts to stick.

Procedures for removing smoke contamination from electronic equipment involve such means as detergents, solvents, neutralizing agents, ultrasonic vibrations, and clean air jets. These procedures are largely empirical rather than scientifically based, and they are not always effective; indeed, sometimes they may provide only a temporary fix.

Thermal Damage

The potential for heat from a fire to ignite and pyrolyze products and to burn skin was discussed in the Heat Transfer and Combustion, Fire, and Flammability chapters.

In a large fire, sufficient heat can be generated to trigger the structure’s collapse. This is what happened to World Trade Center Building 7 on September 11, 2001 [18]. A presentation on structural weakening and collapse is beyond the scope of this text. References [19] and [20] provide thorough introductions to this subject, and the Executive Summary in Reference [18] provides a rendering of an unexpected outcome that changed people’s perception of the vulnerability of tall buildings to fire.

Although this collapse of a steel structure (World Trade Center Building 7) solely due to fire was unprecedented, structural damage to and collapse of wooden structures does sometimes happen following a severe fire. This outcome results from thermal weakening and burning of, for example, the floor joists. Reference [21] documents the performance of several wooden floor assemblies during a severe fire.

The Limiting Hazard Concept

When designing or repurposing a building is undertaken, the building code specifies fire safety requirements that must be met, to which the owner adds functional and economic objectives and perhaps additional safety enhancements. It is then prudent to conduct a (semi-)quantitative assessment of the expected performance of the building in an emergency to highlight the potential vulnerabilities. In addressing these issues, it makes sense to focus on the most important shortcoming.

This chapter and the preceding chapters have addressed the multiple threats to people in a building subjected to a fire: incapacitation from narcotic gases, irritant gases, and skin burns; reduced visibility, affecting the efficiency of actions to escape or find a place of refuge; and structural collapse. As the fire progresses, each of these threats grows, with the growth rate being different in different locations in the building and being experienced differently by each person. The effect on a person depends on such factors as his or her location, physical condition, and means (or possibility) of moving. At some point in time, one of these threats reaches a level at which the person’s life is in peril. This 
limiting hazard
 is the hazard that is most urgently considered in a revised fire protection plan.

Assessment of fire hazards and risks can be performed using a series of calculations or a computational model of a fire in a building. (See the chapter on Computational Modeling of Fires.) In doing this assessment, it is a good idea to let the model continue to run to determine whether other threats reach perilous levels almost concurrently with the limiting hazard. For example, when a room reaches flashover, the toxicological, thermal, and visual threats to people in the adjacent room all increase quickly. Preventing flashover—for example, by choice and extent of furnishings or installing automatic sprinklers—might be more effective as a means to address these threats than locating the exit signs near floor level, which might be a reasonable approach if the limiting hazard were simply visibility through smoke. One could also construct lists and perform limiting hazard analyses for other undesirable outcomes of a fire, such as an interrupted operation (e.g., on a stock exchange trading floor), extended or permanent loss of use of the building (e.g., a home), and loss of sensitive equipment (e.g., a computer).

WRAP-UP

Chapter Summary

•   Smoke inhalation is a factor in most fire deaths in the United States, especially those that result from post-flashover fires.

•   A single exposure to fire smoke can have both acute effects (e.g., death, incapacitation, impaired mobility, and reduced clarity of thinking) and postexposure effects (e.g., lung damage). Fire fighters can experience chronic effects from multiple exposures to smoke.

•   Most of the toxic gases fall into two classes.

•   Asphyxiant (narcotic) gases, mainly CO and HCN, deprive the body of oxygen. The effects are dose related—that is, a function of the volume fraction of the gas and the duration of exposure. Oxygen depletion and carbon dioxide increase the effect.

•   Irritant gases can have acute, instantaneous effects on the senses and upper respiratory tract as well as long-term effects on the lungs. Prominent irritant gases include the halogen acids, nitrogen oxides, formaldehyde, and acrolein.

•   The fractional effective dose concept enables combining the effects of the individual toxic gases and estimating the intensity and length of exposure to smoke that would result in incapacitation or death.

•   Experiments with laboratory animals have shown that the toxic potency of smoke generally can be estimated from the contributions of fewer than 10 toxic gases.

•   To a first approximation, one can use an average toxic potency value in engineering calculations. However, the toxic potencies of smoke from a large number of materials are significantly worse than the average value.

•   Smoke can cause nonthermal damage to electronic circuitry contacts and building structural components.

•   The heat from a fire can cause skin burns and can weaken a structure to the point of collapse.

•   Using the full range of knowledge to identify the limiting hazard(s) in the event of a fire can help developers and owners of structures implement the most effective fire protection technology.

Key Terms

acute toxic effect(s) The effect(s) on a person during exposure to smoke from a single fire.

asphyxiant gas (narcotic gas) A gas whose inhalation can cause an adverse physiological effect due to lack of oxygen.

chronic toxic effect(s) The accumulated damage from exposure to smoke in multiple fires.

fractional effective concentration (FEC) The sum of the volume fraction of each irritant gas divided by its concentration that causes a harmful effect in the average person.

fractional effective dose (FED) The accumulated product of the volume fractions of narcotic gases and the time interval over which they are inhaled divided by the volume fraction-time product that causes incapacitation, death, or any other harmful effect in the average person.

Haber’s rule The empirical finding that, for a particular gas and toxicological effect, the product of the volume fraction and the exposure time is a constant.

irritant gas A gas that causes a physiological effect by affecting the eyes and/or upper respiratory tract.

limiting hazard The fire threat that first reaches a level at which a person’s life is in peril.

N-gas model The empirical finding that death or incapacitation from smoke inhalation can be attributed to just a few of the numerous components of the smoke.

physical fire model An apparatus, including the operating procedure, test specimen configuration, and combustion environment, that is intended to represent a certain stage of a fire.

postexposure toxic effect(s) The delayed effect(s) on a person attributable to exposure to smoke from a single fire.

Challenging Questions

1.   What are the most important toxic gases in smoke? Which are narcotic gases and which are irritant gases?

2.   What is Haber’s rule?

3.   In a closed room containing 40 m³ of air and an electric fan to promote uniform mixing, how many kilograms of a typical combustible would have to burn to create an atmosphere that might incapacitate people in 30 minutes? Assuming Haber’s rule holds, how many burned kilograms of fuel might cause incapacitation in 10 minutes?

4.   Explain why administering oxygen to fire survivors is so beneficial.

5.   While escaping from a building in which a fire is burning, 40 people pass through a smoky room in which the CO volume fraction is 1000 µL/L, the HCN volume fraction is 50 µL/L, and the CO2 volume fraction is 0.02. Despite some visible smoke in the room, they cross the room in about 10 seconds. Calculate the fractional effective dose of the narcotic gases each person has accumulated and estimate whether all of them will make it to the exit door.

6.   In the same situation described in Problem 5, the HCl volume fraction was 200 µL/L and the SO2 volume fraction was 15 µL/L. Calculate the fractional effective concentration and estimate whether this might keep people from reaching the exit door. Which of the two effects do you think is the limiting hazard and why?

References

1.   Hall, J. R. Jr., and B. Harwood, (1989). “The National Estimates Approach to Fire Statistics.” Fire Technology 25: 99–113.

2.   Gann, R. G., and N. P. Bryner. (2008). Combustion Products and Their Effects on Life Safety. In: The SFPE Handbook of Fire Protection Engineering, 4th ed., DiNenno, P. J., et al., eds. Quincy, MA: National Fire Protection Association.

3.    Birky, M., B. M. Halpin, Y. H. Caplan, R. S. Fisher, J. M. McAllister, and A. M. Dixon. (1979). “Fire Fatality Study.” Fire and Materials 3: 211–217.

4.   Esposito, F. M., and Y. Alarie. (1988). “Inhalation Toxicology of Carbon Monoxide and Hydrogen Cyanide Gases Released during the Thermal Decomposition of Polymers.” Journal of Fire Sciences 6: 195–242.

5.   Levin, B. C., et al. (1990). “Analysis of Carboxyhemoglobin and Cyanide in Blood from Victims of the Dupont Plaza Hotel Fire in Puerto Rico.” Journal of Forensic Sciences 35: 151–168.

6.   Purser, D. A. (2008). Assessment of Hazards to Occupants from Smoke, Toxic Gases, and Heat. In: The SFPE Handbook of Fire Protection Engineering, 4th ed., DiNenno, P. J., et al., eds. Quincy, MA: National Fire Protection Association.

7.   Babrauskas, V., R. H. Harris, Jr., E. Braun, B. C. Levin, M. Paabo, and R. G. Gann. (1991). The Role of Bench-Scale Test Data in Assessing Real-Scale Fire Toxicity. Technical Note 1284. National Institute of Standards and Technology.

8.   Hartzell, G. E., D. N. Priest, and W. G. Switzer. (1985). “Modeling of Toxicological Effects of Fire Gases: II. Mathematical Modeling of Intoxication of Rats by Carbon Monoxide and Hydrogen Cyanide.” Journal of Fire Sciences 3: 115–128.

9.   Kaplan, H. L., and G. E. Hartzell. (1984). “Modeling of Toxicological Effects of Fire Gases: Incapacitating Effects of Fire Gases.” Journal of Fire Sciences 2: 286–305.

10.   Neviaser, J. L., and R. G. Gann. (2004). “Toxic Potency Values for Smoke from Products and Materials.” Fire Technology 40: 177–200.

11.   ISO 13571: Life-Threatening Components of Fire: Guidelines for the Estimation of Time to Compromised Tenability in Fires. (2012). Geneva, Switzerland: International Standards Organization.

12.   ISO 27638-11: Analysis of Blood for Asphyxiant Toxicants: Carbon Monoxide and Hydrogen Cyanide. (2011). Geneva, Switzerland: International Standards Organization.

13.   Kaplan, H. L., A. F. Grand, W. G. Switzer, D. S. Mitchell, W. R. Rogers, and G. E. Hartzell. (1985). “Effects of Combustion Gases on Escape Performance of the Baboon and the Rat.” Journal of Fire Sciences 3: 228–244.

14.   Kaplan, H. L., A. G. Grand, and G. E. Hartzell. (1983). Combustion Toxicology, Principles and Methods. Lancaster, PA: Technomic Publishing.

15.   Clarke, F. B. (1983, September). “Toxicity of Combustion Products: Current Knowledge.” Fire Journal 84–108.

16.   ISO 16312-1: Guidance for Assessing the Validity of Physical Fire Models for Obtaining Fire Effluent Toxicity Data for Fire Hazard and Risk Assessment. Part 1: Criteria. (2011). Geneva, Switzerland: International Standards Organization.

17.   ISO/TR16312-2: Guidance for Assessing the Validity of Physical Fire Models for Obtaining Fire Effluent Toxicity Data for Fire Hazard and Risk Assessment. Part 2: Evaluation of Individual Physical Fire Models. (2011). Geneva, Switzerland: International Standards Organization.

18.    McAllister, T. P., et al. (2008). Structural Fire Response and Probable Collapse Sequence of World Trade Center Building 7, Federal Building and Fire Safety Investigation of the World Trade Center Disaster. NIST NCSTAR 1-9. Gaithersburg, MD: National Institute of Standards and Technology.

19.   Franssen, J. M., and N. Iwankiw. (2008). Structural Fire Engineering of Building Assemblies and Frames. In: SFPE Handbook of Fire Protection Engineering, 4th ed., DiNenno, P. J., et al., eds. Quincy, MA: National Fire Protection Association.

20.   Cote, A. E., ed. (2008). Fire Protection Handbook, 20th ed. Quincy, MA: National Fire Protection Association, Section 12.

21.   Su, J. Z., N. Bénichou, A. C. Bwalya, G. D. Lougheed, B. C. Taber, P. Leroux, et al. (2008). Fire Performance of Houses: Phase I. Study of Unprotected Floor Assemblies in Basement Fire Scenarios: Summary Report. RR-252. Ottawa, ON: National Research Council Canada.

CHAPTER 9

Fire Characteristics: Solid Combustibles

OBJECTIVES

After studying this chapter, you should be able to:

•

 

  List the three significant differences between the burning of a solid fuel and the burning of gaseous and liquid fuels.

•   Describe the thermal and chemical processes that result in the ignition and burning of a solid.

•   Describe how char formation and melting occur and how they affect the burning rate.

•   List the types of combustible solids.

•   Describe the types of polymers and explain how they gasify.
•   Describe at least four classes of mechanisms by which fire retardant additives act to modify the ignition and burning of solids.

•   Discuss the use of calorimetry to measure the heat-release rates of materials and products.

Introduction

Unexpectedly vigorous burning of solid combustibles has been at the core of some of the pivotal fires in our lifetimes. These include the

1

9

4

4 Hartford circus fire, which involved a tent that was waterproofed with paraffin wax; the 196

7

Apollo 1 fire, where the capsule environment consisted of

100

percent oxygen; the 1986 Dupont Plaza Hotel fire, which was fed by stacked unused furniture; The Station and Kiss nightclub fires (

2

00

3

and 2013, respectively), both of which were fed by foam insulation on the walls and ceilings; and the

200

7 Sofa Super Store fire. These fires, and millions of less spectacular fires in the United States, underscore the importance of understanding solid fuels, the means by which they burn, and the hazards these fires present.

Fire Stages and Metrics

Solids versus Gases and Liquids

The presentation of fire characteristics in this text began with the simplest case, gaseous fuels. Given that flaming combustion is a gas-phase process, a fuel that starts out as a gas needs only a pathway to ignition. Furthermore, the fuel composition remains identical to the composition of the initial gas mixture throughout the fire, simplifying the combustion chemistry. A gas-phase flame spreads by the chemical action of the chain that propagates the necessary atoms and free radicals. Liquid fuels have a similar degree of simplicity in their gas-phase fuel chemistry, which most commonly involves the vapor from the liquid. As with a gaseous fuel, there is often no change in the fuel chemistry over time. The energetics of vaporization and, for liquid mixtures, the preferential evaporation and burning of the lighter component(s), however, require consideration, as they affect the rates of burning and surface flame spread.

The burning of a solid fuel has three significant and consequential differences from the burning of gaseous and liquid fuels:

•   Significant chemical change generally occurs within the solid during burning. This change results in (1) the fuel becoming non-uniform and (2) this lack of uniformity varying with the extent of the burning and, therefore, over time.

•   The emitted volatiles may not have the same chemistry as the virgin solid.

•   The heat transfer to, from, and within the solid requires consideration of both the changes in the fuel surface and the chemical changes that have occurred below the surface.

Materials and Products

At this point, two important terms must be clarified:

•   A material is a single substance. The simplest material is made of a single chemical component, such as a sheet of a pure plastic. Some materials, such as particleboard, are mixtures of chemicals—in this case, ground wood and a binder. Still other materials, such as a fiber-reinforced composite, are nonhomogeneous.

•   A product is (or is similar to) an item that is available commercially, and is alternatively referred to as a commercial product or a finished product. Examples include an electric cable, an upholstered chair, and a carpet. Such items are composed of one or more materials and are typically not chemically homogeneous.

Real combustible items are products. Large-scale tests are used to measure the burning behavior of products, whereas the fire properties of materials are often determined in bench-scale tests. Some bench-scale tests have been used to characterize small mock-ups of products. The relationship between the fire properties of a product and the fire properties of its component materials remains a subject of research. The use of reduced-scale mock-ups of large combustibles (e.g., an upholstered chair) to characterize ease of ignition has achieved some success, but the use of mock-ups to characterize mass burning rate and flame spread rate requires care in design of the mock-up and interpretation of the results.

Pyrolysis

The involvement of a solid fuel in a fire generally begins with radiative or conductive heat decomposing the solid into sufficiently small fragments that the fragments are able to escape the solid surface and become a gas-phase fuel. Radiant heat, for example, might come from a nearby space heater or the flames from an already burning item; conductive heat might come from an overheated electrical component. Convection can contribute to these modes of fire evolution, but generally convective flow temperatures are not hot enough to volatilize a solid. An important exception to this statement is the hot upper layer in a room near or post flashover.

The decomposition process for a solid fuel is called pyrolysis. If no oxygen is present, the process is termed anaerobic pyrolysis. Most commonly, pyrolysis occurs in air, in which case it is called oxidative pyrolysis. Anaerobic pyrolysis is endothermic; that is, heat must be supplied from somewhere for the decomposition reactions to occur. Oxidative pyrolysis is usually endothermic or thermally neutral. Pyrolysis typically stops when the heat source is removed or turned off.

As discussed in the Physical and Chemical Change chapter, pyrolyzing a solid requires raising the material’s temperature to the point where chemical bonds begin breaking, overcoming any phase change that might occur during this heating process, and releasing volatile compounds or molecular fragments. The heat input required to accomplish this feat is the heat of gasification of a material (which has units of kJ/g). It is an important measure of the ease of ignition of a solid and the flammability of a solid, once ignited. If the chemistry of the remaining fuel changes over time, so will the heat of gasification.

If it is important to understand (or reconstruct) a particular segment of a fire (perhaps the spread of a fire from one combustible item to another), it is necessary to know the heat of gasification for a product during that time interval. If the task is to determine the threat posed by a large fire to the structural integrity of the building, it may be sufficient to use a heat of gasification averaged over the burn life of the combustible.

The minimum condition for igniting a solid is the heating of its exterior surface to a high enough temperature that the pyrolysis gases are produced rapidly enough to exceed their lower flammability limit in the space above the surface. Unlike the vapor from a pure liquid, the pyrolyzate is commonly a mixture of many decomposition products. Its composition depends on the chemistry of the solid fuel, the rate of pyrolysis, and the availability of oxygen. As a result, no tables of flammability limits for solid fuels can be developed. Instead, the gasification rate for ignition is experimentally determined for a particular heating scenario; as indicated in the next paragraphs, it is not a unique property of the fuel in the same way that a heat of vaporization is unique to a liquid.

If the solid is being heated by conduction, the heat generally is supplied at a location away from the fuel’s outer surface. For example, an overloaded electrical conductor heats the wire insulation from the inside, with the heat then being transferred to the inside of the cable jacket. The surface temperature, then, may be the lowest temperature in the solid. In this case, the apparent heat of gasification will be higher than for the case where radiant heat is applied to the exposed surface of the cable jacket.

If the solid is being heated by radiation, and if the radiation is entirely absorbed by the top surface of the solid (i.e., the solid is optically thick), the top layer of the solid will decompose first, followed by further decomposition in depth. Some solids, however, are somewhat transparent to infrared radiation. An example is an acrylic window panel. In this case, heating also takes place below the surface. When decomposition occurs below the surface, subsurface bubbles of pyrolyzate form, pass through the hotter (lower-density) outer material, reach the surface, and burst, spurting volatiles into the air. This process can be a more efficient mode of gasification than surface heating.

Whether the incident heat involves radiation or conduction, an increase in the heating rate is likely to increase the pyrolysis rate. Conductive heat is important in some ignition modes, but is rarely the principal contributor to fire spread and burning intensity. Chemical kinetic principles, as discussed in the Physical and Chemical Change chapter, reveal that the pyrolysis reaction rate increases rapidly with increasing temperature. Under normal indoor ventilation conditions, gasification rates on the order of just a few grams per square meter per second are needed to achieve an ignitable mixture in air.

Ignition to Flaming Combustion

For most organic solids, a temperature between

5

20 °F and 750 °F (270 °C and 400 °C) is necessary for piloted ignition. As with gaseous and liquid fuels, unpiloted ignition or autoignition is possible if the surface reaches a sufficiently high temperature. For example, when wood is heated radiatively, piloted ignition (initiated by a flame maintained near the surface) occurs when the wood surface reaches a temperature of 570 °F to 750 °F (

300

°C to 400 °C), while the same surface must be heated to about 1100 °F (600 °C) to induce autoignition.

The minimum radiative flux that must impinge on a solid to make it ignitable by a pilot flame has been measured for many materials [1]. These values range from 10 kW/m² to 40 kW/m² depending on the nature of the material, including its chemical constituents, reflectivity, size, and orientation with respect to the radiative source. For fluxes in excess of the minimum value, the time to ignition decreases as the flux increases (

Figure 9-1
).

In most serious fires, more than one combustible product is involved. The first product might be ignited by, for example, a candle. The flames from this product grow and generate far more heat than the original ignition source provided. As a result, there are two distinct modes for the ignition of a second combustible product:

•   The first mode is piloted ignition. If the second product is close to the already burning item, the radiation from the flames pyrolyzes the surface material(s) of the second item. The combination of pyrolysis gases from the two items creates a flammable fuel–air mixture in the space between the two products. The flames extend along this flammable mixture and become attached to the second item.

•   If the second product is farther away from the burning item, it can still become involved by unpiloted radiative ignition. As in the first mode, the thermal radiation from the flames from the burning item pyrolyzes the surface material(s) of the second item. Continuing irradiance leads to increasingly higher surface temperatures, and the pyrolyzate autoignites. This process can be enhanced by the pyrolyzate from the second item absorbing some of the flame radiation from the first item. This heat absorption adds to the temperature rise of the pyrolyzate and shortens the time required to generate a sufficient concentration of flame-propagating free radicals.

During the piloted and unpiloted ignition of solid fuels, potential surface heat losses can affect whether and when ignition will occur. These losses are analogous to the difference between the flash point and the fire point for liquid fuels:

1.  The surface can lose heat by conduction into the interior of the solid. Such a loss will occur if the solid is heated quickly at the surface. If the solid starts to flame, but the heat source is removed, the high temperature at the surface is dissipated by conduction into the depth of the solid. The heat feedback from the flame is insufficient to maintain the surface temperature, and the flame goes out. By contrast, if a thick solid is heated very gradually or from within, when its surface reaches the ignition temperature, its interior is already quite hot and will not drain heat very fast from the surface. Only a little heat feedback from the flame is needed to sustain a flammable concentration of pyrolyzate. Thus, the preignition heating rate and mode are important in achieving sustained ignition of thick solids.

Figure 9-1 Effect of radiative flux intensity on time to achieve piloted ignition [1].

Data from: Babrauskas, V., Ignition Handbook, Fire Science Publishers, Issaquah, WA, 2003.

2.  The surface can lose heat by radiating it away to cooler surroundings. This is best demonstrated by an example. If you place a small burner between two large pieces of wood whose surfaces are facing each other (

Figure 9-2

), the surface will ignite when the temperature of the two wood surfaces approaches 750 °F (400 °C). The surfaces, which quickly become charred and black, radiate energy to each other with an intensity characteristic of a 750 °F (400 °C) black body. When the burner is removed, this reciprocal radiation/absorption continues. There is no net radiative heat loss from either surface, and the flaming can continue.

If the same test is repeated with a single piece of wood, as in the second panel of Figure 9-2, either the surface will not ignite or it will take a larger burner to cause the ignition. In this scenario, there is no incident thermal radiation from a facing surface, as in the two-piece case. When the burner is removed, the heat continues to be radiated to the surroundings, and the surface cools. Soon the surface is not hot enough to generate a flammable concentration of pyrolyzate.

Ignition to Nonflaming Combustion

As presented in the Combustion Fire and Flammability chapter, nonflaming combustion, also called smoldering or glowing combustion, can occur in a material with the following properties:

•   The material is initially porous—that is, it has a large interior surface area as well as numerous internal “tunnels” that enable the diffusion of oxygen to those surfaces.

•   The material’s interior surfaces support exothermic reaction with oxygen, producing or maintaining a self-supporting, porous, carbonaceous char.

•   The material is a good insulator—that is, heat generated by the surface reaction accumulates within the material and is not efficiently lost to the surroundings.

Figure 9-2 Effect of radiative enhancement on sustained ignition. The flames in the left panel are sustained when the burner is removed; the flames in the right panel are not.

Because of the importance of oxygen diffusion, we might expect the smoldering rate to be dependent on the ambient oxygen volume percentage, and smoldering combustion does occur more readily in oxygen-enriched air than in normal air. Data from Reference [2] show a fourfold increase in the smoldering rate of cellulose rods when the oxygen was increased from 21 percent by volume to 96 percent by volume.

Smoldering can be started by a nonflaming ignition source such as a lit cigarette dropped onto an easy chair or an overheated electrical cable passing through a wood stud. When such an ignition source is applied to the material, the local temperature increases by as much as hundreds of kelvins above room temperature, and the reactivity of oxygen with the material surface begins at that temperature. If the smoldering is to self-sustain or progress, the rate at which heat is generated must exceed the rate at which heat is diffused away. When the ignition source is located in the interior of the material, significant insulation surrounds the hot spot and the early reaction may proceed at a fairly slow rate. In contrast, if the ignition occurs at the surface of the material, a higher heat-loss rate must be overcome, and the initial smoldering reaction must be faster.

The thermal process of spontaneous ignition followed by self-heating is similar, albeit with one major difference: the starting temperature. The classic example is the self-ignition of a haystack [3]. Very little happens unless the moisture content of the hay is greater than approximately 25 percent, which could result from rain or relative humidity of the air near 90 percent. Under humid conditions, aerobic fungi and bacteria grow in the hay at normal outdoor temperatures, generating heat biologically. If the haystack is thermally thick (a few meters across) and heat losses are small, then the biologically generated heat will increase the interior temperature of the haystack to about 165 °F (75 °C), perhaps taking several days or weeks. Above this temperature, the organisms are no longer active. At 165 °F (75 °C), however, the rate at which oxygen reacts with the decomposition products of the hay (which were formed during the biological heating) is significant. This chemical heat raises the temperature of the haystack’s interior. As the temperature continues to rise, however slowly, the temperature-dependent oxidation rate accelerates, resulting in 
thermal runaway
. After several more weeks, the temperature could rise to the point where flaming ignition occurs spontaneously. This outcome could also happen if the pile were disturbed (e.g., with a pitchfork), bringing fresh air in contact with the glowing region in the interior or if the glowing zone progressed from the interior to the surface.

The key factor in this scenario is that the initial rate of heat generation can be very slow if the insulation is effective at trapping the heat (

Figure 9-3

). The slope of the green curve in Figure 9-3 depicting the rate of heat generation is initially small but increases progressively with increasing temperature, as is characteristic of chemical reactions. The two heat-loss curves, for a faster cooling rate (red) and a slower cooling rate (blue), are approximately linear (constant slope) because the rate of convective cooling is directly proportional to the difference between the temperature of the warm object and the temperature of the surroundings.

At the initial temperature, the rate of heat generation is positive, but the rate of cooling is zero because the object is initially at the same temperature as the surroundings. As a result, the temperature of the material must rise, and must continue to rise until the cooling curve intersects the heating curve. This intersection point, marked by a black dot in Figure 9-3, corresponds to a balance of heating and cooling, such that no further temperature increase occurs. At the rapid heat-loss rate, the green and red curves intersect, and the heating ceases; that is, the green curve would stop at this temperature. The slower, blue heat-loss curve never intersects the heat-generation curve, and the material proceeds toward sustained burning, as shown in Figure 9-3.

Few materials burst into flame spontaneously. In fact, most common solid materials react so slowly with oxygen at normal temperatures that the self-heating, if measurable at all, usually amounts to a temperature increase of no more than one or two kelvins. Table A-10 in the Fire Protection Handbook [4], however, lists 75 substances that are capable of hazardous spontaneous heating. Among the most dangerous are rags or other fibrous materials in contact with corn oil, fish oils (e.g., cod liver oil), linseed oil, pine oil, soybean oil, tung oil, or any unsaturated oil. Such oils are reactive with oxygen at room temperature. In addition, the rags or fibrous material provide an extensive surface area for the oil–oxygen reaction to take place, and they confine the heat, permitting the temperature to rise. By contrast, saturated oils, such as petroleum-derived oils (common lubricating oil or heating oil) do not cause spontaneous heating. Common materials that are prone to spontaneous ignition, if stored in bulk, include charcoal briquettes, low-grade coal, and some types of animal feed.

Figure 9-3 Relative heat-generation and heat-loss rates, which may sometimes lead to runaway heating and self-ignition.

Char Formation and Melting

After a solid has been ignited and the flame has begun to spread across its surface, two distinct categories of burning behavior are apparent. One class of materials, including woods and certain plastics, burns with the formation of a growing surface char layer. The other class of materials, which includes many of the more common plastics (polyethylenes, polystyrenes, and acrylics1), burns with either no char or a small amount of surface char that blackens the fuel surface but never builds up to a thick layer. The importance of char formation is seen in 

Figure 9-4

, which illustrates heat-release rate versus time for a char-forming material (particleboard) and a non-char-forming material (polymethylmethacrylate [PMMA]).

Figure 9-4 Heat release rates versus time for particleboard and acrylic (polymethylmethacrylate) samples under imposed radiative heat fluxes of 25 kW/m² and 50 kW/m².

Note

In terms of its chemical and physical behavior during burning, PMMA is one of the simplest materials. First, during pyrolysis, nearly all of the mass that is volatilized is in the form of methylmethacrylate (MMA). Second, until it reaches its melting point, PMMA does not melt or drip; it sublimes. For these reasons, PMMA is one of the most convenient materials for developing simple combustion models. Very few solid materials exhibit such ideal behavior, so combustion modeling for realistic materials constitutes to be an active field of research.

Texture: Eky Studio/ShutterStock, Inc.; Steel: © 

Sharpshot/Dreamstime.com

The char formed when the particleboard is heated has a structure similar to that of graphite (pencil lead), a very stable form of carbon. In graphite, the carbon atoms are connected in adjacent six-membered rings (polycyclic structure), forming the orderly structure shown in 
Figure 9-5
. In a fire-generated char, irregularities are present in the array of carbon atoms, and one hydrogen atom is typically bonded to one out of every five or six carbon atoms.

Most of the readily pyrolyzed and flammable material is gone at the point that char has formed. The char is brittle, with a very porous, cellular structure, with thin walls and a large fraction of open space. (Think of a household sponge that is rigid, rather than flexible.)

The char is not a good conductor of heat and protects the subsurface material from the heat of the flames. As the char becomes thicker, it progressively slows the rate of conductive heat transfer from the flames above the surface to the virgin material below the surface. This insulation factor reduces the endothermic pyrolysis of this material to form combustible gases and, therefore, slows the rate of burning. This can be seen as the decreasing heat-release rates for the particleboard specimens in Figure 9-4. (The final small increase in the curves near the time of flameout is artificial. The test specimens were mounted on an insulated sheet, which became heated during the test. The last few millimeters of the particleboard specimen were heated both from the front and from the back, so they burned out more quickly.)

The particleboard curves have been smoothed. In a test, as the char gets thicker, it develops cracks and fissures that can provide narrow pathways to the underlying material, releasing brief spurts of flammable vapor before the pathways char over. This mirrors the charring in an actual fire.

The rate of char formation of woods has been reported to be proportional to the radiant heat flux impinging on the surface [5]. For a typical radiant heat flux of 30 kW/m², which might exist just under a flame, the average charring rate would be approximately 0.025 in./min (0.01 mm/s).

Noncharring combustibles generally melt while burning, so there is no insulating layer to provide thermal protection for the subsurface material. In some cases, the melt is very viscous, and little flowing occurs. In other cases (e.g., with some polyethylenes, polypropylenes, and polystyrenes), the melt has a watery consistency. Such materials tend to burn at a high rate throughout the burning period until the fuel is consumed, as exemplified by the acrylic samples in Figure 9-4. This high burning rate can be enhanced if burning drops of molten plastic fall or flow downward, providing a means of spreading the fire. Noncharring combustibles are generally more hazardous than charring combustibles.

Figure 9-5 Ball and stick portrayal of the structure of graphite. The black balls are all carbon atoms.

Mass Burning and Flame Spread

Mass Burning Rate

Once a solid combustible is ignited, its contribution to a fire’s intensity is determined by three related parameters:

•   The rate at which a unit area of the burning surface is consumed

•   The rate at which flames spread over the fuel surface, increasing the burning surface area

•   The combustible’s ability to ignite other combustible items in the vicinity

The first of these, the rate of mass consumption for a defined segment of a burning combustible, is the subject of this section.

The rate of burning of a material or product is expressed as a mass burning rate (g/m2•s) or heat-release rate (kW/m2). If the incident radiative flux to the specimen is increased, both charring and noncharring materials burn more rapidly.

When a material is burning steadily, there is a heat balance at the surface: The net heat to the surface just balances the heat needed to keep supplying fuel to the flame. The net heat to the surface is the heat flux from the flame to the surface minus the rate at which heat is lost by reradiation from the hot surface to the cold surroundings, with both terms expressed in kW/m2 (which is equivalent to kJ/m2•s). The rate of heat absorption per unit surface area (kJ/m2•s) required to sustain a flow of combustible pyrolyzate is the product of the mass rate of gasification per unit surface area (g/m2•s) and the heat of gasification (kJ/g).

Early in the combustion of the first item burning in a compartment, the heat input is derived solely from its own flames. As the burning surface increases, the radiative feedback to the surface at the edge of the flames is approximately 30 kW/m2. As the fire reaches approximately 250 kW in intensity in a room of normal residential dimensions, the hot fire gases in the upper layer of the room and the flame-heated walls reach temperatures where their black body radiation to the product surface is appreciable. Recalling 

Equation 5-5

, and assuming that the soot is radiating as a black body (emissivity = 1), the heat flux per unit surface area (kW/m2) is given by 

Equation 9-1

:

For an upper layer that has reached 500K, the radiant flux to the burning product will be approximately 3.5 kW/m2, which is small compared to the flame radiation to the surface. At 800 K, a temperature indicative of imminent flashover, the radiant flux will be approximately 23 kW/m2. Post flashover, the upper-layer temperature can reach approximately 1100K, which emits a radiant flux of 83kW/m2.

Table 9-1
 presents some values for heats of gasification of a number of combustible solids. Three observations are worth noting:

1.  The range of the values is quite wide, from 1.19 kJ/g to 3.74 kJ/g.

2.  The chemical composition within a generic type of material may vary considerably, so it comes as no surprise that the data show variation in heats of gasification of samples in each of the rows where there are multiple samples from different sources.

3.  If these heats of gasification (1.19 kJ/g to 3.74 kJ/g) are compared with the heats of combustion of the same materials (15 kJ/g to 44 kJ/g), it is apparent that only a small portion of the heat released by burning must return to the pyrolyzing solid to maintain a continuing supply of combustible vapor to the flame.

Table 9-1 Heat of Gasification for Selected Solids [6]

1

1

1

1

2

2

1

1

1

1

Material Type	Heat of Gasification (kJ/g)	Number of Materials Tested
acrylonitrile-butadiene-styrene (ABS)	3.23									1
corrugated paper	2.21
Douglas fir	1.8 2
nylon 6/61	2.35
phenolic plastic	1.64
polyesters (PETs) with glass fibers	1.39 to 1.75			2
polyethylenes (PEs)	1.75 to 2.32
polyisocyanurate (PIC) foams	1.52 to 3.74
polymethylmethacrylate (PMMA)	1.63
polyoxymethylene (POM)	2.43
polystyrene (PS), granular	1.70
polystyrene foams	1.31 to 1.9 4	5
polyurethane (PU) foams	1.19 to 2.0 5	7
polyvinylchloride (PVC), rigid	2.47

Reproduced from: Tewarson, A., “Generation of Heat and Chemical Compounds of Fires,” 

Chapter 3

–4, SFPE Handbook of Fire Protection Engineering, 4th Edition, DiNenno, P.J. et al, eds., National Fire Protection Association, Quincy, MA, 2008.

Table 9-2 Heat Balance for a Horizontal PMMA Slab Burning in Air [7]

Component

Value

Conditions

Flame base

0.305 m × 0.311 m

Mass-burning rate

1.00 g/s

Surface temperature

725 °F; 385 °C

Heat of gasification

1.61 kJ/g

Heat of combustion

24.9 kJ/g

Combustion efficiency

85 %

Heat-release rate

21.1 kW

Radiative fraction

0.34

Heat flux from flame to surface

By radiation

1.91 kW (73 %)

By convection

0.72 kW (27 %)

Total

2.63 kW

Heat absorbed by gasification

1.61 kW (61 %)

Surface re-radiation loss

1.02 kW (39 %)

Total

2.63 kW

Reproduced from: deRis, J., “Fire Radiation-A Review,” Proceedings of the Combustion Institute 17, 1003–1016, (1979). Copyright Elsevier.

Table 9-2

 shows a detailed energy balance for a 30.5 cm by 31.1 cm (approximately 1 ft by 1 ft) slab of PMMA burning in a horizontal configuration. Approximately three-fourths of the heat transfer from the flame to the surface occurs by flame radiation, and the remaining one-fourth involves convection from the flame gases just above the surface. Almost half of the total heat transfer to the surface is re-radiated from the surface. Thus radiation is important for a fire of this size.

If the test piece of PMMA had been larger, then, because of a thicker flame, the flame radiation flux per unit area would have been even greater, and the burning rate per unit area would have been larger. 

Figure 9-6

 illustrates this effect by showing burning rates for horizontal PMMA slabs of various sizes.

Given the data in 

Figure 9-8

, what would happen to the burning rate per unit area if a very large slab of PMMA (e.g., 16 ft (5 m) square), were burned? The temperature in this very large flame would be about the same as in a smaller one, because, for a heat source at a given temperature, there is an upper limit to the radiant intensity it can generate. Therefore, the rising curve in Figure 9-6 would level off at some burning rate. This limit for PMMA has not been measured, but it has been estimated to be between 30 g/m² s and 60 g/m² s.

Even for a piece of PMMA within the size range shown in Figure 9-6, the mass burning rate can depend on other factors. For example, radiative feedback from hot walls and ceiling or from another flame would increase the burning rate. Reduced oxygen in the air entering the flame due to dilution with combustion products would reduce the burning rate.

Figure 9-6 Mass burning rates of square, horizontal PMMA slabs of various sizes [8].

As expected, a material other than PMMA would likely burn at a different rate. For instance, a 0.305 m by 0.305 m slab of a POM would release heat at less than one-half the rate of PMMA, while the same size slab of a PS would release heat 50 percent faster than PMMA and 3.3 times as fast as POM [7]. These differences reflect differences in the thermal decomposition mechanism of the material, leading to differences in flame radiation, in surface re-radiation, and in heat of gasification.

Flame Spread Rate

Generally, a fire originates at a discrete location on the surface of a combustible product, and spreads from there. For residential fires, this initial ignition area is often small because the most common ignition sources are cigarettes, faulty or improperly located electric devices, and heating apparatus. Once the flame has progressed a short distance from the source, its rate of flame spread becomes independent of the ignition source, but is dependent on four other variables:

1.  Orientation of spread

2.  Degree of radiative preheating

3.  The magnitude and direction of any external air flow

4.  Thermal thickness of the solid

The rate of fire spread in a horizontal or downward direction over the surface of a thermally thick solid is generally very slow (“creeping flame spread”), on the order of a few hundredths of a millimeter per second (0.1 in./min), unless the surface has been preheated. For example, if PMMA is preheated for 1 minute with a radiative flux of 20 kW/m², the downward spread rate increases from 0.05 mm/s to 0.5 mm/s—that is, it increases by a factor of 10. More prolonged or more intense preheating can produce further acceleration, perhaps by at least another factor of 10. If the burning material melts and drips, this presents an alternative means for spreading flames downward. In such cases, spread rates can be much faster.

Flame spread in an upward direction is far more rapid than horizontal or downward spread. Furthermore, the flame accelerates as it spreads upward. The reasons for this difference in behavior are as follows.

The hot gases in a flame rise though natural buoyancy. As they do, fresh air is drawn into the base of the flame. For downward or horizontal spread, the direction of this air flow pushes the flame away from the unburned material, as shown in the left and center panels of 
Figure 9-7
. Thus, very little of the flame radiation impinges on the adjacent fuel surface, providing little heat for decomposition and gasification. Even if the surface is heated to some degree by the flame radiation, it is simultaneously cooled by the approaching air.

Conversely, with upward spread, the flame is in close contact with the not-yet-ignited portion of the combustible and can preheat it efficiently by both convection and radiation, so rapid upward spread can be expected. Furthermore, as the flame spreads upward, it becomes taller. Its greater length and thickness promote radiative heat transfer, and its greater length and higher gas velocity promote convective heat transfer. Thus, the upward spread rate increases progressively. On high walls, flame spread speeds of several meters per second are possible.

Figure 9-7 Downward, horizontal, and upward flame spread and direction of buoyancy-induced airflow at the base of the flame. The heavy line on the surface denotes the burning area.

A powerful factor enhancing the flame spread rate is the presence of an additional air flow in the direction that the flame is already moving. This flow could be due to cross-ventilation from open windows or could result from a mechanical ventilation system. This co-flowing air extends the flame, increasing the fuel surface area subject to the flame’s thermal radiation and thus boosting the generation rate of pyrolysis gases. The air flow also provides the additional oxygen needed to combust these gases.

An external air flow counter to the direction of flame spread has an opposite effect. It pushes the flames back toward fuel that is already burning or has already been consumed. This decreases the fuel surface area subject to flame irradiation, slowing the spread rate. However, the increased supply of oxygen can increase the mass burning rate of the fuel surface that is already burning.

The fourth factor affecting flame spread, in addition to orientation, radiative preheating, and air flow, is the thermal thickness of the solid. Downward spread experiments with thermally thin cardboard and fabric samples have shown that the rate of spread is inversely proportional to the physical thickness. This behavior can be predicted from the concept that, for the flame to spread, it must heat the adjacent material to the ignition temperature. Conductive heat transfer through the thickness of thin specimens is fast because the heat does not have far to travel. Thus, if the material is twice as thick, it will take twice as long to be heated, and the flame will advance half as quickly. This phenomenon explains why kindling wood is used to start a campfire. As a consequence of this principle, flames can spread very rapidly over a material composed of thin elements, such as a pile of wood shavings or tissue paper.

Effects of Uncommon Fire Environments

In special environments, such as in the vicinity of pure oxygen being used for medical purposes, the percentage of oxygen by volume exceeds the percentage in normal air. This higher percentage of oxygen greatly increases the fire hazard. Although the heat capacities of oxygen and

nitrogen

are comparable, the increased percentage of oxygen present makes the combustion reactions proceed at a faster rate, in turn generating heat at a faster rate. The combination of faster chemistry and no change in thermal mass leads to a higher flame temperature. The higher flame temperature has several consequences:

•   The rate of heat transfer from the flame to the surroundings is greater, increasing both the flame spread rate and the potential for igniting nearby combustibles.

•   The flame is less easily quenched by still cold, adjacent fuel surfaces, so the flame can spread more rapidly.

•   The process of soot formation in a turbulent diffusion flame is enhanced by higher temperature, so the flame becomes sootier and emits more radiation.

•   Certain solids that will not burn in normal air, unless preheated, will burn in oxygen-enriched air.

Figure 9-8

 shows how the rate of horizontal flame spread increases with increasing oxygen percentage for charring and noncharring materials. The magnitude of the change in the rate of flame spread depends on the thickness and orientation of the specimen. The burning rate per unit area also increases at higher oxygen concentrations, as shown in 
Figure 9-9
.

In some situations, the air pressure is not the 101 kPa found at sea level. The cabin of a commercial aircraft flying at high altitude, for example, is normally pressurized to an absolute pressure of approximately 75 kPa (0.75 atm), and the pressure in Denver, the “mile-high city,” is similarly lower than 1 atm. Hyperbaric chambers for medical treatment can operate at pressures of as much as 300 kPa, sometimes at atmospheres containing 100 percent oxygen.

If the volume percentage of oxygen in the environment remains 21 percent, then the flame temperature will be nearly the same for atmospheric-and non-atmospheric-pressure fires. Thus, the differences in fire behavior are not as profound as when the percentage of oxygen changes.

Figure 9-8 Effect of oxygen percentage in the atmosphere on horizontal flame spread over surfaces [9, 11].

Data from: Lastrina, F.A., Magee, R.S., and McAlevy, F.R., “Flame Spread Over Fuel Beds: Solid Phase Energy Considerations,” Proceedings of the Combustion Institute 13, 935–948, (1971).

Figure 9-9 Effect of oxygen percentage on the burning rate on small, thin, horizontal samples of plastics and heptane [6].

Reproduced from: Tewarson, A., “Generation of Heat and Chemical Compounds of Fires,” Chapter 3–4, SFPE Handbook of Fire Protection Engineering, 4th Edition, DiNenno, P.J. et al, eds., National Fire Protection Association, Quincy, MA, 2008.

The rate of flame spread over solid surfaces increases with increasing pressure. For instance, the rates of horizontal spread over thick slabs of PMMA and polystyrene vary with the 0.82 power and 0.76 power of pressure, respectively [9]. Other fire properties, such as ignition temperature, tendency to produce smoke, flame radiation, and toxic gas concentrations, also vary somewhat with pressure.

Most atmospheres that contain enough oxygen to be breathable will support combustion. Nevertheless, it is fascinating to note that a mixture of 10 percent oxygen and 90 percent nitrogen at 2 atmospheres absolute pressure is easily breathable, but will not support combustion of most materials.

The advent of travel though outer space has introduced the dynamic of reduced gravity into fire safety considerations related to the astronauts’ environment. A fire in a vehicle orbiting Earth or continuing to further planets would behave differently from an ordinary fire on Earth. There is a near absence of gravitational force in these vehicles. (The term microgravity is used instead of zero gravity because of very small residual gravitational effects.) Thus, the fire effluent would not be buoyant and would not rise from the combustion zone to make room for the fresh air needed to continue the oxidation. However, experiments in the space shuttle have shown that flames in this environment do not smother themselves, but rather continue to burn by a diffusion process. A wax candle, for example, has continued to burn for 45 minutes. It takes less energy to ignite a solid in low gravity because of the reduced convective cooling of the heated surface by the surrounding cold atmosphere. Flame spread rates across surfaces and burning rates are somewhat slower, but the forced convection imposed by spacecraft ventilation systems can accelerate burning rate and flame spread. Flames are generally cooler, are somewhat sootier, and have different shapes from flames in Earth’s gravity. The burning of some materials in low gravity might result in hot globules of the material drifting in all directions, possibly starting other fires [10].

Test Methods for Measuring Ignition, Burning Rate, and Flame Spread Rate

From the preceding discussion, it should be clear that ignition and heat release and flame spread rates are sensitive to the magnitude of the incident thermal radiation. Accordingly, a test method intended to provide meaningful, and not misleading, information about a material or product must simulate the radiation correctly.

Figure 9-10
 shows how the heat-release rate of a slab of material can react to the imposed radiative flux. Note the difference in magnitude of the heat release rates for each material as the external flux increases by 10 kW/m² or 20 kW/m². In the case of the wood specimens, a single thick slab will not even continue to burn except with an external radiative flux.

Thermal radiation may come from four sources: the flames from the burning product, the flames from other burning objects, the hot smoke (generally in the upper layer of the room), and the hot ceiling and upper walls. Although the specimens in small-scale test methods are typically of the order of 0.1 meter in dimension, the first of these sources might be replicated. However, the other three sources are not directly related to the test specimen; their values vary depending on the fire scenario.

Several apparatus have been developed that impose a variable radiant flux on a specimen [12]. In these apparatus, the thermal radiation comes from electric heating elements or gas-fired panels. Much effort has been expended on relating the results of these laboratory tests to the results from burning of large combustibles, with some success.

One such apparatus is the basis for ASTM E 1354, Standard Test Method for Heat and Visible Smoke Release Rates for Materials and Products Using an Oxygen Consumption Calorimeter (
Figure 9-11
). In this apparatus, a uniform radiant flux is imposed on a specimen up to 10 cm × 10 cm in area and 5 cm thick. Because the thermal radiation source is shaped like a truncated cone, the apparatus is colloquially called the “cone calorimeter.” The specimen is weighed continuously during a test. The rate of heat release is measured using oxygen consumption calorimetry, whose underlying principle is presented later in this chapter. Dividing the rate of heat release by the rate of mass loss and comparing this number with the heat of complete combustion of the specimen provides an indicator of the combustion efficiency—that is, what fraction of the theoretical heat release is actually generated. Furthermore, by testing specimens at different values of the irradiance from the cone heater, one can determine ignition delay times and the critical radiant flux for ignition.

For decades, the Oxygen Index Test [14] has been used as an indicator of changes in flammability as the formulation of a material is changed. In this test, a mixture of oxygen and nitrogen flows upward over a vertical, pencil-sized sample. The initial volume percent of oxygen is sufficiently high that the specimen can be ignited at the top. The volume percent of oxygen is then reduced until the flame goes out. The percentage of oxygen by volume at which the flame is extinguished is called the 
limiting oxygen index (LOI)
. A high LOI value indicates a material that is more difficult to ignite and/or burns less vigorously, if at all. The results of the Oxygen Index Test do not correlate consistently with ignition delay time, flame spread rate, or heat release rate. Furthermore, under the conditions of the Oxygen Index Test, some known-combustible materials (e.g., red oak) do not burn in air. (If the temperature of test specimens were increased by heating the air flow or with radiant heaters, the measured LOI values would decrease dramatically.) Nonetheless, LOI values, which are available for many materials [4], are frequently used to estimate relative flammability.2

Figure 9-10 Heat release rates of solids as influenced by externally imposed radiation [5, 13]. The magnitude of the heat release rate per unit area depends on the size and orientation of the sample.

Data from: Lastrina, F.A., Magee, R.S., and McAlevy, F.R., “Flame Spread Over Fuel Beds: Solid Phase Energy Considerations,” Proceedings of the Combustion Institute 13: 935–948, (1971).

Figure 9-11 Schematic of the cone calorimeter.

Combustible Solids

Cellulosic and Other Natural Materials

Until the middle of the 20th century, nearly all combustibles in residences, offices, and other structures were made of natural materials. Mattresses and upholstered chairs were stuffed with cotton and horsehair and covered with cotton, wool, and leather; rugs were made of cotton and wool; draperies were woven from cotton, wool, and silk. These materials (with the exception of horsehair) remain popular today, although synthetic materials are now more common in soft goods. Woods and wood products continue to be prevalent as the hard materials in furniture, bookcases, and similar items.

From a fire viewpoint, the most important of these natural materials are 
cellulosics
—that is, materials whose molecular chemistry is based on cellulose. Cellulose, which can be isolated from cotton or chemically synthesized, is a condensation polymer of glucose, C6H12O6, a form of sugar. The chemical structures of glucose and cellulose are shown in 

Figure 9-12

.

Figure 9-12 The structures of glucose and cellulose.

A 
polymer
 is a long string of repeated chemical building blocks. A 
condensation polymer
 is formed by eliminating a small molecule—in this case, water (H2O)—each time another link is added to the polymer chain. As an example, consider this sequence for the formation of cellulose from glucose:

where n generally exceeds 20,000 C6H10O5 units in each molecule of cellulose. If cellulose is boiled in an acidic solution, it decomposes, the water is restored, and glucose is a product. Cotton consists of more than 90 percent cellulose.

Woods contain only 40 percent to 50 percent cellulose. The process of making paper from wood removes much of the noncellulosic material; as a consequence, paper has a fairly high cellulose content. The exact percentage depends on the type of paper.

Many types of wood exist, and they exhibit substantial variations in their physical properties. For example, the densities of ponderosa pine (a softwood) and white oak (a hardwood) are 0.42 g/m³ and 0.73 g/m³, respectively. The chemical composition of dry woods also varies, depending on the type of wood and the growing conditions:

•   40 to 50 percent cellulose

•   18 to 35 percent lignin

•   10 to 30 percent hemicelluloses

•   5 to 20 percent “extractives” (e.g., oils, tars, gums)

•   

0.2

to 1 percent minerals

Woods are somewhat porous in structure and can absorb large amounts of moisture. After long contact with dry air at 68 °F (20 °C) and 20 percent relative humidity, a typical wood has a moisture content of about 5 percent. In extreme conditions, this moisture content can increase to as much as 25 percent. The fire behavior of a wood, especially its ease of ignition, is influenced by the moisture content. The heat absorbed by a wood product in an incipient fire first eliminates the absorbed water, a process that occurs at or below the boiling point of water (i.e., 212 °F; 100 °C). The temperature of the wood stalls in this low temperature range and does not resume its rise until the water is gone. Consequently, it takes a lot of heat to raise the temperature of the wood to approximately 480 °F (250 °C), at which point cellulose gasification begins, with ignition then following. The same process slows the rate of flame spread and the mass burning rate of a moist wood or wood product. Conversely, drier wood has less water to eliminate and burns faster. This is why wildland fires grow so rapidly when there has been little rain and when they are spurred by hot, dry winds.

Note

Lignin, a cross-linked polymer (explained later in this chapter), acts to bind the cellulose fibers together and add strength. (When brown wrapping paper is made from wood, not all the lignin is removed, providing for a stronger paper. Much more of the lignin is removed in making tissue paper.) When heated, lignin decomposes in a manner different from cellulose. Cellulose starts to decompose at about 480 °F (250 °C), and is mostly decomposed at 700 °F (370 °C), leaving a small amount of char behind. Lignin starts to decompose at temperatures below 570 °F (300 °C), but even after prolonged heating at high temperatures in an inert atmosphere, roughly half of the original mass remains as char. Hemicelluloses are polymers of glucose and other sugars, and have a much lower molecular weight than cellulose. One of these molecules can contain several hundred sugar units instead of tens of thousands of such units.

Texture: Eky Studio/ShutterStock, Inc.; Steel: © Sharpshot/Dreamstime.com

Looking at the structure of cellulose in Figure 9-12 and adding the knowledge that wood products contain other polymeric components, we might expect that the gasification of woods would be a complex process. Indeed, the hundreds of pyrolysis products of woods include

carbon dioxide

, water, the pleasantly odorous gases from a fireplace, toxic and irritant gases such as carbon monoxide and acrolein (covered in the Smoke and Heat Hazards chapter), the creosote deposited in chimneys, and the charred remains often seen on the fire ground. The relative proportions of the various species depend on the type of wood, the heating conditions, and the availability of oxygen.

Figure 9-13
 shows the gasification of a wood when this material is heated slowly in an inert (oxygen-free) atmosphere. This gasification process occurs over a fairly wide range of temperatures, leaving a substantial char residue (approximately 25 percent of the original mass).

The net heat of combustion of dry woods, approximately 18 kJ/g, reflects both the heat of combustion of the volatiles produced during pyrolysis (approximately 14 kJ/g of volatiles) and the heat of combustion of the residual char (approximately 34 kJ/g of char). In an actual fire, the char usually burns at a later time than the volatiles, if it burns at all.

Other natural materials in general use include wool, leather, and silk—all animal products with high protein content. The proteins are polymeric molecules with the monomer units (amino acids, An) connected by peptide linkages:

Upon thermal decomposition of these materials, the gases evolved include ammonia (NH3), amines (e.g., CH3NH2), and some hydrogen cyanide (HCN). Wool also contains sulfur, so its pyrolysis generates sulfur-containing compounds. Generally, these animal-based materials are appreciably less flammable than cellulosic materials, but they do burn under high heat fluxes and produce nitrogen-containing irritant and toxic gases and, in the case of wool, sulfur-containing gases, in addition to the usual toxic species found in cellulosic fires.

Figure 9-13 Gasification of wood during slow heating [15].

Data from: Kirk-Othmer Encyclopedia of Chemical Technology, 5th edition, Kroschwitz, J.I., ed., J. Wiley, New York, 2007.

Synthetic Polymeric Materials

Definitions

Following World War II, many products made of human-made polymers entered the marketplace. These materials are not limited to those chemistries that have evolved in nature; and they exhibit diverse physical, visual, and chemical properties. Today, synthetic polymeric materials dominate such product lines as upholstered furniture, paint, carpet, appliance housings, and clothing.

As noted in the previous section, a polymer is a long strand of repeated chemical building blocks. The building block is called a 
monomer
. A molecule consisting of a few monomers is called an 
oligomer
. More specifically, a molecule consisting of two monomers is a dimer, three monomers combine to form a trimer, and so on.

Polymers can be sorted in different ways:

•   Addition and condensation polymers. This classification is based on the nature of the chemical synthesis. These polymers are the subject of the following sections.

•   Linear, branched, or cross-linked polymers. This classification is based on the three geometric configurations of the assembled monomer units.

•   Homopolymer or copolymer. In a 
homopolymer
, all of the building blocks are identical. A 
copolymer
 consists of two or more different monomers.

•   Thermoplastic, thermoset, and elastomeric polymers. This classification is based on polymers’ use properties. 
Thermoplastic polymers
 can be melted and remolded. 
Thermoset
 
polymers
 are cured irreversibly once they are in their final shape. 
Elastomeric polymers
, also called elastomers or rubbers, realign their strands when stressed.

Addition Polymers

The building of an 
addition polymer
 involves the addition of successive monomeric units without elimination of any molecule or fragment. The earliest addition polymer, a polyethylene, was first synthesized in 1898, with a practical synthesis developed in 1933. Both developments were accito and twisted around each other, held together by weak attractive forces. The melting of polyethylene, a thermoplastic vinyl polymer, is the result of the absorbed heat overcoming these attractive forces.

The monomer of an addition polymer has a carbon–carbon double bond. Ethylene (C2H4) is the simplest of these monomers. Its polymer chain comprises a string of pairs of carbon atoms, each singly bonded to its neighboring carbon atoms and to two hydrogen atoms. (See the Physical and Chemical Change chapter.) After the synthesis of the polymer, the carbon atoms in the pairs are indistinguishable; hence, this polymer was originally called polymethylene. Polyethylenes are used in such diverse products as plumbing piping, bubble wrap, and plastic bottles. The harder polyethylenes can be used in machine parts and artificial joints.

Polyethylene is the simplest of an important family of additional polymers called 
vinyl polymers
, so named because of the presence of the H2CH fragment, called the vinyl radical. The carbon on the right has only three bonds to it. Binding an atom or molecular fragment to this site enables the formation of a variety of vinyl monomers. 
Figure 9-14
 shows a number of these substituted vinyl monomers and indicates the commercially important polymers formed from them.

Vinyl monomers can form linear polymers that consist of long chains of carbon atoms (the backbone) with hydrogen atoms and pendant groups attached. The term “linear” means that the chains do not contain any branches. However, the carbon atoms do not actually lie on a straight line; the chains are twisted and coiled. A piece of polyethylene consists of a very large number of polymer strands, lying next to and twisted around each other, held together by weak attractive forces. The melting of polyethylene, a thermoplastic vinyl polymer, is the result of the absorbed heat overcoming these attractive forces.

The adjacent strands can be tied together chemically to form a three-dimensional network and thereby reduce the flexibility, increase the strength or hardness, or increase the melting point of the polymer. This process is called cross-linking. For vinyl polymers, cross-linking is accomplished by adding a small percentage of a monomer containing two double bonds per molecule (a cross-linking agent) to the principal monomer. For example, if 1 percent of divinylbenzene monomer, CH2CH—(CH6H4)—CHCH2 is added to styrene (C6H5—CHCH2) monomer, then linkages will form during polymerization that tie together nearby polystyrene chains at various random points. Similarly, the “B” in the well-known ABS copolymer is butadiene, CH2CH—CHCH2, which acts to cross-link the otherwise linear chains of the mixture of the two vinyl monomers acrylonitrile (“A”) and styrene (“S”). ABS is used in waste pipes, automotive trim components, medical devices, enclosures for electrical and electronic assemblies, safety helmets, canoes, luggage, and small appliances housings.

Cross-linking has an important stabilizing effect on the fire behavior of a polymer. A linear polymer tends to decompose by thermal fracturing of the backbone—that is, by breaking one of the carbon–carbon bonds. The smaller fragments, generally monomers or oligomers containing 2 to 12 carbon atoms, gasify. When their concentration in air exceeds the lower flammability limit, ignition or flame spread can occur. All of the carbon– carbon bonds in the backbone of a vinyl polymer are approximately the same strength. Eventually, the backbone becomes nearly completely fragmented into gasified species, leaving little or no residue. By contrast, a cross-linked polymer can be visualized as a ladder. Even after breaking a rung or an upright, connections continue to hold the ladder together. For a fragment of a cross-linked polymer to break off and gasify, multiple carbon– carbon bonds must be fractured, which in turn requires a higher temperature. At the higher temperature, other bonds can be broken or rearranged, providing an opportunity for the formation of the graphitic carbon structure of a charred residue. The carbon in the char is not readily gasified, so the mass of combustible volatiles is reduced.

Figure 9-14 Some vinyl monomers capable of forming vinyl polymers by linear addition.

Figures 9-15

 and 

9-16

 show some other monomers that are not built on a vinyl structure, but that also serve as precursors to addition polymers.

The pyrolysis of specific addition polymers, described generically earlier in this chapter, differs depending on the polymer. Three basic modes of decomposition are possible:

1.  Unzipping. The backbone breakage preferentially occurs at the linkage to the end monomer unit. The composition of the pyrolyzate is thus dominated by monomer. Examples of polymers that unzip include PMMA, POM, and PTFE.

2.  Random scission. The backbone breaks at random locations, and the smaller oligomers vaporize. The pyrolysis of polyethylene and polypropylene follows this pattern.

3.  Elimination. Polyvinyl chloride (PVC) pyrolyzes in a manner different from either unzipping or random scission. At approximately 482 °F (250 °C), the HCl molecule splits out from the chain and gasifies. The remaining backbone, with unbound sites at nearly all carbon atoms, rearranges to form a stable graphitic char. Very little of the carbon-containing mass undergoes gasification, so pure PVC is very difficult to ignite or spread flames. Unfortunately, HCl is toxic and corrosive.

Note

Pure PVC is a rigid material, and substantial proportions of plasticizers are often added to it to impart flexibility. The resulting plasticized PVC can be used as electric cable insulation or vinyl furniture covering. In general, the plasticizers volatilize or decompose, which increases the flammability of the commercial PVC product. However, the released HCl is a significant flame inhibitor, so even plasticized PVC products are difficult to burn.

Texture: Eky Studio/ShutterStock, Inc.; Steel: © Sharpshot/Dreamstime.com

A few other addition polymers merit mention.

•   Modacrylics are copolymers of acrylonitrile and either vinyl chloride or vinylidene chloride (CH2Cl2). The introduction of the chlorine atoms improves the fire performance relative to ordinary acrylic polymers. Uses of such polymers include synthetic fur, rugs and carpets, work clothing, and wigs.

•   Totally fluorinated polymers, such as the polytetrafluoroethylenes, are almost nonflammable. (They do burn in pure oxygen.) Partial fluorination of a polymer does not improve fire performance as much as chlorination or bromination. The incorporation of bromine in a polymer is more effective against fire than chlorination.

•   Polyurethanes are used widely both in solid form and as foam, with the foam being either flexible or rigid. These materials’ burning properties vary widely, depending on their physical and chemical properties.

•   Isocyanurate cross-linked rigid foams, used for insulation, are less flammable than flexible foams.

Condensation Polymers

Although some condensation polymers were processed in the 19th century, the first commercially successful synthesized condensation polymers date to the 1930s. Neoprene and nylon originally were brand names, but are now generic names for families of polymers (and, therefore, are not capitalized). The first polyesters date to the same decade. Neoprenes are used as cell phone sleeves, knee braces, electrical insulation, and automotive belts. Nylons appear in carpets, clothing, tents, automotive cams and bearings, and stockings (the original use). Polyesters are used in clothing, upholstery fabrics, rope, and pillow stuffing.

As described earlier in this chapter, a condensation polymer is formed by eliminating a small molecule each time another link is added to the polymer chain. 
Figure 9-17
 shows some examples of condensation polymers. Each monomer must contain two functional groups for a chain to be constructed. These polymers involve combinations of organic acids (—COOH) with alcohols (—OH), or organic acids with amines (—NH2), with subsequent elimination of H2O. Cellulose, discussed previously, is a condensation polymer. Note that aramids (commercially Nomex and Kevlar) have greater thermal stability than nylons because of the aromatic carbon rings in the polymer chain. Protective gear for fire fighters, flight crews, and racecar drivers contains aramids.

Figure 9-15 Some nonvinyl monomers capable of forming addition polymers.

Data from: Kirk-Othmer Encyclopedia of Chemical Technology, 5th edition, Kroschwitz, J.I., ed., J. Wiley, New York, 2007.

Figure 9-16 Formation of polyurethanes and polyisocyanurates (addition polymers). R and R’ denote various organic radicals. Partial substitution of a trifunctional isocyanurate permits cross-linking.

The polymers presented so far have been thermoplastics, but an array of thermoset polymers is also available. These extensively crosslinked materials do not melt, but char when heated. Urea (NH2CONH2)–formaldehyde (HCHO) polymers are used as decorative laminates, as the binder in particleboard, and as additives to impart wrinkle resistance to fabrics. Phenol (C6H5OH)–formaldehyde polymers, also called phenolics, are used in circuit boards and hard molded products, such as countertops. Melamine [C3N3(NH2)3]–formaldehyde polymers, also called melamines, are used as decorative laminates and kitchen utensils.

Figure 9-17 Reactions leading to condensation polymers.

Many additional polymers exist. Information about specific polymers can be found in the Kirk-Othmer Encyclopedia of Chemical Technology [15].

Commercial plastics are not pure polymers; they typically contain substantial proportions of additives to obtain the desired mechanical, thermal, electrical, and visual properties, or inexpensive fillers to reduce their cost. Owing to these additives and fillers, the properties and combustion behavior of a commercial plastic may not always match those of the polymer from which it is made. Also, polymers can be made into fibers and fabrics, elastomers (rubbery materials), foams, (flexible or rigid), coatings, films, or solid plastics; each of these forms has different fire properties.

Fire Retardants

Certain properties of solid materials characterize each material’s contribution to fire hazard, including the material’s ease of ignition to both flaming and smoldering combustion, mass burning rate, flame spread rate, soot production, and persistence of glowing combustion following the extinguishment of flames. Some materials and products have been identified as significant contributors to numerous or particularly disastrous fires, and tests have been developed to identify and discourage/prohibit the use of these materials.

Some materials inherently do well in these tests. Many otherwise desirable polymeric materials, both natural and synthetic, do not. So that they can pass the tests, the latter materials are modified by the addition of chemicals called 
fire retardants
. The resulting material might be explicitly designated as “fire retardant,” although many products do not bear such labels. Note that no fire safety regulations actually require the use of fire retardant (FR) chemicals.

Instead, these chemicals are added to materials to expand consumer choice by enabling a wider range of materials to be used for a particular product. This demand has led to the worldwide production of FR chemicals with a value in excess of $4 billion (7 billion kg) annually. Approximately half of that volume consists of halogen (bromine and chlorine) compounds, with the rest being mainly alumina trihydrate, antimony oxide, phosphorus-containing compounds, and boron-containing compounds.

Most fire retardants provide resistance to small ignition sources such as a cigarette, match, or spark. For carpeting, the ignition source is an incendiary tablet; hence the relevant test is called the “pill test.” For expansive applications such as wall coverings or mattresses, testing for rates of heat release, flame spread, or smoke production involves a large flame as the ignition source, providing ample opportunity for ignition.

People generally choose products for properties other than low flammability—for example, flexibility, durability, strength, electrical and thermal insulation, color variety and pattern, resilience, and texture. Large amounts of additives can compromise these properties and add to the purchase price; thus FR chemicals are typically added to a material at or near the minimal level required to pass the fire test and enter the marketplace. Under more severe fire conditions (e.g., high radiant flux or elevated oxygen concentration), then, the fire retardant effectiveness may be significantly lessened or even negated.

All but a few fire tests are performed with new materials or products. Should a product’s formulation change with use or age, the fire performance would also change. For example, if a FR additive was not bound to the host material, it could migrate over time. Some segments of the product would then become richer in the additive, while other segments would contain less than the original concentration and might not show the degree of fire protection expected from the original test performance.

FR additives act to modify the ignition and burning of solids through at least four known mechanisms. Within these classes, the chemical details are not always completely understood. In some cases, a fire retardant or mixture of fire retardants has been found to act through more than one of these mechanisms. Thus the formulation of a FR “cocktail” for a particular polymer and a particular application is often empirically developed. The four mechanism classes are as follows:

Note

Fire retardant chemicals are currently the subject of extensive public discussion. Some critics claim that these additives are harmful to people and the environment worldwide and are ineffective in reducing product flammability, and thus are unnecessary. Others contend that fire retardants have saved lives and reduced fire losses, that the evidence for harm is limited to a very few commercial chemicals, and that the potential harm is outweighed by the fire loss prevention. To date, no systematic, scientific appraisal of the risks and benefits of all the various types of FR additives has been published. However, for a few FR chemicals, sufficient evidence has been gathered to ban them from sale or to withdraw them from the marketplace.

Texture: Eky Studio/ShutterStock, Inc.; Steel: © Sharpshot/Dreamstime.com

1.  Char formation (residue enhancement). The additive promotes the formation of large, stable molecules within the solid, resulting in a carbonaceous char or a tarry substance. This promotion occurs faster than the breakdown of the polymer into small, volatile fragments that can ignite. Some metal-containing and phosphorus-containing compounds operate this way. The more effective this type of FR additive is, the smaller the mass of carbon oxidized and the lower the heat release and heat release rate.

2.  Gas-phase flame inhibition. The additive releases gases that slow or extinguish the gaseous combustion reactions by dilution, cooling, and/or chemical interference with chain branching or propagation reactions. Such additives often involve halogen, metal, or phosphorus atoms. Slowing or termination of the combustion chemistry can lead to enhanced escape of the products of incomplete combustion, such as soot or carbon monoxide. Should the flames be extinguished, the decrease in the burned mass outweighs the modest increase in incomplete combustion products.

3.  Solid-phase heat absorption. The additive decomposes endothermically, absorbing heat that otherwise would have been available to decompose the host polymer. Hydrated alumina (alumina trihydrate, Al2O3 · 3H2O) and limestone (CaCO3) operate in this manner. When heated, the compound decomposes with absorption of heat and release of an inert gas (H2O or CO2, respectively, in the two examples here), which cools and dilutes the flame:

The released gas also provides some flame retardancy by diluting the ambient oxygen.

4.  Barrier formation. The additive forms a glaze or a foam over the fuel surface, which to some degree isolates the subsurface material from the flame above. Char formation also provides a barrier. Some phosphorus-containing compounds act by this mechanism.

As discussed in the Fire Fighting Chemicals chapter, externally applied fire suppressants work by similar mechanisms.

The following are some examples of the application of FR additives:

•   Woods are impregnated with inorganic salts to promote char formation. The positive ions of the most effective salts are ammonium (NH4+), sodium (Na+), potassium (K+), and zinc (Zn+2), while the most effective negative ions are phosphate (PO4)–3, borate (BO3)–3, silicate (SiO4)–4, sulfate (SO4)–2, and sulfamate (NH2SO3)–. Woods are also impregnated with organic compounds containing phosphorus, boron, halogens, or nitrogen (usually as NH2 compounds).

•   Cellulosic materials are filled with borax and/or boric acid to reduce the tendency to smolder or flame, respectively.

•   Rigid solid materials are coated with fire retardant paints, including intumescent paints, which expand into a foam when heated. The expanded foam shields the host materials from flame radiation and slows the transport of pyrolyzate to the flame. These FR coatings, however, often do not survive long exposure to wet or very humid atmospheres. They are also subject to abrasion.

•   Synthetic polymers are copolymerized with small proportions of bromine-containing monomers.

•   Vinyl polymers achieve fire retardant characteristics through a combination of a halogenated organic compound and antimony oxide (Sb2O3). The current understanding is that the effectiveness of this combination results from formation of a volatile antimony halide or antimony oxyhalide (e.g., SbOCl), which inhibits the gaseous combustion reactions.

References [16] and [17] at the end of this chapter provide further information on the chemistry and uses of fire retardants.

In some applications, the required fire performance cannot be achieved with levels of fire retardant additives that do not interfere with other required properties of the material or product. In those cases, a product may be wrapped with a 
barrier material
, also called a fire-blocking layer. Most often, these materials slow or prevent the passage of pyrolyzed material to fresh air and/or absorb the flame radiation without passing it to the vulnerable materials within.

Composite Materials and Furnishings

Most commonly encountered combustible items are 
composite
 in nature, rather than consisting of a single material. For example, an upholstered chair might consist of a wood frame, a polyurethane foam pad, a layer of a polyester fiber, and a wool cover fabric. A rug might consist of a nylon pile, jute padding, and a (nonslip) rubber backing. The copper conductors in an electric cable might be wrapped in a nylon insulation layer, and the cable protected with an outer sheath of plasticized polyvinyl chloride.

Furthermore, the geometry of these products differs sharply from the simple specimens that are of necessity tested in most small-scale flammability apparatus. Surfaces are not always flat, some products are typically covered with other products (e.g., mattresses with bedclothes), and there may be joints where (like or unlike) materials meet. Some products (e.g., wall insulation) would not immediately be exposed to a fire in a room.

It is not yet possible to determine the fire behavior of composite products from knowledge of their constituent materials. To the degree that it is practical, their burning characteristics are determined by conducting full-scale fire tests with the actual products and realistically simulating an ignition source and radiative environment.

No systematic measurements have been completed regarding the ease of flaming ignition (minimum heat or duration of applied heat) of full-scale combustibles such as wall linings, upholstered furniture, and mattresses. Fortunately, ignition studies with a small representative portion of an item can sometimes be conducted [18]. 
Figure 9-18
 shows how a small mock-up (segments of fabric over padding) is used to determine whether smoldering by a lit cigarette will ignite an upholstery fabric [19].

Historically, it was not possible to estimate the rate of heat release of a complex item such as a piece of upholstered furniture, including the period of fire growth, the period of near-maximum heat release, and the burnout period, from small-scale tests. Instead, researchers built large (up to room-scale and beyond) calorimeters to measure the burning rate versus time of individual furnishing items, wall coverings, and other room contents. In these calorimeters, the combustible item is placed on a load cell (to measure the specimen mass), which is located underneath a collection hood connected to an exhaust duct. The exhaust clears the combustion products from the laboratory and, as will be discussed shortly, provides a location for certain needed measurements.

Figure 9-18 Mock-up design for cigarette ignition resistance testing of fabrics for use in upholstered chairs.

Reproduced from: NFPA 260, Standard Methods of Test and Classification System for Cigarette Ignition Resistance of Components of Upholstered Furniture, National Fire Protection Association, Quincy, MA, 2013.

The earliest versions of these calorimeters estimated the heat release rate by weighing the combustible item during the test and multiplying the mass loss rate by the heat of combustion of the item. This technique had several serious drawbacks. First, obtaining the heat of combustion for an item such as a chair is not straightforward. To do so, the tester needs to separate the chair into its component materials, weigh each material, determine its individual heat of combustion, and then calculate a weighted sum of all of these data. Second, for a product that consists of an assembly of materials, the individual materials burn at different rates during a test. Because each material also has a different heat of combustion, the instantaneous mass loss cannot be correlated with the heat-release rate at that time. Third, any unburned residue almost certainly has a chemistry, and thus a heat of combustion, that is different from that of the original product. Fourth, the mass loss rate gives no information about the portion of the combustible item that vaporizes but fails to burn completely. Finally, if the combustible includes a chemical element that burns to form a solid oxide (e.g., aluminum, boron, silicon), then a mass gain rather than a mass loss might result, which would confuse the interpretation of the data.

The next stage in the evolution of large-scale calorimeters involved measurement of the volumetric flow and temperature rise (relative to ambient temperature) of the gas in the exhaust duct, followed by calculation of the convective component of the heat-release rate. Simultaneously, measurements were made of the radiative component of the heat-release rate, with one or more radiometers viewing the burning object from the side. This procedure required the assumption that the radiation was isotropic—that is, it was emitted equally in all directions. Some of the convective heat was lost to the surroundings. Moreover, for tests conducted in a room, the walls and ceiling of the test room could become sufficiently hot that their radiation could not be discounted. A time constant problem also became apparent: the radiative heat reached the radiometers nearly instantaneously, but the temperature was measured far enough into the exhaust duct that the exhaust gases were well mixed. As a consequence, there was a time delay relative to the radiation measurement. This time delay varied, because the gas velocity in the duct changed with temperature.

These uncertainties in the heat release rate determinations were greatly reduced with the advent of 
oxygen consumption calorimetry
. This technique is based on the underlying principle that for almost all common combustibles, and for mixtures of combustibles, approximately 13.1 kJ of heat is released per gram of oxygen consumed [20]. This value varies only approximately ±5 percent for different materials, and the rule holds even if incomplete combustion occurs [21]. Thus the basic required measurements in oxygen consumption calorimetry comprise continuous gas flow and temperature in the duct and the oxygen concentration in that flow. 
Figure 9-19
 shows such an example of such a calorimeter. These devices now exist in fire research and test laboratories around the world. Depending on the design, they can measure fires with heat release rates up to 20 MW.

Figure 9-19 Example of a large fire calorimeter.

The advent of such calorimeters has made possible the regulation of furnishings based on their heat release rates. California requires that upholstered furniture used in high-risk occupancies have a peak heat release rate of less than 80 kW [22]. The U.S. Consumer Product Safety Commission requires that all mattress/foundation sets have peak heat release rates less than 200 kW during the first 30 minutes of burning [23].

Table 9-3
 shows peak rates of heat release for a variety of full-size objects burned under calorimeters like those described previously, as well as some indication of the duration of intense burning. Reference [24] provides additional data on heat release rates.

It is still not possible to predict the ignitability and burning behavior of all combustibles from knowledge of their chemistry or using data from small-scale tests, with one exception: researchers can estimate heat-release rates for upholstered furniture. Due to the importance of these furnishings to fire hazard and risk, extensive research has led to empirical methods for estimating the peak heat-release rate, the time to this peak, and the total heat released. The source of the burning data for these estimations is the cone calorimeter [25].

Acid–Base Pairs

Alkaline substances, also known as caustics or bases, will react with any of the numerous organic and inorganic acids, or even with water, releasing substantial amounts of heat, which could cause other nearby materials to ignite. The most common alkalis are sodium hydroxide (lye) and calcium oxide (lime or quicklime). When wet, these alkalis can generate hydrogen when in contact with aluminum or galvanized steel (zinc). The carbides of lithium, sodium, potassium, calcium, and barium react with water to form acetylene, a highly flammable compound.

Table 9-3 Typical Heat-Release Rates for Various Combustible Items

 

 

300

 

Item	Peak Rate of Heat Release (MW)	Approximate Time of Intense Burning (s)
Mattress/foundation (U.S. twin, pre-2006)	> 2*
Mattress/foundation (U.S. twin, compliant with 16 CFR 1633)	≈ 0.1 (estimated)
Cathode ray tube television, FR plastic cabinet (U.S.)	0.2		300
Upholstered chair (PU foam, polyolefin fabric)	2.0	100
Upholstered chair (California TB133 compliant)	< 0.08
Stack of six steel-frame, polypropylene chairs	1.9
Wooden dresser	1.8	200
Curtains (two, 2.13 m × 1.25 m)	0.13 to 1.2
Automobile	1.2 to 8.3	1000
School bus	29 to 34	5000

Source: Reprinted with permission from NFPA 260, Standard Methods of Test and Classification System for Cigarette Ignition Resistance of Components of Upholstered Furniture, National Fire Protection Association, Quincy, MA, 2013.

Metals

The combustion of metals remains rare in residential, office, or vehicle fires. However, extremely hazardous fires involving metals can occur in some commercial, warfare, and pyrotechnic settings. Contributing factors to metal-based fires include a large metal surface area and potent ignition sources.

The oxidation of a metal is a highly exothermic process. The heats of combustion approximate the values for many synthetic polymers. Thus, in a thermodynamically controlled world, metals would make significant contributions to fires and fire losses. In the real world, however, certain processes reduce these contributions.

The thermal conductivity of metals is very high, about 1000 times that of the most common organic polymers. Thus, it is almost impossible to heat the surface of a massive piece of metal to a high temperature without simultaneously heating the entire mass to nearly the same temperature. Accordingly, massive pieces of metal rarely burn, except after being heated at high intensity or for a long period.

However, if a metal were in the form of a fine powder and suspended in air, even a small flame or spark could heat the nearby particles, and the dissipation of heat away from the surface would be minimal, as would conduction to other particles. Indeed, most metal powders are combustible. In certain cases, the metal need not be as fine as a powder to burn, but rather could take the form of chips or shavings, as from machining.

Metals can be divided into two categories: those that burn on their surface and those that burn in the vapor phase. 

Table 9-4

 presents examples of both types.

The first four elements listed in Table 9-4 are believed to burn on the surface. All have extremely high melting points (TMP) and boiling points (TBP)—higher than their flame temperatures, in fact. The flame temperature of a metal is limited by the boiling point of its oxide because some of the heat of metal oxidation is spent on vaporizing the oxide rather than raising the temperature of the entire combustion system. In fact, because the oxides have very high heats of vaporization, the flames generate only enough heat for partial vaporization of the oxides.

Table 9-4 Metals That Burn on Their Surface and Metals That Burn in the Vapor Phase [27]

* Boron and silicon are actually metalloids, with properties between those of metals and nonmetals.
Republished with permission of Taylor and Francis Group LLC Books, from Handbook of Chemistry and Physics, 92th ed., Haynes, W.M., ed., Section 4, CRC Press, Boca Raton, FL, 2011; permission conveyed through Copyright Clearance Center, Inc. The Handbook is updated annually. For further information, see 

http://www.hbcpnetbase.com

.

The last six metals listed in Table 9-4 vaporize while burning. As a general rule, materials that burn on the surface burn more slowly than materials that can vaporize first because the vapors can more readily contact the surrounding oxygen in the air. The surfaces can also become coated with the metal oxide, reducing oxygen contact with the elemental metal.

Another general characteristic of metals—especially hot metals—is that most of them can react rapidly and exothermically with water to form hydrogen. For example,

2Al + 3H2Oliq → Al2O3 + 3H2 + 819 kJ

Consequently, if a partially wet metal is burning, fire fighters must deal with a hydrogen fire as well as a metal fire. Also, water under the surface of hot molten metal will turn rapidly into

steam

and erupt, throwing molten metal around.

Some details about the combustion of specific metals follow [28]. Information regarding hazardous metals and a wide variety of other materials can be found in Reference [29].

Magnesium in the form of powder, ribbons, or shavings can ignite under some conditions at about 932 °F (500 °C). A massive piece of magnesium must be heated to its melting point (1202 °F; 650 °C) for ignition to occur, as has happened in military combat operations. Some magnesium alloys have lower ignition temperatures. Magnesium chips wet with animal or vegetable oils have been known to ignite spontaneously. Molten magnesium in contact with iron oxide (rusted iron) produces a highly energetic thermite reaction3:

3 Mg + Fe2O3 → 3 MgO + 2 Fe + 971 kJ

When finely divided, magnesium burns in an atmosphere of any of the following pure gases, to yield the products shown:

 

steam	MgO + H2
carbon dioxide	MgO + CO
nitrogen	Mg3N2
halon 1301 (CF3Br)	MgF2, MgFBr, MgBr2, C (This reaction releases more heat than magnesium powder burning in oxygen.)

Magnesium powder mixed with polytetrafluoroethylene, (C2F4)n, even under an inert gas such as helium or argon, will burn to form MgF2 and carbon.

Clearly, the choice of suppressants when fighting a magnesium fire is severely limited. Extinguishing agents are discussed in the Fire Fighting Chemicals chapter.

Aluminum has a much higher boiling point than magnesium, so it cannot be ignited as readily. However, aluminum powder, flakes, very fine chips, and shavings can be ignited and, once ignited, will burn like magnesium chips or powder. Aluminum powder is a major ingredient in solid propellant rocket fuels, and it burns rapidly in rockets to give a very high exhaust temperature. Aluminum powder also reacts violently with halogenated liquids, including common solvents (e.g., trichloroethylene, dry cleaning fluid) and fire suppressants.

Aluminum reacts with iron oxide in a thermite reaction producing, for example, temperatures high enough to melt steel. The relatively low melting point of aluminum favors melting in a fire; consequently, fire investigators use it as a temperature marker. Aluminum will not burn in nitrogen.

Iron and steel generally do not burn in air, but can burn in pure oxygen. Fine steel wool or steel dust in air can be ignited with a torch. Pure iron powder, when exposed to air for the first time after manufacture, can ignite spontaneously.

Massive pieces of titanium generally cannot be ignited. In spite of this element’s high melting point and high boiling point, fine turnings and thin chips of titanium can be ignited with a match. Once ignited, vigorous burning results. Titanium dust clouds can ignite in air when heated. Like magnesium, titanium dust will burn in pure carbon dioxide or pure nitrogen.

The alkali metals—lithium, sodium, and potassium—have some unusual fire properties. These metals have low melting points (366 °F, 208 °F, and 144 °F, [86 °C, 98 °C, and 62 °C], respectively). Sodium and potassium, on contact with water at room temperature, generate hydrogen exothermically and burst into flame spontaneously. Lithium reacts more slowly with water, without bursting into flame. These alkali metals can be ignited by heating in dry air. Once ignited, they burn vigorously, producing dense white clouds of metal oxide particles. Lithium differs from sodium and potassium in that it can burn in pure nitrogen. Alkali metals often are stored under kerosene or oil to isolate them from air and moisture. Alkali metals react violently with halogenated hydrocarbons or sulfuric acid.

NaK is a sodium–potassium alloy with a very low melting point; it is a liquid in the vicinity of room temperature. NaK is used as a heat-transfer fluid. Its fire properties resemble those of sodium and potassium, but its reactions are more vigorous. Reference [27] discusses the fire properties of zirconium, calcium, zinc, uranium, and plutonium.

Exothermic Materials

An 
exothermic material
 comprises either a pure substance or a mixture of substances that can undergo chemical reactions that liberate heat without requiring oxygen from the air. Such a material poses a fire or explosion hazard. Moreover, if such a material becomes involved in a fire, it usually will cause easier ignition, faster fire growth, higher flame temperatures, and more difficult extinguishment. In some cases, it enables propagating combustion even when the oxygen fraction is too low for other combustibles to burn.

Numerous such materials can be found among industrial chemicals, in multiple forms—gases, liquids, and solids.4 The strategy for fighting a fire involving an exothermic material is specific to the material and depends on the quantity involved, likelihood of explosion, toxicity, solubility in water, and compatibility with extinguishing agents, among other factors. References [30] and [31] provide guidelines for dealing with many exothermic materials.

Exothermic compounds release heat through four processes:

1.  Thermal instability. These materials decompose into smaller molecules while releasing heat. Generally, the number of moles of gas released is larger than the number of moles of the source compound. Combined with the heat release, this volumetric increase can result in a severe pressure wave. Examples of materials that exhibit thermal instability include ozone (O3), nitrous oxide (N2O), acetylene (C2H2), hydrogen peroxide (H2O2), hydrazine (N2H4), ethylene oxide (C2H4O), lead azide (Pb(N3)2), diborane (B2H6), and methyl hydrazine (N2H3CH3).

In addition, ethers have the capability of forming peroxides after storage for a month or more. Isopropyl ether (C3H7—O—C3H7) is more susceptible to peroxide formation than other ethers; detonation can occur when this peroxide is heated. Many other peroxides, such as the widely used benzoyl peroxide, (C6H5)—(CO)—O—O—(CO)—(C6H5), can undergo rapid exothermic decomposition.

2.  Self-polymerization. When organic monomers combine to form polymers, the polymerization process releases heat. In controlled polymerization, enthalpy decreases as heat is released, generally by the circulation of a coolant. In uncontrolled polymerization, heat is not removed, and excessive temperatures result. (Examples of polymeric reactions were given earlier in this chapter.)

3.  Intramolecular oxidation–reduction. Some molecules contain a group of atoms with oxidizing power, such as a nitro group (—NO2) or a peroxide group (—O—O—), and a group of carbon and hydrogen atoms that can be oxidized. Such molecules often are used as explosives or rocket propellants. 
Figure 9-20
 shows three examples and the products of their complete combustion. Like the thermally unstable compounds, each of these molecules reacts to form additional moles of gas as well as heat.

4.  Oxidizing agent in contact with a reducing agent. Unlike the previously mentioned compounds, these oxidizing agents (e.g., potassium nitrate, sodium chlorate, and fluorine) are not capable of exothermic decomposition. They are mixed with an oxidizable material (reducing agent), which can consist of almost any organic material or metal. These combinations are used as explosives, pyrotechnics, and solid propellants for rockets. Examples include:

•   Black powder: KNO3 + 3/2 C + 1/2 S → CO2, CO, N2, K2S, K2SO4, K2CO3, S

•   Explosive: n NH4NO3+ 1/3 (CH2)n(fuel oil) → n [N2 + 7/3 H2O + 1/3 CO2]

•   Pyrotechnic delay mixture: BaCrO4 + B → BaO + 1/2 B2O2 + 1/2 Cr2O3

•   Solid propellant: n(NH4ClO4 + Al) + 1.4 (CH2)n → N2, H2, H2O, HCl, Al2O3, CO, CO2

Such mixtures are prepared with extreme care and presumably are stored in bunkers or isolated places.

A more commonly encountered problem relates to oxidizers that have not been deliberately mixed with anything, but might become mixed by accident, such as in the course of a fire. Any of these oxidizers, when it comes in contact with common materials such as paper, cotton, wood, plastics, hydrocarbons, alcohols, vegetable and animal oils and fats, sulfur, and metals, can react exothermically even in the absence of air, although in some cases an elevated temperature would be needed to initiate reaction. The more common classes of oxidizers include the following:

•   Nitric acid and its salts (nitrates)

•   Perchloric acid and its salts (perchlorates)

•   Chromic acid and its salts (chromates and dichromates)

•   Permanganic acid and its salts (permanganates)

•   Fluorine, chlorine, bromine, and iodine (in order of decreasing reactivity)

•   Inorganic peroxides

•   Calcium hypochlorite

•   Potassium persulfate

•   Manganese dioxide

•   Chlorate, bromate, and iodate salts

Figure 9-20 Examples of molecules that undergo internal oxidation–reduction reactions.

WRAP-UP

Chapter Summary

•   The pyrolysis of a solid to form gaseous fragments can be characterized by a heat of gasification. During pyrolysis, chemical change occurs within the solid, and the chemistry of the volatiles is rarely the same as the chemistry of the solid.

•   The ignition of the pyrolyzate from a solid is similar to the ignition of gases and vapors, except that heat losses from the fuel surface to the interior and to the surroundings can delay or prevent the ignition.

•   For a fuel to smolder or self-heat, it must be porous, its interior surfaces must react exothermically reaction with oxygen, and the material must be a good insulator.

•   Surface char formation slows burning by insulating the subsurface material from the flame radiation and by sequestering carbon.

•   For solids that melt during burning, dripping or running of the liquid increases the fire spread rate due to the increase in burning surface. The flaming liquid can also ignite other combustibles on contact.

•   The burning rate depends on the radiant flux to the surface. Early in a fire, this flux comes from the local flames; later, most of the flux is derived from the hot surroundings.

•   The flame spread rate depends on whether the spread direction is upward (fastest), lateral (slow), or downward (slow, except when dripping occurs); the degree of radiative preheating; the thermal thickness of the solid; the ambient oxygen percentage; and the direction of any air flow.

•   Nearly all common solid combustibles are polymers, which consist of strings of identical building blocks. Polymers are classified based on whether they are formed by addition or by condensation; synthesized by nature or by humans; linear or cross-linked; composed of a single building block or multiple building blocks; or thermoplastic, thermoset, or elastomeric.

•   Many formulations of each of the families of plastics exist, as do many different types of woods. The burning properties of these solid combustibles can vary considerably.

•   Fire retardants are added to combustible materials to help them meet flammability criteria. These additives work by promoting char formation, inhibiting the flame chemistry, absorbing heat in the solid, and/or forming a barrier between the flame and the fuel surface.

•   Most common combustible items consist of multiple materials. We cannot yet quantify the fire behavior of a composite from its constituent materials. Instead, the fire performance is measured by burning the entire item or estimated from tests of small specimens of the materials or a representative composite.

•   Oxygen consumption calorimetry is a tool for determining the rate of heat release from a product (or specimens from a product).

•   Exothermic materials and finely divided metals pose unusual and potentially severe fire hazards, but are found almost exclusively in nonresidential settings.

Key Terms

addition polymer A polymer formed without the loss of any atom or molecule.

fire barrier material A protective layer that prevents or severely retards the contribution to a fire from any subsurface material(s).

cellulosic material A material composed entirely or mostly of a polymer of glucose.

composite A product consisting of more than one material, or a material consisting of multiple solids that remain physically distinct.

condensation polymer A polymer that splits out small molecules, usually water, during its formation.

copolymer A polymer made of two or more different monomers.

cross-linked polymer A polymer in which the long chains are bonded to one another at intermediate points.

exothermic material A material that can undergo chemical reaction that releases heat without an additional oxidizer, such as oxygen from the air.

fire retardant A chemical additive that slows the ignition and/or burning rate of a material.

homopolymer A polymer that contains only one type of repeat unit.

limiting oxygen index (LOI) The minimum volume percent of oxygen that will support flaming of a material, as measured in a standard apparatus.

monomer A small molecule that can combine with other molecules of the same kind (or different kinds) to form a repeating chain molecule, or polymer.

oligomer A molecule consisting of a few monomer units.

oxygen consumption calorimetry The determination of heat release rate in combustion using measurement of the depletion of oxygen from the incoming air.

polymer A large molecule consisting of very many repeated units, called monomers.

thermal runaway Self-heating which rapidly accelerates to high temperatures.

thermoplastic polymer A polymer that softens upon heating and returns to its original state upon cooling.

thermoset polymer A polymer that, upon heating, undergoes irreversible change.

vinyl polymer A polymer synthesized from monomers that contain a carbon-carbon double bond.

Challenging Questions

1.   What is the difference between a material and a product?

2.   Why does it generally take more enthalpy to ignite a solid than a gas or liquid?

3.   What is the difference between spontaneous ignition and piloted ignition of a solid?

4.   If a pilot flame is present, how intense must a thermal radiative flux be to ignite wood? How does this flux compare with the intensity of sunlight (approximately 0.7 kW/m² at noon in the tropics)?

5.   Why will a single wooden log, if ignited, soon self-extinguish, while a group of logs near each other continues burning?

6.   How does the formation of a char layer affect the burning of a solid?

7.   Why is upward flame spread over a vertical surface more rapid than downward flame spread?

8.   Why do flames spread more rapidly on thin materials?

9.   The acrylic top of a table is burning, and the flame has spread to the edges. Of the heat transferred from the flame to the table surface (to supply the heat of gasification), which fraction is radiative and what which is convective?

10.   What are the upper and lower extremes of the moisture content of wood?

11.  What do woods, papers, and cottons have in common?

12.   What is the difference between a linear polymer and a cross-linked polymer?

13.   What is the Oxygen Index Test? Why are the results of this test of limited use?

14.   What is the cone calorimeter? How does it capture the effect of a fire environment on a small test specimen?

15.   What are the four ways in which fire retardants can act? Which type(s) of fire retardant would be effective for a material exposed to persistent radiative heat of high intensity, such as from a nearby fire?

16.   Under which conditions can the presence of fire retardants cause problems?

17.   The heat of combustion of aluminum is similar to (actually a little lower than) the heat of combustion of common woods. Why is an aluminum frying pan less of a fire hazard than the overhead wooden kitchen cabinets if the food in the pan ignites?

18.   Why is it not advisable to apply water to burning metals?

19.   The tendency of oily rags to ignite spontaneously depends on the type of oil. Why are many vegetable oils and fish oils far more dangerous than mineral oils or lubricating oils?

20.   Why can certain liquids and gases, such as ethylene oxide or acetylene, burn in the complete absence of oxygen?

21.   Is it true that any artificial atmosphere that will support life will also support combustion of common materials?

References

1.   Babrauskas, V. (2003). Ignition Handbook, Issaquah, WA: Fire Science Publishers,.

2.   Moussa, N. A., T. Y. Toong, and S. Backer. (1973). “An Experimental Investigation of Flame-Spreading Mechanisms Over Textile Materials.” Combustion Science & Technology 8: 165–175.

3.   Bowes, P. C. (1984). Self-Heating: Evaluating and Controlling the Hazards. New York, NY: Elsevier.

4.   Cote, A. E., ed. (2008). Tables and Charts: Chapter 6.17. In: Fire Protection Handbook, 20th ed. Quincy, MA: National Fire Protection Association.

5.   Butler, C. P. (1971). Notes on Charring Rates in Wood. Fire Research Note 896. Borehamwood, Herefordshire, UK: British Fire Research Section.

6.   Tewarson, A. (2008). Generation of Heat and Chemical Compounds of Fires. In: SFPE Handbook of Fire Protection Engineering, 4th ed., DiNenno, P. J., et al., eds. Quincy, MA: National Fire Protection Association.

7.   de Ris, J. (1979). Fire Radiation: A Review.” Proceedings of the Combustion Institute 17: 1003–1016.

8.   Unpublished data. Norwood, MA: FM Global.

9.   McAlevy, R. F., and R. S. Magee. (1969). “The Mechanism of Flame Spreading over the Surface of Igniting Condensed Phase Materials.” Proceedings of the Combustion Institute 12: 215–227.

10.   Proceedings of International Microgravity Combustion Workshops. Washington, DC: National Aeronautics and Space Administration.

11.   Lastrina, F. A., R. S. Magee, and F. R. McAlevy. (1971). “Flame Spread over Fuel Beds: Solid Phase Energy Considerations.” Proceedings of the Combustion Institute 13: 935–948

12.   ASTM Fire Standards and Related Materials, 7th ed. (2007). West Conshohocken, PA: ASTM International.

13.   Magee, R. S., and R. D. Reitz (1975). “Extinguishment of Radiation-Augmented Plastic Fires by Water Sprays.” Proceedings of the Combustion Institute 15: 337–347.

14.   ASTM D 2863: Standard Test Method for Measuring the Minimum Oxygen Concentration to Support Candle-like Combustion of Plastics (Oxygen Index). (1991). West Conshohocken, PA: ASTM International.

15.   Kroschwitz, J. I., ed. (2007). Kirk-Othmer Encyclopedia of Chemical Technology, 5th ed. New York, NY: John Wiley.

16.   Lyons, J. W. (1970). The Chemistry and Uses of Fire Retardants. New York, NY: Wiley-Interscience.

17.   Wilkie, C. A., and A. B. Morgan, eds. (2009). Fire Retardancy of Polymeric Materials. Boca Raton, FL: CRC Press.

18.   Gann, R. G., R. H. Harris, Jr., J. F. Krasny, R. S. Levine, H. Mitler, and T. J. Ohlemiller. (1988). Effect of Cigarette Characteristics on the Ignition of Soft Furnishings. NBS Technical Note 1241. Gaithersburg, MD: National Bureau of Standards.

19.   NFPA 260: Standard Methods of Test and Classification System for Cigarette Ignition Resistance of Components of Upholstered Furniture. (2013). Quincy, MA: National Fire Protection Association.

20.   Huggett, C. (1980). “Estimation of the Rate of Heat Release by Means of Oxygen Consumption Measurements.” Fire and Materials 4: 61–65.

21.   Krause, R. F., and R. G. Gann. (1980). “Rate of Heat Release Measurements Using Oxygen Consumption.” Journal of Fire and Flammability 11: 117–130.

22.   Flammability Test Procedure for Seating Furniture for Use in Public Occupancies. Technical Bulletin 133. (1991). State of California, Bureau of Home Furnishings and Thermal Insulation.

23.   Standard for the Flammability (Open Flame) of Mattress Sets. (2006). 16 CFR Part 1633, U.S. Consumer Product Safety Commission.

24.   Babrauskas, V. (2008). Heat Release Rates. In: SFPE Handbook of Fire Protection Engineering, 4th ed., DiNenno, P. J., et al., eds. Quincy, MA: National Fire Protection Association.

25.   Babrauskas, V. (2008). Upholstered Furniture and Mattresses. In: Fire Protection Handbook, 20th ed., Cote, A. E., ed. Quincy, MA: National Fire Protection Association.

26.   Ohlemiller, T. J., and R. G. Gann. (2002). Estimating Reduced Fire Risk Resulting from an Improved Mattress Flammability Standard. Technical Note 1446. Gaithersburg, MD: National Institute of Standards and Technology.

27.   Haynes, W. M., ed. (2011). Handbook of Chemistry and Physics, 92th ed. Boca Raton, FL: CRC Press, Section 4. The Handbook is updated annually. For further information, see 

http://www.hbcpnetbase.com/

.

28.   Christman, T. (2008). Metals. In: Fire Protection Handbook, 20th ed., Cote, A. E., ed. Quincy, MA: National Fire Protection Association.

29.   Baker, C. J. (2006). Fire Fighter’s Handbook of Hazardous Materials, 7th ed. Sudbury, MA: Jones and Bartlett.

30.   Fire Protection Guide to Hazardous Materials, 11th ed. (2010). Quincy, MA: National Fire Protection Association.

31.   Lewis, R. J. (2012). Sax’s Dangerous Properties of Industrial Materials. New York, NY: Wiley.

 

1 There are many formulations in each of these families of plastics, as well as many different types of woods. The burning properties of these formulations can vary considerably. For this reason, it is important to obtain flammability data for the specific formulation of interest. Estimating the properties of one formulation from data for other formulations requires knowledge of the degradation mechanisms of the formulations and is beyond the scope of this book.

2 In this, and other, tables of LOI data, many of the entries consist of a single number for an entire class of polymers, such as polyethylenes. There are many formulations of polyethylene, so the compiled values should be used with caution.

3 In a thermite reaction, a metal and an oxide of a different metal react exothermically, with the oxygen “changing partners.”

CHAPTER
8

Fire Characteristics: Liquid Combustibles

OBJECTIVES

After studying this chapter, you should be able to:

•

 

  Describe the flash point, fire point, and autoignition temperature of a flammable liquid.

•   List the three classes of flammable liquids, based on flash point and potential ambient temperatures.

•   Define the linear burning rate of a pool of liquid and explain why it varies with the diameter of the pool.

•   Describe the physical considerations that affect the rate of flame spread of flammable liquids.

•   Explain boilover.

•   Explain a boiling liquid/expanding vapor explosion (BLEVE).

Introduction

How many movies have you seen in which a lit cigarette ignites a

gasoline

spill? In reality, this is not a very likely occurrence. A cigarette is a small ignition source, and the spill is a large heat sink. Most likely, the gasoline will cool and quench the cigarette before the cigarette heats the gasoline—but that rather simple outcome would not serve the filmmaker’s intent to have the dramatic effect of flames destroying a car, a house, or a service station.

Liquids do ignite and burn, sometimes quite violently. The chemistry of a burning liquid is the chemistry of its vapor. If a compartment contains an ignitable volume fraction of n-hexane, the chemistry does not “know” whether the fuel entered the compartment as a gas or whether a small amount of spilled liquid vaporized. Thus, the combustion chemistry of burning gases discussed in the previous chapter provides a sufficient basis for understanding the chemistry of burning liquids. However, physical considerations also affect the ignitability and rate of flame spread of flammable liquids as well as the tactics used to limit these hazards. These are the topics covered in this chapter.

Ignition of Liquids: Flash Point, Fire Point, and Autoignition Temperature

It is the vapor of a liquid that burns; therefore, the principal property of a flammable liquid that affects its susceptibility to ignition is the ease with which the molecules vaporize to form a gaseous fuel–air mixture that is within the liquid’s flammability limits.

Placing a match just above a pool of a flammable liquid will not lead to ignition unless the vapor concentration exceeds the lower flammable limit of that vapor in air. In the Physical and Chemical Change chapter, we saw that at 32 °F (

0

°C), the vapor pressure of

m

ethanol

is 3.96 kPa (3.92 atm). The total pressure is 101 kPa (1 atm), so the volume percent, or mole percent, of methanol vapor in the air just above the liquid surface is

100

· 3.96/101 = 3.9 percent by volume. In the Fire Characteristics: Gaseous Combustibles chapter, we saw that the lower limit of flammability of methanol vapor in air is 6.7 percent by volume. Therefore, methanol should not be, and is not, flammable at 32 °F (0 °C).

The vapor pressures of liquids, however, increase sharply with increasing temperature. If a match flame were held next to a small enough quantity of methanol liquid for a long enough time (and if the match did not burn out), the liquid would be heated to

>54

°F (12 °C). At this temperature, the vapor pressure of methanol is 7.17 kPa, and the percent by volume is 7.1 percent. This exceeds the lean flammability limit, so the liquid would ignite.

One of the standard tests for ignition of liquids, ASTM D92 [1], replicates this behavior. A small, open cup of cold methanol is heated gradually from below. A small flame from a tiny burner is passed across the liquid surface every 10 seconds. When the liquid reaches about 54 °F (12 °C), a flame moves rapidly across the surface, consuming the methanol vapor above the surface. After a fraction of a second, no further combustion would occur, because the combustible vapor has been consumed, and the heat transfer from the small flame is too little to overcome the evaporative cooling of the liquid surface and sustain the vaporization. By the time that additional vaporization can restore the original vapor concentration, the burner has been removed, and an ignition source is no longer present. Ten seconds later, the burner is passed over the liquid again, and the sequence repeats itself. The minimum temperature at which this behavior occurs (54 °F, 12 °C for methanol) is called the flash point of the liquid.

When the methanol is heated further (usually 10 °F to 30 °F, or 5 °C to 15 °C higher than the flash point), and the ignition flame is applied from time to time, combustion is sustained after removal of the ignition source. At this temperature, called the fire point of the liquid, the liquid temperature is high enough to maintain a supply of vapor as fast as it is consumed by the flame.

Multiple tests for flash point and fire point temperatures have been developed, which yield some variation in the measured values. These discrepancies arise because the thermal and flow environment above the liquid depends on four properties:

•   The intensity and size of the ignition source

•   The length of time for which the ignition source is held over the liquid

•   The rate of heating of the liquid

•   The degree of air movement over the liquid

Nevertheless, the measured values, especially the flash point, are widely used as guides to the safe handling of liquids. The flash point, being lower than the fire point, is a more conservative value to use.

Liquids can be divided into classes (which are divided further into subclasses) based on their flash points [2]:

1.  Class I: Liquids with flash points below 100 °F (38 °C) an indoor temperature that could be reached sometime during the year.

2.  Class II: Liquids with flash points at or above 100 °F (38 °C) but below 1

40

°F (60 °C), a temperature that could be reached with only a modest degree of heating.

3.  Class III: Liquids with flash points at or above 140 °F (60 °C), a temperature that would require considerable heating.

Table 8-1 Flash Points of Some Common Liquids

 

Flash Point

*

 

 

°C

°F

Class I Liquids

gasoline

–45.5

–50

ethyl ether (anesthetic)

–28.9

–20

n-hexane

–3.9

25

JP-4 (jet aviation fuel)

†

†

–18

0

acetone

–17.8

0

toluene

4.4

40 methanol

12.2

54 ethanol

12.8

55

turpentine†

35

95

Class II Liquids

No. 2 fuel oil (domestic)

††

>38

>100

diesel fuel†

40 to 55

104 to 131

Jet A (jet aviation fuel)††

47

117

kerosene

37.8

100

No. 5 fuel oil††

>54

>130

Class III Liquids

JP-5 (aviation jet fuel)††

66

151

SAE No. 10 lube oil††

171

340

tricresyl phosphate††

243

469

* Data from Reference [3], except as noted.

† Data from Reference [2]. This value was obtained using a closed-up method, which typically gives flash point values that are 5 °C to 10 °C lower than open-cup values.

†† Data from Reference [4]. These data were obtained using a closed-up method, which typically gives flash point values 5 °C to 10 °C lower than open-cup values.

Table 8-1

 lists flash points for some common liquids. Notice the wide range, from –50 °F to 469 °F (–45.5 °C to +243 °C). Note also that these values are meaningful only for a bulk liquid. If a liquid with even a high flash point is formulated as a spray or a foam, is released with air present, and comes into contact with even a small ignition flame, the tiny amount of liquid in contact will immediately be heated to a temperature higher than its flash point and will start burning. The combustion enthalpy released will vaporize the surrounding spray or foam, and the fire will propagate (spread).

As an example of the fire potential of liquids in these three classes, consider a liquid spill on a summer day when the ground has been heated by the sun to 95 °F (35 °C). If the spill consisted of n-hexane, a Class I liquid, there would be a race between the wind dispersing this volatile chemical and the introduction of an ignition source. On a still day, the vapor would ignite; on a very windy day, it probably would not. If the spill consisted of kerosene, a Class II liquid, a fire hazard would exist only if the liquid was exposed to an additional heat source capable of raising the temperature of some part of the liquid by at least 25 °F (14 °C). JP-5, a Class III liquid, was designed to burn well in jet engines, but poses a significantly lower ignition risk during handling and storage than JP-4, a Class I liquid.

Fire points and flash points depend heavily on pressure. The flash point is the temperature at which the vapor pressure of the liquid equals the lower flammability limit. If the atmospheric pressure were 50 kPa instead of 100 kPa, the vapor pressure of the liquid need be only one-half as great to achieve the lower flammable limit. Therefore, flash points and fire points are lower than normal at pressures below atmospheric pressure and higher than normal at pressures above atmospheric pressure. This variation is very important in assessing the flammability in, for example, the vapor space of an aircraft fuel tank. It should also be considered in cities at higher elevations, such as Denver and Albuquerque, where the average atmospheric pressure is approximately 80 kPa (0.8 atm).

Autoignition, in which a vapor–air mixture is ignited strictly by heating, was discussed in the Fire Characteristics: Gaseous Combustibles chapter. Autoignition temperatures are typically hundreds of kelvins higher than flash points or fire points.

Burning Rates of Liquid Pools

Once a pool of a flammable liquid is ignited, the flames generally spread to cover the full surface area of the pool. The liquid will then burn at a more or less steady rate until the liquid is consumed. (Some very-low-volatility, high-fire-point fluids burn locally, perhaps with small, irregularly moving flames. There is some discussion of these later in this chapter.)

The rate of burning of a liquid pool is often expressed as a linear burning rate (in mm/s)—that is, the rate at which the surface of the pool recedes. The following discussion assumes that the liquid is sufficiently deep that steady burning can be established. Treatment of shallower pools is beyond the scope of this book.

The linear burning rate is readily converted to a linear mass burning rate (in kg/m²-s) by multiplying the linear burning rate by the density of the liquid (in kg/m³) and dividing the product by 1000 (to yield mm/m). Then, obtaining the heat of combustion of the liquid (J/g) from a handbook, the rate of heat release per unit area can be calculated, assuming complete combustion. The total heat release rate for the pool is the rate of heat release per unit area multiplied by the surface area of the pool.

To illustrate the magnitude of these rates, a pool of gasoline 3.3 ft (1 m) in diameter and 25 mm (1 in.) deep will be consumed in approximately 4 minutes. The average linear burning rate is 25 mm/240 s or about 0.1 mm/s. Gasoline (and many hydrocarbons of similar molecular weight) has a density of approximately 700 kg/m3 and a heat of combustion of approximately 45 MJ/kg. The surface area of the pool is πd2/4 =0.79 m2. Using these values, the mass burning rate of the pool can be calculated as about 0.06 kg/s. The average rate of heat release for this pool fire would be about 2.5 MJ/s or 2500 kW.

The linear burning rate of a pool of liquid depends not only on the nature of the liquid but also on the diameter of the pool. 
Figure 8-1
 shows the linear burning rate of n-hexane as a function of pool diameter from about 0.2 m to 2.5 m.

Figure 8-1 Effect of pool diameter on the linear burning rate of n-hexane [4].

Note

The burning rate of a pool smaller than 0.2 m in diameter is of less interest in fire protection, as its hazard would be limited to its role as an ignition source. For fires of such small diameter, the flames heat the lip of the container, and the hot container heats the adjacent liquid, increasing the local evaporation rate and, therefore, the overall burning rate. As the diameter increases, the central area of the pool becomes larger relative to the area of the ring near the lip, the edge effect becomes less important, and the linear burning rate decreases.

Texture: Eky Studio/ShutterStock, Inc.; Steel: © 

Sharpshot/Dreamstime.com

For pools larger than 0.2 m in diameter, the burning rate increases with increasing pool diameter, reaching a limit at a diameter of approximately 3 m. The reason for this increase becomes apparent when we consider the three factors that control the burning rate of a liquid pool.

First, little of the flame radiation is needed to vaporize enough liquid fuel to sustain the fire. Each gram of n-hexane that burns releases 44,860 J. The rate of burning of the n-hexane is controlled by its rate of vaporization. To vaporize n-hexane, the latent heat of vaporization—that is, 371 J/g—must be supplied. Therefore, a little less than 1 percent of the combustion energy must return to the n-hexane surface, through the rising vapors, to maintain the vapor supply to the flame. (The bulk of the heat is convected upward with the combustion products and radiated sideways and upward.)

Second, the flame over a pool that is less than 0.2 m in diameter is a laminar diffusion flame. Its combustion is relatively complete, so little soot forms. The flame is optically thin, the emissivity of the flame is low, and the radiant intensity to the fuel surface is low. This small amount of radiation, combined with some conduction from the lip, suffices to vaporize enough liquid to keep the flame burning at a low level.

Third, the flame radiation to the surface increases with increasing pool diameter, up to a limit. At these larger diameters, the flames evolve from laminar to increasingly turbulent in nature. As this happens, more soot forms, and the optical thickness of the flame increases. 

Figure 8-2

 shows how the thermal radiation intensity increases with increasing sootiness.

The lowest curve in Figure 8-2 simulates an optically thin flame. As more soot is “added” to the flame, the radiant intensity increases. The “average” soot curve reaches a limiting value at about 2 m physical thickness. At this point, the flame is optically thick, and a further increase in the physical thickness will not increase the radiant output. The same is true for adding more soot (as in the “high soot” flame, which has eight times the soot of the “very low soot” flame)—this extra soot simply ensures that the “high soot” flame reaches the radiative limit at a smaller flame diameter.

Linear burning rates have been measured for other liquids [4] and have been extrapolated to limiting values for large pool sizes. For these pools, the flames are optically thick, so the energy radiated to the surface is constant. Thus one might expect that the linear burning rate would depend on how easy it is to vaporize the liquid. 
Figure 8-3
 demonstrates this relationship. For these five liquids, the linear burning rate correlates with the reciprocal of the latent heat of vaporization per unit volume (the product of the latent heat of vaporization per gram and the density). The slope of the correlation line is about 3 J/s•mm² (30 kW/m²). Thus the radiation from optically thick flames of any of the five combustibles imposes an average heat flux of approximately 30 kW/m² on the liquid surface, regardless of the chemical nature of the combustible.

Figure 8-2 Calculated radiative intensity coming from a hot, semi-transparent, sooty gas of various thicknesses, for three soot levels.

Figure 8-3 Burning rates of large liquid pools versus inverse of product of latent heat of vaporization (L) and density (p).

Flame Spread Rates over Liquid Surfaces

The previous section dealt with burning rates of burning liquid pools, where the flame covers the entire surface. In these scenarios, the spread over the surface is fast compared to the surface regression rate. This section addresses situations in which the rate of spread of the flame over the surface, after a local region of the surface has been ignited, is important.

Figure 8-4
 shows data for the flame spread rate over the surface of a liquid, n-butanol (C4H9OH). (Note the logarithmic scale on the vertical axis.) This compound has a flash point of 110 °F (43 °C), as measured by the open-cup method described earlier. Above this temperature, the flame spread rate is 2 m/s (6.5 ft/s) and is independent of the liquid temperature. Below 110 °F (43 °C), the flame spread rate heavily depends on the liquid temperature. Indeed, at 68 °F (20 °C), the flame spread rate is only 1/100 of the value at 122 °F (50 °C).

Figure 8-4 Rate of flame spread over the surface of n-butanol [5].

This behavior is typical of flammable fluids. A flame will spread, albeit very slowly, over a liquid whose temperature is well below its flash point. To start the burning, the liquid must be heated locally to above its flash point. Then, if the radiative heat transfer from the flame is sufficient to heat the adjacent cold liquid (and induce convection currents in it), the flame will spread.

For a fluid whose temperature is well above its flash point, a combustible vapor concentration in excess of the lower flammability limit exists over the surface before the arrival of the flame, and the flame rapidly covers the entire surface. If the liquid is too warm, the vapor concentration just over the surface will exceed the upper flammability limit. In that case, air circulation over the liquid pool will decrease the combustible concentration with increasing height above the surface, and somewhere there will be a zone containing a near-stoichiometric mixture.

The direction and velocity of any air flow can have a large impact on flame spread over a liquid pool. A (co-flow) breeze in the same direction as the flame spread will increase the spread downwind. A (counterflow) breeze opposite to the flame spread direction will decrease and perhaps even halt the flame spread.

Hazards of Liquid Fuel Fires

With this understanding of the characteristics of fires of liquid fuels, it is helpful to identify five categories of liquid fires that constitute nearly all the encountered configurations.

1.  A pool of liquid, such as an open tank or the result of a spill. If the temperature of the liquid exceeds its fire point, it can be ignited and will sustain flaming. Given the proper equipment, fires on stagnant liquids are routinely extinguished using techniques such as those discussed in the Fire Fighting Chemicals chapter.

A serious outcome can result from boilover from a pool fire. The key condition for boilover is the presence of two immiscible layers:

•   An upper layer of a low density combustible liquid

•   A lower layer of a higher-density fluid whose boiling point can be exceeded due to heat transfer from the upper fluid layer

Although this phenomenon has been recognized for more than a century, boilover was not understood fully until about 25 years ago [6].

The classic case of boilover begins with a burning pool of hydrocarbon fuel. This liquid source could be as small as a deep-fat fryer or as large as a petroleum storage tank at an oil refinery. Hydrocarbon fuels consist of a mixture of highly volatile and slightly volatile components. A heated sample would start to burn at a fuel temperature below 212 °F (100 °C). As the fire burns in the open container, the more volatile components are driven out of the topmost few millimeters of liquid, and the temperature of this “slice” rises to approximately 572 °F (300 °C). Heat is conducted downward through the liquid to the next few millimeters, causing gasification of volatile components in that slice. The volatiles form bubbles below the surface. The motion of these bubbles greatly accelerates the mixing of the hotter upper fuel and the cooler lower fuel. Enhanced by this bubble-induced mixing, a hot zone spreads downward through the fuel.

Seeing the burning fuel, a nearby person applies water to quench the flames. The water, being of higher density than the hydrocarbon fuel, rapidly sinks to the bottom of the container. Now there are two layers that match the description given earlier: fuel on top and water on the bottom.

As the water sinks, it encounters hot oil at temperatures well above 212 °F (100 °C), and the subsurface water starts to boil vigorously. The volume of 1 kg of hot water vapor is more than 1000 times the volume of 1 kg of liquid water. Under the best of circumstances, the expanding water pushes up on the burning oil, which then overflows the container and spreads the fire. More seriously, the rapidly expanding water can propel blobs of the hot, burning oil out of the container. These airborne blobs have more surface area than the original pool surface, so the rate of oil combustion increases, heating the air surrounding the blobs and further accelerating their dispersion. The outcome is a rapid expansion of the fire outside the pan or tank. The consequences for nearby people or combustibles can be disastrous.

2.  A flowing liquid, such as from an overflowing or rapidly leaking tank. As in the first category, if the temperature of the liquid exceeds its fire point, the material can be ignited. Flowing liquid fires are very difficult, and sometimes impossible, to extinguish as long as the liquid continues to flow and the flames continue to move.

3.  A spray from a small orifice at high pressure (e.g., from a leaking hydraulic fluid line). For fires involving liquids in the form of sprays, the fire point is not a relevant measure of flammability. A pool of domestic (No. 2) fuel oil, a Class II fuel, at 68 °F (20 °C) cannot be ignited with a match (unless a wick is present). However, when the same oil in the form of a spray or foam comes into contact with even a very small ignition flame, the tiny amount of liquid in contact will immediately be heated to above its flash point and will start burning. The combustion energy released will vaporize the surrounding spray or foam, and the fire will propagate (spread). The concept of flammability limits still applies, and the values for a given chemical are similar to those for a vapor–air mixture of the same chemical. Because the flames are spatially linked to the orifice, these stationary fires can be attacked by relieving the fluid pressure and thereby turning a spray into a flow of potentially lower hazard, by inerting the environment, or by applying a gas-phase active fire suppressant.

4.  A thin liquid layer drawn up by capillarity (wicking action) over the surface of a porous medium, such as a fabric or paper. A wick can consist of any nonmelting porous material that the liquid is capable of wetting and that is in contact with the pool of liquid. The liquid is drawn up the wick by surface tension (capillarity), and the wick becomes covered with a thin film of the liquid. (As an example, immerse one corner of a handkerchief in a glass of water and observe what happens.) When an ignition source is applied to the wick (such as a match to a candlewick), the thin film of liquid is heated rapidly to above its fire point and it ignites. As it burns, additional liquid is drawn up the wick and feeds the fire. Such a fire is in itself quite small. However, if the flame from this small fire comes in contact with a large liquid pool or a combustible solid, its heat could eventually warm the fuel immediately adjacent to it so that the fire would spread from the wick to a portion of the adjacent fuel, ultimately growing into a much larger fire.

5.  A confined liquid in a pressure vessel, heated by an external fire. Liquefied gases, such as propane, are stored in tanks that are designed to withstand high pressures from within. The compression of a gas to form a liquid enables storage of approximately 1000 times the mass of the chemical in the tank volume. At atmospheric pressure, propane boils at –44 °F (–42 °C); its vapor pressure is 101 kPa (1 atm) at this temperature. At 77 °F (25 °C), its vapor pressure is 960 kPa (9.6 atm); and on a hot day at 100 °F (38 °C), its vapor pressure is 1320 kPa (13 atm). Thus the tank is designed to withstand at least this pressure. It is fitted with a pressure relief valve, which opens if the liquid overheats and generates an excessive pressure. A typical setting for the pressure relief valve on a propane tank is about 2000 kPa.

If a fire should be burning near the outside of the tank, the temperature of the tank will rise due to the radiative and convective heat transfer from the flames. The temperature, and thus the pressure, of the propane inside this container will rise due to conduction through the tank wall. The liquid propane, which was initially above its boiling point, now boils vigorously.

As the wall of the steel tank gets hot, its tensile strength diminishes. (At about 930 °F (500 °C), the yield point of steel is approximately half of its normal value.) The combination of increased internal pressure and weakened steel ruptures the tank. The liquid contents, which were boiling at perhaps 10 atm, are now suddenly at 1 atm. A sudden expansion of the vapor occurs, resulting in an enormous eruption of vapor and liquid aerosol called a boiling liquid/expanding vapor explosion (BLEVE; pronounced “blev-ee”).

Note

As can be seen from the name, a BLEVE is a change of state from a liquid to a gas, leading to a major pressure increase and (catastrophic) failure of the container. The use of the word “explosion” is actually misleading, because an explosion involves chain branching chemistry. The liquid whose expansion leads to a BLEVE need not be flammable [7]. For example, BLEVEs of water and liquid nitrogen are possible.

Texture: Eky Studio/ShutterStock, Inc.; Steel: © Sharpshot/Dreamstime.com

The potential for loss of life and property can be increased significantly if the expelled fluid is ignited by contact with the external fire. This can lead to a burning cloud that can extend as much as hundreds of meters in diameter and propel pieces of the container as far as 1 kilometer. An example is shown in

Figure 8-5 The fireball formed from an ignited BLEVE. The small, dark object above the plume is a helicopter in the background.

© Ivan Cholakov/ShutterStock, Inc.

WRAP-UP

Chapter Summary

•   Three temperatures characterize the ability to ignite a liquid. The flash point is the temperature at which piloted ignition occurs, but is not sustained. The flash point is the temperature at which the vapor pressure of the liquid equals the lower flammability limit. At the slightly higher fire point, the flame is sustained. The autoignition temperature is hundreds of kelvins higher and applies to unpiloted ignition.

•   Class I liquids have flash points below 100 °F (38 °C), Class II liquids have flash points between 100 °F and 140 °F (38 °C and 60 °C), and Class III liquids have flash points at or above 140 °F (60 °C).

•   For a liquid pool of diameter less than about 0.2 m, the linear burning rate (surface regression rate) decreases as the pool diameter increases. In this pool-size range, fuel vaporization is affected by the container edges, which are heated by flame radiation. For pool diameters greater than 0.2 m, the burning rate rises with increasing diameter. The flames become turbulent and sootier, so flame radiation increases and vaporizes the fuel faster. A limit to the burning rate is reached at pool diameters near 2 m, where the flames are optically thick and the flame radiation to the fuel surface is no longer increasing.

•   If the liquid is at or above its flash point, the flame spread rate is fast, and the entire pool is engulfed within seconds. In such a case, the liquid is already evaporating sufficiently to reach the lower flammability limit in the vapor phase over the fuel surface. As the liquid temperature decreases, flame radiation must both heat the liquid to the flash point temperature and supply the heat of vaporization. As a result, the flame spread rate decreases sharply.

•   Liquid fuel fires generally fit into five categories: (1) a liquid pool, (2) a flowing liquid, (3) a spray from a small orifice at high pressure, (4) a thin liquid layer drawn up by wicking action over the surface of a porous medium, and (5) a liquid confined in a pressure vessel that is heated by an external fire.

•   Two consequences of fires are particularly hazardous: boilover and BLEVE. A boilover is a special case of a liquid pool, which can occur in an open container with flames over the upper of two immiscible liquid layers. If the lower layer, which has a higher density, has a lower boiling point, it can expand violently when its temperature reaches its boiling point.

•   A BLEVE (a liquid confined in a pressure vessel that is heated by an external fire) is also a result of the sudden expansion of a liquid when it vaporizes to a gaseous state. When this expansion occurs because the wall of the pressure vessel is weakened due to heating by the external fire, the vessel wall ruptures catastrophically. If the fluid is flammable and is ignited by the external fire, the flames are accelerated, and the potential damage is multiplied.

Key Terms

boiling liquid/expanding vapor explosion (BLEVE) A violent pressure release that occurs when a closed container of liquefied gas is heated externally, resulting in vaporization of the liquid, creating an internal pressure that exceeds the strength of the container material.

boilover The rapid overflow or expulsion of burning liquid fuel from an open container when water, located below the fuel surface, boils and expands.

fire point The minimum temperature to which a liquid must be heated, in a standardized apparatus, so that sustained combustion results when a small pilot flame is applied, as long as the liquid is at normal atmospheric pressure.

flash point The minimum temperature to which a liquid must be heated, in a standardized apparatus, so that a transient flame moves over the liquid when a small pilot flame is applied.

linear burning rate The rate at which the surface of a liquid pool recedes as it burns.

Challenging Questions

1.   From experiments, we know that the flammability limits of benzene are 1 percent by volume and 8 percent by volume, and that benzene’s temperature (°C) at a given vapor pressure (kPa) is approximated by T = [1200/(5.8 – log Pvap)] – 220. Calculate the flash point.

2.   Under which conditions can a liquid burn if its ambient temperature is below its flash point?

3.   For a chemical storage warehouse in a climate where building temperatures have never exceeded 90 °F, but where power outages are frequent, should safety precautions be taken for Class I liquids? Why? What if only Class II liquids were stored in the warehouse?

4.   n-Hexane is being stored in a cylindrical container (1 m in diameter, 1 m tall) that is located inside the same chemical storage warehouse described in Problem 3. The owner is concerned that, in the event of a leak in the container, the fluid might flow across the floor to the electrical box. In such a case, a local flammable vapor–air mixture might arise and be ignited by the electrical box. The warehouse owner builds a dam around the container, with the walls being deep enough to hold the entire contents. The dam is 3 m in diameter and 12 cm high, and the container stands above the liquid level.

A.  What is the maximum linear burning rate, mass burning rate, and rate of heat release should a major leak occur?

B.  A calculation indicates that a heat release rate of 10 MW will cause serious damage to the warehouse. So, instead of one large dam, the owner constructs 10 small dams, each separated from the other. If all 10 pools should ignite in the event of a major leak, what is the maximum linear burning rate, mass burning rate, and rate of heat release?

5.   How does the rate of flame spread over the surface of a liquid depend on the flash point?

6.   Under which conditions could boilover occur?

7.   Under which conditions could a BLEVE occur?

References

1.   ASTM D92-11: Standard Test Method for Flash and Fire Points by Cleveland Open Cup Tester. (2011). West Conshohocken, PA: ASTM International.

2.   Slye, O. M., Jr. (2008). Flammable and Combustible Liquids. In: Fire Protection Handbook, 20th ed., A. E. Cote, ed. Quincy, MA: National Fire Protection Association, Chapter 6.12.

3.   Rosales, K., and J. M. Stoltzfus. (2008). Oxygen-Enriched Atmospheres. In: Fire Protection Handbook, 20th ed., A. E. Cote, ed. Quincy, MA: National Fire Protection Association, Chapter 9.17.

4.   Kuchta, J. M. (1985). Investigation of Fire and Explosion Accidents in the Chemical, Mining, and Fuel-Related Industries: A Manual (Bulletin 680). Washington, DC: U.S Bureau of Mines.

5.   Burgoyne, J. H., and A. F. Roberts. (1968). “The Spread of Flame across a Liquid Surface.” Proceedings of the Royal Society, London, A308: 39–79.

6.   Hasegawa, K. (1989). Experimental Study on the Mechanism of Hot Zone Formation in Open-Tank Fires. In: Fire Safety Science: Proceedings of the Second International Symposium. New York, NY: Hemisphere, pp. 221–230.

7.   Peterson, D. F. (2001, April 1). “BLEVE: Facts, Risk Factors, and Fallacies.” Fire Engineering.