2
Assessment of Research on Alternatives for Halon

The Montreal Protocol and its amendments have led to bans on the production of halogenated alkanes that can cause depletion of stratospheric ozone. Thus the very effective fire- and explosion-suppression compounds, the halons, are being phased out, as well as the chlorofluorocarbons (CFCs) so prevalently used in refrigeration and cleaning systems. The use of halons as fire and explosion suppressants is well established, and their superior performance in these applications is supported by a wealth of documentation dating from before 1950. A comprehensive review of the existing literature is not attempted here, but the reader is referred to six reviews that summarize this topic and that contain more than a thousand references directly relevant to halons.1,2,3,4,5,6

The use of halons for fire suppression in military systems was the result of a study by the U.S. Army in the late 1940s,7 and no searches for new compounds had been conducted since then until the Montreal Protocol was signed. The halon-based fire suppression systems first used in aircraft and ships were CCl4 (halon 104) and then CH3Br (halon 1001). A shortcoming of these original systems was their toxicity in manned spaces.

Research on halons has been actively pursued since the 1940s at least, and halons 1301 and 1211, which are currently the most widely used total flooding and streaming agents, respectively, were selected from among a wide variety of candidate agents based on the following criteria: demonstrated efficacy in suppressing combustion, acceptable toxicological properties, availability in volume at acceptable cost, storage stability for at least 5 years in any climate, materials compatibility, low electrical conductivity, and acceptable activity of the compound or its combustion products (chemicals and/or particulates formed in the flame during extinguishment that may be toxic, corrosive, or both) in terms of low residue and low corrosion of metals. These are the same requirements that were noted in 1950 and 1951 in Air Force and Corps of Engineers documents.

Based on the suggestion by Molina and Rowland8 in 1974 that the presence of chlorinated halocarbons in the stratosphere may catalyze the destruction of stratospheric ozone, a new requirement was added, namely, that any agent must have acceptable environmental characteristics as determined by ozone depletion potential (ODP). Halons 1301 and 1211 both have a relatively high ODP, and domestic manufacture has been terminated.

The fire fighting community in general, and the U.S. Navy in particular, are interested in identifying alternative fire suppression agents that are toxicologically and environmentally acceptable, that are as efficacious as halons in suppressing fire, and that can be deployed in existing equipment designed for halons. This latter requirement necessitates a close match between the physical properties of halon compounds and any alternative agent that will replace them. The physical properties of an agent affect the mechanisms of dispersion and thus can profoundly influence overall effectiveness. For example, halon 1301 is stored as a liquid under pressure (N2, 600 psi) but vaporizes upon release and is dispersed as a gas. Gas-phase dispersion is a key factor in its effectiveness as a total flooding agent. Alternative agents with higher boiling points tend to be less effective overall.9 The coexistence of liquid and gas in delivery lines complicates the dynamics of distribution, and incomplete vaporization at the nozzle can greatly impede delivery of a total flooding agent to the flame source.

This chapter describes fire suppression models and the complex chemical and physical processes involved in understanding fires and their extinguishment, outlines the chemistry of manufacture of alternative agents currently proposed by industrial suppliers, and indicates the relative performance of commercially available halon replacements. In addition, it discusses the importance of toxicological considerations in evaluating chemical candidates for replacing halon and gives a brief overview of studies conducted to identify candidate halon alternatives and further elucidate mechanisms. The chapter



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 6
--> 2 Assessment of Research on Alternatives for Halon The Montreal Protocol and its amendments have led to bans on the production of halogenated alkanes that can cause depletion of stratospheric ozone. Thus the very effective fire- and explosion-suppression compounds, the halons, are being phased out, as well as the chlorofluorocarbons (CFCs) so prevalently used in refrigeration and cleaning systems. The use of halons as fire and explosion suppressants is well established, and their superior performance in these applications is supported by a wealth of documentation dating from before 1950. A comprehensive review of the existing literature is not attempted here, but the reader is referred to six reviews that summarize this topic and that contain more than a thousand references directly relevant to halons.1,2,3,4,5,6 The use of halons for fire suppression in military systems was the result of a study by the U.S. Army in the late 1940s,7 and no searches for new compounds had been conducted since then until the Montreal Protocol was signed. The halon-based fire suppression systems first used in aircraft and ships were CCl4 (halon 104) and then CH3Br (halon 1001). A shortcoming of these original systems was their toxicity in manned spaces. Research on halons has been actively pursued since the 1940s at least, and halons 1301 and 1211, which are currently the most widely used total flooding and streaming agents, respectively, were selected from among a wide variety of candidate agents based on the following criteria: demonstrated efficacy in suppressing combustion, acceptable toxicological properties, availability in volume at acceptable cost, storage stability for at least 5 years in any climate, materials compatibility, low electrical conductivity, and acceptable activity of the compound or its combustion products (chemicals and/or particulates formed in the flame during extinguishment that may be toxic, corrosive, or both) in terms of low residue and low corrosion of metals. These are the same requirements that were noted in 1950 and 1951 in Air Force and Corps of Engineers documents. Based on the suggestion by Molina and Rowland8 in 1974 that the presence of chlorinated halocarbons in the stratosphere may catalyze the destruction of stratospheric ozone, a new requirement was added, namely, that any agent must have acceptable environmental characteristics as determined by ozone depletion potential (ODP). Halons 1301 and 1211 both have a relatively high ODP, and domestic manufacture has been terminated. The fire fighting community in general, and the U.S. Navy in particular, are interested in identifying alternative fire suppression agents that are toxicologically and environmentally acceptable, that are as efficacious as halons in suppressing fire, and that can be deployed in existing equipment designed for halons. This latter requirement necessitates a close match between the physical properties of halon compounds and any alternative agent that will replace them. The physical properties of an agent affect the mechanisms of dispersion and thus can profoundly influence overall effectiveness. For example, halon 1301 is stored as a liquid under pressure (N2, 600 psi) but vaporizes upon release and is dispersed as a gas. Gas-phase dispersion is a key factor in its effectiveness as a total flooding agent. Alternative agents with higher boiling points tend to be less effective overall.9 The coexistence of liquid and gas in delivery lines complicates the dynamics of distribution, and incomplete vaporization at the nozzle can greatly impede delivery of a total flooding agent to the flame source. This chapter describes fire suppression models and the complex chemical and physical processes involved in understanding fires and their extinguishment, outlines the chemistry of manufacture of alternative agents currently proposed by industrial suppliers, and indicates the relative performance of commercially available halon replacements. In addition, it discusses the importance of toxicological considerations in evaluating chemical candidates for replacing halon and gives a brief overview of studies conducted to identify candidate halon alternatives and further elucidate mechanisms. The chapter

OCR for page 6
--> closes with the committee's findings regarding (1) the status of activities directed toward finding alternatives for halon as a fire suppressant and (2) the potential for discovering an environmentally acceptable equivalent. Chemical and Physical Considerations in Evaluating Alternative Fire Suppression Agents Several large chemical manufacturing concerns have invested heavily in the search for economically viable halon replacements, although the compounds investigated tend to be related to existing commercial products or their precursors. The National Institute of Standards and Technology (NIST) survey of candidate agents encompasses a somewhat broader range of compounds,10 including compounds whose study will help reveal the chemistry and physics of extinguishment. Fire Suppression Models—Chemical and Physical Mechanisms There are a number of ways in which a fire can be extinguished. The simplest introductory concept for understanding the behavior of a fire is the fire triangle, which has three sides: fuel, oxygen, and heat. If any leg of the triangle can be removed, the fire can be extinguished. Fires are categorized as being either flaming combustion or smoldering. The former is predominantly a gas-phase phenomenon and is characterized by the emission of visible and infrared (heat) radiation. The latter type of fire occurs when solids such as plastics burn, and heterogeneous reactions at or near the surface are important. Flaming combustion, the primary target of halon 1301, generates more heat and consequently is more dangerous in the short term. Smoldering often generates more toxic gas emissions and can be more difficult to extinguish. A fire can be extinguished by either physical or chemical mechanisms. Physical suppression mechanisms involve removing at least one leg of the fire triangle. (1) By smothering or blanketing the fire, the fuel and air are separated. An example of such a method is the use of foam extinguishers. (2) If the heat source is removed, the fire can also be suppressed. Thus, methods that cool the flame are important extinguishing techniques. For example, an agent with a high heat capacity can cool the flame by absorbing heat or can undergo a phase change that also requires heat. The most important parameters are the heat capacity and/or the latent heat of vaporization; experiments have shown that the thermal conductivity (how fast heat is transferred) is of lower importance. (3) Mechanical means such as forcing a gas over the flame at high velocity can extinguish a fire by separating the fuel and the air or the fuel and the heat. (4) For liquid or solid fuels, it is possible to place an agent that absorbs thermal radiation between the surface of the fuel and the flame. This prevents the generation of gaseous fuel and is known as flame radiation blockage. It is also possible to have chemical suppression of a flame. This occurs when an agent or its degradation products interfere with the chain reaction that is critical to sustaining combustion. When chain carriers in or near the reaction zone are removed, the chains are disrupted and the fire cannot sustain itself. It is possible to combine both physical and chemical effects in an agent, and in fact many of the best suppressants operate by both mechanisms. It is difficult to separate out the physical and chemical aspects of flame suppression. For example, an agent can remove heat by undergoing an endothermic decomposition process. If the decomposition products are inert, this is considered to be a physical process. However, if the products are reactive, then such products can contribute to a chemical suppression mechanism. The most likely radical chain species to be removed by a reagent or its degradation products are atomic hydrogen and oxygen and the OH radical. There are three types of fuels for fires: gas, liquid, and solids. Gaseous fuels are the easiest to understand as they can readily participate in chain reactions. For liquids, the fuel, in general, needs to be vaporized from the liquid surface, and so the amount of fuel is determined by the rate of vaporization and the surface area of the available liquid. However, because the flame generates heat that can affect the

OCR for page 6
--> amount vaporized, there is a complex coupling between the flame and the fuel source. Solid fuels are the most difficult to understand. The heat from the overhead flame (flame above the surface) is critical to generating the gases needed for flaming combustion. In some cases, the solid fuel can melt, forming a liquid that can be vaporized. In others, complex surface chemistry occurs, leading to the release of gaseous fuels that can feed the flame as well as to changes on the surface such as charring, which can block the emission of radiation. As is obvious from this brief discussion, there are complex flows that must be treated together with the chemistry if one is to really understand flames and fires. In the simpler case, the flow is smooth or laminar, and the adjacent layers of the fluid do not mix by mechanical means. If, however, there is random mixing of the fluids, a turbulent flow has been generated. In general vortices will be generated over a wide range of spatial scales, leading to complex mixing behavior that is not well understood. In order to better understand how to design new fire suppression agents, it is necessary to understand some basic concepts about combustion. Given a fuel, we can write down its reaction with oxygen and calculate how much energy is released (the heat of combustion), providing we know the heats of formation of the reactants and products. For a hydrocarbon, the reaction is usually written as Given the heat of combustion and the composition of the reaction mixture, it is possible to calculate the maximum or adiabatic flame temperature. The assumption that goes into this calculation is that all of the heat goes into heating the product gases and any other gases that are present. This calculation requires knowing the heat capacities of all of the gases. The heats of formation of all of the species are needed as well as the heat capacities in order to calculate the temperature dependence of the heat of reaction. From the expression ΔG = -RTlnK where K is the desired equilibrium constant and ΔG is the free energy, one needs to know the free energy of the process. From ΔG = ΔH - T ΔS, it is clear that one needs not only the enthalpy of the process (ΔH) but also the entropy (ΔS) of the process. Again, the entropies are needed for all important species. The overall equilibrium composition can then be calculated by an iterative procedure. Tables of thermodynamic properties exist, and methods have been developed for computing missing information.11,12,13,14 The reaction thermodynamics described above describe what happens at equilibrium but do not predict how fast the system will reach equilibrium. In order to determine the speed of the process, kinetic information is required. From the global reaction mechanism, we need to write down a reaction mechanism based on individual reaction steps, each of which is a fundamental chemical process, a unimolecular, bimolecular, or termolecular reaction. Then we have to determine the kinetics of each fundamental reaction step and use the rate constants to solve for a global kinetic rate. This is a complex process because much of the required data is not known. However, if the data are not available, methods exist for estimation.15 The types of reactions important in the combustion process are based on radicals.16 If the chain reactions have long chain lengths, then the flame can continue to exist. It is important to note that combustion temperatures tend to be high, >800 K, so that generation of radicals is more important than loss due to recombination. The first reactions in the chain are the initiation reactions that lead to the initial formation of radicals. Most initiation reactions involve breaking a chemical bond and thus have high activation energies. They tend to be slow even at flame temperatures. The chain propagation reactions consume fuel or oxidizer but do not change the number of radicals; hence the chain continues. The chain may branch, leading to an increased number of radicals and hence a higher global reaction rate. These reactions generally consume the oxidizer, in most cases O2. As most of the fuels are molecules containing carbon and hydrogen, H is an important radical in the chain. The most important branching reaction is

OCR for page 6
--> which has an activation energy of 17 kcal/mol (due for the most part to the fact that it is an endothermic process). Note that two radicals are produced for each radical consumed. Reactions of O atoms with fuel are important, as two radicals are often produced for each O consumed. Termination reactions are those that reduce the number of active radicals or lead to formation of less reactive radicals. Many of these reactions are termolecular in nature, involving an unreactive third body such as N2, the predominant gas in the atmosphere, and they increase in rate with decreasing temperature. By understanding the reaction mechanism, it should be possible to design a reagent that can interfere with the production of an important radical and hence reduce the chain length. For example, reaction (2) has a competing reaction (3), which produces a less reactive radical where M is a third body, At 1000 K, these reactions have comparable rates, and at lower temperatures, the three-body process dominates. If H atoms are removed and the temperature can be lowered, the chains will be shorter. The effects of thermochemistry and kinetics can be combined as the concept of thermal balance. If the flame is to continue, the energy released by the combustion process must keep the temperature high enough so that branching reactions dominate over termination reactions. With the advent of modem high-performance computers and advanced software, it is now possible to model some of the characteristics of flames that are extremely difficult to measure.17 Thus computational models have been developed to describe both the flame and the suppressant. The first types of models are those that describe how the agent is released. The agent must be vaporized and dispersed quickly throughout the region where the fire is occurring. The models that have been developed include the design of the holding vessel for the suppressant and most importantly the design of the exit orifice and the flow across it. This allows one to also include the use of pressurizing agents such as N2. One can then measure the agent discharge to validate the model. Furthermore, by using modem computational fluid dynamics codes, it is also possible to use models to look at how the gas flows over a complicated interior space such as that in an engine room. Work has also been done to model simple flames, including transport and important chemical reactions. An important advance has been to use chemical reactor models such as plug flow and continuous stirred tank reactor models to predict the efficiency of suppressants. These provide good initial estimates of the chemical efficacy of a suppressant. The whole area of the modeling of reactive chemical transport is growing, and it should be possible within a few years to perform quite accurate simulations of simple flames and suppressants, given the required input data. However, it is still quite difficult to model turbulent flows and complex kinetics, and such studies will require at least teraflop computers and the design of new computational methods for treating turbulence. Studies of this type are important so that the significant investments being made in other areas of research can be leveraged to provide the information needed to design new suppressants. The models can also be used to predict concentrations of harmful by-products such as hydrofluoric acid. As noted above, one needs to have a wide variety of thermodynamic data for designing new suppressants, especially if one is looking for chain inhibitors. It is time consuming and expensive to measure the needed thermodynamic properties if they are not known, especially for radicals. As one looks at a new agent whose thermochemistry is not known, one needs to either estimate or calculate its properties.18 Estimation methods are showing dramatic improvements as the available databases expand. However, it is often easier to calculate thermochemical properties from ab initio quantum chemical methods than it is to measure them. Given a molecular structure, one can calculate absolute heats of formation either rigorously, with simple empirical corrections, or from idealized reactions based on known thermochemical properties. The accuracy of the results can usually be fine-tuned depending on the cost of the calculation, but it is often possible to predict heats of formation to within ±2 kcal/mol. It is just as straightforward to do this for radicals as for stable species. It is even easier to calculate the heat capacities (Cp) and entropies, as the geometries and vibrational spectra of the species need to be

OCR for page 6
--> calculated to predict heats of formation. Statistical mechanics can then be used to provide Cp and ΔS as a function of temperature. Thus the global thermochemistry for a given mechanism can be predicted.19 Kinetic measurements can be made accurately, but these can be quite difficult and time consuming. Furthermore, it is often difficult to measure rates involving complex radicals or reactions that are very slow. It is also possible to calculate the rate constants from ab initio quantum chemistry and kinetic theories such as transition state theory or unimolecular rate theories. Such calculations are more difficult than the thermochemical ones, but the calculated results can often be scaled to known experimental results to extend the measured data. It is still difficult to predict rate constants to better than a factor of about five, although for simple cases much higher accuracy has been obtained.20,21 Halons suppress fires by both physical and chemical mechanisms. With halons, the bromine atoms are critical to disrupting the radical chains. Other agents that release halogens can have the same effect. Chlorine radicals are less effective than bromine at flame suppression, whereas iodine radicals are more effective. Unfortunately, the same halogen atoms (bromine, chlorine, and iodine) that contribute to the effectiveness of halogenated compounds, including halons, as fire suppression agents also contribute to their ozone-depleting potential. Alternative agents that do not contain halogens other than fluorine tend to be physical action agents only and therefore are generally less effective than halons. Alternative agents that contain iodine tend to be readily tropodegradable; that is, they react in the lower atmosphere through photolysis or some other mechanism to produce stable compounds that will not deliver iodine radicals to the stratosphere. These compounds, while retaining the effective chemical mechanism of fire suppression common to halons, tend to be either unstable with respect to storage and material compatibility, or toxic, or both. There is prior work for some suppressants in terms of their fire extinguishing mechanisms. These include the perfluorocarbons FC-116 (C2F6), FC-218 (C3F8), and FC-3-1 (C4F10); the hydrofluorocarbons HFC-125 (CHF2CF3), HFC-227 (CF3CHFCF3), HFC-318 (CF3CHFCHFCF3), HFC-3110, HFC-32/125, HFC-134a (CH2FCF3), and HFC-236fa (CF3CH2CF3); the hydrochlorofluorocarbons HCFC-22 (CHF2Cl) and HCFC-124 (CHClFCF3), and CF3I. None of these is as good a suppressant as halon 1301 in all respects. Measured or calculated data on C1 and C2 HFCs has been used in extensive predictions about the behavior of such compounds as suppressants. Compounds predicted from the modeling search to be possible candidates for replacing halons include CF2HCl (HCFC-22), CF2=CClF, CF2=CFCF3, CF2=CFBr, CF 2=CHBr, CF3I, SiF4, and NF3. As discussed above, fire suppression agents act through both physical and chemical mechanisms. Physical action can extinguish flames by dilution and by providing a heat sink. Chemical action follows from interaction of the agent with the chemistry of the flame. Huggett22 found empirically that liquid fuels would not bum if the heat capacity of the mixture exceeded about 50 cal/deg per mole O2. Thus, a crude measure of the concentration of agent necessary to extinguish a flame by physical effects alone can be estimated by calculating the concentration that will raise the heat capacity of the surrounding atmosphere to this level. Table 2.123 gives the result of this calculation for several agents of interest. The difference between the observed extinguishment concentration and the calculated concentration is interpreted as a measure of the chemical effect. Sheinson and coworkers24 have carried through a more sophisticated analysis along these lines and have been able to measure the physical and chemical contributions to fire suppression for various agents. For example, for CF3Br the effect is 20% physical, 25% chemical owing to CF3, and 55% chemical owing to Br. For CF3Cl, a less effective agent, the effect is 40% physical, 50% chemical owing to CF3, and 10% chemical owing to Cl. CF4 has a physical effect comparable to that of the other substituted methanes, but no chemical effect owing to the strength of the CF bond.

OCR for page 6
--> Table 2.1 Concentration of Agent Needed for Physical Extinguishment Halocarbon Number Chemical Formula Concentration for Extinguishment Difference     (calculated) (observed) (calc. - obs.) H-1301 CF3Br 17.6 2.9 14.7 IFC-13I1 CF3I — 3.0 — HCF-23 CF3H 23.0 12.6 10.4 HCFC-22 CF2HCl 21.3 11.6 9.7 HCFC-124 CF3-CHClF 13.5 6.7 6.8 PFC-14 CF4 20.0 13.8 6.2 HFC-236fa CF3-CH2-CF3 11.1 5.6 5.5 HCFC-123 CF3CHCl2 12.9 7.5 5.4 PFC-116 CF3-CF3 12.5 7.8 4.7 HFC-125 CF3-CHF2 13.8 9.4 4.4 HFC-227ea CF3-CFH-CF3 10.3 6.3 4.0 PFC-218 CF3-CF2-CF3 9.7 6.1 3.6 HFC-134a CF3-CFH2 14.1 10.5 3.6 PFC-3-1-10 CF3-CF2-CF2-CF3 7.4 5.0 2.4 Manufacturability of Commercially Available Halocarbon Fire Suppression Agents The manufacture of halon replacements is typically more complex than the manufacture of halons themselves. The chemical attribute that generally renders halon-like replacements easily degradable in the lower atmosphere, the hydrogen atom substituent, complicates the manufacturing process because it enables parasitic processes such as formation of significant by-products or the inactivation of catalysts used in production. Synthesis of halocarbon replacement agents often requires the use of toxic, corrosive, or otherwise difficult-to-handle precursor materials as well as severe reactor conditions. Nonetheless, economies of scale allow many of these compounds to be offered at reasonable prices, and several agents are already commercially available in quantity. Other proposed alternatives are available only in small quantities and at significantly higher prices and are thus classified as specialty chemicals. If one of these agents were to be widely adopted, it is possible that increased demand would stimulate production and lead to greater availability at lower cost. Tables 2.2 and 2.3 list commercially available halocarbon replacements for halons in total flooding and streaming applications, respectively. The toxicological and environmental properties of the listed compounds have been well characterized.

OCR for page 6
--> Table 2.2 Commercially Available Replacement Agents for Total Flooding Applications Halocarbon Number Chemical Formula Supplier Designation PFC-218 CF3-CF2-CF3 3M CEA-308™ PFC-3-1-10 CF3-CF2-CF2-CF3 3M CEA-410™ HFC-23 CF3H DuPont FE-13™ HCFC-125 CF3-CHF2 DuPont FE-25™ HFC-227ea CF3-CFH-CF3 Great Lakes Chemical FM-200™ HFC-236fa CF3-CH2-CF3 DuPont FE-36™ HCFC-124 CF3-CHClF DuPont FE-241™ HCFC Blend A, a blend of:   North American Fire Guardian NAF-S-II HCFC-22 CF2HCl   HCFC-123 CF3CHCl2   HCFC-124 CF3CHClF   IFC-13I1 CF3I Pacific Science Triode™ Table 2.3 Commercially Available Replacement Agents for Streaming Applications Halocarbon Number Chemical Formula Supplier Designation PFC-5-1-14 CF3-(CF2)4- CF3 3M CEA-614™ HFC-227ea CF3-CFH-CF3 Great Lakes Chemical FM-200™ HFC-236fa CF3-CH2-CF3 DuPont FE-36™ HCFC-123 CF3CHCl2 DuPont FE-232™ HCFC-124 CF3CHClF DuPont FE-241™ HCFC Blend B, which is primarily HCFC-123 (see above)   American Pacific Halotron™ HCFC Blend C, a blend of:   North American Fire Guardian NAF-P-III™ HCFC-134a CF3CH2F   HCFC-123 CF3CHCl2   HCFC-124 CF3CHClF   IFC-13I1 CF3I Pacific Science Triode™ None of the agents listed in Tables 2.2 and 2.3 contain bromine, which is the chemically active species in halons. However, there is some evidence to suggest that the agents listed do effect some degree of chemical suppression. This could be owing to the production of CF3 radicals, which may serve to chemically inhibit combustion.25

OCR for page 6
--> Synthesis of Alternative Fire Suppression Agents This section describes the known routes of commercial manufacture of the agents of interest as replacements for halon. The synthesis of these agents requires the safe handling of exceedingly hazardous and toxic materials and intermediates. High pressures and temperatures are also required for synthesis, thus adding to the hazard. Exotic alloys must be used in the contraction of reactor vessels, because more common and less expensive materials, such as stainless steel, are subject to unacceptably high corrosion rates. HCFC-22 (CF2HCl) and HFC-23 (CF3H) are both produced from chloroform by a stepwise halogen exchange reaction. This process is typically carried out in the liquid phase utilizing a Lewis acid catalyst and hydrogen fluoride as the fluorine source.26 HCFC-123 (CF3CHCl2) can be produced commercially in at least three ways. Perchloroethylene is reacted with excess hydrogen fluoride in the presence of a Lewis acid catalyst to produce HCFC-123 directly, 27 equation (4). Alternatively, CFC-113a (CF3CCl3) can be hydrodechlorinated in the presence of hydrogen and a noble-metal catalyst,28 equation (5). The third choice is to selectively chlorinate HCFC-133a (CF3CH2Cl) with elemental chlorine in the presence of a metal salt catalyst,29 equation (6). HCFC-124 (CF3CHClF) and HFC-125 (CF3CHF2) can be co-produced using HCFC-123 (CF3CHCl2) as the starting material.30 In this vapor-phase process, excess hydrogen fluoride and HCFC-123 are passed over a transition metal halide or oxide catalyst at elevated temperatures, equation (7). If HCFC-124 is not desired, HCFC-125 can be made the sole product by increasing the temperature above 300°C and increasing the HF to HCFC-123 mole ratio in the feed stream,31 equation (8). HFC-227ea (CF3CFHCF3) is produced in a three-step process beginning with HCFC-22 (CF2HCl).32 In the first step, HCFC-22 is pyrolized to produce tetrafluoroethylene and hydrogen chloride, equation (9). The tetrafluoroethylene is then pyrolized to hexafluoropropylene, 33 equation (10). Finally, hydrogen fluoride is added to yield HFC-227ea, 34 equation (11).

OCR for page 6
--> HFC-236fa (CF3CH2CF3) can be produced by at least three different processes. The first process, described by equations (12) and (13), involves two steps.35 In the first step, perchloropropylene is reacted with hydrogen fluoride to produce CFC-216aa (CF3CCl2CF3) and HCFC-226da (CF3CHClCF3). These intermediates are then hydrodechlorinated to produce HFC-236fa in the second step. The second process36 also involves two steps, equations (14) and (15). In the first step, vinylene chloride is condensed with carbon tetrachloride to yield HCC-230fa (CCl3CH2CCl3). HCC-230fa is then treated with hydrogen fluoride in the liquid phase in the presence of a Lewis acid catalyst to give HFC-236fa. Alternatively, this step can be performed in the vapor phase.37 HFC-236fa can also be produced from perfluoroisobutylene methyl ether, equation (16). Perfluorocarbons are commercially produced by two different process technologies. In the first process, equation (17), a hydrocarbon is passed over a hot, agitated bed of cobalt trifluoride where it is convened to the desired perfluorocarbon. The cobalt trifluoride is convened into cobalt difluoride during the process and is regenerated in a separate treatment with elemental fluorine.38 The second perfluorocarbon process involves electrochemical fluorination, i.e., the electrolysis of hydrocarbons in anhydrous hydrogen fluoride, known as the Simmons process.39 Nickel anodes and nickel or steel cathodes are used. This method is limited to starting materials that have appreciable solubility in hydrogen fluoride. For volatile materials with little solubility in hydrogen fluoride, a complementary method that uses porous carbon anodes and a (KF)2HF electrolyte has been developed that is known as the Phillips process.40 IFC-13I1 (CF3I) can be conveniently produced on the laboratory scale 41 by heating the silver salt of trifluoroacetic acid in the presence of elemental iodine. In principle, heating trifluoroacetyl chloride in the presence of potassium iodide,42 equation (18), represents an industrially viable route to obtaining IFC-13I1. In all of the chemistry described above, many of the materials are toxic or corrosive or both. Handling and reaction on a commercial scale are therefore generally expensive. The manufacture of these materials would not be undertaken without the promise of financial reward.

OCR for page 6
--> Table 2.4 Relative Performance of Commercially Available Halon-Like Replacement Agents Compared to Halon 1301 CFC Number Chemical Formula Extinguishment (volume % required) H-1301 CF3Br 3.5 IFC-13I1 CF3I 3.0 PFC-3-1-10 CF3-CF2-CF2-CF3 5.0 HFC-236fa CF3-CH2-CF3 5.4 PFC-218 CF3-CF2- CF3 5.6 HFC-227ea CF3-CFH-CF3 6.3 HCFC-124 CF3-CHClF 7.0 HCFC-123 CF3CHCl2 7.5 PFC-116 CF3-CF3 7.7 HCFC-123 a CF3-CHCl2 8.1 HFC-125 CF3-CHF2 9.1 HFC-236ea CF3-CH2-CF3 10.2 HFC-134a CF3-CFH2 10.5 HCFC-22 CF2-HCl 12.7 HCF-23 CF3H 13.0 PFC-14 CF4 18.2   SOURCE: The DuPont Company. Effectiveness of Alternative Commercially Available Fire Suppression Agents The various agents discussed above have been evaluated using a standard cup burner test with n-heptane fuel.43 Table 2.4 shows the results for those agents and others, ranked in order of performance, where performance is based on the volume fraction required to extinguish the flame—the most effective agent is that which requires the lowest volume for extinguishment. Halon 1301 is included for comparison. Agents are also compared on a per mole and per weight basis normalized to halon 1301. Toxicology Issues Toxicology is a key aspect of the effort to identify suitable halon replacements. Although the Navy evacuates personnel from spaces to be flooded with halon 1301, there is always the possibility of accidental discharge, and the hazard must be carefully assessed. The toxicology of halon alternatives has been studied, and protocols for agent testing and use have been delineated. In evaluating the toxicology of chemical candidates for replacing halons as fire suppressants, the key consideration is that potential exposures, although infrequent, will possibly be at high levels. The population exposed will be under stress and possibly engaging in a high level of physical activity. If effects are seen in studies focusing on brief, high-level exposures, additional follow-up studies may be

OCR for page 6
--> desirable to further characterize and understand them. In addition, some consideration should be given to carcinogenic potential. The initial study of a candidate replacement for halon should be an acute inhalation toxicity study. The study should be robust enough to define the dose response curve as well as the median lethal concentration. Typically, these studies involve single exposures to groups of five male and five female rats for a period of 4 hours. Mice can also be used and can be the species of choice when the substance being tested is in short supply. Although 4 hours is a long period of time relative to the brief exposure (typically 1 to 2 minutes) expected under conditions of actual use, this time period has become a comparative standard and allows for selection of substances based on a common protocol. Should lethality be a concern, an additional study for a shorter time period could subsequently be conducted. Recently, more attention has been given to predicting the effects of brief exposures to suppressants, and protocols have been developed that include exposure periods of 1 hour, 30 minutes, and even 10 minutes. Information on exposure for brief periods is being used by the U.S. EPA-sponsored Federal Advisory Committee on Acute Exposure Guidance levels,44 and more information on the effects of brief exposures will likely be available from future studies. In addition to observations for lethality, detailed clinical observations of test animals should be conducted during and after exposure. It is common to observe animals at hourly intervals during and immediately after exposure, and then daily for the remainder of a study, to allow determination of other effects such as central nervous system depression (many of the chemicals used as fire suppressants are anesthetics); neurotoxic effects, such as development of convulsions or impaired movement; or irritation of the eyes or respiratory system. Surviving test animals are held for 2 weeks for observation of postexposure effects and recovery. Body weights are measured frequently and can be sensitive indicators of an adverse response to exposure. Some key organs, such as the lungs, kidney, liver, testes, and heart, may be weighed and examined microscopically for abnormal changes, although such examination rarely produces useful information because test animals tend to recover during the 14-day observation period. If only minimal effects of exposure are seen in test animals, the next phase of testing can be initiated. If the compound being tested shows marked toxicity, it may be eliminated from consideration as a flame suppressant in occupied areas. If the substance exhibits intermediate toxicity, additional testing may be warranted. If the chemical at relatively low levels induces sleep (central nervous system depression), it may be desirable to evaluate it for anesthetic potential to determine if its use could create a hazard by reducing the alertness of people exposed to it. If signs of neurotoxicity, such as impaired movement or convulsions, are seen, a study designed to look at those endpoints may be indicated. A basic protocol has been developed by the Organisation for Economic Cooperation and Development 45 and can be used as a reference point. In this application it may be desirable to add some of the parameters described above. Information on a candidate substance's potential to cause skin46 and eye irritation47 may also be valuable for use in protecting maintenance personnel and those other individuals going into an area after discharge of a fire suppressant and coming into contact with walls, furniture, and other items that have been contaminated. Many of the chemicals used as flame suppressants have the potential to sensitize the heart to the effects of adrenaline. A protocol for evaluating this endpoint has been developed48 and was reviewed recently by the National Research Council's Committee on Toxicology. This is a sensitive test because administration of adrenaline by injection increases blood levels well above those seen under conditions of stress.49,* If a response is seen at a concentration lower than one that could be encountered following discharge of a flame suppressant, this effect should be carefully considered. While there has been extensive discussion about the most appropriate method for extrapolation of this data for use in human *   Briefly, in this study, dogs were exposed to the test compound at a set exposure level for a period of 5 minutes, during which time the blood level of the chemical approached near-equilibration levels. The test dogs were then given an injection of adrenaline at a level just below that necessary to cause a cardiac arrhythmia (a rapid, irregular, potentially fatal heart beat). Each dog's EKG pattern was continuously measured during this procedure. The appearance of a response to the adrenaline injection was taken as a positive response.

OCR for page 6
--> risk assessment, in fact, estimates of exposures associated with the development of cardiac arrhythmias in humans have always been much higher than the threshold seen in dogs. This has led to a direct application of the no-effect levels observed in these dog studies to human risk assessments. In the example above with halon 1301, the recently reported no-effect levels are 75,000 to 100,000 ppm in two dog studies. Comparing this to the projected levels for flooding at 50,000 to 70,000 ppm indicates that the application levels are appropriate for use in an occupied area. It is desirable to conduct a multiple-exposure study to determine what effects may result from repeated low-level exposures to fire extinguishing agents, such as are encountered in the workplace or during maintenance operations, especially if indications of toxicity were seen at exposure levels approaching those that might be encountered under actual discharge or use conditions. Either a 4-week50 or 13-week 51 exposure period would be appropriate. The 13-week design may be more desirable if frequent exposures are anticipated. Such a study normally involves a series of daily 6-hour exposures for either 5 or 7 days per week using four groups of 10 male and 10 female rats in each group. Recovery groups of five animals per sex may be included. These animals are held for 2 weeks following the final exposure to assess the animal's potential to recover from the effects of exposure. During the study, in addition to measurements of body weight and clinical observations, complete blood counts and serum chemistry analysis are conducted. At the end, the animals are sacrificed and all major organs are examined for effects of the exposure. If a recovery group has been included, the same parameters are also evaluated for this group. When there is the possibility of pregnant women being exposed repeatedly to a chemical agent for fire suppression, it would be desirable to conduct a developmental toxicity study.52 Information on the potential of a chemical agent to cause either birth defects or developmental toxicity can be obtained from an inhalation developmental toxicity study. In this type of study, pregnant animals, usually rats, mice, or rabbits, are exposed to the chemical during the organ-development phase of pregnancy. At the end of the pregnancy the offspring are evaluated for abnormalities that could have been caused by the mother's exposure to the chemical. The most critical findings would be frank birth defects; however, other effects such as birth weight are also evaluated. A single species, either the rat or mouse, should provide adequate information. In the event of a positive finding, a second study, with rabbits, should be considered. The results from these studies, along with an estimate of potential exposure levels, should be considered in determining the potential risk to pregnant women exposed to chemical fire suppressants during maintenance or other procedures and in area discharges. Mutagenicity is the last element that should be considered in evaluating the toxicology of candidate replacements for halon. Although a carcinogenic study would not be a part of an evaluation program, conducting a few studies to evaluate mutagenic potential would aid in our understanding of carcinogenic potential. Two that are particularly useful are the Ames assay and the human lymphocyte chromosome aberration assay. Both are run with cell cultures. In the Ames assay, bacteria are exposed to a test chemical and their potential to revert to a form capable of growing in a medium lacking a normally essential nutrient is evaluated. For this change to occur, the bacteria must mutate. In the second assay, the direct effect of exposure to a chemical on the chromosomes in a human cell line is evaluated using a microscope. Breaks in the chromosomes are positive indicators. Other genetic studies, including those involving whole animals, such as the micronucleus assay, may be considered for special programs. While other specialized studies involving metabolism, uptake, and pharmacokinetics may be indicated from time to time based on preliminary results, their selection would be compound specific. These studies are usually conducted when the levels of concern from one or more of the toxicology studies are near the potential human exposure level. They answer questions such as whether a test animal metabolizes a chemical in the same way as a human does, and whether the dose of a chemical actually reaching an organ system of concern would be higher or lower in a human compared with a test animal exposed at the same airborne level. The results from such studies allow for a more precise assessment of risk for humans exposed to particular chemical compounds. Generally, a program looking

OCR for page 6
--> Table 2.5 Summary of Selected Toxicity Data for Possible Halon Substitutes   4-hr LC50 (ppm)   Developmental Toxicity Genetic Assays Substitute       Ames Chromosome Aberration HCFC-22a >200,000 50,000 Possible slight effect rat—50,000 Not active Not active       No effect rabbit—50,000     HCFC-123b 32,000 20,000 No adverse effects rat and rabbit Not active Positive (not active, 5 other genetic assays) HCFC-124c >360,000 25,000 No adverse effects rat and rabbit Not active Not active HFC-125d >800,000 75,000 No adverse effects rat and rabbit Not active Not active HFC-134ae >500,000 80,000 No adverse effects rat and rabbit Not active Not active HFC-227eaf >800,000 100,000 No adverse effects rat and rabbit Not active Not active HFC-236fag >457,000 150,000 No adverse effects rat and rabbit Not active Not active a HCFC-22. Joint Assessment of Commodity Chemicals No. 9 Chlorodifluoromethane. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium (October, 1989). b HCFC-123. Joint Assessment of Commodity Chemicals No. 33 1, 1-Dichloro-2,2,2-trifluoroethane. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium (February 1996). c HCFC-124. Joint Assessment of Commodity Chemicals No. 25 1-Chloro-1,2,2,2-tetrafluoroethane. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium (July 1994). d HCFC-125. Joint Assessment of Commodity Chemicals No. 24 Pentafluoroethane. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium (May 1994). e HFC 134a. Joint Assessment of Commodity Chemicals No. 31 1, 1,2-Tetrafluoroethane. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels, Belgium (February 1995). f Personal communication from Great Lakes Chemical Corp., One Great Lakes Blvd., W. Lafayette, IN 47906. g Toxicologic Profile of the Alternative Refrigerant, HFC-236FA, R. Valentine, D.A. Keller, et al. The Toxicologist, No. 30 1, 2, Abstract 1489 (March 1996). at acute inhalation toxicity, cardiac sensitization, potential developmental toxicity, and possibly the effects of repeat exposure, along with a limited genetics screen, should be adequate to characterize the toxicity of potential substitute flame suppressant materials. Table 2.5 lists several results of such toxicology studies for some of the compounds of interest. On exposure to flames, halons and halocarbon alternatives react to yield toxic chemicals. CO and CO2 are present in fire atmospheres generally, but halocarbons produce HF, HCl, HBr, and COF2. These chemicals are produced during extinguishment with halon 1301, but in much smaller quantities. The amounts of these dangerous compounds can be minimized by reducing the time to extinguish the fire, increasing the amount of agent employed, and discharging the agent rapidly into the fire volume. The importance of this source of toxicity depends on cool-down time before reentry of the affected space. Design in this connection will be agent and practice specific.53,54

OCR for page 6
--> Brief Overview of Studies on Halon Substitutes and Mechanisms In 1986, the U.S. Air Force recognized the potential for regulation of halons and initiated a program to identify halon replacements. This program demonstrated that halon replacements with decreased or zero ODP could be obtained by (1) eliminating bromine and (2) substituting hydrogen for fluorine to decrease atmospheric lifetime. 55 The Air Force program identified HCFCs, HFCs, and PFCs as the most likely near-term replacements. The National Institute of Standards and Technology (NIST) issued a report in 1990 whose objective was to initiate a systematic search for halon replacements.56 This report is firmly grounded in the science and technology of combustion and current understanding of the role of halogenated compounds in ozone depletion in the stratosphere. The centerpiece of the publication is a list of 103 chemicals ''covering a range of chemical and physical principles thought to affect flame suppression capability.'' The list includes saturated halocarbons, halogenated ketones, esters and anhydrides, unsaturated halocarbons, halogenated ethers, sulfur halides, compounds containing phosphorus, silicon, germanium and metals, and inert gases. The list contains both candidate halon replacement agents and compounds chosen to test principles of fire suppression. While the NIST list is not claimed to be exhaustively inclusive, it is sufficiently extensive to suggest that the most obvious classes of alternative agents are included. The list includes most of the commercially available agents listed in Tables 2.2 and 2.3, omitting HFC-227ea and HFC-236fa. These compounds were later considered by NIST for aircraft nacelle and dry bay applications. 57,58 Flame suppressants that do not endanger the ozone layer either do not contain the heavy halogens (chlorine, bromine, or iodine) or have atmospheric lives so short that they do not reach the stratosphere. Perfluorocarbons satisfy the first condition but do not offer chemical flame suppression activity and are somewhat less effective in extinguishing fires. Compounds that contain the heavy halogens but are destroyed before they reach the stratosphere are of special interest as halon replacement agents. Substitution of H for a halogen, increasing the dipole moment, and shifting the absorption toward the red are strategies pursued in the NIST list, including compounds containing hydride, ketone, ester, anhydride, and double or triple bonds. In connection with halon 1301 replacement agents, only a limited number of compounds have matching physical properties, i.e., are storable as a liquid at normal ambient temperatures, but have a high vapor pressure and so evaporate rapidly on discharge. Of the 103 compounds listed by NIST, only 31 have boiling points below 0°C (the b.p. for halon 1301 is -68°C), and some of these are toxic or have interest only for determining mechanisms. A low boiling point and high vapor pressure strongly indicate compounds of low molecular weight and/or high fluorine content. The point is that the number of candidate halocarbon halon replacement agents is not very large. Tapscott and coworkers59 are testing various classes of compounds as candidates for total-flooding fire suppression, including brominated fluoro ethers, alkenes and aromatics, amines, and carbonyls, as well as phosphorus and silicon compounds. Some of these materials have good fire suppression characteristics, but satisfactory performance in terms of toxicity, materials compatibility, and other properties remains to be proved. Findings FINDING: After reviewing research, development, toxicology, and engineering activities directed toward finding alternative and replacement agents for halon 1301 and halon 1211, the committee finds that in this context, the relevant aspects of the problem are being studied effectively and a comprehensive body of scientific and engineering knowledge is being developed, and the committee has identified no obvious gaps in these important efforts. FINDING: It is unlikely that a drop-in replacement agent will be discovered that will exhibit all of the beneficial properties of halon 1301 and not also exhibit a significant environmental impact .

OCR for page 6
--> References 1. An Appraisal of Halogenated Fire Extinguishing Systems, National Academy of Sciences, Washington D.C. (1972). 2. Halogenated Fire Suppressants, R.G. Gann, Ed., American Chemical Society, Washington D.C. (1975). 3. W.M. Pitts, M.R. Nyden, R.G. Gann, W.G. Mallard, and W. Tsang, Construction of an Exploratory List of Chemicals to Initiate the Search for Halon Alternatives, NIST Technical Note 1279, U.S. Department of Commerce, Washington D.C. (1990). 4. Evaluation of Alternative In-Flight Fire Suppressants for Full-Scale Testing in Simulated Aircraft Engine Nacelles and Dry Bays, R.G. Gann and W.M. Pitts, Eds., U.S. Department of Commerce, Washington D.C. (1994). Note particularly Chapter 5: "Flame Inhibition Chemistry and the Search for Additional Fire Fighting Chemicals," pp. 467-641. 5. Halon Replacements, A.W. Miziolek and W. Tsang, Eds., American Chemical Society, Washington D.C. (1995). 6. Fire Suppression System Performance of Alternative Agents in Aircraft Engine and Dry Bay Laboratory Simulations, R.G. Gann, Ed., SP890, U.S. Department of Commerce, Washington D.C. (1995). 7. J.F. Malcolm, NFPA Quarterly 45, 119 (1951). 8. M.J. Molina and F.S. Rowland, Nature 249, 810 (1974). 9. Evaluation of Alternative In-Flight Fire Suppressants for Full-Scale Testing in Simulated Aircraft Engine Nacelles and Dry Bays, R.G. Gann and W.M. Pitts, Eds., U.S. Department of Commerce, Washington D.C. (1994), p. 305. 10. W.M. Pitts, M.R. Nyden, R.G. Gann, W.G. Mallard, and W. Tsang, Construction of an Exploratory List of Chemicals to Initiate the Search for Halon Alternatives, NIST Technical Note 1279, U.S. Department of Commerce, Washington D.C. (1990). 11. D.A. Dixon and K.D. Dobbs, "Advanced Materials," Transactions of the Material Research Society of Japan 18A (1993); D.A. Dixon and K.D. Dobbs, Ecomaterial, V, R. Yamato et al., Eds., Elsevier, Amsterdam , (1994) p. 373. 12. D.A. Dixon, J. Phys. Chem. 92, 86 (1988). 13. D.A. Dixon and B.E. Smart, Chem. Eng. Commun. 98, 173 (1990). 14. F. Burgess, Jr., M.R. Zachariah, W. Tsang, and P.R. Westmoreland, "Thermochemical and Chemical Kinetic Data for Fluorinated Hydrocarbons," NIST Technical Note 1412, U.S. Department of Commerce Washington, D.C. (1995). 15. K.D. Dobbs, D.A. Dixon, and A. Koornicki, J. Phys. Chem. 98, 8852 (1993). 16. Pitts et al., NIST Technical Note 1279 (1990), pp. 4-8 17. M.D. Smooke and V. Giovanagigli, Int. J. Supercomput. Appl. 5, No. 4, 34 (1991). 18. E.S. Domelski and E.D. Hearing, J. Phys. Chem. Ref. Data 22, 805 (1993). 19. An Appraisal of Halogenated Fire Extinguishing Systems, National Academy of Sciences, Washington D.C. (1972). 20. An Appraisal of Halogenated Fire Extinguishing Systems, National Academy of Sciences, Washington D.C. (1972). 21. Halogenated Fire Suppressants, R.G. Gann, Ed., American Chemical Society, Washington D.C. (1975). 22. C. Huggett, Combustion and Flame 20, 140 (1973). 23. R. Fernandez, private communication. 24. R.S. Sheinson, J.E. Penner-Hall, and D. Indritz, Fire Safety J. 15, 437 (1989). 25. R.S. Sheinson, J.E. Penner-Hall, and D. Indritz, Fire Safety J. 15, 437 (1989). 26. Daudu and M.A. Youker, U.S. Patents 2,005,705 and 2,005,709 (1935). 27. L.E. Manzer and V.N.M. Rao, U.S. Patent 4,766,260 to du Pont (1985). 28. S. Morikawa, M. Yoshitake, and S. Tatematsu, Japanese Patent Appl. 1-319440 (1989). 29. Y. Furutaka, Y. Homota, and T. Honda, Japanese Patent Appl. 1-2900638 to Daikin (1989). 30. B. Cheminal, E. Lacroiz and A. Lantz, European Patent Appl. 638535 to Elf Autochem (1994). 31. P. Cuzzato, A. Masiero, and F. Rinaldi, European Patent Appl. 638535 to Ausimont (1994). 32. J.M. Hamilton, Jr., in Advances in Fluorine Chemistry, Vol. 3, Butterworth, London (1963). 33. D.A. Nelson, U.S. Patent 2,758,138 to du Pont (1956); E.H. Ten Eyck and G.P. Larson, U.S. Patent 2,970,176 to du Pont (1961); H. Niimiya, British Patent 1,016,016 to Daikin (1963); N.E. West, U.S. Patent 3,873,630 to du Pont (1972). 34. S.P. Von Halasz, German Patent 2,712,732 to Hoechst (1978).

OCR for page 6
--> 35. K. Gassen et al., U.S. Patent 5,171,901 to Bayer (1992). 36. S. Morikawa, M. Yoshitake, and S. Tatematsu, Japanese Patent Appl. 1-319440 (1989). 37. M.S. Nappa and A.C. Sievert, U.S. Patent 5,414,165 to du Pont (1954). 38. R.J. Kingdom, G.D. Bond, and W.L. Linton, British Patent 1,281,822 to ISC (1970). 39. W.V. Childs et al., U.S. Patent 5,387,323 to 3M (1991). 40. R.A. Paul and M.B. Howard, U.S. Patent 3,840,445 to Phillips Petroleum (1975). 41. P. Tarrant, J. Fluorine Chem. 25, 69 (1984). 42. C. Krespan, J. Org. Chem. 23, 2016 (1958). 43. NFPA Standard on Clean Agent Fire Extinguishing Systems, P.J. DiNenno, Ed. (1996), p. 65. 44. National Advisory Committee on Acute Exposure Guidelines Levels, U.S. EPA Office of Prevention, Pesticides and Toxic Substances, 60 FR 55376, October 31 (1995). 45. Acute Toxicity (Inhalation), Official Journal of the European Communities, English Edition, Legislation 35, L383A/117, December 29 (1992). 46. Acute Toxicity (Skin Irritation), Official Journal of the European Communities, English Edition, Legislation 35, L383A, December 29 (1992). 47. Acute Toxicity (Eye Irritation), Official Journal of the European Communities, English Edition, Legislation 35, L383A, December 29 (1992). 48. C.F. Reinhardt, A. Azar, M.E. Maxfield, P.E. Smith, Jr., and L.S. Mullin, "Cardiac Arrhythmias and Aerosol 'Sniffing,"' Archives of Environmental Health 22, 265, February (1971). 49. Committee on Toxicology, National Research Council, Toxicology of Alternatives to Chlorofluorocarbons: HFC-134a and HCFC-123, Chapter 2, pp. 13-22, National Academy Press, Washington, D.C. (1996). 50. Repeated Dose (28 Days) Toxicity (Inhalation), Official Journal of the European Communities, English Edition, Legislation 35, L383A/140, December 29 (1992). 51. Organisation for Economic Cooperation and Development (OECD). Subchronic Inhalation Toxicity: 90-day Study. Method 413. OECD Guideline for Testing of Chemicals, 12 May (1981). 52. Organisation for Economic Cooperation and Development (OECD). Teratology. Method 414. OECD Guideline for Testing of Chemicals, 12 May (1981). 53. S.R. Skaggs, T.A. Moore and R.E. Tapscott, p. 99 in Halon Replacements , A.W. Miziolek and W. Tsang, Eds, American Chemical Society, Washington, D.C. (1995). 54. R. Sheinson, ibid., p. 175. 55. R.E. Tapscott, et al., Five reports from Air Force Engineering and Services Laboratory, Tyndall Air Force Base, Florida (1987-1990). Also issued by New Mexico Engineering Research Institute: WA3-31(103.21), SS2.06(1), WA3-79(3.21), WA3-78(3.21), SS2.06(2). 56. W.M. Pitts, M.R. Nyden, R.G. Gann, W.G. Mallard, and W. Tsang, Construction of an Exploratory List of Chemicals to Initiate the Search for Halon Alternatives, NIST Technical Note 1279, U.S. Department of Commerce, Washington D.C. (1990). 57. Evaluation of Alternative In-Flight Fire Suppressants for Full-Scale Testing in Simulated Aircraft Engine Nacelles and Dry Bays, R.G. Gann and W.M. Pitts, Eds., U.S. Department of Commerce, Washington D.C. (1994). Note particularly Chapter 5: "Flame Inhibition Chemistry and the Search for Additional Fire Fighting Chemicals," pp. 467-641. 58. Halon Replacements, A.W. Miziolek and W. Tsang, Eds., American Chemical Society, Washington D.C. (1995). 59. T.A. Moore, C.A. Weitz, and R.E. Tapscott, pp. 551-567 in Proceedings of the Halon Options Technical Working Conference, Albuquerque, N. Mex., May 7-9 (1996).