accidents in these groups, including fabric flammability standards, redesign of ignition sources and public education. However, it is important to remember that fabric flammability standards alone won't solve the problem. For example, the apparel fire accident problem for males of ages 6-65 is closely tied to the use and misuse of flammable liquids. Since this is largely a behavioral problem and is difficult to approach from a technological basis, it would seem that a broadly based educational program may be the most effective means of preventing this type of accident. Further investigation is needed to effectively determine the hazard involved in various types of fire accidents and to consider the cost-benefit aspects of the various means of protecting people from fire injury. In addition, more detailed information is necessary for the development and implementation of effective educational programs. In relation to these needs, further analysis is in progress to define activity, reaction and garment parameter patterns in greater detail and to determine the relationship of these factors to each other and to the severity of the resultant burn injury. NATIONAL BUREAU OF STANDARDS SPECIAL PUBLICATION 411, Fire Safety Research, Proceedings of a Symposium Held at NBS, Gaithersburg, Md., August 22, 1973, (Issued November 1974) CHEMICAL ASPECTS OF FLAME INHIBITION1 John W. Hastie National Bureau of Standards, Washington, D.C. The role played by inorganic chemical additives in fire retardancy and flame inhibition is considered. Particular attention is given to the molecular level aspects of commercially important systems containing compounds of antimony and halogens. For purposes of discussion, consider a fire where the fuel is supplied by pyrolysis of a polymeric substrate, such as a nylon or polyester carpet. The gases in the immediate vicinity of the carpet will usually consist of a fuel-rich mixture of pyrolysed and vaporized hydrocarbons together with entrained air. At a somewhat greater distance from the decomposing substrate, this fuel-air mixture may support combustion in the form of a flame. Now, it is well known that one can extinguish such a fire, either by adding chemicals to the polymer substrate or externally throwing chemicals at the fire such as with an extinguisher containing halocarbons or alkali metal carbonates. The present discussion is concerned with the mode of inhibition that occurs when chemicals are added to the substrate. What we have to find out is how these chemicals are released from a pyrolysing substrate and then how the released molecular species interact with the flame reactions. The important reactions in flames are radical reactions and one can chemically affect the flame propagating radicals which are mainly hydrogen atoms, hydroxyl radicals and oxygen atoms. By reducing the steady state radical concentration, we can inhibit the flame and even extinguish it. So that is the problem that we are addressing ourselves to. I should mention that the main chemicals that are currently being used by industry are compounds of antimony, phosphorus and chlorides or bromides. We have been studying the mechanisms by which compounds containing these elements inhibit fires, but in this paper I am just going to discuss the antimony system. There are two modes by which such chemicals can affect fires. They can alter the course of decomposition of the polymer such that a less flammable mixture is produced, or they can release chemicals to the vapor phase which actually poison the flame, and it is this latter mode of action which I am going to emphasize. The antimony oxide halogen system is usually comprised of a mixture of antimony oxide solid and an organic halogen compound such as polyvinylchloride or some chlorinated paraffin. This mixture is added to the substrate that we want to fire retard. It has widespread use in industry; however, the mechanism by which it operates as a fire retardant is not very well understood. We have, therefore, chosen this as a model system for extensive study. At the molecular level, without going into technical details, we are able to monitor the pyrolysis chemistry by using a mass spectrometer as a detector for the different vapor species. Our principal observations from this kind of research are that the hydrogen chloride which is released from the decomposing organic chloride additive interacts with the antimony oxide solid over a temperature interval of about 250-500°C, which is an interval characteristic for 1A more extensive discussion of the subject matter dealt with here may be found in the following reference: Hastie, J. W., Molecular Basis of Flame Inhibition, J. Res. Nat. Bur. Stand. (U.S.), 77A (Phys. and Chem.), No. 6, 733-754 (Nov.Dec. 1973). the pyrolysis of many commercial polymers. The main products of this interaction have been found to be gaseous antimony trichloride and water vapor. Also, during the course of this interaction we find that solid oxychlorides are formed. These oxychlorides really control the rate at which the antimony trichloride is released to the vapor phase. This is demonstrated in figure 1. The curves A, B and C give a measure of the rate of release of antimony trichloride to the gas phase at different temperatures. We have started out with an antimony oxychloride solid and gone through a heating sequence with a very small rate so that we could attempt to approach an equilibrium condition. The important thing about this observation, i.e., figure 1, is that the antimony trichloride is released in the form of three bursts, rather than being released all at once. If the trichloride was released just at one temperature, like 250°, we would rapidly lose all of the additive that had been introduced into the polymer, and the chances of the system being an effective fire retardant would be reduced because the extinguishant is only being released over a very narrow temperature interval which may not correspond with the temperature at which the polymer is decomposing. The rule of thumb here is that we want the vapor active ingredient to be released at about the same time as the polymer is producing flammable fuel, and in many cases that is up in the temperature interval of about 300 to 400°C. To summarize the results of an extensive study on this system the release of antimony trichloride is controlled by the formation of solid oxychloride phases which are listed in table 1. This is fairly peculiar to the antimony oxide system. One would not expect to find a similar behavior in many other oxide-halogen fire retardant systems that one could conceive of. Figure 1. Pressure variation of the vapor species SbC13 (curves A, B and C) (curve D) as a function of temperature and time using solid Sbocl as a source and a heating rate of about 0.2°C min-1. and Sb.O 406 A B Table 1. Antimony Oxide Halogen Substrate Reactions ~550°C Now, let me address the problem of how this antimony halide vapor actually affects the flame. In studying the chemistry of these additives in flames, there are two approaches that we can use, either a macroscopic level of observation which would involve measuring such things as flame speed, temperature, combustion limits and initial and final composition or we can look at the microscopic level which involves looking at all the molecular species present in the flame, and for this we would use optical and mass spectroscopic techniques. The important thing here is that if we can characterize the microscopic character of these flames and also the additives in the flames, then by use of kinetics and thermodynamics we can actually calculate these macroscopic properties. If this calculation agrees with the results of macroscopic measurements, then our model for inhibition which has been developed by the microscopic route is most likely a correct one. This is the approach that we are working towards. Table 2 gives a rating of chemicals as flame inhibitors. The large numbers mean that only a small amount of inhibitor is required to reduce the velocity of a particular laboratory flame. Note that antimony trichloride has a rating of about 26, relative to say, a value of only 0.8 for carbon dioxide. Thus, there are many chemicals that are much more effective than chemicals which are currently being used as extinguishants and, in particular, some of the heavy metal systems are extremely effective. It is conceivable that such systems may be used in future retardant formulations. By looking at the molecular level processes of these chemicals in flames, one can eventually understand why an individual chemical has the particular effectiveness that it does on flame speed. Now, in order to get at the molecular basis, we need to develop a system for analyzing flames at one atmosphere and to be able to extract very active species, such as radicals, from these flames. An apparatus developed for this purpose is shown schematically in figure 2; basically, it is a molecular beam mass spectrometer. The sampling of flames is carried out through a small pin hole in the tip of a conical probe into a vacuum system. The gas then freezes into a molecular beam-form and passes into the mass spectrometer where it is mass analyzed. The position of the burner, and hence the flame, can be varied relative to the sampling probe. The mass spectrometer then tells us what the distribution of molecular species is in a laboratory flame. Typical results are given in figure 3. The zero is the burner tip and the luminous reaction zone for this flame is at the 9 mm distance position. From this, we see that one can monitor the decomposition of the fuel, the depletion of oxygen and the appearance of combustion products. More importantly, using molecular beam sampling, we can look at the distribution of radical species which are controlling the speed of these flames as shown in figure 4. The most important species in fuel-rich hydrocarbon flames are the hydrogen atom, hydroxyl radical and the methyl radical. Note that the hydrogen atom appears well upstream of the flame. It is diffusing back upstream and most of the pre-flame chemistry is probably the chemistry of this hydrogen atom. |