Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
AN INTRODUCTION TO COMBUSTION CHEMISTRY W. G BERL Applied Physics Laboratory, The Johns Hopkins University Introdwtion The question we are asked to discuss is* What major contributions to reducmg fire losses and protection costs have come and can be expected to come from insights into the chemtcal aspects of fires' Also unphed m this question is a corollary if areas of ignorance exist, should they be cleared up by the fire research commumty or can one expect their solution to come from elsewhere' I raise this pomt because only a small fraction of the research on the physical and chemical prmciples of fires has been directly supported thus far by sources concerned with fire suppression. For example, what is known about the structure of premixed flames was obtained largely m support of jet engme development; understandmg of ignition of matenals by radiation was a consequence of work on defense against nuclear attacks; and the charactenzation of materials with respect to their behavior m fires is largely the responsibihty of chemical manufacturers. Five participants will address themselves in some detail to specific chemical aspects of fires. My purpose is to stress what one might achieve m practice with an orgamzed, close look at "Fire Research" topics, particularly those related to urban/mdustnal circumstances. Understandmg of prmciples, one feels, is essential in order to permit prediction of risks; to assess the adequacy of responses and to compare their effectiveness with what is theoretically required; to minimize un- necessary losses or meffective prevention expenses; to reduce the scale of the counterattack in extinguishment; and, not least, to permit operations with reason- able safety m environments that are potentially hazardous. Fire problems are largely chemical-engmeermg m nature. The basic phenomenon is a chemical reaction, strongly coupled with physical (flmd and heat flow) inter- actions which often may be dommant m determimng the course of the fire. Smce so many chemical substances, simple and complex, qualify as fuels, an enormous variety of special situations exist. However, the goals of a research effort are common, i.e., to develop principles of prediction (how will a fire develop?); to explore principles of initiation and suppression (how can a fire be prevented or extinguished?); and to set down design cntena that wiU lower fire risks and fire damage. Added to these science/engmeermg-oriented problems are other important facets concerned with (1) the analysis of the fire services operations; (2) the training of professionals and the education of the pubhc; (3) the economics of alternate solu- tions; (4) the acquisition and evaluation of meanmgful statistics. Each topic represents an essential support to the ultimate goal of more rational control of fires. Being closely related to actual fire situations, they deserve detailed attention. Comhmtion Reactions I will describe, on the basis of a model fire situation, where chemical influences play a significant role and where physical effects predommate. The question I want 153
154 F I R E E E S E A R C H to ask is: Is it possible to predict the tune history of a fire from information that might be found m an appropriate design handbook? I want first, however, to place fires mto the more general contejct of reactions grouped together under the terms "Combustion and Oxidation." One cannot help but be astonished by the extraordmary subtleties which nature (and man) has discovered to convert carbon/hydrogen compounds mto (mostly) H2O and CO2 and to utihze the energy hberated m this process for useful purposes Most basic (and chemically most complex) is the process of "respiration" m plants and animals by which carbon/hydrogen contammg substances (protems, carbohydrates, fats) are ultimately converted to carbon dioxide and water (and a small number of "waste products") while simultaneously generatmg heat and an untold number of compounds of essential use to the hvmg orgamsms. This oxida- tion, takmg place efficiently only near room temperature, represents the largest chemical conversion process on earth and is still several orders of magmtude larger than the controlled and uncontrolled chemical combustion reactions that occur at higher temperatures. In these metabohc reactions the fuel substrates are converted by numerous small interlocked steps, with oxygen addition and water elmunation, mto simple carbonyl- containmg orgamc acids, 3-6 carbon atoms m length. Carbon dioxide is ehmmated with the help of specific enzymes from some of these acids (mainly the jS-keto acids) and replaced, m turn, by acetate, a common product of the "m vivo" degrada- tion of carbohydrates, fats or protems. The oxygen, in turn, is reduced to H2O m a number of steps mvolvmg iron m various valence states, bound to coenzymes. The energy hberated m this set of reactions is, m turn, transferred to the fuel degrada- tion reactions which require it m a number of endothermic steps This set of reactions m hvmg organisms leads to the same products as the combustion reactions, but the detailed pathway bears no resemblance whatever At more elevated temperature (a few degrees above room temperature for the highly reactive boron hydrides, and a few himdred degrees for hydrocarbons) a new oxidation regime exists m which reactions proceed at reasonable rates, ac- compamed by "wave" phenomena in which flow processes and chemical reactions couple. In such "cool" flames the emphasis is still predommantly on complex chemical reaction pathways, although they are much less mtncate than respiration. In the past 40 years, msights mto these oxidation pathways of hydrocarbons have grown substantially (thanks, m large part, to the development of refined analjdiical techmques) so that the mterlocked reactions and product distributions are be- commg reasonably well understood. The reactions are still so varied and mterwoven (74 products having been identffied in the oxidation of 3-methyl pentane") that a prediction of rates of product formation from flrst prmciples is not possible. I t should be noted that the final products, although containmg some CO2 /CO /H2O, are far removed from the thermochemically most stable chermcal equihbrium state One may wonder about the practical value of such studies. They have had a profound effect on the large-scale synthesis of many chemical compounds by partial oxidation from hydrocarbon raw material. Quite unexpectedly, they have also given insight into air pollution problems from the mcomplete oxidation of hydro- carbons in practical devices since several of the reaction products will, with oxygen, react further photochemically to form "smog " A gap of several hundred degrees exists between "hot" flames m which oxidation reactions go essentially to completion and the "cool" flames just described. This is
ABSTRACTS AND R E V I E W S 155 due to experimental difficulties which do not permit sufficiently rapid heating and cooling to study the behavior of complex systems under isothermal conditions, although shock tube mvestigations are beginning to fill the gap.' In the regime of "hot" flame reactions we distmgmsh two pathways for trans- formmg reactants mto products- In "premixed" systems m which fuel and oxidizer are mixed on a molecular scale (or, as m monopropellants, exists m the same mole- cule) flame propagation is made possible m a steep temperature and composition gradient by an elegant combmation of heat flow and diffusion from the hot com- bustion products mto the unreacted mixture, coupled with chemical reactions within the combustion wave. The pathways are fewer m number and complexity com- pared to the "cool" flames where small changes m condition can have profound effects on product distribution Premixmg need not be confined to the molecular level. Nonhomogeneities com- parable m size with the reaction zone thickness will behave similar to the homogene- ous case. As the "unmixedness" of fuel and oxidizer mcreases and, an added com- plication, when hquid or sohd fuel phases are involved, then mixmg, heat transfer and radiation contributions become an mcreasmgly important part of the reaction and frequently dommate the rate of the combustion process. In most fire situations we deal with such coarsely distnbuted fuel-oxidizer mixtures. As a consequence, m normal fires the chemical reactions do not control the rate of the overall conversion. Thus, many of the phenomena of on-gomg fires (particularly their propagation and steady-state burning rates) are only weakly coupled with the chenustry of the conversion reaction and depend mstead on the mtrmsic nature of the reactants, their thermochemical properties, the mixmg and heat transfer conditions, the degree of subdivision, etc. We will deal with specific examples m the next section Ftre Situation The fire situation I want to discuss consists of a compartment that contams* (1) A relatively easily ignited fuel source (which may be a hquid fuel m a pan; or a contmuous array of sohds, such as paper strips; a discontmuous array of sohds, such as vertical fibers, or mterconnected sohds as exemplified by wood cribs); (2) A flat ceilmg of a combustible matenal (wood), (3) A floor of similar matenal (fireproofed by a chemical treatment). The compartment has wmdows and ventilation ducts. I t is equipped with fire de- tecting devices, contains a supply of fire mhibitors (chemical or water) and a hvmg orgamsm A fire m such systems would have the followmg history An energy mput of adequate size near the fuel-air mterface mitiates a flame. This flame will propagate across the fuel source. After reachmg the fuel boundary a steady-state fire situation wiU result The products of combustion will leave the fuel source, formmg a convection column. This column will mteract with the combustible ceilmg, be deflected and igmte the ceihng matenal. A flame will be formed under the ceilmg which, m turn, can igmte the combustible floor. The rates of fuel consumption, the nature of the combustion products and the pattern of gas flow will be mfluenced by the wall openmgs, the dimensions of the structure and the fuel source, which, in turn, affect the effectiveness of the coun- termeasures.
156 F I R E R E S E A R C H The questions I want to explore are: What can be said about the speed of progress of a fire m such a situation? Can effects of changes m fuel arrays be predicted? How do vanations m compartment dimensions affect fire progress? Can the tem- peratures and the gas compositions inside the compartment be predicted? What are the effects of a fire on the organism? I t will be no surprise to learn that an overall analysis is not possible, neither now nor very soon. A promismg beginnmg has been made on some of the mdividual steps. But many design parameters are nussmg as well as knowledge of the mteractions among the various steps. I will summarize where work has been done or is m progress on the component parts of the problem and emphasize those aspects where chemical parameters appear to be important or dommant. Ignition Igmtion energy requirements for homogeneous gas mixtures are known for a vanety of conditions as are radiant energy requirements for the ignition of sohds.* While the detailed chemical and physical pnnciples responsible for the establish- ment of a propagatmg combustion "wave" are only moderately clear (for example, the steps m the igmtion of a radiation-igmted sohd are not even quahtatively under- stood), the energy requirements are well known. Thus, much information on mi'Tiimiim electric spark or radiation ignition requirements is on hand and techniques for measuring them are well established although httle has been pubhshed about energy requirements in the presence of combustion inhibitors or retardants. If igmtion is due to convective or conductive heat transfer, the available design m- formation is much more limited. In view of the convenience of direct measurement, no strong motivation exists to pursue a "first principles" approach m order to solve practical igmtion prob- lems. However, smce the chemical characteristics of the fuel (and of the oxidizer) enter strongly mto the igmtion energy requirements and the igmtion limits beyond which no seif-sustainmg reactions are possible no matter what the size of the igmtion source, a better understandmg of the igmtion process would provide a useful guide for the selection of effective inhibitors or fire retardants. Propagation The rate of flame propagation across a Uquid surface or a soUd fuel array is a complex phenomenon m which chemical parameters play by-and-large a secondary role (except in the case of propagation for hquids of sufficiently high vapor pressure that provide mixtures withm the flammabibty range of the fuel). In the common cases, flame propagation rates are controlled by convective or radiant heat transfer processes, which determine fuel evaporation from hqmds and fuel generation from sohds. Detailed analyses of the propagation of lammar diffusion flames across sohds have been made by McAlevy, by Tanfa and by de Ris.^ The analysis for propaga- tion across thin and thick fuel slabs is based on the assumption that the gas phase chenucal reaction is not rate lumtmg except near and at the propagation limits. Generation of combustible gases from the sohd is caused by heat transfer m front of the diffusion flame which mtroduces conductive heat transfer, radiation, fuel
ABSTRACTS AND E E V I E W S 157 vaponzation, the thermal properties of the fuel and the temperature of the flame mto the analysis With hqmds an additional flow process of convective mmng of heated fuel with cold hqmd occurs ahead of the flame' which mfluences the vapor generation rate No analysis of more complex fuel arrays (such as sohd fuel cnbs) has as yet been made, but chenucal reactions appear adequately fast and do not set the limit on the rate of flame propagation, as long as the system is situated well withm the flammabdity regime As m the igmtion case or the steady-state burning (next section) near the propagation linut, chemical factors can play a very important role. Steady-State Burmng The bummg of hqmd fuel pools and of sohd structures, once flame has spread from a locahzed igmtion source to the fuel boundary, accounts for the bulk of the heat generation and the nature of the combustion products m typical fire situations. The studies of Bhnov and Khudiakov on the rate of bummg of hquid fuel pools have recently been extended and reanalyzed.' There is httle doubt that, m the re- gion where bummg proceeds the chemical processes are too rapid to be rate hmitmg. Fuel generation and the mmng of fuel vapor and air determme the overall fuel consumption rates Scale effects are pronounced since transitions from laminar to turbulent flow affecting the imxing of fuel and oxidizer, set in with pan dimen- sions of practical sizes Also, the heat transfer mechanism to the fuel bed changes from convection to mixed convection-radiation with mcrease m scale due to changes in the optical thickness of the radiatmg gas column. I f combustion takes place withm an enclosure contammg a limited amount of air but which commumcates with an unlimited air supply via flow-hmitmg open- mgs (wmdows or ducts) the bummg rates wUl be strongly affected by the "ventila- tion rate" as set by these constnctions.* Very pronounced effects on bummg rates, gas composition and circulation withm the enclosure have been quahtatively observed. While the flmd dynamic and heat transfer effects play a dommant role m most situations of common fire experience one should not lose sight of the fact that such fires can be extmgmshed by mterfermg with the evaporation process (by coohng), by lowermg the oxygen partial pressure (by mert gas admixture or lowenng of total pressure) or by the addition of inhibiting chemicals. The two last interferences clearly affect the chemical reaction rates by modifymg the balance of heat genera- tion m the gas phase and heat transfer to the fuel. These limit conditions, so important m the practical extmgmshment of fires, are not now amenable to analysis nor will there be a gmde to further development unless the mechamsm of combustion limits (mcludmg the dommant mechamsms m the absence and presence of chemical mhibitors) are understood. The availabihty of counterflow diffusion flames' has opened up a promismg avenue for such studies leadmg to extmgmshment under reahstic conditions. I t has already been demon- strated that diffusion flames can be extmgmshed either by thermal quenchmg m the presence of a heat smk or due to chemical limitations on the combustion rate. Those mterestmg studies could readily be extended to mvestigate the influence of chemical inhibitors and, thus, lead to a much better understandmg of the events takmg place at or near the limits of flame propagation.
158 FIRE HESEABCH Convection Column In the absence of confining walls, the properties of turbulent buoyant colmnns above a source of heat are reasonably well understood. Sunilarly, the composition and temperature profiles of convection columns above bunung fuels have also been adequately analyzed.'" An mterestmg problem m fire propagation arises rf the combustion takes place inside an enclosure. The flow of hot gases under a potentially combustible ceihng of the enclosure has an important bearmg on two important items it determmes the response of devices mstalled to detect the presence of fires in the enclosure; it also determmes how rapidly igmtion of the ceihng due to convective heatmg will take place. Prehmmary studies are being made" on the time-temperature and velocity of buoyant pliunes unpmgmg on horizontal surfaces This is normally an aerodynamic process, without chemical mputs However, as soon as the ceihng material begms to decompose the processes leadmg to eventual igmtion are strongly affected by the chemical properties of the ceilmg, and by the chenucal composition of the buoyant plume. Celling Fires A steady flame, bummg below a honzontal combustible ceilmg, is a diffusion flame m which flmd dynamics and heat transfer to the fuel determme the bummg rate, although limits on chemical rate may be responsible for extmction." Such celling fires have been mvestigated by de Ris who presents good evidence that turbulent free convection from the flame to the fuel is responsible for bummg rates, which are set by the mass-transfer driving force that characterizes the heat flow from the combustion gases to the fuel source, takmg mto consideration the effect of mass addition from the decomposmg ceihng on the transport properties. Such a flame is a source of radiation which may lead to further fire spread on surfaces (floor, side wall) which, m turn, will undergo thermal decomposition and igmtion. Igmtion at a distance is commonly observed (flashover) and may be mitiated either by flame propagation in the gas phase if sufl&cient combustible gases accumulate in the enclosure or may be caused by igmtion at or near the heated surfaces as products from thermal decomposition mix with the ambient atmosphere. Detectors A large number of detector designs are conceivable and have been built and tested. Those that depend on measurements of physical constants such as tempera- ture, temperature gradients, hght emission, are, of course, unmfluenced by the chemistry of the combustion reaction. In view of the presence of large amounts of CO m most fire situations, its detection is a good clue of potential fire hazards and is apphcable m situations where direct physical measurements of flame proper- ties are impractical. This is particularly tme for large buildmgs with many mter- connected cubicles, makmg use of a common air supply. Other chemically-related indications of fires are based on the detection of smoke or of odors typical of thermal decomposition products of common buildmg materials.
ABSTRACTS AND REVIEWS 159 Inhibition and Extinction Existing fire can be brought under control if ignition and propagation are pre- vented. In the present example this can occur in several ways and by several modes The effect of water is predommantly through its mterference with the gasification of sohd fuels It also can affect the temperature of the fire convection column and, thereby, mterfere with the igmtion of ceihngs or waUs by convection. Water drop- lets, m addition, may reduce radiant energy transfer from flames to umgmted fuels. It IS doubtful that water vapor has a substantial specific effect on the chemical reactions m the gas-phase combustion zone. Chemical mhibitors, on the other hand, exhibit their mfluence largely by mter- fenng with the chemical reaction m the gas phase" and may also mteract with the reaction mechamsm m the sohd fuel as it decomposes The behavior of fire retardants is still poorly understood Some exert their mfluence m the gas phase to which they are transported with the combustible constituents of the fuel." Others produce a physical barrier that alters the heat flow to the fuel. The effects are predommantly chemical m nature. Chemical Effect of Fires on Livmg Bemgs Exposure of hvmg beings to hot gases can be lethal or produce serious burn mjuries. Many fatahties or severe sicknesses can also be produced by exposure to the chemical products of combustion The state of knowledge m this field is rudi- mentary. While the toxic effects of carbon monoxide are well estabhshed, the possible mjurious effects of other products of combustion are not weU documented. This IS, in part, due to the difiiculty to specify the chemical compositions of the products of combustion of thermally decomposmg (but not burmng) sohds, or of complex fuels bummg m oxygen-Iimited atmospheres, particularly under conditions where the chemical reactions do not reach eqmhbnum temperatures and compositions. Conclusions What conclusions can we reach' "Normal" fire behavior is strongly affected by the flmd dynamics and heat flow parameters of the combustion, by the physical properties of the fuels and by the dimensions of the fire and of the structure withm which it IS contained Propagation across surfaces,' m ducts," across arrays of various sizes and subdivisions," faU mto this category, as do the properties of convection columns, the mfluence of wmdows, etc. Chemical reactions m the gas phase are too rapid to have any limitmg effect on the overall process. However, flame imtiation (igmtion) and termmation (suppression) are strongly affected by the nature and rates of chemical reactions Sunilarly, the nature and generation of combustible gases from sohds is strongly mfluenced by chemical factors. Their unquestioned complexity makes it a difficult task to elucidate their effects. However, they should be pursued if the goal of fire research is to minimize fire occurrences and to maximize the effectiveness of fire suppression and retardation. References 1 KAHLSON, P Inlroduclion to Modem Biochemistry, Academic Press, 1965 2 BABAT, P , CuLLis, C F AND POLLARD, R T ⢠"Studies of the Combustion of Branched-Chain
160 PraB RESEARCH Hydrocarbona," Thirteenth Symposium (International) on Combustion, p. 179, The Combustion Institute (1971). 3. JACHIMOWSKI, C . J . AND HOUGHTON, W. M.: "Shock-Tube Study of the Initiation Process in the Hydrogen-Oxygen Reaction," Combust. Flame 17, 26 (1971) 4. MARTIN, S "Diffusion Controlled Igmtion of Cellulosic Matenals by Intense Radiant Energy," Tenth Symposium (Intemattonal) on Combustion, p. 877, The Combustion Institute (1965) 5. TABIPA, G . S , D E L NOTARIO, P P . AND TORRALBO, A M.: "On the Process of Flame Spread- mg over the Surface of Plastic Fuels in an Oxidizmg Atmosphere," Twelfth Symposium {Inter- Tiational) on Combustion, p 229, The Combustion Institute (1969). D B RIS, J . : "Spread of a Lammar Diffusion Flame," ibid, p 241 LASTRINA, F . A , MAGBE, R S AND MCAIBVT, R . F . I l l : 'Tlame Spread over Fuel Beds: Sohd-Phase Energy Considerations," Thirteenth Symposium (International) on Combustion, p 935, The Combustion Institute (1971). 6 SiRiGNANO, W. A AND CLASSMAN, I ⢠'Tlame Spreadmg above Liquid Fuels, Surface-tension- dnven Flows," Combust Sci Technol 1, 307 (1970) BuRQOTNB, J . H AND ROBERTS, A F . "The Spread of a Flame across a Liqmd Surface I . The Induction Period," Proc Roy Soc 308, 39 (1968), "II Steady-state Condition," ibid. 308, 56 (1968), "III A Theoretical Model," ibid 308, 69 (1968). 7. BLINOV, V I AND KHUDIAKOV, G N . Diffusive Burning of Liquids, Izdatel'stevo Akademu Nauk SSSR, Moscow (1961) D B RIS, J D AND ORLOFF, L Combustion and Flame (to be published) CoRLETT, -R C-ibid U, 351 (1970) 8. HESSELDEN, A J M , THOMAS, P H AND LAW, M.* "Burmng Rate of Ventilation Controlled Fires m Compartments," Fu-e Technol 6, 123 (1970) GROSS, D AND ROBERTSON, A F "Experimental Fires m Enclosures," Tenth Symposium (IntenuUionaT) on Combustion, p 931, The Combustion Institute (1965). TEWARSON, A Pnvate Commumcation. 9 TsuJi, H AND YAMAOKA, I "Structure Analysis of Counterflow Diffusion Flames m the For- ward Stagnation Region of a Porous Cyhnder," Thirteenth Symposium (International) on Combustion, p 723, The Combustion Institute (1971) MiLNB, T . A, GREEN, C L AND BENSON, D K ⢠"The Use of the Counterflow Diffusion Flame m Inhibition Effectiveness of Gaseous and Powdered Agents," Combust. Flame IS, 255 (1970) 10. THOMAS, P . H ⢠"The Size of Flames from Natural Fires," Ninth Symposium (International) on Combustion, p 844, The Combustion Institute (1963). 11. FRIEDMAN, R ⢠Pnvate Commumcation 12 ORLOFF, L . AND D E Rra, J . : "Cellular and Turbulent Ceihng Fires," Combustion and Flame (to be pubhshed). 13. FRISTROM, R M AND SAWYER, R . F . ' 'Tlame Inhibition Chemistry," AGARD 37th Pro- pulsion Panel Meetmg, May 1971. 14 FBNIMORE, C . P. AND MARTIN, F . J . : "Modes of Inhibiting Polymer Flammabihty," Combust. Flame 10, 295 (1966). 15 D B RIS, J ⢠"Duct Fires," Combust Sci Technol. 2, 239 (1971) 16. EMMONS, H W AND SHBN, T . 'Tire Spread in Paper Arrays," Thirteenth Symposium (Inter- national) on Combustion, p. 917, The Combustion Institute (1971)
