Fire research abstracts and reviews: Volume 13, 1971 (1971)

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.

CHEMICAL THERMODYNAMICS AND FIRE PROBLEMS DANIEL R. STULL Thermal Research Laboratory, The Dow Chemical Company Introduction Throughout history there has been a contmuing struggle to subordinate fire to the service of man, but the mastery is mcomplete The usual concept of a fire is the umon of a fuel with oxygen, but m its most general aspect, we must regard fire as a chemical reaction The study of chemical reactions is a relatively recent activity of man that has provided many benefits, but agam the mastery is mcom- plete, smce we still expenence reactions that rage out of control A fire is a special kind of reaction; it is spontaneous, it evolves heat, it is not reversible, and the reaction products are generally smaller molecules and of greater number than are the reactant molecules. The past few centuries have witnessed a continually accelerating growth of chemical manufacturmg, and of the study of chenucal reactions Thus, for milhons of years fires mvolved the combustion of natural matenals grass, wood, oils, fibers, foodstuffs, and the like. Chemical manufacturmg mtroduced chemical fires feedmg on man made chemicals and resultmg in a larger number of fire types and generatmg a wider range of fire problems. For example, the combustion of wood m air has happened often enough so that there is a reasonable body of expenence available to aid m extmgmshmg such a blaze. On the other hand, a leak m an iron pipelme transporting chlorme under pressure has been known to result m a fire which generates copious clouds of feme chlonde as the pipe bums away. In this case, expenence in extmguishing such a blaze is limited. The corrosiveness of the reaction products raises some new problems for the first time The world of today witnesses an increasing variety of fires involving both natural and man made matenals in combmations that are mcreasmgly complex. The job of the modem fire fighter involves somethmg more than soakmg the blaze with water to cool it below the ignition temperature. It is necessary to know what matenals are burning in order to properly extmguish a complex fire. It is necessary to have some idea of the degree of reactivity of the reactants so the fire fighter can do his work at close range or maintain a safe distance if the reactants may detonate. It IS necessary to know if the products of the fire are toxic, or present some un- healthy condition to the humans who are exposed. Society must also decide how much safety it is w i l l i n g to pay for, because it costs money to fight fires and brmg them under control. After all, fires or chemical reactions proceed according to natural laws, most of which are known, but have not yet been apphed in sufficient detail. It is my behef that this is one area where progress can be made and the rate of progress will be related to the amount of national treasure directed at this problem. 161

162 PIBE RESEARCH The Significance of Thermochemistry The word THERMODYNAMICS denotes the dynamic or mechamcal action of heat, while the word THERMOCHEMISTRY expresses the relations existmg between chemical action and heat The laws of thermodynamics and thermo- chemistry are Imked together and govem the behavior of all the tangible matenal m the umverse The first and second laws of thermodynamics were developed and proven dunng the nmeteenth century while the laws of thermochemistry are a development of the twentieth century' which partially accounts for the fact that they have not been as widely apphed as is fully warranted. Thermochemistry is a quantitative science m which a man of action can have the satisfaction of domg somethmg no one else has done, and of makmg a tme contnbution to world knowl- edge. The work area is the frontier where what is known meets that which is not known. It is on this frontier that new knowledge can be put to work, and where problem-solvmg takes place As new experimental thermochemical data are gen- erated, new msight is brought to the fields studied. Note too, that thermochemistry rests on the measurement of heat and that calonmetry has been adequately de- veloped to meet our current needs The Basic Thermodynamic Properties The basic thermodynamic properties are energy E, entropy S, pressure P, absolute temperature T, and volume V. The forces between the atoms that compose any kmd of matter are related to the energy, while the arrangement of those same atoms is related to the entropy. Quantitative information on the behavior and stabihty of a chenucal substance can be obtamed by hnkmg the energy and entropy with the absolute temperature, while these three properties can be derived entirely from well known laboratory measurements.* These thermodynamic quantities are related thus The Gibbs energy (G) is defined as G=E+PV-TS, (1) while the enthalpy or heat content (H) is defined as H^E+PV; (2) hence, G = H-TS. (3) When the reactants and products of a constant temperature reaction are m their standard states, we can wnte AG° = AH''-TAS°. ( 4 ) A general reaction at equihbrium can be symbohzed aA+bB ^ cC+dD (5) where A, B, C, and D represent the thermodynamic concentrations of the com- ponents, and a, b, c, and d mdicate the number of moles present at eqmhbnum. The eqmhbnum constant {K) for this reaction is given by ^ (activity of C)°(activity of Py (activity of yl)''(activity of B)'*

ABSTRACTS AND REVIEWS 163 This equihbrium constant K for the reaction is related to the standard Gibbs energy change by -RT]is.K=LG° = ̂ H''-T ^S°, (7) where R is the umversal gas constant. Eearrangmg, we have hi K = - AHyRT+AS°/R. (8) This relationship shows that the atoms present m a reaction will prefer the molec- ular configurations m which the entropy is maximized and m which the energy is algebraically mmmuzed Maximum entropy is achieved by the molecular configurations with the largest number of states available to the system, thus providmg more "freedom" for the system. Mmimum energy is achieved by the molecular structures in which the atoms are most strongly bound to each other, thus providmg the maximum "sta- bihty" for the system. In evaluating the equihbrium constant K for a reaction, the result IS a compromise between these two opposmg factors* stabihty and freedom. The stabihty term is more dommant at low temperatures, but at high tempera- tures the equihbnum is more heavily mfluenced by the freedom term AS°. Apphca- tion of these thermochemical prmciples to real problems requires the thermodynamic data for the substances mvolved m the problem. Rapid Computation of Chemical Equilibria The past two decades have been marked by the very active development of elec- tromc digital computers. Their present state of development permits the detailed solution of many problems that were formerly regarded as too extensive to con- template. Today, lengthy computations of reaction equihbna based upon chemical thermodynamic data are commonplace ' Costs of such calculations are low enough to permit widespread employment of this technology. A descnption of the capabih- ties of a specific computer procedure will give a better idea of what can be ac- complished. In a specific chemical equihbnum problem, the reactants mteract to form the natural products which at equihbnum will turn out to be those products possessing the largest Gibbs energy change from the reactants. The mput to the computer consists of (a) the number and kmd of atoms m each mole of reactant, (b) the heat of formation for each reactant, (c) the quantity of each reactant present, (d) the pressure and temperature where the equihbna are to be calculated, or the maximum temperature that the reactants can achieve, (e) an estimate of the maximum temperature reached (Tmax), and (f) the name of each reactant. Stored m the computer is a massive list* of possible products with a complete tabulation of thermodynamic properties of each product. The computer matches the elements in each reactant against the elements in each product species m the stored product list and discards the species havmg elements not m the mput reac- tant set. The remamder constitute a hst of possible reaction products, which are then ordered accordmg to the Gibbs energy at T m o x . The computer then selects

1 6 4 FIRE RESEARCH the most stable products at T m a x and balances the stoichiometry of the reactants and products. Once the product composition is established, the enthalpy of the products from 2 9 8 ° K to T m a x is compared with the enthalpy of the reaction. If these two enthalpies are unequal, the computer selects Qi new T m a x and contmues iteratmg until these two enthalpies are equal. From the final equihbrium composi- tion at 1 atm total pressure, the final number of moles can be calculated, together with the enthalpy and entropy change of the reaction, and the pressure developed by compressmg the products from 1 mole of reactants mto 1 molar volume at r „ a x This IS one smgle set of options that can be calculated by a versatile program at a cost commensurate to one-tenth of a man hour of tune. The Methane-Nitrogen-Oxygen System As an example, detailed calculations of the methane-nitrogen-oxygen system were calculated. Figure 1 shows the flammabihty diagram for the system at 26°C, 1 atm pressure, and dehneates the compositions which are flammable The series of compositions given m Table I were mput to the computer with the mstructions to calculate Tmax and the enthalpy generated for each composition. Figure 2 is a plot of the calculated adiabatic equihbnum flame temperatures; the dotted Ime mdicates the experimental boundary between the flammable and nonflammable mixtures. Note the steep temperature-methane concentration gradients; 2 0 0 ° is eqmvalent to a change of 1 mole % methane concentration. A lean-hmit flame temperature of 1 6 7 5 ° K ± 2 0 0 ° K will define the boundary between flammable and nonflammable compositions. This is m agreement with the temperature found by Fenn and Calcote'm their kmetic studies. They have shown that this lean-hmit boundary is related so closely to the activation energy for methane combustion that practical activation energies for combustion can be derived from the observed adiabatic equihbnum flame temperature of the lean-hmit mixture. This is a very unportant relationship between kmetics and thermodynamics. Figure 3 shows the calculated heat m kcal/100 grams released by eqmhbnum combustion of the compositions m the methane-mtrogen-oxygen system. The highest heat releases occur along the methane-oxygen binary as expected. The Vinyl Chlonde-Air System Chlormated hydrocarbons are widely used for many purposes and are shipped m tank-car quantities. At about 2 : 4 5 P M . , Thursday, September 1 1 , 1 9 6 9 , ' near Glendora, Mississippi, there was a railway wreck in which 1 5 cars were derailed Eight of these were tank cars contammg vmyl chlonde, one of which was ruptured and was reported to be leakmg at 8 P M., but there was no fire At about 9 P M . the vmyl chloride became ignited from sparks caused by a Mississippi utihties crew trying to remove three limbs from nearby power hnes. At 6 : 4 5 A.M., Friday, one car of vmyl chlonde exploded, causmg another car to be ruptured and igmted Due to some erroneous information (that bummg vmyl chlonde would produce phosgene) approximately thirty thousand people were evacuated from theu: homes unnecessarily. Mississippi National Guard troops were called m and martial law was put in force in the area A special Army Team from Ft McClelland, Alabama, Chemical School reported finding no phosgene m the area, and informed the

ABSTRACTS AND REVIEWS 165 L J 60 > 50 a = Stoichiometric for combustion to CO2 and b= Minimum flammable oxygen concentration c = Lower flammable limit in oir d = Upper flammable limit in air F R O M : M. G . Z A B E T A K I S , U . S B U R . M I N E S B U L L . 6 2 7 , ( 1 9 6 5 ) . F I G 1 Flammabihty diagram for the system methane, mtrogen, oxygen at atmosphenc pressure, 26°C Mississippi National Guard that a phosgene hazard did not exist. In the vicinity of the wreck, some Lvestock were found dead by returning residents However, a report issued later indicated that the Lvestock deaths were not phosgene-connected. After all fires were extinguished, of the eight vinyl chloride cars involved, one ex- ploded, two burned, and five remained intact and were recovered. The vmyl chlonde-air system was studied over the whole combustion range. First, vmyl chloride was equihbrated with its thermodynamically most stable products from 300° to 1500°C. The thermal decomposition of vmyl chloride takes

166 FIRE RESEARCH TABLE I Temperature and enthalpy generated by flammable compositions in the methane-mtrogen-oxygen system CHi N, 0, Enthalpy CH, N, 0, Enthalpy (mole) (mole) (mole) (°K) (kcal/100 g) (mole) (mole) (mole) CK) (kcal/100 g) 0 040 0 000 0 960 1237 -24 5 0 200 0 000 0 800 2863 -94 3 0 040 0 100 0 860 1341 -24 8 0 200 0 100 0 700 2858 -95 0 0 040 0 200 0 760 1245 -25 1 0 200 0 200 0 600 2855 -95 7 0 040 0 300 0 660 1250 -25 4 0 200 0 300 0 500 2847 -96 3 0 040 0 400 0 560 1254 -25 7 0 200 0 400 0 400 2814 -95 8 0 040 0 500 0 460 1259 -26 1 0 200 0 500 0 300 2619 -88 2 0 040 0 600 0 360 1264 -26 4 0 200 0 600 0 200 1700 -51 5 0 040 0 700 0 260 1269 -26 8 0 200 0 700 0 100 889 -20 7 0 040 0 800 0 160 1274 -27 2 0 300 0 000 0 700 3033 -117 6 0 040 0 900 0 060 983 -18 9 0 300 0 100 0 600 3009 -117 6 0 060 0 000 0 940 1637 -37 0 0 300 0 200 0 500 2956 -116 1 0 060 0 100 0 840 1640 -37 4 0 300 0 300 0 400 2774 -108 3 0 060 0 200 0 740 1645 -37 8 0 300 0 400 0 300 2129 -78 6 0 060 0 300 0 640 1650 -38 3 0 300 0 500 0 200 1127 -34 4 0 060 0 400 0 540 1656 -38 8 0 400 0 000 0 600 3031 -135 7 0 060 0 500 0 440 1662 -39 4 0 400 0 100 0 500 2869 -128 1 0 060 0 600 0 340 1669 -39 9 0 400 0 200 0 400 2425 -105 0 0 060 0 700 0 240 1677 -40 5 0 400 0 300 0 300 1596 -62 6 0 060 0 800 0 140 1686 -41 2 0 400 0 400 0 200 981 -32 0 0 060 0 840 0 100 1410 -32 5 0 500 0 000 0 500 2615 -130 2 0 100 0 000 0 900 2274 -60 4 0 500 0 100 0 400 1971' -92 7 0 100 0 100 0 800 2268 -60 7 0 500 0 200 0 300 1134 -45 7 0 100 0 200 0 700 2272 -61 4 0 600 0 000 0 400 1524 -75 8 0 100 0 300 0 600 2278 -62 2 0 600 0 100 0 300 1028 -44 2 0 100 0 400 0 500 2286 -63 0 0 700 0 000 0 300 1000 -47 8 0 100 0 500 0 400 2296 -63 9 0 095 0 715 0 190 2223 -62 2- 0 100 0 600 0 300 2307 -64 9 0 086 0 793 0 121 1503 -36 9«' 0 100 0 700 0 200 2282 -64 6 0 050 0 751 0 200 1481 -33 9° 0 100 0 800 0 100 1128 -25 5 0 150 0 672 0 179 1774 -50 6* " Stoichiometric for combustion to COi and HjO. *> Minimum flammable oxygen concentration " Lower flammable limit m air. ^ Upper flammable hmit m air the followmg pattern as the temperature is raised. 2H2C=CHC1 2HC1+CH4+3C At the higher temperatures, the methane cracks CH4 -> CH3+H C + 2 H 2 , formmg methyl groups, atomic hydrogen, diatomic hydrogen, and graphite, and the

ABSTRACTS AND REVIEWS 167 3000 h T ' K 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000 800 80 70 60 50 40 30 20 10 0 MOLE % NITROGEN I 1 I I I I L _ 10 20 30 40 50 60 70 MOLE % METHANE F I G 2 Flame temperatures for the system CH<-Nj-Oj

168 FIRE RESEARCH hydrogen chloride decomposes to HCl H-fCl, formmg atomic hydrogen and chlorine. These relationships are shown graphically in Figure 4 Note that at no time is free diatomic chlorme formed m the presence of hydrogen Atomic chlorme and atomic hydrogen are formed m equal amoimts from the decomposition of hydrogen chlonde and -reach 1 ppm at 1000°C, 10 ppm at 1200°C. The decomposition of methane mto sohd graphite, and hydrogen and methyl radicals is also clearly shown Combustion of vmyl chlonde with stoichiometric air requires enough air to con- tam 2 5 moles of oxygen, which is accompanied by 10 0 moles of mtrogen. H2C = CHCl-|-[2 5O2+10.0N2] -» HCl+2 CO2-I-H2O-I-IO 0 N2 This five-element (C, H, CI, 0, N) system is much more complex, but is readily solved by the computer for the products that are thermodynamically the most stable, and are shown over the temperature range 300° to 1500°C m Figure 5. With excess oxygen, the major products are mdeed mtrogen, water, carbon dioxide, and hydrogen chlonde There are trace amoimts of diatomic hydrogen (10 to 30 ppm), 140 h KCAL/IOOG 120 h 100 h 80 60 40 20 0 70 60 50 40 30 20 MOLE % NITROGEN J I \ I I 10 20 30 40 50 60 MOLE % METHANE F I G 3 Heat released by the Bystem CH<-Nr-Ot 70

ABSTRACTS AND REVIEWS 1 6 9 MOLE FRACTION GASEOUS PRODUCTS IN PPM 1,000,000 100,000 10,000 1,000 GRAPHITE (SOLID) HYDROGEN CHLORIDE DIATOMIC HYDROGEN METHANE ATOMIC CHLORINE ATOMIC HYDROGEN 400 600 800 1000 1200 1400 TEMPERATURE, ' C Fia. 4. Thermodynaimcally calculated decomposition of vmyl chlonde.

170 FIRE RESEARCH and carbon monoxide (1 to 30 ppm) below 800°C, which increase in concentration as the temperature is raised; at higher temperatures, there are trace amounts of atomic chlorme (1 to 1000 ppm), diatomic chlonne (1 to 18 ppm), diatomic oxygen (1 to 300 ppm), mtnc oxide (1 to 150 ppm) and hydroxĵ l (1 to 100 ppm). The sys- tem does form some phosgene, but it is of the order of 0.1 ppm. Calculations were also made for underoxidized conditions contammg 0 25, 0 50, and 0.75 stoichiometnc air. The results for the species present at concentrations of 1 ppm or greater are given on seven figures tracmg the separate constituents for clanty. Figure 6 shows how the concentration of sohd graphite decreases as the air 1,000,000 NITROGEN CARBON DIOXIDE 100,000 WATER and HYDROGEN CHLORIDE 10,000 MOLE FRACTION GASEOUS PRODUCTS IN PPM 1,000 ATOMIC CHLORINE DIATOMIC HYDROGEN NITRIC OXIDE DIATOMIC CHLORINE CARBON MONOXIDE HYDROXYL 400 1400 600 800 1000 1200 TEMPERATURE, "C Fio. 5. Thermodynamically calculated combustion of vmyl chlonde with stoichiometnc air.

ABSTRACTS AND REVIEWS 171 MOLES SOLID GRAPHITE PER MOLE OF VINYLCHLORIDE IN PPM (SOLID) GRAPHITE NO AIR 1,000,000 : . + 0.25 AIR 100,000 10,000 + 0.50 AIR 1,000 1 1+0.75 AIR 100 10 1 1 li 1 1 1 1 1 400 600 800 1000 1200 1400 TEMPERATURE, ' C Fia 6 Thermodynamically calculated vmyl chlonde-air equilibria.

172 F I R E RESEARCH is increased. With stoichiometric air, the graphite is all oxidized to carbon dioxide and there is no soot. Absence of soot thus means complete combustion. Figure 7 shows the increase of carbon dioxide as the quantity of air nses to the stoichiometnc amount. Figure 8 shows the decrease of carbon monoxide as the air mcreases, reachmg its lowest value at the stoichiometnc pomt. Also mcluded is the mcrease of the hydroxyl radical as the concentration of air rises to the stoichiometnc pomt. Figure 9 shows the expected rise in the mtrogen concentration as the air mcreases Also shown is the decrease m methane concentration as the air nses to the stoi- MOLE FRACTION GASEOUS PRODUCTS IN PPM CARBON DIOXIDE + 1.0 AIR 100,000 ^ • ^ ^ + 0.75 AIR ^ 10,000 \ +0.50 AIR 1,000 100 - \ + 0.25 AIR 10 1 1 1 1 1 1 1 400 600 800 1000 1200 TEMPERATURE, "C F I G . 7. Thermodynamically calculated vmyl chlonde-air eqmhbna. 1400

ABSTRACTS AND REVIEWS 173 MOLE FRACTION GASEOUS PRODUCTS IN PPM 100,000 10,000 1,000 h 100 h 10 I CARBON MONOXIDE + 0.25 AIR + 0.50 AIR + 0.75 AIR + 1.0 AIR> / 0 / / HYDROXYL / + 1.0 AIR V + 0 75AIR +0.50 A I R 7 - ^ / I / I / Y / A 800 1000 1200 TEMPERATURE. ' C 400 600 Fio 8. Thermodynamically calculated vinyl chlonde-au equibbna. 1400 chiometric point. With the methane is also shown the high temperature methyl radical, which of course parallels the methane concentration. Figure 10 shows the rise of water concentration as the concentration of oxygen mcreases. Smce the system contains nitrogen and hydrogen, the concentrations of ammoma are shown and reach a maximum at about 400°C. Hydrocyamc acid synthesized at high temperatures by the reaction CH«-f-NH8+1.5 O2 3 H2O-I-HCN

174 EIRB BESEABCH MOLE FRACTION GASEOUS PRODUCTS IN PPM 1,000,000 100,000 10,000 1,000 100 10 J I T R O G E N ^ 0 . 2 5 AIR + 0.50 AIR \ \ > + 0 . 7 5 AIR • V ^ ^ + 1 . 0 0 AIR \ \ METHANE \ \ \ + N 0 AIR ^ \ ^ . < f 0 . 2 5 AIR \ \ + 0 . 5 0 AIR " ^ ' " ^ \ \ + 0 .75 AIR \ \ \ \ METHYL \ \ + N O A I R ^ - . - ' 4 0 0 600 800 1000 1200 1400 T E M P E R A T U R E , ' C F I G . 9 Thermodynatnically calculated vinyl chlonde-air eqmlibna

ABSTRACTS AND REVIEWS 175 IS present in the 0.25 and 0 50 air calculations At stoichiometnc air, mtnc oxide is also formed. Figure 11 shows the concentrations of both atomic and diatomic hydrogen, and how they decrease as the au- rises m concentration Figure 12 shows the concentra- tions of hydrogen chloride, and how they decrease as air is added. Also shown is the concentration of atomic chlorme, and how it decreases as au" is added. Note, too, that with stoichiometric air, diatomic chlorme is formed (1 to 18 ppm). Further study will show why phosgene is not found m these equihbnum mixtures at concentrations above about 0.1 ppm Figure 13 shows the thermodynamicaUy MOLE FRACTION GASEOUS PRODUCTS IN PPM 100,000 10,000 - 1,000 WATER + 1.00 AIR V +0.75 AIR \ \ ' + 0 . 5 0 AIR \ + 0 . 2 5 AIR AMMONIA + 0 .25 AIR + 0 .50 AIR + 0 .75 AIR , NITRIC / - O X I D E + 1.0 AIR HYDROCYANIC ACID+ 0 25 AIR~y + 0.50 Al R 4 0 0 6 0 0 8 0 0 1000 1200 T E M P E R A T U R E , "C Fia 10 Thermodynanucally calculated vinyl chloride-air equilibna 1400

176 F I B E BESEABCH MOLE FRACTION GASEOUS PRODUCTS IN PPM 100,000 10,000 1,000 100 + 0 25 AIR + 0.50 AIR + 0 .75 A R 10 DIATOMIC HYDROGEN +N0 AIR ATOMIC HYDROGEN + N0 AIR - + 0.25 AIR- /V/V / / / y \ V / / / + 0 . 5 0 AIR ^ V / % 0 . 7 5 AIR / / / r / / / 4 0 0 6 0 0 800 1000 1200 TE"MPERATURE, ' C 1400 F I G 11 Thermodynamically calculated vinyl chlonde-air eqmlibna.

ABSTRACTS AND REVIEWS 177 calculated equilibria between phosgene, carbon monoxide, and chlorine, according to the equation COCI2 5=i CO+CI2. The plot shows that phosgene is stable below about 200°C and, as the temperature is raised above 200°C, the phosgene begins to decompose to chlorine and carbon monoxide, and is all substantially decomposed by about 1000°C. Note that m order to form phosgene, the system must contam both chlorme and carbon monoxide. 1,000,000 - 100.000 10.000 MOLE FRACTION GASEOUS PRODUCTS IN PPM 1,000 h 100 10 h HYDROGEN CHLORIDE + NO AIR + 0 25 AIR + Q ? Q A | R \ ^ + 0.75 ^ + 1.00 AIR AIR + 1.0 AIR ATOMIC C H L O R I D E ^ X / NO AIR. / + 0.75 A I R ^ " ^ ' CHLORINE + LO AIR / ' < / / ^ j . n 2 5 A I R AIR _ _ L 4 0 0 600 800 1000 1200 T E M P E R A T U R E , ' C 1400 F I G 12. Thennodynamically calculated vinyl chlond&-air equilibna.

178 FIBE EESEAECH CARBON MONOXIDE FRACTION PHOSGENE CHLORINE 4 0 0 600 8 0 0 TEMPERATURE, "C 1000 1200 F I G . 13. Thermodynamically calculated equihbna between phosgene, carbon monoxide, and chlonne Reference to the above figures shows that, m the case of vinyl chloride, there are 3 hydrogens to every chlonne, and that hydrogen chloride is the most stable species that the chlorme can form Note too that the highest carbon monoxide concentra- tions occur under oxygen deficient conditions, but these are the same conditions that also yield high hydrogen concentrations Thus, under the lean oxygen condi- tions required to form carbon monoxide, there is no free chlorme to form phosgene Likewise, under high oxygen concentrations where free chlorme is formed, the carbon monoxide has been oxidized to carbon dioxide. On October 23, 1969, a meetmg was called by the Manufacturmg Chenusts' Association to accumulate authontative information on the combustion products of Annyl chloride monomer. After an mtensive laboratory investigation by the manufacturers, on January 14, 1970, they reported: "Phosgene is a very mmor combustion product from vmyl chloride monomer. Very small quantities of phosgene (20-40 ppm) can be produced when vmyl chloride monomer is bummg under very specific conditions which mvolves premixmg vmyl chloride monomer with oxygen Without premixmg, no detectable amount of phosgene is produced. The detection of phosgene m the presence of gross quantities of hydrogen chlonde produced from the bummg of vmyl chlonde monomer is an extremely difficult analytical test. When vmyl chlonde monomer is burned m the

ABSTRACTS AND REVIEWS 179 presence of air, an almost quantitative yield of hydrogen chlonde is obtained. The only other combustion products of any consequence are carbon monoxide, carbon dioxide, and water. Under diffusion conditions, the combustion of vmyl chlonde monomer produces a very sooty flame Approximately 10% of the available carbon is converted to carbon black." These findmgs were later pubhshed.' Identifying Potential Chemical Reaction Hazards The behavior of chemicals m fire problems is closely bound to the mtrmsic ener- gies of the molecules mvolved. Reactions that are capable of releasmg energy are potentially hazardous, and the greater the energy release, the greater is the potential hazard. Therefore, a study of combustion or chemical reactions producmg heat from a chemical thermodynamic viewpomt will prove rewardmg, and has been my mam mterest for the past several years. Once mitiated, the energy generated may be derived from three sources. The first of these is the release of stored energy from a smgle compound. Acetylene is a good example of this case Thermochenustry teaches that the elements graphite and hy- drogen at 298°K can be synthesized into acetylene gas at 298°K by storing 54 19 kcal per mole in the acetylene molecule (see Figure 14) This acetylene molecule with its tnple bond and its 54 19 kcal of stored energy can be thought of as similar to a tightly wound sprmg. Once released, the stored energy is added to the system, largely as heat Thus, the decomposition of acetylene releases 48.2 kcal which heats the constituent elements, graphite and hydrogen adiabatically to 2898°K. This seK decomposition of acetylene is the primary hazard. Now, if this graphite and hydrogen heated to 2898°K is released m oxygen, it would undergo complete oxidation, releasing 116 9 kcal, and produce a senes of most stable oxidation products heated adiabatically to 3314°K. The expansion of these gaseous products to 3314°K would produce the maximum destruction acetylene can produce The difference between this totally oxidized condition and the self decomposition is regarded as the secondary hazard of acetylene These two quantities, primary and secondary hazards, are the basis of a method of potential hazard evaluation. The second method of energy generation is the decomposition of a compound formmg more stable products Ammomum mtrate is an example of this case (see Figure 15). The synthesis of a mole of sohd ammomum mtrate from its constituent elements at 298°K is accomphshed by the release of 87 3 kcal This, m itself, is qmte a degradation m the absolute energy scale Upon appropriate activation, sohd ammomum nitrate may decompose mto more stable products, gaseous mtro- gen, oxygen, and water heated to 1246°K by the further release of 28 3 kcal. For ammomum mtrate, this decomposition is the primary hazard, and smce the prod- ucts contam excess oxygen, there is no secondary hazard. The third method of energy generation is the energy-producing reaction of two or more chemical species. Octane and oxygen will be used as the example reaction. Thermochemically speakmg, a mole of liqmd octane can be synthesized from its elements at 298°K by the release of 60 kcal (see Figure 16) Self decomposition of bquid octane to form graphite and methane and hydrogen gases heated to 591''K by the further release of 17.2 kcal, is the primary hazard. Complete oxidation of hqmd octane produces the most stable oxidation products heated to 3102°K by the release of 587 2 kcal, and wiU produce the maximum destruction this system can

180 FIRE RESEARCH 298 'K • C2H2 GAS C2H2+2 1/2 O2 S E r L F 1 MAXIMUM DECOMPOSITION DESTRUCTION +54 .2 - 6 2 . 7 S Y N T H E S I S + 54.19 KCAL 4 - 4 8 . 3 K C A L 2 9 8 ' K PRIMARY HAZARD C O M P L E T E OXIDATION 2C + .94 H2+ . I2H 2C + H2 E L E M E N T S 2 8 9 8 ' K . T D SECONDARY HAZARD -116 .9 KCAL 3 3 I 4 ' K To . I 6H2+ .35 H + .46H2O+ .41 OH .55O2+.5O 0 + . 5 2 C 0 2 + I .48C0 F i a 14 Thennochemistry of acetylene.

ABSTRACTS AND REVIEWS 181 2 9 8 °K - 8 7 . 3 2 9 8 ° K - 1 1 5 . 6 I 2 4 6 ' K TD N 2 + I 1/2 O 2 + 2 H2 E L E M E N T S S Y N T H E S I S - 8 7 . 3 K C A L NH4NO3 S O L I D S E L F D E C O M P O S I T I O N - 2 8 . 3 K C A L I PRIMARY HAZARD N2+ 1/2 O2 + 2 H2O F I G 16 Thermochemistry of ammonium mtrate

182 0 - 6 0 FIRE RESEARCH 298»K 8 C + 9 H2 E L E M E N T S SYNTHESIS - 6 0 KCAL 2 9 8 ' K 5 9 r K T D C s H i e LIQUID S E L F DECOMPOSITION - I 7 . 2 ^ K C A L PRIMARY HAZARD . 4 H 2 + 4 . 3 CH4 + 3 . 7 C SECONDARY HAZARD CeHia + 121 /2 02 MAXIMUM DESTRUCTION C O M P L E T E OXIDATION - 5 8 7 . 2 KCAL 6 4 7 3I02''K To 1.23 H2+ I.IOH + 6.21 H2O + 2.02 OH 2.22 O2+ l.lOO + 3.24 CO2 + 4.76 CO F I G 16 Thennochemistry of octane.

ABSTRACTS AND REVIEWS 183 produce. The secondary hazard is the difference between the self decomposition (denoted by subscnpt d) and the complete oxidation (denoted by subscript o). These energies and product gas temperatures are calculated from the thermo- dynamic properties of these systems, and are thus completely scientific and di- vorced from human emotions In an earher study,* a correlation of the values of the prunary and secondary hazards with the National Fire Protection Association Chemical Reactivity Rat- mg,' as detailed m Table I I , was presented. Contmued work with this concept has shown that the correlation based on the self decomposition temperature Td and the temperature of the fully oxidized products To is the most useful for potential hazard prediction. To mtroduce the concept here, twenty compounds plus acetylene and ammonium mtrate, whose N F P A Chemical Reactivity Ratmgs have been as- signed, were selected, their Td and To values calculated (see Table I I I ) , and the values of the prunary and secondary hazards plotted m Fig 17. Observation of the data shows that the safest materials, N F P A Chemical Reac- tivity =0, are to be found m the upper left-hand comer, while the most potentially hazardous matenals, rated 4, are in the lower right-hand comer. Thus, a band from methane to nitroglycerme covers the range from safest to the most potentially hazardous matenals There is, however, considerable overlap m the ratmgs. Of this data, N F P A says "This material is not an official standard of the N F P A ; it is only a compilation of data from various authoritative sources presented for information as a guide." Propyl alcohol faUs between octane and acetone, and seems to be out of place for a 1 ratmg. The other three 1-rated matenals, butyraldehyde, epi- chlorohydrin, and vmyl chloride, spread over qmte a wide area. Acetaldehyde, ethylene, styrene, and vmyl acetate are clustered m the same area, and seem out of place smce they carry a 2 ratmg The 3-rated materials are all qmte hazardous, and cover a wide range from 4-mtroanilme to ammomum nitrate to acetylene. T A B L E I I NFPA Chemical Reactivity ratmg 0 Matenals which are normally stable even under fire-exposure conditions, and are not re- active with water 1 Normally stable materials that may become unstable at elevated temperature and pressures, or which may react with water with some release of energy but not violently. 2 Matenals which are normally unstable and readily undergo violent chenucal change, but do not detonate Includes materials which can undergo chemical change with rapid release of energy at normal temperatures and pressures or which can undergo violent chenucal change at elevated temperatures and pressures Also mcludes matenals which may react violently with water, or which may form potentially explosive mixtures with water. 3 Materials which are capable of detonation or of explosive decomposition or of explosive reaction, but which require a strong imtiatmg source or which must be heated imder con- finement before imtiation Includes materials which are sensitive to thermal or mechamcal shock at elevated temperatures and pressures or which react explosively with water without requirmg heat or confinement 4 Matenals which are readily capable of detonation or of explosive decomposition or explosive reaction at normal temperatures and pressure. Includes matenals which are sensitive to mechamcal or localised thermal shock.

184 FIBE BESEABCH T A B L E I I I Calculated decomposition and complete oxidation temperatures for NFPA Chemical Reactivity rated compounds State NFPA To-Ti ^1. Acetone 0) 0 704 2331 2 Methane (g) 0 298 2756 3 Methyl alcohol (1) 0 692 2189 4. Octane 0) 0 591 2511 5 Butyraldehyde (g) 1 1039 2104 6 Epichlorohydnn (1) 1 975 2056 7. Propanol 0) 1 615 2401 8 Vinyl chlonde (1) 1 1374 1714 9. Acetaldehyde (1) 2 815 2201 10 Ethylene (g) 2 1005 2170 11 Styrene 0) 2 923 2268 12 Vinyl acetate (1) 2 844 2179 13 Acetylene (g) 3 2898 443 14 Ammomum mtrate (8) 3 1242 0 15 Cellulose mtrate (8) 3 2213 641 16 2,4-Dmitroanilme (g) 3 1904 1219 17. 4-Nitroamhne (8) 3 1103 2013 18 Nitroethane 0) 3 1064 1943 19. Nitroglycenne (1) 4 2859 0 20 Nitromethane (1) 4 2418 509 21. Propargyl bronude (1) 4 1898 1266 22. 2,4,6-Tnmtrotoluene (8) 4 2025 1084 g=ga3, l = hqmd, s = solid Interspersed m this area are the 4-rated materials, potentially the most hazardous matenals. This type of calculation provides an objective method of segregatmg any ma- terial mto its relative degree of potential hazard from the least hazardous materials to the potentially most hazardous materials. By such an ordermg process, items havmg the same degree of potential hazard will be close to each other, permittmg handlmg information on known matenals to be extended and apphed to matenals where no handlmg information is available. Once the fire fighter is informed of the degree of potential hazard he is deahng with, he can operate more safely and ef- fectively than before. I t is, therefore, cntically necessary for the chemicals and chemically denved matenals of modem commerce to be studied from the energy pomt of view, and to calculate the degree of potential chemical reaction hazard appropnate for each matenal. Real Chemical Reaction Hazards I t must not be forgotten that thermodynamics treats equihbnum systems and takes no cogmzance of time. Thus, the calculated energy releases, and the product gas temperatures reached will represent the equihbnum case, and may require

ABSTRACTS AND REVIEWS 185 relatively long times to achieve. I t is for this reason that these calculations are spoken of m the potential sense. We are hving, however, m a real world and not a potential world. In a potential world, an energy release may cover years and cause no trouble whatever, but m a real world, the same energy may be released m 1 mmute or m a fraction of a millisecond with altogether shattermg results The important factor that is also required tn addition to the chemical thermodynamic information is the time-rate of energy release by the reactmg system, or the overall kmetic parameters that govern the energy release. This is an area where chemical thermodynamics and chemical kmetics must be linked together to provide real answers for problems of the real world as we know it today. Jomt studies between these two disciplmes will prove to be very fruitful. NFPA CHEMICAL REACTIVITY 0 1 2 3 4 M E T H A N E • O C T A N E P R O P Y L A L C O H O L A C E T O N E M E T H Y L S T Y R E N E A i r n H n i - - E T H Y L E N E . ^ i l . . ^ ^ M v r , / K A * B U T Y R A L D E H Y D E A C E T A L D E H Y D E /V"^ 4 - N I T R O A N I L I N E V I N Y L A C E T A T E / > N I T R O E T H A N E E P I C H L O R O H Y D R I N A V I N Y L C H L O R I D E M P R O P A R G Y L B R O M I D E « 2 , 4 - O I N I T R O A N I L I N E » 2 , 4 , 6 - T R I N I T R O T O L U E N E N I T i ^ O C E L L U L O S E « N I T R O M E T H A N E A C E T Y L E N E AMMONIUM N I T R A T E N I T R O G L Y C E R I N E 5 0 0 1000 1500 2000 2500 3 0 0 0 DECOMPOSITION T E M P E R A T U R E T d ' K F I G 17 Comparison of primary and secondary hazard ratings with the N F P A chemical re- activity rating

186 FIRE RESEARCH References 1 G N L E W I S AND M RANDALI., revised by K S P I T Z E R AND L B R E W E R Thennodynamics, 2nd ed , McGraw-Hill Book Company (1961) 2 D R S T U L L The American Scientist, 59, 734^743 (1971) 3 D. R C R U I S E - J Phys Chem 68, 3797-3802 (1964) 4 JANAF THERMOCHEMICAL T A B L E S , second edition, National Standard Reference Data System Report No NSRDS-NBS 37, Issued June 1971 For sale by the Superintendent of Documents, U S Government PrmtmgOflBce, Washington, D C , 20402 (Order by Catalog No C13 48 37) 6. J B F E N N AND H F . C A L C O T E FouHh Symposium on Combustion, pp 231-239, WiUiams and Wilkins Co (1953) 6 R D K O O L E R AND D . L D O W E L L Preprint 14B presented at the Loss Prevention Symposium, 68th Annual AIChE Meeting, Houston, Texas, March 1-3, 1971 See also The National Ob- server, (September 15, 1969) 7 M M O'MARA, L B C R I D E R AND R L D A N I E L Amer Ind Hyg Assoc J SS, 153-156 (1971) 8 D R S T U L L Chem Engr Progress, Loss PREVENTION 4, 16-24 (1970) 9 Fire Protection Gmde on Hazardous Materials, 3rd ed (1969), Section 325M, "Fire-Hazard Properties of Liqmds, Gases, and Volatile Sohds," National Fire Protection Association, 60 Batterymarch St , Boston, Mass 02110