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.

CONDENSED PHASE COMBUSTION C H E M I S T R Y LEO A. WALL Naltonal Bureau of Standards In this article we shall endeavor to survey the pyrolj^ic processes occumng in polymeric melts on the assumption that our comprehension of these processes are highly important if not cntical for control of igmtion and burning of matenals. P Y R O L Y S I S O F P O L Y M E R S U N D E R H I G H V A C U U M A comprehensive review of the pyrolytic degradation of polymers has been made by Madorsky ̂ Other reviews covermg pyrolytic and many other types of polymer degradation are also available For vmyl and related polymers, a large vanation m the number of pyrolysis products' from a given polymer occur (see Table 1). A few decompose to yield 100% pure monomer. The results in Table 1 are for certam experimental conditions, high vacuum, thm samples and low residence time m the furnace, and are essen- tially the pnmary products produced m the melt phase In Table 2, overall activa- tion energies are hsted I t is seen that these cover a large range and that the larger values are close to 80 kcal/mole, which is an approximate upper hmit for the carbon-carbon single bond One may suspect then that these decompositions have been imtiated by homolytic bond rupture of the tj^pical bonds m the polymer cham. In general it is now qmte evident that the Rice-Herzfield mechanism modified for apphcation to condensed phase polymenc systems can be apphed to many polymeric decompositions The energetics of decomposition can thus be discussed in terms of the elementary steps Initiation: Random Q. Ri-j-\-Rj E n d Q. Propagation* fi. R^-X+M Transfer. R,+Q, Q,+R, R, »A;8 Qi-k+Rk Termination: Combination Rt-\-R, QH, Disproportionation Rt-]-Rj 204

ABSTRACTS AND REVIEWS 205 TABLE 1 Yield of monomer m the pyrolysis of some organic polymers m a vacuum (In per cent of total volatibzed) (Ref 1) Temperature Yield of range. monomer, Polymer "C % Polymethylene 335-450 0 03 Polyethylene 393-444 0 03 Polypropylene 328-410 0 17 Polymethylaerylate 292-399 0 7 Hydrogenated polystyrene 335-391 1 Poly (propylene oxide), atactic 270-550 2 8 Poly (propylene oxide), isotactic 295-355 3 55 Poly (ethylene oxide) 324r-363 3 9 Polyisobutylene 288-425 18 1 Polychlorotnfluoroethylene 347-415 25 8 Poly-j9-deuterostyrene 34&-384 39 7 Polystyrene 366-375 40 6 Poly-m-methylstyrene 309-399 44 4 Poly-o-deuterostyrene 334-387 68 4 Poly-a, /S, /J-tnfluorostjrene 333-382 72 0 Polymethylmethacrylate 24&-354 91 4 Polytetrafluoroethylene 504-517 96 6 Poly-a-methylstyrene 25»-349 100 Polyoxymethylene Below 200 100 Here, Q, is the number of polymer molecules composed of i monomer M umts, and Rt IS the number of radicals havmg i monomer umts. The rate of volatilization of the polymer would m general be given by: L (1) where "M" symbolizes total volatUes in terms of monomer segments, and L is the smallest degree of polymerization that must chemically decompose in order to vaporize. If ki were zero, and only monomer of molecular weight rn volatilizes, then dM/dt=kJt, (2) where i2=5^, . The fractional rate of weight loss C, based on initial weight of sample Wo, is: 1 dC/dt= {m/wo)lcJt. (3) For the purpose of discussing the key characteristics of these reactions, we shall present a treatment of a simple case. One of the simplest mechanisms is composed of propagation, random initiation, and disproportionation. The steady-state condi- tion then follows: hZpV{t)/m-]=kJi^/V{t). (4)

206 F f f i E R E S E A R C H T A B L E 2 Activation energies of thennal degradation of some organic polymers m a vacuum (Rcf. 1) Temperature Activation Molecular range, energy, Polymer weight "C kcal/mole Phenolic resm 331 5-355 18 Atactic poly (propylene oxide) 16,000 265-285 20 Polymethylmethacrylate 150,000 22&-256 30 Polymethylacrylate — 271-286 34 Poly (ethylene terephthalate) — 336-356 38 Isotactic poly (propylene oxide) 215,000 285-300 45 Cellulose triacetate — 283-306 45 Poly (ethylene oxide) 9,000-10,000 320-335 46 Pol3Tsobutylene 1,500,000 306-326 49 Hydrogenated polystyrene 82,000 321-336 49 Cellulose — 261-291 50 Polybenzyl 4,300 386-416 50 Polymethylmethacrylate 5,100,000 296-311 52 Polystyrene 230,000 318-348 55 Poly-a-methylstyrene 350,000 228 8-275 5 55 Poly-onieuterostyrene Bigh 321-341 55 Poly-/3-deuterostyTene High 326-346 56 Poly-77}-methylst3rrene 450,000 318 5-338 5 56 Polyisoprene — 291-306 58 Polychlorotnfluoroethylene 100,000 331 8-371 57 Poljfpropylene — 336-366 58 Polyethylene 20,000 360-392 63 Poly-oc/S-iS-tnfluorostvrene 300,000 333-382 64 Poly-2,3,4,5,6-pentariuorostyrene — 395-410 65 Polymethylene High 345-396 72 Poly-p-xylylene — 401-411 73 Polytetrafluoroethylene — 423/5-513 80 5 I n the units used R/V{t) is the steady-state concentration of radicals. Thus, we have the number of radicals given by /e = [(fci/fc4)(p/m)]'«F(0. (5) I t is seen to be proportional to the volume or weight of the sample (p is the density of the polymer) The final rate equation is dC/dt = h l i h / h ) ( m / p ) ] ' « ( l - C ) . (6) Next, we define the zip length Z, which m this special case is identical with the kmetic cham length Z = m / l {2kiR^)/V (<)] = hV it)/2kji = {m/kjcipyi\ (7) I n more comphcated cases, the zip length is the ratio of the rate of propagation to termmation plus transfer.

ABSTRACTS AND R E V I E W S 207 In this rare case, it is seen that Z is a constant. Also, since we have restricted termination to disproportionation, we have already assumed that Z « P „ , where P„ is the number average degree of polymerization. By taking the weight loss to be entu-ely the result of monomer production by propagation we have assumed also that Z > 1 . We next determine the variation of P„ with the extent of reaction. For the purpose of finding the molecular weight dependence on conversion, we write a differential equation for the vanation in the total number of polymer molecules 2 With the extent of reaction dQ/dt=kJi:'/V{t). (8) Dmdmg by Eq (3), dQ/dC = IkJtyV (OfcjK] (wo/m) = (wo/m){h/h) IR/Vm Replacmg R/V{t) from E q (7) and integratmg, we find that Qi-Qo=(wo/m){C/2Z) (9) or [w{t)/mPn(<)]-CWmP„(o)] = (wo/m) (C/2Z) , and then ( l -C) /P„(<) -CPn(o) ] - ' = (7/2Z. (10) Here, Pn(0) and Pn(0 are the number average degrees of polymerization imtially and at time t, respectively Transformmg E q . (10) to p , . P n ( 0 ) [ l - C ] ^"^'^ (1+CP„(0)/2Z]C}- ^''^ I t IS seen that m the limit of CPn(0)/Z3—>0, PB(C) decreases Imearly with (1 —C). On the other hand, as the ratio takes on larger positive values, Pn(C) decreases more rapidly Variations of the above treatment are, on the whole, quite satisfactory for in- terpretmg the mechamsm of polymer decomposition of the type that produce mainly monomer. Catalytic effects resultmg from small concentrations of impurities oi abnormal structures m the cham which lead to rapid imtiation, change the behavior for some polymers drastically as one studies different samples, prepared and processed by different methods Sensitivity of the decomposition to impurities, peroxides, ultra- violet hght, gamma rays, and other agents is proportional, of course, to the kinetic cham length. By mcorporating transfer m the mechanism, the treatment can be apphed suc- cessfully to most all vmyl-type polymers such as those in Table 1, with some exceptions—for example, polyvinyl chloride and polyvmyl acetate. Table 3 lists zip lengths deduced from measurements of the decomposition be- havior of four polymers The two with the large zip length, ~10', have relatively

208 R E S E A K C H T A B L E 3 Zip lengths for polymer depolymenzation at 1%/niin rates Polymer Zip lengths Methyl methacrylate ~10» a-methyl styrene ~10' Styrene ~ 5 Tetrafluoroethylene ~ 3 low thermal stability. For the polymethyl methacrylate a variety of rates and activation energies (30-52 kcal/mole) have been reported. E N E R G E T I C S O F P O L Y M E R D E C O M P O S I T I O N From the rate equation for polymer decomposition, E q . (6), it is seen that meas- urement of the temperature dependence of the rate yields an overall activation energy Eo which is related to the activation energies of the elementary processes by E,=HEi-E,)+E,. (12) By defimtion, the bond-dissociation energy is D{R—R)=Ei—Ei Also, Ei = Hp+E-2, where Hp is the heat of polymerization and E-i is the activation energy for the addition of monomer to a polymer radical. Table 4 presents values for E-2 taken from a recent review^ of the thermochemical measurements pertment to polymerization processes and Table 5 presents values for the heat of polymeriza- tion Also m Table 4 are values of Ea, the activation for termination of large radicals m polymerization systems Thus, these Ei values are for radicals m low-viscosity hqmd systems at or near 25''C The fact that the Et values are positive fimte quan- tities IS assumed to be due to diffusion control, since activation energies for radical termination m the gas phase are effectively zero In the degradation processes, the systems are high-viscosity polymer melts at 200°-400°C and Ei values run 20-40 kcal/mole (1 cal=4,184 joules) * Equation (12) now can be formulated as E,=hLD{R-Rn+Hp+E.2, or DiR-R)=2£E,-Hp-E.i\. (13) Smce E-2 values are small and not readily available, we may approximate values ofD{R-R) from DiR-R)=2lEo-Hp-52. (14) Assuming Eo is mdependent of temperature the estimated values of D{R—R) would be for 25°C. The results are not accurate enough for this factor to concern us. Apphcation of E q . (14) to several polymers which appear to imtiate at random give the results m Table 6. The D{R—R) values are quite large and support the view that m the cases listed carbon-carbon bond rupture is the initiatmg process for depolymenzation. The high value of 90 kcal/mole for poly-a-methyl styrene may mdicate an abnormally high E-2, while that for polyethylene most likely

T A B L E 4 ^ Absolute rate constants (at 25°C), activation energies, and PZ factors for propagation and termination in radical polymerization (Ref 7) 2 a Propagation Termmation H i P Z X I O " ' fc.XlO"' hp, hter Ei, kcal hter mole"' hter mole"' Et, kcal hter mole"' Monomer mole"* sec"' mole"' sec"' sec"' mole"' sec"' Ethylene 470 (83°C) 50 6 100 (83°C) Styrene 20 6 4 0 45 0 23 1 9 0 06 Methyl methacrylate 269 4 7 0 09 1 5 1 2 0 11 Vmyl acetate 980 5 0 3 2 2 0 3 2 3 7 Vmyl chloride 6200 3 7 0 33 1200 4 2 Methyl acrylate 720 7 1 10 Vmyhdene chlonde 8 6 0 02 <ert-Butyl methacrylate 350 4 4 1 4 1 1 Acrylomtrile 1450 52 4 1 3 200 0 5 5 4 3300 Acrylanude 18,000 1 45 Butadiene 20 9 12 Isoprene 10 10 12 2-Vinyl p3rndine 96 8 0 89 5 4-Vmyl pyndme 12 Nitroethylene 14,000 « 5 8 CO

210 F I B E B E S E A B C H T A B L E 5 Standard enthalpies of polymerization (hquid-crystal) (Ref 9) -Af f - l c , Monomer kcal/mole Tetrafluoroethylene 41 5 37 Ethylene 25 88, gc 21 2 3,3-Dichloromethyl 1-oxacy clobutane 20 2 Vinyl acetate 21 2 Butadiene 17 6 Propylene 24 89 19 5 Butene-1 19 0 Isoprene 17 9 Styrene 16 7 Methyl methacrylate 13 2 Ethyl methacrylate 13 8 Isobutylene 12 9 Formaldehyde 13 2 7 4 tt-methyl styrene 8 4 Tetrahydrofuran 5 3 Sulfur - 3 17 Selemum - 2 27 Cyclopropane 27 0 Cyclobutane 25 1 Cyclopentane 5 2 Cyclohexane - 0 7 Cyclooctane 8 3 results from the fact that it does not depolymenze to monomer and, hence, E q (14) does not strictly apply. Polyethylene volatilizes at 410°C by a random scission process having an overall rate constant A;o = 10" exp(—72,000/RT) sec~' Stress-relaxation measurements' of a crosshnked polyethylene in the region of 300°C indicate that the chains rupture with a rate constant k = KP" exp(-73,300/flr). The similarity in the two results indicates that an identical process determines the results of both experiments Instead of calculating the value of D(R—R) for polyethylene from its activation energy for random degradation, it is more informative to assume that the random decomposition of polyethylene occurs by a chain reaction process with transfer as the propagating process Taking this process to initiate by a random carbon bond rupture with. D (R—R) =80, then we calculate an activation energy of 34 kcal/mole for the transfer process in the melt O T H E R T Y P E S O F P O L Y M E R D E C O M P O S I T I O N Table 7 lists some other polymers which decompose by mechanisms qmte dif- ferent from those so far discussed These produce to a greater or lesser degree char

ABSTRACTS AND R E V I E W S 211 and certam volatile products. There are m mam two types of decompositions. In one, side groups split from the chain leaving a charred residue, for example, polyvmyl chlonde which dehydrochlonnates, and cellulose which can with the aid of catalysts be made to dehydrate leavmg a carbon residue. The second type is those contaimng mostly aromatic group hnks in the polymer chain These have been shown to cross-bnk and, hence, ultimately form carbon and various small volatile species The polymers m Tables 1 and 2 decompose endothermically. Those m Table 7 tend to decompose exothermically. There are polymers which are exo- thermic m polymerization and exothermic m decompositionThe dehydration of pure cellulose is exothermic by 29 4 kcal per mole of water produced, yet it de- composes mamly to levoglucosan and flammable tars, a process that is probably endothermic at 25°C Table 8 lists thermodynamic quantities for vanous conceivable decomposition processes for polyethylene and polytetrafluoroethylene Thus, thermodynamic considerations alone are insufficient for predicting the thermal decomposition behavior of polymers, since as is also likely for cellulose, the actual observed decomposition mechanisms (d and / m Table 8) have positive free ener- gies at 25''C while decomposition processes with large negative free energies are not observed. Heats of polymenzation, Table 5, are very useful for choosmg con- ditions of polymenzation, but are of less value in predictmg thermal stabihty.' I N D U C E D A N D O X I D A T I V E D E G R A D A T I O N Some polymers particularly those with long zip lengths are sensitive to electro- magnetic radiation and decompose at temperatures 100 or more degrees lower when irradiated Such polymers when irradiated at room temperature or lower can also develop concentrations of free radicals and will give off a small burst of volatile monomer on subsequent heating to a modest temperature The most stable samples of polymethylmethacrylate decompose to monomer with an activation energy of 52 kcal/mole If the imtiation process is by a radiation mechanism, the activation energy for decomposition would be given by E q (12), with El set equal to zero, E (photo) =E2-^Et (15) For polymethylmethacrylate, this gives (photo) =9 kcal/mole, and would occur readily at or near room temperature. For poly-a-methyl styrene and polytri- fluoroethylene, (photo) values of 25 and 13, respectively, have been estimated.' Fortunately, few polymers appear to have very long kmetic cham lengths. T A B L E 6 Estimated dissociation energies, D{R-R), for polymer bonds and initiation rates for depolymenzation (8) PoljTner D(R-R), kcal/mole fci, sec"' Tetrafluoroethylene 80 10»exp(-120,000/ f iT) a-Methyl styrene 90 10"exp(-65,000/B7') Methyl methacrylate 68 — Ethylene (90) [fc, = 10" exp(-72,000/flr)]

212 F I R E R E S E A R C H T A B L E 7 Pnncipal products from pyrolysis of polymers (Ref 2) Polymer Structural umt Pnncipal volatile products Cellulose VmyUdene chlonde Vmyl chlonde Vinyl acetate Chloroprene Acrylonitnle p-Xylene Benzyl Phenyl CHjOH I C - 0 / H \ — C — H HC—O— \ H H / C — C 0 0 H H — C H j C C l r - — C H , — C H C l — O O C — C H , I — C H , — C H — H - C H r - C = C — C H r - Cl H H C N Levoglucosan, CO, CO.; HjO -CH, i<( ^ C H , - C H , . <z> HCl HCl , C .H. CH,COOH HCl Small amounts of H C N , acrylomtnle 4% of xylene, toluene, benzene, meth- ylstyrene, methylethylbenzene 7% of toluene, benzene, xylene H , Polymer decomposition may also be catalyzed or inhibited by gases.* Small catalytic quantities of oxidant gases will act essentially as electromagnetic radiation and imtiate depolymenzation with near-zero activation energy In a bummg process, then, one might anticipate a lower activation for polymer decomposition because of the presence of the oxidant. Recently, we have studied'* the thermal decomposition of an aromatic polymer (poly[i\r,Ar-(p,p'-oxydiphenylene) pyromelhtimide]) under 1 to 150 mm pressure of oxygen, and over the temperature range of 480" to 520°C. We found the much

ABSTRACTS AND R E V I E W S 213 faster oxygen-induced rate of initial weight loss to depend upon the 0 4 power of the oxygen pressure One would, by postulating a chain reaction and oxidative imtiation, anticipate an 0 5-power dependence The oxidation of hydrocarbons at low pressure when measured by oxygen consumption normally has a first-power dependence upon oxygen pressure The physical deterioration of polymers by oxidative aging processes"'" has long been of concern to the polymer field. Dependmg on the polymer and condi- tions, most polymers oxidize slowly at ambient temperatures and antioxidants are incorporated in polymeric articles Oxidative aging will, m general, be expected to enhance the ignition probabihty and combustibihty of polymers Studies of pyrolytic decomposition are normally earned out on highly punfied materials and, thus, data such as listed m Tables 1 and 2 are not hkely to be always characteristic of aged materials T A B L E 8 Thermodynamic quantities* for decomposition of polyethylene and polytetrafluoroethylene (Ref 8) AS, Aff, cal/mole/ Af, kcal/mole °K kcal/mole -CHiCH, •CH4=CH, b - » l / 3 C , H , + H , 22 35 16 46 34 07 34 28 12 19 6 23 +2 C (Graphite)+2 H , 9 85 46 76 - 4 10 -m-Alkenes [(l/18)CH,(CHj)8>CH=CH,] 1 09 2 03 0 49 +C (Graphite)+CH4 - 8 04 27 49 -16 23 -CFjCF, •CFsCF, g +2 0 (Graphite)+2 F , 46 194 45 71 33 172 *C (Graphite)+CF4 - 2 7 35 - 3 8 * With the exception of graphite all the substances are taken to be m the "ideal" gas state at 298 2°K For the high polymers, this "ideal" gaa state is evidently for the polymer m its most extended conformation (Ref 18)

214 F I B E R E S E A R C H Burning of Polymers Fnedman'5 has pointed out that plastics burn with several times the intensity of cellulose and that the heat of combustion per gram (or cc) for polyethylene is greater than that for cellulose by over a factor of 2 Also, cellulose absorbs water and tends to char while polyethylene is lyophobic and burns without char produc- tion These variations in behavior and properties suggest that a comparison of enthalpies of combustion and approximate reaction temperatures may be helpful. In Table 9 the enthalpies for the burmng of polyethylene with three different oxidants are compared along with data for cellulose All the species are taken in the gas phase, except for carbon The values obtamed with carbon as graphite and carbon as vapor are both hsted The well known observation that oxidative attack on the hydrogen in hydrocarbon occurs more easily than attack on carbon suggests the separation of the burmng process into two separate processes, one for the burn- ing off of the hydrogen and one for a hypothetically sequential combustion of the carbon The reactions are compared on the basis of one polymer carbon. Hence, the factor 1/6 for the cellulose We would have preferred to have compared the heats of combustion with activation energies for the same process Smce such data was lackmg in most instances, we used general information and experience to deduce comparative reaction temperature The approximate reaction temperature for the hydrogen oxidation is hundreds of degrees below that for graphite oxidation. It is evident from the table that the usefulness of chlorme-containmg substances as flame retardants is related to the relative ease with which chlorine will attack hydrogen and the relative difficulty it has m reacting with carbon This serves to emphasize agam considerations of importance to flammabdity long known from studies on the decomposition of cellulose," where the presence of small amounts of shghtly acidic or lomc substances will catalyze the dehydration of cellulose The size of the exothernucity of the dehydration process suggests that this process should occur more readily than it has been observed." Heats of Vaponzatton and Decomposition It has been mdicated that heats of gasification," presumably the sum of the heats for vaporization and decomposition, of plastics are important in the mathe- matical analyses of burmng systems In these treatments, a parameter called the fi-number is defined as where r is the stoichiometric mass fuel-air ratio, H and L the respective heats of combustion and "vaporization" per umt mass, Cg and Ce heat capacities of gas and liqmd, and T„ Ta, and T, are, respectively, the temperatures of surface of the burmng material, the ambient air, and the mtenor of the matenal. Thus the fi-number is the ratio of energy released by the burmng to the energy required to vaporize the fuel Values of L for polymers are, m general, untabulated as such For polymers, we may take L to be the sum of the heats of molecular decomposition Hd, and molecular vaporization H, L=Hd+H, (17)

ABSTRACTS AND R E V I E W S 215 T A B L E 9 Comparison of enthalpies (kcals) of combustion-t3rpe reactions A H „ kcal ~Reaction T, °C Carbon as (carbon as Reaction Graphite Gas graphite) —CH,—-1-1/2 0,-*C-l-H,0 C + 0 , - C O , —CHr—|-F , -»C+2HF C + 2 F,->CF4 —CHr-+Cl2->C+2 HCl C + 2 C h ^ C C U 1/6 C,H,o06-vC-1-5/6 H2O C+02->C0, -52 9 -94 1 -147 0 -123 5 -220 0 -343 5 - 3 9 2 -25 5 -64 7 -14 0 -94 1 -108 1 118 8 -265 8 48 2 -391 7 132 5 -197 2 157 7 -265 8 100 500 low, <0 300 200 high, >1000 300 (cat <250) 500 H , + l / 2 0,-*H,0 -57 8 If the material decomposes mto small molecules, then it is likely that Ha^H, Conversely, if it decomposes into large molecules then However, smce L here is umts of heat per gram, this effect will not be pronounced For the relatively few polymers which decompose (depolymerize) back to the original monomer we have values for Hd smce, for these systems, Hd is equal to the enthalpy of polymerization, Table 5 Since the values m Table 5 are for the process at 25°C, we calculated L values at 25''C. Heats of vaporization at 25°C were estimated except where values were in the hteiature These L values are hsted in Table 10 along with values for polystyrene and polyethylene For these polymers, which decompose to products other than monomer, we deduced heat of decomposi- tion in several ways For polyethylene, we had previously calculated a value of the enthalpy of decomposition; see Table 8, Reaction (d). This was based on the con- cept that this decomposition can be approximated by the indicated Reaction (d). Here, we assume that the products predominantly a vanety of n-alkanes and w-alkenes can be represented by a n-Cze'Hn alkene (molecular weight, 504). A second calculation was made assummg the products were representable by a n-CisHw alkene (MW, 252) I n Table 10 we list the molecular weight (or average) of the decomposition products For the alkenes the heats of vaporization would be comparable to those for the corresponding alkanes We have been mvestigatmg the

216 TIRE R E S E A R C H T A B L E 10 Estimated heats of gasification for some polymers at ~25°C L (cal/g) Polymer (g/mole) Calc Exptl , Ref (17) Methyl methacrylate 100 232 380 at 460°C Formaldehyde 30 630 440 at 160°C a-Methyl styrene 118 158 Tetrafluoroethylene 100 415 Stjrrene (198) 241 Ethylene (d) (504) 102 Ethylene (d') (252) 141 Ethylene (b) (21) 683 Ethylene (a) 28 924 vaporization of linear alkanes'' and find that their heats of vaponzation m kcal/mole can be approximated by Aff, = 3 21 n»»-0 0193 r-f-2 92, (18) when n is the number of carbon atoms m the alkane and T is in "K. The 2/3 power dependence is mterpreted to mdicate that these flexible molecules coil up and vaporize as spheres Thus, the mcremental heat of vaporization contmually de- creases with size of molecule In Table 5 we have data on various hypothetical decomposition processes for polyethylene, we also calculated L values shown m Table 10 for decompositions a and b As the temperature mcreases, one anticipates that the decomposition mechamsm will shift from d to d', b and a (see Table 10) For poIystjTene, which decomposes to give 42% monomer, the heat of polymeriza- tion and the heat of vaporization of styrene, weighted by the actual number of bond ruptures, were used for the calculations The measured L values of Dmitnev et al " are hsted in Table 10 for comparison Their value of 440 cal/gm polyformaldehyde is exactly the molecular heat of polymerization divided by 30, the molecular weight of the monomer At the decomposition temperature of 160°C, formaldehyde would be above its cntical temperature and there would be no heat of vaporization; on the other hand, one would expect the heat of polymerization to increase. The higher experi- mental value for methyl methacrylate suggests that other decomposition processes are becoming important rather than the decomposition to monomer. The tempera- ture (460°C) is several hundred degrees above that used in earher degradation studies * We have for some time considered it desirable to measure heats of decomposition However, available instruments, such as the differential scanmng calorimeter, were not completely smtable, partly because the temperature range did not extend as high as we wished to go and partly because of problems m handlmg and measuring the volatile products Several polymers (for example, polyvmylidene fluoride) which decompose to hydrogen fluoride and char probably have exothermic heats of de-

ABSTRACTS AND R E V I E W S 217 composition For polyvinyhdene fluonde, one estimates an exothermic heat of ap- proximately 20-30 kcal/64 g. A C T I V A T I O N E N E R G I E S F R O M R E G R E S S I O N R A T E S An extensive study of the burning of polymers has been earned out by Blazowski, Cole, and McAlevy " An expenmental arrangement was used which permitted measurements to be made of the length of polymer rod consumed per time mterval and the surface temperature of the burmng polymer From such measurements, determination of activation energy for polymer consumption under burmng condi- tions were made. Results were reported for six polymers Their results are listed m the last column of Table 11 for direct comparison with the activation energies reported'-' for polymer pyrolysis based on weight loss of polymer heated under vacuum. From the point of view of this writer, the results show good agreement. An expenmental value for the activation energy of Delnn is not available However, from its heat of polymerization (13 2 kcal/mole) and low decomposition tempera- ture, one anticipates an activation energy between 20-30 kcal/mole. The pyrolytic activation energy for nylon is low; but, m pyrolytic studies, nylons behaved very erratically, having activation energies and rates that depended upon the method of sample treatment. On the other hand, the pyrolytic Ea for polypropylene is high I n view of the many expenmental effects possible in both pyrolytic and regression rate studies, the agreement is very good I n studies of polymer pjTolysis, the weight loss is measured isothermally and the sample is uniformly held at a given temperature, i e , there is no temperature gradient. The rate of weight loss is not often dependent upon the conversion to the first power but may, as in polyethylene and polystyrene, have a contmually varymg dependence upon conversion and hence upon the weight of sample remaimng However, the rate of weight loss imtially (—dw/dt)o is always observed to be pro- portional to the imtial weight Wo At any other pomt of conversion dunng a given expenment, we also find that i-dw/dtJc^'Kwc. Now, what one expects to find in regression rate studies is that the rate of regression (—dl/dt) would be related simply to rate of weight loss by -dl/dt= -dw/pAdt=KwJpAa = Kh, T A B L E 11 Comparison of activation energies from pyrolysis of polymers (Ref 1) and regression rates (Ref 19) Polymer Pyrolysis Regression rate Methyl methacrylate 32-52 42-48 Ethylene 72 73 Propylene 58 40 Formaldehyde (Delnn) (>20) 24 Nylon 6,6 42 65 I C R P G Urethane 30

218 F I R E R E S E A R C H where p and A are density and burning area, respectively By assuming a constant active weight Wa at the measured surface temperature as the source of volatile fuel, we can equate dl/dt=Kla, where la is the active thiclcness for the pyrolyzmg skm. It is interestmg to note that, with available data,"'"-' one finds that h values of 1-2X10-=' mm for polymethylmethacrylate, and 1-2X10"' for polyethylene, and are relatively constant with regression rate. A final comment is that while polymers volatiUze at a rate proportional to the weight of matenal, hqmds (and also crystals) vaporize at a rate proportional to surface area This difference should perhaps be kept m mind, smce in many other aspects amorphous polymers are considered analogous to hqmds C O N C L U D I N G DISCUSSION There is a considerable amount of pertinent thermochemical and kinetic data on reactions in the condensed phase of polymer melts that is apphcable to problems of flammability and combustibihty I t appears, however, that no extensive effort has been made to apply available data and also it is not clearly evident what in- formation IS important and required It seems evident that the main problem is the lack of experimental procedures and theory for characterizing, measuring, mvesti- gatmg, and standardizing fire phenomena Fire damage is often the result of un- foreseen factors or unanticipated situations The areas of polymei degradation which are most pertinent to the fire area would include 1 Volatile products at temperatures and rates near those occurring during bummg 2 Heats of decompo-sition measurements, particularly m the higher tempera- ture region 3 Studies of oxidant induced weight loss and oxidation or combustion under oxygen-deficient conditions 4 The variation in combustibility and volatile products with aging of materials One can calculate that the rugs and padding in a house have enough combustible fuel to consume fifty times the oxygen in the house Also, if only half the oxygen is consumed and some asphyxiative*" gases are produced, not to mention carbon monoxide, the condition is l e t h a l T h e important factor is obviously the speed of burning and igmtibihty of the matenals. Such properties are likely to be enhanced with deteriorated or aged materials 5 Study of pyrolysis, combustibility, and igmtibility of mixtures of matenals, blends, and bulk samples simply placed m the same furnace, for instance The volatile species from one substance may react vigorously with another sub- stance or its volatiles Another not unhkely possibihty is that volatiles from one polymer will mitiate or catalyze the decomposition of another polymer, producing fuel of low flash point, while a third polymer provides igmtion bv rapid exothermic decomposition. Our basic knowledge is quite hmited m the area of matenals compatibility for fire safety This is beconung an extremely pertment item because modern tech- nology IS introducing an extreme vanety of matenals into the field; consider, for

ABSTRACTS AND R E V I E W S 219 example, the variety of substances in textiles, garments, composite structures, and building materials. References I S L MADORSKT Thermal Degradation of Orgamc Polymers, Interscience, 1964 2. L A WALL Chap V tn Analytical Chemistry of Polymers, Part I I (G M Khne, Ed ), Inter- science, 1962 3 L A WALL AND J H FLTNN Rubber Chem and Technol SS, 1157 (1962) 4 L A WALL 'Tyrolysis of Fluoropolymere," Chap 12 ire Fluoropolymers (L A Wall, E d ), Interscience, to be published m 1971 5 H H G JELLINEK Degradat'On of Polymers, Academic Press, 1955 6 N GBASSIE Chemistry of High Polymer Degradation Processes, Interscience, 1956 7 R M JosHi AND B J ZwOLiNSKi Chap 8 tn Vmyl Polymenzation (G E Ham, Ed ), Marcel Dekker, 1967. S L A WALL S P E J 16 (1960) 9 H Y u AND L A WALL Polymer Preprints (ACS) 6, 940 (1965) 10 L A WALL, L J FETTERS, AND S STRAUS Polymer Letters 5, 721 (1967) I I P R E J COWLEY AND H W MELVILLE, Proc Roy Soc 2^0, 461 (1951), 2;^, 320 (1952) 12 L A WALL AND S STRAUS unpubhshed 13 E M BBVILACQUA Chap 18 tn Autoxidation and Antioxidants, Vol I I (W O Lundberg, E d ) , Interscience, 1962 14. M TBYON AND L A WALL Chapt 19 ira Autoxidation and Antioxidants, Vol I I (W O Lundberg, Ed ), Interscience, 1962 15 R FRIEDMAN "Aerodynamic and Modehng Techniques for Prediction of Plastic Burmng Rates" 16 S L MADORSKT, V E HART, AND S STRAUS J Research Natl Bur Stds 66, 343 (1956) 17 B M DMrrBiBV, 0 A KOCHBTOV, V B ULTBIN, AND A S SHTEINBERG Fiz Goren Vzryva 5, 26 (1969), Translation No 2426, The Johns Hopkms University—Apphed Physics Laborar tory, Silver Sprmg, Md, July 1970 18 L A WALL, J H FLTNN, AND S STRAUS J Phys Chem 74, 3237 (1970) 19 W A BLAZOWSKI, R B COLE, AND R F MCALEVT, I I I Tech Rept to ONR M E - R T 71004 (June 1971) 20 I N EiNHORN, J D SBADER, C M M i E L F E r r n , AND W 0 DRAKE Tech Rept NASA U T E C - M S E 71-039