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.

CONTRIBUTIONS OF ANALYTICAL CHEMISTRY TO FIRE PROBLEMS H O M E R W. C A R H A R T Naval Research Laboratory, Waahtngton, D C I N T R O D U C T I O N One can readily picture early cave man, havmg "caught" a fire, stanng at the flame and wondermg what it was. With time, more cmhzed man would come to consider flre as one of the four basic elements from which all matter was composed. His lack of understandmg of fire would eventually lead him to mvent a special property, phlogiston, in order to explain the visible flame. I t was not until the latter part of the 18th century, however, that, with the beginnmgs of modern chemistry and the begmmngs of guanhtative chemical analysis, man began to ac- quire imderstandmg of the chemical nature of fire. Smce then, man has come a long way m developmg his understandmg of fire, but the complexities of the chemi- cal processes mvolved, their speed and their dependence on composition and environment leave not-so-pnmitive man still stanng at a flame and wondermg what it really is. Fure IS exceedmgly important to man m two ways, first, because he has learned (and seems to be doomed to have to contmually releam) that it can be so very destructive, and, second, because he has partially learned to harness it to do his biddmg, and with it is now hterally movmg mountams. For these reasons he has apphed a great effort to learn more about it, and still is. In order to imderstand what fire is, and how it behaves and propagates, a great number of scientific disciphnes must be invoked, not the least of wbach is analytical chemistry. Again, however, because fire is so complex, it is exceedmgly difficult to separate these disciphnes discretely, and in looking at the vanous approaches to the study of fire it is readily apparent that analj^ical chemistry is not only the handmaiden of other disciphnes but is intimately interwoven with them. The determination of products, mtermediates, imstable species, highly reactive species, excited states, etc., dunng the combustion process must contmue to remain one of the cntical keys to greater understandmg of the mystenes of fire, imderstandmg which, hopefully, will lead to greater control of fire. Though analytical chemistry has already contnbuted a great deal towards our understandmg to date, one can readily see that the challenges to the analyst and analytical chemistry m the future are even more demandmg and excitmg. W A N T E D A N D U N W A N T E D F I R E Fires come in two great classes—wanted and unwanted. In the former category fall man's comforts and joys, fires for cooking and warmth, fires for power and locomotion, fires to go to the moon and back, but m this category also fall man's hates, fires for battle, to kill, to bum property and maim humans. I n the second 220

ABSTRACTS AND R E V I E W S 221 category fall man's fears, the uncontrolled burning of his home, his forests, his goods, his loved ones. And though much of the chemistry of a fire may remam the same, the approach m studymg it—or controlhng it—may take a different tack, dependmg on whether the fire is useful or destructive, and dependmg on what the mdmdual researcher wants to learn. Because of the complexity of fires, the re- searcher often chooses, and may limit himself to, a certam aspect of combustion to suit his own cunosity or purpose. For example, he may want to study the chemi- cal reactions occumng m the preflame stages of combustion so that he may learn how to control succeedmg stages, or he may want to study the flame itself with all its "hot" species such as free radicals, ions, etc., m order to get at the kinetics and mechamsms of combustion, or he may want to study products formed in fires, particularly those that might be highly toxic, in order to evolve techniques for protectmg life, or he may be interested m determimng the best way to suppress a fire, or he may want to learn why some matenals igmte spontaneously under one set of conditions but not under others which on the surface appear to be more severe. The complexity of fire is well illustrated by this apparent anomaly as shown m Table 1 m which it is seen that with mcreasmg temperature a fuel may exhibit altematmg zones of igmtion and apparent non-igmtion. Such behavior is hardly m keepmg with the classical concept that the rates of chemical reactions increase with increasing temperature. The ultimate goal m all of these studies however, 18 to achieve better means of enhancing wanted fires or controlhng or prevent- mg unwanted fires. The study of fire is further compbcated by the fact that there are so many different kmds of fires, based not only on different fuels and oxidizers, e g , rockets, T A B L E 1 Zones of spontaneous igmtion (pos ) and nonigmtion (neg) (m air) Navy special fuel oil (Sample NRL-71-7)* n-Hexadecene** Temperature range ("F) Result Temperature range (°F) Result Below 473 Neg Below 493 Neg. 473-479 Pos 493-505 Pos 480^86 Neg 506-518 Neg. 487-542 Pos 519-613 Pos 543-548 Neg 614-691 Neg 549-571 Pos Above 692 Pos 672-580 Neg 581-585 Pos 586-632 Neg 633-659 Pos 660-666 Neg Above 667 Pos • ASTM Method D 286-66T •* 21 ml. reaction chamber

222 F I B E R E S E A R C H metals, metal-halocarbon, etc., but also baaed on enviroiunent (with the same fuel and oxidizer), e g., an explodmg oil tanker and an automobile engme. Thus, the demands on analytical chemistry to evolve appropnate techmques apphcable to study of the particular fire m question become very acute. In this regard, re- searchers have been mventive mdeed, and all sorts of approaches have been tried, many very successfully. The hterature is replete with discussions of these, but to a large extent the approaches used are described m articles and books devoted to what is learned from the analytical results rather than to the analyses themselves.* T E C H N I Q U E S I t is unpossible to cover all the techmques used and theur many vanations. A few are mentioned below.** Wet Methods Although m the past few decades analytical chemistry has tended to become more and more highly mstnimented, wet methods are still used, for example for followmg the concentrations of peroxides. Orsat Both wet and dry. Tedious but useful when only a few species are mvolved Old wet method for O2, CO2, etc., still earned on some ships for flue gas analysis. Afoss Spectrometry An exceedingly useful tool for the analysis of both stable and imstable species. Particularly useful when coupled to a small computer. Unstable species, such as ions, can be measured by having the inlet to the spectrometer be the du'ect probe into the flame (Knudson probe). Gas Chromatography A tool developed m the past decade and a half which has proven to be exceed- mgly useful for analysis of stable species It is imusually versatile, accurate and sensitive and can analyze very small samples. Coupled to a small computer it can yield a great deal of information relatively qmckly. An example is shown m Figs. 1 and 2, Fig. 1 bemg the chromatogram of a sample taken from a butane cool flame and Fig. 2 the computer prmt-out, which mcludes retention times and concen- trations. * For this reason, no attempt has been made to cit« references or to give a bibliography The extent of the ht«rature on the subject is shown by the many well known books on fire, combustion, etc, by the biennial pubUcation of the Symposia (International) on Combustion, by the vanous journals on the subject, by the extensive abstracts m Fire Research Abstracts and Reviews, etc •* The author is highly indebted to the excellent treatise 'Tlame Structure" by Fristrom and Westenberg for much of this part of the paper, and the reader is referred to this work for more details.

ABSTRACTS AND R E V I E W S 223 n 20 24 TIME, MIN. F i a 1 Chromatogramof a sample taken immediately downstream of a butane cool flame Gas Chromatography-Mass Spectrometry The combination of a gas chromatograph and a mass spectrometer m tandem gives the analyst an exceptionally powerful tool for stable species, especially for complex mixtures. Normally m a very complex mixture, such as that occurrmg m early stages of a flame or from pyrolysis or mcomplete combustion, data from a mass spectrometer alone is difficult to analyze But, if the mixture is first passed through a chromatograph, the sizes of the eluted peaks tell how much of a given component is present, the retention time gives an mdex of what it might be, but if a mass spectrometer is coupled to the chromatograph essentially as though it were a detector, identification is usually imeqmvocal. Even m mixtures where separation I S mcomplete on the chromatographic column, the few components that may be present m a smgle peak can still be identified by the spectrometer. Absorption Spectroscopy This is a particularly useful tool to look mto the flame itself because unhke many sampling probes it does not disturb or distort the flame. It has proven par- ticularly useful for study of free radicals, especially OH. Absorption spectroscopy, both tr and uv (very httle m the visible) has also proven useful for condensed samples taken from flames by probes, etc. Flame Photometry Particularly useful for identification of species that emit hght m the flame as a result of the reactions taking place, e g., excited E C H O m cool flames. I t gives considerable insight mto these reactions and where they occur, but data are diffi- cult to mterpret quantitatively.

224 F I B E H E S E A K C H METHOD? C A L L E S T D 14 R E A D Y METHOD? METD E S T D 14 RUN? 10(5 PORAPAK S E X H A U S T ANAL COMPONENT R T . T M . ABSVOL% ++++++++ 0^17 2 6 4 1 * E - 7 ++++++++ 0 1 1 7 7 2 1 0 * E - 7 iffiTHANE 0 1 4 3 1054-!=E-4 ETMEME 0 2 4 7 3 U 3 - E - 4 E T H A \ ' E 0 2 8 8 7 0 9 2 * E - 6 P R O P S N E 0 6 8 5 1 5 7 8 * E - 4 PROPA>IE 0 7 4 6 4 5 2 9 * E - 6 A C E T A L D E 1 3 5 7 S 9 0 9 + E - 4 I B U T A N E 1 6 0 4 2 S 7 1 + E - 5 I B U T E N E 1 7 0 5 6 2 8 0 - ' E - 4 N3UTANE 1 8 2 5 8 1 1 8 * E - 3 22DMPH0P 2 4 6 1 2 3 2 5 * E - 5 ++++++++ 2 5 2 0 1 1 4 8 * E - 5 P R O P I A L D 2 6 0 4 3 1 4 8 ^ E - 5 A C E T O N E 2 6 8 7 3 2 1 5 t E - 5 9MBUTANE 2 8 6 9 2 2 6 5 * E - 5 D E T E T N E R 3 0 1 1 8025=^E- 6 ++++++++ 3 3 1 9 6 5 6 9 A E - 6 ++++++++ 3 6 3 5 9 3 3 l ! ^ E - 5 T H F U R A N 3 7 7 5 4 5 3 8 " - E - 5 I f l E X E N E 4 0 2 6 8 2 4 1 * E - 6 ++++++++ 4 2 2 0 1 6 8 7 + E - 5 »cAEND O F R U N : = * ^IETHOD? F I G 2 Computer pnnt-out of data obtained m Fig 1 Tracers and Scavengers A variety of tracers can be used very successfully for studying reactions in fires and combustion. These include the use of stable isotopes (e g., deuterated com- pounds, compounds containmg 0", etc.) or radioactive isotopes, or compoimds added to a flame, such as N2O which reacts with both H and 0 atoms (to give N2 and NO) and hence can be used to measure theu- concentrations. Scavengers can also be used, especially in microprobes, m which the scavenger is added m excess immediately after sampbng at the onfice to quench radicals or other un- stable species.

A B S T R A C T S A N D R E V I E W S 225 ESB Spectroscopy Electron spin resonance can be used to measure free radicals and atoms in flames due to their having an impaired electron. Direct measurements (flames m the resonant cavity) are hard to interpret but probmg has been successful. Photography and Schlxeren These can be very useful for recordmg positiomng of flames, showmg gradients, fast reactions, mtensities, etc , and are used extensively. Ions The electrical properties of flames and fires are particularly mtriguing and the detenmnation of charged species is a subject all its own. Many different kmds of measurmg devices have been used, mass spectrometry, colhsion cross sections, Langmuir probes, radio-frequency absorption, photography, etc. Special studies have been made of ions m rocket exhausts and theu: effect on communications black-outs. TOA Thermal gravimetric analysis—coupled with gas analysis—is a useful tool for measurmg pyrolysis, particularly of sohds, and can be used to give mdices of flammabihty potential and behavior. Reactors and Burners The Vertical Tvbe Reactor. Many, many different reactors, burners and other devices have been used for studymg fires and flames Of special mterest (at least to the author, smce it was developed m his laboratory) is the Vertical Tube Re- actor in which flames can be separated mto three distmct lummous stages separated m space by non-lummous regions. This is illustrated m Fig 3. The beauty of such a reactor is that it makes studymg and probmg the flame much easier, and also allows for the insertion of scavengers or reactive species mto the different stages of the flame processes. Probing Probing to obtam vahd samples is exceedingly difficult. In samphng flames by use of probes, one must not disturb the flame, and must also assure hunself that reactions do not take place m the probe itself. Probes can be designed and used to collect meanmgful samples, such as the microprobe described by Fnstrom and Westenberg (loc. cit.), but the researcher must be forewarned that it is not easy. A P P L I C A T I O N S A N D S T U D I E S The' apphcations and contributions of analytical chemistry to fire problems and research are legion. It must be emphasized that analytical chemistry per se may not be the end objective but without it the end objective could not be achieved. Only a few examples can be given and m only one or two of these wiU we permit ourselves the luxury of an excursion mto detaal.

226 F I R E B E S E A B C H E L E C T R O D E J A C K E T O R A N G E F L A M E B L U E F L A M E N O N - L U M I N O U S Z O N E C O O L F L A M E D I F F U S E R E L E C T R O D E N 2 - I - F U E L O 2 V E R T I C A L T U B E R E A C T O R Fio 3 Schematic of the vertical tube reactor showmg relative positions of flame stages &ajely Safety is an area where analysis has had a very important mipact. For example, as man has acqiured a better imderstandmg between flammabihty charactenstics of hqmd petroleum fuels and their composition he has been able to msist that new designs m refinery techniques be put mto practice so that the fuels would be safer. One need only remember the rashes of fires that used to occur many years ago with kerosene, and even with gasohne, until by virtue of greater control of composition, as measured by different analytical techmques, the properties of such products have been modified so as to make them much safer The story keeps repeating itself as new fuels and oxidizers are developed, especially high energy ones, which for that very reason are often unstable and hence dangerous. The vanous anal3d;ical techmques used to measure properties of fuels and oxidizers often end up as requirements m the specifications used for the purchase and sale of the products, and m order that there shall be no misunderstandmg or disagree- ment between seller and purchaser, these techmques are very carefully worked out. Indeed, whole societies, such as the A S T M , devote a great deal of attention to this very pomt, and are contmually developing new means for measurmg not only composition or performance, but safety as well.

A B S T R A C T S A N D R E V I E W S 227 Fire Suppression Again, analytical chemistry has played a large part in achievmg better materials and techmques for suppressing fires The development of commercial materials such as "Light Water" and "Purple-K-Powder" are greatly abetted by acquirmg understandmg of what is needed composition-wise. Even today, the composition of "Light Water" concentrate is bemg radically modified because analysis has shown that problems in corrosion require it Analytical chemistry has also played a sig- mficant role m the development of vapor phase inhibitors that mterfere or react with free radicals and other cham earners in the combustion process. FlammabilUy This I S related to safety but will be treated separately because it is so cntical in creatmg a potential for fires For example, the characteristics of the vapor space above a hquid petroleum product must be known (hence measured) if one is to design safer fuels and fuel handhng and storage systems A petroleum fuel does not bum in the hquid phase, it must evaporate first. But petroleum fuels are complex mixtures of hundreds of components, so that neither the composition nor properties are the same from batch to batch of a given product. Yet, knowledge about the vapor space is cntical if prevention of fires is to be achieved by control of that space, control which has not always been achieved as attested to by many disastrous explosions, especially m tankers. The Navy is concerned about the properties of JP-5 jet fuel smce an aircraft earner may contam as much as 1 5 miUion gallons of it, and it is not earned m a protected part of the ship. For safety's sake, the specifications for JP-5 state that it shall have a minimum flash pomt of 140°F But this is not a fixed property, and can vary dependmg on the ullage, as can be seen m Fig. 4 This shows the m i n i m u m temperature at which the vapor space is flammable for two JP-5's and for a pure hydrocarbon, n-undecane. The vapor concentration above n-undecane is fixed, smce it is a pure compound, and, hence, so are the flammabihty properties as shown m the figure But the JP-5's, bemg mdefimte mixtures, have varymg flam- mabihty properties with ullage, and are not consistent as shown by the fact that one fuel has a small variation with ullage whereas the other one shows a very marked effect Indeed, extrapolation of the curve for this latter fuel shows that as one approaches zero ullage, the vapor space above the fuel would be flammable at 120''F This shows that this particular fuel, even though it more than meets the specification with a measured flash pomt of 144°F, can be flammable, and hence hazardous, at a much lower temperature. Another unportant pomt from the standpomt of safety deals with the stratifica- tion of fuel vapors m large tanks There is often a barrel or two (or more) of product left m a tank after discharge from a tanker During discharge, air replaces the fuel. I f the product is highly volatile, such as gasolme, JP-4, naphtha, or crude, nght after discharge there will be a very thm layer of fuel vapors unmediately above the hquid that is too fuel-rich to bum Just above that there will be a small layer that is m the flammable range, and above that, the vapor is too lean to bum. With time, the two bottom layers expand until eventually (if enough hqmd exists) the whole space is too nch to bum. A few measurements have been made m real-

228 F I R E R E S E A R C H •C 100 90 80 70 S 60 r 212 O n-UNDECANE ^ JP-5(Na4l71 • J P-5 (No 424) A FLASH POINT (PENSKY-MARTENS) ..-•tSCi" Q n-UNDECANE JP-5(N0 4I71 JP-5 (No 424) 50 30 JP-5 (NO 424) --•3P-5(N0 417) 122 86 SO ULLAGE (%) 70 BO 90 100 F I G 4 Effect of ullage on the lower flammabihty temperature hmit. hfe Situations m tankers, usmg sunphfied techniques, and these have mdicated that it takes days for this process to occur, but details about how large the flammable layer is, where it is, and how long it lasts under the wide variety of fuels and prac- tices that are used are mostly imknown Indeed, the paucity of good sobd informa- tion on this subject is amazing, yet it certainly is important, especially for ships running empty, as shown by the explosions of the three super-tankers off the coast of Afnca. There is a requirement m the specifications for several Naval fuels, called the "explosiveness" test. Essentially, the requirement states that, at 125°F, the con- centration of fuel vapors above the hqmd m a contamer should be less than 50%^ of the lower ffammabihty limit. The present method for this test uses an explosion' meter as the means of analyzmg the vapor concentration, which is also the method commonly used m the field But the response m an explosion meter is dependent on the structure of the hydrocarbon (meters are usually calibrated with n-hexane) and so a new and more accurate method was desirable. In our laboratory we have designed a new analytical tool for measurmg vapor concentrations as a function of their flammabihty (the "flammabihty mdex") based on the hydrogen flame lomzation detector, so popular m gas chromatography. A hydrogen-air flame does not generate ions, but if organic vapors are added, chemi-ions are formed. Most fortmtously, for hydrocarbons the amount of these ions is a function of the molecular weight and concentration present, so that effectively the ionization detector counts carbons present. But, as shown in Fig. 5, for most hydrocarbons (and their mix-

A B S T R A C T S A N D R E V I E W S 229 tures) these are also related to flammabihty. Hence, an analytical scheme has been developed, based on the hydrogen flame ionization detector which measures flammabihty index directly Although still in the bread-board stage, we have already demonstrated the accuracy of the method, as shown in Tables 2 and 3. I t 18 easily seen that this new analytical tool will be apphcable to the study of all the situations descnbed above Detection Devices for the early detection of fires and/or activation of alarms and ex- tmgmshers exist which are based on chemical characteristics of fires and which semi-quantitatively measure such properties as heat, lomzation, hght, etc. They are semi-quantitative m the sense that they must sense a given amoimt or rate of change of a fire property before triggering so as to not give false alarms, and yet do this early enough to allow for corrective action. Such devices are based on thermocouples, low-meltmg sohds, photocells, ion probes, distortion of laser beams, etc , but there is still a real need for better and better detectors. Efficiency and Power Even though mtemal combustion and jet engmes have been developed largely empirically,* extensive use of anal3iiical chemistry has been made m the study and control of the "wanted" fires m such engmes There is always the desire to squeeze that extra erg of energy out of an engine and one must determine what is happemng, and where, m order to do so. The location and eflSciency of bummg m boilers is very important if fuel is to be used economically and the design and operation of sprayer plates is controlled by this. Also, analytical chemistry has been very important to the study of the mechanisms by which tetraethyl lead reduces knockmg, and very clever probmg devices have had to be developed to obtam samples from high speed engmes. There are many other examples T A B L E 2 Flammabihty index of pure hydrocarbons in air Flammabihty index (E) Concentration Lower Flammabihty Calc Hydrocarbon C, % v/v L, % v/v Expt (E=C/L) Methane 1 66 5 0 0 26 0 33 j-Butane 0 45 1 8 0 25 0 25 z-Butane 0 90 1 8 0 60 0 50 z-Butane 1 35 1 8 0 75 0 75 7i-Hexane 0 499 1 2 0 42 0 42 n-Octane 0 200 0 89 0 22 0 22 * Indeed, had we waited until we really understood the full chenustry of combustion before we designed these engmes, we would still all be walking, such is the complexity of the combustion process

230 F I B E B E S E A B C H E = C / L (1) p = k^nC (2) Eliminating C between Equations (1) and (2): E = p/k^nL (3) nL = kg = constant (4) E = p/kjkg (5) E = kp (6) W h e r e E = Flammabihty Index C = Concentration of Hydrocarbon Vapor (%v/v) L = Lower Flammabihty Limit (% v/v) p = Response of Flame Ionization Detector n = Carbon Number k j , kg, and k are constants F I G 5 Equations showmg relationship of flammability index with flame lomzation detector response T H E F U T U R E The analytical chemist now faces many demandmg but mterestmg challenges m man's mvolvement with fires. One of his primary aims, obviously, is to obtam better control over both wanted and unwanted fires (and their products) by achiev- mg better understandmg of them. The foUowmg are some examples of areas of challenge. PolliUion Now that the pubhc has "discovered" pollution and is pressmg Congress mto domg somethmg about it, techmques for measurmg pollution from all sorts of fires will become m much greater demand. Most of these will have to be aimed at the "wanted" fire (mcludmg those designed to study and control unwanted fires or to tram people m combattmg them) smce people are still conditioned to be somewhat philosophical about imwanted fires and to accept their consequences as still bemg part of the disaster. Therefore, we are talking largely of wanted fires for the generation of heat and power. The trends m demands for the future are already quite obvious As the hd on pollution is clamped ever tighter, there will be a need to develop analytical techmques that will be more and more sensi- tive, accurate, simple, cheap and rugged. For example, we are beconung much

A B S T R A C T S A N D R E V I E W S 231 more mterested in simple devices for measurmg not only the total oxides of mtro- gen, but how much of which ones. I t is also becoming important that we do this accurately down to the fractional ppm range as opposed to, say, determinmg CO2 m percentage values. For some materials, such as ozone, we will be mterested m the parts per bilhon range The tune is commg when pubhc authonties with a mmimum of techmcal trainmg will want to have such devices available for en- forcement purposes, for example at automobile mspection stations The problem of disposal of wastes is also becommg more acute and the possi- bihty of mcineratmg them on a large scale will have to be considered. Instead of open bummg, special mcmerators will be reqmred and a high efficiency of bummg becomes mandatory. Considenng the mcreasmg complexity of materials and the increasmg use of elements and molecular moieties not found m nature, analytical methods for more varieties of pollutants will be needed, m addition to the present CO2, CO, NOi, SO2, H X , H C N , smoke, orgamcs, etc For example, there is an mcreasmg use of fluorme m articles of commerce. What and how much do we look for if these find their way mto mcmerators? If one looks at the trends m concentrations of pollutants that will be allowed, it becomes apparent that the only direction they will be gomg wiU be down. This results from increased knowledge of subtle nonlethal effects of low concentrations The need for mcreased sensitivity m analytical methods is apparent. The need for continuous monitormg (as m stationary power plants) also becomes apparent And as we contmue to discover the need for control of additional matenals (e g., certam types of organics), the need for analytical techniques for these new substances will also constitute a challenge. T A B L E 3 Flammabihty index of hydrocarbon mixtures m air n- Alkanes (% v/v) Other hydrocarbons Percent flammabihty index CH4 CeHi4 C7H1J CsHu Cflll20 CJ0H22 Expt Calc. 1 2 — — 0 22 0 20 0 055 71 76 1 2 — 0 25 0 22 0 20 — — — 85 90 — — 0 50 — 0 20 — — — 63 64 — 0 28 0 25 0 22 0 20 — — — 87 85 — 0 33 0 29 0 26 0 14 0 13 — — 105 105 — — 0 46 — — — — 0 078% Decahn 49 49 — 0 28 0 25 0 22 — — — 0 063% n-Octene-2 70 71 — 0 28 0 25 0 22 — — — 0 062% 2,3,4- 70 71 T M P — — 0 42 0 37 — — 0 14% Benzene 94 94 — — 0 25 0 22 0 20 — — 0 14% Meth cy- 77 78 0 32 clopent — 0 29 0 13 0 049 0 076% n- 94 95 Hexene-1

232 FIRE RESEARCH Toxicity Many people are killed by fires due to "suffocation," "asphyxiation," "smoke inhalation," etc., without the real cause of death bemg pm-pomted. Also, it has been observed m a number of fires, particularly m relatively confined spaces, that some men (without respiratory protection) will fight the fire, help clean up and, except for the usual symptoms of exposure, appear to be m good health only to die of pulmonary edema a few days later. Agam the true culpnt may not be known. Although the toxic compounds generated may be essentially the same as those that must be considered under pollution, we are now concerned with acute rather than chrome toxicity. The problems of analysis are different m that sensitivity is not as important, but the vanety of matenals may be greater due to mcomplete combustion especially of materials that may not bum "clean" such as plastics, pamts, polymers, etc , to yield H C l , H C N , added CO, H F , N0», unsaturated aldehydes, other orgamcs, etc Also the concentration of potentially lethal agents may change markedly and qmckly as new and different matenals are attacked by the fire, or as the oxygen gets used up The challenge to the analytical chemist to give us the nght answers is very real, answers that we can rely on to provide m- creased and qmck protection to the people. For example, the Navy now issues chemical protective masks to all of its men on ships. Can these be used and under what conditions, and when should they not be used? Should additional stop-gap protection be provided' These are important issues. Unusual Atmospheres Man seems to be determmed to encapsulate himself and to place such capsules m a very hostJe environment, environments that would be lethal if man were to be dumped mto them suddenly. Examples are space ships, submarmes, sea-labs, and even commercial jet aircraft at 30,000-pIus feet.* We are even senu-encapsulat- mg ourselves m not-so-hostile environments but m such fashion that egress is difficult Access to large amounts of fresh air are himted because of recirculation of air-conditioned air Ships and high-nse buildmgs are examples of this class. A fire started m such systems could easily be feeding on highly vitiated air m a very short tune, and the chemical course of the fire could easily deviate markedly as the fire progresses. In submarmes, space ships, divmg capsules, hyperbanc (Oj-ennched) medicme, etc , the actual composition of the atmosphere is con- trolled—it IS no longer air. Nitrogen may be substituted by hehum, oxygen con- centrations might vary from 4% to 100% and pressures from 7 to 0.3 atm as shown for SeaLab I I and Apollo m Table 4.** Yet, most of our fire experience * Most of us who nde these au-craft don't stop to reahze how hostile the environment really is just across the thm skm of the aircraft should we suddenly be dumped mto ita really sub-zero temperatures, too httle oxygen partial pressure to support most of us, a violent depressunzation and a blast of what air is there gomg by at 600 mph ** Many of us hke to compare the combustion of foods m the body with combustion of orgamcs m fires, in that a certam level of oxygen is needed for both However, once we get mto these unusual environments, a very important difference appears In order to sustam life comfortably, we need a partial pressure of oxygen of about 0 2-0 3 atm, regardless of the total pressure This means that the concentration of oxygen must be adjusted accordmgly But from the standpoint of fires, it is the concentration of oxygen that is important, much more so than the partial pressure Thus, in the environment inside SeaLab I I , the aquanauts tried to hght matches but with no success— 4% oxygen is much too low to sustam such a fire

ooo A B S T R A C T S A N D R E V I E W S T A B L E 4 Oxygen levels in various environments Environment Total pressure (atm) Concentration of 0 , (%) Partial pressure (atm) Apollo Normal S E A L A B I I 1/3 1 7 100 21 4 0 3 0 2 0 3 has been m air. We have already leamed much about fire behavior m such en- vironments through studies by NASA, Air Force, Navy, FAA, and others. One of the outstandmg features of such fires is that rates of bummg are markedly m- creased even with relatively small oxygen ennchments, and that materials that appear to be flame resistant m air (e g , fireproofed fabncs) wiU bum merrdy at higher oxygen levels In 100% oxygen, even Teflon will "bum," and common combustibles bum ahnost explosively as was demonstrated so disastrously in the Apollo fire m 1967. Frres m capsules are particularly bad because there is no place for the occupants to run. The composition of fire products and their potential toxicity are also gomg to depend on the particular atmosphere Problems of detec- tion, analysis and control of such fires (partly by studymg fire properties as a function of composition of matenals) again challenge the analytical chemist to provide new techmques and new knowledge. Control The challenge of control of fires, both wanted and unwanted, contmues to exist. Means to achieve control can be reached either empirically or by developing understandmg and knowledge I t is in the latter area that analytical chemistry can play a vital part. As man contmues to try to improve the efficiency of his engmes and power and heatmg plants, he must learn still more about what is happemng and when and where In studymg suppression of unwanted fires, he aims to leam more of the mechamsms mvolved so that he can leam to modify the course of the oxidation (e.g., what free radicals are formed, how they propagate cham reactions, which ones are most amenable to reaction with mhibitors, etc ) . Detection In the case of unwanted fires, the sooner these are discovered, the better the chance of nunimizmg their damage. Therefore, the challenge contmues to exist to develop new and better detectors, and the analytical chemist must contmue to improve, adapt and mvent new techmques. Probing, Sampling, Doping, Scavenging I t has already been mentioned that probmg, samphng, "dopmg" (m the sense of addmg selected reactive species into chosen parts of the flame stmcture) and scavengmg (m the sense of removing reactive species from a flame by reaction to give identifiable products) are all very powerful analytical tools for study of flames. I t has also been mentioned that these are not easy. But, if properly designed and

234 F I B E B E S E A B C H apphed, they can give a wealth of information not otherwise obtamable. Different forms of each are possible, for example, probmg might mclude use of laser beams, ion probes, sound pulses, etc. Samplmg unphes more than the collection of stable species. I t includes sampling directly mto measunng devices such as the Knudson leak mto a mass spectrometer. Dopmg can be by a very wide vanety of agents and can be mtroduced mto different parts of the flame (the Vertical Tube Reactor lends itself particularly well to this). Scavengmg, bemg closely related to dopmg, can also be by a wide vanety of agents (to essentially "freeze" a reaction or a species in place so that it can be sampled and identified). Because fires are so complex, and parts of it so transient, all of these techmques must be designed and applied with great judgment so that the true events are not masked by spurious second events caused by the technique itself. But, because the pay-off is great, elaboration of such techmques will not be demed, thus posmg a real challenge to the imagmation and mventiveness of the analytical chemist * The "Non-Fire" A great deal of effort, regulations, monies, eqmpments, etc., are expended to mmimize the ravages of unwanted fires. The ultimate solution to the problem of course, is total prevention, that is, use materials, practices, etc, that will not allow a fire to happen m the first place Or, to put it m another way, to have "non- fires." One might define a "non-fire" as a fire that would normally have happened at a given time and place, just as fires do today, except that this particular event did not happen because we had developed enough knowledge (and had used it) to mclude mto the system the proper matenals, geometry, practices, behavior, etc to prevent it. One trouble with "non-fires" is that it is exceedmgly difficult to prove that we did mdeed have a "non-fire," because hfe goes on as before, and "who notices?" For this reason, it is much more difficult to sohcit research support from sponsors to study non-fires, than it is to get support for study of, say, ex- tmgmshment after a particularly disastrous conflagration. This, coupled with a certam amount of mental lazmess or lack of imagmation on the part of us re- searchers, tends to push us mto expendmg our major efforts and dollars m research- mg the last big fire (fires of the past) rather than the next big fire and "non-fire" (fires and "non-fires" of the future) It is another way of saying that the stop sign at the crossmg goes up after the accident In a broad sense one might also thmk of a "non-fire" as bemg a fire that had an mcipient beginnmg but was quenched so quickly that no damage resulted These, of course, would result because automatic detectors and extmgmshers were present and worked Obviously these would be and are installed m particularly hazardous areas where we anticipate trouble. In order to have "non-fires" we certamly need a lot more knowledge, acquired through research and development m many, many disciphnes, not just analytical chemistry, and m particular how to apply the knowledge thus gamed mto our everyday hvmg and environment. The control m the composition of materials, * In this regard, the individual who pursues such endeavors may think of himself as some other kmd of chemist or mvestigator and is really looking for answers that these techmques will yield, but in the process, whether he knows it or not, he la an analytical chemist.

A B S T R A C T S A N D R E V I E W S 235 how these are used, how they are juxtaposed, how they are handled, what the contribution of the human element is, the precautions and preventive measures taken, etc., etc., obviously all come mto play m desigmng a "non-fire," and must be mcluded m our attempts to reach the utopia of "non-fires," a utopia which is certamly worth stnvmg for. In conclusion, wouldn't it be great if we could say that next Tuesday afternoon we will have a non-fire at the Blank refinery and on Thursday mommg Tanker X will not blow up and, perhaps that at this very moment, m this very buildmg, we are havmg a "non-fire," because we were smart enough?