Mines and Bunkers: Volume 10, Fire Safety Aspects of Polymeric Materials (1980)

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.

APPE N D I X C SMOKE H AZ ARD S AND ME ASURE ME NTS OF SMOKE OPACITY Excerpt from J. R. Gaskill, "Smoke Hazards and Their Measurement; A Researcher's Viewpoint, Journal of Fire and Flammability, 4 ( 1 973) : 279-298 

M I N E S A N D B U N K ERS [ Smoke is defi ned ) as the a i rborne products evolved when a mate r i a l decom· poses by py rolysis or comb ust i o n . S moke may conta i n gases, l iq u id or so l id parti· c l es . or any combi n at i on of these . Proper t y H a z• r d Meas urable 1. O pa c 1 t y H 1 nders E s c.pe Y es • nd Re1cue 2 Lachr y � t or y I nd uces Pa n 1 c No l r r o t an t Be 1 n g S t ud 1 ed 3 T O X ICi t y I nca pa C i ta tes No I Oo r e c t l K1lls B e 1 n g S t ud 1ed 4. T o x 1 C1 t y A n o x 1a No l l nd u e c t l Be1 n g S t ud 1 ed !) Hea t Sears No Resp. System Be 1 n g S t ud 1ed 6 Synerg1sm Comb o n ed E f f e ct s No Beong S t ud 1ed SMO K E OPAC I T Y S moke opacity, or l ight obscurat i o n , is commonly measu red by determ i n i ng the atten uat ion of l ight from a sou rce th rough a col u m n of smoke onto a photo· e lect r i c ce l l . Table 4 sh ows the methods co m m o n l y used in this country for t h i s purpose. The Ste i ner T unnel ( ASTM E ·84 1 . o r ig i na l l y desig ned to measure the spread of fl ame across a ce i l i ng surface . h as been ada pted tc measu re the obscu ra· t ion of smoke as it passes through the e x i t f l u e . At the present t i me, th i s i s the o nl y s m o k e test common l y accepted . H owever, i t h a s been subject to cr i t icism both because of the l ocat i on of the sam ple ( some th i n k that wal l m ou nt i ng or floor mou nt i ng wou l d be preferable i n some cases) and because it represents a l i m i ted set of f i re pa rameters. The X P 2 Chamber (AST M · D 284 3 1 was deve l oped and i s u sed for measur i ng s moke dens i ty from bu r n i ng pl astics. I t has been cr i t i c i zed both because of the s m a l l s i ze of th e sample invol ved and a l so the fact th at it represents a single set of f i re cond i t ions . The N BS Chamber a nd the L L L mod i f i cation both measure smoke dens i ty - l ig h t obscurat ion - by s ubject i ng the sample to rad i a n t heat ( py r o l y s i s) or to rad i ant heat i n the presence of a p i l o t f l ame ( py ro l y s i s pl us com bust i on ) . The L L L mod i f i ca t i on to da t e has co nsi sted o f add i ng a ven t i l a t i o n capab i l i ty ; and we a re cu rren t l y deve l op i ng a h igher rad i ant heat source . B oth th e N B S Chamber and the L L L mod i f icat ion h ave not prese n t l y been accepted as standard me thods. but a re used by a l a • ge nu mber of l aborator i es throughout the cou n t ry . 1 54 

T.tlltt 4. CompMrlon of Smoktt Tttll Syltttml ftN Mttnurmg Smoktt Ob�eurtttron St.,ner Tunnel XP2 Chamber N BS Chamber L A L Chamber S.m .,.. : Sou Large Small S m� l l Small Area !ex posed 36 I t ' 1 in . 1 std 6.6 on ' 6.6 o n 1 14 o n 1 ponoblel ThockneK Va r oable - 0. 1 - 1 o n . 0 .002 1 on 0 002 1 o n T es t d u r ii t i O n l mo n l 10 4 .;30 usue l l y 30 us ... l l y Hut : Sou r ce F la me F la me A a d o a n t + optional f lame Same liS N B S a F le x obohty • SO% 40 pso to 5 pso + 20% 800 % Ventolet oon : Rate 240 hnear I t / mi n None Noneb V ar t able F lex ibo h t y t 35'11o Possible None b 0 t o 20 chang.. s /hr Hea t · t r a nsler mode Pr o met o l y convect o o n Convec t l o n /radoa t oon A ad o a t o o n + some convec· Sa me as N B S t oon o f l la m o "9 )> -,;) S mo k e · produc t l on mOde P y r olysos + combus· Combus t o o n - total P y r olysos w o t h o u t f l a me . Same a s N B S . h ogher hr.. o s -,;) m t oon pr ogr esso ng o nvo lveme n t t o pat t o a l pyrolysos + combuu o on may resul t o n mos t l y com· z a long surf ace . Some onvolvemen t . w o t h f l a me -both on sur · buu oon smokes 2 pene t r a t i o n . f ace + penetra t o o n . X (") Smoke measu rernent : Met hod I n tegr a t ed ra t e I n tegr a t ed r a t e A c c u m u l a t •on ; ma x • mu m Same as N B S r a t e os measured . as os obs c ur a t i o n t o me A epor t o ng Area under obscu ra· S D A ( s moke dens o t y Ma x denso t y ; ma x r a t e Same as N B S . ma t " " " I t o o n vs t i me curve r a t o ngl o n 'llo of smoke a nd t ome. obscur a t o o n smoke obscuration o nO.: a compared t o that for obscurat o o n · t • me t • me . s u m of SO l 's f or vartous 'r ed oa k . curve. l o re pa r a me t er s . F or e p;�rameters possoble 2 1 2 8 E qv opment : Cost $40.000 $1 000 $4000 $4500 Por tabohty No Yes No No Wor k space lf t l 20 x 30 3 X 5 I Movabl e I I Movable I 8 N o t t h e • • •ndard method. b u t posttb le w t t h the eQ u •pmen t . 8l I> N BS h es r ecen t l y added e vent i letoon cepet> i h t y . the f le a obilt t y is t h us - BOO'• 

M I NES A N D B U N K ER S L L L Approach At Lawrence L i vermore Labo ratory we take the view that w i t h i n the l i fe t i me of a f i re, the e xposure of any materi a l or mate r ial system of i n te res t can be e x ­ pressed b y a ser ies of bracketing parameters ( see F igure 1 ) ; i . e . , the mate rial m a y be exposed to a l ow heat or a h igh heat or someth ing in between . It may be ex posed to fl ame or no flame. It may be subjected to no ven t i l ation, to mi nor vent i l ation, or to considerable venti l at i on . The variable in the f i re regi men is the k i nd . th i ck ness, a nd 3tti tude of the materi a l . VEKT I LAT I ON H I GH HEAT Fl.lt£ ~ Tlfo[ / �� �'<£' F L.»'£ LClol HEAT Variable - Materia l : Kind Th ickness Attitude F i gu re 1 . Boundary co nd itions of a fire. 1 56 

APP E N D I X C We have exposed o ver 1 00 d i fferent material systems of l i m i ted th ick ness ( up to one inch) and i n one atti tude ( vertica l ) to w hat we term " low radi ant heat" in the presence or absence of f lame wi th no venti lation, or with vent i la t i on rates up to 20 a i r changes i n an hour. O u r f i nd i ngs have been reported i n the l i terature ( 7 , 8 ] . However, a brief descri ption of the methods u sed and some of the sa l ient resu l ts may be of interest. F igu re 2 shows a p icture of the L L L Density Chamber, wh ich cons ists of an 1 8 cubic foot a l u m i n u m box, 3 feet h igh by 3 feet w ide by 2 feet deep. A 3 by 3 i nch - sq ua re sample of the material u nder test, mounted i n a metal frame and held ver t i cal ly, is slid in front of a rad i an t heat source operating so that the flux on the surface of the ex posed specimen is 2.5 watts per square centi meter. As the sample pyrolyzes it generates smoke w h i ch rises and i ntercepts a vertical l ig h t beam located at the top of the chamber and focused onto a photoelectric ce l l i n the bottom. The loss i n l ight transm ission is measured by a recorder operati ng th rough an a m p l i f i ca· t i on syste m . For the f l a m i ng e x posure cond i t i o n , a series of six sma l l p i l o t fl ames a re positioned at the bottom face of the sample about one-fourth of an i nch away i n order to ign i te any fl ammable species emi tted by the decom posing s pecimen. I n tests where venti lation i s a factor, a i r i s adm i tted th rough a slot i n a hor i zontal tube l ocated i n the l ower right·hand edge of the ch ambe r and is exhausted t h ro ugh a port located i n the upper left·hand back corner of the chamber. Table 5 shows the various data obtai ned i n test i ng a material i n our chamber, the ca lcul ated val ues used, and the L LL in -house smoke standards empl oyed to rate the obscuration p ropert i es of smok e f rom various material . O f i :-terest i s 0 5 , the s pecific optical density of the smoke . The laws of physics define the opt ical density of a med i um as the l oga rithm of the reci procal of the l ight transm i tted through the med i u m . That i s to say, if the l ight transmi tted th rough a med i u m is 1 0% of that i nc ident u pon it, the optical density is 1 ; i f the l ight transm i tted is 1 %, the optical density i s 1 0, etc. The spec ific opti cal density i s a calcul ated value that red uces the area smok i ng, the vol ume i nvo lved, and the l ight path all to unity. I n other words, it i s the opt ica l densi ty that wo uld be obta ined if one sq uare u n i t of mate rial is evo l v i ng smoke i nto a vol u me of one cubic unit and the l ight is transm itted th rough a path of one l inear u n i t . The u t i l ity of th is specific optical density, 0 5 , w i l l be d iscussed later. O ther va l ues of i nterest are O m , the ma x i m u m specific optical densi ty ob· tai ned i n a test; T m , the time it occurs; A m , the ma x imum rate of change of specific opt i cal dens i ty ; T m the t i me at wh i ch it occu rs; and T-1 6, the t i me at , • wh ich the specific opt i cal density reaches a va l ue of 1 6. Tests by Shern ( 9 ] have i nd i cated that masked observers found i t d ifficu l t to see th rough smoke w i th a specific opt i cal density of 1 6. One add i t ional va l ue wh ich we have found useful is the smoke obscu rat i on i ndex I SO I ) . This is defi ned by G ross and R obertson ( 9 ] as being proportional to the prod u ct of the ma x i m u m smoke density and i ts rate of r i se and i nd i rectly proportional to obscuration t i me ; i .e. , T- 1 6. T he mathemati cal derivation of the S O I is gi ven i n References 8 and 9 and i s expressed as shown in Table 5. 1 57 

M I NES A N D B U N K ERS Tabl' 5. Kry to Symbols U"d and IlL Smo•' CHns• tY Stanthrds Sv mbol Oe l o n o t oon v 1 00 IX Speco l o c O P t o Q I Oens o t y = - A L I Iog o o T l V. A. L A e5pec t o ve l y , cha mber vo l u me . ex p05ed samplt' area, and l e ng t h of l o ght pa t h - a l l o n con s o u e n t u n o t s F or t he L L L Chambe r V IA L " 1 32 T L o g h t t r a n 5 m o u o o n - percent Om Ma x o mu m Os a u a o ned on a t e s t A rn Ma x o mu m d /d t i Os l - m o n u t es ' !aver aged over 2 mon I Trm T t mt:' a t wh •ch A m o c c u r s - m t n u t e s T 16 T • me a t wh e c h Ds 1 6. ' " a tt•st - mt n u t es [ SOl S m o k l' Ot..o s c u • a t o o n l nd � • OmA / T 16 O' m 1 ;�� ;- �� SOl . T 16 To • To 7 To J To.J To 9. T 0 7 . e T ' '"" l m o n . I t o r a c h 09 Om. 0 7 Om. e t c . r e wect o vcl � t" h L L L S m ok e S t andards I H· m Va l t w � a n c1 Ot>SCt t p t t o n 0 • � • fi..1. t Jt. • r n u rn S m o t.. •· 7!> :? !> !> 0 1 00 400 · 4 00 Ot· ns• h h qh l moO•'' •' ' ' ' ' " d••nst· V�f 'V t..rll o k • · d t ' f l \ 1' \ fTI O k t ' s n1 o lr.. t• d e n s .- s mo k t• T 1 6 V o \ u rl l Ob\C t J f rl l t O n 10 10 5 !) 1 1 V I ' ' \ \ I O V\ \ l tJ\1\. modP o a t t' l v vf'r v t a u S tnO J.. o · • \ nV"' k e> • f ns r s mo k P r smokeo SOl Smo• • 0 s !) 10 1 0 30 · 30 01 J, (. I t l ,t 1 o () l s,, t , l t 1 0h.t b 1 \ JH O tMl> l \ ha / a r d O U \ I n t h' • \ 1110 k • sa t .· s m o k t · h,t/,t t dous smoke s mo l u A l so shown i n Table 5 are t he i n -house smoke standards used at l l l to attach descr i pt i ve terms to va l ues obtai ned for ma x i m u m smoke density, obscu ra­ t ion t i me, and smoke obscuration i nde x . These val ues, the i r segregat i o n , and the i r 1 58 

APP EN D I X C terminology have been based on our e x perience in the f i e ld , p l us consu l tation wi th others s i m i larly engaged. L L L Resu l ts A typica l smoke density vs time set of curves for red oak , the standard material used i n the Stei ner T unnel , is shown i n F igure 3. Note that u nd e r non flam· i ng cond it ions i n a closed chamber ( or room ) , smoke density r i ses slow l y f i rst and then more rapidly to a va l ue at wh ich i t is considered to be very dense. Under flaming e x posure, a considerably l ower val ue i s ach ieved . By plott i ng the max imum smoke de nsities obta i ned agai nst various vent i la t ion rates, the values for red oak are typica l ly shown in F igure 4. As can be see n , it is q u i te easy to clear away the smoke u nder the f l am i ng cond i t ion . but i t is less so u nder the nonfl a m i ng ex pos ure . On F igure 5 are show n the smoke dens i ty - t i me curves for th ree . so l i d , one· fourth inch thick transparent acry l ics. U nder nonfl a m i ng expos ure the f i re reta rdant 1 59 

M I NE S AND B U N K E R S 400 Non · f lomint F l o m ont .. .. T111 C m i n l 20 17 .. � 300 Om 39 5 75 ... R111 C m i n" 1 1 5 4 1 5 0 .. Tm 15 � � r,. ( m i n i 4.1 0 200 .!! SOl 55 .. u • a !! 1 00 cr 0 0 4 8 12 16 20 24 Time , men Figure 3 . Typiul ��nob IHI/MopmMJt from r«< Nit. soo Vent e l o t eo n , N o n f l o m o ng F' lo m ong ... C ho n t t t / h r Om SOl � SO l Oi 400 • 0 395 3 5 75 c • 3 325 2 2 55 285 1 5 .. 6 so 0 I 0 !t 10 205 9 20 t 300 20 1 25 3 s 0 .!! !: .. • A .. ' 200 E ; 0 � � 100 0 0 18 0 3 6 t 12 IS Ventila tion, o i r c h O I\ ft s l hr 1 60 

APP E N D I X C I J - f lome rttl stont, U V obsor b i n t 1 4 - I\ t o t r e s i s t o n�, U V o bs o r bint 1 1 - I\ t o t r t s i s t o n t , U V t r o n s m lt t lnt Autoitni tion '-'- - - - - - - 1 ' ' ' .. , -:: 400 I .. •.---- c .. � 0 .':i .. � 0 ... ... ·c:; .. � "' .. 0 8 16 Time, min r ' "' " r S . Smolr• tt.wlop,.nr fTom rhTH c/HT KTylia. variety e x h i b i ted a cu rve si m i l ar to that of red oak. The other two mater ials showed a smoke of noticeably l ess density . U nder f l am i ng ex posure, the f i re retardant variety qu ick ly y i e lded a very dense smoke. The other two smoked as fast but to a smal ler degree. Of particu lar i nterest is the autoign i t ion tendencies of two of the materials. In over half the tests m ade, these sampl es ejected f l am i ng species as shown onto the coi l s of the rad i ant heater and set themsel ves on f i re, thus produc· i ng a m uch denser smoke th an when the phenomena d i d not occur. In F igure 6 are shown the effects of vent i l at i on on the max i m u m smoke dens ities of these acry l ics a nd one t ransparent styrene materi a l of the same thick· ness. N ote that u nder the py rolysis cond i t i o ns, the acry l i cs behaved i n a matter s i m i l ar to wood. The max imum smoke density atta i ned i s rapid ly l owe red w i th i ncreasi ng venti lat ion . On the other hand, vent i lation does not seem to help the smoke density u nder f l a m i ng cond i t i ons. In fact for two of the mater ials, a 1 61 

M I NES AND B U N K ERS I :S- o c r rlic , f lome rnist011t , U V o lltorlliftt 1 4 -oc r r lic , llto t r t t i t t ollt , U V olltorll•nt 1 7 - o c r r l ic , lltot r t t o t f oll t , U V tront m i t t lnt .. .. • 1 6 · t�tl r s t r r e "' i .. ! u • • .. E 200 l ·;; 0 a O L---�--��--�----_.---J 0 6 9 12 15 18 21 Vt n t i l o t ioll , oir c 110 11 9 t t / llr F itutt 6. Eff«:t of ,., til•tion on m•ximum -•• .,.,;ry of , ,.. «'Yiicl Mid on• ,,.,,.,.. moderate vent i l a t i ng rate of 6 a i r changes an h o u r seemed t o i ncrease t h e max i m u m smoke density . I n t h e cases of t h e sty rene materi a l a n d the f i re-retardant acry l ic, venti lation wa s completely i neffect i ve in clear i ng away the smoke as far as maxi· mu m dens i ty is concerned. F igu re 7, showi ng the smoke development from clear, rigid po l y v i n y l ch loride in two th ick nesses, i s i n te resti ng i n that unde r the flam i ng cond i tion it does not seem to make any d i fference whether the mate r i al s is one-fourth or one-eighth i n ch thick. The data shown on F igure 8 po i n t u p the need for testi ng mate r i a l s systems in the ma nner i n wh i ch they a re go i ng to be used . Curves obtained show the smoke· den sity t i me cu rves, for a po lyester/epox ide coat i ng on th ree·eighth i n ch gypsum wal l boa rd . U nder the nonfl am i ng exposu re , the results were as ex pected. The 1 62 

600 r-----,--r-- 100 1 I 67- On I cement boorct , norm o i l y cond i t i oned I I Nonf lom i ng 50 ... .. ... 0 c !' .. .. L d "; " ' �mol '0 c .. ... 0 -- u 0 "' " .. 50 as > .. .. 'V � ... cs 'V 0 0 m .':! � a s . bore subs t r o t e z .. .. !2 u .. X .. 24b • ... u Ill 0 Ill - 200 .. 6 6 - 0n wo l l boor d , nor m o l l y c ond it oOIItd 0 "' 100 I I I 0 I J,....-; ... as' 1 00 ? 4 o - 1 /4 i n . t h i c k tJ - 1 1 0 in. t hick 50 1- I / 1 ,_- as cs 0 4 8 12 16 20 0 4 8 12 L_ I I Time , "' '" 16 24 32 Tim� , min r '<I""' 7 Smok l' r/l'vFiopmf'tt l from • cf�.,. ,,,d polyvmyf ch/ofld�. F ogure 8. Smolltt dtt wtopn���n r o f polyttstttrlttpo•idtt co.ti"f. - � ' 

M INES AND BUNKERS coated speci men y ielded a sl ightly denser smoke t h a n t h e ba re specimen. However, the opposi te wa s found to be true under f l am i ng ex posure when the specimens were ° tested under the normal cond ition i ng or preheated to 60 C to remove moisture before the test. Tests of the same coati ng appl ied to cement board showed some· what d i fferent resu lts. It should be noted th at each curve shown represents at least two repl icate tests with good agreement between rep l i cates. As stated above, we have exami ned and tested over 1 00 d ifferent materials i n t h e L L L smoke ch ambe r . O u r f i n d i ngs t o date - under t h e 2 . 5 W/cm 2 rad iant heat f l u x ex posure - ca n be summarized as fol l ows : 1 . Woods, i ncl uding sol id woods, plywood, and other ce l l ulosi cs, show curves s i m i lar to those for red oak in the flami ng and nonf l am i ng e x posures, both with and w i tho ut ven t i l ation. H owever, each particular product or material has its own char· acte ristic ma x i mum smoke density value. An exception should be noted i n the case of one wood with two d i ffe ren t f i re retarda nt treatments. In th is case, denser s mokes were o bta i ne d under the f l am i ng cond iti ons for the material wh ich had been fi re retarded. 2 . Plastics may be d i v ided into two broad categories: a few wh ich do not prod uce visi ble smoke u nder either flam i ng or nonfl a m i ng ex posure ; a nd t h e vast major i ty wh i ch can be d i vided i n to two further classi fications. O f those materials, wh i ch do produce smoke, exposu re to heat a l one y i e lds a broad spectrum of smoke densities. Some behave much l i ke wood , slowly bu i ld i ng u p to a h igh de nsity; others wi l l bu i l d up fairly rapi d l y , but to about the same density. In the presence of heat and f l ame, however, we h ave observed two separate phenomena: ( a ) plastics wh ich tend to burn clean ly are sim i l ar to wood under s i m i l ar condi t i ons . ( b) Those that do not burn cleanly ; i . e . , are f i re retarded i n one way or an· othe r, rapidly evolve dense smokes. These are not read i l y cleared away by ven t i l at· i ng. A q uestion arises as to how these results may be used. We grant that these data are obt a i ned i n a s ma l l -scale test and we agree w i th Christian ( 1 0 ) that, " N o s i ngle smoke rating n umber shou ld be expected to d e f i n e relative smoke hazard s o f materials i n al l situations." F urthermore, we poi n t out that t h e necessarily large· sca l e correl a t i ng tests needed to ver i fy the a p p l i ca b i l ity of the L L L Chamber tests a re yet to be done. Neverthe less, we feel that the resu lts can be u sed w i th some j udgment by e m p l oy i ng the nomograph shown in F igure 9 to eval uate the opac i ty hazard of a materials system. Th is ch art, the original of which was developed by G ross [ 9 ) . plots the speci fic opt ica l density of the smoke at the ti me of i n terest aga i nst the room geometrical factor ; i . e the vo l u me of the room, the area of t h e .• mater i a l smoki ng, and the l ight path or the distance between the observer's eye and an e x i t sign. By putting i n the appropri ate n u m e ri cal values for the area of the material i n 1 64 

APPE N D I X C its i ntended appl ication, the spec i f i c optical density for the t i me of interest as determined by a test, and i nc l ud i ng the volume of the room and the d i stance from an observer or v icti m , as you please, to the exit s ign ; one ca n determ ine whether or not an opacity hazard ex its from smoke i nvol v i ng th i s mater i al . References 7. J. R . Gask i l l and C. R . Veith, "Smoke O pacity from Certa i n Woods and Pl astics," Fire Technology, 4 ( 1 9681 , pp. 1 85-1 95. 8. J . R . Gask i l l , "Smoke D evelopment in Poly mers D u ring Pyrolysis or Combus· tion," J. Fire & Flammability, 1 ( 1 9701 . pp. 1 83-2 1 6. 9. ( a) J. H. Shern , "Smoke Contribution F a ctor in F i re Hazard C l assification of B u i ldi ng M aterials." ( b) D . G ross, J. J. Loftus and A . F. Robertson, " Method for Measu r i ng S moke from B u r n i ng M aterials: A merican Society for Testing and Materials, ( P h i l ade l ph ia STP - 422, 1 9661 F i re Test Methods - Rest ra i nt and Smoke. 1 0. W . J. Christian a nd T . E. Waterman , "Abi l ity of Smai i ·Scale Tests to Pred ict Smoke Prod uct i on," Fire Technology, 7 ( 1 97 1 1 , pp . 332·344 . 1 65 ..