A STUDY OF THE FINE STRUCTURE OF ~ARB~NA~E~~S SOLIDS BY ~~~AS~R~~~~~T~ OF TRUE AND APPARENT DENSITIES PART I. COllLS BY ROSALIND E. I?K.kNKLIN Received 18th Novernbev, 1948 The true density of a series of coals was nicasured with helium gas alld apparent densities were measured with methanol, water, n-hexane and lxuzenc. lqu1ds. From the results obtained, the following conclusions were dr:~~~~, Helium fills rapidly and completely the pore space of coals ground to pass a 72 B.S. sieve, and measures the true rieiiiity of the coals. coals ground to pass a 72 1I.S. sieve is filled by nriethanol alitzost conlpletely in a few hours. There is a c,ontraction oi 2.6 r: IO-* ~rn.~ for each cni.$ of surface covered by methanol. Water, n-hexane and benzene fill the pores space of some low rank coals practically coinpletely-. 'i'he apparent densities of such coals in these liquids are high owing to the contraction which accompanies adsorption, ?z-~Iexane and benzene penetrate only very siow~y into the pore space of somp. foals owing to the relatively large diameters of thc molecules of thl.ic liqnids. Therc is no appreciable volrnne of closed pores in coals. The accessibility of the pore space of a coal to li rank in a n:anner similar to the porosity and adso high porosity hxve the most open pore structure. niimei-oils fine coIistric,tions, arid thr variation in t space from one coal to another is relntcil to a vsri constrictions rather than in the mean dianteter of the pores. 'l'he TI iilifi of t constrictions is of the s:inic ortler as thc diameters of si[tiple niolecules arid s~iiatlest in coals containing betwccii 89 "/o and 93 yo carbon. The pore space ____I--. The Apparent Density of Porous Solids.- -It is well kno\vn that the apparcnt density of iiiiely porous solids and o F solids possessing lage specific surfaces is li~giily dq>endent on the metlioii of nicasurenienf The variatjo:: may t,e attributed in the main to two faciors which in- fluence the 1-txsults in opposite directions. T'alues greater than the true. density of the solid may rcsult from the de,. rease in volurne which aG- companies adsorption of the filling fluid ; alter-natively, slow or in- complete penetration of the pres by the fluid mny lead to lox>- apparenf' density vd71cS. In thc latter cast: a density drift (or increase of apparelit? density with time) is frequently observed. These consider:~tioris ap a wide range of organic and inorganic colloiclal materials and ar illustrated by the mimy measiirenieriis which have becn Iiiaile on cha A selection of these is given in 'i'able I ; it is clcar thxt very 1-aried res may be obtained with a single clirzrcoal, and also that the valnes obtalne with a given series of liquids niay fall into different order when differel charcoals are used. The results quoted in Table I show that both ineomplctc penetration of tile pores and the contraction due to adsorption may be important Thus the apparent dcnsity valties are intiniately related to the fine structtire of the solid investigated ; they clepenil not only 011 the Po-' perties of the liquicl and the dcnsity of the solid, but also on the n and extent of the surface of the solid, the size and accessibility 0 pores, and the extent to which the accessibility of the pores and mR" surfaces may be influenced by deformation of the solid rcsultiilg from interaction with the liquid. 27.1 ROSALIND E. FKANKL,IN 27 5 object of the present work was to use measurements of true and t densities to lilvestigate the inner structure of coals, the vari- structure with rank (see later), and the nature of the Interacilon coal suifaces and certain liquids. T~\BLB I - APPARLST DE~SITIT s or CJIARCOAL~ IN Lr~u~ns in . isulphtdc . lcohol . . .. Artivntdd coconut charcoal .. 1.84 1'90 2'01 2.06 2'1 I - Firth Coconut Siigar charcoal ~ chxcoal ~- Corriez Sugar charcoal 1.88 1.6j 1'97 ._ - - ty Measurements made with Helium.- le measurement of the '' true '' density of a been dictated by its 5Illall moleculai diameter netinte into very fine pores, arid I,y iti small g 111 negligible adiorption on solids at room +ever, self-evident that a " true " density x and fine-structured material such ai coal. p(1rcs which are iuaccessible tri all fluids y be iiecessaiy to rlpfiiie arbitlalily whether the '' tru, " volume ol the 1iiatcrm.i. by IIerni,ii~s,~ density 15 essentially egulaiities of nioleciilar dimcniion5 c,riinot pack ~~olt.culc\ irlto the ~i\~~~ktbl~ sp~. d ib such that '' pres " of a fcm Angstioriis fraction ol thc total voluine, thrii nieLtsurc- ecular dimneter ZA) will givt an appairnt e or fiindnnir.iitd signifi~ance. fleriiiaiis cellulose " there 15 no longer a conclusive pccial prcfci eiice shonld bc givcri to the hat it would ha\ e any particular physicd resulting from the prt sent work, how ever, liuin may be uscd to intasii~e ;L \vcll-definr.d bly he callrtl the true density. In par- here Chat a:thoiigh the fine stiucture and to vaiy widtly irom one coal to another, in is a Iiinction of chcmit ~1 composition hcliu~n IiiLusures the true density of coals of Coal.-Coals of different rank form a iesentirig different (though iiot ncccisarily nd Ewng, J. Amer Clzrnc. SOL, 1921, 43, 1787. Awier Chem Sor , 1920, 42, 391. er, zbcd , 1938, 60, 2695 on to the Physccs of Cellulose Ezbres (Elyevier Publishing 276 FINE STRUCTURE OF CARBOKACEOUS SOLTIE successive) stages in tlie process of `` coalification " which started with the clecay of plant material iii a peat-bog and. led to the formation of lignites, sub-bituminous coals, anthracites and peranthracites. In- vestigations of the colloidal structuie of coals are described in a series of papers presented to a Conference on the " Ultra-finc Structuie of Coals and Cokes," and the subject has also been reviewed by Banghnm.8 Griffith and IIirst 8 measured the heat of wetting in methanol of some two hundred coals, arid, with the aid of adsorption data clue to Maggs,Io gave evidence that the sztvfulce area of coals acccssihle to methanol varies from about zo to zoo m,2/g. The relation between the surface ai-ea and rank of coals is sliown.i~i Fig. I, where the heat of wetting in methanol is plotted against the volatile content of the coal." 0 5 1`1 15 20 25 3,' i:, .,o qj 50 y,, Volatilr hiatter Dry, AENSITIES OF THE C'OALS ~ Coal __ A 8 C D I: Ii G n 1 K I M Heliurii leilaity g. jcm.3) 1.64j "517 1.497 1.361 "337 I'jI' 1'293 1.301 1.302 1'305 "3'7 1.341 0'0 0'0 0'0 0'4 0'0 0.0 0'0 0'0 0'0 0'0 0'0 0'0 Methanol __ ~ Uensity after 24 hr. g. jcm. 3) Water Densit). aftrr 2q hi. g jcm.3) 1.630 1'47.5 1.31P 1 '30.5 1'291 1.2q7 "307 1'333 1.326 1.328 1'370 1,488 ~- Drift 2 hr. to 24 hi.) (% 0'0 0'0 0'0 0'0 3'0 0 ' 0 0'0 0'0 0'0 0'0 0 ' 0 0'0 __ %-Hexane . Drift ( 70 2 hr. t 0 24 hr.: 0.8 1'9 6.3 i, ' 3 0'3 0'7 I .s 0.8 1.8 2'9 2'0 2'2 _- neilzene 1- e True and Apparent Densities of Twelve Coals.-The app es and density drifts of the twelve coals described in Table I1 red in methanol, water, n-liexane and benzene, and true densities were stilts are given in Table 111, ant1 in Fig. 3 the carbon content of the coals. The main feature is in every case tlie highest. d in benzcne are in general lower than the true ties in n-hexane and in benzene differ appreciably, Apparent densities in water are lower than the true density for high rank nd higher for low rank coals. Thc change occurs between 82 yo and arbon content. Two low rank coals :LX exceptions. lue given by n-hexane is the lower. l7 Wandless and Macrae, Fuel, 1934, 13, 4. 280 FINE STRUCTURE OF CARBONACEOUS SOLIDS TABLE IV.--INFLUENCE OF PARTICLE SIZE ON APP.kRxxr DRNSITY Coal F K Size (I) Through 14 B.S. 011 1/32 in. . . . (2) Through 72 13,s. 011 I00 B.S. . . . (3) Through 72 B.S. . . . (4) Through 240 H.S. . (I) Through 14 B.S. on 1/32 in. , . . (2) Through ;z B.S. on 100 B.S. . . . (3) Tlirough 72 B.S. . . . (4) Through 240 U.S. . Helium Density (g./c111.3) 1.293 1'299 1.311 1'307 1'292 1'29.1 1'305 1'33.5 0'0 0'0 0'0 0'0 0'0 0'0 0'0 0'0 Methanol ___. lensity after g. /cm.3) 2 $ klT. 1.31 j 1'337 1'333 1'335 1'373 1.371 1.3S6 1.380 Water 0'9 0'0 0'0 0'0 0'0 0.0 0'0 0'0 .- . Carbon per cen% n-Hexane ~. Iensity after 24 hr. 9. jcni.3) __ 1.273 1.271 1.281 1.291 .___ 1.316 1.318 7'330 1'339 _- - 3rift 1r. to Yo, 2 4k) - 0'0 0'4 0'3 0.5 -7 2'2 2'1 1.8 1.8 FIG. 3 ROSALINIJ E. FKANKILIN 281 (5) (a) There is, in general, no density drift in helium. One coal showed a small drift. (b) There is no density drift in water. (6) Density drifts in methanol are generally sniall (0.2-0.6 TI). (d) Considerable density drifts in whexane and herizcne are observed RUE Dtr.?.sITIEs.--The significance of the values obtained for the true of coals is discussed elsewhere.' It is shown thrtt for all the coals in- i.d the true specific volume is a linear function of the liydrogeii content, rapohtion gives 1.85 g./~rn.~ for the deiisity of a " coal " of zero hydrogen . This relationship indicates that there i.; a charactcridic inolecular 'c packing which is common to all coals and very djffcrent from that of hese measurements provide a general survey of the way in which the ap- densities of coals vary l\-ith rank and with the nature of the fluid used to e the measurement. The following series of subsidiary iiieasiireiiieiits were e to assist in elucidating the results. I) To ascertain the significance of the low apparent densities of some coals in water, %-hexane and benzene, the Innip densities of the coals c, I>, 17, H and K were measured as describetl above. and from the lump densities and the densities in helium the true porosities were calculated. The results are given in Table IV-. (2) To investigate further the density drifts and low apparent density values, the dependence of the 1-esults on the particle size of the coal was studied. Results obtained with coals F and K prepared in four size grades ax-e given in Table V. with some co:ils. (6). (density, 2.26 g./~rn.~). TAB1.E \'.---POJtOsITY OF c:OALS Coal C D I.' 31 IC Density * of Dry Lump (g./cm.3) 1.364 1.313 1.301 1'214 1'151 Porosity Calc. from Density in Helium (%) 10'1 4'7 3'2 7'7 12.3 * Densities I' tt corrected for niiriei-a1 matter -The strong influence of the chei~~ical character of the cod surface on the values of the apparent dcnsities in liquids was confiinied by measurcments made on oxidisod coals. Samples of coals F and K, ground to pass 72 B.S.S., were spread in thin layers in a well-ventilated air oven at 105~ C for '36 days. The densities and heats of wetting before and after oxidation are given in Table VI. WhLE VI.-IKFLUENCE OF OXIDATION ox APPARF.ST UJ~KSITIFS AND Density g. 1cm.a. 0.0 1'333 0.6 2.0 1.291 0'0 I"1.00 0'7 3'4 1.357 0.0 1.387 0.4 16.6 1.326 _-_________ 0'0 I 1.577 1'1 22.6 1'479 I ___ {eat of 6.) hr.) seat Of ,vet- ting :tal./ 9.) 1.8 2'3 4'9 I'j 281 FINE STRUCTURE OF CARBONACEOUS SOLIL7S Discussion Apparent Densities in Methanol : Compression of the Adsorbed Liquid.-The apparent density in mctha no1 exceeds the density in hclium for all the twelve coals investigated. It has been shomn above that helium penetrates rapidly and completely into the pores of coals ground to pass a $2 B.S. sieve. hloreover, measurements 011 the two widely ciifferent coals F and K (Table V) show that the porcs of coals ground to pass a 72 B.S. sieve ai-e also completely fil!zd by niethanol in 24 hr. ; density drifts 111 methanol are sinall and decI-f:ase with decrcasinb particli: size, arid r1i:nsity values altcr 24 fir. are practically inciepended of particle size. The difference between tlie density values obtainecl \pit>, helium and with methanol must therefore be attributed to the contractioli which accompanies adsorption. The contraction per g. of coal is given by (VHe-V~lleolI), where VI,, is the specific volume in helinrri and V,,,,, tht! apparent specjhc volume in methanol. This functiou is found to be directly proportiorial to the heat of wetting of the coal in methanol, for elevcri of the txvelve coals investigated (Fig. 4), the contraction being 0.0026 c~n.~ per caloric heat of wetting. Only coal J is rcpi-escnted by a point lying wcll above the line, but the heat of wetting of this coal was liberated exceptioiially sloivly and the value recorded is known from other cvideiice to represent a gross under-estimate.* Maggs 11) has shown that the heat of wetting of coals in methalid is I cal./Io m.z of surface, and this is the same as the figure obtained by HUT OF WETTING IN METHANOL, cal./g. FIG. 4. Bangliam for charcoal. The heat of wetting in metllano1 may therefore be taken as a measure of the specific surface of tlie coals, modified cods * Owing to the limitations of tire calorimeter used, lients of wett~ng coda oxily be measured during the first IO to 20 rrtin. after immersion. Spe;;$' volunles measured after z hr. immersion r;iiher than after 21 hr., were thcrefd@ used in obtaining the values of (I'/He-V>nmH) shown in 1;ig. 4. ROSALIND E. FRANKLIN 283 coals discussed here. The linear relationship 111 Fig. 4 then shows contraction due to adsorption is proportional to the surfacr area The faLt that the line passes through tire origin confirms thnt 1 and helium penetrate to an approximately equal extent iiito important to note th'tt the contraction pcr unit aiea of adsorbing IS found to be of equal magnitude for cliarcoali, anthracitcs, low rank and highly oxidised coals in spite of the widely differing and chcmical properties ol these rnaterl,ils. 'llr contraction refore, be attributed either to incipient solutiori or to ~ontiac- e'tt of wetting of I cal to represent IO ni of surface tlie action is 2.6 x IO-@ ~111.~ per cm of surface, oi about lume which -would he occupied by a monolayer of niethanol of the solid if no contraction occurled. Thus, even if the re spread evenly throughout the absorbed material (coals iivalent of from 3 to j molecular layers of methanol at ssiiii: at 25' C), it is surprisingly l'trgc. Xea5ureInents oali and celluloiic materials indicate, however, that the on observed when methanol ii adsorbed on coali is not ave calculated the hydrostatic prcssure required to contr,iction in the adsorbed film, but It seems that have little ineaning since the adsoihed film arid the Its formation are essentially aniqotropic, and the ies of tlic film difter froin those of the bulk liquid. of the volume occupied by watei vapour on ginic vapours on chaico'il l9 ha\ e sho\%n that the occ ins for miall qunntitiei ol vapoui adsorbed, nlic ii ssure IS very smdl and tlie \olurnc change niust be ntirely to tlic close approxh of the adzoibate molc~- atom? of the adtorbent. It ietm5 probable, therefore, 21 t of the largc contrdctron obi-tivtcl when ~iiethaiiol in the direction pei penclicular to the acliorbiiig arcnt Den4ties in R and Benzcne.--1 Jle appLrtiit den. itiei of coals in ?i-heu'tne LL ne show ~~niidcr.rz111~ cliitts, and ii aic loir~r than thc cor1( >pond- ities in hcliurn for all ex of thc coJi inQcitigxted Pent- of tne pores by these liquidi is slow ant1 incorriplc.te, ancl tlie of penetration \ami with the rank of the coal. In tlie coals u, F, tlie lrqiiicls ar? almost completely cxclucktl from the pores . the are exceptioiially slo~, arid for coal u thr appatnt iltiisity in xane after 2 hr ~~-nintrsion is still very near the limp iic nsity. The sties of the anthiacitei in n-hexane are much lov,ei than in helium, comparison M it11 the lump dc.riiitY shows that thr liquids penetrate CoIlsiderable part of the pole spaci Tlic ~oal? K and 31, for which areiit deiisitics in n li~x~irie are higher than ln htl111111, arc both law ak of high porokity It appcai5, therefoie, that the acceiiibility pores of coal5 to n hex,inc md benzeiic. is at a iiiiniiiium for coals ning between 89 %) and 91 yo carbon, and incrtrtsti foi coals of or of lower rank, bciiig grcatest for loir rank coal> of high poiosity. , the accesszbalzty is cloicly related to the poros2iy. ne and benfene readily Trct the surface5 of coals, and the vis- 1 he poor peuctTating 1101\ei of the liquids canliot, thciefore, be attiibiitcd either to a n~ettmg angie greater than goo obtaiiid after 24 Iir n-hexane ii verj- low. Stanlm and Seborg, J Emforth ancl Dcvrles, I 1937, 41, 1007. I'hysic. Chew., 193 j, J. Anaev. Chcm. Soc., and lianson, 284 FINE STRUCTURE OF CARBONACEOUS SOLIDS or to high viscosity, but must be associated rathei with the relatively large size of the benzene and Pz-hexaiie molecules. The nlolecule of n-hexane is considerably larger than that of benzene, and, where the apparent densities of a coal in the two liquids differ appreciably, the ?z-hexane value is the lower. The above results therefore suggmt that the width of the pore of constrictions in the pore system in coals is of the same order as diameters of the molecules used for the density measurements 'r fine pores or constrictions are smallest in coals coiitaining between 8 and 93 yo carbon. In the coal D the pores or conztrictioris are so na that even helium (molecular cliameter z A) penetrates soniewkat slo a small denzity drift being observed. are always associated with long, slow drifts. This was with other organic liquids of relati\ ely large molecul heavy oil supplied for a Hyvac rotaiy pump showed a drift, indicating slow penetration. This behavioui attributed to slow distortion of the coal substance by the in the ad.;orbed films, and is in sharp contrast with that coals; the more rigid structure of the latter resists clefor sorbed films at 25' C, and, if the inolecules of a giveii hqui to enter the pores, then the liquid ib totally excluded and there is 110 density drift.20 Apparent Densities in Water.--The apparent density in wate greater than the density in helium for coals of less than about 84 yo car content. For high rank coal\, on tlie other hand, it is iriterriied between the values obtained with helium and n liexarir, being close the low, rz-hexane value for coals coiilcniiing between 89 96 and 93 carbon, and nearly equal to the helium value for the highest r anthracite (1 16. 3). accessibility of the pores and inner suifaces of coali to nnter ib greatest for 10% rank coal-, of high porosity, pdsses through a inininium for coal? tontaiiiing bctueeu 89 o/o and 93 yo carbon, aiid iiicieniei again wits increasing rmk amoiig the aiithr, cif es to fill completely the poxrs of high rank co cannot be due to the I of the water riiolccule, since tlic molecule rcLeth,tnol is larger, and appdrerit dtiisitici in methanol are high Rlo okei, the abseiice of any deiislty dritt suggcsts that tlit cduse of 1 appaient deiisities in water is different from that of the low value tai~it-d with n-hexane and benzene. This difference is alm reveale the measure~neiits made un samplcs of coals > arid I< ground to v sizes (Table V). Finer grinding increases the appdrciit density of coals 111 n-hexane, and the drift for coal li is also increased. on the other hand, the appaient density of coal F rciiinins constan that of coal K is incieascd, but in neither case is a density drift introduc It scems, therefore, th,tt the low apparent densities of high rdn iii nater is an effect aisociated with their pool uett,Lbility Tha external surfaces of ~oals D aiid F (91.7 ";o .ind 89 7 "/o carbon respect are not easily wetted by water shown when attempts were ma measure their densities by immersing the sample (not previously evac in the boiling liquid in a simple density bottle. wetted even after prolonged boiling, arid apparent dtnsities as 0.7 g were recorded.* With coals c and K, on the otliei ha It may be noted that tlic low apparent densities in n-hex lhus, as in the cLLse of n-hexane arid beiizene, 'The failure of watt In ?lie paiticles were 2o Frmklin, Coal h'sc , 1946, p 37 * Striking proof of the wnv in uhir h poor wettability (due to d large contact) may mfluence appatent deniitles 111 water evcn after evacuat obtamed with a iamplc of tlic comm6rcial carbon black 1'33 lhe mater1 evacuated at 90' C for thrce days, after which time the particles adhe one another, and the sub5tance could be shaken about in the glass bulb 1x1 bm ROSALIND E. FRAKKLIN 255 resnlts agreed well with the apparent densities measured in water after evacuation There is an apparent contradictioii between these results and some later work by Maggs 21 who found that even high rank coals adiorb at saturation pressure, sufficient water vapour to fill all the pores. The &sclepancy may perhaps be due to thc fact that Mags used water zla3011y vhereas in the prescnt work tlie coal nas always plctced 111 direct tact with the liquid Iligh rank coals contain adsorbed methane, last traces of which are difficult to remove, and it may t)c that water our can replace the gas rnolcciile by niolccule, wheieas the presence Dinlensions of the Pores in Coals.-Cornpatisor1 of the poiosity coals as nieaiured by King and \Vilkiiis with the heat of wetting in thanol a? measurcd by Griffith and Eiirst suggests that the porosity )roximately proportional to the surface area It follows that, for lven structural model, the mean diamctsi of the pores is approxini- tlie same in all coals. lCIeasuremeiits of the porosity of tlic coals F, €1, and K confirm this observation Assuming that I cd. of heat ttliig in methanol corresponds to IO m ol surface, then, if the poi^ e 111 coals conhizts in unifoim, cylindrical, non mtei5ccting pmes, lean pore diameter in all codl~ is abo~il 40 A. Alternati\ely, if the nsists of equally spaced cubes of coal substance, tlie distance of ion between the cubes would be about 12 A. ?he climeiisions of ions of coiitinuous coal substance wi~uld, of course, vxy with the area of the coal. as betii shouii, howe.i.cr, that coals do iiot bchave as if they all uniform poies whose size IS iriclipendciit o: iaiilc The pores contaning bttl$eeii 89 yo and 03 9, carbon ale the least easily ted by liquids and arnorig both low I ,ink cads and anthracites lily inzrcasc s with incrcaiirig poiosity. Similar obirivdtions inadz by (~iaha111 22 mho 11as pointed out that the I)( irneability ne and to rrioisture Laries widelj from ~iie coal to another, the of maiiy anthiac ite5 aiid sonzc low r~nh co Is being kJTV cr lon rai~li coal5 and. also the Iiiglitst iarih diiilirruite ha\e watei impedes its cscctpc mean pore diametci in all co,ili ii ,ippro me~ibility vane5 wit1 ly, it must be prc.surilet1 ig tlitir 1~iigth ijnt contain iine coiiiti ictioni and that of thc pore spa~e to lic1uiili and g,iies i-, diterrnined by quericy of thesc coIistii( tion5 rather than by the mean On this hypotheiii, the poies arc nioit Iiighly constricted ing betwecn 89 "/b ai:d 93 yo carbon, uhile lcm rank coals have the rriost open poi e iti uct tirc. atioii of the hie poiei niiist be intiinatel> ir1,ited to thc the coal niicelles aie houiid together and the variation of ture with the rank of the coril must be associ,itcd with the icellar structure which wcur dui~ng coalificdtron. If the rface area and p Irositj which occurs TXith incicasing lank 0 yo Laibon be attributed to an increase iii the sne of tlic the dPcrcwmig acCCi5ibihtp of the pores is in no way ex- rcsulti suggest, rather, that the changes in pore structure Its deriilty III helium nl '. When attempt> uert Inadc to I~C.L\IIIC tlx. density In vatcr ed, and appioxim.ite mtasiirernenti gavt values of dbOllt o 8 g /cm.', atmospliPii~ piesyurt 7% <+, not sufficient to force the \\atcyr into the aces betvetn the looiely aggegratecl particles. Grifiitli aid hLigg5, I-UVU~~J SOC Dm ZISSZO~S, 1948, 3, 29. m, Coizfeieme on the UltvnJliie Struclwr of Cools and Coke3 (B C.U.K.A., heinisp!ieiiial lump with a prilv shiny surface. 286 TIIE I(IKETICS 01' TEIE TI