A STUDY OF THE FINE STRUCTURE OF CARBONACEOUS SOLIDS BY MEASUREMENTS OF TRUE AND APPARENT DENSITIES PART IL--CARBONIZED COALS By Rosartnp E. FRANKLIN Received 15th February, 1949 Measurements of the true and apparent densities and adsorptive properties of coals carbonized at 600-1,650° C have been used to study the variation of the colloidal structure with the teimperature of carbonization. It is shown that the true density increases with increasing carbonization temperature, reaching about 2-1 g./om3 at 1,100-1,200° C, The porosity of particles small enough to pass a 240 B.S. sieve is large (from 7 % to 23 %) and increases with increasing carbonization temperature in the range 600° to 1,000° C. The specific surface increases with increasing carbonization temperature between 600° and 800° C, The accessibility of the pores and inner surfaces to liquids and gases decreases with increasing carbonization temperé ture in the range 600° to 1,600° C and is governed by the width of fine constrictioD in the pore system rather than by the mean diameter of pores which are unifom along their length. As a result of this pore structure, the solids function as molecular sieves, the width of the constrictions in the pore system being of the same order as the diameters of the simple molecules used for the density ROSALIND E. FRANKLIN 669 surements (i.e. 2 to 6 A). This molecular sieve structure must be of im- ance in determining both the chemical behaviour of the solid and the Xmposition of the gas evolved during carbonization, The industrial behaviour of coke prepared from any given coal is ily dependent on the maximum carbonization temperature. In ticular, there is loss of reactivity when the temperature of carbon- on rises above 650°C (the exact temperature varies from coal to ),.and the change is reflected in many of the physical and chemical erties Of the material. For example, the adsorption of carbon le, the reactivity to sulphuric acid, the dispersibility in sulphuric the ignition temperature and the “ combustibility ’’ of a number arbonized coals all decrease rapidly with increasing carbonization mperature above 7oo° C.4 In an attempt to investigate the underlying ural changes, Cannon, Griffith and Hirst? made an extensive y -of the internal surface areas of carbonized coals, as measured by © heat of wetting in methanol. They found that there was apparently arked loss of inner surface above 550°-650° C. This change in struc- was further investigated by Maggs,? who studied the adsorption of yl, ethyl and propyl alcohol vapours and showed that the apparent ease in surface area resulted from a decrease in the accessibility of ér surfaces. This is in accord with the earlier work of Macpherson, nd Sinnatt,4 who found that the moisture content of laboratory- nized coals after exposure to the atmosphere for 48 hr. was greatest mples carbonized at about 7oo° C and that the rate of adsorption Gesorption of moisture decreased rapidly with further increase in mization temperature. ie measurements of the true and apparent densities of carbonized ven in this paper make clearer the nature of the structural changes ccur between 600° and 1,000° C and which result in the changes erties inentioned above. In particular, it will be shown that ized coals contain a large volume of very fine pores and a large al’surface area, even after heating to temperatures above 1,000° C, osity of the solid, due to these fine pores alone, increases with sing carbonization temperature at Icast up to 1,000° C, The loss iyity and the apparent decrease in porosity and in internal surface nse from the diminished accessibility of the pores. fine-structure porosity of a carbonized coal may be as high as Jt must be emphasized, however, that it is in no way related to ubble-structure developed by coking coals in what is called the a. Tange. It is, in fact, much smatler in the products obtained cere coking coals than in those from anthracite and low-rank, caking coals which give no bubble-structure, Moreover, the art of the porosity is developed at temperatures above the nge, when the carbonized coal has already acquired a rigid and no further bubble-formation can occur. siderable number of measurements of the apparent densities of d. coals have been previously reported.5 Although the influence carbonization temperature and method of density measurement “RO case systematically investigated, the results ail appear to be ith. Metropolitan Gas Co., The Solid Products of Carbonisation of Coal O: Ww 4). mn, Griffith and Hirst, Proc, Conf. on Uliva-fine Structure of Coals B.C.U.R.A., 1944), p. I31. * Maggs, ibid., p. 147. on, Slater and Sinnatt, Fuel, 1928, 7, 444. ad. Eng. Chem., 1922, 14, 1047; Drakeley and Wilkins, J. Soc. :2931, 50, 3311; Hiles and Mott, Fuel, 1937, 16, 64; Smith and pnd, Eng. Chem., 1942, 34, 43; Milner, Spivey and Coff, J. Chem, Soc., 670 FINE STRUCTURE OF CARBONACEOUS SOLIDS consistent with those of the present wotk. This suggests that structuray changes similar to those described below occur during the carbonizatioy of all coals, and are not peculiar to the specimens studied here. Experimental Materials .—CaRBONIZED CoaLs. four coals. ANTHRACITE C: Samples were prepared from the following S. Wales anthracite, Group AI, 94:2 % carbon (dry. mineral-free). ” Coa. F: 5S. Wales coking coal, Meta-bituminous, 89-7 % carbon’ (dry, mineral-free). Coal H: Yorkshire caking coal, Meta-hgnitons, 83°5 % carbon (dry, mineral-free). Coat K: Northumberland weakly caking coal, . Meta-lignitous, 82-4 % (dry, mineral-free). For carbonization at temperatures up to 1020°C samples ground to pags a 72 B.S. sieve were heated in nitrogen in Nichrome-wound furnaces at 5° C/ming and held for 2 hr. at the maximum temperature. For carbonization at temperatures higher than ro20° C a molybdenum., wound tube furnace was used. The sample was first carbonized at 1000? G as above, and, if coherent, re-ground to pass a 72 B.S. sieve. Tt was then packed into a $in.-bore carbon tube through which nitrogen was passed. The molybdenum furnace was brought to the desired maximum temperature, and the tube containing the sample was then advanced through the furnace in steps so that each part of the sample remained for 2 hr. in the zone of maximum temperature. During the initial stages, and again at the completion of thé process, the closed end of a carbon sighting-tube was situated in the zone of maximum temperature, and the temperature was measured with an optical pyrometer. Analyses.--Proximate and ultimate analyses of the carbonized samples are given in Tables I to IV. TABLE J.—ANALYSIS OF CARBONIZED COAL C Proximate Analysis Ultimate Analysis (Parr’s basis) Carboniza- | 777777" ~ rs as — ~~ tion Temp., i . Volatile °C i Volatile ave then. | , Oxygen Moisture Jess dixed | Ash vate Carbon Hydro- Nitrogen 8 Moisture candle ash-tree) & | Errors oe |e i | ~ -— | ----—-- | — | 26 47 90:6 21 4-9 9472 2g 1:2 I°7 305 1:6 4°9 gig 2k 5'r | Qqrl 2:8 Io 21 410 2:0 4°35 gi2 2°3 4°77 | 94°76 27 o-9 8 510 22 43 Q1-3 22) 4:5 | O47 2-6 oO" 18 605 29 379 Olt 21 4:1 | 94°7 24 o8 21 7E9 $4 3°¢ 90°5 24 3°20 | 95°5 17 a" 19 805 5:0 25 gos 2-0 2-7 96-3 I'o o8 19 gro 49 | 15 OL5 2°1 16 | 96:9 o6 0-6 | 179; 1,OL5 ot | O83 | 97°3 2:3 03 | 98-6 os o-6 03 | Liquids.—METHANOL, BENZENE, ACETONE, ETHER, CHLOROFORM, CARBON TETRACHLORIDE: Analar reagents were used. n-HEXANE boiling at 67-69° was supplied by B.D.H. CARBON DISULPHIDE was purified by treatment with potassium permanganate, mercury and mercuric sulphate according to the method of Hammick and Howard * and was subsequently dried and distilled. Gases.—HyproGeN was prepared by electrolysis of 10 % caustic potash solution saturated with baryta, and was purified by diffusion through a palladium tube, OxYGEN, MeTHANE.—Cylinder gases were used without purification. ® Hammick and Howard, /. Chem. Soc., 1932, 2915. ROSALIND E. FRANKLIN 671 TABLE II.—Awnarysis or CARBONIZED COAL F Proximate Analysis | Ultimate Analysis (Patr’s basis) . Volatile | | Volatile * . . i ; Oxygen Moisture less Cn Ash vane Carbon Hydro- Nitrogen} i Moisture ash- tree) gen Errors cone | Se -- | jp | --—— 08 24°0 quy 4I | 25:2 | 89-7 570 | 18 3°5 06 19°6 73°3 45 207 | 89-7 4°7 18 3°8 1:2 12-0 81-7 5:1 12°8 | or-2 3°2 18 3:8 3°0 6-0 85-2 5°3 6:6 | 93-9 274 13 rg 374 4° 87-1 5°5 F4 | 953 U7 rg ier o7 rg girl 6-3 2°0 97°5 o6 I: o3 TABLE IIL—Awnatysis of CARBONIZED CoaL Proximate Analysis Ultimate Analysis (Parr’s basis) | vomtiel | dvente| P Volatile ¢ | rola re | ole | Oxygen Moisture less Petia Ash nal Carbon Hydro: Nitrogen! ch Moisture a n ash-tiee) ger | Errors 59 | 34°9 | 57°3 | TO | 379 | 835 | 54 | 22 | Bg 1:6 34°0 6295 rg 35°3 | 87:8 51 ig | 10-2 { 13 219 74°5 23 | 227 | 848 4°5 230, B 4 24 Ti3 833 3°0 | 11g | 88-9 31 | 2-2 | 58 33 _ — 22) — 90-7 “| 2-2 — — 56 4:2 87-6 2:6 4°60 | 93-6 Les 20 7 29 Gr 2°8 88-6 275 3:1 94°7 V2 |) 20 | 27 G2 _- 30 — 9575 08 | a Go — | -- | 29 — | 9665 0-8 | — 28 | --+ — 29 | ~ | y6°8 | o6 | rol ows 2-2 | a | 3°4 ~~ | 99°77 FOF | a 13 | og Gig | 34 | og | 96-7 | Org | 5 | O'4 I i 1 TABLE IV.-—Awatysis of CARBONIZED CoaL kK Proximate Analysis | Ultimate Analysis (Parr’s basis) | vette! — | dvowne | Volatile ee 1 atte | Oxygen Moisture! — Jess Cet Ash — Carbon Hydro- jNitrogen + Moisture arbo asb-free) gen Errors a a | fof. | | {oo 10-5 33'8 54°7 Io 38:2 | 82-4 | 51 Ig | 10°6 270 33°5 63°4 lr 34°6 82-2 51 20 | 107 22 24°9 72°0 og 26:0 | 83-2 4°5 2-2 | IO'L wg 13-7 | 83:0 Iv4 14:2 88-4 3°3 2:3 6-0 3°6 8-9 86-0 15 94 | Go-7 2-9 2°4 | 4:0 2-6 7:2 88-3 Ig 75 Q2t 2°5 2:3 | 3 _- — — rg _- g2:1 2:2 — ~— 5°4 5°6 87-3 eg 60 | 95°53) 46 2-0 | Ir ; 6-2 4° 88-0 7 4°5 | O71 o7 P53 | O7 6-7 21 | 89-0 2:2 2°30) 97-5 o"4 | 1-8 | 03 —t | | 672 FINE STRUCTURE OF CARBONACEOUS SOLIDS Measurement of Density.—The apparatus and methods used for thé measurements of densities in helium and in liquids have been described in 3 preceding paper.? Density ‘‘ Drift ’’.—The “ density drift,’’ or increase of apparent densit with time, was investigated by making measurements 2 hr. and 24 hr. afte; immersing the solid in the liquid. In helium, measurements were normally continued until no further drift occurred in 24 hr.; in a few cases the densit drift in helium continued for many days and was not followed to completion: Correction for Mineral Matter.—-The density values recorded have in all cases been corrected for the mineral content of the samples by use of the correction formula of Wandless and Macrae.’ Although the mineral content of carbonized coals is somewhat greater than that of raw coals, the uncertainty in the density correction is less since the composition of the mineral matte; differs less from that of the ash whose density was determined. The correction. did not exceed 3 % for any of the carbonized coals used. Results APPARENT DENSITIES OF COALS CARBONIZED AT TEMPERATURES Up To 1650° C.—Measurements of the densities of coals c, F, H and K carbonized at temperatures from 300-1650° C were made with helium, methanol, water and n-hexane. With coal Hu, carbon disulphide was also used and with coal K i f I J T b-2°O — denen st = - ee As” AT A ied Carborizalion lemperature, °C. 00 [300800 700 \900 Fic, 1.—-Anthracite c. x Helium, O Methanol. (i Water. A n-Hexane. benzene. In Fig. 1 to 4, densities are plotted against carbonization temper ature for temperatures up to r1oz0° C. The full-line curves in the Figures represent density values after 24 hr. immersion in the liquids and the final values obtained with helium, and the broken line curves represent apparent densities ? Pranklin, Trans Faraday Soc., 1949, 45, 274. ® Wandless and Macrae, fuel, 1934, 13, 4. ROSALIND E. FRANKLIN 673 p= I ] | | A #-Hexane. 702 poo 1920 ; 700 j00 Fic, 2,—Coal F, Helium. © Methanol, Cc] Water. Tt T T T | t- 2°O ” + ty w 3 re § Q b~ fp ae of Qo a Asccenfs™” Carborizalion lemperalire, LVv0O 300 4800 ,7ao 900 © Flelium, © Methanol. Fic. 3.—Coal n. [1 Water. A 2-Hexane. -+ Carbon disulphide. 674 FINE STRUCTURE OF CARBONACEOUS SOLIDS after 2-hr. immersion in liquids and the first recorded values in helium. height of the shaded areas is a measure of the density drifts. recorded here and elsewhere for samples which show a large drift in helium refer to measurements made as quickly as possible after admitting the gas to the sample ; they have only qualitative sign:licance.*) qT I ] I t-2-0 a i. nen Tit (The initial values fo? pe? pee 722 I 900 Fic 4 —Coal kK x Helium. © Methanol. C] Water (\ n-Hexane. A Benzene. TEMPERATURES BELOW 500° C.—Below 500° C the variation of apparent: density with carbonization temperature is different for each of the four coals investigated. For the anthracite c there is no appreciable change. The ap: parent densities of coal F in water, of coal uw in water, n-hexane and helium, and of coal x in water and in #-hexane are slightly decreased by heating to 300°.G, aud apparent densities in methanol in all cases increase continuously with-in- creasing carbonization temperature. TEMPERATURE RANGE 500° C To 600° C,—Between 500° C and 600° C’an increase in the apparent densities of the anthracite c sets in and the densities of the other coals all increase. TEMPERATURE RANGE 600° C to roz0° C. (i) Coals c, H and K.~Fig. 1 to 4 show clearly that large and significant structural changes occur in all four coals between 600°C and rooo’C, The re sults for coals c, H and K may be summarized as follows. . n-Hexane gives a low density value which increases continuously with creasing carbonization temperature and shows no appreciable drift. obtained with benzene for coal K are similar. The re Helium gives a density whiéh is from 7% to 23 % in excess of the n-hexane value and which increases DOK 70 0 rapidly with increasing carbonization temperature. When the temperate of treatment rises above about goo° C the density in helium begins to show.so! drift, and at rooo° C the drift is large and the initial density is only slightly i excess of the density in n-hexane. *In all cases sufficient time elapsed to ensure that thermal equilibfi between the sample and the thermostat was established before the helt was admitted. ROSALIND E, FRANKLIN 675 he density in methanol exceeds that in helium for all carbonization tem- res up to 700° C. Between 6oo° C and 750° C a density drift in methanol is’ to appear and increases with increasing carbonization temperature. out 800° C the density in methanol (measured, after 24-hr. immersion) S through a maximum, and between 800° and goo° C it decreases sharply to ne approximately equal to the m-hexane value. At 1,000° C the densities waethc anol and in #-hexane are equal and show no drift. the variation of the apparent density in water is similar to that observed vmethano’ except that the corresponding changes occur about 100° C Up to 800° C the density in water shows no drift and is only slightly yan that in helium. Betweéen 800° and goo” Ca density drift ‘develops, nd.the density after 24- -hr. immersion is at a maximum atabout goo° C. After }onization at 1,000° C the densities of the coals c and K in water are equal the low, w-hexane values and show no drift, while with coal H the density aly slightly higher than the n-hexane value and the dritt is small. “Carbon disulphide gives apparent densities for the carbonized coal H which dentical with the methanol values for carbonization temperatures up to a0°.C. Between 600° and goo° C the apparent densities in carbon disulphide ather higher than in methanol, although a density drift is observed in roximately the same temperature range for the two liquids. After carbon- wation at 1,000° C the apparent density in carbon disulphide is equal to the jow,-n-hexane value and shows no drift. 2s) Coal ¥.—The density changes which occur in the coal F in the temper- re range 600-1,000° C are fundamentally similar to those described above for ‘coals C, H, and K, but differ significantly in two respects. Firstly, the density of this coal in n-hexane increases more rapidly with éarbonization temperature, and the difference between the densities in helium and n-hexane for any one carbonization temperature is less; in methanol and water “the maxima which appear in the curves for the other three coals are fattened into strong inflexions. Secondly, all the changes occur about 100° C t for the coal F than for the coalsc, Hand x. ‘The drift in density in methanol éts.in below 600° C and at 800° € the density in methanol has fallen to the phexane value. At goo° C the density in water is equal to that in m-hexane, Avdensity drift in helium begins below 800° C and at 1,000° C the density in helium too has fallen to the »-hexane level and shows no drift. “oeThe results described above show clearly that after heating to 600° there is, I four coals, a large pore volume which is accessible to helium but inaccessible [ T T T T i ee 2 20 ee © T DENSITY, G/CM. e T t i { : 1 500 730 1000 1250 soo CARBONIZATION TEMPERATURE °C Fic. 5.—Coal Fr, x Helium Ly n-Hexane exane. As the carbonization temperature is increased above 600° this glume is increased ; at the same time, however, it becomes less accessible ids, _Methanol, carbon disulphide and water are successively excluded, [ter carbonization at 1 000°, helium too is excluded from the pores of sand penetrates only slowly in the other coals. 676 FINE STRUCTURE OF CARBONACEOUS SOLIDS The low values obtained with n-hexane for the coals carbonized above boot are sharply defined and the apparent density in this liquid appears to represent a real characteristic of the solid material. There is no appreciable drift, and when apparent densities in water, methanol and carbon disulphide fall to the n-hexane level, these too show no drift. On the other hand, apparent densities intermediate between the n-hexane and helium values are associated with large drifts, showing that penetration of the pores by the liquids is slow and incomplete RANGE 1,000° C to 1,650° C.—The general pattern of the results outlined above is further clarified by measurements made on samples carbonized 3¢ tempcratures above 1,000° C. Results for coals Fr and H are shown in Fig: and 6, For each of the three coals c, F and H carbonized at 1,000°, the densitiag T T T T | 20 DENSITY, G/CM? oa on L 1 i 500 TSO {OOO t250 b00 CARBONIZATION TEMPERATURE ,°C Fic. 6,—Coal u, x Helium. A n-Hexane. given by methanol, water and »-hexans are equal and show no drift, and this is true also of higher carbonization temperatures. For the coal F, the density in helium, too, continues equal to the value obtained with the liquids. With increasing carbonization temperature from 1,000° C to 1,650° C the apparent density of the carbonized coal ¥ in helium and in liquids increases from 1.87 g./em3 to 2-04 g./em.s, The density of the carbonized coal H, measured in methanol, water or #- hexane, increases only slowly above 1,000°, reaching 1:75 g./om.? at 1,600°.G The density in helium at first increases, and then decreases to the low value given by the liquids.* The highest densities in helium are associated with large and prolonged drifts. The maximum value recorded was 2:03 g./cm.$ for a carbonization temperature of 1,100-1,130° but since the drift was not followed to completion, this figure is lower than the true density of the sample ; a higher value was obtained when the sample was more finely ground (see below). The Influence of Particle Size on Apparent Density.—All the results described above were obtained with samples ground to pass a 72 B.S, sieve. Measurements were also made on samples of coal u carbonized at temperatures above 600° C and ground to pass a 240 B.S. sieve. The results are shown I Table V. The density in methanol of the sample carbonized at 805° C and the densities in helium of those carbonized at 1,r00° and 1,250° C were considerably * The density in helium of the anthracite c carbonized at 1,450° C shows no drift and is equal to the -hexane value (obtained by interpolation) for coal H carbonized at the same temperature. Since the apparent densities of the anthracite ¢ and of the coal H carbonized between 700° and 1,000° C and measured in helium, water and w-hexane are almost identical (see Fig. 1 and 3), it is prob: able that the density of coal # in helium after carbonization at 1,450° C wows be equal to the low #-hexane value and would have no drift. This result wis taken into account in drawing the broken part of the helium curve in Fig ROSALIND E. FRANKLIN 677 tacreased by reducing the particle size. On the other hand, the results obtained neh n-hexane for samples carbonized above 800° C and with helium for those w ared below 1,000° C were not altered. Thus, where no density drift was Peerved the density was not increased by grinding to the lower size limit, The density drift in helium of the sample carbonized at 1,000° C was completed in considerably less than 24 hr., and in this case, too, the density measured after 4 br. was independent of the particle size. Yor samples which showed large and prolonged density drifts, the apparent density after a given time was highly dependent on the particle size of the material. TABLE V.—INFLUENCE OF PARTICLE SIZE ON APPARENT DENSITY OF CARBONIZED COAL H ee Helium Methanol | n-Hexane ae —| ——__ Gazbonization Size Density} yp... Density Drift, Drift, Density Drift, Temp., after pee Drift, after % he after % ensity, © " (2hr.| (2 hr. ’ |(2 br. a4 br... g./om.3 ° 24 br, to24| to 72 | 24 hr, to 24 g./om, g./em.4 hr.) hr) g. fem, hr.) 600 72 BS. | 1-542 | 1542 oo 1-600 | O-2 | —- | 1-364 | 0-3 240 B.S. | 1-543 | 1-543 oo — —- f = 1-384 | 0-6 805 72 B.S. | 1-848 | 1-848 o-o r76r | 7-9 | ro-q] 4-558] o-o 240 BS. | 1-861 | 1-861 oo 1816 | 741 gro | 4-558 | or 1,000 72 B.S. | 2-001 | 2-004 108"7 1°655 | OO] -—~ | 1-663 | oo 240 B.S. | 2-005 | 2-005 7-6 — — |} -— | 1665 | oo 1j100-1,130 972 B.S. | 1-982 | 2-029 13°4 — —}]-—- no - 240 B.S. | 2-028 | 2-087 15°8 ~ — ff -— ao y245°1,250} 72 BS. | 1-875 _ 8-0 1723 | OO} — | 1-724 | oro {to 24 hr.) j 240 B.S. | 1-979 ! 27056 Tig — | ae fo | —— | — | The maximum density value recorded was 2-09 g./cm.3 for coal H carbonized at aj100° C, The true density of the material exceeds this value, since a slight drift was still observable after 9 days. Influence of Molecular Size.—-The order in which n-hexane, methanol, swater and helium are excluded from the pores as the carbonization temperature is raised is the same for each of the four widely different coals c, F, H and kK. Mis, in fact, the order of decreasing molecular size. Further measurements were made on the carbonized coal 1 with acetone, ether, chloroform and carbon tetrachloride ; these, together with the results obtained with carbon disulphide ior-coal H and with benzene for coal K confirm that molecular size js the principal iactor determining the power of organic liquids to penetrate into the pores of carbonized coals. The results obtained with 8 liquids and with helium for coal Secarbonized at temperatures between goo” and 1,000° C are given in Table VI, together with the molecular volumes of the liquids at 25°C. The 4 liquids whichchave the largest molecular volumes give apparent densities approximately equal to the low, m-hexane values. No apparent densities appreciably lower than-the n-hexane values have been observed. Adsorption of Gases on Carbonized Coal.—The density of fine-structured Catbonaceous solids cannot be measured in gases other than helium since all other §ases are to some extent adsorbed at room temperature. However, since Physical-adsorption is a very rapid process, measurements of the apparent rate bf adsorption may serve to study qualitatively the rate of penetration of gases into the fine pores in carbonized coals. With this object, the room-temperature adsorption of hydrogen, methane and oxygen on coal # carbonized at 600°, B00" and i,000° C was briefly investigated. The apparatus and method were the same as those used for the measurements of densities with helium. The true volume of the sample was first determined with helium, which was then Pumped off, and the measurements repeated with one of the other gases. The itference between the apparent volumes of the sample in this gas and in helium Eeve the volume of the gas adsorbed, 24 ® 678 PINE STRUCTURE OF CARBONACEOUS SOLIDS TABLE VI aon | | a | Moleculag 7 Olumae. of Liquid Vv Material Temp. 700° C | 750°C | Bao? © j 950° C | 1,005° C cm. at 25° oe Helium . » | Density, g./eom.3] 1-64*] 1-831} 1-861] 1-96*! 1-998 — Diift, % — oo oo oo 10°6 Water. . | Density after 24 hr., g./em.4 r62* | 1°75* | 1-808] 1-710 1695 18 Drift, % (2 hr. to 24 hr.) — — ol 28 0-9 Methanol . | Density after 24 hr., g./em3 r-70*} 1-80*! 1-759 | 1-64*] 1-660 4t Drift, % (2 hr. to 24 hr.) —- — 79 — 0-2 Carbon di- Density after 24 sulphide. hr., g./em3 1°757| 1°828) 1-855] 1-663] 1-656 61 Drift, % (2 hr. | to 24 hr.) o7 4°73 | 4-2 Ig oo Acetone . | Density after 24 hr., g./em? — i810: — 1-629 -—— 74 Drift, % (2 hr. ! to 24 hr.) — ro — jor — Chloroforin . | Density after 24 hr., g./em. rr-q8o} m552| — 1°659 81 Drift, % (2 hr. to 24 hr.) oo + $ OO —- oo Carbon tetra- Density after 24 chloride . hr. g./em.§ -- —- 1560] =~- 17660 99 Drift, % (2 hr. to 24 br.) _ — o-o — oro Ether. . | Density after 24 hr., g./om.3 — 19533], — — oo 105 Drift, % (2 hr. | to 24 hr.) ; O3 | —- - n-llexane . | Density after 24 | | | | | hr, g./em.* T47*] 1°53*! 1-550| 1-65* | 1:663) 128 Drift, % (2 hr. | to 24 hr.) —_ — | O-2 -- oo * Obtained by interpolation. Hyprocen: ‘The quantities of hydrogen adsorbed on the three samples are shown graphically in Fig. 7. The heat of adsorption, calculated from measurements made at 15° and 25° C, was about 1,800 val./g.-mol. Since the isotherms shown in Fig. 7 are approximately linear, and the heat of adsorption is the same for each sample, the slopes of the isotherms give a relative measure of the specific surfaces. Between 600° and 800° C the specific surface increases by about 4o %. : Adsorption equilibrium was established immediately (i.e. in less than $.min the time required for making the first measurement) on the samples carbon ized at 600° and 800° C. On the 1,000° C sample, a rapid initial adsorption was fullowed by a large adsorption drift, and equilibrium was established within 24 hr. Jetween 800° and 1,000° there is a small decrease in specific surface and the good agreement between the results obtained with the 72 mesh and the 240 mesh sample suggests that this decrease is real. MetHane: The results obtained with methane are given in ‘lable VII The heat of adsorption on the sample carbonized at 610°, calculated from measurements made between 12° and 30°, was 5,800 cal./g.-mol. . Much larger quantities of methane were adsorbed than of hydrogen, but the process was considerably slower. With the sample carbonized at 619 G there was a large initial adsorption, &1 % of the total occurring in the first ROSALIND E, FRANKLIN 679 ie equilibrium was established within 5 hr. On the 810° C sample less 0%, of the total adsorption occurred in the first 2 min., and 5 days were [x COAL H CARBONISED AT 100% 62895 © COAL H CARBONISEO AT 8OS% (72 BSS la COAL H CARBONISED AT 1000% (72 BSS O) COAL H CARBONISED AT IOOO % (240 BSS. i a al NTR fo "5 i a | J x or et a a Q 1 8 Ok et JAZ. - 0-05 Va | PRESSURE, (cm. Hg.) a | 1 iG. 7, raguired for the establishment of equilibrium. Adsorption on the 1,010°C sample was very slow; a measurement made immediately after adinitting the #43 gavean apparent density little different from that in n-hexane, TABLE VII.—Ansorprion of METHANE ON CARBONIZED Coat 1 Carbonization Equilibrium | s tae im Volume Adsorbed Temperature, Pressure , | om.3/g. at N.t.p. | rs i Or5 610 ee 6:86 2°33 a — a a 4 OxvcEn : Oxygen was adsorbed in larger quantity than methane, and tlt adsorption proceeded, on the whole, more rapidly. Tor example, on the fample carbonized at 810° C, 55 % of the gas admitted was adsorbed in the mae Peete compared with 1% for a similar quantity of methane. ‘The pttiteats Sorption was, however, always followed by a long slow drift which is var oniveg at least m part, to slow chemical adsorption, On the sample thane at toro? C the initial measurement with oxygen gave, as with ry, yh apparent density approximately equal to the w-hexane value. coat’ a (eoove results show that the accessibility of the pores of the carbonized treae.. 2th of the four gases (helium, hydrogen, methane and oxygen) de- cS with increasing carbonization temperature, and, as was found with 680 PINE STRUCTURE OF CARBONACEOUS SOLIDS liquids, the larger the molecule of the gas, the lower the carbonization temper. ature for which it is first excluded from the pores. in Table VIII. TABLE VIII The results are summarized Coal u Carbonized at Molecular Gas Diam. —_ i oT 600° C 800° C 1,000 °C Helium . 19 No drift, pene- No drift, pene- Large drift, pene- tration complete | tration complete | tration complete within 24 hr, Hydrogen 24 No drift, pene- No drift, pene- | Large drift, pene. tration complete ; tration complete | tration and ad- and adsorption and adsorption sorption com- rapid rapid | plete within 24 he, Oxygen 3:0 Large rapid ad- As for 600° C | Penetration and sorption followed ; but rather | adsorption very by long slow slower. 55 % | slow drift. $2 % of total ad- | of total ad- sorption oc- | sorption oc- curred in first | curred in first 2 min. | 2 min. ' Methane 4:0 Adsorption equi- ] Adsorption equi- | Penetration and librium: estab- librium practic- | adsorption very lished in 5 hr. ally established ' slow sr % of total in 5 days. | adsorption oc- 1% of total j curred in first adsorption oc- | 2 min, curred in first 2 min, | i ! It may be noted that the ease with which different fluids penetrate into the pores of carbonized coal increases with decreasing molecular size for both liquids and gases taken in a single seri Although strictly comparable values of the molecular diameters of all the fluids are not available, the general relation- ship is clear, and provides striking confirmation that the penetrating power does in fact depend primary on molecular diameter and not on any other property of the liquids. Specific Surface.—The specific surface of carbonized coals and its vati- ation with the temperature of carbonization has been studied for a wide range of coals by Cannon, Griffith and Hirst,? who measured the heat of wetting of the materials i: methanol. Further measurements on the carbonized coals c, F, Hand K have been made by Grillith, and the results are shown in Fig... The curves conform to the types previously described by Cannon, Griffith and tlirst, who have discussed their interpretation. In all cases the heat of wetting decreases with increasing carbonization temperature above 550-650° C. Maggs * has shown that this does not necessarily indicate a decrease in the true specific surface, since the accessibility of the pores also decreases—a result which is amply: confirmed by the present work. Adsorption of hydrogen has shown that the increase in specific surface which occurs between 450° and 600? C and which ts detected by methanol continues, also, at higher ternperatures. The heat, of wetting in methanol of coal u carbonized at 600° C is 12 cal./g.; om increasing the carbonization temperature to 800° C the heat of wetting is reduced to 2 cal,/g but adsorption of hydrogen is increased by about 40 %. Cannon, Griffith and Flirst ? have emphasized that the internal surface of coals is never entirely lost during carbonization and that coals which have large inner surfaces yield, in general, carbonized products with large inner sur faces, The present work provides further evidence that the fine-pore structure of a carbonized coal is directly related to that of the raw coal, Of the four coals investigated, coal K has the greatest porosity 7 and yields the most highly porous carbonized products. The raw coal F has a lower porosity and smaller specific surface than the coals c, H and x, and also a more highly constricted ROSALIND E, FRANKLIN 681 evsystem ;? after carbonization, the fine-structure porosity, surface area vaccessibility of the pores are all less than for the other coals carbonized bite same temperature. pox fe COAL H AN COAL K ‘ 200 400 600 BOO 1Ooo 200 400, 600 800 O00 CARBONIZATION TEMPERATURE, C Fic. 8. ine-Structure Porosity of Coals Carbonized above 600° C,—Tt s.from the results described above that the low apparent density ob- or, carbonized coals with #-hexane or with other liquids of large mole- size is substantially independent of the experimental conditions, and is lizdefined and significant property of the solids. There is considerable nee. that it is approximately equal to the lump density of the individual ticles... Not only is it constant for a wide range of liquids, but it was also 1! when samples of coal a pre-heated to 600° and 800° were exposed rat room temperature for 1 hr. after evacuation and before admitting quid. The presence of air would obviously impede the diffusion of a liquid fa . pores, and the same treatment did, in fact, reduce considerably the tent density of the same samples in methanol. Moreover, Maggs has asured the adsorption of #-hexane on coal K carbonized at Goo®, The amount gibed at saturation pressure at 25° C corresponds approximately to that 1 to, form a monolayer on the external surface of the powder. im may be considered to measure the true density of carbonized coals \¢ carbonization temperature is not so high that the gas is excluded from ra part of the pore space. This condition is fulfilled for coals c, rbonized at temperatures up to goo° C and for coal ¥ at temperatures °C. Tt follows that the fine-structure porosity of these materials is P = (dye — dnex)/Tre, e.and dnex are the densities in helium and #-hexane respectively. Aeeg."9, P is plotted against carbonization temperature, It is seen that nghout: the range of carbonization temperature for which the true density asurable (by helium) the fine-structure porosity increases with increasing ture for all four coals. the fine-structure porosity increases while the accessibility of the This leads to the conclusion that the accessibility of the pores the size and frequency of fine constrictions or “ bottle-necks ” ystem, and not by the mean diameter of pores which are uniform : alternative assumption of uniform pores would require asing temperature, the diameter of the pores steadily decreased 682 FINE STRUCTURE OF CARBONACEOUS SOLIDS while their number increased rapidly, new pores formed at any one temperature being narrower than those formed at a lower temperature. “In such a system the increase in porosity would be accompanied by a still larger increase in specifig surface. It has been shown, liowever, that between 800° and 1,000° C the specific surface decreases while the porosity increases. The structural changes which accompany increasing carbonization temper. ature between 600° and 1,000° are associated with the loss of volatile decom. position products from a rigid solid. The existence of the decomposition products (mainly CO, CH, and H,) implies that there is a re-arrangement of surface atoms or small groups of atoms within the solid. This may be sufficient to account for the creation of a more highly T I I I I I constricted pore system resultip; I. + from increased intermicellar con. tacts. A similar effect may per- J haps account for the slight de- tS crease in surface area between 800° and 1,000° by elimination of ay x surface roughness. The increase 3 in fine-structure porosity, on the 2 9 Q other hand, shows that the rigid. 4 ity of the structure is too great ° (or the carbonization temperature too low) to permit such large. s 5 s ° scale re-arrangement as would be 3 a necessary to compensate for the S ° e _| loss of volatile matter aud the x increase in true density. Molecular Sieve Properties v s we ° - of Carbonized Coals.—-The a volume of fine pores in coals |» Lyn carbonized at 1,000° C or at 8 4 higher temperatures may amount 4 g to more than 20 % of the volume LL -+| of the solid. For carbonization temperatures between 600° and ° 1,000° C the solids behave as 4 — molecular sieves; in any given : sample all molecules larger than / a certain size ure completely Carbonizaton femperature. “| excluded from the pores. Car- 5 bonized coals may be compared [990 {600 {700 1890 {800 _ 100% in this respect with the eotites Pia. 9. studied by Barrer,® although their OCoalc. [Coal ¥, 6 Coalu. x Coal k. adsorption capacity is much less The sieve ‘‘ mesh ”’ depends both on the starting material and on the temperature of carbonization, and in the materials studied here it is never large enough to permit appreciable penetration of liquids with molecules as large as those of n-hexane and benzene. The molecular sieve properties of carbonized coals must obviously be of considerable importance in determining the chemical behaviour of the materials, and probably also in determining the course of the later stages of the carboniz- ation process. The principal volatile products of carbonization above 600° C are methane, hydrogen and carbon monoxide, and, above 7oo°® C, hydrogen and methane, the proportion of hydrogen increasing with increasing temperature. Tt has been shown that methane penetrates the pore structure only slowly after carbonization at 600°C and still more slowly after carbonization at higher temperatures, whereas hydrogen penetrates freely after carbonization at 600 and 800° C. The very different rates of diffusion of the gaseous products may well have an important influence on the final composition of the gas evolved. fo This work was carried out at the British Coal Utilization Research Association, and the author is grateful to Dr. Db. H. Bangham for his interest and advice. Laboratoire Central des Services Chimiques de Tl’ Etat, 12 qguat Henri LV, Paris, IV, France. * Barrer, Diffusion in and through Solids (C.U.P., 1941).