84
Energy & Fuels 1987,1, 84-88
Since the IR spectrum of the microbial products shows an insignificant amount of ester carbonyl, the 175-185 ppm band can be attributed mainly to carboxylate carbon. The level of carboxylate carbon is much higher than that observed for lignites12and is consistent with extensive oxidation of the lignite in near-surface seams that give rise to Leonardite. Other structural features include a significant percentage of substituted aryl carbon (129-148 ppm),B protonated aryl c a r k n (110-129 ppm): a broad mixture of alkyl functional groups (11-60 ppm), and a lack of significant straight-chain alkanes (30 f 1 ppm). The process of microbial digestion may have led to further oxidation of the lignite: the IR spectrum of the product (Figure4B) $indicates that the aliphatic stretching band around 2950 cm-' was detectably reduced in the product by comparison with that of the starting coal (Figure 4A). The 170(t1750-cm-' shoulder, corresponding to ester or ketone fuhtional groups, was also reduced in the product, perhaps because of further oxidation or ester hydrolysis. The broad IR absorption bands in the 3600-3000 and 1650-cm-' regions indicate the presence of OH arid a carboxylate or quinone functionality. The proton NMR spectrum of the acid-precipitated material indicates broad regions of aromatic and aliphatic hydrogen, in about equal amounts. While the 'H NMR spectrum obtained directly from the freeze-dried product (12)Franz, J. A.;Camaioni, D.M. Fuel 1984,63,990-1002.
accounts for 95% of the hydrogen detected by elemental analysis, absorption of some water by the hygroscopic product may have contributed to the observed HDO band. We have demonstrated the exchange of counterions in these materials by acid precipitation (Table 11)and suggest that such exchange may allow flexibility with regard to solubility and other characteristics. For example, a more "soapy" material might be obtained by increasing the sodium counterion concentration. These soapy materials might be useful as energy-rich surfactants to enhance the quality of coal/water slurries. Also, ion-exchange results indicate that unwanted trace metals might be exchangeable for other, more desirable counterions. The bioconverted materials are more analogous in structure to the feed coals than are thermochemically derived materials, but the products obtained to date are of lower quality than many thermochemically derived coal-conversion materials. However, it is also likely that the bioconverted materials might eventually be much less expensive to produce. Further research will be required to determine if there is a place in the energy market for such "bargain basement" liquid fuels.
Acknowledgment. This work was performed under Contract No. 23112-06617 with Electric Power Research Institute, Palo Alto, CA. The X-ray fluorescence and X-ray diffraction were performed by Ronald W. Sanders. We are grateful to Dr. Robert Botto, Chemistry Division, Argonne National Laboratory, Argonne, IL, for providing the 13C NMR analysis of the solid Leonardite.
Modification of Coal by Subcritical Steam: Pyrolysis and Extraction Yields R. A. GrafP and S. D. Brandes The Clean Fuels Institute, The City College of The City University of New York, New York, New York 10031 Received July 29, 1986. Revised Manuscript Received September 29, 1986 Exposure of Illinois No. 6 coal to 50 atm of steam at temperatures between 320 and 360 "C produces changes in coal of considerable significance. These changes have the effect of dramatically improving the yields of liquids obtained subsequently either by flash pyrolysis or by solvent extraction under mild conditions. Similar treatment of the coal in a helium atmosphere has no effect. If the steamtreated sample is exposed to ambient air, both the pyrolysis and extraction yields fall to the levels observed for untreated coal. Evidence so far available suggests that coal is partially depolymerized by this steam exposure.
Introduction Methods for disrupting the cross-linked structure of coal are of value for providing new information about the physical and chemical nature of this material. Moreover, a procedure of low cost may form the basis for improved processes for manufacturing synthetic fuels. In this paper we report the effects of exposing coal to subcritical steam below 360 "C. This mild treatment has the pronounced effect of increasing yields of gases and liquids in subsequent processing, giving the appearance of having partially depolymerized the material. When coal and steam are jointly considered, the temperature range 300-350 "C assumes importance for two 0887-0624/87/2501-0084$01.50/0
reasons: First, this is a threshold region for thermal processes where the softening of plastic coals is incipient (vide infra). Second, a high degree of mutual solubility between water and aromatic hydrocarbons can be expected at about 100 atm.' In addition, there is evidence (though at higher temperatures and pressures) that steam is reactive with carbonaceous materials (vide infra). In our study coal was exposed to 50 atm of steam at temperatures from about 300 to 370 "C and tested in two types of experiments. In the first, pretreated coal was pyrolyzed in 50 atm of steam. In the second, pretreated (1) Schneider, G.M.Adu. Chem. Phys. 1970, 17,1.
0 1987 American Chemical Society
Energy &Fuels, Vol. 1, No. 1, 1987 85
Modification of Coal by Subcritical S t e a m HELIUM
A STEAM SUPEY%
SAMPLE
-THERMOCOUF I N J E C T I O N 'LE TUBE
REACTOR QUARTZ CHIPS
/+THERMOCOUPLE
WELL
Figure 1. Injection reactor for coal pretreatment and pyrolysis.
coal was extracted with pyridine. In both cases the resulting increase in yield was substantial.
Experimental Methods Steam Pyrolysis. In these studies a batch injection reactor is used to pyrolyze the coal (Figure 1). In order to prevent slow and premature coal heating, the sample is held in an external injection tube while the reactor is brought to temperature under a flow of superheated steam. The reactor is constructed of 2.5 cm i.d., 3.8 cm 0.d. 304 stainless-steel pipe and is 25 cm in length. A trap of -8 10 mesh quartz chips, 2.5 cm deep, supported by a drilled steel plate is located at the center of the reador. Heating is provided by an external furnace and electrical windings. Temperature is read from a thermocouple in a well inserted from the bottom of the reactor and in contact with the trap. This reading is corrected for the temperature difference across the bed measured in separate experiments at operating conditions. Coal, held in the injection t u b (5.1 mm id.) by two quartz wool plugs, is injected into the reactor by a pulse of helium obtained from a 32.5 cm3pulse tank charged to 82 atm by opening a solenoid valve for a nominal 0.1 s. Upon injection into the incoming stream of superheated steam, the sample is carried downward and deposited on the trap. It is rapidly heated by conduction from steam, by radiation from the reactor walls, and by direct contact with the trap. Volatiles released from coal are carried by flowing steam to the reactor outlet and expanded to atmospheric pressure with a residence time of 3-5 s. A fraction of the product gas stream is conducted to an on-line mass spectrometer for analysis. The only carbon-containing gases observed in significant amounts (over 1 %) are CO, COz,and CHI. The reaction is terminated by flooding the reactor with helium. Following steam pyrolysis, the residual char is burned with oxygen at 14 atm and the amount of COz produced is used to determine the quantity of carbon remaining in the reactor (the amount of CO produced has been found to be negligible). This information is used to construct a carbon balance. Total volatiles yield is obtained by subtracting carbon determined in residual char from the carbon content of the coal sample charged. The difference between the carbon yield of pyrolysis gases and total volatiles yield is ascribed to liquids. Both pretreated and raw coal are injected into the reactor and pyrolyzed by following identical procedures as described above. Pretreatment when desired, is carried out in a preliminary step on the coal charged to the injection tube. An auxiliary heating mantle is placed around the injection tube, which is then heated to above the steam condensation point (264 "C) while charged with helium. Steam is admitted to the reactor, and its flow through the coal sample is regulated by the vent valve above the injection tube. Temperature during pretreatment is read with
+
a thermocouple spot-welded to the outside of the injection tube and controlled by adjustments to the power supplied to the auxiliary heater and the steam flow. Pretreatment is terminated by lowering the temperature while flushing with helium. The sample may then be injected a t any time without any intervening exposure to air. In separate experiments it has been determined that a 7% weight loss occurs by volatilization during pretreatment. This is reported as part of the total volatiles yield. Even though some fraction of volatile material produced during pretreatment is condensable liquid, this has not been included in the liquid yields given below. Tests were conducted on a batch of Illinois No. 6 coal ground under inert atmosphere to pass 200 mesh, having the following elemental analysis (wt %) on a moisture-free basis. Anal. Found C, 69.1; H, 5.2; N, 1.2; organic S, 3.0; organic 0,9.2; mineral matter, 12.3. (We thank Dr. Ronald Liotta of Exxon Research and Engineering Co., Clinton, NJ, for this sample, its preparation, and analysis.) Coal was charged to the reactor with a moisture content of 2.3%. From 150 to 250 mg of coal was charged to the reactor in each run. Charges larger than 250 mg give lower liquid yields, possibly exceeding the ability of the system to rapidly heat injected material. Further details of equipment and methods are given in Graff et aL2 Solvent Extraction. A single batch of Illinois No. 6 coal was ground under nitrogen to pass 200 mesh and used for all extractions. Samples were pretreated by heating them in a batch reactor under a constant flow of helium a t 50 atm to 300 "C. Steam at 50 atm was then admitted and the temperature brought to the desired pretreatment conditions within 3 min. After 15 min of exposure to steam, electricity to the heaters was shut off and a flow of helium was resumed until there was no detectable moisture in the outlet of the pretreater. The coal was then sealed under 50 atm of helium and transferred to a nitrogen-filled glovebag. Pretreatment conducted in helium mimicked the steam pretreatments as closely as possible. All extractions (except in a few cases where deliberate exposure to air was tested) were carried out in a nitrogen-filled glovebag at room temperature. The pretreated coal was removed from the pretreatment tube, crushed in a mortar, and weighed in the glovebag. Samples of about 1.5 g were transferred to 250-mL Erlenmeyer flasks, and 100 mL of pyridine was added. If the coal sample was to be exposed to air, the flask was removed from the glovebag, the contents were stirred, and then the pyridine was added. The contents of each flask were stirred with a Tefloncovered stirring bar for 15 min. The stirring was halted, and the particles were allowed to settle to the bottom of the flask. The solution was decanted onto a weighed filter paper. An additional 100 mL of fresh pyridine was added to the residue in the flask and the solution was stirred another 15 min. The procedure for decanting was repeated, using the same filter paper. A final volume of fresh pyridine, 150 mL, was added, and the solution was stirred for another 15 min. The stirring plate was turned off, and the sample was loosely covered and left for 16 h in pyridine. The solution was then quickly agitated and filtered through the weighed filter paper. The filter paper was left to dry for about 2 h and any remaining pyridine was removed in a vacuum oven at 90 "C for 2 h. The residue left in the filter was weighed and the extract yield calculated on the initial weight of coal charged to the pretreater. The extract yield, therefore, includes the loss of volatile material evolved during pretreatment.
Results Steam Pyrolysis. Figure 2 shows yields obtained from Illinois No. 6 coal over a range of pyrolysis temperatures both with and without pretreatment. In each case pyrolysis was carried out in pure steam at 50 atm of pressure under rapid heating conditions. When coal is not pre(2) Graff, R. A.; LaCava, A. I. "Studies Toward Improved Techniques for Gasifying Coal"; final report for 1980-1984, U.S.Department of Energy Contract No. DE-AC-21-80MC14875, 1984.
86 Energy & Fuels, Vol. 1, No. 1, 1987
Gruff and Brandes
Table I. Steam Pyrolysis of Pretreated Illinois No. 6 Coal with Prompt Injection run no. E113 E114 E115 E117
time, min 15 15 5 5
pretreatment temp range, "C 301-348 260-353 313-340 313-370
wt loss, %
(7) (7) (7)
pyrolysis temp, "C 733 723 718 746
CHa 3.9 2.4 1.8
co 12.7 3.7 1.6
carbon conversion, % CO, liquids 12.2 35.9 10.1 54.9 5.6 49.4
total volatiles 71.7 78.1 65.4 82.5
Values in parentheses were estimated from direct measurements on other samples.
Table 11. Effect of Storage under Air and Helium on Steam Pyrolysis Yields from Pretreated Illinois No. 6 Coal pretreatment ~
run no. E93 E95 E98 E120 E126 E127 a
time, min 30 30 30 15 15 15
temp, range, O C 265-339 289-335 289-335 288-377 311-358 299-362
storage time medium 12 days air 4 days air 25 days air 4h He 2 min air 18 h He
~
wt loss," %
7.8 (8)
(7) (7)
pyrolysis temn "C 720 714 724 713 720 712
CH, 2.9 2.0 3.9 3.4
carbon conversion, % CO, liauids total volatiles 13.7 25.7 55.5 20.7 14.8 53.4 44.5 74.9 7.4 20.2 25.5 64.0 2.1 12.7 61.3 86.5
CO 5.4 5.4
Values in parentheses were estimated from direct measurements on other samples. loo
90
-
80 -
8
-
70-
z
-
60w 50-
5
-
8
-
0
-
z 40-
9
30-
-
20
-
10 -
'650
700 750 800 REACTION TEMPERATURE PCI
Figure 2. Steam pyrolysis yields from raw (open points) and pretreated (closed points) Illinois No. 6 coal a t 50 atm.
treated, no more than about 23% of its carbon is converted to liquid (this is higher than has been reported in the literature for pyrolysis in an inert atmosphere at the same pressure3). Pretreated samples in Figure 2 were exposed to 50 atm of steam for 30 min at temperatures from 300 to 360 "C. By this pretreatment, at a pyrolysis temperature of about 740 "C, the liquid yield is more than doubled and the total volatiles yield is increased by about 20 5%. While some variation in temperature during pretreatment could not be avoided, it is possible to give approximate limits within which the procedure is effective. Steam was admitted when the saturation temperature had been safely passed, usually at about 300 "C, If the sample was not then allowed to reach at least 320 "C, little, if any, yield improvement resulted. If a temperature much above 360 "C was reached at any time, the benefits of pretreatment were lost. These limits are likely peculiar to the coal, the pressure, and the exposure time employed and are expected to be different if any of these parameters are changed. For commercial processes shorter pretreatment times will be desirable. This possibility has been tested in a few exploratory runs (Table I). Two runs were made for 15 min of pretreatment and two for 5 min of pretreatment. (3) Howard, J. B. Hydropyrolysis, in Chemistry of Coal Utilization, Second Supplemental Volume Elliot, M., Ed.; Wiley: New York, 1981; p 665 ff.
Yields are more sensitive to pretreatment temperature at these shorter times than for the 30-min pretreatments. In run E113, the sample was pretreated for 15 min at temperatures mostly around 320 "C. Both total volatiles and liquid yields were considerably lower than for 30-min pretreatments. Yields were brought into agreement with the 30-min pretreatment in run E114 by keeping the pretreatment temperature mostly in the 33Ck340 "C range. A 5-min pretreatment, mostly a t about 340 "C, gave low volatiles yield but good liquids yield (E115). When the sample was held at about 350 "C for the 5-min pretreatment, the volatiles yield was raised to the value obtained in 30 min; the liquid yield was not determined (E117). The equipment used here is not suitable for studies of pretreatments less than 5 min in duration. However, results so far suggest that good liquid yields can be maintained with even shorter pretreatments. A second series of runs were made to test the effect of sample storage between pretreatment and pyrolysis (Table 11). In runs E93, E95, and E98, batch samples of about 1 g were pretreated. After the samples were cooled in helium, the light agglomerates that had formed were easily broken up and the samples stored in small vials. During handling and storage, no precautions were taken to exclude air. After the time period noted, the sample was reacted in the usual way. All samples show volatiles yields less than those for untreated coal and liquid yields about the same as those for untreated coal. One sample (E126) was contacted with air for only 2 min after cooling in helium. Volatiles yield was slightly higher and liquids yield the same as those for untreated coal. It seems possible, however, to store pretreated coal under inert gas. Storage for 4 h in a helium atmosphere (Table 11) gave the same volatiles yield as pretreated and promptly injected coal (E120, no liquids yield available). A sample stored for 18 h in a helium atmosphere gave higher yields of both volatiles and liquids than promptly injected coal (E127). However, this sample was heated to 226 "C and may have received further exposure to steam in the minutes before injection; verification of the results of run E127 will be required. Six coal samples have been pretreated in helium at 50 atm of pressure. In the temperature range 315-350 "C an increase in volatiles yield is observed over that obtained from raw coal. Whether this increase is as much as that obtained from steam is not yet certain; the one available liquid yield is not better than that for untreated coal.
-
Modification of Coal by Subcritical Steam
Energy & Fuels, Vol. I, No. 1, 1987 87
region for coal. Temperatures between 300 and 350 "C form a threshold region for thermal processes in coal. Plastic coals begin to soften or melt in this range."14 A STEAM TREATMENT 7 small but significant literature points to important changes W 30 in the physical and chemical properties of coal occurring below 350 "C. Kirov and Stevens13 state that there is a breakdown of hydrogen bonding at 300 "C and substantiate this by noting the work of Maher et al.,15 who found K HELIUM the acidity of coal to decrease at this temperature. In W TREATMENT vacuum pyrolysis, Holden and Robb16 observed methane and alkylphenol evolution starting at 320 "C and believed that this temperature marked the onset of chemical dePRETREATMENT TEMPERPTURE PC) composition (in contrast to the release of physically bound Figure 3. Yields from pyridine extraction of raw and pretreated species released at lower temperatures). The amount and coals. kind of material extractable from coal is affected by lowtemperature heating in an inert gas" and in Solvent Extraction. Extraction of coal with various Aromatic extractables are increased after treatment at solvents has been used by other researchers to determine the amount of cross-linking in the coal ~ t r u c t u r e . ~ , ~ about 350 0C.24 Extract yield increases with extraction temperature, but it is significant that a jump in yield from However, the severity of the extraction can alter the 40 to 85% occurs between 300 and 320 0C.25 Additional amount and kind of material e~tracted.~.'Using too harsh evidence for low-temperature thermal activity is found in an extraction procedure might mask whatever changes had differential thermal analysis and differential scanning occurred in the coal due to the treatment. To avoid this, ca10rimetry.l~26-28 extractions were carried out at room temperature in a Thermal effects occurring in an inert gas, however, strong solvent, pyridine, for a short amount of time. Excannot fully explain our observations. For pyrolysis, traction yields from raw and pretreated coal are compared pretreatment in steam is far more effective than in an inert in Figure 3. Raw coal, when extracted in a nitrogen atgas. For extraction, inert-gas pretreatment decreases yield. mosphere as described above, gives values averaging about A further indication of the difference between steam and 17%. Through those points representing samples preinert gas is in the degree of agglomeration occurring during treated in steam and extracted under a nitrogen blanket pretreatment. No agglomeration is observed as a result can be drawn a curve that puts the maximum yield at a of pretreatment in inert gas even up to 375 "C. In steam, pretreatment temperature of 340-350 "C. on the other hand, light agglomeration is already observPyrolysis yields from pretreated coal depend on how the able at 320 "C. coal is handled after pretreatment. Pretreated coal exOne way in which steam may depolymerize coal is by posed to air prior to pyrolysis lost all the beneficial effects the disruption of hydrogen bonds. Hydrogen bonds are of the pretreatment. Liquid yields fell to the level observed important cross-links in coal. L a r ~ e states n ~ ~ that in 11for unpretreated coal. To test for this effect in extraction, linois No. 6 coal these are four times as numerous as coseveral batches of coal were pretreated and then divided valent ones. Hydrogen bonds rupture at 300 "C.13 By under a nitrogen atmosphere. Part of the sample was entering coal as a hydrogen-bonding solvent (see below for removed from the glovebag and extracted in air. In nearly evidence of the solubility of water in coal), water may assist all cases the yields from air-exposed samples were reduced to those observed for untreated coal. Although further work will be required for confirmation, the adverse effect of air on extraction yields parallels those previously ob(8) Loison, R.; Peytary, A.; Boyer, A. F.; Grilloy, R. In Chemistry of Coal Utilization, Supplemental Volume; Lowery, H. H., Ed.; Wiley: New tained in pyrolysis. York, 1963; p 150. The final piece of information which comes from these (9) Habermehl, D.; Orywal, F.; Bayer, H. D. In Chemistry of Coal experiments is that pretreatment in steam produces, on Utilization, Second Supplemental Volume;.Elliot, M., Ed.; Wiley: New York, 1981; p 317. subsequent extraction, a greater yield of soluble material (10) Anderson, W. C.; Roberts, J. J. SOC. Chem. Ind., London 1898, than coal pretreated in helium (Figure 3), a trend similar 17, 1013. to that observed for pyrolysis (although data for the latter (11) Brown, H. R.; Waters, P. L. Fuel 1966, 45, 41. are more limited). In all cases, even up to 375 "C, the coal (12) Illingworth, R. S . Fuel 1922, 1, 3. (13) Kirov, N. Y.; Stevens, J. N. Physical Aspects of Coal Carbonipretreated in helium remained nonagglomerated, unlike zation; Research Monograph; University of New South Wales: Sidney the steam pretreated coal, which is lightly agglomerated Australia, 1967; Chapter 12, p 58. at 320 "C. This is evidence that the change in coal pro(14) Neavel, R. C. Coal Sci. 1982, 1, 1. (15) Maher, T. P.; Harris, J. M.; Yohe, G. R. Rep. Inuest.-Ill., State duced by steam pretreatment is more than just a thermal Geol. Suru. 1959, No. 212. alteration. 50
Discussion
This work has established that pretreatment is effective for Illinois No. 6 coal at 50 atm only if carried out between about 320 and 360 "C. This is an interesting temperature (4) Peppas, N. A.; Hill-Lievense, M. E.; Hooker, D. T.; Lucht, L. M. "MacromolecularStructural Changes in Bituminous Coals During Extraction and Solubilization",Annual Report for Sept 1980-August 1981, U.S. Department of Energy Contract No. FG22-80PC30222, 1981. (5) Hombach, H.-P. Fuel 1980,59, 465-470. (6) Friedel, R. A.; Shultz, J. L.; Sharkey, A. G., Jr. Fuel 1968, 47, 403-407. (7) Ruberto, R. G.; Cronauer, D. C. Fuel Process. Technol. 1979, 2, 215-219.
(16) Holden, H. W.; Robb, J. C. Fuel 1960, 39, 39. (17) Dryden, G. C.; Parkhurst, K. S. Fuel 1955, 34, 363. (18) Hargen, J. J. SOC.Chem. Ind., London. 1914, 33, 389. (19) Guin, J. et al. Prepr. Pap.-Am. Chem. Soc., Diu.Fuel Chem. 1975, 20, 66. (20) Blessing, J.; Ross, D. A C S S y m p . Ser. 1971, No. 71, 172. (21) Bartle, K. D. et al. Fuel 1979, 58, 423. (22) Bartle, K. D. et al. Inst. Chem. Eng. S y m p . Series 1980,62, B1. (23) Squires, T. G.; Aida, T.; Chen, Y.-Y.;Smith, B. F. Prepr. Pap.Am. Chem. SOC.,Diu. Fuel Chem. 1983,28, 228. (24) Radke, M.; Willsch, H.; Leythaeuser, D. Geochim. Cosmochim. Acta 1982, 46, 1831. (25) Heredy, L. A. and Fugassi, Adu. Chem. Ser. 1966, 55, 448-459. (26) Glass, H. D. Fuel 1955, 34, 253. (27) Berkowitz, N. Fuel 1957,36, 355. (28) Mahajan, 0.;Tonita, A.; Walker, P. L., Jr. Fuel 1976,55,63-69. (29) Larsen, J. W. Prepr. Pap.-Am. Chem. Soc., Diu. Fuel Chem. Prepr. 1985, 30(4); 444.
88 Energy & Fuels, Vol. 1, No. 1, 1987
in the disruption of these bonds and prevent them from re-forming. Evidence for reaction of steam with carbonaceous materials at near pretreatment conditions is in the literature. In steam at 400 "C and 184 atm, the ether link in anisole reacts at a rate five times that for thermal rupture.30 If anisole is a model for ether-aryl linkages in coal, these data imply that steam has the ability to bring about depolymerization of the coal structure by cleaving what are otherwise thermally stable bonds. Ross31has further found that steam, under the same conditions, removes 40% of the coal's oxygen. In steam cracking of shale oil liquids at temperatures to 425 "C and pressures to 100 atm, Ty1efj2demonstrated a significant reduction in the average molecular weight of compounds with more than 25 carbon atoms and a simultaneous hydrogenation of the material. There is also a reduction in the sulfur and nitrogen content of the oil. The important point recognized by Tyler, as well as Allred33and Lewan et is that steam is not merely an inert medium for hydrocarbon reactions but rather a highly reactive and participatory agent. Since observations of the phase behavior of water and coal or coal liquids a t appropriate conditions are not available, it must be inferred from data on steam-hydrocarbon systems. Data for binary water-hydrocarbon systems are reviewed by Schneider.' Although these substances have little mutual solubility at ordinary conditions, even complete miscibility may be reached well below the critical point of water. In the naphthalenewater system, for example, the critical solution temperature is as low as 310 "C and 100 atm. Even though miscibility may not be complete at pretreatment conditions (the pressure is lower), wide regions of solubility are likely. This would have the following consequences: to the extent that water dissolves in plastic or molten coal, mobility is increased, thus improving transfer of internal donatable hydrogen. It also places water into intimate contact with coal for chemical reaction and makes ionic reactions possible. To the extent that large aromatic hydrocarbon fragments dissolve in a water phase, the rate of their repolymerization is reduced by dilution. (30) Ross, D. S.; McMillen, D. F.; Ogier, W. C.; Fleming, R. H.; Hum, G. M. "Exploratory Study of Coal Conversion Chemistry"; Quarterly Report No. 4; U.S. Department of Energy Contract No. DE-AC2281PC40785;Feb 19-May 18, 1982; p 26. (31) Ross, D. S. et al. "Exploratory Study of Coal Conversion Chemistry"; Quarterly Report No. 9; U.S. Department of Energy Contract No. DE-AC22-81PC40785;May 19-Aug 18, 1983; p 19. (32) Tyler, A. L. Oil Shale Symp. Proc. 1981,14th,212-219. (33) Allred. V. D. Oil Shale SvmD. Proc. 1979.12th. 241. (34) Lewan, M. D., Winters, C:; McDonald,'J. H. h e n c e (Washington,D.C.) 1979,203,897.
Graff and Brandes
Conclusions Two tests of steam-treated coal, flash pyrolysis in 50 atm of steam and extraction with pyridine, not only showed improved yields, but also showed parallel results in all other respects as well. The completely different nature of these two experiments provides a firm basis for the pretreatment effect. The literature supports a number of possible explanations for the pretreatment effect. The temperature range between 300 and 350 "C is certainly a very interesting one for coal. It is a threshold region for thermal processes where the softening of plastic coals is incipient. There is ample evidence that changes in chemical structure are induced merely by heating coal to these temperatures or below. With evidence of its reactivity in the literature, especially toward oxygen functional groups in coal and a model compound, we may no longer assume steam to be inert. Either a decrease in hydrogen bonding or oxygen cross-linking, resulting in partial depolymerization of raw coal, offers an explanation of increased liquid yields. The latter mechanism also offers a link to another experimental observation: pretreated coal is very sensitive to exposure to ambient air. The complex phase behavior of steam-hydrocarbon systems opens several interesting possibilities for physical and chemical interaction. Even though pretreatment conditions are subcritical, we may expect a condensed aqueous phase and a high degree of mutual solubility. Water may dissolve in coal, greatly increasing reactivity. While the literature offers several avenues for explaining our empirical observations, it does not allow us to choose among these or to provide a detailed description of the process. This must await further experimental work. From the experimental results and discussion above, we conclude that important changes in the structure of coal occur during treatment with steam at temperatures between 320 and 350 "C. It seems likely that coal is partially depolymerized during pretreatment, resulting in higher yields of volatiles in a subsequent pyrolysis or extraction. If this is indeed the case, steam pretreatment would lend itself well, as a first step, to a variety of coal conversion processes. Not only would it increase liquid yield, but it should also give lighter products. The increased yields demonstrated in steam pyrolysis conducted without added hydrogen or catalyst may raise this process to levels of commercial importance. Acknowledgment. Much credit is due to John Bodnaruk for construction and development of the experimental equipment. This work was supported by the U.S. Department of Energy under Contracts DE-AC2180MC14875 and 21315.
