Energy & Fuels 1994,8, 83-87
83
Effect of Catalyst Dispersion on Coal Liquefaction with Iron Catalysts? A. V. Cugini,* D. Krastman, D. V. Martello, E. F. Frommell, A. W. Wells, and G. D. Holder$ Pittsburgh Energy Technology Center, US.Department of Energy, Pittsburgh, Pennsylvania 15236 Received July 12, 1993. Revised Manuscript Received October 26, 1993"
The effectiveness of fine particle iron catalysts for first-stage coal liquefaction is influenced by the dispersion of the catalyst. Dispersion can be characterized by catalyst surface area, particle size, and crystallite size. Increasing catalyst dispersion by increasing catalyst surface area, and decreasing catalyst particle size and/or decreasingcatalyst crystallite size,results in higher levels of coal conversion to soluble products. For iron systems, a distinction must be made between the catalyst precursor and the active catalyst. Coal impregnation with iron catalyst precursors resulb in higher coal conversion than simple mixing of powdered catalyst precursors with coal. Impregnation of iron precursors onto coal affects catalyst dispersion by maintaining the fine particle size of the precursor during the transformation to the active catalyst phase. Agglomeration to larger particle size catalysts can occur if the iron catalyst precursors are physically mixed with the coal.
Introduction Dispersed (unsupported) catalysts have been used in first stage direct coal liquefaction studies.' Compared to supported catalysts, dispersed 'catalysts offer several advantages for first-stage coal liquefaction, including the absence of aging due to their short residence time in the reactor and, for inexpensive catalysts, such as iron, ease of disposal. Dispersed catalysts are typically introduced in the form of a precursor which is converted to the active form upon heating to the reaction temperature. The method of precursor addition and the physical properties of the precursor can affect the ultimate dispersion and activity of the catalyst which is formed during liquefaction. Intimate contact between coal and the catalyst precursor during the initial stages of coal liquefaction (where the catalyst is converted to an active form) may lead to greater dispersion of the active catalyst. Optimizing the physical properties of the catalyst precursor, i.e., increased surface area, smaller particle size, or smaller crystallite size, may also lead to a more dispersed, active phase catalyst. The extent of catalyst dispersion is often treated qualitatively. Studies have shown that methods of catalyst preparation that should result in enhanced levels of catalyst dispersion also result in the highest catalyst activities as measured by coal conversion to soluble or distillable products.14 It has been reported that the t Reference in this paper to any specific commercial product or service is to facilitate understanding and does not necessarily imply its endorsement or favoring by the United States Department of Energy. t Department of Chemical and Petroleum Engineering, University of Pittsburgh, Pittsburgh, P A 15261. Abstract published in Aduance ACS Abstracts, December 1, 1993. (1)Derbyshire, F. J. 'Catalysis in Coal Liquefaction: New Directions for Research"; IEA CR/08, IEA Coal Research, June 1988. (2)Cugini, A. V.;Ruether, J.; Krastman, D.; Cillo, D. L.; Balsone, V.; Smith, D. N. Prepr. Pap.-Am. Chem. Soc., Diu.Fuel Chem. 1988,33(I), 6. (3)Utz, B. R.; Cugini, A. V. US. Patent 5,096,570,1992. (4)Derbyshire, F. J.; Davis, A.; Lin, R.; Stansberry, P. G.; Terrer, M. T. Fuel Process. Technol. 1986,12, 127. (5)Joseph, J. Fuel 1991,70, 459.
effectiveness of iron catalysts can be improved by decreasing the particle size of the catalyst precursor.9-ll However, agglomeration of small catalyst particles has been observed under liquefaction conditions when the precursor is converted to the active catalyst, leading to larger particle catalysts that exhibit relatively low levels of activity.12-15 The presence of coal during catalyst activation and pretreatment of the catalyst precursor with sulfates can act to mitigate this effect.lk16 Another method of enhancing the activity of iron catalysts is to improve the contact between the coal and catalyst. Studies have shown that catalyst pretreatments that result in enhanced levels of contact between coal and the catalyst precursor result in higher levels of coal ~ o n v e r s i o n . ~ ~An~ Jexplanation ~-~~ for the improved coal conversion is that higher levels of catalyst dispersion may result from the enhanced contact. The enhanced contact (6)Artok, L.; Davis, A.; Mitchell, G. D.; Schobert, H. H. Fuel 1992,71, 981. (7)Barry, H.F.; Mitchell, P. C. H. Coal Liquefaction with Molybdenum Catalysts. InProceedings: lthlnternutional Conferenceon the Chemistry and Uses of Molybdenum; Climax Molybdenum Co.: Ann Arbor, MI, 1982;p 179. (8) Cugini, A. V. The Effect of Catalyst Dispersion and Coal-Catalyst Contacting on the Observed Activity of Unsupported Catalysts for Coal Liquefaction. Ph.D. Dissertation, School of Engineering, University of Pittsburgh, 1993. (9)Pregermain, S.Fuel Process. Technol. 1986,12,155. (IO)Andrea, M.; Charcosset, H.; Chiche, P.; Dvignon, L.; DjegaMaradassou, G.; Joly, J. P.; Simone, P. Fuel 1983,62,69. (11)Fukuyama, T.; Okada, T.; Takekawa, T.; Matsubara, K.; Moriguchi, S. Proc.: Int. Conf. Coal Sci. 1985, 181. (12)Srinivasan, R.; Keough, R. A.; Davis, B. H. Prepr. Pap-Am. Chem. SOC.,Diu.Fuel Chem. 1992,37 (3),1265. (13)Bacaud, R. Fuel Process. Technol. 1991,28,203. (14)Djega-Maradassou, G.; Besson, M.; Brodzik, D.; Charcosset, H.; Huu, T. V.; Varloud, J. Fuel Process. Technol. 1986,12, 143. (15)Stephens, H.P.; Stohl, F.; Padrick, T. D. Proc.: Int. Conf. Coal Sci. 1981,368. (16) Huffman, G. P.; Ganguly, B.; Zhao, J.; Rao, K. R. P. M.; Shah, N.; Feng, Z.; Huggins, F. E.; Taghiei, M. M.; Lu, F.; Wender, I.; Pradhan, V. R.; Tierney, J.; Seehra, M. S.; Ibrahim, M. M.; Shabtai, J.; Eying, E. M. Energy Fuels 1993,7, 285. (17)Garg, D.; Givens, E. N. Fuel Process. Technol. 1983, 7,59. (18)Mitra, J. R.; Chowdhury, P. B.; Mukherjee, D. K. Fuel Process. Technol. 1986,8,283. (19)Cugini, A. V.;Krastman, D.; Lett, R. G.; Balsone, V. Catal. Today, in press.
0887-0624/94/2508-0083$04.50/00 1994 American Chemical Society
Cugini et al.
84 Energy &Fuels, Vol. 8, No. I, 1994 Table 1. Analyses of Coal Feeds Blind Canyon Black Thunder DECS-6 proximate analysis (wt % , as recd) 19.2 4.7 moisture 34.8 42.4 volatile matter 47.3 40.6 fixed carbon 5.6 5.4 ash ultimate analysis (wt %, moisture free) 68.2 76.5 carbon 5.9 hydrogen 4.8 1.5 1.0 nitrogen 0.4 sulfur 0.4 9.9 18.8 oxygen (difference) 5.8 6.8 ash sulfur forms (wt % ) 0.02 0.01 sulfate 0.02 0.04 pyritic 0.41 0.30 organic
Illinois No. 6 4.2
36.9 48.2 10.7
70.2 4.8 0.9 3.1
9.9 11.1
0.03 1.20
1.90
or interaction of coal with iron catalyst precursors can also suppress the growth of iron catalyst particles,13J4 thereby promoting a high level of catalyst dispersion. Cugini et al.19 have shown that an incipient wetness technique is preferable to a technique in which excess water is used during the impregnation of coal with an iron catalyst precursor. The incipient wetness technique results in better contact between the iron catalyst precursor and the coal and seems to promote better catalyst dispersion. Recent work at PETC has centered on the development of an iron catalyst precursor that produces fine particles of high activity catalyst upon conversion to the active, sulfided phase. The procedure, reported previ~usly,~ results in the precipitation of hydrated iron oxide (FeOOH) directly on the coal surface. Failure of the FeOOH to be in intimate contact with the coal surface resulted in a lower level of catalyst activity as measured by coal conversion. The present study is an investigation of the surface area and particle size changes that occur during the transformation of the FeOOH precursor to the active sulfide phase, presumably pyrrhotite. The effect of temperature, precursor, and impregnation on the size of the particles formed is also discussed. Experimental Section Feed Coals. Experiments were conducted with Illinois No. 6 (from the Burning Star mine), Blind Canyon bituminous (DECS-6, from the DOE/Penn State Coal Sample Bank), and Black Thunder (from Campbell County Wyoming) subbituminous coals. Properties of the feed coals are presented in Table 1.
Catalyst Precursor Preparation. The catalyst precursor was added to the reactor as a dry powder, in an aqueous solution, or by precipitation onto the coal. Hydrated iron oxide (FeOOH) was dispersed onto the feed coals by an incipient wetness impregnation/precipitation approach.3 The precursors tested include powdered FenO,, aqueous ferric nitrate (from Fisher Scientific Co.), aqueous ferrous sulfate (from Fisher Scientific Co.), and powdered FeOOH. A sample of powdered Fe203, with a nominal particle size of 1pm (from Spang and Co.) was added as a dry powder. High surface area powdered FeOOH was prepared by precipitating FeOOH from an aqueous solution of ferric nitrate by the dropwise addition of ammonium hydroxide at room temperature. The precipitate was collected on a 0.45pm filter, vacuum dried at 40 "C, and ground to a powder. The Nz BET surface area of the FeOOH prepared in this manner was 138 m2/g. Coal was also impregnated with FeOOH. The impregnation procedure, described previously,3involved mixing
coalwithanaqueousferricnitratesolutionatthe incipientwetness ratio of aqueous solution to coal of approximately 1:1by weight. The wetted coal was then mixed with an excess solution of ammonium hydroxide to cause FeOOH to precipitate onto the coal from the ferric nitrate. The impregnated coal was collected on a 0.45-pm filter and was vacuum dried at 40 "C. Characterization of Catalyst Precursor Surface Interaction with Coal. Characterization of the interaction of the FeOOH with the coal resulting from the incipient wetness impregnation technique described above was done using a surface derivatization technique described previously.20This technique provides a measure of the surface hydroxyl groups on the coal by exposing the coal to hexamethyldisilizane (HMDS) which attaches to the OH groups as trimethylsilyl ether. The difference in the number of derivatized OH groups between the original coal and the coal loaded with iron (as FeOOH) was used to calculate the extent to which hydroxyl groups were blocked by the iron catalyst precursor. Liquefaction Studies. The effectivenessof catalysts formed from each precursor was determined by placing each precursor, together with reactants, in a 40-mL tubular microautoclave reactor. Experiments were conducted by adding 3.3 g of coal (-200 mesh) to the reactor with 6.6 g of PANASOL (a mixture of alkylated naphthalenes obtained from Crowley Chemical). Elemental sulfur (0.1 g) was added to the reactor to sulfide the catalyst precursors. The reactor was charged with lo00 psig (6.9 MPa) of hydrogen and sealed. The pressurized reactor was then heated to the liquefaction temperature (425 "C) in a fluidized sandbath. The heating period lasted 30-40 min. Following the liquefactionperiod (30min), the reactor was cooled rapidly (within 2 min) by immersion in cold water and depressurized. Coal conversion was calculated from the solubility of the coal-derived producta in tetrahydrofuran (THF)and in heptane as determined by a pressure filtration technique.21 Preparation of FeOOH Supported on Carbon Black. In several experiments, FeOOH was impregnated onto carbon black rather than coal to eliminate interferences from the indigenous pyrite in the coal on subsequent analyses. The carbon black was Raven 22 Powder obtained from Columbian Chemicals Co. The impregnation of carbon black with FeOOH was the same as the procedure described above for impregnation of coal with FeOOH. Catalyst Characterization. A series of iron sulfide catalysts were prepared from Fe203, FeOOH, aqueous ferric nitrate solutions, aqueous ferrous sulfate solutions, and FeOOH supported onto carbon black. The iron sulfide was prepared by adding the precursor to a 1-Lautoclave containing400 g of tetralin (obtained from Mallinkrodt, Inc.) and sufficient precursor to produce 4 g of iron sulfide (pyrrhotite). To convert the precursor to catalyst, the mixture was heated to one of a series of temperatures ranging from 150 to 400 "C and held for 0.5 h at 2500psig (17.3MPa)withH2/3% H2Spassingthroughthereactor at 4scfh (standard cubic feet per hour). The iron sulfidecatalysts were recovered as the residue in a THF extraction of the reactor products. The extracted particulate sampleswere examined using both X-ray diffraction (XRD) and transmission electron microscopy (TEM) to determine the iron sulfide phases present and the average sulfide crystallite and particle size. X-ray Diffraction. Sample preparation for XRD examination consisted of grinding the extracted (catalyst) residue in an agate mortar and pressing it into a flat, shallow cavity in a silica glass slide. A Rigaku computer-controlled diffractometer was used for these studies. Copper K a radiation was used with a graphite monochromator to remove the Cu KO, W, and background X-rays. A software correction was used to remove effects from Cu Ka2. A 20 scan rate of 0.03 deg per 2 s was used to generate the powder diffraction patterns. The resulting powder patterns were referenced to the Search Manual of the International Center for Diffraction Data.22 (20)Wells, A. W.; Frank, R. F.; Waldner, K. Prep?. Pap.-Am. Chem. SOC.,Diu.Fuel Chem. 1993,38 (l),311. (21)Utz, B. R.;Narain, N. K.; Appell, H. R.; Blaustein,B. D. Coal and Coal Products: Analytical Characterization Techniques; Fuller, Jr., E. L., Ed.; ACS Symp. Ser. 205;American Chemical Society: Washington, DC, 1982;p 225.
Effect of Catalyst Dispersion on Coal Liquefaction Table 2. Effect of Iron Precursor Surface Area on Coal Conversion of DECS-6 Blind Canyon Coal at 426 OC, 0.6 h, 1000 psig (cold) of H e 5000 ppm Fe, 0.1 g of S added to 9.9 g of a 2 1 Mixture of PANASOL to DECS-6 Coal coal conversion (%) to precursor precursor surface area (m2/g) THF sols heptane sols none none 58 30 FeOOH 138 66 34 Fez03 6 73 35 The approximate iron sulfide crystallite size was determined using the Scherrer equation:28L = KX/(@cos €9, where L is the mean dimension of the hkl crystal planes, K is a constant approximately equal to 1, related to the crystal shape and to the way p and L are defined, X = 0.154 059 8,@ is the 26 full width half-maximum (fwhm) width of the diffraction peak in radians, and 6 is the hkl diffraction peak position. Transmission Electron Microscopy. Sample preparation for TEM consisted of dispersing milligram quantities of the extracted liquefaction residue in electron microscopy grade acetone. Three minutes of sonication using a microtip ultrasonic probe dispersed most of the particles for observation of electron transparent single crystals and crystal agglomeratesin the TEM. After sonication, one drop of the mixture was placed on a FORMVAR-filmed TEM grid with a Pasteur pipet. The excess liquid was immediately removed by wicking onto the corner of a lint-free tissue. The grids were coated with =20 nm of carbon, based on a red interference color, in a Ladd Research Industries Vacuum Evaporator. A JEOL 200CX TEM with a La& electron source waa used for brightfield and darkfield imaging, and for selected area diffraction (SAD);200-kV excitation and a camera length of 137 cm were used for electron diffraction. Indexing of the SAD patterns was performed to verify that the crystals observed were the iron sulfide particles of interest. A MOOSstandard was used to calibrate the camera length of the TEM. For electron diffraction, a particle of interest was located in the field of view at 100 OOOX, and isolated using the selected area aperture. The microscope was adjusted for electron diffraction conditions and the sample was tilted until a hexagonal net representative of the 001 planes was displayed. The patterns were indexed using the Bragg equation for electron diffraction: RdMl= A&, whereR is the radialdistance between the transmitted and diffracted beam spots, dhkl is the interplanar spacing for the hkl planes, X, is the relativistic wavelength of the electron beam at 200 kV (0.002 51 nm), and L is the calibrated camera length (137 cm). BET Surface Area. The surface areas of the samples were determined from the adsorption isotherms for nitrogen measured at 77 K using the procedure developed by Brunauer, Emmett and Tellerau In these surface area calculations, a value of 0.162 nm2 was used for the cross-sectional area of nitrogen. A Coulter Omnisorb lOOCX gas sorption analyzer was used to measure the adsorption isotherms. Prior to this determination, the samples were outgassed for approximately 18 h at 95 O C under a vacuum of less than 104 Torr. For the adsorption measurements, the instrument was operated in the fixed-dose,static-flowvolumetric mode. Equilibration of the sample with the nitrogen dose was defined aa five consecutive readings with a variability of less than 0.1 Torr over a 3-8 interval.
Results and Discussion Catalyst Surface Area. The effect of iron oxide surface area on catalyst activity was investigated using powdered FeOOH with a surface area of 138 m2/g and micronized Fez03 with a surface area of 6 m2/g. Table 2 (22)Powder Diffraction File, JCPDS, International Center for Diffraction Data, 1990,1601 Park Lane, Swarthmore, PA 19061-2389. (23)King, H.E.;Alexander, L. E., X-Ray Diffraction Procedures for Polycrystalline and Amorphous Materials, 2nd ed.; John Wiley & Sons: New York, 1974;Chapter 9. (24)Brunauer, S.;Emmett, P. H.; Teller, E.. J. Am. Chem. SOC.1938, 60,309.
Energy & Fuels, Vol. 8, No. 1, 1994 86 Table 3. Effect of Precursor Type and Surface Area on the Resulting Iron Sulfide Surface Area Produced in a l-L Autoclave at 400 OC, 0.6 h, 2600 psig of Hd3% HzS surface area iron sulfide catalyst precursor (m2/g) surface area (mZ/g) micronized Fez03 powder 6 9 FeOOH 138 17 aqueous ferric nitrate n/a 30 aqueous ferrous sulfate n/a 32 Table 4. Effect of the Surface Area of Iron Sulfide Produced from Aqueous and Powdered Precursors on the Coal Conversion of DECS-6 Blind Canyon Coal at 426 OC, 0.6 h, 1000 psig (cold) of Hz, 6000 ppm Be, 0.1 g of S Added to 9.9 g of a 2:l Mixture of PANASOL to DECS-6 Coal coal conversion (7%) to iron sulfide THF heptane precursor used to generate iron sulfide surface area (m2/g) sols sols none none 58 30 micronized Fez03 powder 9 73 31 powdered FeOOH 17 70 31 30 80 38 aqueous ferric nitrate aqueous ferric sulfate 32 76 35
gives the effect of precursor surface area (iron oxide) on coal conversion. The precursor surface area does not appear to be important in determining the extent of coal conversion. This was not unexpected, since the precursor undergoes a chemical reaction to form the catalyst. The surface area of the catalyst itself is the important variable in determining the extent of conversion. Consequently, a series of tests were conducted to determine the effect of catalyst precursor type and surface area on the surface area of the catalyst (pyrrhotite) formed at 400 "C. The catalyst was formed from the precursor in tetralin with an HdHzS atmosphere as described in the Experimental Section. BET surface area measurements were conducted on the resulting pyrrhotite. Table 3 presents the surface areas of the resulting catalysts (pyrrhotites). The BET surface areas of the resulting pyrrhotites were significantly different than those of the original iron oxides. For the high surface area precursor, the surface area decreased from 138 to 17 m2/g,while the low surface area precursor slightly increased its surface area from 6 to 9 m2/g. The similarity of the surface areas of the pyrrhotites resulting from the solid iron oxide precursors helps to explain the similar coal conversions observed with each, although the surface areas of active catalysts in the liquefaction system containing coal may be slightly different. The surface areas of the pyrrhotites produced from aqueous precursors were both about 30 m2/g,which is greater than those from either of the iron oxide precursors. The next series of tests was aimed at investigating the effect of pyrrhotite surface area on coal conversion. Table 4 presents coal conversion as a function of iron sulfide surface area. There is not a significant difference in coal conversion for the different iron sulfide (pyrrhotite) catalysts. However, it appears that the higher surface area pyrrhotites, formed from the aqueous iron precursors, resulted in higher coal conversions than the pyrrhotites from the powdered iron oxide precursors. A broader range of pyrrhotite surface area needs to be tested to verify that a correlation exists between pyrrhotite surface area and coal liquefaction activity. Liquefaction Tests with Iron Catalysts Supported on Coal. The effectiveness of catalysts formed from FeOOH depends on the method by whichFeOOH is added to the system. Table 5 compares coal conversion using
86 Energy & Fuels, Vol. 8, No. 1, 1994
Cugini et at.
Table 5. Effect of FeOOH Mode of Addition on Coal Conversion of DECS-6 Blind Canyon Coal, Illinois No. 6, and Black Thunder Coal at 425 "C, 0.5 h, 1000 psig (cold) of H2,5000 ppm Fe, 0.1 g of S Added to 9.9 g of a 2:l Mixture of PANASOL to Coal precursor Blind Canyon none physically mixed FeOOH impregnated FeOOH Illinois No. 6 none physically mixed FeOOH impregnated FeOOH Black Thunder none physically mixed FeOOH impregnated FeOOH
Table 6. Estimated Crystallite Size from XRD Line Broadening of Iron Sulfides Generated from Carbon-Supported and Unsupported FeOOH reaction T ("C) phasealhkl plane crystallite size (nm) Iron Sulfide Catalyst Supported on Raven 22 Carbon Black 150 amorphous MI101 14 200 17 200 MI101 MI101 20 300 17 PI206 24 PI208 350 19 PI220 PI206 26 400 24 PI220
coal conversion (% to THF sols heptane sols
58 66 85
30 34 41
53 67 83
25 40 50
54 64 79
300 30 35 39 350
physically mixed FeOOH with impregnated FeOOH. The impregnated FeOOH is more active and results in higher coal conversion than the physically mixed FeOOH. As shown in Table 3, the surface area of the iron sulfide formed from powdered FeOOH (not impregnated) in tetralin was lower than the surface area of its precursor. This large reduction in surface area produced a catalyst which resulted in lower coal conversions compared to the conversionsobtained when the precursor was precipitated onto the coal. Characterization of Iron Catalysts Supported on Carbon. The surface area and crystallite size for the catalyst formed from FeOOH supported on the coal may not be the same as those measured for the catalysts formed from unsupported FeOOH. The pyrrhotite formed from coal-impregnated FeOOH is not easily characterized because of the presence of indigenous pyrite and other crystalline material in the coal. Therefore, a separate preparation of impregnated carbon black was prepared in order to see what effect impregnation has on crystallite size. A 2:l mixture of tetralin and impregnated carbon black (with no coal present) was subjected to liquefaction conditions and recovered by THF extraction. X-ray powder diffraction patterns of residue material from THF-extraction were compared with tabulated standards. The X-ray results are found in Table 6. The results indicated that the major iron sulfide phase for the carbon supported catalyst produced at temperatures below 300 "Cwas similar to marcasite (orthorhombic). However, at temperatures of 300 "C and above, X-ray peaks for marcasite decreased in intensity and disappeared, while those for pyrrhotite appeared. The unsupported iron sulfide phase only showed XRD peaks for pyrrhotite (hexagonal). Based on X-ray line broadening, average crystallite sizes were estimated for the identified phases (Table 6). The average size of the marcasite 101planes (face diagonal) in the supported iron sulfide samples increased as the temperature was increased from 200 to 300 "C. In a similar fashion,the average size of the pyrrhotite unit cell diagonal planes increased as the temperature was raised from 300 to 350 "C for the iron sulfide catalyst on carbon. Above 350 "Cthe average size of the pyrrhotite unit cell diagonal planes increased slightly. The temperature at which the pyrrhotite is initially formed (approximately 300 "C) may be the preferred activation temperature for the catalyst. This was confirmed in a separate study.lg Figures 1and 2 are darkfield TEM images of the iron sulfides produced at 300 and 350 "C,respectively. The
400 400 a
Unsupported Iron Sulfide Catalyst PI208 PI220 PI220 PI220 PI208 PI220 PI220 PI203 PI220 PI220 PI205 PI220
45 31 35 35 62 50 51 49 43 42 69 56
M = marcasite, P = pyrrhotite. ~~
I!
Figure 1. Iron sulfide catalyst on carbon black, 300 "C reaction temperature.
I
r
Figure 2. Iron sulfide catalyst on carbon black, 350 "C reaction temperature.
bright particle in the center of each electron micrograph is a strongly diffracting pyrrhotite crystal based on SAD patterns. The particles are agglomerates measuring around 20 to 40 nm. Note the carbonaceousmatrix around the iron sulfide particles.
Effect of Catalyst Dispersion on Coal Liquefaction
Energy & Fuels, Vol. 8, No. 1,1994 87 Table 7. Comparison of Surface Derivatization Reaulta of FeOOH-Impregnated Illinois No. 6 Coal with the Predicted Results from Yamashita's Models PPm Fe (as FeOOH) loading 2 500
5000 10 000 20 OOO 25 000 50 OOO
Figure 3. Unsupported pyrrhotite particle, 350 "C reaction temperature.
Figure 4. Selected area electron diffractionpattern of pyrrhotite particle, 001 plane.
Characterization of Unsupported Iron Catalysts. For the unsupported pyrrhotite samples, a general trend of increasing crystallite size with increasing reaction temperature is shown in the XRD data of Table 6. Figure 3 is a brightfield transmission electron micrograph of iron sulfide produced at 350 "C. This hexagonal pyrrhotite platelet measures about 225 nm across the 001 plane. A representative SAD pattern of a pyrrhotite platelet is shown in Figure 4. Note the regular hexagonal array from the 001 (basal plane) orientation. This pattern and similar ones were indexed using a MOOSstandard to calibrate the TEM camera length. Comparison of Supported with Unsupported Iron Catalysts. Comparingthe supported versus unsupported iron sulfides, it appears that, for a given reaction temperature, the iron sulfide on carbon consisted of smaller crystallites than the unsupported catalyst. This general size difference is evident comparing the TEM images of supported versus unsupported iron sulfideparticles, taken at 100 000X. The difference in size may be attributed to some type of interaction between the iron sulfide crystals and the carbon support that effectively constrains the growth of the crystals. A similar effect may be expected for FeOOH impregnated on coal. Surface Interaction between Catalyst Precursor and Coal. According to a model of Yamashita, the iron catalyst precursor, FeOOH, binds directly with coal hydroxyl groups on the surface of the coal.25 It seemed likely that, if FeOOH were chemically bonded to coal surfacehydroxyl groups,then the affected hydroxylgroups (25) Yamashita, H.; Ohtauka, Y.; Yoshida, S.;Tomita, A. Energy Fuels 1989,3,686.
mmoles loading
mmoles OH/g blocked by model measurement
0.0062 0.0125 0.025 0.0500 0.0625 0.1250
0.013 0.025 0.050 0.100 0.120 0.250
0.02 0.03 0.07 0.10 loading limit 0.10 0.10
would be blocked from derivatization with gaseous hexamethyldisilazane (HMDS). To evaluate the effectiveness of the iron loading procedure described above, a series of FeOOH-impregnated Illinois No. 6 coal samples were prepared at iron loadings ranging from 2500 to 50000 ppm Fe. These samples were subjected to gas-phase silylation. To calculate the surface hydroxyl groups blocked by the added iron in the FeOOH impregnation, the results for the FeOOH-impregnated samples were subtracted from the surface hydroxyl content of the nonimpregnated control sample and compared to the Yamashita model. These results are summarized in Table 7. These results indicate that the number of "free" hydroxyl groups decreases with increasing iron loading until the available groups are saturated. This indicates that the iron, as FeOOH, may be binding to surface hydroxyl groups. The adsorption of the FeOOH to the surface of the coal may help to stabilize the iron so that it does not agglomerate when converted to its active form. This would explain why the pyrrhotite formed from the supported FeOOH is smaller than the pyrrhotite formed from the unsupported FeOOH. The saturation point, 20 000 ppm Fe, also represents the point at which the added iron is no longer bound to the coal. A separate series of results, not reported, indicated that the iron added beyond 20 000 ppm was less active. At higher loadings, the iron was visible as a separate entity, distinct from the coal. Conclusions In the absence of a carbonaceous support, the transformation of FeOOH to iron sulfide results in a loss of surface area, possibly due to agglomeration. This has been previously documented1215 and shows that catalyst precursor surface area does not affect coal conversion in the limited range of surface areas investigated. However, an increase in the surface area of the actual (iron sulfide) catalyst does improve liquefaction yields. It is recommended that iron sulfide preparations with a broader range of surface areas be investigated. The presence of a carbonaceous support for FeOOH tends to mitigate particle size growth and favors the formation of smaller particle size iron sulfide catalysts that are likely to have higher specific surface areas. One of the effects of impregnation of the FeOOH is the generation of smaller iron sulfide particles in the system. It is likely that the same effect occurs in coal impregnated with FeOOH, since addition of the precursor through impregnation results in higher coal conversions. Acknowledgment. The authors thank Sidney S. Pollack and Neil Johnson for their assistance in the X-ray diffraction study and Robert A. Ference for his assistance in the BET study. They also thank Raymond Bernarding and Theodore Jordan for their assistance in the liquefaction operations.
