Energy & Fuels 1989,3, 421-425
42 1
High Liquid Yields from Bituminous Coal via Hydropyrolysis with Dispersed Catalysts C. E. Snape*$tand C. Bolton Coal Research Establishment, British Coal, Stoke Orchard, Cheltenham, Gloucestershire GL52 4RZ, U.K.
R. G. Dosch and H. P. Stephens* Sandia National Laboratories,. Albuquerque, New Mexico 87185 Received December 27, 1988. Revised Manuscript Received March 23, 1989
Results of catalytic hydropyrolysis experiments using a U.K. bituminous coal with several catalysts and dispersion techniques are described. Single-stage catalytic tests carried out in a bench-scale fixed-bed reactor operated at 150 bar and 500 "C have produced daf coal basis tar yields greater than 60%, with weight ratios of tar to hydrocarbon gases up to 80% higher than those observed in an uncatalyzed case. Use of well-dispersed sulfided molybdenum or hydrous titanium oxide catalysts coated directly on the coal is superior to use of Lewis acids and alumina-supported hydrogenation catalysts in terms of tar yields achieved and reduction of the amount of light hydrocarbon gases produced. In addition to the single-stage catalytic hydropyrolysis tests, results of catalytic two-stage experiments demonstrated that the tar produced by the first stage is readily upgraded to distillate containing low concentrations of heteroatoms.
Introduction In some respects, coal hydropyrolysis is potentially a more attractive route for the production of liquid fuels than direct liquefaction techniques utilizing solvents. It offers a process configuration that avoids the use of a recycle solvent, which constitutes up to two-thirds of the reactor feed stream for direct liquefaction. However, development of technology for feeding ground coal into a high-pressure reactor is required, and historically, pyrolysis processes have been associated with low conversions of coal to liquid making char the prinicpal product for which some application, apart from providing process heat and hydrogen, must be found. Moreover, large quantities of methane are usually produced during hydropyrolysis a t temperatures above 600 "C, which are required to achieve substantial liquid yields. Consequently, hydrogen consumptions are high. Previously reported results44 of two-stage hydropyrolysis studies, in which primary tar vapors are passed through a catalyst bed, have shown that daf coal basis yield of 20-25% distillate with only 3-470 methane can be achieved by noncatalytic hydropyrolysis of bituminous coal. For those experiments hydropyrolysis was carried out a t a pressure of about 150 bar and a temperature of 500 "C, followed by catalytic upgrading of the tar a t a temperature of 400 "C using a commercially available Ni-Mo/alumina hydrotreating catalyst. Such catalysts have been widely used in coal liquefaction to upgrade primary liquefaction products to distillable liquids.' Previous ~ o r k established *~~ that coals could be catalytically hydrogenated in batch reactors in the absence of a solvent to give high conversions to pyridine-soluble materials. Impregnated molybdenum catalysts were among the most active studied. More recent worklo has demonstrated that dispersed sulfided Mo is effective under mild conditions; over 50% chloroform-soluble material can be generated from bituminous and subbituminous coals 'Present address: Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow G1 lXL,Scotland.
0887-0624/89/2503-0421$01.50/0
Table I. Analysis of Linby Coal Proximate Analysis moisture (as received) ash (dry basis) volatile matter (daf basis)
9.8% 6.9%
37.9% 83.0% 5.5%
Ultimate Analysis (Dmmf Basis) C 1.0% S (organic) H 1.9% N
0
8.7%
Maceral Analysis (Volume Basis) 10% exinite 74% vitrinite 16% inertinite
by reaction at 400 "C and a cold-charge hydrogen pressure of 70 bar. Lewis acids, such as zinc and stannous chlorides, have been shown to enhance liquid yields in hydropyrolysis,"J2 but relatively large concentrations of catalyst are required. Other studies have shown that Mo could be (1)Gavalas, G. R. Coal Pyrolysis; Elsevier: Amsterdam, 1985. (2)Chakrabartty, S. K.; du Plessis, M. P. Modern Coal Pyrolysis; Information Series 95; Alberta Research Council: Edmonton, Canada, 1985. (3)Furfari, S. IEA Report No. ICTIS/TRZO, 1982. (4)Snape, C. E.; Martin,T. G. Proceedings of the Round Table Meeting; Commission of the European Communities: Brussels, 1987; EUR 10588. D 167. ( 5 ) Bolt& C.; Snape, C. E.; Stephens, H. P. Prepr. Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1987, 32(1), 617. (6)Stephens, H.P.; Bolton, C.; Snape, C. E. 2987 Internationul Conference on Coal Science; Coal Science Technology -. 11; Eleevier: Amsterdam, 1987;p 703. (7) Derbyshire, F. Catalysis in Direct Coal Liquefaction: New Directions for Research; IEA CR 08; IEA Coal Research London, 1988. (8)Weller, S.:Peliuetz, M. Friedman, S.; Storch, H. H. Ind. Eng. Chem. 1950, 42,330. (9) 622.
d.:
Hawk, C. 0.: Hiteahue, R. W. Bull.-U.S. Bur. Mines 1965, No.
(10)Derbyshire, F. J.; Davis, A,; Lin, R.; Stansberry, P. C.; Terrer, M. T. Fuel Process. Technol. 1986,12, 127. (11)Wood, R.E.; Wiser, W. H. Ind. Eng. Chem. Process Des. Deu. 1976, 15, 144.
(12)Kershaw, J. R.;Barrass, G.; Gray, D. Fuel Process. Technol. 1980,
3, 115.
0 1989 American Chemical Society
422 Energy & Fuels, Vol. 3, No. 3, 1989
Snape et a1.
Table 11. Product Distributions Resulting from Single-Stage Hydropyrolysis Experimentsa product distribn, wt % (daf coal basis) selectivic2-c4 t Y (prod catalyst (wt % active hydrowt ratio metal (daf coal basis)) tar CH, carbons char (tar/g.as)) none 4 26 3 60 3.7 7 ZnC12 (5%) 36 6 45 2.8 7 SnC12 (1%) 54 6 30 4.1 4 Ni-Mo/ Alumina (3.6 ?%) 42 3 42 6.0 4 CoNiMoHTO bulk 38 2 46 6.3
CONTROL THERMOCOUPLE
PRESSURE HYDROGEN
COALKATALYST BED
ELECTRICAL CONNECTORS
WIRE WOOL PLUG MUFFLE FURNACE
CATALYST BED
WIRE WOOL PLUG
(2.9%)
PdHTO bulk (2.0%) NiHTO coating (0.5%) NiHTO coating (1.0%) PdHTO coating (0.7%) PdHTO coating (1.6%) MoS2 (0.6%)
47 50 53 64 62 59
3 4 4 3 4 4
4 6 6 8 7 6
35 32 29 16 20 20
6.7 5.0 5.3 5.8 5.6 5.9
Experiments performed at 500 "C final hydropyrolysis temperature and 150 bar pressure. Products included -6% water and
CONTROL THERMOCOUPLE DRYICE COOLED TRAP
-
TO GAS COLLECTION AND MEASURING SYSTEM
Figure 1. Diagram of two-stage hydropyrolysis reactor.
a
-2%
co + cop
used to achieve hydropyrolysis yields exceeding the coal proximate volatiles content.13J4 However, sulfided Mo, which is thought to be the most active form for coal conversion, was not used. Moreover due to the high temperatures and pressures used, secondary reactions could not be controlled, resulting in a low selectivity to liquid products. In this paper, we report the results of catalytic hydropyrolysis experiments in which greater than 60% daf coal basis yields of tar are produced with weight ratios of tar to hydrocarbon gases up to 80% higher than observed in an uncatalyzed reaction. The work has also shown that hydrogenation catalysts dispersed directly on the coal are superior to Lewis acids and alumina-supported hydrogenation catalysts mixed with the coal in terms of tar yields achieved and reduction of the amount of light hydrocarbon gases produced. In addition, we have performed two-stage hydropyrolysis tests which demonstrate that the tar produced is readily upgraded to distillate containing low concentrations of heteroatoms.
Experimental Section Coal a n d Catalysts. A high-volatile United Kingdom bituminous coal (Linby),ground to 75-150-pm particle size range, was used for these tests. Analysis of the coal is given in Table I. Catalysts added to the coal for the hydropyrolysis experiments described below were (1) zinc and stannous chloride Lewis acids, (2) a commercial Ni-Mo/alumina, 3) bulk hydrous titanium oxide (HTO) formu1ation~'~J~ (Pd and Co-Ni-Mo), (4) coatings of Niand PdHTOs and (5) sulfided Mo. Lewis acid catalysts were impregnated into the coal from aqueous solution. Powdered (-200 mesh) alumina-supported and bulk HTO catalysts were physically mixed with the coal, the catalyst weight being 20% daf coal. Niand PdHTO catalyst coatings were dispersed by coating the coal with a thin layer of sodium hydrous titanium oxide15J6(NaHTO) followed by contact of the NaHTO-coated coal with aqueous solutions of either Ni or Pd. This resulted in the active metal begin incorporated into the hydrous titanate coating via ion exchange for Na'. Sulfided Mo catalyst was dispersed by wetting the coal with solutions of (NH4)2M~S4 or (NH4)2M002S2, which (13) Hiteshue, R. W.; Friedman, S.; Madden, R. Rep. Invest.-US. Bur. Mznes 1962, No. 6027; 1962, No.6125;1964, No.6376; 1964, No. 6470. (14) Schroeder, W. C. U.S.Patents 3,030,297, 1962, 3,152,063 1964, 3,926,775, 1975. (15) Dosch, R. G.; Stephens, H. P.; Stohl,F. V., U.S.Patent 4,511,455, 1985. (16) Stephens, H. P.;.Dosch,R. G.; Stohl, F. V. Ind. Eng. Chem. Prod. Res. Dev. 1985, 24, 15.
rapidly decompose to MoS2 at about 200 "C during the heat-up period in the hydropyrolysis reactor. The amounts of active metals in the coal-catalyst mixtures are given in Table 11. For the two-stage tests, a presulfided commercially available 10% Mo, 2.5% Ni on alumina catalyst in the form of 1.6 mm diameter by 4 mm length extrudates was used in the second stage of the hydropyrolysis reactor. Apparatus and Procedure. A diagram of the hydropyrolysis reactor used for the tests is shown in Figure 1. The fixed-bed tubular down-flow reactor is divided into two heated zones, thereby allowing two-stage operation. The upper section, used for hydropyrolysis of coal or coal-catalyst mixtures, is heated by passing high currents between the two electrical contacts on the reactor tube. The resistive heating provides a uniform temperature along the hydropyrolysis zone. Temperature was monitored by a thermocouple in the coal bed. For all of these experiments either 10 g of neat coal or 5 g of the coal-catalyst mixture was combined with 10 g of sand (to avoid plugging the reactor during hydropyrolysis),packed into the reactor, and supported between two wire wool plugs. The lower section, heated by a muffle furnace, can be used for hydrotreating tar vapors produced in the hydropyrolysis section. Liquid products, including water, were condensed in the dry ice trap, and gaseous products were collected for analysis. The types of experiments and pertinent conditions are briefly described below. Procedures for operation of the hydropyrolysis reactor are reported in greater detail el~ewhere.49~ To investigate the influence of temperature and pressure on conversions with dispersed catalyst, initial single-stage tests with coal impregnated with sulfided Mo were carried out a t 475-550 "C and 1-200 bar. For comparison, uncatalyzed tests were performed a t 100,150, and 250 bar at 500 "C and 150 bar a t 600 "C. Following completion of these tests, single-stage hydropyrolysis experiments to compare the activity of the various catalysts were carried out a t 150 bar of hydrogen pressure and the relatively low temperature of 500 "C. The hydropyrolysiszone was heated a t a rate of 5 "C/s from ambient temperature and held at temperature for 10 min. Hydrogen was passed downward through the reactor a t a flow rate equivalent to 15 L/min a t standard temperature and pressure. In addition, two two-state experiments were performed: the first with a noncatalytic hydropyrolysis stage followed by the presulfided Ni-Mo/alumina catalytic upgrading stage5v6operated at 400 "C and the second with dispersed MoS2catalyst (0.6% Mo, daf coal basis) in the hydropyrolysis stage followed again by presulfided Ni-Mo/alumina catalyst in the second stage. A flow rate equivalent to 5 L/min at standard temperature and pressure was used with 8 g of hydrotreating catalyst in these tests to give an estimated vapor residence time of 10 s for the second stage. Product Recovery and Analysis. Gas from the reactor was analyzed for C1-C4 hydrocarbons, CO, and COz. After each experiment, char and ash remaining in the reactor tube were removed and weighed. Tar and liquor (for the single-stage experiments) or liquid products (for the two-stage experiments) collected in the reactor product cold trap were first weighed and then recovered with dichloromethane (DCM). Water was removed
Energy & Fuels, Vol. 3, No. 3, 1989 423
High Liquid Yields from Bituminous Coal 100
80
60
40 L
I
20
0
450
" 475
500
525
550
575
600
0
50
100
150
200
250
FINAL TEMPERATURE ( O C ) Figure 2. Effect of temperature on product distribution for the hydropyrolysis of Linby coal at 150 bar pressure: Mo-catalyzed (0.6%) yields, (m, -) tar,+(^, -) char, and (0, -) Cl-C4 hydrocarbons; uncatalyzed yeld, (A,- - -) tar.
HYDROGEN PRESSURE (bars) Figure 3. Effect of hydrogen pressure on product distribution for the hydropyrolysis of Linby coal at 5 OC/s to 500 "C: Mo-) C& catalyzed (0.6%) yields, (m, -) tar,(0,-) Char,and (0, hydrocarbons; uncatalyzed yield, (A,- - -) tar.
by phase-separating paper and weighed, and the DCM was evaporated to give tar samples for analysis. The daf coal basis percentage yield of tar for each experiment was calculated as the weight of material recovered in the cold trap less the weight of water produced. The tars were separated into toluene insolubles, asphaltenes (toluene soluble, n-pentaneinsoluble), and n-pentane solubles and were analyzed by a number of techniques. Analyses for oxygen, phenolic OH contents by enthalpimetric titration, aromatic hydrogen contents by 'H NMR, and number-average molecular weight by vapor-pressure osmometry and size exclusion chromatographicanalyses were carried out as described previo~sly.~'n~~ For the two-stage experiments, the liquid products exiting the second stage of the reactor were recovered neat in order to determine the fraction of light naphtha, measured by gas chromatography as the concentration of constituents with boiling points below that of o-xylene (150"C).
(Figures 2 and 3) are comparable to the inertinite content of Linby coal (Table I), it is tempting to assume that the chars obtained from these and the high conversion experiments with the other catalysts are composed mainly of inertinite-derived material. Impact of First-Stage Hydropyrolysis Catalysts. The results of the investigation of the influence of temperature and pressure on conversions with the dispersed MoS2catalyst were used to set the conditions for the single-stage tests with the other catalysts. Conditions of a final hydropyrolysis temperature of 500 "C and a hydrogen pressure of 150 bar were chosen to achieve a high selectivity to tar a t a reasonably low operating pressure. At higher temperatures, up to 525 OC, somewhat higher tar yields could be achieved, but a t the expense of higher hydrocarbon gas yields. The yields of tar, methane, C2-C4 hydrocarbons, and char from the single-stage catalytic hydropyrolysis experiments a t these conditions are listed in Table 11, together with the yields of these products from the corresponding noncatalytic test. As can be seen in Table 11, all of the catalyzed experiments produced significantly more tar than the noncatalytic case. Examination of the results of the various catalysts indicates that tar yields and product selectivity are dependent on the catalyst or active metal employed and the technique used to disperse the catalyst with the coal. The greatest tar yields, 5944%, were achieved with well-dispersed catalysts containing metals known to be active for hydrotreating and coal conversion. The effect of the active metal may be seen by comparison of the experiments with Ni and Pd HTO coatings and MoSP As anticipated, the Pd and Mo catalysts gave higher tar yields (59-64%) than the Ni catalysts (-50%). With respect to dispersion, PdHTO catalysts coated onto the coal and the MoS2dispersed from aqueous solution gave higher yields than the PdHTO and Ni-Mo/alumina finely divided catalyst powders that were physically mixed with the coal, even though the coal basis amount of active metal was significantly greater for the tests with powdered catalysts. The Lewis acid catalysts, ZnClz and SnCl,, dispersed onto the coal from aqueous solution, produced tar yields of 36% and 54%, respectively, despite a 5-fold greater catalyst weight for the ZnC1, experiment. Although SnC1, is a weaker Lewis acid than ZnC12,and consequently has lower activity for cracking reactions, it is probably a much more effective catalyst for hydrogenation.' Results of experiments performed by varying the concentration of the Ni and Pd for the HTO-coated coal
Results and Discussion Effects of Temperature and Pressure. Figure 2 shows the effect of temperature on product distribution a t 150 bar of hydrogen pressure and a Mo loading of 0.6% daf coal. Tar yields of 5045% over the temperature range 475-550 "C are all considerably higher than the 26-34% yields obtained for the uncatalyzed experiments at 500 and 600 "C, respectively. The maximum tar yield of 65% for the Mo-catalyzed experiments was achieved a t a temperature of 520 "C, which is considerably lower than that of 620 "C for peak tar yield in the absence of catalyst.lg The yield of C1-C4 hydrocarbons increases monotonically across the temperature range investigated, and at temperatures above about 520 "C, this increase appears to be a t the expense of tar production. The effect of hydrogen pressure on product distribution for hydropyrolysis performed at 500 "C and between 1and 200 bar, shown in Figure 3, is much more pronounced with the dispersed MoS2than without catalyst. Tar yields for the catalyzed experiments reached 60% a t 200 bar. Provided the superficial gas velocity was kept constant:g@ tar yields for the experiments without catalyst increased monotonically from 24% to 36% over a pressure range of 100-250 bar. Because the char yields obtained from the MoS2-catalyzed experiments with the higher tar yields ~~
(17)Snape, C. E.;Bartle, K. D. Fuel 1984,63,883. (18)Bolton. C.:Riemer, C.: Snape, C.E.;Derbyshire, F. J.; Terrer, M. T. h e l 1988,67,.901. (19)"Production of Distillate Fuels and Chemical Feedstocks from Coal";Final Report on ECSC Project No. 7220-EC/817,1982. (20) Bolton, C.; Snape, C.E.;OBrien, R. J.;Kandiyoti, R. Fuel 1987, 66,1443.
Snape et al.
424 Energy & Fuels, Vol. 3, No. 3, 1989
Table 111. Composition of Primary HydropyrolysisTar and Second-Stage Product for Experiments with and without a Catalyzed First Stage' uncatalyzed first stage
catalyzed first stageb
% toluene insolubles % asphaltenes % n-pentane solubles
84.3 6.8 5.3 1.5 0.9 4.3 0.97 36 250 3 40 57
86.1 7.0 3.6 1.7 0.4 3.0 0.98 33.5 270 2 29 69
second-stage product,' wt % of product C H OH ppm of N H/C atomic ratio % aromatic H of total H light naphtha, wt % of daf coal
84.0 13.6 0.01 7 1.94 4 8.3
88.0 11.9 0.05 33 1.62 7 11.2
primary tar comp, wt % of product C H
0 N S
OH H/C atomic ratio % aromatic H of tot. H
Mn
"First stage conditions: 500 O C , 150 atm of hydrogen pressure. Second stage conditions: 400 O C , 150 atm of hydrogen pressure. bFirst stage catalyst: 0.4%, coal basis, MoS2. cSecond stage catalyst: presulfided Ni-Mo/alumina.
samples showed that doubling the concentration of the active metal had an insignificant effect on the conversion to tar. These results suggest that further experiments are needed to determine the minimum amount of active metal required to achieve high yields. Indeed, we have performed experiments similar to those reported here that demonstrate concentrations of Mo as low as 0.1% can be used without sacrificing tar yield.21 With respect to hydrocarbon gas formation (Table 11), most of the catalytic experiments produced 2-4% methane, which was nearly equal to that observed for the noncatalytic experiment (3%). However, the Lewis acid catalysts produced significantly more methane (6%). The yield of C2-C4 hydrocarbon gases produced by the noncatalyzed case was 4%; the catalyzed experiments produced yields from 4% to 8%. Because liquids are the desired hydropyrolysis products and gaseous hydrocarbons are not only of low value but result in wasteful consumption of hydrogen, a useful figure of merit, shown in Table 11, for comparison of the efficiency of conversion is the selectivity defined in terms of the weight ratio of tar to gas yield. Compared to the selectivity for the noncatalyzed experiment, 3.7, the selectivities for the Lewis acid catalysts, 2.8and 4.1,a t best offer little improvement. However, the other catalysts achieved selectivities of 5.0-6.7, amounting to relative improvements in selectivity ranging from 35% to 80%. Composition of Tars and Hydrotreated Products. The two-stage experiments were performed to determine if the tar produced in greater yields from catalyzed hydropyrolysis can be hydrotreated in a second stage to produce a high-quality liquid product. Table I11 summarizes the compositional data for the primary tars produced by uncatalyzed and MoS2-catalyzedhydropyrolysis and the liquid products resulting from second-stage vapor-phase hydrotreatment of the primary tars with pre(21) 'Direct Conversion of Coal to Chemical Feedstock"; Final Report on ECSC Project No. 1220-EC:/827.
I
A Loa*:w
t I
I
I
I
1
I
1
I
I
10
12
14
16
18
20
22
24
ELUTION TIME (min) Figure 4. Size exclusion chromatogram of hydropyrolysis tars produced at 500 O C and 150 bar of hydrogen pressure: (A) without catalyst; (B)with MoSz catalyst.
sulfided Ni-Mo/alumina catalyst. As can be seen from Tables I1 and 111, although tar yield for the catalyzed experiment was more than double that for the uncatalyzed case, apart from the differences in oxygen and sulfur contents, the compositions of both tars were similar. The tar obtained with MoS2contained less oxygen and sulfur and had a slightly higher number average molecular weight as determined by vapor pressure osmometry and as also evidenced by size exclusion chromatograms of the two tars depicted in Figure 4. The higher pentane solubility of the tar obtained with MoS2 may be attributed to the lower concentration of polar OH groups.lB However, the similarity in the two tars suggests that there is some degree of intramaceral uniformity in the structure of vitrinite, which is undoubtedly the major contributing material to the additional tar obtained with catalyst. From the compositions of the primary tars produced at 500 "C and 150 bar pressure, it was estimated that the hydrogen consumption with dispersed MoS2 was approximately 3% compared to about 1.5% without catalyst. Although most of the increased hydrogen consumption for the catalyzed experiments was due to the greater tar production, some of the additional hydrogen was obviously required for heteroatom removal and C-C bond cleavage resulting in hydrocarbon gases. It is probable that the reactions of form tar are preceded by hydrogenation a t relatively low temperatures (
