Coal liquefaction catalysis

Coal liquefaction

catalysis

Iron pyrite and hydrogen Robert

M. Baldwin

sulphide

and Stephen

Vinciguerra

Chemical and Petroleum-Refining Engineering Department, Colorado School of Mines, Golden, CO 80401, USA (Received 11 March 1982; revised 14 July 1982)

An unreactive hvC bituminous coal has been hydrogenated in a batch-stirred reactor using pyrite, hydrogen sulphide, and pyrite+ hydrogen sulphide as catalysts. The data indicate that H,S is an active homogeneous catalyst for coal liquefaction, and suggest that pyrite may be acting indirectly as a catalytic agent via H2S release. (KeWods:

coal; catalysts; liquefaction;

bituminous coal)

Direct hydrogenation processes for coal liquefaction are generally hampered by two problems: slow formation of distillate oils at ‘mild’ conditions and poor hydrogen utilization efficiency due to formation of light hydrocarbon gases (C,-C,). Unfortunately, raising the temperature to increase the rate of oil formation causes the rate of light hydrocarbon formation to greatly accelerate, leading to even poorer hydrogen efficiency and hence increased processing costs. To overcome these problems, catalysts may be used which are selective for hydrogenation and which accelerate this rate so that lower temperatures may be employed. It is well known that the reactivity of certain coals liquefied by the I. G. Farben process during the 1930s in Germany was enhanced by addition of iron and/or sulphur to the feed slurry’. In the early 197Os, Wright and Severson reported that minerals present in bituminous coal mineral matter served as hydrogenation catalysts. Additionally, Na, K and Fe were found to catalyse CO-steam lignite liquefaction. Subsequent to these discoveries, research on disposable catalyst liquefaction of coal was initiated to identify the active catalytic species in coal mineral matter. Mukherjee and Chowdhury3 found increasing conversion with increasing mineral content and identified iron pyrite as the active catalyst. They also indicated a synergistic effect between pyrite and organic sulphur. Extensive research by Guin et aL4-’ and Tarrer et al.* on the catalytic activity of coal minerals clearly established pyrite as a hydrogenation catalyst and further identified other catalytic agents present in coal mineral matter. Hamrin’ investigated hydrodesulphurization of model compounds with coal minerals as catalysts and Granoff et al.“, in batch autoclave studies of mineral-matter catalysis, demonstrated the effect of pyrite on product distribution and illustrated the magnitude of the observed catalytic effect on net oils formation. More recently, Mossbauer analysis of liquefaction residues by Montano et al.’ ’ - ’ 3 has led to a greater understanding of the 00162361/83/050498-04%3.00 @I 1983 Butterworth& Co. (Publishers) Ltd

498

FUEL,

1983,

Vol62,

May

behaviour of iron/sulphur species under liquefaction conditions. Attar and Martin14 speculated that an iron sulphide intermediate (between FeS,,,, and FeS,) is the active pyrite-derived catalyst whilst recently, Stenberg’ 5 reported on the use of hydrogen sulphide as a homogeneous catalyst for liquefaction of lignite. This Paper presents the results of batch autoclave hydrogenation experiments where H,S, pyrite, and H,S + pyrite were used as liquefaction catalysts. Products were analysed for oils, asphaltenes, and preasphaltenes and the effect of additives on the coal liquefaction rate as well as product distribution was determined. EXPERIMENTAL An hvC bituminous coal from the Wadge seam of the Energy Fuels Mine near Yampa, Colorado (Rocky Mountain coal province) was hydrogenated in a 300ml stirred-batch reactor. The coal employed was similar to sample PSOC-233 from the DOE/Penn State Coal Sample Bank, and was chosen because of its low inherent pyrite and total sulphur content and relatively poor thermal liquefaction reactivity. An analysis of the coal sample is presented in Table 1. Ash analysis showed 47 wt% and 33 wt% SiO, and A1203 respectively and 3.3 wt% total iron as Fe,O,. Titanium as TiO, was 1.05 wt% of the ash; no other potential catalysts were indicated in the ash analysis. Pyrite was separated from a Kentucky bituminous coal from the Colonial Mine near Madisonville, KY, by first grinding the parent coal then separating pyrite from coal on a shaker table. The pyrite was analysed by X-ray diffraction and Miissbauer spectroscopy and found to be a mixture of pyrite and marcasite. The surface area of the pyrite concentrate, as determined by the BET nitrogen adsorption method, was l.9m2g-1’b. The hydrogenation experiments were carried out in an autoclave modified to permit rapid injection of coal into a preheated pressurized reactor. Experiments were carried

Coal liquefaction catalysis: R. M. Baldwin and S. Vinciguerra

all products were calculated based on a 100% recovery of the product slurry. RESULTS AND DISCUSSION 80-

60z

E .ac 403 P Y * 20-

, I

I , I I

o-

P

_

Figure 1

Conversion to THF-soluble

Figure 2

Preasphaltene yield

materials

Data for coal hydrogenation in the presence of added H,S, pyrite and H,S + pyrite are shown in Figures Z-4 along with baseline data for hydrogenation without the use of additives. Baseline runs were replicated to establish experimental reliability and one ‘memory effect’ run was performed between each run where pyrite was employed as an additive to check for changes in the thermal reactivity of the coal. What is perhaps most striking about these data is the influence of added H,S in the absence of added pyrite. The percentages of H,S refer to the mole y0 H,S in the gas atmosphere before heating and reaction (initial total pressure = 2.76 MPa at room temperature). A 56% increase in overall conversion is indicated at the 10 min residence time (2,5 and 10% H,S level), with a 21% increase found at 60 min when gaseous H,S alone (5 and 10% level) is added to the reaction gas atmosphere. The predominant influence on product distribution is in the preasphaltene fraction, especially at the short residence time. Clearly, H,S is acting as a catalyst for coal conversion at these conditions. Concentrations above 2%

Figure 3

out at 380°C in a 1O:l (weight ratio) excess of tetralin as the vehicle, 13.79MPa total pressure (at temperature). Two different reaction times were used, 10 and 60min, to test for rate effects due to the additives employed. Runs with pyrite added were made by injecting a coal:pyrite mixture (lOwt% pyrite) into the reactor. Coal and pyrite were ground to 100% <74pm to minimize diffusional mass transfer resistances in the system. The injection vessel was loaded with a paste consisting of 20 g coal/40g tetralin/2 g pyrite, while the balance of the tetralin vehicle was pre-heated in the reactor. At the end of a run, the reactor was quenched by rapid forced-convection cooling. Reaction product gases were analysed for hydrocarbons through C,, CO, CO,, H, and H,S and the liquid products were analysed by selective solvent fractionation @SF) to separate oils, asphaltenes, preasphaltenes and insoluble organic matter (IOM). Pentane, toluene and tetrahydrofuran were the solvents used in the SSF product workup. Inorganics in the liquefaction residue were analysed by Mossbauer spectroscopy to determine pyrite/pyrrhotite and yields of

Asphaltene yield

30 , -

0

Figur s 4

Oil yield

FUEL, 1983, Vol 62, May

499

Coal liquefaction catalysis: R. M. Baldwin and S. Vinciguerra Table f Analysis of coal sample’ Ultimate analysis (wt %. as received) 69.0 C H 5.3 N 1.7 S 0.5 Ash 5.6 17.8 0 (diff.)

Sulphur forms (wt %I Pyritic 0.05 Sulphitic nil Organic 0.45 ’ Penn State Reference Fuels; State: Colorado;

No.: PSOC-233; Rank: hvCb

Proximate analysis (wt %, as received) Moisture Volatile matter Ash Fixed carbon

5.8 36.9 5.6 51.6

Maceral distribution (vol %, dmmf) Vitrinite lnertinite Liptinite

88.8 6.5 4.7

Seam: Wadge; Mine: Energy

H,S in the initial gas phase mixture do not seem to appreciably increase the conversion at lOmin, but a substantial effect is present with increasing H,S concentration at 60 min. The function of H S in this case may be as a homogeneous catalyst. Rebick’ 3 has reported a catalytic effect of H,S on n-hexadecane pyrolysis, and attributed the effect to catalysis of hydrogen transfer. Since the early stages of coal liquefaction are thought to proceed via free radicals, a similar effect may be operative here. Free radicals formed rapidly by initial pyrolysis of the coal matrix could interact with H,S in the following manner: RI-R,+R;+R; R;+

H,S+R,

-H + HS

HS+R,-R,+R,-Rz.+H,S

HS-+Ri*+RiSH where R, -R, =coal macromolecule. Similar reactions could be written for radical R;. This mechanism also predicts that sulphur could become incorporated in the lower molecular weight products of reaction (RiSH). Preliminary analysis of SSF samples has shown a very small increase in total sulphur in the oil, asphaltene and preasphaltene fractions and a very large (factor of 2 to 3) increase in total sulphur in the THF-insolubles (IOM plus mineral matter). When pyrite alone (2 g) was employed as the additive, total conversion results were found to be comparable to the case with 2% H,S and no added pyrite. The increase in conversion at short residence time was again reflected most strongly in the preasphaltene fraction. Gas analyses on the reaction products after cool-down showed that H,S levels were about 0.03% when pyrite alone was the additive, and 0.3% when 2% H,S was the additive. No H,S was present in the reactor product gas for the baseline runs (no additives). The effect on conversion to THF-soluble materials noted upon addition of pyrite may thus be explained by two entirely different catalytic mechanisms: a heterogeneous catalytic mechanism involving catalysis promoted by reduced pyrite; or a homogeneous catalytic mechanism involving catalysis of hydrogen transfer by H,S released from the hydrogenated pyrite. Pyrite reduction to iron sulphide has previously been shown to be very rapid at the temperatures employed in this study’*. A synergistic effect seems to be present when both H,S and pyrite are added to the reactor, as evidenced by the

500

FUEL, 1983, Vol62,

May

significantly reduced preasphaltene content and enhanced oil yield as the H,S content is increased in the presence of added pyrite. This may reflect maintenance of the catalytically active iron sulphide species in the reactor or simply reflect increased hydrogen transfer catalysis by H,S. Although reduced iron pyrite has been purported by many researchers to be a catalyst responsible for enhanced reactivity of coal towards liquefaction, it is difficult to see how a substance with such a low surface area ( < 2 m* g- ‘) can function efficiently as a hydrogenation catalyst. A weak heterogeneous effect is indicated by the preasphaltene data at 60 min residence time for pyrite + H,S and the oil data for pyrite at 10 min. As shown in Figure 2, the yield of preasphaltenes decreases steadily with increasing H,S content in the presence of loo/, added pyrite. As shown in Figure 4, only pyrite alone is effective in increasing the yield of pentane-soluble oils at the short the reaction mass. Mossbauer spectroscopy was used to follow the iron sulphide stoichiometry in the liquefaction residues. Results of these analyses are shown in Table 2. Obviously, the final pyrite/pyrrhotite mixture present in the liquefaction residue is a very strong function of both residence time and H,S partial pressure. However, very little variation in the stoichiometry of the pyrrhotite was observed with a change in H,S partial pressure. Although the data in Table 2 are presented as pyritefpyrrhotite fractions, the Mossbauer spectra indicate that the nonmagnetic phase (reported as pyrite) is not comprised of pure pyrite (Fe&). The iron/sulphur stoichiometry of this non-magnetic phase cannot be determined with the precision of the magnetic phase. It is also possible that theinfluence of added H,S was to sulphide non-pyritic iron in the indigenous coal mineral matter. Ash and forms of sulphur analysis on the parent coal indicated that 50% of the total iron was present as iron pyrite (including marcasite), with the remainder of the iron being present in a non-sulphided state. Hydrogen sulphide in the reaction gas atmosphere would quickly sulphide any non-pyritic iron and thus generate additional quantities of the active catalyst. This hypothesis would explain the enhanced conversion found with H,S only added, as well as the large increase in total sulphur found in the THF-insolubles. Unfortunately, the small sample size of this fraction precluded analysis for forms of sulphur. Such information would aid in elucidating whether a homogeneous or heterogeneous catalytic effect is operative at these conditions.

Table 2

Mossbauer results

Residence time8 (min)

H,S content (%I

Non-magnetic phase, pyrite fraction f%)

Magnetic phase, pyrrhotite fraction (%J

Pyrrhotite X-value (FexS)

10 10 10 10 60 60 60 60

0 2 5 10 0 2 5 10

29 26 30 41 25 16 24 27

71 74 70 59 75 84 76 73

0.892 0.894 0.891 0.885 0.891 0.899 0.891 0.886

a All runs with 10% added pyrite

Coal liquefaction catalysis: R. tW. Baldwin and 23. Vinciguerra

ACKNOWLEDGEMENTS The authors would like to thank Dr D. W. Williamson of Colorado School of Mines, Physics Department, for the Mijssbauer analyses on coal and liquefaction residues. Appreciation is also extended to the Pittsburgh and Midway Coal Mining Company for providing the highpyrite coal and the DOE/Penn State coal sample bank for supplying characterization data. This work was carried out under DOE contract DE-AC22-79ETi4881, whose support is gratefully acknowledged.

6

8

9 10 11 12 13 14

REFERENCES 1 2 3 4 5

Wu, W. R. K. and Starch, H. H. ‘Hydrogenation of Coal and Tar’, US Bureau of Mines Bull., 1968, 633 Wright, C. H. and Severson, D. E. Am. Chem. Sot. Diu. Fuel Chem., Preprints 1972, 16(2), 68 Mukherjee, D. K. and Chowdhury, P. B. Fuel I976,55,4 Guin, J. A., Tarrer, A. R., Lee, J. M., Lo, L. and Curtis, C. W. Ind. Eng. Chem. Proc. Des. Deo. 1979, IS, 371 Guin, J. A., Tarter, A. R., Lee, J. M., Van Brackle, H. F. and

15

16

17 18

Curtis, C. W. Ind. Eng. Chem. Proc. Des. Deu. 1979, 18, 631 Guin, J. A., Lee, J. M., Fan, C. W., Curtis, C. W., Lloyd, J. L. and Tarrer, A. R. Ind. Eng. Chem. Proc. Des. Deu. 1980, 19,440 Guin, J. A., Tarrer, A. R., Prather, J. W., Johnson, D. R. and Lee, J. M. Ind. Eng. Chem. Proc. Des. Dev. 1978, 17, 118 Tarrer, A. R., Guin, J. A., Pitt% W. S., Henley, J. P., Prather, J. W. and Styles, G. A. Am. Chem. Sot. Div. Fuel Chem., Preprints 1976, 21, 59 Morooka, S. and Hamrin, C. E. Chem. Eng. Sci. 1977,32, 125 Granoff, B. and Thomas, M. G. Am. Chem. Sot. Div. Fuel Chem., Preprints 1977, 22, 183 Montana, P. A. and Granoff, B. Fuel 1980.59,214 Montano, P. A., Bommannaver, A. S. and Shah, V. Fuel 1981,60, 703 Montana, P. A., Vaishnava, P. P., King, J. A. and Eisentrout, E. N. Fuel 1981,60, 712 Attar, A. and Martin, J. B. Am. Chem. Sot. Div. Fuel Chem., Preprints 1981, 25, Ooo Stenberg, V. I. ‘Low RankCoal Liquefaction Employing H,S as a Homogeneous Catalyst’, presented at National AIChE Meeting, Orlando, March 1982 Eaton, W. J. ‘Disposable Additive Screening Study in Coal Liquefaction’, unpublished MS Thesis, Arthur Lakes Library, Colorado School of Mines, 1981 Rebick, C. Ind. Eng. Chem. Fundam. 1981,20, 54 Gertenbach, D. D., Baldwin, R. M., Bain, R. L., Gary, J. H. and Golden, J. 0. Ind. Eng. Chem. Proc. Des. Dee. 1979, 18, 102

FUEL, 1983, Vol 62, May

501