- Email: [email protected]
Liquefaction of bituminous coal at moderate temperatures
Liquefaction temperatures Raymund
of bituminous
P. Skowronski
coal at moderate
and Laszlo A. Heredy
Rockwell international, Rocketdyne Division, 6633 Canoga Avenue, CA 9 7303, USA (Received 14 July 7986; revised 23 February 1987)
Canoga
Park,
Coal hydrogenation was investigated in the temperature range 275 to 325°C in order to minimize the number of thermal side reactions that take place. Gas-phase hydrogen was used in batch experiments without an added donor solvent, to avoid the additional analytical complexities introduced by such a solvent. It was found that significant oil yields (up to 72% of the daf coal) can be obtained from the hydrogenation of bituminous coal at 325°C. Furthermore, at this temperature, the data indicate that cleavage of certain C-O bonds may have an important role in oil formation. The metal surfaces of the liner and impeller of the autoclave had a strong catalytic effect on the liquefaction reactions under these conditions. The oil yield was 48 % when the metallic surfaces were exposed and only 19 % when these components were coated with glass. Catalysis by nickel, applied as nickel acetate impregnated into the coal, gave higher overall conversion, lower oil yield, and a more saturated oil product than catalysis by the autoclave surfaces. (Keywords: bituminous coal; liquefaction; moderate temperatures; catalysis)
There are several incentives for conducting coal liquefaction at moderate temperatures. Better product mixes can be expected from operating at lower temperatures, because such conditions favour liquids formation and the reduction of gas byproducts and coke formation. Materials requirements would be less severe and thermal efficiency probably could be improved. Research at moderate temperatures is relevant to subsequent scale-up. In addition, such research is valuable for investigating the chemistry of coal liquefaction, because at moderate temperatures much of the carbon skeleton of the coal is expected to survive with little change and to be recovered in the liquid products. Several investigations have been conducted recently on coal liquefaction at moderate temperatures’. Hydrogen donor solvents were used in most of these studies. This experimental technique, however, produces complications when investigation of liquefaction chemistry is the main objective. The donor solvent interferes with investigation of the reaction chemistry and analysis of the reaction products by forming adducts with the coalderived intermediates. Furthermore, it is difficult to characterize separately the donor solvent and its derivatives in the complex product mixture. At moderate reaction temperatures, internal hydrogen rearrangement may play a relatively role in stabilization. more significant radical Consequently, hydrogenating coal with gas-phase hydrogen and without a donor solvent at moderate temperature offers significant advantages in understanding the mechanisms that occur. Hydrogenation of coal without using a donor solvent has been investigated at the US Bureau of Mines’. In that work, however, the liquefaction reactions were conducted at 400 to 5OO”C, where side reactions leading to gas and coke formation are quite extensive. The objective of our research was to explore hydrogenation without using a CO16-2361/87/121642-04$3.00 0 1987 Butterworth & Co. (Publishers) Ltd.
1642
FUEL, 1987, Vol 66, December
donor solvent at temperatures reactions are minimal.
where
the thermal
side
EXPERIMENTAL The hydrogenation experiments were conducted in a 1 litre stirred autoclave equipped with a liner and an anchor-shaped impeller, which was positioned as close to the liner surface as possible. The use of this impeller improves gas-solid contact and mass transfer and maintains effective stirring during the initial softening of the coal particles when caking can otherwise occur. In each experiment, 25g of high volatile bituminous coal was used. The coal was ground to - 200 mesh and dried at 115°C in vucuo before use. A liner made of Inconel 600 and an impeller made of 316 SS were used in Experiments 1A through 1E and in Experiment 2. The liner and impeller were coated with glass for Experiments 3 and 4 to eliminate the catalytic effect of the metal surface. The liner and impeller were coated by West Coast Porcelain. A slurry of finely ground glass was sprayed onto each, and the glass subsequently was melted at 925°C. Two sets of hydrogenation experiments were conducted using two bituminous coals of slightly different rank. The objective of the first set was to investigate reactions in the 275 to 325°C range using 1 h and 48 h reaction times. The objective of the second set was to explore the effect of coal composition (rank) and two coal conversion oil catalysts on and product characteristics using a 325°C reaction temperature and a 48 h reaction time. The first set of experiments was conducted using a Pittsburgh seam bituminous coal of the following composition (wt %, daf): C, 80.1; H, 5.1; S, 3.6; N, 1.6; and 0, 9.6. The dry coal yielded 11.2 wt % ash. A Pittsburgh seam bituminous coal of higher C and lower 0
Liquefaction Table 1
Test conditions
of bituminous
and product
coal at moderate
distributions
Product distribution (wt %, daf)
Experimentb
Temp. (“C)
Time (h)
Gas
THF soluble
Product appearance
Untreated coal 1A fB (Dz) 1c 1D 1E
212 271 310 325 325
7.0 1.0 1.0 1.0 48.0
0.4 0.3 1.0 1.7 5.8
5 8 15 22 29 77’
Dry powder Dry powder Dry powder Black solid Shiny black solid Thick dark oil
4 Loss-free basis b All experiments were conducted at approximately 20 MPa c Most of this THF-soluble product was hexane-soluble oil (72 wt % based on daf coal)
content was used in the second set of experiments. The elemental composition of this coal (wt %, daf) was as follows:C,82.1;H,5.9;S,3.3;N, 1.6;and0,7.1.Theash yield of the dry coal was 8.2 wt %. Each experiment was conducted at a final hydrogen pressure of approximately 20 MPa. The product gases were measured and analysed by gas chromatography. The hydrogenated product was Soxhlet-extracted with benzene. (In Experiment 2, a sizeable fraction of the oil distilled from the liner during the reaction time and was collected as condensate outside of the liner.) The benzene solution was concentrated by distillation and then poured into excess hexane to precipitate the asphaltenes. After filtration, the hexane solution was distilled to recover the oil fraction. The benzene-insoluble residue was Soxhlet-extracted with THF and separated into a THF-soluble fraction and an insoluble residue. This procedure was based on a separation method recommended by Mima et ~1.~. The solvent-separated fractions were analysed for elemental composition. Furthermore, a detailed characterization of the product oils was made using simulated distillation, vapour phase osmometry (v.P.o.), field ionization mass spectrometry (f.i.m.s.), and proton nuclear magnetic resonance (n.m.r.) spectrometry. A portion of the product from Experiment 1B was combusted, and the resulting water was analysed by mass spectrometry (Shrader Analytical Labs, Inc.) to determine its protium and deuterium contents. RESULTS
AND
DISCUSSION
The test conditions and product yields of the first set of experiments are presented in Table 1. Very little reaction was observed by visual inspection or chemical analysis in a 1 h reaction period at the lowest test temperature (275°C). The appearance of the dry powder product very closely resembled that of the starting coal. Most of the gas products consisted of carbon oxides; the gaseous hydrocarbon (CH,) yield was only 0.04 wt %. The THFsoluble extract increased from 5 to 8 wt %. An experiment made using D, under the same conditions (lB), however, showed that a significant Hexchange had taken place at 275°C. About 15% of the protium (‘H) in the coal was replaced by deuterium (‘H or D). In this experiment, a sizable increase in the THF
temperatures:
R. P. Skowronski
and L. A. Heredy
solubility also was observed. This increased solubility may be significant. However, additional experiments with deuterium are needed to reach a firm conclusion regarding a possible isotope effect. In the experiments conducted at 310 and 325°C there were significant increases in the THF solubility. Finally, when the reaction time was increased to 48 h at 325°C the THF solubility of the product increased to 77 wt %. Hexane-soluble oil constituted 72 wt % of the recovered organic material (Table 2). This result demonstrates that higher temperatures (i.e. >325”C) are not necessary to obtain high oil yields. The yield of the gas products was less than 5 wt % of the total soluble products in the first three experiments, and increased only slightly to about 7 wt y0 as the temperature was increased from 271 to 325°C and the reaction time from 1 h to 48 h. This indicates that the product liquids were stable under these experimental conditions. The product yields from the second set of experiments (2 through 4) are listed in Table 2. These experiments were designed to test the effect of coal composition (rank) and two selected catalytic conditions on coal conversion and product oil characteristics4. Table 2 also includes the product yields from the lower rank coal (Experiment 1E) for comparison. The product yields from Experiments 1E and 2 show that the lower rank coal used in Experiment 1E was much more reactive, particularly with regard to the conversion of the asphaltenes to oil. The evaluation of Experiments 2 and 3 shows that the autoclave’s metal surfaces have a strong catalytic effect on oil production. The results of Experiment 4 indicate that nickel very effectively catalyses the conversion of coal to soluble products but is somewhat less effective than the autoclave surfaces in catalysing oil production. Gas formation was remarkably low in all four experiments. It appears that gas formation is proportional to oil production; the gas yield was z 10 wt % of the oil yield in each experiment. An example of a detailed gas composition is presented in Table 3. The elemental composition of the oil products is given in Table 4. The part of the liquid that distilled out of the liner in Experiment 2 is shown in Table 4 as ‘distilled liquid’. The H/C atomic ratios ranged from 1.13 to 1.24,
Table 2 liquefaction
Product distributions from experiments” (wt %. daf)b
Experiment Coal rank (% C, daf) (% 0, daf) Catalyst
Conversion Product distributions Gas Oil Asphaltene Preasphaltene Residue
1E
moderate
temperature
2
3
4
82 7.1
82 7.1
82 7.1
Yes (liner/ (impeller)
No
Yes (nickelacetate)
83
87
80
94
6 72 5
4 48 18 17 13
3 19 33 25 20
4 40 28 22 6
80 9.6 Yes (liner/ impeller)
’ All hydrogenation experiments 48 h reaction time * Loss-free basis
made at 325°C
- 20 MPa, and using
FUEL, 1987, Vol 66, December
1643
Liquefaction
of bituminous
coal at moderate
temperatures:
with the oil from the nickel-catalysed reaction having the highest value. The oil produced from lower rank coal (Experiment IE) had the same H/C ratio as the oil product from the higher ranking coal (Experiment 2), even though the oil yield in Experiment 1E was 50% higher. The oil Experiment 1E had a significantly lower oxygen content than the oil from Experiment 2. This difference may indicate that the cleavage of certain C-O bonds has an important role in oil formation. The hydrogen type distributions of the oil products were determined by proton n.m.r. spectroscopy. Figure 1 shows the spectrum of the oil from Experiment 1E as an example. The integration results are presented in Table 5. In agreement with the elemental analysis data, the n.m.r. integration values show that the oil made by nickel catalysis (Experiment 4) had significantly more saturated structures than the oil formed via the catalytic effect ofthe
Table 3
Product
gas composition” Yield ‘(wt oA coal)
Gas component CH, C,H, C,H, C,Hs i-C,H
1.76 Trace 1.23 1.05 0.06 0.70 0.23 0.29 0.23 0.23 0.00
1,,
r&H,, i-C,H,, n-CsH,, co CO, H,Sb
’ Experiment 1E b H,S probably reacted with autoclave the end of the run
Elemental
autoclave surfaces (Experiment 2). These results plus the fact that the oil yield was greater in Experiment 2 than in Experiment 4 indicate that the catalytic effect of the autoclave surfaces preferentially catalysed hydrocracking reactions, leading to oil formation, while nickel catalysis resulted in more hydrogen uptake, leading to the formation of saturated structures. F.i.m.s. spectra of the oils from Experiments 2 through 4 were recorded. Profiles of these product oils have a roughly Gaussian shape. The more intense peaks were located in the range M/Z = 100 to 500. Strong peaks were noted in all spectra at M/Z = 178,192,206,220,234, and higher. This series of values occurs every 14 mass units and corresponds to anthracenelphenanthrene and their higher molecular weight alkylated homologues. The series M/Z = 154, 168, 182, 196, and higher was present and corresponds to diphenyl/acenaphthene and their higher homologues. The 154 peak, however, is relatively weak. Peaks at 202, 216, 230, 244, and higher were present and correspond to pyrene and its alkylated homologues. V.p.0. was used to determine the number-average molecular weights of the oils from Experiments 2 through 4. The molecular weights ranged from about 250 to 500. The number-average molecular weights when determined by f.i.m.s. ranged from about 340 to 400. A similar difference between such molecular weights obtained by v.p.0. and f.i.m.s. has been observed by others’. The oil products from Experiments 2 through 4 were
compositions
wall and thus was not present at
of product
liquids (wt %)
Sample
C
H
N
0
S
Total
H/C
Experiment lE, oil Experiment 2, distilled liquid Experiment 2, oil Experiment 3, oil Experiment 4, oil
88.2
8.4
1.1
1.8
0.6
100.1
1.14
86.3 86.1 87.2 86.5
8.1 8.2 8.6 8.9
1.2 0.9 0.8 0.9
3.6 4.2 2.7 2.7
0.4 0.4 0.7 0.6
99.6 99.8 100.0 99.6
1.13 1.14 1.18 1.24
Proton-n.m.r.
integration
10
6
6
Figure
1
Proton
n.m.r. spectrum
Hydrogen
Condensed aromatic Aromatic single ring + phenolic OH a-Aliphatic P-Aliphatic y-Aliphatic ‘Includes
1644
2.3 y0 contribution
SHIFT (6)
of oil fraction
from Experiment
1E
values
Chemical type
2
4
CHEMICAL
Experiment
Hydrogen
1
T
Composition
Table 5
and L. A. Heredy
5.78
Total
Table 4
R. P. Skowronski
number
type distribution
(wt %)
shift 3 (Oil)
4 (Oil)
13.6
9.2
7.8
13.8” 30.6 33.2 8.8
9.6 33.4 36.9 10.9
8.5 29.1 39.3 15.3
(ppm)
1E (Oil)
2 (Liquid)
2 (Oil)
7.2-9.0
14.1
12.9
6.3-7.2 2.e4.2 l&2.0 0.551 .o
6.1 35.7 35.4 8.7
11.9 32.8 34.4 8.0
from 4.2 to 5.2 ppm region
FUEL, 1987, Vol 66, December
Liquefaction Table 6
Moderate
temperature
of bituminous
U~YS~Stwo-stage
coal at moderate
liquefaction-comparison
temperatures:
of hydrogen
R. P. Skowronski
and L. A. Heredy
types Hydrogen
distribution
(wt %) Aliphatics
Oil product
from
liquefaction” Moderate temperature Two-stage liquefaction* Extract’ Process solvent’
Yield (wt %)
Condensed aromatic
12
14
41 48
27 26
Single ring aromatic
ci
B
Y
6
36
35
9
12 10
28 24
29 37
4 3
’ Pittsburgh seam coal, C: 80.1 wt % b Illinois No. 6 coal, C: 79.2 wt % (from Ref. 6) c 454°C distillate
further characterized by simulated distillation. These chromatograms had the overall appearance expected for coal liquids. In each, there were large, discrete peaks as well as a broad envelope. Each oil had an undetermined amount of residue, since material was still coming off the column at each endpoint. The endpoint was about 550°C based on aromatic standards. The hydrogen type distributions of an oil fraction obtained by moderate temperature hydrogenation and of distilled oil fractions of extract and process solvent samples from an integrated two-stage liquefaction pilot plant are compared in Table 6. In accordance with thermodynamic considerations, the oil from moderate temperature hydrogenation is much less aromatic and contains more saturated structures than the pilot plant products, which were produced at significantly higher temperatures.
SUMMARY
by nickel, applied as nickel acetate (4) Catalysis impregnated into the coal, gave significantly different results from those obtained in the metal-lined reactor. Overall conversion to soluble products was higher using the nickel catalyst (94 versus 87 wt %). However, nickel catalysis gave lower conversion to oil (40 uersus 48 wt %). oil produced in the nickelcatalysed (5) The hydrogenation was much more aliphatic in character than the product from the metal-surfacecatalysed reactions; the atiphatic to aromatic proton ratios were 5.1 and 2.7, respectively. Also, the nickelcatalysed product had a significantly higher (p +~)/a aliphatic proton ratio than the metal-catalysed product (1.9 uersus 1.4). These data indicate that nickel catalysis strongly promoted hydrogenation of aromatic rings, while the metal components more effectively catalysed hydrocracking reactions, which resulted in oil formation.
AND CONCLUSIONS
Liquefaction of two bituminous coals was explored to obtain information on the chemistry of liquefaction. Tests were conducted in the temperature range 275 to 325°C and at approximately 20 MPa hydrogen pressure. No donor solvent was added to the reactant to simplify product analysis and the evaluation of the results. The results and conclusions are summarized below.
(1) High conversion
to oil (up to 72 ~1%) can be obtained at 325°C with few side reactions resulting in gas or coke formation. At this temperature, cleavage of certain C-0 bonds may have an important role in oil formation. (2) Of the two bituminous coals tested, the lower rank coal (C: 80 wt %) was much more reactive and gave 50% higher oil yield than the higher rank coal (C: 82 wt %). The oil yields were 72 and 48 wt %, respectively. It would be of interest to conduct similar moderate temperature experiments at pressures lower than 20 MPa. (3) It was shown in experiments made with the higher rank coal (C: 82 wt ‘A) that the metal components of the autoclave (liner and impeller) had a strong catalytic effect on the liquefaction reaction. When the metal parts of the autoclave were replaced with a glass-coated liner and impeller, the oil yield was reduced from 48 to 19 wt %.
ACKNOWLEDGEMENTS The authors wish to thank Dr Kwang E. Chung of the Rockwell International Science Center for his assistance with the autoclave experiments and Dr Kwang E. Chung and Mr Joseph J. Ratto for recording the proton n.m.r. spectra. In addition, the field ionization mass spectrometry work done by Dr Ripudaman Malhotra of SRI International is gratefully acknowledged. This research was supported by the Electric Power Research Institute, the US Department of Energy, and Rockwell International Corporation. REFERENCES Stansberry, P. G., Lin, R., Lee, C. W., Terrer, M. T., Derbyshire, F. I. and Davis, A. Am. Chem. Sot. Div. Fuel Chem., Prep. 1986,31(4), 111 and references therein Hawk, C. 0. and Hitashue, R. W. US Bureau of Mines, Bull. 1963, No. 622 Mima, M. J., Schultz, H. and McKinstry, W. E. PERC/RI-76/6, September 1976 Heredy, L. A. and Skowronski, R. P. Proceedings of the Symposium on the Chemistry of Coal Liquefaction and Catalysis, Hokkaido University, Sapporo 060, Japan, 17-20 March 1985 Collins, C. J., Triolo, R. and Lietzke, M. M. Fuel 1984, 63, 1202 Neuworth, M. B., Heredy, L. A., McCoy, L. R., Skowronski, R. P. and Ratto, J. J. ‘Chemistry of the Extractive Phase of Two-Stage Coal Liquefaction’, Final Report to US DOE, WP-8SWOO113, February 1985, pp. 20, 34
FUEL, 1987, Vol 66, December
1645
of bituminous
P. Skowronski
coal at moderate
and Laszlo A. Heredy
Rockwell international, Rocketdyne Division, 6633 Canoga Avenue, CA 9 7303, USA (Received 14 July 7986; revised 23 February 1987)
Canoga
Park,
Coal hydrogenation was investigated in the temperature range 275 to 325°C in order to minimize the number of thermal side reactions that take place. Gas-phase hydrogen was used in batch experiments without an added donor solvent, to avoid the additional analytical complexities introduced by such a solvent. It was found that significant oil yields (up to 72% of the daf coal) can be obtained from the hydrogenation of bituminous coal at 325°C. Furthermore, at this temperature, the data indicate that cleavage of certain C-O bonds may have an important role in oil formation. The metal surfaces of the liner and impeller of the autoclave had a strong catalytic effect on the liquefaction reactions under these conditions. The oil yield was 48 % when the metallic surfaces were exposed and only 19 % when these components were coated with glass. Catalysis by nickel, applied as nickel acetate impregnated into the coal, gave higher overall conversion, lower oil yield, and a more saturated oil product than catalysis by the autoclave surfaces. (Keywords: bituminous coal; liquefaction; moderate temperatures; catalysis)
There are several incentives for conducting coal liquefaction at moderate temperatures. Better product mixes can be expected from operating at lower temperatures, because such conditions favour liquids formation and the reduction of gas byproducts and coke formation. Materials requirements would be less severe and thermal efficiency probably could be improved. Research at moderate temperatures is relevant to subsequent scale-up. In addition, such research is valuable for investigating the chemistry of coal liquefaction, because at moderate temperatures much of the carbon skeleton of the coal is expected to survive with little change and to be recovered in the liquid products. Several investigations have been conducted recently on coal liquefaction at moderate temperatures’. Hydrogen donor solvents were used in most of these studies. This experimental technique, however, produces complications when investigation of liquefaction chemistry is the main objective. The donor solvent interferes with investigation of the reaction chemistry and analysis of the reaction products by forming adducts with the coalderived intermediates. Furthermore, it is difficult to characterize separately the donor solvent and its derivatives in the complex product mixture. At moderate reaction temperatures, internal hydrogen rearrangement may play a relatively role in stabilization. more significant radical Consequently, hydrogenating coal with gas-phase hydrogen and without a donor solvent at moderate temperature offers significant advantages in understanding the mechanisms that occur. Hydrogenation of coal without using a donor solvent has been investigated at the US Bureau of Mines’. In that work, however, the liquefaction reactions were conducted at 400 to 5OO”C, where side reactions leading to gas and coke formation are quite extensive. The objective of our research was to explore hydrogenation without using a CO16-2361/87/121642-04$3.00 0 1987 Butterworth & Co. (Publishers) Ltd.
1642
FUEL, 1987, Vol 66, December
donor solvent at temperatures reactions are minimal.
where
the thermal
side
EXPERIMENTAL The hydrogenation experiments were conducted in a 1 litre stirred autoclave equipped with a liner and an anchor-shaped impeller, which was positioned as close to the liner surface as possible. The use of this impeller improves gas-solid contact and mass transfer and maintains effective stirring during the initial softening of the coal particles when caking can otherwise occur. In each experiment, 25g of high volatile bituminous coal was used. The coal was ground to - 200 mesh and dried at 115°C in vucuo before use. A liner made of Inconel 600 and an impeller made of 316 SS were used in Experiments 1A through 1E and in Experiment 2. The liner and impeller were coated with glass for Experiments 3 and 4 to eliminate the catalytic effect of the metal surface. The liner and impeller were coated by West Coast Porcelain. A slurry of finely ground glass was sprayed onto each, and the glass subsequently was melted at 925°C. Two sets of hydrogenation experiments were conducted using two bituminous coals of slightly different rank. The objective of the first set was to investigate reactions in the 275 to 325°C range using 1 h and 48 h reaction times. The objective of the second set was to explore the effect of coal composition (rank) and two coal conversion oil catalysts on and product characteristics using a 325°C reaction temperature and a 48 h reaction time. The first set of experiments was conducted using a Pittsburgh seam bituminous coal of the following composition (wt %, daf): C, 80.1; H, 5.1; S, 3.6; N, 1.6; and 0, 9.6. The dry coal yielded 11.2 wt % ash. A Pittsburgh seam bituminous coal of higher C and lower 0
Liquefaction Table 1
Test conditions
of bituminous
and product
coal at moderate
distributions
Product distribution (wt %, daf)
Experimentb
Temp. (“C)
Time (h)
Gas
THF soluble
Product appearance
Untreated coal 1A fB (Dz) 1c 1D 1E
212 271 310 325 325
7.0 1.0 1.0 1.0 48.0
0.4 0.3 1.0 1.7 5.8
5 8 15 22 29 77’
Dry powder Dry powder Dry powder Black solid Shiny black solid Thick dark oil
4 Loss-free basis b All experiments were conducted at approximately 20 MPa c Most of this THF-soluble product was hexane-soluble oil (72 wt % based on daf coal)
content was used in the second set of experiments. The elemental composition of this coal (wt %, daf) was as follows:C,82.1;H,5.9;S,3.3;N, 1.6;and0,7.1.Theash yield of the dry coal was 8.2 wt %. Each experiment was conducted at a final hydrogen pressure of approximately 20 MPa. The product gases were measured and analysed by gas chromatography. The hydrogenated product was Soxhlet-extracted with benzene. (In Experiment 2, a sizeable fraction of the oil distilled from the liner during the reaction time and was collected as condensate outside of the liner.) The benzene solution was concentrated by distillation and then poured into excess hexane to precipitate the asphaltenes. After filtration, the hexane solution was distilled to recover the oil fraction. The benzene-insoluble residue was Soxhlet-extracted with THF and separated into a THF-soluble fraction and an insoluble residue. This procedure was based on a separation method recommended by Mima et ~1.~. The solvent-separated fractions were analysed for elemental composition. Furthermore, a detailed characterization of the product oils was made using simulated distillation, vapour phase osmometry (v.P.o.), field ionization mass spectrometry (f.i.m.s.), and proton nuclear magnetic resonance (n.m.r.) spectrometry. A portion of the product from Experiment 1B was combusted, and the resulting water was analysed by mass spectrometry (Shrader Analytical Labs, Inc.) to determine its protium and deuterium contents. RESULTS
AND
DISCUSSION
The test conditions and product yields of the first set of experiments are presented in Table 1. Very little reaction was observed by visual inspection or chemical analysis in a 1 h reaction period at the lowest test temperature (275°C). The appearance of the dry powder product very closely resembled that of the starting coal. Most of the gas products consisted of carbon oxides; the gaseous hydrocarbon (CH,) yield was only 0.04 wt %. The THFsoluble extract increased from 5 to 8 wt %. An experiment made using D, under the same conditions (lB), however, showed that a significant Hexchange had taken place at 275°C. About 15% of the protium (‘H) in the coal was replaced by deuterium (‘H or D). In this experiment, a sizable increase in the THF
temperatures:
R. P. Skowronski
and L. A. Heredy
solubility also was observed. This increased solubility may be significant. However, additional experiments with deuterium are needed to reach a firm conclusion regarding a possible isotope effect. In the experiments conducted at 310 and 325°C there were significant increases in the THF solubility. Finally, when the reaction time was increased to 48 h at 325°C the THF solubility of the product increased to 77 wt %. Hexane-soluble oil constituted 72 wt % of the recovered organic material (Table 2). This result demonstrates that higher temperatures (i.e. >325”C) are not necessary to obtain high oil yields. The yield of the gas products was less than 5 wt % of the total soluble products in the first three experiments, and increased only slightly to about 7 wt y0 as the temperature was increased from 271 to 325°C and the reaction time from 1 h to 48 h. This indicates that the product liquids were stable under these experimental conditions. The product yields from the second set of experiments (2 through 4) are listed in Table 2. These experiments were designed to test the effect of coal composition (rank) and two selected catalytic conditions on coal conversion and product oil characteristics4. Table 2 also includes the product yields from the lower rank coal (Experiment 1E) for comparison. The product yields from Experiments 1E and 2 show that the lower rank coal used in Experiment 1E was much more reactive, particularly with regard to the conversion of the asphaltenes to oil. The evaluation of Experiments 2 and 3 shows that the autoclave’s metal surfaces have a strong catalytic effect on oil production. The results of Experiment 4 indicate that nickel very effectively catalyses the conversion of coal to soluble products but is somewhat less effective than the autoclave surfaces in catalysing oil production. Gas formation was remarkably low in all four experiments. It appears that gas formation is proportional to oil production; the gas yield was z 10 wt % of the oil yield in each experiment. An example of a detailed gas composition is presented in Table 3. The elemental composition of the oil products is given in Table 4. The part of the liquid that distilled out of the liner in Experiment 2 is shown in Table 4 as ‘distilled liquid’. The H/C atomic ratios ranged from 1.13 to 1.24,
Table 2 liquefaction
Product distributions from experiments” (wt %. daf)b
Experiment Coal rank (% C, daf) (% 0, daf) Catalyst
Conversion Product distributions Gas Oil Asphaltene Preasphaltene Residue
1E
moderate
temperature
2
3
4
82 7.1
82 7.1
82 7.1
Yes (liner/ (impeller)
No
Yes (nickelacetate)
83
87
80
94
6 72 5
4 48 18 17 13
3 19 33 25 20
4 40 28 22 6
80 9.6 Yes (liner/ impeller)
’ All hydrogenation experiments 48 h reaction time * Loss-free basis
made at 325°C
- 20 MPa, and using
FUEL, 1987, Vol 66, December
1643
Liquefaction
of bituminous
coal at moderate
temperatures:
with the oil from the nickel-catalysed reaction having the highest value. The oil produced from lower rank coal (Experiment IE) had the same H/C ratio as the oil product from the higher ranking coal (Experiment 2), even though the oil yield in Experiment 1E was 50% higher. The oil Experiment 1E had a significantly lower oxygen content than the oil from Experiment 2. This difference may indicate that the cleavage of certain C-O bonds has an important role in oil formation. The hydrogen type distributions of the oil products were determined by proton n.m.r. spectroscopy. Figure 1 shows the spectrum of the oil from Experiment 1E as an example. The integration results are presented in Table 5. In agreement with the elemental analysis data, the n.m.r. integration values show that the oil made by nickel catalysis (Experiment 4) had significantly more saturated structures than the oil formed via the catalytic effect ofthe
Table 3
Product
gas composition” Yield ‘(wt oA coal)
Gas component CH, C,H, C,H, C,Hs i-C,H
1.76 Trace 1.23 1.05 0.06 0.70 0.23 0.29 0.23 0.23 0.00
1,,
r&H,, i-C,H,, n-CsH,, co CO, H,Sb
’ Experiment 1E b H,S probably reacted with autoclave the end of the run
Elemental
autoclave surfaces (Experiment 2). These results plus the fact that the oil yield was greater in Experiment 2 than in Experiment 4 indicate that the catalytic effect of the autoclave surfaces preferentially catalysed hydrocracking reactions, leading to oil formation, while nickel catalysis resulted in more hydrogen uptake, leading to the formation of saturated structures. F.i.m.s. spectra of the oils from Experiments 2 through 4 were recorded. Profiles of these product oils have a roughly Gaussian shape. The more intense peaks were located in the range M/Z = 100 to 500. Strong peaks were noted in all spectra at M/Z = 178,192,206,220,234, and higher. This series of values occurs every 14 mass units and corresponds to anthracenelphenanthrene and their higher molecular weight alkylated homologues. The series M/Z = 154, 168, 182, 196, and higher was present and corresponds to diphenyl/acenaphthene and their higher homologues. The 154 peak, however, is relatively weak. Peaks at 202, 216, 230, 244, and higher were present and correspond to pyrene and its alkylated homologues. V.p.0. was used to determine the number-average molecular weights of the oils from Experiments 2 through 4. The molecular weights ranged from about 250 to 500. The number-average molecular weights when determined by f.i.m.s. ranged from about 340 to 400. A similar difference between such molecular weights obtained by v.p.0. and f.i.m.s. has been observed by others’. The oil products from Experiments 2 through 4 were
compositions
wall and thus was not present at
of product
liquids (wt %)
Sample
C
H
N
0
S
Total
H/C
Experiment lE, oil Experiment 2, distilled liquid Experiment 2, oil Experiment 3, oil Experiment 4, oil
88.2
8.4
1.1
1.8
0.6
100.1
1.14
86.3 86.1 87.2 86.5
8.1 8.2 8.6 8.9
1.2 0.9 0.8 0.9
3.6 4.2 2.7 2.7
0.4 0.4 0.7 0.6
99.6 99.8 100.0 99.6
1.13 1.14 1.18 1.24
Proton-n.m.r.
integration
10
6
6
Figure
1
Proton
n.m.r. spectrum
Hydrogen
Condensed aromatic Aromatic single ring + phenolic OH a-Aliphatic P-Aliphatic y-Aliphatic ‘Includes
1644
2.3 y0 contribution
SHIFT (6)
of oil fraction
from Experiment
1E
values
Chemical type
2
4
CHEMICAL
Experiment
Hydrogen
1
T
Composition
Table 5
and L. A. Heredy
5.78
Total
Table 4
R. P. Skowronski
number
type distribution
(wt %)
shift 3 (Oil)
4 (Oil)
13.6
9.2
7.8
13.8” 30.6 33.2 8.8
9.6 33.4 36.9 10.9
8.5 29.1 39.3 15.3
(ppm)
1E (Oil)
2 (Liquid)
2 (Oil)
7.2-9.0
14.1
12.9
6.3-7.2 2.e4.2 l&2.0 0.551 .o
6.1 35.7 35.4 8.7
11.9 32.8 34.4 8.0
from 4.2 to 5.2 ppm region
FUEL, 1987, Vol 66, December
Liquefaction Table 6
Moderate
temperature
of bituminous
U~YS~Stwo-stage
coal at moderate
liquefaction-comparison
temperatures:
of hydrogen
R. P. Skowronski
and L. A. Heredy
types Hydrogen
distribution
(wt %) Aliphatics
Oil product
from
liquefaction” Moderate temperature Two-stage liquefaction* Extract’ Process solvent’
Yield (wt %)
Condensed aromatic
12
14
41 48
27 26
Single ring aromatic
ci
B
Y
6
36
35
9
12 10
28 24
29 37
4 3
’ Pittsburgh seam coal, C: 80.1 wt % b Illinois No. 6 coal, C: 79.2 wt % (from Ref. 6) c 454°C distillate
further characterized by simulated distillation. These chromatograms had the overall appearance expected for coal liquids. In each, there were large, discrete peaks as well as a broad envelope. Each oil had an undetermined amount of residue, since material was still coming off the column at each endpoint. The endpoint was about 550°C based on aromatic standards. The hydrogen type distributions of an oil fraction obtained by moderate temperature hydrogenation and of distilled oil fractions of extract and process solvent samples from an integrated two-stage liquefaction pilot plant are compared in Table 6. In accordance with thermodynamic considerations, the oil from moderate temperature hydrogenation is much less aromatic and contains more saturated structures than the pilot plant products, which were produced at significantly higher temperatures.
SUMMARY
by nickel, applied as nickel acetate (4) Catalysis impregnated into the coal, gave significantly different results from those obtained in the metal-lined reactor. Overall conversion to soluble products was higher using the nickel catalyst (94 versus 87 wt %). However, nickel catalysis gave lower conversion to oil (40 uersus 48 wt %). oil produced in the nickelcatalysed (5) The hydrogenation was much more aliphatic in character than the product from the metal-surfacecatalysed reactions; the atiphatic to aromatic proton ratios were 5.1 and 2.7, respectively. Also, the nickelcatalysed product had a significantly higher (p +~)/a aliphatic proton ratio than the metal-catalysed product (1.9 uersus 1.4). These data indicate that nickel catalysis strongly promoted hydrogenation of aromatic rings, while the metal components more effectively catalysed hydrocracking reactions, which resulted in oil formation.
AND CONCLUSIONS
Liquefaction of two bituminous coals was explored to obtain information on the chemistry of liquefaction. Tests were conducted in the temperature range 275 to 325°C and at approximately 20 MPa hydrogen pressure. No donor solvent was added to the reactant to simplify product analysis and the evaluation of the results. The results and conclusions are summarized below.
(1) High conversion
to oil (up to 72 ~1%) can be obtained at 325°C with few side reactions resulting in gas or coke formation. At this temperature, cleavage of certain C-0 bonds may have an important role in oil formation. (2) Of the two bituminous coals tested, the lower rank coal (C: 80 wt %) was much more reactive and gave 50% higher oil yield than the higher rank coal (C: 82 wt %). The oil yields were 72 and 48 wt %, respectively. It would be of interest to conduct similar moderate temperature experiments at pressures lower than 20 MPa. (3) It was shown in experiments made with the higher rank coal (C: 82 wt ‘A) that the metal components of the autoclave (liner and impeller) had a strong catalytic effect on the liquefaction reaction. When the metal parts of the autoclave were replaced with a glass-coated liner and impeller, the oil yield was reduced from 48 to 19 wt %.
ACKNOWLEDGEMENTS The authors wish to thank Dr Kwang E. Chung of the Rockwell International Science Center for his assistance with the autoclave experiments and Dr Kwang E. Chung and Mr Joseph J. Ratto for recording the proton n.m.r. spectra. In addition, the field ionization mass spectrometry work done by Dr Ripudaman Malhotra of SRI International is gratefully acknowledged. This research was supported by the Electric Power Research Institute, the US Department of Energy, and Rockwell International Corporation. REFERENCES Stansberry, P. G., Lin, R., Lee, C. W., Terrer, M. T., Derbyshire, F. I. and Davis, A. Am. Chem. Sot. Div. Fuel Chem., Prep. 1986,31(4), 111 and references therein Hawk, C. 0. and Hitashue, R. W. US Bureau of Mines, Bull. 1963, No. 622 Mima, M. J., Schultz, H. and McKinstry, W. E. PERC/RI-76/6, September 1976 Heredy, L. A. and Skowronski, R. P. Proceedings of the Symposium on the Chemistry of Coal Liquefaction and Catalysis, Hokkaido University, Sapporo 060, Japan, 17-20 March 1985 Collins, C. J., Triolo, R. and Lietzke, M. M. Fuel 1984, 63, 1202 Neuworth, M. B., Heredy, L. A., McCoy, L. R., Skowronski, R. P. and Ratto, J. J. ‘Chemistry of the Extractive Phase of Two-Stage Coal Liquefaction’, Final Report to US DOE, WP-8SWOO113, February 1985, pp. 20, 34
FUEL, 1987, Vol 66, December
1645