Energy & Fuels 1989,3, 127-131
127
Coprocessing of Coal and Residuum under Low-Severity Reaction Conditions: Effect of Basic Nitrogen Promotorst R. L. Miller,” G. F. Giacomelli, K. J. McHugh, and R. M. Baldwin Chemical Engineering and Petroleum Refining Department, Colorado School of Mines, Golden, Colorado 80401 Received August 2, 1988. Revised Manuscript Received November 3, 1988
Coprocessing batch reactor experiments have been carried out to evaluate the effectiveness of seven model nitrogen compounds in promoting coal conversion and distillate production under low-severity reaction conditions. Addition of dipropylamine and 1,2,3,4-tetrahydroquinolineto Cold Lake atmospheric residuum improved Kentucky No. 9 coal conversion by 53 and 41 wt %, respectively, with minimal nitrogen incorporation or thermal degradation losses. Reaction severity was found to strongly influence distillate yield and nitrogen incorporation but impacted coal conversion to a much lesser extent. This observation suggests the use of a low-severity coal dissolution step promoted by nitrogen compound addition followed by a catalytic upgrading step to maximize distillate yield.
Background The concept of simultaneouslyconverting both coal and non-coal-derived residual oil to distillate products has intrigued researchers for many years. This type of process, termed coprocessing, has several potential advantages over conventional direct liquefaction processes including conversion of two low-grade feeds to higher quality liquid products, reduction or elimination of recycle solvent requirements, and improved economics for smaller plant sizes. Furthermore, observed synergistic interactions between coal and heavy oil frequently result in distillate yields and coal conversion levels greater than expected from individually processing coal and heavy oil. However, coprocessing suffers from several technical problems that must be solved to improve economics and encourage commercial development. Most petroleum heavy oils are less aromatic than coal-derived liquids and, not surprisingly, are rather poor coal dissolution and hydrogen donor solvents. To compensate, prior solvolysis research by Yamada,’ Yan,2 and M o s ~ h o p e d i s and ~ * ~ coprocessing studies by HRI,5 UOP,6 Chevron,’ CANMET,8 Kerr-M~Gee,~ Lummus Crest,lo the Pittsburgh Energy Technology Center (PETC),” and CUrtis12have all relied upon severe reaction conditions and/or use of expensive heterogeneous catalysts to achieve sufficiently high levels of coal conversion. This practice causes increased hydrogen consumption, excessive cracking of distillate to gases, and accelerated catalyst fouling. Improving coal dissolution rates at lower severity reaction conditions where distillate product selectivity is maximized is central to coprocessing technology development. This approach requires developing and evaluating methods for promoting feed coal reactivity under lowseverity conditions. With the exception of a two-stage coprocessing scheme developed by Ignasiak13J4in which coal and bitumen are reacted noncatalytically in a low-severity first stage followed by upgrading in a catalytic hydrotreater, the studies described above have attempted to compensate for poor residual oil solvent quality by applying the “heat and beat” Presented at the Symposium on Coal-Derived FuelsCoprocessing, 195th National Meeting of the American Chemical Society and 3rd Chemical Congress of North America, Toronto, Ontario, Canada, June 5-10, 1988.
approach-the use of excessive reaction temperatures and pressures, long reaction times, and expensive heterogeneous catalysts. We feel this is an approach with limited potential and instead have focused on developing methods to efficiently coprocess coal and residual oil under lower severity conditions. Our efforts have been driven by many technical and economic incentives for reducing coprocessing severity, including the following: reduced hydrocarbon gas make resulting in reduced feed gas consumption, improved distillate selectivity, and enhanced hydrogen utilization efficiency; suppressed retrogression of primary coprocessing products resulting in enhanced distillate yields and residuum product quality; production of high-boiling residuum which is less refractory and thus more amenable to upgrading in a conventional catalytic hydrocracker; substitution of less expensive off-the-shelf ~
~~
(1)Yamada, Y.; Honda, H.; Oil, S.; Tsutaui, H. J . Fuel SOC. Jpn. 1974, 53, 1052. (2)Yan, T. Y.; Espenscheid, W. F. Fuel Process. Technol. 1983,7,121. (3)Moschopedis, S.E. Fuel 1980,59,67. (4)Moschopedis, S. E.Liq. Fuels Technol. 1984,2, 177. (5)Duddy, J. E.; Harris, E. C.; Smith, T. 0. Coal/Oil Co-Proceeaing Program-Status Report. In Proceedings of the DOE Direct Liquefaction Contractors’ Reuiew Meeting; PETC: Pittsburgh, PA, 1987;pp 448-474. (6)Gatais, J. G. Prepr. Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1986, 31(4),181. (7)Shinn, J. H.; Dahlberg, A. J.; Kuehler, C. W.; Rosenthal, J. W. The Chevron Co-Refining Process. In Proceedings of the Ninth EPRI Contractors’ Conference on Coal Liquefaction; EPRI: Palo Alto, CA, 1984; pp 33-1-33-15. (8)Fouda, S.A.; Kelly, 3. F.; Rahimi, P. M. Prepr. Pap.-Am. Chem. Soc., Diu. Fuel Chem. 1988,33(1),178. (9)Rhodes, D. E. Comparison of Coal and Bitumen-Coal Process Configurations. In Proceedings of the Tenth EPRI Contractors’ Conference on Clean Liquid and Solid Fuels; E P R I Palo Alto, CA, 1986, pp 8-1-8-21. (10)Greene, M.; Gupta, A,; Moon, W. Prepr. Pap.-Am. Chem. Soc., Diu.Fuel Chem. 1986,31(4),208. (11)Cugini, A. V.;Ruether, J. A.; Cillo, D. L.; Krastman, D.; Smith, D. N.; Balsone, V. Prepr. Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1988, 33(1),6. (12)Curtis, C. W.;Cassell, F. N. Energy Fuels 1988,2, 1-8. (13)Ignasiak, B.; Kovacik, G.; Ohuchi, T.; Lewkowicz, L.; du Plessis, M. P. Two-Stage Liquefaction of Subbituminous Alberta Coals in NonDonor Solvents. In Proceedings of the CANMET Coal Conuersion Contractors’ Reuiew Meeting; CANMET Ottawa, Canada, 1984;pp 385-395. (14)Lee, L. K.;Ignasiak, B. Prepr. Pap.-Am. Chem.SOC.,Diu. Fuel Chem. 1988,33(1),20.
0887-0624/89/2503-0127$01.50/00 1989 American Chemical Society
128 Energy & Fuels, Vol. 3, No. 2, 1989
Miller et al.
Table I. Ultimate Analysis (wt %, Dry Basis) of Feed Coals Kentuckv No. 9 Wvodak carbon 69.7 61.4 4.7 4.0 hydrogen nitrogen 1.4 1.0 4.2 0.7 sulfur oxygen (diff) 9.2 18.6 10.8 14.3 ash total
100.0
100.0
vessels, piping, and pumps in place of expensive, customdesigned units; less severe slurry-handling and materials-of-construction problems associated with lower operating temperatures and pressures. Conventional coal dissolution theories have held that coal must be heated to temperatures near 400-500 "C causing thermal rupture of various labile bonds within the three-dimensional cross-linked coal structure and formation of free-radical intermediates. Stabilization of the free radicals with hydrogen radicals results in products of lower molecular weight than the parent coal. However, work by several research groups, particularly PETC,15the North Dakota Energy Research Center,16Stanford Research Institute," Carbon Resources, Inc.,18 and the Colorado School of mine^'^*^ has demonstrated the ability to dissolve coal at much milder conditions than previously thought possible. Generally, we can conclude from this work that (1) if treated carefully, coals of all ranks are inherently very reactive, and (2) certain bonds in the coal structure can be broken or partially disrupted much more efficiently via selective chemical action than by thermolysis. In searching for effective coal dissolution promotors to be evaluated under low-severity reaction conditions, we concluded that on the basis of results from high-severity studies, basic nitrogen compounds warranted extensive study. Atherton and Kulik21*22 summarized data from several high-severity liquefaction studies using 1,2,3,4tetrahydroquinoline (THQ) in which THF coal conversions in the range of 85-100 w t 7% (maf basis) were obtained with Wyodak subbituminous and Illinois No. 6 bituminous coals at reaction temperatures of 400-450 "C. However, distillate yields from these experiments were much lower than expected, and nitrogen material balance measurements indicated nearly complete THQ incorporation into nondistillable products. Thus, while basic nitrogen compounds appear attractive as coal dissolution promotors, the incorporation problem limits their usefulness in high-severity liquefaction processes. We hypothesized that utilizing basic nitrogen compounds at lower severity could reduce chemical adduction and possibly physical entrapment of nitrogen compounds but that effectiveness toward promoting coal dissolution would remain high. This paper describes results of batch reactor experiments designed to evaluate the (15) Bockrath, B. C.; Illig, E. G. Prepr. Pap.-Am. Chem. Soc., Diu. Fuel Chem. 1984, 29(5), 76. (16) Sondreal,E. A.; Wilson, W. G.; Stenberg,V. I. Fuel 1982,61,925. (17) Ross, D. S.; Nguyen, Q. C.; Hum, G. P. Fuel 1984, 63, 1206. (18) Porter, C. R.; Kaesz, H. G. The ChemCoal Process for the Chemical Transformationof Low Rank Coal. In Proceedings of the Thirteenth Biennial Lignite Symposium; DOE: Grand Forks, ND,19%; pp 357-365. (19) Miller, R. L.; Baldwin, R.M. Prepr. Pap.-Am. Chem. Soc., Diu. Fuel Chem. 1986,31(4),152. (20) Miller, R. L. Liquefaction Coprocessing: A Study in Critical Vartables; EPRI Final Report for Project 2383-1, October 1987, EPRI: Palo Alto, CA, 1987. (21) Atherton, L. F.; Kulik, C. J. Advanced Coal Liquefaction. Paper presented at the Los Angeles AIChE Annual Meeting, November 1982. (22) Atherton, L. F.; Kulik, C. J. Coal Liquefaction Chemistry. Paper presented at the Anaheim AIChE National Meeting, May 1984.
Table 11. Properties of Cold Lake Atmospheric Residuum Distillate Fraction (wt %) water 0.0 177 "C0.4 177-260 "C 4.1 260-343 "C 4.1 10.9 343-454 " C 80.5 454 "C+ Ultimate Analysis (wt % Dry Basis) carbon 84.2 hydrogen 12.0 1.0 nitrogen sulfur 1.1 oxygen (diff) 1.7 ash 0.0 Solvent Solubility (wt %) tetrahydrofuran 100.0 to1uene 100.0 cyclohexane 100.0 Table 111. Nitrogen Compound Additives compound aqueous pKb code symbol 8.97 N1 THQ dipropylamine 3.00 N2 7,b-benzoquinoline 9.20 N3 piperidine 2.88 N4 4-piperidinopyridine 8.04 N5 5,6-benzoquinoline 10.05 N6 diphenylamine 13.23 N7 Table IV. Coprocessinn Reaction Conditions low high severity severity reaction temp, "C 344 440 feed gas CO/HzO,Hz H2 initial pressure, psig 850 1000 maximum pressure, psig 1500 2000 time at temp, min 30 30 coal/A-8/nitrogen compound feed ratio 1/1.5/0.5 1/1.5/0.5 ~~
use of homogeneous basic nitrogen compounds in promoting coal conversion and increasing distillate yields during low-severity coprocessing.
Experimental Section Kentucky No. 9 bituminous and Wyodak subbituminous coals were used as feed coals in batch reactor coprocessing experiments. Ultimate analysis data for these coals are presented in Table I. Coal samples were vacuum-dried to less than 1.0 w t % of moisture content and stored under a nitrogen blanket before use. Cold Lake atmospheric residuum was used as the coprocessing heavy oil. Available characterization data for this solvent are shown in Table 11. This heavy oil has been extensively utilizedz0 as a coprocessing feedstock and has performed well under low-severity reaction conditions. The seven model nitrogen compounds listed in Table I11 were evaluated as coal dissolution promotors. Each compound was ACS reagent grade and was used without further purification. These compounds were chosen to provide a wide range of base strength as measured by aqueous pKbvalues. For convenience, the coding system shown in Table III will occasionally be used to refer to each of the seven model nitrogen compounds. Runs were completed by using either commercial grade hydrogen or carbon monoxide. Each feed gas contained 1.0 mol % of krypton used as an inert tracer for material balance calculations. Distilled water in an amount equal to 50 w t % of the MAF feed coal was added to runs using carbon monoxide feed gas. Table IV lists the reaction conditions used in this study. A majority of the runs were completed under the low-severity conditions shown. However, for comparison purposes, several m were also completed under conventional high-severity conditions. Coprocessing experiments were performed by using a 300 cm3 Autoclave Magnedrive I1 batch reactor interfaced to an Apple IIe personal computer for temperature control. The computer
Energy & Fuels, Vol. 3, No. 2, 1989 129
Coprocessing of Coal a n d Residuum was also routinely used for temperature and pressure data acquisition during a run. Reactor heatup time to 344 "C reaction temperature was approximately 30 min. At the end of a run,water was circulated through cooling coils immersed in the reactor contents, allowing cooldown to temperatures below 200 "C in less than 20 minutes. After the reactor was cooled, gaseous products were recovered in evacuated stainless-steel vessels and analyzed for light hydrocarbons and carbon oxide gases by using a Carle Model l l l H gas chromatograph. The krypton tracer concentration was also measured and used as a tie element for determining product gas yields. Liquid-olid product slurry was removed from the reactor by using toluene as a wash solvent and quantitatively centrifuged to separate solids from the liquid product (termed "decant oil"). The residue (termed "centrifuge residue") contained unconverted coal and mineral matter coated with liquid product. Toluene added during slurry recovery and centrifugation was quantitatively removed during distillation. Decant oil samples were distilled to a 454 "C endpoint with an ASTM-type microdistillation apparatus. Portions of the centrifuge residue and 454 "C+ decant oil residuum were extracted in a Soxhlet extraction apparatus with cyclohexane, toluene, and tetrahydrofuran (THF). Standard elemental analyses for carbon, hydrogen, nitrogen, sulfur, and ash content were performed on liquid and solid product samples, and oxygen content was determined by difference. Detailed quantitative measurements of individual model nitrogen compounds in 454 "C- distillate samples were performed by using an HP 5890 capillary gas chromatograph interfaced to an H P 5790B quadrupole mass spectrometer. These data helped to provide a measure of nitrogen compound losses by thermal degradation or incorporation. Details of the experimental procedures used in this work have been r e p ~ r t e d . * ~ v ~ ~ These data were used to compute coal conversion to T H F soluble products, C4-454 "C distillate yield and the extent of nitrogen incorporation into residual products for each experiment by using eq 1-3. Coal conversion and distillate yield are reported on an maf coal basis, while the extent of nitrogen incorporation is computed as a percentage of nitrogen compound fed. In eq 3, the amount of incorporated nitrogen compound is estimated by measuring the difference in nitrogen contents for the coprocessing run and a "base case" run under the same reaction conditions but without nitrogen compound addition. This estimate is an approximation, since the extent of coal conversion varies from run to run. However, product nitrogen distribution is only a weak function of coal conversion level, and the approximation is generally a good one. coal conversion (wt 90 of maf coal fed) =
wf- wp x Wf
100 (1)
where Wf = weight of maf THF-insoluble coal charged and W = weight of maf THF-insoluble products recovered (not correctej for nitrogen incorporation). C4-454 "C distillate yield (wt % maf coal fed) = Wout
- Win
Wf
x 100 (2)
where Win = weight of C4-454 OC distillate charged and Wout = weight of C4-454 "C distillate products recovered. extent of nitrogen incorporation (wt % of nitrogen compound fed) =
Nout
- Nbaee
x 100 (3)
Nfeed
where Nout= weight of nitrogen contained in nondistillable product fractions (THF insolubles and 454 "C+ T H F solubles), Nbase= weight of nitrogen contained in nondistillable product fractions (THF insolubles and 454 "C+ T H F solubles) in base case experiment without nitrogen compound addition, and Nfd (23) Giacomelli, G. F. Low Severity Coal/Oil Co-Processing Using Model Nitrogen Compounds. M.S. Thesis T-3447, Colorado School of Mines, Golden, CO, July 1987. (24) McHugh, K. J. Exploratory Studies in Enhanced Low Severity Coal Liquefaction. M.S. Thesis T-3528, Colorado School of Mines, Golden, CO, May 1988.
Table V. Coal Conversion, Distillate Yield, and Incorporation Results from Low-Severity Coprocessing of Kentucky No. 9 Coal Using Cold Lake Residuum and Nitrogen Compound Additives" nitrogen coal distillate nitrogen, compound conversion, yield, incorporation,b added wt% wt% wt% none 16.4 7.4 0.0 THQ
dipropylamine 7,8-benzoquinoline piperidine 4-piperidinopyridine 5,6-benzoquinoline diphenylamine
57.4
69.5 44.3
15.1 24.2
3.1 2.4
0.1
6.0
31.2 17.4 47.7
-26.8 -33.9 -7.4
15.5 19.4 7.1
30.9
-35.8
8.3
OReaction Conditions: Low severity, C O / H 2 0 feed gas. bThis includes both physical entrapment and chemical adduction. = weight of nitrogen contained in nitrogen compound fed to the reactor. Data reproducibility was within A2.1 w t % for coal conversion, 13.2 w t % for distillate yield, and A0.9 wt % for nitrogen incorporation with use of the experimental techniques described above.
Results and Discussion Effect of Model Nitrogen Compound Addition. Low-severity coprocessing experiments using K e n t u c k y No. 9 coal and CO/H20 as reducing agent were completed to evaluate the seven model nitrogen compounds listed i n Table I11 as coal dissolution promotors. Results from these r u n s are summarized in T a b l e V. Only 16.4 wt % of the coal was converted u n d e r low-severity conditions i n the absence of a n y dissolution promotors, and distillate production was correspondingly low. Addition of 1,2,3,4tetrahydroquinoline (THQ) and dipropylamine improved coal conversion to 57.4% and 69.5%, respectively. Lesser improvement was noted for 7,8-benzoquinoline, piperidine, 5,6-benzoquinoline, and diphenylamine, while addition of 4-piperidinopyridine d i d not affect coal conversion. Distillate yield results also varied widely from 24.2 and 15.1 wt % for dipropylamine and THQ addition, respectively, to large negative yields i n the range of -26.8 to -35.6 wt % for piperidine, 4-piperidinopyridine, and diphenylamine. Nitrogen compound adduction was f o u n d to be a b o u t 3 w t % or less for THQ and dipropylamine but greater than 15 wt % for piperidine and 4-piperidinopyridine. Taken together, these data show that a wide range of process behavior can be expected for different homogeneous nitrogen compound additives. Both THQ and dipropylamine gave significant conversion and distillate yield enhancement with little incorporation into residual products. Cronauer= found that THQ was a much better coal liquefaction solvent than quinoline or other non-hydrogen-donor nitrogen compounds, largely due to THQ's ability to donate hydrogen d u r i n g dissolution. Tagaya26 and KazimiZ7reported that primary alkyl amines such as n-butylamine enhanced coal liquefaction, primarily b y ether cleavage and hydroxyl substitution reactions. O u r results support these observations and also show that other nitrogen compound classes s u c h as benzoquinolines and piperidines (including 4-piperidinopyridine) are much less effective. Conversion and yield results from Table V are plotted in Figures 1 and 2 as functions of the nitrogen compound (25) McNeil, R. I.; Young, D. C.; Cronauer, D. C. Fuel 1983,62,806. (26) Tagaya, H.; Sugai, J.; Onuki, M.; Chiba, K. Energy Fuels 1987,
1 , 397. (27) Kazimi, F.; Chen, W. Y.; Chen, J. K.; Whitney, R. R.; Zimny, B. Prepr. Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1985, 30(4),402.
130 Energy & Fuels, Vol. 3, No. 2, 1989
Miller et al. 0 Cool Conversion
I
bBc
e;rs
lncorporotion
n
L o w Extent of Incorporotion
NZ
8-
20-
0
1
High Extent of Incorporation
0
0
"
0
"
" " 2
"
" 4
"
'
6
8
10
12
14 Nl/LS
pKb of Nitrogen Compound Additive
Figure 1. Effect of nitrogen compound aqueous pKb on coal conversion from low-severitycoprocessing of Kentucky No. 9 coal with CO feed gas.
1
.-
VI
N2
VI 0
fQ z
E + L 0 ._ >
20
I
L o w Extent of Incorporation
h
-
0 -
p1
c - -20 -
High Extent of Incorporation N4
.-z0
N7
a
-40 2
4
6
8
10
12
14
pKb of Nitrogen Compound Additive
Figure 2. Effect of nitrogen compound aqueous pKb on distillate yield from low-severity coprocessing of Kentucky No. 9 coal with CO feed gas.
aqueous PKb value. In each plot, the data can be arranged into one group for which a low to moderate (e10 wt %) extent of nitrogen incorporation was observed, and a second group for which a high ( > l o wt % ) extent of nitrogen incorporation was observed. These groupings are only semiquantitative in nature and should not be considered a predictive tool for assessing performance of compounds other than the seven shown. Nevertheless, the data clearly show that coal conversion is enhanced with increasing nitrogen compound base strength (decreasing PKb value). This result indicates the importance of acid site attack via nitrogen substitution reactions with hydroxyl functionalities within the coal structure during low-severity dissolution.2e As shown in Figure 2, distillate yield also increased with decreasing nitrogen compound PKb value. This result reflects the observation reported by Millerz0and others that increased coal conversion during coprocessing increases distillate production from the residual oil via a free-radical attack mechanism (one possible explanation for any "positive synergism" seen during coprocessing) in addition to any distillate that may be coal-derived. Thus, addition of nitrogen compounds such as THQ or dipropylamine indirectly improve distillate yields by enhancing the rate and extent of coal conversion. Of course, any nitrogen compound incorporation into nondistillable products will have a direct negative impact on observed distillate yields. The extent of nitrogen incorporation showed only a general positive trend with PKb value. This result was not unexpected since no simple parameter such as PKb should be able to describe the complex and still poorly understood
Nl/HS
NZ/LS
NZ/HS
Nitrogen Compound/Reoction Severity
Figure 3. Effect of THQ ( N l ) or dipropylamine (N2) addition and reaction severity on coprocessing of Kentucky No. 9 coal with hydrogen feed gas (LS = low reaction severity, HS = high reaction
severity).
physical and chemical interactions between nitrogen species and coal. Effect of Reaction Severity. Both physical entrapment and chemical adduction (including hydrogen bonding and C-C and C-N bond formation) can contribute to nitrogen compound incorporation into residual coprocessing products. We hypothesized that lowering the reaction temperature would reduce or possibly eliminate the fraction of nitrogen compounds chemically adducted and thus increase the likelihood of recovering and recycling nitrogen compound additives in a continuous coprocessing scheme. To test this hypothesis, a series of experiments was completed using Kentucky No. 9 bituminous coal, Cold Lake residuum, and hydrogen gas under low- and high-severity reaction conditions. Both THQ and dipropylamine were used as nitrogen compound additives. Results from these runs are summarized in Figure 3. Increased reaction severity caused a coal conversion increase of approximately 22 wt % in the THQ (Nl) runs but only about a 5 wt % increase in the dipropylamine (N2) runs. As expected, distillate yield was even more sensitive to reaction severity with yield increases of about 30 and 53 wt % for the THQ and dipropylamine runs, respectively. Much of this increase can be attributed to additional cracking of the Cold Lake residuum at high reaction severity. Taken alone, these conversion and yield data seem to favor high-severity coprocessing as the more attractive processing option. However, as shown in Figure 3, reaction severity also strongly influences the fate of nitrogen compound additives during coprocessing. THQ incorporation increased nearly 13-fold (from 2.9 to 36.9 wt %), while dipropylamine incorporation increased nearly 6 times (from 2.7 to 15.4 wt %) with probable contributions from both physical entrapment and chemical adduction. Studies are presently under way to study the effect of reaction severity on each of these mechanisms for nitrogen compound loss. In addition, about 5 wt 70of the THQ and nearly 80 wt % of the dipropylamine initially fed to the reactor was cracked to ammonia and hydrocarbon gases under high-severity conditions but no cracking losses were noted for low reaction severity. In the absence of these losses, high-severity distillate yields would be even larger than shown in Figure 3, but little THQ or dipropylamine would be available for recycle in a continuous process. In our opinion, a more promising scenario involves achieving a high level of coal dissolution under low-severity conditions by using a nitrogen promotor such as THQ or dipropylamine (or a process-derived nitrogen-rich stream) where incorporation and degradation losses are minimized
Energy & Fuels, Vol. 3, No. 2, 1989 131
Coprocessing of Coal and Residuum 0 Cool Conversion
E 100
[29 Distillate Yield
'E54 lncorporotion
I
n WY/H~
W/CO
Feed Cool/Feed Gas
Figure 4. Effect of feed coal and feed gas on low-severity coprocessing with dipropylamine addition (Ky = Kentucky No. 9 coal, Wy = Wyodak coal).
followed by a more conventional second-stage catalytic hydrocracker to process the residual first-stage coprocessing products. On the basis of previous studies in our laboratory, this residuum should be very reactive and more easily converted to distillate products.20 This process scheme is currently being evaluated in our laboratory, and results will be reported when they become available. Effect of Coal Rank and Feed Gas Composition. Low-severity coprocessing runs were completed to study the effects of coal rank (subbituminousvs bituminous) and feed gas (H2 vs CO/H,O) on the performance of dipropylamine as a coal dissolution promotor. Results of these experiments are summarized in Figure 4. In the Kentucky No. 9 coal runs, use of hydrogen gas slightly improved coal conversion, but distillate yield and extent of nitrogen incorporation were unaffected. Conversely, a significant improvement in both conversion and distillate yield were observed with Wyodak coal when CO/H20 was used, a result attributable to favorable water gas shift chemistry with high oxygen content coals.ls Nitrogen incorporation in the Wyodak runs was about double that of
the Kentucky runs, once again suggesting the importance of interactions between nitrogen species and acidic oxygen sites in the coal. The same trends were noted in a similar set of experiments using THQ as nitrogen compound additive.
Conclusions Results from low-severity coprocessing experimentsshow that selected nitrogen model compound additives enhance coal conversion and distillate yield without significant thermal degradation or incorporation into residual products. THQ and dipropylamine addition improved Kentucky No. 9 coal conversion by 41 and 53 w t % (maf basis), respectively, with nitrogen compound incorporation of about 3 wt % or less. Decreased reaction severity significantly lowered the extent of nitrogen incorporation, presumably by reducing both chemical adduction reactions and physical entrapment. As expected, distillate yields under low-severity conditions were much lower than high-severity yields, but coal conversion was affected to a much lesser extent. This trend suggests the possibility of developing a two-stage coprocessing scheme in which a low-severity first stage using nitrogen compound coal dissolution promotors is followed by a more traditional catalytic upgrading second stage to improve distillate yields. Selective fractionation and recycle of the nitrogen promotors would also be an integral part of the process. Dipropylamine or THQ addition improved coal conversion and distillate yields with both Kentucky NJ. 9 bituminous and Wyodak subbituminous coals. However, nitrogen incorporation was more extensive with the higher oxygen content Wyodak coal. Acknowledgment. We acknowledge the financial support of the US.Department of Energy under Grant NO. DE-FG22-86PC90909. Registry No. THQ, 635-46-1; dipropylamine, 142-84-7; 7,8benzoquinoline, 230-27-3; piperidine, 110-89-4; 4-piperidinopyridine, 2767-90-0;5,6-benzoquinoline,85-02-9;diphenylamine, 122-39-4.