Conclusions
Delayed coking appears promising as a method of increasing the commercial value of lignite pitch. The technique yields three products: coke, oil, and gas. The coke is potentially valuable as an aggregate for metallurgical electrodes, other graphite products, and a low-sulfur fuel. The oil, upon distillation, is a valuable chemical intermediate. Gas resulting from the coking of the pitch, following stripping to remove the CB compounds, is a possible substitute for natural gas or a source of hydrogen by the steam-reforming process. Ethylene could also be recovered from the gas stream and used as a raw material.
Berber, J. S. Rice, R. L., Fortney, D. R., IND. ENG. RES. DEVELOP. 6,197 (1967a). CHEM.PROD. Berber, J. S., Rice, R. L., Hiser, A. L., Wainwright, H. W., Bur. Mines. Rept. Invest. 6916, 17 pp., 1967b. Dell, M. B., Ind. Eng. Chem. 51, 1297 (1959). Donnelly, F. J., Barbour, L. T., Hydrocarbon Process. Petrol. Refiner. 4 5 , 2 2 1 (1966). Martin, S.W., Wills, L. E., in “Advances in Petroleum Chemistry and Refining,” Vol. 11, J. J. McKetta, Jr., and K. A. Kobe, Eds., pp. 367-87, Interscience., Xew York, 1959. RECEIVED for review April 1, 1968 ACCEPTED August 13,1968
Literature Cited
Berber, J. S., Rahfuse, R. V., Wainwright, H. W., IND. ENG.CHEM.PROD. RES. DEVELOP. 4 , 2 4 2 (1965).
Division of Fuel Chemistry, 155th Meeting, ACS, San Francisco, Calif., April 1968.
REFINING OF PYROLYTIC SHALE OIL D.
P .
MONTGOMERY
Phillips Petroleum Company, Bartlesuille, Okla. 74003
A pyrolytic shale oil was refined into marketable products to determine the severity of hydrodenitrogenation needed for sustained refining and the yields and qualities 6f the refined products. The refining steps were: delayed coking, with recycle of heavy coker distillate; hydrodenitrogenation of coker distillate, followed by fractionation; further hydrodenitrogenation of naphtha; reforming of nitrogen-free naphtha; catalytic cracking of gas oil from the hydrodenitrogenated coker distillate. Temperature and pressure maximums of 950’ F. and 1500 p.s.i.g. were well within the scope of current refinery practice. Nitrogen content of the reformer feedstock was 1.5 p.p.m., and that of the cracking stock was 100 p.p.m. (basic); thus, both were adequately free of nitrogen to ensure long catalyst life. Reformates and catalytic gasolines of 98 octane number were produced. The hydrodenitrogenated coker distillate yielded a large volume of specification grade diesel fuel. Yields and product distributions for each refining step were determined, along with properties of the finished liquids.
BECAUSE of revived interest by both industry and government in the production of petroleum products from oil shale, a pilot-scale study of the refining of pyrolytic shale oil was made by Phillips Petroleum Co. This study used current refining techniques and incorporated no novel or unproved refining steps. Only commercially available catalysts were employed. The question to be answered was not whether shale oil can be refined to a synthetic crude and then to finished products, but what refining severity is required to reduce 274
I&EC PRODUCT RESEARCH A N D DEVELOPMENT
the nitrogen to a satisfactory level for continuous processing in a conventional refinery, and what yields and qualities of the products result. Domestic oil shale locations and the problems associated with recovering oil from these shales have been presented (Oil Gas J , 1964), and recent refining developments were discussed (Chem Eng Progr., 1966). Pyrolytic shale oils are characteristically high in concentrations of nitrogen, sulfur, and oxygen; concentrations are typically 2, 1, and 1%,respectively. Because nitrogen is a potent poison for
modern, acidic type refining catalysts, its removal is of paramount importance. The oils are black, waxy, and foul smelling, with high pour points and viscosities, and high concentrations of olefins and diolefins make them unstable. A detailed description of shale oil is given by Hellwig et al. (1962). Facilities for refining may be nonexistent a t shale retort locations, since known oil shales are located in remote areas that are mountainous, arid, and cold. For this study the most economic sequence of refining steps was postulated to be recycle coking and hydrostabilization a t the retort, followed by pipeline delivery to a refinery near a marketing area, where the oil would be hydrodenitrogenated ( H D N ) in one or more passes and fractionated, and distillates requiring further nitrogen reduction would be separately processed. Naphtha, a distillate of a boiling range that encompasses alkylated naphthenes and does not exceed the end point of gasoline (180" to 400" F . ) , would be reformed to increase octane number by dehydrogenation of naphthenes to aromatics and isomerization of paraffins. Heavier distillates would be catalytically cracked to products of lower boiling ranges. I n pilot plant operations we found that one-pass H D N of the hydrostabilized coker distillate was sufficient for distillates heavier than naphtha. In fact, the 400" to 650" F. distillate was a satisfactory diesel fuel without additional processing. Both 400+" and 650+" F . distillates were catalytically cracked. (A second-pass H D N of the hydrostabilized coker distillate was carried out, but the fractions were not further processed.) Figure 1 is a block diagram of the processing scheme. R a w , Pyrolytic Shale
Oil
Feedstock
The shale oil used as raw feed was obtained by pyrolysis of a Colorado oil shale (Table I ) . Determination of the consumption of hydrogen in many of the refining steps is important. Since this determination was based upon a balance of the hydrogen contained in feeds and products, unusual accuracy was required in analyses of the liquids processed. This was attained by comparing the nuclear magnetic resonance spectra of protons in feedstocks and products and relating these
Figure 1. Process scheme
spectra to those of pure hydrocarbons of approximately the same hydrogen content. A statistical study of this procedure by Silas (1967) has shown that hydrogen content can be determined to +0.055, a t a 9 5 L confidence ~ level. Pilot Plant and Procedure
Coking and Hydrostabilization. Raw shale oil was pumped through a preheater into the bottom of the coking vessel, made from a 20-inch length of 8-inch diameter pipe and closed with weld caps. The coker was vertically mounted in a two-section furnace, with each section separately controlled. Within the coker was a concentric thermowell for determining the temperature a t any height. Heaters were set so that incoming feed and outgoing distillate were in the range 800" to 850" F. Coker distillate passed through a heated line to the flash still; temperature in the flash zone was 665°F. Product gases were cooled in an external condenser and the condensate was metered, sampled, and collected. A regulator ahead of the gas meter held the still pressure a t 10 to 12 p.s.i.g. and set coker pressure. Distillate was condensed and flowed into a water knockout. from where
Table I. Coking Conditions: 14 p.s.i.g., 810' F., 0.208 vol.":coker vo1.-hr.
Yields, 5' Hydrogen Hydrogen sulfide Methane Ethylene Ethane Propylene Propane Butylenes Isobutane n-Butane Pentane and heavier Coke
Rau Shale Oil (R.S.O.) Wt. Vol. 0.09 0.11 1.39 0.42 1.16 0.67 0.87 0.67 0.14 0.44 81.43 12.61 100.00
... ... ... ... ...
...
1.o 0.2
0.7 85.5
... 87.4
Properties
R.S.O.b API gravity Pour point Vis., 100" F., cs. Vis., 210" F., cs. Aniline point Mol. wt. Br. no. Carbon res.. wt. Composition, wt. C C
H S 0 N
19.5 +75.
Coker dist.'
Coke
3.08
26.9 +40. 5.05 1.55 83.6 252 45.3 0.4"
84.6 10.84 0.61
84.3 11.30 0.54
89.5 3.4 0.32
1.5 2.2
1.2 2.0
2.34 4.5 0.55
... ... ...
...
...
Ash Coke drum filled in 3%hours. ' N o free water u,as observed in the m u shale oil. 'Not stabilized. From another run
VOL. 7 NO. 4 DECEMBER 1968
275
it was pumped to the hydrostabilizer. Bottoms were continuously recycled to the coker. Coker distillate was mixed with metered hydrogen, preheated to 450" F., and fed to the hydrostabilizer. Reactants passed down-flow through a 1-gallon reactor filled with presulfided catalyst, Filtrol 475-8. A temperature increase of 30°F. was noted as reactants passed through the bed. Effluent was cooled and the gas and liquid phases were separated in the chamber of a liquid-level controller. Pressure was maintained a t 500 p.s.i.g. by action of a regulator on gas from the separator (75°F.). This gas was sampled and metered, as was the gas evolved from the liquid discharged from the separator. The liquid product was sampled and held for H D N operations. Hydrodenitrogenation. The composited, hydrostabilized liquid was mixed with 5200 standard cubic feet of hydrogen per barrel, preheated to 750" F., and passed downflow over the bed of presulfided Harshaw 4303-E catalyst, heated by a five-zone furnace. A sharp increase in temperature of 50" F. occurred a t the bed top. The remainder of the reactor was virtually isothermal a t 812°F. and 1500 p.s.i.g. The space velocity was 1.0 liquid volume per hour. T o prevent effluent lines from plugging with deposits of ammonium sulfides, water was injected into the effluent a t a rate of 50 ml. per hour. High pressure gas from the liquid-gas separator (ambient temperature) was sampled and metered, and similar measurements were made on the gas desorbed when pressure on the liquid was reduced to atmospheric. As 10-gallon portions of liquid product accumulated, each portion was fractionated into gas (butanes and lighter), 60" to 180°F. and 180" to 400°F. distillates, and 400+" F. bottoms. The 180" to 400+" F. distillate was feed for naphtha H D N and reforming, and the 400+" F. bottoms were catalytically cracked. Corresponding distillates and bottoms from each distillation were composited and analyzed. A 10-gallon portion of the 400+"F. bottoms was further fractionated a t 20-mm. of pressure to yield a 400" to 650°F. diesel fuel and 650+"F. bottoms for catalytic cracking. Hydrodenitrogenation of Saphtha. The 180" to 400" F. distillate obtained by fractionation of the initially hydrodenitrogenated oil comprised the feedstock naphtha. Other than the use of a fresh portion of catalyst (presulfided), and a bed temperature of 700"F., the apparatus and procedure were the same as in the initial H D N operation. Since little nitrogen and sulfur remained in the naphtha feedstock water was not injected into the reactor effluent. Reforming of Nitrogen-Free Naphtha. The total liquid product from naphtha hydrodenitrogenation was used as feedstock without stabilization (fractionation to remove hydrocarbons lighter than pentanes). A 300-ml. bed of fresh U.O.P. R-5 catalyst was used, and the reactants were preheated in a tube furnace to reaction temperature. All tests were carried out a t 500 p.s.i.g., with nominal values of 8 to 1 for hydrogen to naphtha (molar) and 3.0 for liquid hourly space velocity. Reactor effluent was cooled by exchange with tap water and charged to a gas-liquid separator a t 75" F. Samples of reformates were caught a t 500 p.s.i.g. and stabilized to yield butane-free reformates and gases composed of butanes and lighter components. Tests were performed in the chronological sequence 2, 3, 4, 5, 6, and 1 of Table VII. Cracking of Gas Oils. Both 400+" and 650+" F. oil from the first-pass H D N were catalytically cracked. Properties 276
I&EC PRODUCT RESEARCH A N D DEVELOPMENT
of these oils are given in Table IV. The catalyst was an equilibrium catalyst withdrawn from a refinery FCC (fluid catalytic cracking) reactor and was composed of 75% Filtrol-800 and 25% of a mixture of other catalysts. The confined bed cracking apparatus was described by Johnson and Stark (1953). Measured quantities of steam and oil were pumped into the catalyst bed, and noncondensed products were metered and analyzed by gasliquid chromatography (GLC). Coke deposits were determined by analysis of the regeneration gases. After fractionation of condensed products, fractionation gases were analyzed by GLC. Cracking was carried out a t 10 p.s.i.g., a nominal steamoil ratio of 15 pounds per barrel, and cycle times to accomplish a value of 1.0 weight % for the mean carbon content of the catalyst. Space velocities were varied to determine the WHSV conversion relationship, as shown in Figure 2 . The 650+"F. oil was cracked a t 906°F.; the more refractory 400+"F. oil was cracked a t 914°F. Results a n d Discussion
Coking and Hydrostabilization. Table I contains data characterizing the raw shale oil, conditions of coking, and yields and properties for products of the coker. The porous, friable coke was easily removed from the coker. Hydrostabilizer yields and product properties are given in Table 11.
The reported data are for 32 hours of operation. Recovered gases plus hydrostabilized liquid equalled 100.8 weight % of the 214.5 pounds of raw shale oil fed. Catalyst coke was not determined. Charging raw shale oil to the flash still rather than to the coker would reduce the volume of coker feed. Also, shale oil components of coker distillate boiling range would not undergo degradation in the coker. However, this mode of operation could not be practiced because an insoluble, organic material separated from the flashstill bottoms and plugged transfer lines. The nature of this insoluble matter was not determined. Coker distillate with a higher end point would be possible, but this might make hydrodenitrogenation more difficult. Over-all yields would probably be little affected by increasing coker distillate end point. The coker distillate was immediately hydrostabilized in the pilot plant to insure that its properties would not change radically while awaiting further processing. Commercially, this step may not be required because retained samples of coker
I
/
I
Figure 2. FCC conversion v s . reciprocal space velocity
Table II. Hydrostabilization
Conditions: average of 470" F. (4550 inlet, 485" outlet), 500 p.s.i.g., 0.89LHSV, 1150 standard cu. ft. hydrogem bbl. coker distillate
Yields
5; Coher Distillate Vol.
wt. Hydrogen Water Hydrogen sulfide Methane Ethane Propylene Propane Butylenes Isobutane n-Butane Pentanes and heavier Total
-0.71" 0.07 0.02 0.03 0.01 0.01 0.02 0.01 0.01 0.01 100.52 100.71
0.1
101.7 101.8
5; R.S.O.
wt.
Properties o f C;Liquid
Vol.
29.0
API gravity Pour point, ' F. Vis. 100" F. cs. Vis. 210" I?. cs. Aniline point, O F. Br. no. Composition C H
-0.58 0.06 0.02 0.02 0.01 0.01 0.02 0.01 0.01 0.01 81.8
i t40.
4.51 1.46 94.7 37.2 84.3 11.93 0.49 0.83 2.0
S 0
86.8
S
Distillation icorr. to 1 atm.) 120' F. 330 412 476 516 561 608 655 708 769 900
IBP l0Cc 20CC
30'c 40'; 50'c 60' c
ioc;
goLC 9OCC EP
Hydrogen consumption 414 standard cu ft bbl of feed
distillate did not show any obvious changes, such as precipitates or increased viscosity, after 6 months of storage in glass. Elimination of hydrostabilization, however, may require a higher severity operation in the subsequent hydrodenitrogenation step and would certainly increase hydrogen consumption during H D N . If coking and hydrostabilization take place a t the site of the retort, three economies may be possible. First, unsold coke could be mixed with raw shale charged to the retort so that coke combustion could provide heat otherwise derived by burning of shale, and thus prevent loss of oil precursors in the shale. Second, steam reforming of coker (flash) gas could provide hydrogen in excess of that needed for hydrostabilization. Third, additional heat could be obtained by adding air to the retort gases, followed by burning of carbon monoxide and hydrocarbons in the retort gas. Hydrodenitrogenation of Liquid Product from Hydrostabilization. Yields and properties of the various fractions from H D N are given in Tables 111 and IV. A material balance was calculated for 6.75-hour period, during which products exceeded hydrocarbon feed by 4.1 weight %. Hydrogen consumption and yields of hydrogen sulfide, ammonia, and water were calculated from feed and product contents of hydrogen, sulfur, nitrogen, and oxygen, respectively. Carbon on the catalyst was not determined until the second pass of H D N operation was completed (not reported in detail because of its ineffectiveness). At that time carbon content of the catalyst was 2.2%. N o decline in catalyst activity was observed. The value of this H D N process can be seen in the high qualities of the product fractions, particularly the diesel fuel (400" to 650°F.). N o further processing would
be required to market the diesel fuel, and its volume was 41% of the raw shale oil. Either the 400+" or 650+" F. bottoms could be charged to a catalytic cracker although it may prove desirable to market the 400" to 650°F. as diesel fuel. The light gasoline (Cj,-180"K.) from coker distillate H D N was of a modest octane number (82.2 RON + 3 ) , but the small volume of this fraction would not upset the pooled octane number of a refinery. The conditions for this hydrodenitrogenation process were suggested from unpublished results of Kertamus
Table 111. Hydrodenitrogenation and Fractionation of Hydrostabilized Coker Distillate
Conditions: 1500 p.s.i.g., 812" F., 1.002 LHSV, 5200 standard cu. ft. hydrogen! bbl. of feed
Yields i; R.S.O.
' i Feed
wt. Hydrogenn Water Hydrogen sulfide Ammonia Methane Ethane Propane Isobutane n-Butane Ci-1803F. 180'-400" F. 400-650" F . 650+0 F , Total
-3.41 0.92 0.31 2.42 1.21 1.21 1.16 1.96 0.91 3.29 30.78 44.96 14.28 103.41
Vol.
Wt.
Vol.
3.1 1.4 4.3 34.9 47.3 14.8 105.8
-2.79 0.84 0.25 1.99 0.99 0.99 0.96 1.61 0.77 2.71 24.96 36.86 11.66 84.59
2.7 1.2 3.7 30.1 41.3 12.9 91.9
'Hydrogen consumption 1970 standard cu. ft bbl
VOL. 7 N O . 4 D E C E M B E R 1 9 6 8
277
Table IV. Properties of Hydrodenitrogenated liquid Products
Total Liquid Productn API gravity C, wt. '7c H , wt Yc N,total p.p.m. wt. N,basic, p.p.m. wt. S, p.p.m. wt. Research octane no., unleaded Research octane no., +3 ml. TEL Cetane no. Pour point, F. Saturates' Aromaticsb Olefins' Paraffins Monocyclic naphthenes Dicyclic naphthenes Benzenes Heavy aromatics ASTM color Vis., 100" F. sus. Vis., 21O'F. cs. Mol. wt. Carbon res., ram., Full range Carbon res., rams., 10% bms. Copper corrosion Flash point, F., Pensky-Martin Storage stability, Dupont (Nalco) pad rating
42.7 86.06 13.84 917
... 60
...
... ... ... 69.8 27.5 2.7
...
... ...
... ... ...
Cs-180" F . 180"-400" F.
76.8
...
... 10
...
...
110 61.4 82.2
10 32.8 61.4
... ...
... ... ... ...
...
...
... ... ... ...
...
... ...
... ...
... ...
...
...
... ... ... ... ... ...
...
51.3 85.79 14.19 240
47.7 34.5 2.4 13.2 2.2
...
... ... ...
...
...
(1966). However, catalyst studies by Falk and Berg (1965) indicated that nickel-tungsten promoters were best for H D N . Our results validated and extended these studies to show that high concentrations of these promoters gave best H D N . Since plant costs would be reduced if only one H D N catalyst were utilized, such a single catalyst should have maximum H D N activity, and for this reason Harshaw 4303-E was chosen. After 100 hours of operation, a snap sample of total liquid product contained 917 p.p.m. of nitrogen; this was comparable to the combined, total product and indicates no catalyst aging took place. Although the results obtained were satisfactory for the purposes of our study, better yields and properties might be realized after further exploratory study. For instance, most of the hydrogen consumed (1970 standard cu. ft./ bbl.) was taken up by saturation reactions, not by direct reactions of nitrogen elimination. The hydrogen contents of feed and products bear this out; it was also evidenced by the heat released. Had the feed nitrogen been present as pyridine, for instance, only 830 standard cu. ft. of hydrogen would have been required to convert the pyridine to pentane and ammonia. Naphtha Hydrodenitrogenation. Table V cites the conditions and yields of products from the naphtha hydrodenitrogenation. The properties of the liquid product were similar to the naphtha charge except for the nitrogen content which was reduced from 240 to 1.5 p.p.m. (see Tables IV and VI). The liquid product was not stabilized, because unstabilized liquid from first-pass H D N contained only 0.4%; propane plus butanes, and gas from the present l & E C PRODUCT RESEARCH A N D DEVELOPMENT
cc
400"-650" F .
36.1
37.5
...
... ...
...
710 415 11
...
... ... +45 67.8 26.2 6.0
...
... ... ... ...
...
...
...
100.
...
4000 Max.
... ...
... ...
... ...
50 Min. -15 Max.
50.2 -15
+95.
... ... ... ...
... ... ... ... ...
... ...
...
...
...
1.46 234 0.03
... ...
... ...
34.9
30
neu
L 0.P
R-5 catalyst.
Test Temp., F.' LHSV Moles H? naphtha Pressure, p.s.i.g. I
1
2
3
806 2.96 8.35 500
829 2.89 8.52 500
853 2.96 8.21 500
4
6
3
878 2.96 8.59 500
908 2.95 8.72 500
954 2.96 8.59 500
Yields ~
Wt. Hydrogen Methane Ethane Propane Isohutane n-Butane Pentanes and heavier Total Hydrogen prod. standard cu. ft. bhl. Yield,-Oct. (liquid vel. i x RON +3 ml. T E L ) x 10 '
Vol.
1.81 0.30 0.89 1.48 0.98 94.54 100.00
1.3
91.5 -
905
92.8
Wt.
Vol.
1.84 0.48 0.96 1.T8 0.06 2.43
0.1 3.2
U't.
Vol.
1.89 0.82 1.22 2.52 0.98 2.41
1.3 3.1
Wt.
2.05 1.26 2.24 4.32 1.12 3.03
Vu/
Wt
1.5 4.0
1.86 2.13 4.03 6.70 1.97 4.26
Vol.
Wt.
Vol
2.7 5.6
1.01 3.49 6.98 11.50 2.67 5.71
3.6 10.1
92.45 89.6 90.16 ~ 87.1 85.98 82.8~ 79.05 mi D- . -3 100.00 92.9 100.00 91.5 100.00 88.3 100.00 83.8 920
946
1028
931
~~
66.64 62.8 100.00 76.5 507
(
59.0
56.7
78.1
77.8
T4.0
64.1
" M e a n reading o f nine ecml? spaced TCb in catalyst bed. ~~
VOL. 7 NO. 4 DECEMBER 1968
279
Test
IBP
224 251 265 2 79 291 304 318 333 346 358 374 388 416 48.
5 10 20 30 40 50 60 70 80 90 95
EP API gravity . Octane nos.' RON, clear RON, +3 MON, clear MON, +3
51.6 64.4 52.0 64.4
100
960
2
3
4
5
6
192 226 251 274 282 296 304 326 342 358 378 401 430 48.3
161 212 239 262 276 292 307 322 338 354 375 399 437 47.7
136 188 216 246 264 280 296 314 328 346 371 402 449 47
129 167 189 220 245 265 282 299 318 338 366 401 452 45.5
120 176 191 214 235 253 271 289 308 328 358 385 422 43.5
69.0 85.6 65.0 78.2
74.9 89.7 68.8 82.2
90.9 98.0 82.4 87.6
97.6 102.1 86.7 91.4
1
4
980
Table VIII. Reforming of Naphtha-Properties of Reformates
83.1 94.0 75.2 85.2
940
L.
920
-
rn 3 +
