GREEKLETTERS = effectiveness factor = catalyst particle density (g/ml)
7
pp
Z
= moles summation
7
= time factor (g hr/mol)
r,,,
= tortuosity factor
6
= modified Thiele modulus
SUBSCRIPTS D = p-DEB E = p-EVB H = hydrogen L = light hydrocarbons V = p-DVB literature Cited
Amos, J. L., Rubens, L. C., Hornbacher, H., in “Styrene, Itma Polymers, Copolymers and Derivatives,” Reinhold, New York, N. Y., 1952, Chap 16, pp 709-35.
Balandin, A. A., Shuikin, N. I. Marukyan M., Brusov, 1. I., Seimovich, R. G., Lavrovskaia, T. K., dikhailovskii, V. R., Zh. Prikl. Khim.., 32. f14.54). . ~2566 -.__ , __,_ j _ l
Bogdanova, 0.K., Balandin, A. A., Belomestnykh, I. P., Izv.
Akad. Nauk SSSR Ser. Khim., 2100 (12) (1963). CarrA, S., Forni, L., lnd. Eng. Chem. Process Des. Develop., 4 , 281 (1965). Corrigan, T. E., Chem. Eng., 62, 199 (April 1955). Huang-Minlon, J.Amer. Chem. Soc., 68, 2487 (1946). Mowry, D. T., Renoll, AT., Huber, W. F., ibid., 68, 1105 (1946). Satterfield, C. N., “Mass Transfer in Heterogeneous Catalysis,” MIT Press, Cambridge, Mass., 1970, Chap 4, pp 164-207. Sultanova, A. I., Nagiev, T. bl., Azerb. Khim. Zh., 85 (4) (1964). Virasoro, E., Capeletti, R., Rev. Fac. Ing. Quim. Unav. hrac. Lztoral (Santa Fe, Arg), 29, 95 (1960). Yang, K. H., Hougen, 0. A., Chem. Eng. Prog., 46, 146 (1950).
RECEIVED for review December 22, 1970 ACCEPTED March 29, 1971 The financial aid of the Italian Consiglio Nazionale delle Ricerche is acknowledged.
Hydrogenation of COED Process Coal-Derived Oils Harry E. Jacobs Atlantic Rich$eld Co., Harvey, 111.
J. F. Jones and R. 1. Eddinger’ FMC Corp., Princeton, N . J . 08540
Oil produced by the COED process at FMC’s process development unit wos hydrogenated in bench units at Atlantic Richfield’s R&D laboratories. The hydrogenation was conducted at 3000 psig over a fixed bed of NiMo catalyst. The response to hydrogen varied with the source of the coal. Hydrogen consumption, and oxygen, nitrogen, and sulfur removal were correlated with operating conditions. The properties of the hydrogenated products were examined and were suitable for processing in conventional petroleum refinery units.
T h e Office of Coal Research, Department of the Interior, has sponsored several projects having as one of their objectives, the production of oil from coal. One of these, Project COED, conducted by FMC Corp., produces oil by low temperature pyrolysis of coal. The Atlantic Richfield Co. (XRCO), which had been conducting hydrogenation experiments on coal-derived oils, was requested by the Office of Coal Research to hydrotreat several COED oils. Accordingly, the following work was carried out by ARCO in cooperation with
FMC. Raw oils from the COED process have a low hydrogen coiitent and high concentrations of osygen, nitrogen, and sulfur. Hence, hydrotreating is required before the oils can be used in present day petroleum refining processes. I n the COED process, the raw oil product is taken from the reactors as a vapor, leaving the residual char to be withdrawn separately as a fluidized solid. Any entrained char in the condensed raw oil is removed by filtration. Consequently, the COED oil, being virtually solids-free, can be hydrotreated in fixed-bed reactors. I n this study the hydrotreating was accomplished in a continuous downflow system with a 100-cc catalyst bed. To whom correspondence should be addressed. 558 Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971
The catalyst used was HDS-3A1manufactured by American Cyanamid. I t is a l/lo-in. estrudate containing 3.070 NiO and 15% ;\loo3on alumina. Coal-Derived Oil Hydrogenation Feeds
COED oils from three coals were charged. Their coniposition and gravity are shown in Table I. Included also is a typical Venezuelan petroleum crude for comparison. M70rk reported earlier (White et al., 1968) was for COED oil from the Utah coal only. The Utah and Illinois coals are high-volatile B bituminous coals. Pittsburgh-seam is a high-volatile A bituminous coal. The major differences between the COED raw oils arc the lower density, lower sulfur, and higher hydrogen content of the Utah oiI. The Illinois KO.6 oil is highest in sulfur and oxygen. Hydrotreating Conditions
All of the hydrogeiiation experiments were conducted a t 3000 psig. Hydrogen flow rates were generally between 8000 and 12,000 scflbbl. Variations in hydrogen rates did not have a significant effect on the hydrogenation reactions. The process variables were space velocity and temperature. The
Table 1. Properties of Hydrogenation Feeds
%C
Oil source
Utah A-seam IllinoisNo. 6 Pittsburgh-seam Venesuelancrude
%H
%0 %N
83.3 8.5 6.8 80.1 7 . 2 9 . 1 83.3 7.1 6.6 86.0 12.6 . . .
1.1 1.1 1.2 0.2
%S
Density, 'API
0.3 2.5 1.8 1.2
-3.5 -13.1 -12.3 32.7
PRESSURE 3000 P S I 5
6.6% OXYGEN
ILLINOIS N 0 . 6 9 . 1 Yo OXYGEN
6 8
latter ranged from 65O-85O0F, and space velocities (WHSV) of 0.3-3.0 lb/hr of oil per lb of catalyst were used. Hydrogen consumption was between 1500 and 4500 scf/ bbl. The exothermic reaction resulted in a n uneven temperature profile in the catalyst bed. I n R typical run, the temperature rose 65" in the first 2-4 in. of the bed. The temperature then declined gradually in the remaining bed as the heat sink of the reactor and furnace removed heat from the system. The temperatures used for correlating the data are arithmetic averages of temperatures taken a t 1-in. intervals in the bed. This average temperature was about 20" less than the maximum temperature.
\ \ HATU 6 . 8 % OXYGEN
996 8
t I
I
0
0.5
I
The data on removal of nitrogen, oxygen, and sulfur are shown in Figures 1 through 3. These data were correlated by simple first-order kinetics by plotting each of the heteroatom removals as a function of reciprocal space velocity times a temperature factor. The temperature factor, F,, shown in the bottom curve of each figure was derived from the experimental data by plotting heteroatom removal a t various levels of space velocity and temperature. The removal of nitrogen from COED oils is shown in Figure 1. The temperature factor indicates an activation
2.0
0.8 TEMPERATURE * F
05
Heteroatom Removal
I
I
1,s
I .o
i.0 TEMPERATURE
i5.
4
x
9.0 104
Figure 2. Oxygen removal
energy of 29,000 Btu/lb mol. This activation energy compares with 16,400 Btu/mol for temperatures above 750°F shown for the coal-derived oil (Utah coal) in the earlier study (White et al., 1968). As shown, to remove nitrogen from the Utah oil requires only two-thirds the severity of the Illinois oil. The Pittsburgh-seam oil requires twice the severity of the Illinois oil.
PRESSURE 3000 PSlG
1.2% NITROGEN
ILLINOIS N 0 . 6
0 Y
98 I
0
0.5
,
1
I
1.0
1.5
WHlSV
2.0
I 2.5
FI 99.6 99.7
0
0.8
I
0.5
TEMPERATURE
I .o
I .5
'F
0.6
8.5
8.0
TEMPERATURE
Figure 1 . Nitrogen removal
4
9.0 x 104
Figure 3. Sulfur removal Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971
559
Table II. Space Velocities for 80'7, Removal Conditions, 720' F and 3000 Psig 80% nitrogen
80% oxygen
80% sulfur
removal, WHSV
removal, WHSV
removal, WHSV
9
1.2 0.8 0.4
1.8 1.5 0.6
1.8 2.6 1.1
8
Utah A-seam Illinois No. 6 Pittsburgh-seam
Similar correlations for oxygen removal and for sulfur removal from COED oils are shown in Figures 2 and 3. Indicated activation energies are 15,000 Btu/lb mol for oxygen and 9500 Btu/lb mol for sulfur removal. I n the first study with Utah oil (White e t al., 1968), the activation energy for oxygen removal at temperatures above 730-50°F was shown as 8500 Btu/lb mol. Correlations for sulfur were not derived. T o remove oxygen from the Pittsburgh-seam oil requires 21/2 times the severity of the Illinois oil. This is shown in Figure 2. Also shown is the need for 80% of the severity of Illinois oil to remove the oxygen from the Utah oil. Similarly, referring t o Figure 3, Pittsburgh-seam oil requires about 21/2 times the severity of Illinois oil to remove sulfur. Sulfur removal from Utah oil is indicated t o be more difficult than from Illinois No. 6 oil. This difference may not be real. Because of the low 0.3% sulfur content of the Utah oil feed, there is more scatter in the data on this feed than for the other two. White e t a1 (1968) showed that there was a marked break at about 730-50°F in the Arrhenius plots for oxygen and nitrogen removal. The activation energies were substantially lower a t the higher temperatures. Qader e t al. (1968) also found that the activation energies for nitrogen and
$
SYMBOL A
OIL SOURCE UTAH ILLINOIS N0.6 PITTSBURGH
1'
4
:
a3 I
i 01 600
I
I
700
800
TEMPERATURE , O F
Figure 5. Gas yields
sulfur removal from a low temperature coal t a r changed a t about 750°F. The general scatter of the data and the method of determining average temperature in this work preclude the identification of different activation energies above and below 750°F. Nitrogen is the most difficult of the heteroatoms t o remove. At 720°F and 1 WHSV, the removals from Illinois No. 6 oil were as follows: nitrogen, 72; oxygen, 91; and sulfur, 98%. The required space velocities for 80% oxygen, nitrogen, and sulfur removal from the three oils are given in Table 11.
i b p,llG, ;,Mr
OIL
UTAH ILLINOIS N 0 . 6
A 30
40
t
C 7 . j
TOTAL SCF Hz/BBL SATURATION FOR 4600
6200
6400
-
k
PRESSURE 3000 PSlG
\
Total hydrogen consumption was also correlated by pseudo first-order kinetics. First the quantity of hydrogen required to convert the oil to cycloparaffins (CnHz,J, and to remove all oxygen, nitrogen, and sulfur as water, ammonia, and hydrogen sulfide was determined. This hydrogen saturation value assumes no cracking.
y,;,,/ I
0
I
I .o
0.5
* I .5
I
2.5
2.0
3.0
Ff
2.0r
0.8
0.6
020
800
780
TEYPERATURE .F 760 740 720
700
880
660
640 62
0.5
9:o
8.5
8:O
TEMPERATURE
f
Hydrogen Consumption
a 104
Figure 4. Percent hydrogen saturation 560 Ind. Eng. Chem. Process Des. Develop., Vol. 10, No. 4, 1971
Oil source
Hydrogen for tota' saturotion, scf/bbl
Utah A-seam Illinois No. 6 Pittsburgh-seam
4600 6200 6400
The observed hydrogen consumption, divided by the quantity needed for total saturation, is reported as the percent hydrogen saturation. This value was then plotted as a function of reciprocal space velocity and temperature in the same manner as for heteroatom removal. As seen in Figure 4, when plotted in this manner, the data from Utah and Illinois No. 6 oils fall on a common line. The Pittsburgh oil requires 21/z times as much severity to obtain a similar degree of saturation. This method of correlating hydrogen consumption ignores any effect the approach to equilibrium may have on the rate of hydrogen usage. The fact that the data plot so well indicates t h a t the reactions are not equilibrium controlled.
Table 111. Properties of Hydrogenated Oil Product
4*;J
40 r
Oil source
7-
L.
Conditions Weight hourly space velocity 0 . 7 5 Temperature, "F 748 Pressure, psig 3000 Hydrogen consumption, scf/bbl 3500 C4+ liquid product O API 31.8 Viscosity, SUS @ 180°F 34.6 Pour point, "F 80 Ramsbottom carbon, wt % 0.44 Hydrogen, wt 70 12.7 Oxygen, wt % 0.13 Nitrogen, w t % 0.05
