Energy & Fuels 1993, 7, 978-980
978
Effect of Ammonia on the Hydrogenation of Phenanthrene during the Hydrodenitrogenation of Quinoline Chung M. Lee and Charles N. Satterfield' Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received August 3, 1993. Revised Manuscript Received August 4, 1 9 9 9
The hydrogenation of phenanthrene in the presence of 24 kPa of ammonia was studied during the hydrodenitrogenation (HDN) of quinoline on a sulfided NiMo/AlaOa catalyst a t 360 "C,6.9 MPa total H2 pressure in a vapor-phase tubular reactor. Tetralin, used as a carrier liquid, was slightly hydrogenated. Ammonia, through competitive adsorption, reduced the formation of phenanthrene hydrogenated products and the hydrocracking of 9,lO-dihydrophenanthene to biphenyl. The effects occur largely when the HDN reactions are substantially completed.
Introduction The effect of ammonia on the hydrogenation of naphthalene and butylbenzene during the hydrodenitrogenation (HDN) of quinoline was recently reported.' The motivation behind those studies and this study was to investigate the possibility of control of hydrogenation selectivity of aromatic hydrocarbons by the addition of ammonia during hydrotreating. The present study focused on the effect of ammonia on the hydrogenation of phenanthrene as a representative three-ring aromatic hydrocarbon. Limited literature is available on the hydrogenation of phenanthrene. Shabtai et aL2 reported on the .hydrogenation of phenanthrene and pyrene in a semibatch autoclave reactor. Under their conditions, they found 1,2,3,4-tetrahydrophenanthrene(P4H), rather than 9,lOdihydrophenanthrene (P2H), to be the principal first product formed. Lemberton and Guisnet3 recommended phenanthrene hydroconversion as a test reaction for studying hydrogenation and cracking properties of coal hydroliquefaction catalysts. In a brief study, LaVopa4 found that the addition of piperidine restricted the amount of hydrogenation of phenanthrene during its purification by hydrotreatment. Girgis and Gates5in a recent detailed review critically analyzed the literature on high-pressure catalytic hydroprocessing with substantial consideration of binary mixtures.
Experimental Section The reactor and catalyst used in this study have been described previously.176 In brief, the reactor was avertical0.52-cm4.d. tube immersed in a fluidized sand bath to maintain isothermality and the catalyst was a commercialNiMo/AlzOs (AmericanCyanamid HDS-3A). Methods of operation and analytical procedures were as before.
* Author to whom correspondenceshould be addressed. e Abstract published
in Aduance ACS Abstracts, October 1, 1993.
(1)Lee, C.M.; Satterfield, C. N. Energy Fuels 1992, 6, 315.
(2)Shabtai, J.;Veluswamy, L.; Oblad, A. G. Prepr. Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1978, 23, 107. (3) Lemberton, J.-L.; Guisnet, M. Appl. Catal. 1984, 13, 181. (4) LaVopa, V. Sc.D. Thesis, MIT,1987. (5)Girgis, M.J.; Gates, B. C. Znd. Eng. Chem. Res. 1991, 30, 2021. ( 6 ) Lee, C. M.; Satterfield, C. N. Energy Fuels 1991, 5 , 163.
0887-0624/93/2507-0978$04.00/0
Reactants. To achieve the desired solubilityof phenanthrene it was necessary to use tetralin (99% pure, Aldrich)as the solvent rather than hexadecane, used in the earlier related study of the hydrogenation of naphthalene. All runs were made with feeds consisting of 0.24 mol/L of phenanthrene (99+ % pure, Aldrich) and 0.024 mol/L of quinoline (99% pure, American TokyoKasei). Propylamine (98% pure, Aldrich) when desired was added at a concentration of 0.24 mol/L to generate in situ an ammoniapartial pressure of 24 kPa in the reactor. Complete HDN of the quantity of quinoline in the feed produces a NHs partial pressure of only 3 kPa, so the effects of NH may be attributed to that deliberately formed via propylamine. A hydrogen sulfide partial pressure of 12 kPa, for maintaining the sulfided state of the catalyst, was generated in situ in the reactor by the addition of 0.12 mol/L of l-dodecanethiol (98% pure, Aldrich) to the feed. Analysis. Liquid samplesof the reactor effluentwere collected downstream from a low-pressure phase separator and analyzed on a Perkin-Elmer Sigma 1B gas chromatograph equipped with dual flame ionization detectors. Concentrations of the products were related to the starting liquid feed. Carbon mass balances on total products compared to phenanthrene disappearance exceeded 93% . Separation of reaction specieswas accomplished by using either Supelco SPB-35 (30 m, 0.25 mm i.d., 0.25 pm) or Supelcowax 10 (30 m, 0.25 mm i.d., 0.25 pm) columns. Phenanthrene and 9,lO-dihydrophenanthrene (P2H) were identified by comparing their retention times with those of pure standards. The presence of other products was identified by GC/MS.
Results All data were obtained a t 360 "C, 6.9MPa total pressure, in the presence of 12 kPa of HzS and 24 kPa of NH3 or its absence, both generated in situ. Space-times, expressed relative to phenanthrene, were varied from 100 to 600 h g of catalyst/mol of phenanthrene. Figure 1 shows a plot of the hydrodenitrogenation of quinoline in the presence of 24 kPa of ammonia and its absence versus space-time. As in the previous studies with naphthalene and butylbenzene, ammonia had only a slight inhibiting effect on the overall removal of nitrogen from quinoline and its intermediate reaction products. Essentially complete HDN was achieved a t the highest spacetimes. (Percent HDN is defined here as degree of conversion of quinoline and N-containing intermediates into hydrocarbons.) Figure 2 shows a plot of concentrations of phenanthrene and its products for the same runs shown in Figure 1.Five 0 1993 American Chemical Society
Hydrogenation of Phenanthrene with Ammonia
Energy & Fuels, Vol. 7, No. 6, 1993 979
100
0.03 C
.-0 z
-
80
0
.--0
e C
.-
m
6
0
-.-m .0
c
0
z
n I w
e
C
40
L
C
0
m
2 a
0.02
C
60
C
i
20
0.01
.-0 0
I
w
U
2 4 k P a of added NH,
0
0.00
0
100
200
300
400
500
600
700
Space Time Relative to Phenanthrene, (hr-gcatjmol of Phenanthrene) Figure 1. Hydrodenitrogenation of quinoline without added NHs in the presence or absence of 24 kPa of NH8, generated in situ.
++ phenanthrene
0.
o+
0 .
vv
P2H P4H
0
0.08
I +
0.60
C
I
-m
-
..C
C
0.50 C
0
0
I
c
600
700
I
0.05
L
0
0.40
0.04 -
0.03 -
C
0 .e 0
0.02
I
U
0.30
-
0.01
-
P
.m .-
500
0.06 -
0
0
e
m r
400
C
1
2 5
300
C
.-0
PHP A A biphenyl
m
.-
200
Space Time Relative to Phenanthrene, (hr-gcat/mol of Phenanthrene) Figure 3. Inhibition bymnmonia of the formation of perhydrophenanthrene when HDN is nearly complete.
-
P8H
100
m U
0.00
0.20
0
-C
12 /
2
I
100
200
300
400
500
600
700
c
0
c
.-
0.10
0
I LL
0.00 0
100 200 300 400 500 600 Space Time Relative to Phenanthrene, (hr gcatlmol of Phenanthrene)
700
Figure 2. Effect of ammonia on the hydrogenation of phenanthrene during the HDN of quinoline at 360 "C,6.9 MPa. Open
symbols represent no added ammonia; filled symbols represent 24 Wa of ammonia generated in situ.
species of hydrogenated phenanthrenes were detected, 9,lO-dihydrophenanthrene (P2H), 1,2,3,4-tetrahydrophenanthrene (P4H), perhydrophenanthrene (PHP), and two species of octahydrophenanthrene (P8H), 1,2,3,4,5,6,7,8-P8H and 1,2,3,4,9,10,11,12-P8H. The two species of P8H were lumped together as one because of poor resolution of their peaks on the GC. Some hydrocracking also occurred, as shown by the formation of biphenyl. From 1to 4 5% of the solvent, tetralin, was hydrogenated to decalin or dehydrogenated to naphthalene, but there was no significant effect of the presence or absence of 24 kPa of ammonia here on the degree of tetralin hydrogenation. NH3 has no significant effect on the phenanthrene hydrogenation network until the HDN reactions are nearly complete as shown in Figures 3 and 4. Figure 3 shows that NH3 suppresses complete hydrogenation to PHP, and
Space Time Relative to Phenanthrene, (hr-gcat/mol of Phenanthrene) Figure 4. Reduction of the amount of hydrocracking to form biphenyl caused by ammonia when HDN is nearly complete. Figure 4 shows that it reduces cracking to biphenyl. These data are from Figure 2, shown on a larger scale for clarity. Recall that these results were obtained with an integral reactor. With a differential reactor the effects of NH3 would presumably be much more marked. Kinetic Model. The most detailed information on the reaction network (Figure 5) appears to be that of Shabtai e t aL2 They studied the hydrogenation of pure phenanthrene over several catalysts, particularly a sulfided NiW/ A1203,a t temperatures of 200-380 "Cand 2900 psig. They concluded that the main primary product is P4H, which is converted t o P8H and then PHP. P2H was formed in considerably lower yield. Moreau and Geneste' took phenanthrene as being hydrogenated to either P2H or P4H by parallel pathways. Subsequent steps were not specified, but they assumed that P2H could be hydrogenated to P8H, whereas our model assumes that P8H is ~~
~
(7) Moreau, C.; Geneste, P. In Theoretical Aspects of Heterogeneowr Catalysis; Moffat,J. B., Ed.;Van Nostrand Reinhold New York, 1990; p 246.
Lee and Satterfield
980 Energy & Fuels, Vol. 7, No. 6,1993
+
+ v Phenanthrene
Dihydrophenanlhrene (PZH)
H
Biphenyl
A
I
C
Tetrahydrophenanthrene
Perhydrophenanthrene
.-n
+7
phenanthrene dihydrophenanthrene tetrahydrophenanthrene octahydrophenanthrene perhydrophenanthrene biphenyl
\+
I
Oclahydrophenanthrenes (PsH)
Figure 5. Kinetic model of the hydrogenationof phenanthrene. Table I. First-Order Rate Constants for the Hydrogenation of Phenanthrene to Products first-order reaction rate constants" added [NHa], ki kz k3 k4 ks kPa 0 0.004 05 0.002 40 0.00304 O.OO0 290 O.OO0 244 24 0.004 03 0.002 10 0.002 70 O.OO0 103 O.OO0 114 a
i
LL
o
,
0.00 0
100
200
2
300
+ v
formed from P4H. Lemberton and Guisnet reported limited studies a t 430 "C and 10MPa of hydrogen pressure using neat phenanthrene on a NiMo/AlzOs catalyst. A variety of cracked products were formed as well as various partially hydrogenated products, but no P H P was formed. Assuming that each successive hydrogenating and hydrocracking step in Figure 5 proceeds via first-order reaction, then first-order rate constants can be calculated for the sets of data with and without added ammonia as shown in Table I. Figures 6 and 7 show the fit of the first-order rate constants.
The results can be well interpreted in terms of competitive adsorption effects between phenanthrene and its hydrogenated products on the one hand and quinoline and ita nitrogen-containing HDN products and NH3 on the other. It is seen that phenanthrene is readily converted to partially hydrogenated products even in the presence of quinoline and its nitrogen products. In our previous study' it was shown that NH3 can inhibit the hydrogenation of tetralin to decalins but here the effect of ammonia on tetralin hydrogenation was negligible largely because of the much higher concentration of tetralin. The effect of added ammonia is insignificant or minor on the hydrogenation of phenanthrene and intermediates until quinoline and its nitrogen-containing HDN products are largely removed (Figures 3 and 4), because of the greater adsorptivity of the latter relative to NH3. Therefore, kl, kz, and k3, which represent upstream behavior, show no significant effect of ammonia. On the other hand k4 and k g , which represent downstream behavior when nitrogencontaining hydrocarbons have been effectivelyeliminated,
500
600
700
Space Time Relative to Phenanthrene, (hr-gcat/mol of Phenanthrene) Figure 6. Space-timeplot of the hydrogenation of phenanthrene with no added ammonia. Solid lines representfiist-orderkinetics.
Units of k = (mol of phenanthrene)/(hg of catalyst).
Discussion
400
0
A
0
100
200
300
dihydrophenanthrene tetrahydrophenanthrene octahydrophenanthrene perhydrophenanthrene biphenyl
400
500
600
700
Space Time Relative to Phenanthrene, (hr-gcat/mol of Phenanthrene) Figure 7. Space-timeplot of the hydrogenation of phenanthrene with 24 kPa of ammonia generated in situ. Solid lines represent first-order kinetics.
show the substantial inhibiting effect of NH3 on hydrocracking and further hydrogenation. The addition of ammonia has little effect on the overall HDN of quinoline. Thus, NH3 addition provides an operating region where complete HDN of quinoline can be achieved with reduced formation of hydrogenated phenanthrenes and biphenyl. The undesirable hydrocracking is presumably suppressed by the effects of ammonia neutralizing the acidic sites on the catalyst. Acknowledgment. This study was supported by the Office of Fossil Energy, US.Department of Energy, under Grant No. DE-FG22-89PC89775. We also appreciate the assistance of Anthony Modestino from the MIT Energy Laboratory for his help in GC/MS analysis.
~
