Ind. Eng. Chem. Fundam., Vol. 17, No. 4, 1978
one inlet and one exit stream these equations are shown to reduce to the familiar expressions for such systems while the application to systems with two inlets and one exit is shown to produce the equations of Treleaven and Tobgy (1971). The definition of the characteristic residence time of the ith feed strea:m used here differs from that of Rosensweig (1966) as does the method of relating these characteristic residence times. Also the equation (17) which relates the overall residence-time frequency function to the flowrates and residence-time frequency functions of specific inlet-outlet pairs differs from eq 9 of Buffham and Kropholler which purports to represent such systems. Nomenclature
c = concentration, mol vel-' f ( t ) = residence time frequency function, time-' F ( t ) = residence time distribution, dimensionless
M = total tracer injected in impulse to inlet stream, mol m = number of feeds n = number of outlets Q = volumetric flowrate, vol time-l t = time V = volume Greek Symbols 7
= mean residence time, time
Subscripts 0 = entrance concentration
1 = relevant to stream 1 2 = relevant to stream 2 e = relevant to an outlet stream i = general index number of feedstream j = general index number of outlet stream
291
h = steady state tracer levels Literature Cited Aris, R., "Compartmental Analysis and Theory of Residence-Time Distribution", D in "Intracellular TransDcft", . . 167, K. B. Warren, Ed.. Academic Press, New York, N.Y., 1966. Asbjornsen, 0. A., A.I.Ch.E.-I. Chem. E. Symp. Ser., 10, 40 (1965). Aubry, C..Villermaux, J., Chem. Eng. Sci., 30, 457 (1975). Anand, U. S., Ph.D. Thesis, University of London, 1967. Buffham. B. A.. KroDholler. H. W.. Math. Biosci.. 6. 179 (1970). . . Devillon, J. C., Ph.D: Thesis, University of Nancy, 1972. Gal-Or, B., Resnick, W., Ind. Eng. Chem. Process Des. Dev., 5, 15 (1966). Hammond, D. C., Mellor, A. M., Comb. Sci. Tech., 2, 67 (1970). Hammond, D. C., Mellor, A. M., Comb. Sci. Tech., 4, 101 (1971). Hammond, D. C., Mellor, A. M., Comb. Sci. Tech., 6, 279 (1973). Hanhart, J., et al., Chem. Eng. Sci., 18, 503 (1963). Hanley, T. R., Ph.D. Thesis, Virginia Polytechnic Institute and State university, 1972. Levenspiel, O., "Chemical Reaction Engineering", Wiley, New York, N.Y., 1972. McCord, J. R., 111, Chem. Eng. Sci.. 27, 1613 (1972). Ostergard, K., Adv. Chem. Eng., 7, 71 (1968). Resnick, W., Gal-Or, B., Adv. Chem. Eng., 7, 295 (1968). Reynier, J. P., Rojey, A., Chem. Eng. J., 3, 187 (1972). Rippel, G. R., et al., Ind. Eng. Chem. Process Des. Dev., 5, 32 (1966). Rosensweig, R. E., Can. J . Chem. Eng., 44, 255 (1966). Schmitz, R. A., Amundson, N. R., Chem. Eng. Sci., 18, 265 (1963). Swithenbank. J., et al.. Roc. XIVth Symp. (Int.) on Combustion, 627 (1973). Thring, M. W., Masdin, E. G., Combust. Name, 3, 123 (1959). Towell. G. D., Ackerman, G. H., Boceedings, Fitth European/Second International Symposium on Chemical Reaction Engng, Elsevier, Amsterdam, 83-1-83-13, 1972. Trambouze, P., Proceedings, Fifth European/Second International Symposium on Chemical Reaction Engng, Elsevier, Amsterdam, AI-I-AI-22, 1972. Trambouze, P., Chem. Eng. Sci., 14, 161 (1961). Treleaven, C. R., Tobgy, A. H., Chem. Eng. Sci., 26, 1259 (1971). Vaclavek, V., Collect. Czech. Chem. Commun., 32, 3646 (1967). Zoulalian, A., Villermaux, J., Adv. Chem. Ser., No. 133, 348 (1974). Zvirin, Y., Shinnar, R., Int. J . Multiphase N o w , 2, 495 (1976a). Zvirin, Y., Shinnar, R., Water Res., 10, 765 (1976b). Zweitering, T. N., Chem. Eng. Sci.. 11, 1 (1959).
Received f o r review December 14, 1977 Accepted May 24, 1978
Hydrogen Transfer Cracking of Dibenzyl in Tetralin and Related Solvents Donald C. Cronauer," Douglas M. Jewell, Yatish T. Shah, and Karen A. Kueser Gulf Research & Development Company, Pittsburgh, Pennsylvania 15230
This paper presents the results of a study of hydrogen transfer cracking of dibenzyl in tetralin and related donor solvents. The experiments were carried out in the temperature range of 400 to 450 OC and a range of contact time frorn 0 to 180 min at 10.3 MPa (1500 psig). Detailed experimentation using various donors with dibenzyl indicates that the controlling reaction is the thermal cracking of dibenzyl to form benzyl radicals. The reaction rate is independent of the donor solvent (tetralin, decalin, mesitylene), but the product distribution is dependent upon the type of solvent. Hydroaromatics rearrange to form methyl indane type compounds, thereby resulting in a solvent of reduced hydrogen donor capacity.
Introduction
One alternative to meet the predicted fuel oil shortage is to convert coal into environmentally acceptable fuel via liquefaction. The kinetics of coal liquefaction are very complex. The experimentation needed to understand the reaction paths is expensive and cumbersome, and results are often masked by product analysis problems. It is, however, now well accepted that during coal liquefaction, the depolymerized coal fragments accept hydrogen from the solvent to produce a host of products. The hydrogen-donating capacity of the solvent thus plays a very important role in the product distribution (Squires, 1976; Neavel, 1976; Ktang et al., 1976; Reuther, 1977; Curran et al., 1967). 0019-7874/78/1017-0291$01.00/0
Coal has a very complex structure consisting of a variety of C-C, C-N, C-0, and C-S bonds. To understand how the hydrogen transfer from a donor solvent to the coal fragments occurs, a systematic model compound study investigating the reactions between a variety of carboncarbon bonds and the donor solvent needs to be undertaken. The present paper illustrates the results of such a study for one model hydrogen-acceptor compound (dibenzyl) with a variety of hydrogen donor solvents. Dibenzyl
0 1978 American Chemical Society
292
Ind. Eng. Chem. Fundam., Vol. 17,No. 4, 1978
Table I. List of Experimental Systems for the Dibenzyl Study acceptor solvent
donor solvent"
atmosphere
dibenzyl (10%) tetralin (90%) NZ dibenzyl (10%) tetralin (90%) HZ dibenzyl (10%) mesitylene (90%) N2 dibenzyl (10%) mesitylene (90%) HZ dibenzyl (10%) dimethyl tetralin (40%) + mesitylene (50%) NZ dibenzyl (10%) H,Ph (40%) + mesitylene (50%) NZ dibenzyl (10%) H,Ph (7%) t tetralin (33%) + MST (50%) NZ dibenzyl (20%) decalin (80%) NZ dibenzyl (10%) tetralin (40%) t decalin (50%) NZ dibenzyl (25%) tetralin (75%) NZ dibenzyl (33%) tetralin (67%) NZ N2 dibenzyl(50%) tetralin (50%) dibenzyl (40%) tetralin (10%)t mesitylene (50%) NZ dibenzyl (10%) tetralin (90%) + wire screen NZ dibenzyl (22%) D, tetralin (29%) + MST (49%) N, a H,Ph, octahydrophenanthrene; MST, mesitylene (1,3,54rimethylbenzene). VENT
O
Table 11. Dibenzyl/Tetralin System: Effect of Adding a Screen (Reactors at 450 ' C in a Nitrogen Atmosphere)
GAS SAMPLE
it!
D R Y ICE T R A P
$:
tetralin, wt %
dibenzyl, wt %
F E E D TANK
nEATERS
temperature, C 400,425,437,450,463,475 450 425,437,450 425,450,475 450 450 450 450 450 450 450 450 450 450 400,425,450
LY
- s+ VACUUM
SAMPLE
Figure 1. Schematic of experimental unit.
represents C-C bonds attached to two aromatic rings. Several donor solvents ranging from poor hydrogen donor activity (e.g., mesitylene) to good hydrogen donor activity (e.g., tetralin) have been examined. The hydrogenolysis of dibenzyl to form toluene has been discussed in the literature (Tsuge et al., 1962; Horrex and Miles, 1951; Yura and Oda, 1941; Fortina et al., 1971; Brower, 1977). However, a detailed discussion of the kinetics of the primary and secondary reactions taking place in the presence of a donor solvent does not appear to have been published. Experimental and Analytical Procedures The kinetic experiments were performed in a 300-cm3 batch stirred autoclave. The unit also had 300-cm3heated feed tank to permit the injection of reactants at elevated temperatures under pressure. A schematic of the experimental setup is shown in Figure 1. In most of the runs, a sample (typically 75 g) of donor was charged to the reactor and rapidly heated to the desired temperature. The typical heat-up time was 60-80 min. A t this temperature, an additional charge (usually 75 g) of donor and acceptor was injected using nitrogen at 10.3 MPa (1500 psig). The injected liquid was typically preheated to 275-300 "C. A t the time of injection, the reactor temperature dropped 5-10 "C, but it recovered within 3 min. Provisions were made to take 5-10-cm3 samples from the reactor during the run with the sample line cooled with dry ice to minimize losses. The donor/acceptor reaction products were analyzed on a high-resolution capillary GLC with a 100-ft SCOT column. Selected samples were also analyzed using a medium-resolution GLC interfaced to a duPont 21-491
run no. screen in place sample time, min
1
2
3
1
2
3
no
no
yes
no
no
yes
3 15 30 60 90 120 150 190
9.5 8.8 7.6 8.4 5.8 2.7 1.9 1.3
9.8 8.7 8.2 6.4
9.4 9.1 7.1 5.5 3.4 2.2 1.4 0.9
-
2.6
-
1.5
88.0 87.4 85.7 84.5 85.8 82.4 83.9 81.5 76.8 79.8 77.4 72.4 7.2 68.4 69.9 66.9 65.3 - 62.2 62.3 63.3 54.9
mass spectrometer. Specific samples were analyzed by conventional and high-pressure liquid chromatography. Further analysis of isolated products was carried out using a CEC-103 low voltage mass spectrometer and a CFT20/FT NMR instrument with both hydrogen and carbon-13 probes. In this study, dibenzyl was used as the hydrogen acceptor because: (a) dibenzyl represents C-C bonds attached to two rings; similar carbon linkages are believed to exist in coal; (b) the dibenzyl C-C bond fractures at coal liquefaction conditions (400-450 "C); ( c ) the products are readily analyzed. The primary considerations in the choice of donor solvents were the following: (a) known ability to donate hydrogen, (b) existence in typical coal liquids, (c) availability, (d) cost, and (e) simplicity of structure for ease of analysis. For this study, the choice was limited to hydroaromatics. Most of the reactions were performed in tetralin. A limited number of reactions were done in 2,6-dimethyl tetralin, decalin, and sym-octahydrophenanthrene to determine the effect of donor structural changes. 1,3,5Trimethylbenzene (mesitylene) was used as an inert solvent in which to test the thermal stability of dibenzyl at 450 "C. Although mesitylene (MST) is thermally stable at this temperature, it reacts with benzyl free radicals. A list of experimental conditions examined in this study is shown in Table I. These experiments were performed for contact times ranging from 0 to 180 min and for temperatures between 400 and 475 "C. Results and Discussion To verify that the wall of the reactor did not catalyze the reaction, a run was made with 10% dibenzyl and 90% tetralin under a nitrogen atmosphere at 450 "C with mesh screens of the same material as the reactor wall added to the reactor. As shown in Table 11, the conversions of both
Ind. Eng. Chem. Fundam., Vol. 17, No. 4, 1978
dibenzyl and tetralin were not changed by the increased metal surface area. The study reported here is, therefore, believed to represent purely thermal hydrogen transfer reactions. 1. Primary Reaction. The primary reaction in this study involves the theirmal cracking of dibenzyl followed by stabilization of the free radicals by reaction with a donor solvent. This reaction can be characterized by dibenzyl
k,
-
2[benzyl]
where CB is the concentration (weight percent) of dibenzyl. The values of k l for all the systems examined in this study are illustrated in the form of an Arrhenius plot in Figure 2. The activation energy for k, was found to be approximately 48 100 cal/g-mol. This activation energy is consistent with that of Horrex and Miles (1951), who reported a value of 48000 cal/g-mol for the gas-phase cracking reaction. The fact that eq 4 correlated the experimental data for all systems implies tha,t the benzyl free radical is very active and readily extracts hydrogen from any available solvent. Furthermore, recombination of benzyl free radicals does not occur a t a significant rate. Combination of the benzyl radical with mesitylene radicals to form adducts can be a significant reaction in the absence of sufficient hydroaromatics. Combination of benzyl radicals with good donor solvent molecules, such as tetralin and octahydrophenanthrene, was not observed. Incidentally, it was not possible to carry out experiments using dibenzyl in the absence of a solvent due to its melting point of 51 "C. 2. Secondary Reactions. Unlike the primary reaction, secondary react ions are strongly dependent upon the structure of the solvent. The product distribution is, therefore, a function of the donor solvent. (a) Reactions with Mesitylene in Nitrogen. When mesitylene (MS'I') is used with dibenzyl under a nitrogen atmosphere, the major Secondary reaction is given by
+ 2 mesitylene
-N2
2 toluene
A
A
0
+
2 [ Benzyfl
1 0 % Dibenzyl t 90%Tetralin i n N 2 10'. Dibenzyl 9O%Tetralln in H 2 10'. Dibenzyl 90% Mesilylene in N 2 1 0 % Dibenzyl t 90% Mesitylene in H 2 10% Dibenzyl + 4 0 w Dimethyl Tetralin + 5 0 % Mes1-J tylene in N, 10% Dlbeniyl +40% HIP+ SO%Mesitylene i n N:, 1 2 0 ~ D i b e n z y l + 8 0 ' f ~ D e c a l i nin N2 10% Dibenzyl 4 0 % Tetralin 5 0 % Decalin in N2
+
+
+
+
(1)
[benzyl] + donor solvent products (2) where [benzyl] represents the active benzyl free radical. With each donor solvent, a number of secondary reactions also occur producing a variety of products. The mechanisms for these secondary reactions are complex and differ with each solvent. They may involve the formation of more free radicals and a series of chain transfer reactions. The effect of the donor solvent on dibenzyl conversion depends on the relative rates of reactions 1 and 2. If the rate of reaction 1 is cornparable to or greater than that of reaction 2, the nature of the donor solvent will affect the rate of dibenzyl conversion. However, if the first reaction is slower, the rate of forination of products from the overall reaction, namely dibenzyl + donor solvent -* products (3) will not be significantly dependent upon the nature of the donor solvent. The foymation of the benzyl free radical (reaction 1) in this case is largely a thermal process, independent of solvent. Once the free radical is formed, it reacts with any donor solvent to form products. Analysis of experimental data indicates that reaction 1 is the limiting reaction for dibenzyl conversion. Thus, for all donor solvents examined in this study, the rate of disappearance of dibenzyl can be expressed as
2[benzyl]
Dlbenzyl3 0 0
+ dimer
(5)
293
i I
33cDibenzyl t 67% Tetralin in N 1 z D i b e n z y l 5 0 a Tetralin in N2
+
1
! I
\J50
I
E :48,10Ocal./g. mole
I
+
103( ~ ' 1 )
RECIPROCAL TEMPERATURE
Figure 2. Arrhenius plot for the primary dibenzyl reaction.
The dimer is considered to be di-mesitylene; however, the products from this reaction may vary. Mesitylene can also react with benzyl radicals to form other adducts
[benzyl ]
CH3
mesitylene
CH3
MST/benzyl adduct
Experimental data indicate that reaction 6 may not be first order with respect to mesitylene concentration. GC-MS studies were performed to identify major products. Some typical toluene and adduct concentrations obtained in these studies are shown in Table 111. These studies identify toluene as the predominant product. (b) Reactions with Mesitylene in Hydrogen. The major secondary reaction is identical with that when the mesitylene/dibenzyl system is used under a nitrogen atmosphere (see reaction 5). The other secondary reaction (6) also occurs. Unlike the previous case, the presence of hydrogen promotes cracking of mesitylene to form xylenes. Typical concentrations of cracked products in this system are illustrated in Table 111. The adduct product for the dibenzyl/MST system in either Hz or N2 is a family of oligomers. Nearly all of the adducts incorporate mesitylene or mesitylene fragments (xylene). The distribution of adducts is dependent upon the reaction time, atmosphere (N2or Hz), concentration
294
Ind. Eng. Chem. Fundam., Vol. 17, No. 4, 1978
Table 111. Dibenzyl/Mesitylene System : Typical Product Distributions (wt 5%) at Various Reaction Temperatures (Reactions with a Feed of 10%DB/90% MST in a Nitrogen Atmosphere) sample time,min
425°C
3 15 30 60 90 120 150 180
0.3 0.7 1.2 2.0 2.6 3.1 3.4
toluene, wt % 437°C 450°C
425°C
0.2 1.5 3.1 5.1 5.9 7.1 7.7 8.0
0.1 1.1 1.7 2.7 4.0 4.7 5.3 5.8
0.3 0.6 1.4 1.8 2.8 3.7 3.6
adducts: wt % 437°C 450°C 0.2 1.4 1.9 3.3 4.5 5.3 5.4 6.3
light ends,b wt % 437°C 450°C
425°C
0.3 1.4 3.6 5.6 6.6 7.6 8.3 8.5
a Components of higher boiling point than mesitylene as detected by GLC. zene, xylenes, and light unknowns).
0.2 0.1 0.4 0.5 0.7 0.7 0.8 0.8
0.5 0.6 0.3 0.2
0.6 1.1 0.3 0.4 0.5 0.6 0.6
0.5 0.5 0.5
Cracked products (benzene, ethyleneben-
16
I
I
I
I
I
I
I
I
I
I
-
L
Mesitylene
Fraction isolated l o r analysis
14 -
Olbenryt
Time
i
20% Dibenzyl and 80% Decal n with N 2
12
-
A
10% D i b l m y l 40% Tctralin and 50% Decalin with
0
10% Dibcnzvl and 90% Tetra1 n w t h H 2
#
10% OiQenzyl and 90% Tetra1 n w w
-
N2
-
N2
i
-
Figure 3. Dibenzyl/mesitylene system in a nitrogen atmosphere: 3-h reaction period.
of acceptor radicals, and concentration of other solvent molecules. One dimer preferentially formed results from abstraction of H from -CH3 group. The above dependence was determined by observing several GLC curves. Figure 3 shows the typical product distribution for the dibenzyl/mesitylene system with a nitrogen atmosphere. Again, toluene and adducts are the primary products. It is noted that, in this case, one particular adduct is preferentially formed. This adduct has been identified as 3,3',5,5'-tetramethyl dibenzyl, which is the primary mesitylene dimer.
2
0 0
20
40
60
80
100
120 140
160
180
TIME, Min.
Figure 4. Naphthalene concentration vs. time for different feeds at 450 "C.
CHJ
H3C'
( c ) Reactions with Decalin in Nitrogen. It is thermodynamically likely that decalin will dehydrogenate to form tetralin. Therefore, the following reactions are possible when using decalin (7)
2 H +@ J=yJ-y- (
decalin
[benzyl]
decalin
toluene
tetralin
- UCH3 @ Q J
(9)
t
f
[benzyl]
tetralin
tetralin
toluene
naphthalene
Decalin can react with dibenzyl by reaction 8 or by a series of reactions (7 and then 9). The series of reactions will
produce a significantly greater amount of naphthalene than will reaction 8. Results from experiments a t 450 "C and 3 h with 20% dibenzyl and 80% decalin showed a low production of naphthalene (1.4 wt %), while an 8% naphthalene yield is observed when tetralin is used as the solvent (see Figure 4). These results imply that reaction 8 is considerably faster than reaction 7 . Experiments were made using a solvent mixture of 40% tetralin and 50% decalin with 10% dibenzyl. Naphthalene production in this case was close to that when using tetralin alone (see Figure 4). This implies that reaction 9 is more favorable than reaction 8, and tetralin is a better hydrogen source than decalin for the benzyl free radicals. This is further substantiated by examining the production of toluene illustrated in Figure 5. As shown in this figure, for a temperature of 450 "C, the toluene production for the dibenzyl/tetralin system is very similar to that for the dibenzyl/tetralin/decalin system. However, both are considerably different from that observed in the dibenzyl/decalin system. The dibenzylidecalin system produces more cracked products than the dibenzyl/tetralin system. This is il-
Ind. Eng. Chem. Fundam., Vol. 17, No. 4, 1978 295 Tetralin -Methyl
lndane
k 2
First o r d e r w i t h respect to T e t r a l i n
I
i
I
)
10-1 30
I
tr
t
0
I
c a y g mole
L 20
-
E
I
E
k
E
L 1OC
0
1
0
20
40
60
80
100 120 140 160 180
TIME Min
;: ;:xaq7,,"Fp M5T
*
1,
%2
0
-
Figure 5. Toluene concentration vs. time for different feeds at 450 "C.
-
E :37,600
cay
-
5 3.
1.35
z
1.375
1.4
1.425
1.45
1.475
1 / ~ IIOJ(oK")
0
G
Figure 7. Arrhenius plot for the reaction of tetralin forming methyl indane.
L I-
z w u z
0 0
systems: 10% dibenzyl with 90% tetralin in a nitrogen atmosphere, 50% tetralin with 50% mesitylene in nitrogen, 40% tetralin with 50% mesitylene and 10% phenanthrene in nitrogen, and 10% dibenzyl + 50% mesitylene + 35% tetralin 5% octahydrophenanthrene in nitrogen. The graph shows that while the presence of phenanthrene slightly alters the value of k 2 , the presence of dibenzyl increases the rate constant by an order of magnitude. Therefore, the rearrangement to methyl indane is significantly promoted by benzyl free radicals. It is also noted that there is a difference in activation energy in the absence or presence of dibenzyl; this is indicative of a change of reaction mechanism. Tetralin rearrangements can be depicted as proceeding through three-membered ring intermediates: if hydrogen is abstracted from the 1,3 sites (dihydronaphthalene is likely to form if abstraction occurs at the 1,2 or 1,4 sites)
+
0
20
40
60
80
100 1 2 0
140 1 6 0 1 8 0
T I M E , Min.
Figure 6. Cracked products (benzene plus light unknowns as determined by GLC) concentration vs. time for different feeds at 450
"C.
lustrated in Figure 6 f'or a temperature of 450 "C. The figure also shows that, as one would expect, the cracking of solvent is also enhanced by the presence of a hydrogen atmosphere. Rearrangement Reactions One of the most undlesirable rearrangement reactions that occurs in the tetralin/dibenzyl system is
tetralin
cledage
methyl indane
The rearranged product,, methyl indane, can in turn crack to indane t
IIa
IIb
$"3
methyl indane
indane
In the absence of reliable information, reaction 10 was assumed to be first order with respect to tetralin concentration. This model fit the experimental data very well. Figure 7 shows Arrhenius plots of k2 for the following
Although intermediate I is a known comr>ound (Goodman and Eastman, 1964), the three-membered ring will cleave at 400 "C to generate radicals IIa or IIb and
296
Ind. Eng. Chem. Fundam., Vol. 17, No. 4, 1978 F
i
l
l
I
I
I
l I
1
t
100
I
'
90% Tetralin + 10% Dibeniyl in H Z 90% Tetraiin + 10% Dibenzyl ~n N2
-
0
4 F o r m a t i o n of dimethyl isomers
A F o r m a t i o n of m e t h y l i s o m e r s
1
0 Formation of m e l h y l lndane
A
-
I s o m e r i z a t i o n of s y m - H 8 P h
from
-
5 i
-
tetralin
-
75% Tetralin + 25% Dibentyl in N?
4
0 I-
-
< a c z w
0
00
0
20
40
60
80 100
1 2 0 140
160 180
TIME, M i n .
Figure 9. Indane concentration vs. time for different feeds at 450 OC.
0
20
40
60
80
100 1 2 0 140
Methyl lndane
160 180
TIME .Min.
T h e r m a l Stability of Donor Solvents In addition to reactions 10 and 11,the following reaction can occur because of the instability of tetralin above 400 "C
tetralin
naphthalene
Other reactions such as tetralin cracking to form toluene or naphthalene cracking to various products are also possible but are normally slow and appear to be negligible. The data indicate (see Figure 9) that for the system of 50% mesitylene with 50% tetralin in nitrogen, methyl indane dealkylation, reaction 11,occurs slowly. For low
lndane
k3
Z e r o order r e a c t i o n
Figure 8. Stability of octahydrophenanthrene at 450 "C.
finally the methyl indanes. The common observation of the 1-methyl indane isomer can be attributed to greater stabilization of the benzyl radical IIa. sym-Octahydrophenanthrene ( H 8 h ) would be expected to follow the same rearrangement as tetralin except with more isomer possibilities. Since dehydrogenation always accompanies rearrangement, isomers of tetrahydrophenanthrene can also be expected. The fact that sym-H8Ph is indeed unstable at 450 "C is illustrated by Figure 8, which shows that after 3 h 60% has isomerized. The feed for this run was 10% sym-HEPh with 90% 1,2,4-trimethylbenzene. Since the methyl and dimethyl isomers, as a group, were resolved by packed-bed GLC and identified by mass spectrometry, their concentrations are also plotted. Smaller amounts of tetrahydrophenanthrene isomers are present but not shown in this figure. For reference, the level of tetralin isomerization to indane is also shown in this figure. These studies are particularly relevant to coal liquefaction processes that depend on good donor solvents. The rearranged hydroaromatics resist further dehydrogenation (do not transfer hydrogen) but readily dealkylate. The reactions are, for practical purposes, irreversible. Since the products are isomers, mass spectrometry cannot detect the nondonor species.
I
I
r
n x
.-
'
\ 1.35
1.375
1.4
I/T
1.425
1.45
1.475
1.5
x~d3(o~-')
Figure 10. Arrhenius plot for the cracking of methyl indane to form indane.
levels of production of indane, it can be characterized by a zero order mechanism. Thus
The above equation also correlates the data with systems of 10% dibenzyl and 90% tetralin in nitrogen; 50% mesitylene, 40% tetralin, and 10% phenanthrene in nitrogen; and 50% mesitylene, 35% tetralin, 5% HEPh,and 10% dibenzyl in nitrogen. The Arrhenius plots for K 3 obtained for these systems are described in Figure 10. As shown, the presence of dibenzyl significantly increases the
Ind. Eng. Chem. Fundam., Vol. 17, No. 4, 1978
297
dCN = K4
dt where K4 is the zero-order rate constant and dCN/dt is the net production rate of naphthalene by reaction 12. The Arrhenius plots for K4 are shown in Figure 11. The rate constant, K4, is only mildly increased by the presence of an acceptor (dibenzyl). Phenanthrene appears to somewhat hinder reaction 12.
h
t
-1
1
4 t 1.35
1.375
1.4
yTx16'
1.425
1.45
1.475
1.5
(OK1)
Figure 11. Arrhenius pllot for t e t r a l i n dehydrogenation t o naphthalene.
rate of reaction 11. Phenanthrene appears to somewhat hinder the reaction rate. Preliminary analyses of data at 450 "C with 10% dibenzyl and 90% tetralin in hydrogen and dibenzyl with tetralin in nitrogen at lhigh dibenzyl concentrations (>25%) indicate that for these systems, reaction 11is no longer zero order (see Figure 9). In these cases, reaction 11 is most likely first order with respect to methyl indane concentration. Also, the rate of indane production with a feed of 90% tetralin plus 10% dibenzyl in a hydrogen atmosphere is double that with the same feed in a nitrogen atmosphere. The naphthalene formation by reaction 12 can be obtained by subtracting: the naphthalene production from reaction 9 from total naphthalene production. This net production is believed to occur by reaction 12. For the tetralin systems 90% tetralin with 10% dibenzyl in nitrogen, 50% tetralin with 50% mesitylene in nitrogen, and 40% tetralin with 50%)mesitylene and 10% phenanthrene in nitrogen, the net production rate of naphthalene follows the rate expression
General Discussion In the temperature, pressure, and contact time ranges commonly encountered in donor solvent thermal coal liquefaction processes, good donor solvents will not directly break carbon-carbon bonds of the type encountered in dibenzyl. This bond must be first thermally broken to form free radicals. Once free radicals are formed, they will readily react with any donor solvent. The nature of the product distribution will, of course, depend upon the nature of the donor solvent. This is an important conclusion since it implies that the operating conditions, such as temperature, pressure, and contact time, may be equally or even more important than the nature of the donor solvent in coal liquefaction processes. Since coal has a complex structure consisting of a variety of carbon-carbon, carbon-nitrogen, carbon-sulfur, and carbon-xygen bonds, further work on the role of donor solvent in breaking of other types of bonds is needed. This work is presently under investigation, and it will be reported in the near future. Acknowledgment Funding of this research study was provided by the US. Energy Research and Development Administration under Project E(49-18)-2305. We wish to also acknowledge our indebtedness to L. Kindley (DOE), H. Podall (DOE), A. B. King, R. G. Goldthwait, and H. G. McIlvried for helpful suggestions. Literature Cited Brower, K. R., F u e l , 56, 245 (1977). Curran, G. P., Struck, R. T., Gorin, E., Ind. Eng. Chem. Process Des. Dev., 6, 166 (1967). Fortina, L. Maggiore, R., Toscano, G.,Ann. Chim. (Rome)81 (4), 283 (1971). Goodman, A. L., Eastman, R . H., J . Am. Chem. SOC., 86, 908 (1964). Horrex, C.,Miles, S . E., Discuss. Faraday SOC.,No. I O , 187 (1951). Kang, C.C., Nogbri, G.,Stewart, N., Am. Chsm. Soc.,Dhr. Fuel Chem., Prepr.,
21 (5),19 (1976). Neavel, R. C., F u e l , 55, 237 (1976). Reuther, J., Ind. Eng. Chern. Process D e s . Dev., 16, 249 (1977). Squires, A. M., "Reaction Paths in Donor Solvent Coal Liquefaction," a paper presented at Conference on Coal @sHication/Liquefaction, Moscow, USSR, (Oct 1976),s p s ~ r e dby US-USSR Trade and Economic Council, Inc., 1976. Tsuge, O., Yanagi, K., Tashro, M., Kwv Tam, 14 (lo),507 (1962);Ghem. Abstr.,
80,433d (1974). Yura,
S.,Oda, R., J. SOC. Chem. Ind. Jpn., 44. 476 (1941).
Received for review D e c e m b e r 19, 1977 Accepted J u n e 30, 1978
