Ind. Eng. Chem. Res. 1987,26, 2204-2211
2204
r . = rank-order parameter for optimization variable x,
R = reflux ratio
R, = minimum reflux ratio RPc = reflux ratio for the product column x = reactor conversion x = equilibrium reactor conversion (Table I) x;”= optimization variable YpH = hydrogen purge composition
Literature Cited Andrecovich, M. J.; Westerberg, A. W. “A Simple Synthesis Method Based on Utility Bounding for Heat-Integrated Distillation Sequences”, AIChE J. 1985, 31, 363. Boland, D.; Hindmarsh, E. “Heat Exchanger Network Improvement”, CEP 1984, 80(7),47. Douglas, J. M. “A Hierarchical Decision Procedure for Process Synthesis”, AIChE J. 1985, 31, 353. Douglas, J. M.; Woodcock, D. C. “Cost Diagrams and the Quick Screening of Alternatives”, Ind. Eng. Chem. Process Des. Deu. 1985, 24, 970. Fisher, W. R.; Doherty, M. F.; Douglas, J. M. “Evaluating Significant Economic Tradeoffs for Process Design and Steady-State Control Optimization Problems”, AIChE J . 1985, 31, 1538. Fisher, W. R.; Douglas, J. M. “Evaluating Process Operability at the Preliminary Design Stage”, Comp. Chem. Eng. 1985, 9, 499. Fisher, W. R.; Doherty, M. F.; Douglas, J. M. “The Interface Between Design and Control”, Ind. Eng. Chem. Res. 1987, in press. Kirkwood, R. L.; Locke, M. H.; Douglas, J. M. “An Expert System for Synthesizing Flowsheets and Optimum Designs”, Comp. Chem. Eng. 1987, in press.
Linnhoff, B.; Vredeveld, D. R. “Pinch Technology Has come of Age”, CEP 1984, 80(7), 33. Linnhoff, B.; Townsend, D. W.; Boland, D.; Hewitt, G. F.; Thomas, B. E. A.; Guy, A. R.; Marsland, R. H.; “A User Guide on Process Integration for the Efficient Use of Energy”, Inst. Chem. Eng. 1982, 1.
McKetta, J. J., Ed. “Encyclopedia of Chemical Processing and Design”, Marcel Dekker: New York, 1977; Vol. 4, p 182. Sacerdoti, E. D., “Planning in a Hierarchy of Abstraction Spaces”, Artif. Intelligence 1974, 5, 115. Steinmetz, F. J.; Chaney, M. D. “Total Plant Energy Integration”, Presented at the Spring National AIChE Meeting, Houston, March 24-28, 1985; Paper 88d. Terrill, D. T.; Douglas, J. M. “Heat Exchanger Network Analysis. 1. Optimization”, Ind. Eng. Chem. Res. 1987, 26, 685. Tjoe, T. N.; Linnhoff, B. “Using Pinch Technology for Process Retrofit”, Chem. Eng. 1986, April 28, 47. Townsend, D. W.; Linnhoff, B. “Surface Area Targets for Heat Exchanger Networks”, Annual Meeting of the Institute of Chemical Engineers, Bath, U.K., April 1984. Westerberg, A. W. ”The Role of Expert System Technology in Design”, Paper presented at the International Symposium on Chemical Reaction Engineering, Philadelphia, May 18-21, 1986; ISCRE 9. Witherell, W. D.; Linnhoff, B. “Pinch Technology Retrofit: A Complex Industrial Application”, Presented at the Spring National AIChE Meeting, Houston, March 24-28, 1985; Paper 88b. Received for review August 15, 1986 Revised manuscript received June 3, 1987 Accepted June 27, 1987
Cocracking and Separate Cracking of Ethane and Naphtha Patrick M. Plehiers and Gilbert F. Ftoment* Laboratorium voor Petrochemische Techniek, Rijksuniversiteit te Gent, B-9000 Gent, Belgium
This paper presents experimental data on the thermal cracking of a naphtha-ethane mixture in a pilot plant, under conditions representative of industrial operation. The effects of the interaction between ethane and naphtha on the ethane conversion and kinetics and on the product distribution are investigated. When naphtha, ethane, and the mixture are cracked under an equal molar dilution, the cocracking yields and selectivities can be quantitatively predicted from the separate cracking data, except for hydrogen, methane, and high molecular weight products. The combined effects of cocracking and partial pressure, occurring when naphtha, ethane, and the mixture are cracked under an equal weight dilution, are such that, if maximum olefins selectivities are desired, separate cracking is to be preferred. Some aspects of coke formation are addressed as well. Introduction Literature data on naphtha-ethane pyrolysis are scarce and very incomplete. Most authors base their conclusions upon one single data point. de Blieck and Goossens (1971a,b)observed that cocracking with naphtha increased, for identical reaction conditions, the ethane conversion. From a comparison of “typical” product distributions for cocracking and separate cracking, Mol (1981) concluded that the interaction between ethane and naphtha in the cracking of a mixture containing about 24% by weight ethane leads to an enhanced ethylene selectivity, enabling a 2 YO savings in naphtha consumption. Propylene and butadiene selectivities, however, were markedly reduced. Nowowiejski et al. (1982) cracked naphtha with ethane in an industrial millisecond furnace. In addition to the millisecond effects, they found the methane and ethylene yields to be favored by cocracking; the C,+ and butenes yields were found to be reduced. No accelerating effect of the naphtha on the ethane cracking was noticed in this case. A considerable influence of the naphtha composition on the deviations of the cocracking yields from additivity 0888-5885/87/2626-2204$01.50/0
was observed, but no further details are given on this issue, however. Clearly, until now, no thorough study of naphtha-ethane cocracking has been published. The work reported in the present paper aimed at a better understanding of the interaction between naphtha and ethane during pyrolysis. I t was investigated how the interaction alters the overall ethane cracking kinetics and the product distributions. Experiments were performed under conditions close to those encountered in industrial practice. Statement of the Problem The aim of a study on cocracking of hydrocarbons is to compare the cocracking yields with those resulting from the mixing of the effluents of separate cracking. Quite often in the literature, “identical reaction conditions” are chosen as a basis for this comparison. Since different hydrocarbons demand completely different operating conditions, comparing yields at equal conditions does not seem very appropriate. Moreover, identical reaction conditions do not guarantee the naphtha and ethane con0 1987 American Chemical Society
Ind. Eng. Chem. Res., Vol. 26, No. 11, 1987 2205 Table I. Partial Pressures in Ethane, Mixture, and Naphtha Cracking under the Same Weight and Molar Dilution (COP = 2 bar, 0 = 0.55 s) exit dilution, inlet pHC, ethane mixture 50/50 mol % naphtha
ethane mixture 50/50 mol 7'0 naphtha
kg/kg 0.288 0.288 0.288
bar 1.44 1.15 0.91
dilution,
inlet pHC, bar 1.15 1.15 1.15
1.000 1.000 1.000
Table 11. Naphtha Characterization av M, g/mol H-to-C molar ratio, mol/mol specific gravity a t 20 OC, g/cm3 ASTM distillation IBP 10% 50% 90% 94% residue loss
37 "C 62 OC 112 "C 153 "C 161 O C 2% 4% (important C4-C6 fraction)
0.46 0.35 0.26
Pc&
PC&
0.01 0.09 0.10 exit
0.01 0.02 0.02
PC$( PC&
PC&
0.37 0.35 0.33
0.01 0.02 0.03
0.01 0.09 0.13
Table 111. Settings of the Most Important Process Variables mixture ethane naphtha B C D E A set 100.0 100.0 naphtha, w t % 76.2 23.8 100.0 100.0 ethane, wt % 0.600 0.288 0.192 0.288 0.288 6, k g / k 6, mol/mol 1.000 1.000 0.480 1.000 1.538 0.55 0.55 0.55 0.55 0.55 0, s 2.00 2.00 2.00 2.00 2.00 COP, bar 750-850 750-880 800-850 750-850 800-850 COT, "C 1
Yieia
w a iw?
I,*)
ETHANE
96.13 2.153 0.7134 PIONA analysis n-paraffins isoparaffins naphthenes aromatics olefins unknown
34.01 wt % 35.82 wt % 21.51 wt % 7.48 wt % 0.07 w t % 1.11 wt %
versions in separate cracking to be equal to the corresponding conversions in mixture cracking. In this way, yields or selectivities are compared at different mixture conversions, and this can only lead to biased conclusions. The only basis from which meaningful conclusionscan be drawn is comparison at equal conversions (Froment et al., 1977,1979); in other words, both the naphtha and ethane conversions in separate cracking should be equal to the conversion of these components in mixture cracking. When naphtha, ethane, and a mixture are cracked under an equal weight dilution, the large differences in molecular weight and expansion between naphtha and ethane lead to quite different hydrocarbon and product partial pressure profiles, as is shown in Table I for a 50/50 molar mixture. It is well-known that the partial pressure profile has an important influence on the cracking yields and selectivities. In studying the cocracking and separate cracking of ethane and naphtha under equal weight dilution, the cocracking effects may be obscured by partial pressure effects. When the three feedstocks are cracked under an equal molar dilution, the partial pressure effects are far less pronounced, as is illustrated in Table I. In this work, the effect of both types of dilution was investigated.
Experimental Program The experimental work was done in the laboratory's pilot plant for thermal cracking. The unit has been described by Van Damme and Froment (1982). The evaporation and preheating section of the furnace was modified to achieve mixing of ethane and naphtha in the gas phase. The analytical equipment and the calibration procedures were described by Dierickx et al. (1986). The ethane was a commercially available, high-purity product, containing only 0.5 wt % impurities, mainly methane and ethylene. The naphtha was a full-range naphtha. Its characterization is given in Table 11. Five sets of experiments were performed. The settings of the most important process variables are listed in Table 111. First, ethane, naphtha, and a 50/50 molar mixture were
ETHANE C O P , 2 bar
0
~rmilene iropylene
3
llelhane
A
nelhane
6 .0 600 6 , 0 288 6 .0600 6 . 0 288
kg 'LQ
4
kg! hg kg!xg kg/kg
2 54
Figure 2. Methane and propylene yields from ethane cracking.
cracked under an equal weight dilution of steam of 0.288 kg/kg (sets A, C, and E). Comparing these product distributions yields the combined effects of cocracking and partial pressure. Then, ethane, naphtha, and the same mixture were cracked under an equal molar dilution of 1.0 mol/mol (sets A, B, and D). Comparison of the product distributions obtained from these experiments allows the effect of cocracking proper, i.e., free from partial pressure effects, to be determined.
Experimental Results For pure ethane cracking (sets B and C), the yields of the most important C4- products (CHI, C2H4,C3H6,and C4H6)and of cyclopentadiene, benzene, and toluene are presented as a function of the ethane conversion in Figures 1-3. For pure naphtha cracking (sets D and E), the same product yields are plotted vs. the naphtha conversion in Figures 4-6. The naphtha conversion was calculated as a weighted and normalized sum of the conversions of the individual naphtha components (Van Damme et al., 1981). The key components account for 82 mol % of the naphtha
2206 Ind. Eng. Chem. Res., Vol. 26, No. 11, 1987 ",el0
13
i
,
, *.'.
1 ETHANE
5
COP
.i
1
bar
~
3
L
5 . 0Eoc *ikD
benzene
ciCIw?lad#ene
E
0286 W h g
6
C600 hg hg
,
'
/
-
i
I r 6
/
i
i-
, ,
2 1ar
i3p
6
0
cycbpenI6dere
b
mzwe
5
0192 hg kg
c
tolaw
6
0 192 hg hg
0192 kg Lo
L
/
Figure 3. Cyclopentadiene, benzene, and toluene yields from ethane cracking.
1
CONVERSlOld
YS
I
EQUIVALENT SPACE -TIME
0 -.={
100
'INM
Xki
L
EO
8L
1oC
Figure 4. Methane and ethylene yields from naphtha cracking.
II
COP
2
0
iiwhlha
6
0
naphtha
6
154
A
ethane
6
106
A
ethane
6
047
100m0klm011
vE,Fol
bar
lbar, ISlmOIe
I
I
150
100
50
-1
Figure 7. Conversions vs. equivalent space-time.
,
,
t
50-
80
60
",,a
t
/ m'
IY"/,
NAPHTHA
COP 0
.2
7 6 2 ETHANE 02.
6.
238
*t
/ '
0 2 8 8 i; ' 2
eflyels
?DO
Figure 5. Propylene and butadiene yields from naphtha cracking.
feed. The relation between the naphtha conversion and the equivalent space-time, VE/Fo,is presented in Figure 7. The mixture conversion was defined as XM
= +NXN,M + +EXE,M
(1)
This conversion, together with the naphtha and ethane conversions in the mixture, is also presented vs. V E / F 0in Figure 7. Product distributions for mixture cracking (set A) are given in Figures 8-10. Ethane Conversion in Cocracking As mentioned in the Introduction, some authors observed a higher conversion of ethane in cocracking than in pure ethane cracking at the same reaction conditions.
4
t
KO
80
7c
10
M
90 Bo
96
4
'ia XL
Ind. Eng. Chem. Res., Vol. 26, No. 11, 1987 2207
I M
Dropylene
A
butadiene
30
60 0
6
0
50
40
eo
70 4
0
XF M
98
43 e
a
,
O
%M XM ,
Figure 9. Propylene and butadiene yields from mixture cracking.
--c 2C
I
40
40
3 .
80
70
60
50
60
YE M
50
93
98 70
XNM
X
Figure 10. Cyclopentadiene, benzene, and toluene yields from mixture cracking.
Figure 11 shows that at low severities, corresponding to a low average temperature level in the reactor, the ethane conversion in cocracking is higher than in pure ethane cracking at equal conditions. At high severities, corresponding to high temperatures, the ethane conversion in cocracking is lower than in pure ethane cracking. In the former case, ethane cracking is accelerated by the high concentration of active chain carrying radicals that are generated from the naphtha at moderate temperatures. In the latter case, when the naphtha is converted for more than 97%, ethane no longer profits from the positive effect of the presence of the naphtha, but instead its conversion is hindered by the reaction products, mainly olefins, which consume rather than produce radicals. This shows that the temperature ranges in which naphtha and ethane crack do not sufficiently overlap in the high conversion range, so that the naphtha cannot produce enough radicals at the temperature level at which they are most useful for an enhanced pyrolysis of ethane. Hence, all changes in process variables that require higher temperatures to achieve the same naphtha conversion (e.g., high dilution, low total pressure, short residence time, and light naphtha) favor ethane conversion. On the other hand, any changes in process variables that widen the gap between the temperature ranges for the cracking of the liquid feedstock and that for ethane (e.g., cocracking of ethane with gas oil) disfavor the ethane conversion and lead to a high and eventually an excessive ethane recycle. From Figure 11, it is clear that naphtha-ethane cocracking does not allow high ethane conversions to be reached. Even for obtaining a moderate 60% conversion, very severe conditions are required.
750
800
839
880 aDOlOXCOT
850
c
Figure 11. Ethane conversion in ethane and mixture cracking.
Kinetics of Ethane Conversion in Cocracking The kinetics of the overall disappearance of ethane in pure ethane cracking and in cocracking with naphtha were determined by the integral method of kinetic analysis. The equivalent space-time (see Figure 7) was used to reduce the data to isothermal and isobaric conditions. The method has been discussed in detail by Froment et al. (1961) and Van Damme et al. (1975). It has previously been shown that the order of the ethane disappearance is 1. For the cracking of pure ethane at low conversions (corresponding to temperatures below 1050 K), an activation energy of 263.7 kJ/mol and a frequency factor of 1.82 X 1013s-l were obtained, in excellent agreement with 259.3 kJ/mol and 1.03 X 1013s-l determined by Froment et al. (1976). For ethane conversions exceeding 30%, obtained at higher temperatures, some retardation of pure ethane cracking through the products occurs. For the activation energy, a lower value of 200.9 kJ/mol was found, since the effect of the reverse reaction was not accounted for in the present calculations. The frequency factor was 1.38 X 1O1O s-l. The kinetic parameters for the global disappearance of ethane in a mixture with naphtha are E = 178.7 kJ/mol and A = 1.20 X lo9 s-l. The low value of the activation energy supports the observations made in the previous paragraph. Comparison of Product Distributions The cocracking effluent composition can be compared with the composition of the product stream that would originate from the mixing of the separate cracking effluents in terms of either yields or selectivities. Mixed effluent product yields are calculated by using a simple additivity rule Yi,Madd
=
Y;,EwE
+ Y~,NwN
(2)
Product selectivities in the mixed effluent of separate cracking have to be calculated from their definition Si,Madd
=
(moles of i formed from the cracking of ethane + moles of i formed from the cracking of naphtha)/ (moles of ethane cracked + moles of naphtha cracked) (3) Si,Madd
=
Si,EXE+E XE+E
+ 8,NXN+N + XN+N
(4)
In (4), the selectivities in pure component cracking can be taken at the conversions reached in mixture cracking. This accounts to a certain extent for the interaction, and a
2208 Ind. Eng. Chem. Res., Vol. 26, No. 11, 1987 Table IV. Deviations of the Cocracking Yields from the GK Yields (Cocracking Minus GK in wt '70 Absolute; COP = 2 bar, 0 = 0.55 8.6 = 0.288 k d k d high low conversion, conversion product XE,M = 40% XE,M = 55% hydrogen -0.20 -0.35 methane +1.20 +1.80 ethylene -0.50 -0z30 propylene -1.30 -0.90 butadiene -0.20 -0.25 butenes -0.40 -0.30 + + cyclopentadiene + + benzene
Yield l r r t b
HYDROGEN
6
COP 2 bar
1 mle
mole
0 mixIUle
4
naphtha
A
-
/
ethane
,/L
-GK
METHANE
field l w t % l
COP
2 bar
6
1 moleimoS
0 mixture
20
I
naphtha
P elhsne
/*
~GK_ .
0
20 50
70
Po
M
40
xf
70
60 90
98
x,
Figure 12. Hydrogen yield comparison at equal molar dilution.
better approximation of the cocracking selectivities is obtained. S,E(XE,M)XE,M+E + %,N(XN,M)XN,M+N S L , M G K=
(5)
+ XN,M+N This way of predicting cocracking selectivities was called "global kinetics based" (further formally denoted as GK selectivities) by Froment et al. (1977). Any step beyond this would require a detailed reaction scheme. A similar approach can be taken for predicting cocracking yields; i.e., in (2)) the yields Yi,E and Yi,N can be taken at xE,M and xN,M, respectively. Hence, yields calculated in this way will be called further global kinetics based yields (GK yields). The GK yields are related to the GK selectivities like the true cocracking yields to the true cocracking selectivities: XE,M+E
MF =-= Y c , ~ G K MixM
Si,MGK
SL,M" YC,~"
5
/ *' 4
*
- A3d"A / A
0
50 70
70
60 80
'h 8a
90
98
%
Figure 13. Methane yield comparison at equal molar dilution. Yield i w t %
1
ETHYLENE COP 2 bar
6
1 mole'mole
50
4A
(6)
MFis the average molecular weight of the naphtha-ethane feed mixture, Mi is the molecular weight of the product, the selectivity of which is being considered, and xMis the mixture conversion. Because of (6), comparing cocracking yields with GK yields is completely analogous to comparing cocracking selectivities with GK Selectivities. If the experimental cocracking yields agree with the GK yields, the interaction between naphtha and ethane is limited to an effect on the conversions. A deviation of the experimental cocracking yields from the GK yields means that there is a strong interaction that cannot be accounted for through global kinetics. Deviations of the cocracking yields from the GK yields obtained from cracking under equal weight dilution are compared in Table IV for a number of important products. For hydrogen, methane, and the C3+products, the deviations are quite pronounced. This may lead to the precipitous conclusion that the cocracking of ethane and naphtha leads to serious interaction between the reacting species, significantly affecting the product spectrum. Comparison at equal weight dilution, even at the same
/
3
20
70
80
M
40 BO
50
90
65 s;yh
10
98
+i
Figure 14. Ethylene yield comparison at equal molar dilution.
conversions, still includes the effect of partial pressure on the selectivities, however. This is not included in the GK
Ind. Eng. Chem. Res., Vol. 26, No. 11, 1987 2209 I PROWLENE
YldU I W t % l
I;
COP
E
6
2 bar
Yleu 1wt.h)
BENZENE
1 molehole
COP
:
2 bar
,
6
:
1 moleimole
/
8
i
mixture
@
I
5
naDhtha
Figure 15. Propylene yield comparison at equal molar dilution.
" x,
d
2p
Yield
II
BUTADIENE
iWt%i
COP
2
2 bar
6
:
1 moleimole
L
-
b
/--&
50
70
80
W.
70
80
90
ep
a
~
8 8 h
Figure 17. Benzene yield comparison a t equal molar dilution.
NAPHTHALENE
COP 2 b a r 6 4 moleimole
-3
0
mixture naDhtha
___
GK
2
40
3L
60
50 70
. ,.
/ v
20
80
50
90
60
8;8
70
98
XNM
Figure 16. Butadiene yield comparison at equal molar dilution.
prediction. Only comparison of the product yields and selectivities obtained from cracking under equal molar dilution can yield the true cocracking effect. The cocracking yields of hydrogen, methane, ethylene, propylene, butadiene, benzene, and naphthalene are compared with the GK yields in Figures 12-18. It now appears that the deviations from the GK values are generally small (about 0.2 wt % absolute), except for high molecular weight products like naphthalene. For those products that are mainly formed through hydrogen abstraction reactions, the deviations are large as well, since global kinetics cannot predict deviations in the concentrations of the abstracting radicals, because they insufficiently reflect the true radical mechanism of the cracking. Depending upon the conversion, the cocracking yield of hydrogen is between 0.20 and 0.35 wt % lower than the GK yield. The cocracking methane yield is between 1.20 and 1.85 wt % higher than the GK yield. These deviations indicate that the H' and the CH3' radical concentrations are strongly affected by cocracking and can be explained as follows.
60
50 70
80
80
70
90
88
XNh
Figure 18. Naphthalene yield comparison a t equal molar dilution.
Most H' radicals are formed through hydrogen abstraction on ethane, followed by the decomposition of the ethyl radical. When a H radical abstracts a hydrogen from ethane, the radical is almost always regenerated through the subsequent decomposition of the ethyl radical. When a H radical abstracts a hydrogen from a naphtha molecule, the radical is rather seldom regenerated. Instead, a CH3' radical is formed. The abstraction of a hydrogen atom from a naphtha molecule by a CH< radical leads, in most cases, to the regeneration of the radical. Hydrogen abstraction from ethane by a CH< radical yields a H'radical instead. Since most of the abstractions take place on naphtha molecules, it may be expected that CH3' radicals are generally regenerated and that H'radicals are in most cases replaced by a CH,' radical. Consequently, the hy-
2210 Ind. Eng. Chem. Res., Vol. 26, No. 11, 1987
I w
".
'y
3-'-
'E
49
I 1
I
Table V. Yield and Selectivity Deviations from the GK Values in the Cocracking of Ethane and Naphtha (Equal Conversions, Cocracking Minus GK) equal molar product dilution equal wt dilution effect of hydrogen abstracting methane + + radicals concn ethylene small uncertain but small HC partial propylene small pressures butadiene small small butenes small + cyclopentadiene small + benzene Table VI. Comparison between Separate and Cocracking for an Industrial Case (6 = 0.288 kg/kg, COP = 2 bar, 0 = 0.55 s) separate re1 % diff cracking cocracking naphtha feed rate 1667 1648 -1.14 total HC feed rate 1809 1801 -0.44 naphtha conversion 98 % 99 70 +1.02 ethane conversion 60 % 55% -8.33 ~~~
'
ETHYLEYE
Cop
oar
I
Figure 19. Partial pressure dependency of the ethylene yield.
Production in 103T/year hydrogen methane ethylene propylene butadiene benzene
.-
-~~
pROPYLEhE
COP
-
e.hwe
13
7
L
aa
,
, 'HC bar)
j
Figure 20. Partial pressure dependency of the propylene yield.
drogen selectivity is decreased, while the methane selectivity is increased. This is valid for any type of naphtha cocracked with ethane, when the naphtha content in the feed is sufficiently high. From Figures 14-17, it may be concluded that the cocracking of ethane and naphtha, in the high conversion range that is of practical interest, has very little effect (about 0.2 wt 70)on olefins and aromatics yields (and relatively an equal effect on the corresponding selectivities), provided that the effects of conversion, which is a part of the interaction that is already accounted for by the use of the GK yields and selectivities, and of residence time and partial pressure are accounted for. In other words, the naphtha-ethane cocracking yields and selectivities can be quantitatively predicted by means of the GK formulas. This also means that partial pressure effects, but not cocracking itself, are responsible for the yield and selectivity deviations that were observed with cocracking under equal weight dilution. The results, obtained from cracking under equal weight dilution, can now be looked a t from a different focus. Figures 19 and 20 show the evolution of respectively the ethylene and propylene yields from naphtha and ethane separate cracking at given conditions ( x N = 97%, xE = 49%) and residence time (0 = 0.55 s) as a function of the hydrocarbon inlet partial pressure or, in other words, of the molar dilution. The experimental cocracking and GK
19.5 271.0 500.0 242.5 72.0 111.0
17.0 288.0 500.0 222.0 67.5 117.0
-12.82 +6.27 -8.45 -6.25 +5.41
yields for the corresponding mixture conversion of 73% ( x ~=, 9770, ~ x E , M = 49%) are represented as well. It is again illustrated that the GK yields closely agree with the experimental cocracking yields, when naphtha, ethane, and the mixture are cracked under an equal molar dilution and at the same conversions. In the case of cracking under equal weight dilution, only the ethylene cocracking yield can be predicted by the GK formula. It follows from these figures that the dependency of the yields in naphtha and ethane pyrolysis upon the hydrocarbon partial pressure determines how the cocracking yields (and, by analogy, selectivities) deviate from the GK yields obtained from cracking under equal weight dilution. Except for ethylene, the partial pressure dependency of the product yields is much more pronounced for naphtha than for ethane cracking, whatever the naphtha composition may be. It can, therefore, be concluded that the cocracking under equal weight dilution of ethane with any naphtha results in yield (selectivity) deviations from GK yields (selectivities) as given in Table V. In other words, if maximum propylene, butenes, and butadiene selectivities are aimed at, separate cracking is to be preferred. Practical Considerations It was shown in the previous paragraphs that the cocracking of ethane and naphtha has a rather negative influence on the product selectivities. Moreover, cocracking cannot be applied to reach high ethane conversions, e.g., to reduce the recycle rate of ethane. Consequently, cocracking can only be advantageous if it allows the naphtha conversion (cracking severity) to be raised with respect to pure naphtha cracking. The higher naphtha conversion then leads to higher ethylene, cyclopentadiene, and aromatic yields, as is illustrated in the following example. First, the yield pattern of a conventional naphtha cracking plant, in which recycle ethane is cracked in separate furnaces, is investigated. Second, part of the naphtha feed is cocracked with the recycle ethane under the same
Ind. Eng. Chem. Res., Vol. 26, No. 11, 1987 2211 weight dilution, residence time, and total pressure as the naphtha but at a higher severity (COT 850 "C instead of 830 "c,XNY 99% instead of 3cN 98%). Table VI shows the results. The same amount of ethylene can be produced by cocracking, using about 1%less feedstock. More aromatics are formed as well, while the production of propylene, butenes, and butadiene is decreased. The limiting factor to such an increase in conversion, however, is the coke yield. Therefore, two additional experiments were performed to investigate the coke formation during the pyrolysis of the naphtha and the mixture under the conditions mentioned in the above example. The reaction conditions were kept constant for 6 h. Afterward, the coke was burned off with a controlled flow of air. At regular time intervals, the carbon oxides were measured with the Carle gas chromatograph. Since the surface-to-volume ratio and the internal wall temperature of the pilot plant differ from those encountered in industrial reactors, the coking data cannot directly be extrapolated to industrial operation, but at least, a trend is indicated. During the cracking of the naphtha (6 = 0.288 kg/kg, 8 = 0.55 s, COP = 2 bar, COT = 830 "C), 0.41 g of carbon is formed per kilogram of hydrocarbon fed. In mixture cracking (6 = 0.288 kg/kg, 8 = 0.55 s, COP = 2 bar, COT = 850 "C), the carbon deposit is almost quadrupled and amounts to 1.60 g per kilogram of hydrocarbon fed. Because of the increased coking, raising the naphtha conversion in cocracking by increasing the temperature is to be rejected. Since the selectivities in cocracking are also less favorable, separate cracking is to be preferred.
Conclusions In the present work, the cocracking of one particular naphtha with ethane has been studied. We arrived at the following conclusions. Cocracking does not allow high ethane conversions to be reached, since the ethane cracking is strongly hindered by the unsaturated reaction products of the naphtha. If ethane, naphtha, and the mixture are cracked under an equal molar dilution, and if the naphtha and ethane conversions in separate cracking correspond to the respective conversions in cocracking, the cocracking yields and selectivities can be quantitatively predicted by means of the GK formulas, except for those products that are mainly formed through hydrogen abstraction reactions (hydrogen and methane) and for high molecular weight products. This means that the major part of the interaction between naphtha and ethane in cocracking is accounted for by the GK formulas. If ethane, naphtha, and the mixture are cracked to corresponding conversions under an equal weight dilution, the combined effects of cocracking and partial pressure result in an adverse deviation of the propylene, butenes, and butadiene selectivities from the GK calculated values. If these are the desired products, separate cracking is to be preferred. The formation of typical secondary products like cyclopentadiene and benzene, on the contrary, is favored. This conclusion is valid for the cocracking of ethane with any naphtha. For the naphtha studied here, the combined cocracking and partial pressure effects adversely influenced the ethylene selectivity as well. This cannot be generalized, however. The ethylene yield is increased by cocracking when the naphtha conversion is raised. It is accompanied by a considerable increase in coke formation.
Acknowledgment Plehiers is grateful to the Belgian Nationaal Fonds voor WetenschappelijkOnderzoek for a Research Assistantship.
Nomenclature A = frequency factor, s-l for a first-order reaction C', = products with m or more carbon atoms C,- = products with n or less carbon atoms COP = coil outlet pressure, bar COT = coil outlet temperature, O C
E = activation energy, kJ/mol F = molar flow rate, mol/s M = molecular weight, g/mol p = pressure, bar S = selectivity VE = equivalent reactor volume, L w = weight fraction x = conversion Y = yield Greek Symbols 6 = dilution, kg/kg or mol/mol $ = mole fraction 0 = residence time, s Subscripts
E = ethane i = component M = mixture or in mixture N = naphtha 0 = initial Superscripts add = additive co = in cocracking GK = on basis of the GK formulas Registry No. Ethane, 74-84-0; ethylene, 74-85-1;butadiene, 106-99-0;propene, 115-07-1;methane, 74-82-8; benzene, 71-43-2; cyclopentadiene,542-92-7; toluene, 108-883;hydrogen, 1333-74-0; naphthalene, 91-20-3.
Literature Cited de Blieck, J. L.; Goossens, A. G. Neue Verfahren in der Chemischen Technik, Dechema Monographien Band 68; Verlag Chemie: Weinheim, FRG, 1971a. de Blieck, J. L.; Goossens, A. G. Hydrocarbon Process. 1971b,50(3), 76-80. Dierickx, J. L.; Plehiers, P. M.; Froment,G. F. J. Chromatogr. 1986, 362,155-174. Froment, G. F.; Pijcke, H.; Goethals, G. Chem. Eng. Sci. 1961,13, 173-179. Froment, G. F.; Van de Steene, B. 0.;Van Damme, P. S.; Narayanan, S.; Goossens, A. G. Ind. Eng. Chem. Process Des. Dev. 1976,15, 495-504. Froment, G. F.; Van de Steene, B. 0.;Vanden Berghe, P. S.; Goossens, A. G. AIChE J . 1977,23,93-106. Froment, G. F.; Van de Steene, B. 0.;Sumedha,0. Oil Gas J . 1979, 77(16), 87-90. Hougen, 0.A,; Watson, K. M. Chemical Process PrinciplesKinetics and Catalysis Vol 111; McGraw-Hill: New York, 1947. Mol, A. Hydrocarbon Process. 1981,60(2),129-131. Nowowiejski, G. B.; Petterson, W. C.; Kii, T.; Suwa, A. Chem. Eng. Prog. 1982,78(5), 49-51. Van Damme, P. S.; Narayanan, S.; Froment, G. F. AIChE J . 1975, 21,1065-1073. Van Damme, P. S.; Froment, G . F.; Balthasar, W. B. Ind. Eng. Chem. Process Des. Dev. 1981,20,366-376. Van Damme, P. S.; Froment, G. F. Chem. Eng. Prog. 1982,78,77-82.
Received f o r review May 1, 1986 Accepted June 23, 1987
