Dehydrocyclization in Platforming G. R. DONALDSON, L. F. PASIK,
AND VLADlMlR HAENSEL
Universal Oil Products Co., Riverside, 111.
A
hydrogenated Fischer-Tropsch naphtha consisting of n-paraffins (CSH18 through CllH24) was Platformed at various conditions to obtain data showing the effect of temperature, pressure, and space velocity on the product distribution. Detailed aromatic yields were obtained and the quality of the nonaromatic Ca+ raffinate was determined. Aromatic production by dehydrocyclization of paraffins proceeds readily and i s favored by higher temperatures, lower pressures, and lower space velocities. A 97.4 F-1 clear octane number C,+ Platformate was produced from the -50 octane number charge. The aromatics produced were shown to be distributed in agreement with the equilibrium data.
T
HE C O P Platforming process ( 1 ) is widely utilized for the upgrading of straight run gasolines and for the production of benzene, toluene, and xylenes. One of the major differences in the feed stocks for the gasoline producing units is the naphtheneparaffin ratio, which has been found t o vary from about 0.25 t o about 6. T h e lower ratios are associated with some of the Michigan, Pennsylvania, East Texas, and Kuwait stocks while the higher ratios are generally exhibited by California, Gulf Coast, and Venezuela gasolines. Since the Platformates derived from widely different charging stocks do not show a large variation in aromatic content, i t follows t h a t the conversion of paraffins to aromatics by dehydrocyclization takes place t o a greater extent with t h e naphthene lean stocks. For some time i t has been believed that dehydrocyclization of paraffins does play a n important role even with higher naphthene content stocks and t h a t during the course of reaction some of the naphthenes undergo ring opening t o produce paraffins, while some of the paraffins undergo ring closure and, subsequently, form aromatics. One of the most elegant ways of proving the extent of each of these reactions in a mixture of hydrocarbons would be by means of tracer technique. However, considerable information can be obtained by selecting a mixture of specific hydrocarbons and subjecting the mixture to*Platforming under a variety of conditions. Although a mixture of Cs-C11 n-paraffins represents only a small part of the usual Platforming charge
Table I.
Charge Stock
Composition, lis. vol. yo Paraffins Olefins Naphthenes Aromatics Composition by fractionation (Figure n-C7Hie and lighter n-CsHia n-CrHzo n-CloHzz n-CulIzr n-CizH1c and heavier Properties F-1 clear octane number Sulfur, turbidimetric, wt. yo Grayity OAPI Q BO0 F. Specific 'gravity @J 60° F. Engler distillation, F. Initial b. p .
3 2 90% E3point Average molecular weight
100
. o
0 0
liq. vol. yo
2 15 43 29 10 1
- 50a 0.0002 64.0 0.7238 2158
288 310 341 352 369 130
-
a Blend of 50% naphtha and 50% of (IO octane number straight run rated 5.0 F-1 clear.
April 1955
stock, i t can be considered as becoming representative of t h e paraffins normally present because of the considerably greater rate of isomerization as compared to the rate of dehydrocyclisation. Thus, a Fischer-Tropsch naphtha of the desired boiling range has been selected as P feed stock for Platforming at a variety of conditions to determine the extent of dehydrocyclization and the nature of the products.
Apparatus and Procedure The charge stock was prepared from a Fischer-Tropsch naphtha product received from Germany. A mild hydrogenation pretreatment was used t o saturate the olefins and t o convert t h e oxygen containing compounds t o paraffins. The resulting change had the properties and the composition shown in Table I and Figure 1. The charge was processed in a bench scale type Platforming unit equipped with a recycle gas compressor and a continuous debutanization column. I n this unit t h e platinum-aluminahalogen catalyst is contained in a stainless steel reaction tube equipped with a n axial thermowell which contains a movable thermocouple. T h e reaction tube is maintained at the desired temperature by means of a n electrically heated block furnace. The hydrocarbon charge is pumped to t h e top of the reaction tube where i t mixes with the hydrogen rich recycle gas and the combined stream is preheated in an internal, spiral section in the reaction tube. The effluent from the reaction tube is cooled to about 60' F. and collected in a high pressure rereiver. The liquid phase is charged to a debutanization column where the C4 and lighter hydrocarbons and dissolved hydrogen are separated from the Cg and heavier Platformate. This fractionation system functions very effectively to produce a butane-free Platformate while retaining more than 97% of the Cs charged to the debutanizer. This amounts to a retention of more than 94% of the total Cs produced. A part of the recycle gas stream from the high pressure receiver is vented by a pressure control system while the remainder is recycled t o the reactor. The compositions of the two gas streams, the excess recycle gas and the debutanizer overhead gas, were determined by mas3 spectrometer analysis. The debutanized Platformate composition wasdetermined by precision fractionation and mass spectrometer analysis through the Cg fraction. The maximum Cq content detected in t h e series of tests was 0.1 wt. %. I n two tests the Ce+ Platformate was separated into formatic and nonaromatic fractions b y silica gel adsorption, and the composition of the aromatic fraction through the Cs isomers was determined by subsequent fractionation and infrared analysis.
INDUSTRIAL A N D ENGINEERING CHEMISTRY
731
ENGINEERING, DESIGN, A N D PROCESS DEVELOPMENT 220
I
I
I
I
I
I
nLC12H&--
I
Table II.
Effect of Temperature at Constant Pressure and Space Velocity
Temperature, C. Yields Hydrogen, std. cu. ft./bbl. of charge Dry gas (Ci Ca), std. cu. ft./bbl. of charge Butanes, lig. vol. 7 of charge Pentanes, liq. vol. of charge Ca+ Platformate, liq. vol. Yo of charge Aromatics, wt. Yo of charge
-
%
160
140
Compositions Liq. vol. % ino in butanes Liq. vol. % is0 in pentanes
I20
L V % OVERHEAD
80
1
I
0
Figure 1.
I
1
20
IO
30
Jo
40
I
80
70
60
90
100
True boiling point charge stock data I
I
I
I
I
I
hT I A
+$
601
I
I
-50
C 5 '
Figure 2.
I
I
0 FLATFORMATE
I
I
2 0 4 0 6 0 8 0 l a , F - l CLEAR OCTANE NUMBER
C6+Platformate I
yield octane range
I
I
I
T
I
/
I
C z
PLATFOMTE
Figure 3.
F-l
CLEAR CClANE
NUMBER
Aromatic yield range
Discussion of Results The effect of operating conditiqns throughout a wide range has been investigated. T h e weight hourly space velocity( WHSV) was varied between 2 and 8, the pressure between 350 and 700 pounds per squareinchgageand the temperature between 450' and 500°C. T h e effects of each of these variables are shown by the data in Tables 11,111,and IV. I n Table I1 the effect of a temperature variationfrom 450' to 500' C. is shown while Tables I11 and IV show the effects of pressure variation and WHSV variations within the abovelimits. The tabulations present the product distributions and the Platformate properties. Figures 2 and 3 show the ranges of the Cj+ Platformate yield and of the aromatic production obtained at given Ca+ Platformate octane numbers. These results show t h a t rather extensive aromatic productions by dehydrocyclization of paraffins takes place over the Platforming catalyst at pressures as high as 700 pounds per square inch gage. They also show t h a t a 90 F-1clear octane number Cs+ Platformate is readily produced from the -50 octane number charge
732
Properties of Cs+ Platformates F-1 clear octane number F-2 clear octane number Aromatic content, w t . % Gravity, OAPI @ 60' F. Specific gravity @ 60' F. Engler distillation, F. @d Initial b. p.
10%
jo% 90% 95% End point
450
500
572 278 10.7 10.8 78.1 30.9
816 447 16.3 14.3 64.9 43.9
30 48
37 56
74.4
97.4 84.2 63.2 49.0 ,7839
...
38.3 57.1 .7503 124 165 262 33 1 346 382
-
110 136
263 340 356 423
stock. T h e aromatic productions obtained compare favorably with the aromatic productions realized in the usual Platforming operations (Figure 9). The upgrading of the parafins by isomerization and hydrocracking is substantial. When the nonaromatic portion is separated from a 97 F-1 clear octane Platformate, t h a t fraction has clear and leaded octane numbers of 51 and 70, respectively. T h e distribution of aromatics at two conditions are given in Tables V, VI, and VI1 along with comparative equilibrium data ( 5 ) . A comparison of these results with the aromatic distributions obtained while processing cumene (2)and with the equilibrium concentrations ( 5 )is the subject of Table VIII. An examination of t h e aromatic distribution results obtained with the n-paraffin charge, Tables V, VI, and VII, reveals two interesting points. The first is the production of sizable yields of lower boiling aromatics from higher boiling n-paraffins. The Cs and lighter aromatics formed amounted t o about one third the total aromatic production while the C8 and lighter paraffins amounted to only about one sixth t h e charge stock. Furthermore, as the space velocity is reduced there is only a relatively small increase in the yield of Clo+ aromatics, compared t o the substantial increase in t h e yields of C7, Cs, and Cs aromatics. The other point is the predominance of the ethyltoluenes compared to the trimethylbenzenes in the Cg aromatic distribution a t the higher space velocity condition and the agreement with the equilibrium distribution between the ethyltoluenes and the trimethylbenzenes a t t h e lower space velocity condition (Table VIII). However, distribution of the individual ethyltoluenes and trimethylbenzenes within each group was unchanged, and in both cases good agreement existed between these results and the equilibrium data ( 5 ) . This indicates t h a t t h e ethyltoluenes are the precursors of the trimethylbeneenes. Thus, i t appears reasonable t h a t in general the alkylaromatics containing the larger alkyl groups are formed first. These alkylaromatics are rehydrogenated t o the corresponding naphthenes; t h e naphthenes are isomerized t o polyalkylcyclohexanes which are then dehydrogenated to the corresponding polyallrylaromatics. The ready establishment of t h e equilibrium concentrations of the individual alkylaromatics with the same alkyl groups is then inherent. The following schematic diagram presents an example of a path by which these aromatics may be formed. I n this case, a '2x1nparaffin undergoes cyclization t o a five-membered ring as proposed by Herrington and Rideal (4). The polyalkylcyclopentane isomerizes t o another polyalkylcyclopentane with a larger number of alkyl groups through a carbonium ion mechanism. The indicated loss of a n ethyl group, t o form ethane, may occur a t this time resulting in a molecular weight reduction. The
INDUSTRIAL AND ENGINEERING CHEMISTRY
Vol. 47, No. 4
GASOLINE PROCESSING
Effect of Pressure at Constant Temperature and Space Velocity
Table 111.
Pressure, lb../sq. inch gage Yields Hydrogen std. cu. ft./bbl. of charge Dry Ras (61 Cs),std. cu. ft./bbl. of charge Butanes, lig. vol. % of charge Pentanes, Iiq. vol. % of charge Ca+ Platformate, liq. vol. % of charge Aromatics, wt. 70of charge
-
350
700
787 347 14.2 13.3 71.1 40.3
210 546 17.2 17.9 66.5 28.1
35 56
39 59
90.7 80.1 53.9 52.5 0.7690
85.2 4i:9 60.8 0.7358
110 146 261 336 360 411
114 134 200 313 330 3 67
Compositions Liq. vol. 70is0 in butanes Liq. vol. 70is0 in pentanes Properties of Cs+ Platformates F-1 clear octane number F-2 clear octane number Aromatic content, wt. % Gravity, 'API @ 60' F. Specific gravity @ 60' F. Enpler distillation, F. @ Initial b.p. 10%
-
-
jo%
90% 95y0 End point
polyalkylcyclopentane produced can isomerize through two general paths t o form methylethylcyclohexanes or trimethylcyclohexanes. These are readily dehydrogenated to the corresponding aromatics. The higher initial production of ethyltoluenes as compared with t h a t of t h e trimethylbenzenes can be explained on the basis of a faster rate of equilibration of the methylethylcyclohexanes t o the ethyltoluenes than the rate of equilibration of the methylethylcyclohexanes to the trimethylcyclohexanes. This agrees with t h e initial preponderance of the ethyltoluenes and the subsequent establishment of the rela-
Table IV.
Effect of Space Velocity a t Constant Temperature and Pressure
Space velocity, WHSV Yields Hydrogen, std. cu. ft./bbl. of charge Dry gas (CI Ca), std. cu. ft./bbl. of charge Butanes, lis. vol. yo of charge Pentanes, liq. vol. % of charge CS+Platformate, liq. vol. Yo of charge Aromatics, wt. yo of charge
I
I
0 PLATFOR?&T€
Figure 4.
27 48
34 55
52.0
73.2 67.3 35.9 58.1 ,7463
88.5
23:s 59.9 ,7393
48:O 56.3 .7535
130 186 277 .326 340 372
117 160 258 328 345 380
120 151 252 344 370 383
E3 point
-
tive equilibrium concentration b e h e e n the ethyltoluenes and the trimethylbenzenes.
CZ
/
C2
0
6
/
c1
f.l
Ethyltoluenes
+ 3H2
'
I
20
60 80 100 OCTANE NUMBER
40
CLEAR
F-I
C 5 + Platformate yield-octane I
ip-
22 39
90%
c1 -50 C$
510 440 15.2 15.8 69.5 34.3
E %
/
L
60 1 335 9.0 11.7 78.3 28.8
Compositions Liq. vol. '70 !so in butanes Liq. vol. yo is0 in pentanes Properties of Cs+ Platformate F-1 clear octane number F-2 clear octane number Aromatic content, wt. 70 Gravity, OAPI @ 60° F. Specific gravity @ 60' F. Engler distillation, ' F. @ Initial b. p. 10%
2
372 193 7.9 7.9 84.5 20.5
-
C2
5
4
8
I
Trimethylbenzenes 3H2
/'
C
+
C
I
I
/
W
Table V.
Aromatic Distributions L
Space velocity, WHSV
Wt. % Total aromatics
Charge
- -2 Wt. 70 Total aromatics Charge
I
-50 C$
Figure 5.
w 20
0 PLATFORMATE
2 0 4 0 6 0 8 0 1 0 0 F-l
CLEAR OCTANE NUMBER
Butane yield, liquid volume charge
I
I
I
I
I
70 of
Table VI.
Xylenes, orthomnta-
Figure 6.
April 1955
PLATFMTE
F-l
C L E M OCTANE NUMBER
Pentane yield, liquid volume charge
0.8 10.0 33.3 81.0 24.9 __ 100.0
0.4 4.4 14.6 13.6
10.9 43.9
I
Space Velocity, WHSV
Cg
1.3 0.4 8.1 2.4 19.2 5.8 38.0 11.5 3 3 . 4 10.0 ~100.0 30.1
€3 en z e n e Toluene Xylenes and ethylbenzene C QAromatics ~ I O +Aromatics Total
70
of
paraEthylbenzene Total
Detailed CSAromatic Yields 8
2
Wt. % Aro- Group, Charge matics Yo 1.4 2.1 1 1
13 5 8
4.7 7.1 3 5
25 37 18
19 2
100
39
E
INDUSTRIAL AND ENGINEERING CHEMISTRY
Wt. yo EquiAro- Group, librium Charge matics Data 3.0 6.1 2 8 2 7 14 6 ~
6.9 13.8 6 4 6 2 333
21 41 19
19 100
23 46 21 L O 100
733
ENGINEERING, DESIGN, AND PROCESS DEVELOPMENT As shown in Tables VI, VII, and VI11 the agreement among the results with the n-paraffin charge at 2 WHSV, the cumene data ( d ) , and the equilibrium data (6) is quite good. About the only exception is the 19% ethylbenzene concentration in the Ca aromatics produced from the n-paraffin charge. This may be the result of ethylbenzene being the precursor of the xylenes in the manner indicated. The relatively greater reactivity of the heavier versus the lighter aromatics would account for the lack of equilibrium isomerization of ethylbenzene t o the xylenes.
C z P L A T F M T E F-1 CLEAR OCTANE NUMBER
Figure 7.
Dry gas, Cl-C3, yield std. cu. ft,/ bbl. of charge
As the path of this isomerization appears to be through the naphthenes a decrease in the hydrogen partial pressure would be expected t o have a detrimental effect on the extent of isomerization. The 8 space velocity data shown was obtained a t a higher pressure than the 2 space velocity data. This pressure difference apparently counteracted the effect of the longer time and the net result was no change in the relative concentrations of ethylbenzene and xylenes between the 8 space velocity and the 2 space velocity results. The isomerization of butane and pentane produced appears t o be limited b y competition from the higher boiling paraffins and aromatics for the catalyst surface. Table I V indicates t h a t the isobutane content of the total butanes and isopentane content of the total pentanes increases from 22 to 39% a t 8 space velocity to 34 and 55%, respectively at 2 space velocity. An increase in the temperature a t constant pressure shows a similar effect (Tabel 11). A smaller trend of this nature is shown when pressure is increased (Table 111). Except for a beneficial effect of increased residence time, a n increase in the total pressure would not be expected to benefit the conversion of light hydrocarbons under conditions when competition for the catalyst surface produces a limiting effect. A comparison of the results obtained with the 100% n-paraffin content charge and the results presented earlier with a 66% paraffin content naphtha (Kuwait) and a 39% paraffin content naphtha (Venezuela) (3) a t the same pressure is shown by Figures 4 to 12. At the 90 F-1 clear CS+Platformate octane number level the yield decreases with a n increase in the paraffin content of the charge from 90 liquid volume % with the 39% paraffin content charge to 69 liquid volume % with the 100% paraffin content charge (Figure 4). As can be observed from Figure 4,
/-
I
I
-32 C5'
Figure 8.
.;
Aromatics, wt.
%
of C 5 + Platformate
"t
40
/
I
1
I I I 100 90 80 PLATFOWWTE YIELD
CC'
I I I I I I I 0 2 0 0 4 0 6 0 8 0 1 0 0 PLATFORMATE F - I CLEAR OCTANE NUMBER
Figure 9.
20
Net aromatic production, wt. charge I
I
%
J
I 70
,L V % ff
60 a-L4RGE
Pentane yield, liquid volume % of charge
Figure 1 1.
Cs+
1
I
0 20 40 60 80 100 PLATFORh44l-E F - I CLEAR OCTANE NUMBER
I
I
I
I
I
I 90
I 80
I 70
600
of
I
I
66O PARAFFINS
/+-loo%
100
90
80
C5+ PLATFORMATE YIELD Figure
734
10.
,
70 60 LV% OF CH4G-E
Butane yield, liquid volume charge
70 of
-
t-i
I
100
C
