J . Chem. Tech. Biotechnol.

1986,36, 519-526

n-Octane Dehydrocyclisation on Monofunctional

and Bifunctional Pt/AI,03 Catalyst
Churchill T. Ako and Alfred A. Susu
Department of Chemical Engineering, University of Lagos, Lagos, Nigeria
(Manuscript received 20 October 1985 and accepted 11 February 1986)

The dehydrocyclisation of n-octane to iso-octane, ethylbenzene and o-, m- and

p-xylene was investigated on monofunctional (non-acidic) and bifunctional (acidic)
Pt/AI,03 catalyst in a microcatalytic reactor with hydrogen as carrier at 1.8 atm and
563473 K. On bifunctional Pt/AI20,, the total conversion of n-octane started from a
high value and decreased with increasing temperature for all pulse sizes investigated.
The primary product of n-octane conversion on acidic Pt/AI,O, was iso-octane. The
product yield-temperature profiles showed a large initial production of iso-octane
which decreased to a minimum as the catalyst temperature increased due to its
conversion to ethylbenzene and o-xylene. On non-acidic Pt/A1203,the total conver-
sion of n-octane increased initially and then went through a maximum as the catalyst
temperature increased. The primary products of the reaction were found to be
ethylbenzene and o-xylene, indicative of the activity of the metal to effect these ring
closure reactions.
Keywords: Normal octane dehydrocyclisation; acidic and non-acidic Pt/A1203
catalysts.

1. Introduction

The dehydrocyclisation of paraffins is one of the important reforming reactions for the production of
high octane gasoline and aromatics for the petrochemical industry. These dehydrocyclisation reac-
tions have been widely studied on chromia-alumina and platinum-alumina catalysts. Most of the
earlier works were carried out largely on supported non-acidic metal oxide or metal catalysts in
continuous vapour-phase reactors where catalyst deactivation must be significant. Some recent
works on the conversion of CbCS alkanes, naphthenes and aromatics on supported platinum were
carried out in microreactors where the surface of the catalyst is maintained in a relatively clean
state.''. I 2
There appears to be more controversy on the mechanism of the dehydrocyclisation process on
supported platinum catalyst. The mechanisms of dehydrocyclisation of normal and branched-chain
heptanes, octanes and trimethyl pentanes on non-acidic chromia-alumina appears to be more
reasonably understood.
In the present investigation, experimental results are presented on the conversion of n-octane on
acidic and non-acidic platinum-alumina catalysts using the pulse technique. These reactions can be
studied for relatively long periods without catalyst deactivation. This may enhance the understand-
ing of the complex reactions occurring during the dehydrocyclisation process.
Since little has been done on the investigation of dehydrocyclisation reactions on acidic Pt/AI2O3,
this work presents results that may help understand the kinetics of n-octane conversion on bifunc-
tional catalysts. Furthermore, it may be possible to delineate the specific contributions of the acidic
and the metal components in the bifunctional catalyst as each of the components is known to catalyse
more than one reaction type.
519
520 C. T. Ako and A. A. Susu

2. Experimental
2.1. Materials
Redistilled n-octane (Fison Scientific) was used in this investigation. Extra pure hydrogen
(Industrial Gases) was passed through a bed of molecular sieve for the removal of traces of water
vapour. A commercial Pt/A1203catalyst was used with the following properties: 0.3% Pt, 0.6% Cl,
180 m2g-' surface area and 0.50 cm'g-' pore volume. The acid component of this catalyst was
removed by thorough washing in a 0.1 M solution of KOH followed by several washings with distilled
water. The catalyst was then washed in hot distilled water and oven dried at 120°Cfor 16 h. A pH test
showed that the acid content was removed by this procedure.
2.2. Analysis
Product samples were analysed chromatographically by means of an on-line Hitachi 64 FID. A
Bentone 34-diisodecyl phthalate column was used for the separation of n-octane, iso-octane,
ethylbenzene, o-, rn- and p-xylene. Quantitative analyses of the product stream were performed
through calibration of known mixtures of the components. The number of moles of all components
were calculated by means of carbon balance.

2.3. Procedure
The apparatus used in this investigation has been described el~ewhere.'~. l4 Briefly, the dual mode
catalytic reaction system (Chemical Data System, Inc.) was operated in the pulse mode. The reactor
was packed with 1 g of either the acidic or non-acidic Pt/Al,O, of particle size 0.1-0.4 mm. Prior to
each experiment, the catalyst was activated under hydrogen flow for 10 h at 693 K. The flow rate of
hydrogen was varied between 20 and 120 cm3 min-' and pulse sizes of reactants were varied between
0.1 and 2 . 0 ~ 1 The
. catalyst temperature was varied between 573 and 673 K while the total system
pressure was kept constant at 1.8 atm. All experimental measurements were made after the catalyst
had reached steady-state activity. This was ensured by sequential pulsing at a given temperature until
the activity of the catalyst reached a steady value. Furthermore, catalyst temperature was randomly
selected in all experimental runs.

3. Results
3.1. Total conversion
The variation of the total conversion of n-octane with sample size and catalyst temperature is shown
in Figures 1 and 2 on acidic and nonacidic Pt/A1203, respectively. Conversion was found to be

'g
0.0

673K
0
lo
I J
0.0 0.5 1.0 I .5 2.0
Pulse size ( p l )
Figure 1. n-octane conversion against pulse size on acidic Pl/AI,O, (H2 carrier, 60 cm'/m)
n-Octane dehydroeyelisation on monofunctional and bifunctional Pt/AI,O, catalyst 521

613 K
633 K
653 K
673 K

Fipn 2. n-octane conversion against pulse size on non-

acidic Pt/AI,O, (Hzcarrier, 60 cm’/m). I I I
0.5 1.0 1.5 2.0
Pulse size (pl)

insensitive to pulse size at sizes above 1.0 p1 for both catalysts in the temperature range investigated.
This ensures that for the sample size of 1 . 0 ~ 1in all subsequent experimentation, equilibrium
chromatography prevailed in the catalyst bed.
Figures 3 and 4 illustrate the variation of the total conversion of n-octane with temperature on
acidic and non-acidic Pt/[Link] conversion-temperature profiles showed significant differences
on the two catalysts. On acidic Pt/AI2O3,total conversion was found to decrease with increasing
catalyst temperature for sample sizes greater than 0.1 pl. However, on non-acidic Pt/A1203, total
conversion was found to increase from zero and passed through a maximum for each of the sample
sizes investigated. The observed initial high conversion of n-octane on acidic P t / A b O , at 573 K
(Figure 3) must have occurred over a narrow temperature range as its value must also increase from
zero. This increase from a low value was not detected in this investigation.
3.2. Product distribution
Figure 5 shows the variation of product yield as a function of temperature on acidic Pt/A1203for a
pulse size of 0.5 microlitre. The product yield for component i is defined as the ratio of the moles of
component i in the product stream to the moles of n-octane reacted. Figure 5 shows that ko-octane
was formed rapidly from n-octane at 573 K because of the intrinsic acidity of the alumina. The iso-
octane was subsequently converted to ethylbenzene and o-xylene to the total exclusion of p- and
m-xylene. For a pulse size of 1 pl, small yields of p- and m-xylene were formed on the acidic catalyst
(see Figure 6). As the pulse size increased, the yields of p- and m-xylene increased although the

.s
f
0
u

Figure 3. n-octane conversion against

temperature for different pulse sizes on
acidic Pt/AI20,. (Hzcarrier. 60 cm3/m; 1 ,
2.0; 2, 1.5; 3, 1.0; 4,0.5; 5 . 0 . 1pi).
573 593 613 633 653 673 693 713
Temperature ( K 1
522 C. T. Ako and A. A. Susu

0.8-

g 0.6-
'i
c
B 0.4-
C
0
5
;0.2-
Figure 4. n-octane conversion against temperature for
different pulse sizes on nonacidic Pt/A1203 (H2 carrier,
60 cm3/m; 1, 2.0; 2. 1.5 and 3,O.S; 4, 0 . 2 ~ 1 ) .

Temperature ( K )

Figure 5. Product yields against temperature for

n-octane conversion on acidic Pt/AI20,4. pulse
size=O.S PI (1, iso-octane; 2, ethylbenzene; 5 . J-xylene).

Temperature ( K )

Figure 6. Product yields against temperature for n-octane

conversion on acidic Pt/AI,03 (1, iso-octane; 2 ethylbenzene;
3 p-xylene; 4, rn-xylene; 5 o-cylene).

predominant products from iso-octane conversion remained ethylbenzene and o-xylene. Results a t
pulse sizes of 1.5 and 2.0 pI resulted in increasing production of p- and m-xylene although the
maximum product yield recorded for either product was about 10%. For all pulse sizes investigated,
the iso-octane yield went through a minimum while the ethylbenzene and the xylene yields each went
through a maximum.
n-Octane dehydrocyclisstion on monofunctional and bifunctional R/AI,O, catalyst

Ngure 7. Product yields against temperature for

n-octane conversion on nonacidic Pt/Ai,03 (2
ethylbenzene; 3. p-xylene; 4, rn-xylene; 5, o-xylene).

Temperature ( K )

On non-acidic Pt/A1203, there were significant differences in the yield-temperature profiles.

Since the acidic component was absent, the catalyst became monofunctional; the acid catalysed
isomerisation of n-octane to iso-octane was completely absent. At a pulse size of 1 . 0 ~ 1(see
Figure 7), ethylbenzene and o-xylene were the predominant products initially although quantities of
p- and m-xylene less than 5% were also produced. As the temperature increased, ethylbenzene went
through a minimum while o-xylene became predominant and passed through a maximum. The yields
of p- and m-xylene were less than 20% at their maxima. The general characteristics of the yield-
temperature profiles on acidic Pt/A1203 were retained except that on non-acidic Pt/A1203, two
products were initially formed (ethylbenzene and o-xylene) compared to one (iso-octane) on acidic
Pt/Al,03. Furthermore, more p- and m-xylene was formed on the non-acidic catalyst.
Although the yield data reported above provided information on primary products of reaction
especially at small pulse sizes, it is necessary to confirm these general observations by considering
product yields versus reciprocal velocity data.
Figure 8 shows the variation of the product yield as a function of reciprocal velocity on acidic
Pt/A120, at 593 K. These plots show very clearly that iso-octane is the primary product of reaction in
agreement with the product yields-temperature plots. All the xylenes are secondary reaction
products. Similar results were obtained for acidic Pt/A1203at 613, 633, 653 and 673 K.
However, on nonacidic Pt/A1203,there are two primary products-ethylbenzene and o-xylene-
although ethylbenzene predominates, as illustrated in Figure 9 at 593 K. This behaviour was
substantially reproduced at 613 and 633 K. However, at 653 K, the only products of the reaction

0.8

0.6
-00,
>
0.4

F9gure8. Product yields against recipro-

cal velocity for n-octane conversion on 0.2
acidic Pt/Al,O, (1, bo-octane; 2
ethylbenzene; 3,p-xylene; 4, rn-xylene; 5, 384
o-xylene).
0.0 1.0 2.O 3.0
W/F [qrnin~rn-~x~o-~)
su C. T. Ako and A. A. Susu

0.0 Lb

Figure 9. Product yields against reciprocal

velocity for n-octane conversion on nonacidic
Pt/A120, (1, ethylbenzene; 2, p-xylene; 3,
rn-xylene; 4. o-xylene).

Figure 10. Product yields against reciprocal velocity for

n-octane conversion on nonacidic Pt/A120, (1, ethylbenzene;
I I I 2, o-xylene).
1.0 2.0 3.0
W / F ( q min cm-3x~0-2)

were ethylbenzene and o-xylene (see Figure 10). This is consistent with the product yields-tempera-
ture profile of Figure 7 which showed no production of p- and m-xylene at 633 K on nonacidic
Pt/AI20,. Initial rate data provided evidence in support of ethylbenzene and o-xylene being primary
products of n-octane dehydrocyclisation on nonacidic Pt/AI20,. If ethylbenzene is the only primary
product, then the ratio of the initial rate of ethylbenzene to o-xylene should be unity at 653 K where
no other products were formed. The observed ratio was 0.67 supporting the argument that o-xylene
is also a primary product as its rate of production exceeded its rate of formation from ethylbenzene.

4. Discussion of results
4.1. Reaction pathways
The reaction pathways for the dehydrocyclisation of n-octane on acidic and nonacidic Pt/AI20, are
complex as they consist of several consecutive and simultaneous elementary reactions. On nonacidic
Pt/A1203,Davis and Venuto9 reported that more than 95% of the aromatic products from n-octane
dehydrocyclisation were those predicted by direct 1.6-ring closure. At 755 K these products were
ethylbenzene and o-xylene. The results of our microreactor investigation on nonacidic Pt/A1203at
lower temperatures confirmed their results. Furthermore, we confirmed that the primary products
of the dehydrocyclisation reaction were ethylbenzene and o-xylene at temperatures between 593 and
653 K. This agreement is significant as the work of Davis and Venuto9 was done without the use of a
carrier gas where rapid carbiding of the platinum surface might have taken place. It may thus be
concluded that surface carbiding does not alter the mechanism of the dehydrocyclisation reaction.
n-octane dehydrocyclisation on monofunctional and bifunctional R/AI,O, catalyst 525

Christoffel and Paal" have shown by using a pulse technique that on a dual-functional Pt/Al,O, in
Hzcarrier at 673"K,n-octane was converted to ethylbenzene and o-xylene. Also, C,-ring naphthenes
were also formed as these were not converted further to C6-ringnaphthenes under their experimen-
tal conditions. Our results are at variance with the above on bi-functional Pt/Al,O,. In the tempera-
ture range of 563-673 K, no C&g naphthenes were observed. More importantly, iso-octane was
found to be the primary product of reaction throughout the temperature range investigated. The
results of our product yields-temperature profiles did show that iso-octane was subsequently
converted to ethylbenzene and o-xylene.
The two previous works discussed above were done at single temperatures with no report of the
effect of temperature on product yields. An important aspect of this work is that product yields of
n-octane dehydrocyclisation on both acidic and nonacidic Pt/AI,03 were investigated in hydrogen
carrier over a wide temperature range. As mentioned earlier, conversion on acidic Pt/A120, was
only detected at temperatures 2573 K where isomerisation of n-octane to iso-octane was predomi-
nant. Since this reaction is very fast, it is not unrealistic that a high yield occurred over a short
temperature range. More importantly, the reciprocal velocity data indicated that iso-octane was the
sole primary product in the temperature range 593 to 673 K. Therefore, the minimum in iso-octane
production on the acidic catalyst could not be due to a decrease in the intrinsic acidity of the alumina.
If the intrinsic acidity of the alumina is actually decreasing, the minimum temperature of iso-octane
yield should be relatively independent of pulse size. This was not observed as the temperature at the
minimum range from 613-650 K with pulse size of 0.5-2.0 pl. It can thus be concluded that iso-
octane decrease was due solely to its subsequent conversion as exhibited by ethylbenzene and
o-xylene yields. This decline may be attributable to the decreasing ability of the catalyst to effect ring
closure reactions as the temperature increased beyond a certain value.
The identity of the primary products of the dehydrocyclisation reaction suggests the following
sequence of single reactions on the bifunctional Pt/AI20,:

ethylbenzene
n-octane-iso-octane
ert /
L o - xy lene

On monofunctional Pt/AI,O,, iso-octane production is absent and direct formation of ethylbenzene

and o-xylene occurs.
4.2. Total conversion
In this study, single reactions characterising monofunctional and bifunctional dehydrocyclisation of
n-octane on Pt/AI20, catalyst have been identified. We shall now attempt to give a kinetic explana-
tion of the decrease in total n-octane activity with temperature. From the results, the total conver-
sion decreased from a high value on the acidic catalyst after the initial fast isomerisation reaction. On
the nonacidic catalyst, however, total activity was observed to go through a maximum.
This behaviour cannot be attributed to the exothermicity of the rate determining step(s). The
mechanism of the dehydrocyclisation reaction probably remains unaltered after the acidity of the
bifunctional Pt/A120, catalyst is removed. This is because the isomerisation reaction is fast com-
pared to subsequent conversion steps. Thus, the dehydrocyclisation step is probably the slowest step
in the sequence. The dehydrocyclisation step is known to be endothermic." The endothermicities of
n-octane conversion to ethylbenzene, 0-,m-andp-xylene were calculated from heats of combustion
dataIhto be 62.0, 130.9. 127.6 and 128.3 kcal/gmol, respectively. No maximum in total conversion
with temperature is thus expected from the above argument.
It is likely that the behaviour is due to the presence of adsorbed hydrogen whose coverage at
different surface temperatures determines the activity-temperature profile. The hydrogenation/
dehydrogenation activity of the metal component is known to be dependent on the coverage of
chemisorbed hydrogen and has been shown to explain maxima in is~merisation'~and
dehydrocyclisation reactions. I'. I*
526 C. T. Ako and A. A. Susu

References
1. Pines, H. The Chemistry of Catalytic Hydrocarbon Conversiom Academic Press, New York, 1981.
2. Pines, H.; Nogueria, L. (1981) Aromatization of heptanes, ethylcyclopentane and cycloheptane over ‘non-acidic’ platinum-
alumina catalyst. Comparison of the mechanisms of aromatization over chromia-alumina and over platinum-alumina
catalysts. 1. Catalysis 70, 391-403.
3. Noguera, L.; Pines, H. (1981) Aromatization of [1-14C] heptane over ‘non-acidic’ platinum-alumina catalyst. J. Catalysis
70, 406408.
4. Davis, B. H. (1971) Hydrocyclization of paraffins. J. Caralysis 23. 34CL354
5. Davis, 8.H. (1971) Hydrocyclization of paraffins. J. Catalysis 23, 355-357.
6. Davis, B. H. (1971) Hydrocyclization of paraffins. J. Catalysis 23, 365373.
7. Davis, B. H.; Venuto, P. B. (1971) Carbon-14 tracer study of the dehydrocyclizationof n-heptane. J. Catalysis 23,337-340.
8. Davis, B. H. (1973) I4Ctracer study of the dehydrocyclization of n-heptane over Pt on non-acidic alumina. J. Caralysis 29,
396403.
9. Davis, B. H. (1969) Paraffin dehydrocyclization. J. Caralysis 15, 36S372.
10. Lester, G. R. (1969) Different mechanisms of aromatization of trimethylpentanes over chromia-alumina and platinum-
alumina catalysts. J. Caralysis 13, 187-192.
11. Christoffel, E. G.; Paal, Z. (1982) Activity and selectivity of bifunctional platinum catalysts in hydrocarbon reactions. J.
Catalysis 73, 3@-44.
12. Paal, Z. (1980) Metal-catalysed cyclization reactions of hydrocarbons. Adv. Caralysis 29, 273-334
13. Susu, A. A.; Ako, C. T. (1984) Comparative kinetic study of ethylbenzane isomerizationon acidic and non-acidic Pt/AI,O,
catalysts. 1. Chem. Tech. Biotechnol. MA,324-334.
14. Susu, A . A.; Ako. C. T. (1985) Ethylbenzene isomerization on bifunctional platinum/alumina. Product yields and pseudo
mass action kinetics. Appl. Catalysis 16, 179-192.
15. Gates, B. C.; Katzer, Schuit, G. C. Chemisrry of Caralytic Processes McGraw-Hill. New York, 1979.
16. Hongen. 0. A.; Watson, K. M.; Ragatz, R. A. Chemical Process Principles John Wiley & Sons,Inc., New York, 1954.
17. Rohrer, J. C.; Hurwitz, H.; Sinfelt, J. C. (1961) Kinetics of the catalytic dehydrocyclization of n-heptane. J. Phys. Chem.
65, 1458-1460.