Pyrolysis of hexadecane

Article abstract of DOI:10.1135/cccc19971057

The kinetics of thermal decomposition of hexadecane was studied in a flow tubular reactor from stainless steel. The experiments were performed in the temperature range of 700 to 780°C for the mass ratio of steam to hydrocarbon 3 : 1. The hexadecane pyrolysis took place according to the first-order reaction with a frequency factor of 3.5 . 10⁹ s^-1 and an activation energy of 162 kJ mol^-1. In the pyrolysis products there were above all 1-alkenes. From alkanes, methane and ethane and less propane were formed in a higher degree. The prevailing compounds are ethene and propene whose amount increases with increasing temperature and residence time. The content of 1-alkenes higher than 1-pentene decreases with increasing conversion which gives evidence of their decomposition owing to their lower stability in comparison with the lighter 1-alkenes. The formation of dienes (1,3-butadiene and propadiene) and benzene also confirmed the course of secondary reactions. The observed higher formation of hydrogen results from the reaction of steam with coke deposited on the walls of the reactor or with hydrocarbon radicals. The evidence of the coke formation is given also by the presence of carbon oxides whose amount grew with the pyrolysis severity. The high content of 1-hexene in comparison with the other higher 1-alkenes is probably caused by the isomerization of alkyl radicals by 1,5-hydrogen transfer.

Full text of DOI:10.1135/cccc19971057

Pyrolysis of Hexadecane
1057
PYROLYSIS OF HEXADECANE
1
2
Elena BARTEKOVA and Martin BAJUS
Department of Petroleum Technology and Petrochemistry, Slovak Technical University,
8
2
12 37 Bratislava, Slovak Republic; e-mail: 1 barteko@chelin.chtf.stuba.sk,
bajus@chelin.chtf.stuba.sk
Received October 9, 1996
Accepted April 22, 1997
The kinetics of thermal decomposition of hexadecane was studied in a flow tubular reactor from
stainless steel. The experiments were performed in the temperature range of 700 to 780 °C for the
mass ratio of steam to hydrocarbon 3 : 1. The hexadecane pyrolysis took place according to the first-
9
–1
–1
order reaction with a frequency factor of 3.5 . 10 s and an activation energy of 162 kJ mol . In the
pyrolysis products there were above all 1-alkenes. From alkanes, methane and ethane and less pro-
pane were formed in a higher degree. The prevailing compounds are ethene and propene whose
amount increases with increasing temperature and residence time. The content of 1-alkenes higher
than 1-pentene decreases with increasing conversion which gives evidence of their decomposition
owing to their lower stability in comparison with the lighter 1-alkenes. The formation of dienes
(
1,3-butadiene and propadiene) and benzene also confirmed the course of secondary reactions. The
observed higher formation of hydrogen results from the reaction of steam with coke deposited on the
walls of the reactor or with hydrocarbon radicals. The evidence of the coke formation is given also
by the presence of carbon oxides whose amount grew with the pyrolysis severity. The high con-
tent of 1-hexene in comparison with the other higher 1-alkenes is probably caused by the isomeriza-
tion of alkyl radicals by 1,5-hydrogen transfer.
Key words: Hexadecane pyrolysis; Kinetics; Alkenes.
On producing alkenes by pyrolysis, various raw materials are used, mainly ethane,
LPG, primary naphthas and gas oil. Most units in Canada and U.S.A. process ethane or
LPG, whereas Europe and Japan already traditionally use naphthas. However, in last
years, an increasing trend is noted in processing the higher-boiling materials, especially
gas oil. The shift in this direction came on because of an insufficient amount of tradi-
tional raw materials used to cover the requirements of ethene production especially
1
when the utilization of primary naphthas for production of motor gasoline is getting
priority. The trend therefore brings about the necessity for the study of kinetics and
formation of products of pyrolysis of gas oil. It was one of main reasons why we chose
the linear alkane – hexadecane – as a model substance corresponding by the number of
carbon atoms to gas oil.
More laboratories have pursued the thermal decomposition of hexadecane. The re-
sults reached are in a decisive extent influenced by the experimental conditions which
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
058
Bartekova, Bajus:
are considerably different. The low-pressure experiments were carried out in the gas
2
3
and liquid phases and the high-pressure ones in the gas phase . The aim of the work
4
dealing with the hexadecane decomposition in the liquid phase was to elucidate the
mechanism of decomposition of the high-molecular alkane under mild temperatures of
3
50 to 440 °C and at a pressure around 2 MPa modelling the conditions of mild crack-
ing of heavy oils (visbreaking).
Hitherto only Depeyre and Fairburn with co-workers have studied the kinetics of
pyrolysis of gas oil and hexadecane under the conditions of pyrolysis near the industrial
5
1
(
short residence times, high temperatures up to 850 °C, approximately atmospheric
pressure).
For this reasons, we carried out our study under conditions of middle-temperature
pyrolysis. In the interval of operating temperatures of 700 to 780 °C (with 20 °C in-
crease in temperature), we determined the kinetic parameters – rate constants, activa-
tion energy and pre-exponential factor. We investigated the selectivity of formation of
the profiling hydrocarbons as a function of temperature and residence time. As a di-
luent, steam was used. It is so even in the industrial pyrolysis where steam is used for
reducing the hydrocarbon partial pressure and shortening the residence times. In this
way, the course of secondary reactions is suppressed resulting in the formation of high-
molecular compounds which are precursors of the coke formation. Steam is also able to
react at higher temperatures with already produced coke deposits on forming carbon
oxides.
EXPERIMENTAL
Pyrolysis Equipment
6
The equipment used was described in detail in our preceding paper . The steel tubular reactor (com-
position: 16.8% Cr, 10.7% Ni, 0.4% Si, 1.1% Mn, 0.09% C) was of length of 750 mm and inside
–
1
diameter of 6 mm. The ratio of inside surface to reactor volume represents 57.1 mm . The hexade-
cane purity (Fluka, Switzerland) was determined by gas chromatography and reached 99.7 mass %.
–
1
The mass flow rate was within 10 to 55 g h . The inaccuracy of measurement was in the range 5 to
2%.
1
Analytical Method
The analyses of the gas and liquid products with the aim to find out the qualitative and quantitative
distribution of single components were made by means of the gas chromatography. In the products
were contained: C to C alkenes, of these above all 1-alkenes, alkanes from methane up to small
2
15
amounts of pentane, and butadiene. In the liquid products, unreacted hexadecane prevailed. When
analysing the hydrocarbon components, a flame-ionization detector (FID) and for non-hydrocarbon
portions (hydrogen, carbon monoxide, carbon dioxide) a thermal-conductivity detector (TCD) were
used. The experimental conditions of the analyses are given in Table I.
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Pyrolysis of Hexadecane
1059
RESULTS AND DISCUSSION
Kinetics
The temperature profile is influenced above all by the characteristics of heating ele-
ments of pyrolytic furnace, space velocity of hydrocarbon material and inert diluent,
heat capacities of reactants and products and endothermicity of the reaction. From these
variables, we could control experimentally only the space velocity and the intensity of
electric current. The radial change of temperature in the reactor as well as the pressure
changes along the reactor were neglected in calculations. The equivalent reactor volume
needed for the determination of reaction time was calculated according to the Hougen–
7
Watson theory . The individual temperatures in the temperature profile were measured
along the reactor with the 20 mm step. The calculated equivalent volume was inserted
6
into the relation for calculating the residence time .
The curves of dependence of conversion on residence time for the temperatures of
7
00 to 780 °C do not start from the zero origin but there is here a certain shift on the
axis of residence time. This shift represents approximately the value of 0.015 s by
which the calculated retention time had to be diminished. Further calculations and dia-
grams were carried out with these adjusted residence times. In this way we reached
TABLE I
Conditions of GC analyses of pyrolysis products
Chromatograph type
Parameter
CHROM IV
gases
CHROM IV
CHROM IV
Hewlett–Packard
Analysis
Column type
Length, m
i.d., mm
Phase
ethene, ethane
CH4, H2, CO, CO2 liquids
packed
3.7
packed
3.7
packed
2.5
capillary
50
3
3
3
0.2
ethylene glycol
butyrate, 20% on
Chromosorb P
Florisil
Supersorb
methylsilicone
0.5 µm
Column
50
50
80
40–180
temperature, °C
Carrier gas
Feed
N2
N2
He
N2
0.2 ml
FID
0.2 ml
FID
1 ml
TCD
0.17 µl
FID
Detector
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
060
Bartekova, Bajus:
approximately the same values of rate constants for measurements at a single tempera-
ture which values had hitherto increasing trend with increasing conversion.
The determination of reaction order is important for a right choice of rate equation
describing the course of the reaction studied. For low conversions, at a constant tem-
perature and pressure profile, the reaction order can be determined as a slope of the rate
6
dependence .
We performed a series of measurements at five temperatures, at the same tempera-
ture and pressure profiles but at different mass hexadecane feeds, i.e., at different
residence times. We report the reaction orders as determined by the method of linear
regression: At 740 °C 1.1, at 760 °C 1.1, and at 780 °C 1.0. These values are in good
agreement with the generally accepted assumption of the first-order reaction for the
alkane pyrolysis. Analogously, the authors of most works dealing with hexadecane py-
3
–5,8
rolysis
found out that the thermal decomposition is controlled by the reaction of
first order despite the fact that the reaction conditions differed considerably.
For a reactor with a plug flow in which a first-order reaction takes place, the follow-
9
ing equation of integral conversion
1
− x
k τ = (1 + ε) ln
− ε x
(1)
1
holds. The expansion factor ε represents a relative change of volume in the system
when changing from the zero (Vx=0) to complete conversion (Vx=1
)
Vx=1 − Vx=0
ε =
.
(2)
Vx=0
The calculated values of equivalent volume, effective residence time and expansion
factor for single measurements along with the conversion (Table II) served for the cal-
culation of rate constants of the thermal decomposition of hexadecane.
The right-hand side of Eq. (1) is illustrated in Fig. 1 as a function of residence time.
The slopes of these dependences represent the rate constants which were calculated by
the method of linear regression. The dependences are linear without indicating a de-
crease of values of rate constants in dependence on conversion. It gives evidence that,
in the region of conversions (from 9 to 50%) where we worked, a marked inhibition of
pyrolysis by reaction products does not take place.
The determined rate constants had the values:
t, °C
k, s
700
7.7
720
10.2
740
14.3
760
21.8
780
35.3
–
1
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Pyrolysis of Hexadecane
1061
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

1
062
Bartekova, Bajus:
Collect. Czech. Chem. Commun. (Vol. 61) (1996)

Pyrolysis of Hexadecane
1063
The magnitude of activation energy calculated by linear regression from the values of
–
1
rate constants in terms of the Arrhenius equation (Fig. 2) is 162 kJ mol and the pre-
9
–1
exponential factor represents 3.5 . 10 s .
The values of published kinetic parameters of thermal decomposition of hexadecane
and other higher linear alkanes differ enough. The activation energies lie in the interval
–
1
1–5,8
of 164 to 290 kJ mol (refs
). The value of activation energy determined by us is
relatively low but it is in a very good agreement with the activation energies of decomposition
–
1
1
–1
of hexadecane, 164.96 kJ mol (ref. ) and decomposition of dodecane, 164 kJ mol
(
1
0
ref. ). This low value is probably caused by the catalytic effect of metallic surface of
the reactor on the hexadecane decomposition.
Composition of Products
The quantity of gaseous products grows with the depth of conversion of hexadecane
(
Fig. 3). At a given conversion, the formation of pyrolytic gases at different tempera-
tures is approximately identical and therefore it depends above all on the hexadecane
conversion. From it follows that the overall yield of gases may be a good indicator of
conversion in pyrolysis of liquid hydrocarbons.
We report a survey of selectivities for five working temperatures (Table II). The
composition of products was expressed for single measurements in mol of the given
component formed from a hundred mol of reacted hexadecane. Under the reaction con-
ditions – conversion up to 50% and atmospheric pressure – above all alkenes were
formed on splitting hexadecane. As to alkanes, methane and ethane were formed in a
3
.6
ln k
3.2
0
.6
5
4
3
2
Y
1
0
0
.4
2
2
.8
.4
.2
.0
0
2.0
0.94
0
0.02
0.04
0.06
0.96
0.98
1.00
3
–1 1.04
τ, s
1
/T .10 , K
FIG. 1
FIG. 2
Graphical representation of Eq. (1) at tempera-
tures 700 °C (1), 720 °C (2), 740 °C (3), 760 °C
Graphical representation of Arrhenius equation
for hexadecane pyrolysis
(
4), 780 °C (5); Y right-hand side of Eq. (1)
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
064
Bartekova, Bajus:
larger extent, in smaller extent propane. We noted only small quantities of butane and
2
pentane. On the contrary, Wu et al. found out that on decomposing hexadecane in the
liquid phase at mild temperatures, even higher alkanes C –C and addition products
3
13
C –C are formed in a considerable amount.
1
8
30
As to alkenes, the dominant position showed ethene and propene. Their selectivity
grows with conversion and with temperature. From higher alkenes are present all 1-al-
kenes up to pentadecene. From 1-alkenes still 1-butene showed a growth with increas-
ing temperature. However, starting from 1-pentene, an opposite effect takes place. It is
caused by the fact that at higher temperatures, the longer 1-alkenes are less stable than
the shorter 1-alkenes. The former decompose on forming lighter 1-alkenes. The higher
hydrocarbons than hexadecane were not found in the reaction products.
The effect of residence time on the formation of profiling reaction products at the tem-
perature of 760 °C is illustrated in Figs 4–6. The prolongation of residence time has a
greater effect on the ethene formation in comparison with propene. Even at low residence
times, we noted the formation of 1,3-butadiene which is formed by secondary reactions.
The formation of methane (Fig. 5) was practically independent of the residence time.
The selectivity of non-hydrocarbon components – hydrogen, carbon monoxide and
carbon dioxide increased with increasing residence time.
The content of higher 1-alkenes decreased with increasing residence time and within
the range of the homologous series (Fig. 6). An exception was observed with 1-pentene
whose content was lower in comparison with 1-hexene.
4
0
250
M
1
A, %
2
00
3
0
0
150
2
100
2
10
50
3
4
0
0
0
20
40
60
0
0.1
0.2
0.3
x, %
τ, s
FIG. 3
FIG. 4
Formation of gas products A (mass %) as a
function of hexadecane conversion x (%) at
temperatures ❍ 700 °C, ❐ 720 °C, ∆ 740 °C,
Formation of lower alkenes and butadiene M
(mol/100 mol) as a function of residence time τ
(s) at 760 °C; ethene (1), propene (2), 1-butene
+ methylpropene (3), 1,3-butadiene (4)
✚
760 °C, ♦ 780 °C
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Pyrolysis of Hexadecane
1065
The application of radical mechanism makes it possible to interpret most realistically
the experimental data of hexadecane pyrolysis. In the propagation step, small radicals
such as methyl, ethyl, propyl and hydrogen atom react with the hexadecane molecule
(
(
reactions (A)–(D)). The abstraction of hydrogen from the primary or secondary carbon
hydrogen transfer reaction) manifests itself by the formation of molecular hydrogen,
methane (Fig. 5), ethane and propane (Table II).
•
•
C H
+
+
H•
C H
16 33
+
+
+
+
H2
(A)
(B)
(C)
(D)
1
6
34
CH3•
C H
C H
CH4
16
33
•
•
•
•
+
C H
C H
2
5
16 33
2
6
+
C H
C H
C H
3
7
16 33
3
8
The radicals produced in these two steps can further decompose on forming alkenes
and smaller free radicals.
Kossiakoff and Rice and Fabus12 assumed that all the ethyl radicals produced in
single reaction steps were changed to ethane by abstracting hydrogen especially from
the reactant (reaction (C)). However, at higher temperatures, even the decomposition of
ethyl radical (reaction (E)) can take place simultaneously with the hydrogen abstrac-
tion. The decomposition monomolecular reaction is supported by decreasing pressure
and increasing temperature.
11
100
M
1
80
6
0
0
2
4
3
4
FIG. 5
2
0
Formation of hydrogen, methane, and carbon
oxides M (mol/100 mol) as a function of residence
time τ (s) at 760 °C; H (1), CH (2), CO (3), CO
2
2
4
0
(
4)
0
0.01
0.02
0.03
τ, s
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
066
Bartekova, Bajus:
•
H•
C H
C H
4
+
(E)
2
5
2
These reactions account for the hydrogen formation in comparison with the Rice–Kos-
siakoff mechanism.
Up to now we have considered that only primary reactions take place during the
hexadecane pyrolysis. On the contrary, already at 700 °C and conversion of 9%, we
noted the formation of butadiene (Table II). It is not a primary product of the hexade-
cane pyrolysis. On increasing the hexadecane conversion, on increasing temperature
(
Table II) or residence time (Fig. 6), the selectivity for 1-alkenes higher than 1-pentene
decreases. Therefore it is necessary to take into account even the course of consecutive
secondary) reactions of the 1-alkenes formed. At higher temperatures, the longer 1-alkenes
(
are less stable than the shorter ones and decompose on forming shorter 1-alkenes.
At 35% conversion of hexadecane (760 °C) we noted for the first time traces of
benzene in the products. The formation of aromates in pyrolysis of alkanes can result
from the oligomerization of short alkenes or Diels–Alder addition.
The experiments were carried out at high temperatures, short residence times, and
strong dilution with water vapour, i.e., under the conditions under which a minimum
number of works dealing with pyrolysis of higher alkanes is published. For one thing as
1
3
an elucidation, we give the comparison with the results from the work of Depeyre .
From the given values of mole per cent of single products, from our measurements at
7
00 °C follows (Table III) that they are in a very good agreement both with the ex-
perimental and with simulated yields of products of pyrolysis of hexadecane. The only
3
0
M
2
1
0
0
2
1
3
4
5
6
FIG. 6
Formation of 1-alkenes M (mol/100 mol) as a
function of residence time τ (s) at 760 °C; C H
(1), C6H12 (2), C7H14 (3), C8H16 (4), C12H24 (5),
0.3 C15H30 (6)
5
10
0
0
0.1
0.2
τ, s
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Pyrolysis of Hexadecane
1067
TABLE III
Comparison of experimental results of hexadecane pyrolysis with published data
mole %
Component
a
b
c
exp
exp
(sim)
H2
10.6
6.6
1.5
25.9
0.3
7.9
4.3
1.1
2.5
3.2
1.7
1.2
1.1
1.0
0.9
0.7
0.6
0.5
0.2
24.6
–
3.6
9.2
4.1
27.1
0.5
9.2
3.9
1.3
3.5
1.2
1.1
1.2
1.2
1.2
1.0
0.8
0.7
0.6
0.2
28.4
(3.0)
(9.2)
(1.9)
(23.4)
(1.2)
(10.7)
(3.5)
(1.9)
(2.8)
(2.1)
(1.7)
(1.3)
(1.1)
(0.9)
(0.7)
(0.6)
(0.4)
(0.3)
(0.2)
(28.5)
–
CH4
C2H6
C2H4
C3H8
C3H6
C4H8
C4H6
C5H10
C6H12
C7H14
C8H16
C9H18
C10H20
C11H22
C12H24
C13H26
C14H28
C15H30
C16H34
Benzene
CO
0.05
2.6
0.8
–
–
–
CO2
–
a
b
13
c
13
t = 700 °C, x = 36.5%, τ = 0.0586 s; ref. ; simulated, t = 650 °C, x = 33.9%, τ = 0.25 s (ref. ).
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

1
068
Bartekova, Bajus:
significant difference is observed as to the hydrogen formation whose content is nearly
triple with regard to the results by Depeyre and as to the ethane formation whose con-
tent is a little lower. This difference can be caused by more factors.
At higher temperatures (our measurements at 700 °C, the published ones at 650 °C),
the dehydrogenation of ethyl radical (Eq. (E)) in comparison with reaction (C) takes
priority. Our experiments were performed in the steel reactor whose wall had a cata-
lytic effect on the course of secondary reactions unlike the quartz reactor used by
Depeyre. In the steel reactor, probably, a part of ethyl radicals changes to methylene
and methyl radicals. The methylene radicals may then react in the presence of the metal
6
with water on forming carbon monoxide and hydrogen . Hydrogen can be also formed
owing to the reaction of steam with hydrocarbon radicals or of steam with coke de-
posited on the reactor wall (Eqs (F) and (G)).
C
C
+
+
H O
CO
+
+
H 2
(F)
(G)
2
2 H O
CO2
2 H2
2
From the comparison of the 1-alkene content in the series from 1-pentene to 1-pen-
tadecene follows that their content decreases with increasing number of carbon atoms
in the molecule, the exception being 1-hexene. Its higher content in comparison with
neighbouring 1-alkenes can be a confirmation of the highest rate of isomerization of
•
terminal alkyl radicals by the 1,5-hydrogen transfer. In this way, the C–C–C–C–C –C–R
radical is formed whose β-scission, 1-hexene is formed with the greater probability, the
longer is the chain in the alkyl R. This our observation is in agreement with the results
1
4
of Billaud and Freund who found a high yield of 1-hexene in pyrolysis of n-decane at
8
10 °C.
SYMBOLS
E
k
P
t
activation energy, kJ mol–1
rate constant of first-order reaction, s
reaction pressure, MPa
–
1
reaction temperature, °C
T
Vx=0
Vx=1
x
ε
τ
reaction temperature, K
reaction mixture volume at zero conversion, dm
reaction mixture volume at unit conversion, dm
conversion of hexadecane
expansion factor
3
3
residence time, s
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Pyrolysis of Hexadecane
1069
REFERENCES
1
2
3
4
5
6
. Fairburn J. A., Behie L. A., Svrcek B. W.: Fuel 69, 1537 (1990).
. Wu G., Katsumura Y., Matsuura C., Ishigure K., Kubo J.: Ind. Eng. Chem. Res. 35, 4747 (1996).
. Khorasheh F., Gray M. R.: Ind. Eng. Chem. Res. 32, 1853 (1993).
. Blouri B., Hamdan F., Herault D.: Ind. Eng. Chem., Process Des. Dev. 24, 30 (1985).
. Depeyre D., Flicoteaux C., Chardaire C.: Ind. Eng. Chem., Process Des. Dev. 24, 1251 (1985).
. Bajus M., Vesely V., Leclercq P. A., Rijks J. A.: Ind. Eng. Chem., Prod. Res. Dev. 18, 30
(
1979).
. Hougen O. A., Watson K. M.: Chemical Process Principles, Vol. III, p. 884. Wiley, New York
947.
7
1
8
9
. Ford T. J.: Ind. Eng. Chem., Fundam. 25, 240 (1986).
. Ilavsky J., Valtyni J., Brunovska A., Surovy J.: Aplikovana chemicka kinetika a teoria reaktorov,
Vol. I. Alfa, Bratislava 1990.
1
1
1
1
1
0. Zhou P., Hollis O. L., Crynes B. L.: Ind. Eng. Chem. Res. 26, 846 (1987).
1. Kossiakoff A., Rice F. O.: J. Am. Chem. Soc. 65, 590 (1943).
2. Fabuss B. M., Smith J. O., Satterfield C. N.: Adv. Pet. Chem. Refin. 9, 157 (1964).
3. Depeyre D., Flicoteaux C.: Ind. Eng. Chem. Res. 30, 1116 (1991).
4. Billaud F., Freund E.: Ind. Eng. Chem., Fundam. 25, 433 (1986).
Collect. Czech. Chem. Commun. (Vol. 62) (1997)