Kinetics and modelling of heptane steam-cracking

Article abstract of DOI:10.2478/s11696-013-0518-2

The kinetics and product distribution during the cracking of heptane in the presence of steam were investigated. The experiments were performed in a flow reactor under atmospheric pressure in a temperature range of 680-760°C with a mass ratio of steam to heptane of 3: 1. The overall decomposition of heptane is represented by a first-order reaction with activation energy of 249.1 kJ mol^-1 and a frequency factor of 3.13 × 1013 s^-1. The reaction products were analysed using gas chromatography, the main product being ethylene. The molecular reaction scheme, which consists of a primary reaction and 24 secondary reactions between primary products, was used for modelling the experimental product yields. The yields of ethylene and hydrogen were in good agreement; however the experimental yields of propylene were higher than the predicted yields.

Full text of DOI:10.2478/s11696-013-0518-2

Chemical Papers
DOI: 10.2478/s11696-013-0518-2
ORIGINAL PAPER
Kinetics and modelling of heptane steam-cracking
a
a
b
b
aNatália Olahová*, Martin Bajus, Elena Hájeková, Lukáš Šugár, Jozef Markoš
aDepartment of Petroleum Technology and Petrochemistry, bDepartment of Chemical and Biochemical Engineering,
Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava,
Radlinského 9, 812 37 Bratislava, Slovakia
Received 26 May 2013; Revised 15 October 2013; Accepted 22 October 2013
The kinetics and product distribution during the cracking of heptane in the presence of steam
were investigated. The experiments were performed in a flow reactor under atmospheric pressure in
a temperature range of 680–760◦C with a mass ratio of steam to heptane of 3 : 1. The overall decom-
position of heptane is represented by a first-order reaction with activation energy of 249.1 kJ mol−1
and a frequency factor of 3.13 × 1013
s
−1. The reaction products were analysed using gas chro-
matography, the main product being ethylene. The molecular reaction scheme, which consists of a
primary reaction and 24 secondary reactions between primary products, was used for modelling the
experimental product yields. The yields of ethylene and hydrogen were in good agreement; however
the experimental yields of propylene were higher than the predicted yields.
c
ꢀ 2013 Institute of Chemistry, Slovak Academy of Sciences
Keywords: steam-cracking, heptane, kinetics, product distribution, molecular model
Introduction
ature, enhance the rate of the radical conversion, in-
crease the flexibility of the pyrolysis process, improve
selectivity, and use a feedstock with different proper-
ties. Among the compounds capable of influencing the
thermal decomposition process are some inorganic or
organic compounds containing nitrogen, oxygen, sul-
phur, and phosphorus (Chakraborty & Kunzru, 2012).
Sulphur-based additives can affect the process either
favourably or unfavourably in various ways (Bajus,
1989; Jazayeri & Karimzadeh, 2011; Reyniers & Fro-
ment, 1995).
Various investigators have attempted to model
the pyrolysis of hydrocarbons at three levels of com-
plexity (Karaba et al., 2013; Berreni & Wang, 2011;
Bounaceur et al., 2002; Dente & Ranzi, 1983; Fabuss et
al., 1964; Savage, 2000). The simplest models present
a relation between the product yields and certain pa-
rameters as a function of average residence time and
reactor temperature. At the other extreme are the
mechanistic models that are based on fundamental
free radical reaction kinetics, represented by a high
incidence of radical and molecular reactions. Stud-
ies carried out by several groups of researchers have
Steam-cracking of hydrocarbons is one of the most
important processes for producing light olefins and
aromatics, being the building blocks in the petro-
chemical industry. Currently, ethylene is produced al-
most exclusively via steam-cracking in tubular reactor
coils installed in externally fired heaters. The proce-
dure is performed at high temperatures (800–900◦C),
short residence times (0.1–0.6 s), and low hydrocarbon
partial pressure in tubular reactors, where the reac-
tor coils are constructed of heat-resistant alloys (Fe–
Cr–Ni). Olefins are mainly obtained by the thermal
fission of hydrocarbons from natural gas and crude
oil. Single-component feedstocks, such as ethane or
propane, or multi-component hydrocarbon feedstocks,
such as natural gas liquids, naphtha, and gas oils from
crude oil, may be used for the production of olefins
(Kapur, 2005).
The advantages and disadvantages of pyrolysis are
well-known. One feasible way of improving the pro-
duction of lower alkenes is pyrolysis in the presence
of chemical initiators to reduce the pyrolysis temper-
*Corresponding author, e-mail: natalia.olahova@UGent.be

ii
N. Olahová et al./Chemical Papers
demonstrated the general free radical nature of the
pyrolysis reactions together with several defined crit-
ical reaction steps. These findings have led to an ex-
tensive compilation of fundamental kinetic constants
applicable to thermal degradation reactions (Allara &
Shaw, 1980; Kopinke et al., 1987; Sundaram & Fro-
ment, 1978).
reaction of 22 species, 28 reactions, and 25 transi-
tion states. Computational fluid dynamics (CFDs)-
based predictions were presented for non-pre-mixed
and partially pre-mixed flames burning vaporised hep-
tane fuel (Katta et al., 2012). Three state-of-the-art
chemical kinetics models were incorporated into a
time-dependent, two-dimensional CFD model known
as UNICORN.
Aribike and Susu (1988a, 1988b) developed a
mechanistic model for heptane pyrolysis from the
product distribution data from a pulse reactor. How-
ever, the experimental and simulated product distri-
butions differed considerably. The kinetics and prod-
uct distribution during the pyrolysis of heptane were
investigated within a temperature range of 680–750◦C,
at atmospheric pressure, under steam as an inert dilu-
ent (Pant & Kunzru, 1996). The product yields from
the experiment could be satisfactorily modelled by a
molecular reaction scheme consisting of a first-order
primary reaction and 24 secondary reactions among
the primary products.
In pyrolysis units, surface reactions are relatively
more important in smaller diameter tubes. Previous
studies (Albright & Tsai, 1983; Bajus & Veselý, 1974)
showed that these surface reactions produced the fol-
lowing results: a) selectivity was higher towards the
production of olefins and diolefins; b) coke was pro-
duced of which part was gasified and the surface reac-
tions involving steam led to the production of CO and
CO2 and also destroyed the integrity of the metal sur-
face due to carburisation reactions; and c) oxidising–
reducing and sulphiding–desulphiding (Bajus, 1989)
steps caused surface roughening and corrosion.
This study was part of a more extensive project
investigating the effect of sulphur compounds on con-
version and coking during the pyrolysis of hydrocar-
bons. Heptane was chosen as a feedstock as it gives the
highest yield of ethylene and conversion in the range
of C5 to C12 alkanes (Bajus & Veselý, 1976). There-
fore, heptane is particularly suitable to study kinetics
and mechanisms of the pyrolysis at different reaction
rates, and in reaction systems in which wall effects
are favourable for the selectivity of process. Moreover,
very few data are available on the kinetics and product
distribution of heptane pyrolysis.
The present study sought to investigate the ef-
fects of temperature and space time on the prod-
uct yields during heptane steam-cracking over a wide
range of conversion. A further objective was to develop
a molecular model based on the experimental product
distributions.
Experimental
A flow pyrolysis reactor was used, with a de-
tailed description of the equipment given in a previous
work (Hájeková & Bajus, 2005). The flow reactor was
manufactured from stainless steel (composition: 23–27
mass % Cr, 18–22 mass % Ni, max. 2 mass % Si, max.
1.5 mass % Mn, max. 0.25 mass % C) and it was of
the tubular type. The tube, 750 mm in length, was
located in an electrically heated furnace. The internal
diameter of the outer tube was 12 mm and the exter-
nal diameter of the inner tube was 6 mm. The smaller
characteristic dimensions in the equipment afford an
extremely high internal surface area-to-volume ratio
(SA/V), here representing 6.65 cm−1. Demountable
thermocouples (Ni/Cr–Ni) were placed in the inner
tube. The flow of heptane was from 15 g h−1 to 35 g
h−1 and that of water from 45 g h−1 to 105 g h−1. The
mass ratio of distilled water to heptane of 3 g g−1 was
constant in each experiment. Once the reaction was
complete, the products from the pyrolysis of heptane
were passed through a water cooler and were cooled
in a freezing trap.
Murata et al. (1973) studied the pyrolysis of hep-
tane at 700◦C and developed a molecular model to
explain the product distribution. Bajus et al. (1979)
investigated the thermal decomposition of heptane in
the presence of steam in a flow reactor in the temper-
ature range of 680–760◦C. The overall reaction was
approximately of first order with activation energy of
195.5 kJ mol−1, no model was proposed. An experi-
mental study of heptane pyrolysis was performed at
low pressure in a temperature range from 507◦C to
1507◦C. The mole fractions measured indicated that
H2, CH4, C2H2, and C2–C6 alkenes were major prod-
ucts of heptane pyrolysis. On the basis of the experi-
mental observation and theoretical calculation, the py-
rolysis channels of mono-molecular dissociation were
proposed to comprehend the pyrolysis process of hep-
tane (Yuan et al., 2011).
The thermal rate constants of the heptane py-
rolysis are also important in the combustion process
because thermal decomposition of fuel inevitably ac-
companies fuel oxidation (Ding et al., 2013). Using
statistical-theoretical methodology, the mechanism of
heptane pyrolysis as well as the temperature and
pressure dependence of the rate constants were stud-
ied in some detail (Ding et al., 2013). The pyrol-
ysis mechanism includes the important primary re-
action, β-scission reaction and H-atom abstraction
During the steam-cracking of heptane, qualita-
tive and quantitative analyses of gaseous and liquid
products were performed using the gas chromatogra-
phy method. The gases obtained by steam-cracking
were analysed on an HP 6890+ gas chromatograph
(Hewlett–Packard, Wilmington, NC, USA) equipped
with three PLOT chromatographic columns: HP

N. Olahová et al./Chemical Papers
iii
PLOT Al2O3 “M” (50 m × 0.53 mm); HP PLOT
Q (30 m × 0.53 mm); HP PLOT Molsieve 5A
(30 m × 0.53 mm), FID and TCD detectors with
three switching valves. The chromatographic system
afforded a detailed separation of the mixture of C1–C5
gaseous hydrocarbons, saturated and unsaturated hy-
drocarbons, hydrogen, and carbon oxides in one sam-
ple.
The analyses of liquids formed during the steam-
cracking of heptane were performed on a CHROM
V gas chromatograph (Laboratorní Přístroje, Prague,
Czech Republic). It was equipped with an FID de-
tector and a polydimethylsiloxane capillary chro-
matographic column DB PETRO (50 m × 0.2 mm
i.d. × 0.5 m). Helium was used as a carrier gas. A
detailed description of the gas chromatographic anal-
ysis of products derived from steam-cracking is given
in a previous work (Hájeková et al., 2007).
Fig. 1. Dependence of reactor temperature
T
on reactor
length L for different reference temperatures, TR/◦C:
680 ( ), 700 ( ), 720 (ꢀ), 740 (×), and 760 (∗).
Theoretical
Reaction kinetics
Space time and reference reaction temperature
The effects of temperature and space time on the
conversion, selectivity and product yields in heptane
pyrolysis at atmospheric pressure were investigated.
The values of the reaction parameters varied in the
following ranges: temperature from 680◦C to 760◦C;
The equivalent volume VR of the reactor was deter-
mined in accordance with Hougen and Watson (1947);
it is defined as the volume which, at the reference tem-
perature TR, gives the same conversion as the real re-
actor with its temperature profile. All data are related
to this temperature.
mass flow-rate of heptane from 15 g h−1 to 35 g h−1
,
flow-rate of water from 30 g h−1 to 105 g h−1, and
space time from 0.055 s to 0.145 s. The inlet steam-
rT dVR = rTdV
(1)
to-heptane ratio was maintained constant at 3 g g−1
.
R
The present study’s primary objective was to in-
vestigate the effects of temperature and space–time re-
lation on product yields from heptane steam-cracking
over a wide conversion range. It was necessary to de-
sign a molecular model from the experimental product
distributions.
ꢀ
ꢁ
ꢂ
V
1
E
ꢀ
ꢁ
VR
=
exp
−
dV
(2)
E
RT
0
exp
−
RTR
The heptane pyrolysis process can be divided into
two steps: i) primary reaction, which involves heptane
decomposition into primary products; ii) secondary re-
actions, which involve consecutive and/or parallel de-
composition and interaction of the primary products
of heptane pyrolysis. Table 1 summarises these reac-
tions. The proposed scheme includes 24 reactions in-
cluding 18 species. The reaction rate constants and
their temperature dependence (given by Arrhenius
equation) have also been calculated and the respec-
tive parameters are given in Table 1.
Using the original experimental data, the aims of
this study were: i) to estimate stoichiometric coeffi-
cients of the primary reaction (Eq. (4)); ii) to esti-
mate the reaction rate constant of the primary reac-
tion, e.g. frequency factor and activation energy; and
iii) to simulate the process under study using a simple
mathematical model of the system investigated using
parameters obtained in the first two steps and kinetic
parameters achieved by Pant and Kunzru (1996) and
to compare the simulated and experimental data.
where E is activation energy, R is universal gas con-
stant, TR is reference temperature, T is measured tem-
perature, VR is equivalent reactor volume, rT is reac-
R
tion rate at reference temperature, and rT is reaction
rate at measured temperature.
In order to calculate the effective reactor volume,
the longitudinal temperature profile in the reactor
must be known (Fig. 1). It was assumed that the
activation energy for the pyrolysis of heptane was
251.2 kJ mol−1. This represents an average value of
activation energy for splitting C—C bonds in hydro-
carbons (Bajus et al., 1979). The equivalent volume
calculated according to Eqs. (1) and (2) was substi-
tuted with the equation for the space time, τ,
˙
τ = VR/VF
(3)
˙
where VF is the volumetric feed flow rate. The refer-
ence temperature was the maximum temperature in
the reactor for a given experiment.

iv
N. Olahová et al./Chemical Papers
Table 1. Reaction scheme for heptane pyrolysis according to Pant and Kunzru (1996)
Pre-exponential factor
Activation energy
kJ mol−1
Reaction
s−1
C7H16 → 0.23H2 + 0.59CH4 + 1.69C2H4 + 0.15C2H6
+ 0.104C3H8 + 0.44C3H6 + 0.13C4H8 + 0.1C4+
C2H6 ↔ C2H4 + H2
+
3.13 × 1013
249.1
4.65 × 1013
7.28 × 1012
1.03 × 109a
3.75 × 1012
7.08 × 1010a
5.89 × 1010
4.69 × 1010
2.54 × 1010a
7.38 × 1012a
2.50 × 1011a
273.0
273.5
172.7
273.1
253.0
214.7
211.8
247.2
268.6
228.1
C3H6 ↔ C2H2 + CH4
C2H2 + C2H4 → C4H6
2C2H6 → C3H8 + CH4
C2H4 + C2H6 → C3H6 + CH4
C3H8 ↔ C3H6 + H2
C3H8 → C2H4 + CH4
C3H8 + C2H4 → C2H6 + C3H6
2C3H6 → 3C2H4
2C3H6 → 0.3CnHb2n−6 + 0.14C4 + 3CH4
C3H6 + C2H6 → C4H8 + CH4
1.00 × 1011a
7.00 × 1012
7.00 × 1014
4.10 × 1012
1.64 × 1012
2.08 × 1011
251.2
249.7
295.9
256.6
261.0
212.2
C4H10 → C3H6 + CH4
C4H10 → 2C2H4 + H2
C4H10 → C2H4 + C2H6
C4H10 ↔ C4H8 + H2
C4H8 → 0.41CnHb2n−6 + 0.19C4+
C4H8 ↔ C4H6 + H2
1.00 × 1010
8.38 × 106a
9.74 × 105a
6.40 × 1011a
1.50 × 106a
9.10 × 1012
2.35 × 1012
4.00 × 1011
209.3
144.7
149.2
242.4
124.3
226.0
230.2
230.2
C2H4 + C2H6 → Bb + 2H2
C4H6 + C3H6 → Tb + 2H2
C4H6 + C4H8 → EBb + 2H2
2C4H6 → STb + 2H2
C4+ → 1.16(C2H4 + C3H6)
C4+ → 1.16(CH4 + C2H6)
C4+ + H2 → 1.16(C4H8 + CH4)
a) In m3 mol−1 −1; b) B: benzene; T: toluene; EB: ethylbenzene; ST: styrene; CnH2n−6: aromatics.
s
Estimation of stochiometric coefficients of the
primary reaction (Eq. (4))
is stoichiometric coefficient, kV denotes the first order
reaction rate constant, and V is volumetric flow-rate.
˙
Initial conditions for the solution of Eq. (5) is:
The primary reaction of heptane steam-cracking
can be expressed as:
˙
˙
τ = 0 : V = VF, CA = CF,A, FA = FF,A
(6)
k1
C7H16 −−−−→ aH2 + bCH4 + cC2H4 + dC2H6 +
where F refers to feed, C is molar concentration, and
F stands for molar feed flow-rate.
After separation of variables from Eq. (5), the fol-
lowing can be obtained:
+ eC3H8 + fC3H6 + gC4H8 + hC4+
(4)
where k1 denotes the first order reaction rate constant.
The stoichiometric coefficients of the primary re-
action from experimentally obtained dependences of
selectivity for individual primary products on the con-
version were determined. The values of selectivity were
extrapolated to the values of conversion close to zero.
If heptane reacts only in the primary reaction, then
the reaction rate of heptane decomposition could be
assumed to be of the first order with respect to the
reactant. The reaction rate constant can be estimated
from the experimental conversion of heptane varying
the space time and initial temperature in the reactor.
The material balance of heptane assuming plug flow
in the reactor is in the form:
XA
ꢂ
˙
V
˙
τ = (−νA) VFkV
dXA
(7)
(1 − XA)
0
The volumetric flow-rate is changed along the re-
actor because of the increase in the total molar flow of
the reaction mixture caused not only by the primary
reaction but also by the secondary reactions. From the
experimental observations, it can be assumed that the
volumetric flow-rate is changed linearly with the con-
version of the heptane:
˙
˙
V = VF (1 + εXA)
(8)
dXA
dτ
(1 − XA)
˙
= (−νA) VFkV
(5)
where the expansion factor, ε, has been estimated from
the known volumetric flow-rate of the reaction mixture
in the reactor inlet (conversion of heptane is zero) and
˙
V
where A refers to heptane, X stands for conversion, ν

N. Olahová et al./Chemical Papers
v
Table 2. Values of rate constants for each temperature of hep-
Parameters of Eq. (13) are given in Table 1.
The molar concentrations needed for the reaction
rates calculation are obtained from:
tane pyrolysis
Pyrolysis temperature/◦C
Rate constant/s−1
Fi
680
700
720
740
760
0.8
1.3
2.3
4.5
8.7
Ci =
(14)
˙
V
Assuming ideal behaviour of the gas mixture, the
volumetric flow in the reactor is computed as follows:
FRT
˙
V =
(15)
P
under conditions when the conversion of heptane is
complete. The value of the expansion factor varied
from 0.10 to 0.15.
with the reaction mixture total molar flow of:
NI
ꢃ
F =
Fi
(16)
VX=1 − VX=0
ε =
(9)
i=1
VX=0
Because there was insufficient information for en-
thalpy balance development, the reaction rate con-
stants in Eq. (13) were calculated using temperature
values interpolated from the experimental tempera-
ture profile for the given position in the reactor. The
same temperature was also used to calculate the vol-
umetric flow-rate using Eq. (15).
Eq. (12) represents a system of differential equa-
tions which were solved using MATLAB, under the
initial conditions:
where V is volume of the reaction mixture.
Assuming the validity of Eq. (8), Eq. (7) can be
integrated resulting in the integrated form of the ma-
terial balance:
1
kVτ = (1 + ε) ln
− εX
(10)
1 − X
Comparing the experimental values of the heptane
conversion with those calculated from Eq. (10), it is
possible to estimate the value of the reaction rate con-
stant of the primary reaction (see Table 2) for each ref-
erence temperature in the reactor and, consequently,
from these data the activation energy and the fre-
quency factor.
τ = 0 : Fi = FFi, i = heptane, water
(17)
the initial molar flow-rate of all the other species was
set to zero.
Mathematical model of reactor
Results and discussion
The mathematical model used to simulate the be-
haviour of the experimental pyrolysis reactor entailed
NI = 18 species and NR = 25 reactions. Assuming the
plug flow in the reactor, the material balance for each
species can be expressed as follows:
As well as the channel geometry and fluid proper-
ties, the viscosity, µ, and Reynolds number, Re, were
determined in the reactor channel. The viscosity val-
ues for all the experiments and temperatures were ob-
tained by calculations proposed by Reid et al. (1988)
and they varied around the value of (5.1 0.1) × 10−5
Pa s. Depending on the viscosity, the Reynolds num-
ber is predicted as:
NR
ꢃ
dFi
dV
˙
j,iξVj
=
ν
,
i = 1, ..., NI
(11)
j=1
˙
dhρU
where ξV is reaction rate, NI is numbers of compo-
nents, and NR is number of reactions.
Re =
(18)
µ
˙
Using the space time dV = VFdτ, Eq. (11) become:
where dh is the channel diameter, ρ is the density of
the fluid, µ is the viscosity of the fluid, and U is the
velocity of the fluid in the channel. The Reynolds num-
bers under the reaction conditions in this study ranged
from 35 to 100 (Fig. 2). Low values of the Reynolds
number indicate the laminar type of flow in the reactor
(Chakraborty & Kunzru, 2012).
The value of the overall reaction order was found
to be one (Bajus et al., 1979). Given this value, the
determination of rate constants and activation energy
NR
ꢃ
dFi
dτ
˙
˙
j,iξVj,
= VF
ν
i = 1, ..., NI
(12)
j=1
Reaction rates are defined as follows:
ꢀ
ꢁ
NI
ꢄ
E
Cia , j = 1, ..., NR
i,j
˙
ξVj = kV∞j exp
−
RT
i=1
(13)

vi
N. Olahová et al./Chemical Papers
Table 3. Product distribution from pyrolysis of heptane
Temperature/◦C
Conversion/%
Space time/s
680
4.8
0.1
680
10.5
0.2
700
6.1
0.1
700
16.8
0.1
720
11.2
0.1
720
29.8
0.1
740
16.6
0.1
740
47.6
0.1
760
28.1
0.1
760
64.3
0.1
Component
Component amount per 100 moles of heptane decomposed/mol
Hydrogen
Methane
Ethane
Ethylene
Propane
Propylene
Butane
1-Butene
2-Butene
23.7
56.6
14.0
108.1
1.3
36.2
0.3
11.6
0.1
26.6
59.4
15.8
178.9
2.2
49.0
0.7
17.0
0.5
26.7
58.3
13.7
181.1
1.1
46.5
0.2
17.6
0.2
31.6
51.1
9.9
164.9
1.0
44.4
0.2
16.9
0.2
32.0
52.3
9.9
169.3
1.1
43.8
0.2
16.8
0.2
1.3
tr
28.9
51.5
9.0
171.4
1.3
48.8
0.3
19.3
0.4
33.3
56.0
9.7
188.5
1.4
51.3
0.2
20.4
0.3
49.6
51.0
8.7
165.9
1.2
47.2
0.2
17.3
0.5
27.4
46.4
7.4
162.8
1.3
44.4
0.2
18.6
0.3
25.7
51.5
7.7
174.5
1.3
48.2
0.2
17.1
0.7
1,3-Butadiene
3-Methylbutane
Pentane
0.6
tr
tr
1.4
tr
0.1
1.3
tr
tr
1.6
tr
tr
2.7
tr
tr
2.4
tr
tr
3.8
0.1
tr
2.9
0.1
tr
5.7
0.1
tr
tr
1-Pentene
trans-2-Pentene
1-Hexene
Carbon monoxide
Carbon dioxide
6.34
tr
1.8
1.2
tr
2.66
tr
0.1
0.4
tr
3.89
0.1
0.7
1.7
tr
8.33
0.1
0.7
0.7
tr
7.28
tr
5.9
0.3
tr
5.34
0.1
2.2
0.6
0.1
4.61
0.1
1.2
0.6
tr
4.36
0.1
1.2
1.1
0.6
7.89
0.1
3.1
0.6
0.1
3.74
0.2
1.0
1.3
0.2
tr = traces
Fig. 3. Gas yield Y as function of heptane conversion X at
different temperatures, T/◦C: 680 ( ), 700 ( ), 720 (ꢀ),
740 (×), and 760 (∗).
Fig. 2. Dependence of equivalent volume of reactor Vr on
Reynolds number, Re, at 680◦C.
it was presumed that the conversion of heptane fol-
lows the first order reaction. The values of the rate
constants were estimated by linear regression from ex-
perimental data and the rate constants for five tem-
peratures are listed in Table 2. From these rate con-
stants, the activation energy and the frequency factor
were obtained, being found to be 249.1 kJ mol−1 and
the product yields and selectivity were investigated.
Among the liquid and gaseous products formed in
the pyrolysis of heptane, there are saturated, olefinic,
and aromatic hydrocarbons of different types. Under
the experimental conditions, the key pyrolysis product
was ethylene; other major products were hydrogen,
methane, propylene, and 1-butene, whereas the mi-
nor products include 1,3-butadiene, ethane, propane,
butane, 1-pentene, 1-hexene, carbon monoxide, car-
bon dioxide, and benzene (at high heptane conversion)
(Table 3). The gaseous pyrolysis products have an av-
3.13 × 1013
s
−1. The value of activation energy cal-
culated from the pyrolysis of heptane is in excellent
agreement with the presumed value (251.2 kJ mol−1),
which was used in the calculations of equivalent vol-
ume (Eq. (2)).
erage molecular mass of approximately 28.7 g mol−1
.
It is clear that the amount of gas formed during steam-
cracking of heptane increases with the increasing tem-
perature and conversion (Fig. 3).
Composition of products
The effects of the temperature and space time on
Selectivity is defined as the number of moles of

N. Olahová et al./Chemical Papers
vii
Table 4. Comparison of product selectivities (amount of product in moles per 100 moles of heptane decomposed) with published
literature in pyrolysis of heptane
Product selectivity, S · 102
H2
CH4
C2H4
C2H6
C3H6
1-C4H8
1-C5H10
1-C6H12
Kossiakoff and Rice (1943)
Murata et al. (1973)
Bajus et al. (1979)
Appleby et al. (1947)
Pant and Kunzru (1996)
Chakraborty and Kunzru (2009)
Present results
–
43
51
5
–
18
23
47
47
50
46
42
43
60
96
125
117
84
122
89
53
17
6
46
10
35
15
36
44
32
41
31
44
44
27
27
21
33
21
25
13
17
15
13
11a
18
9
7
6
–
11
6
15
10a
170
–
a) C5 and C6 combined.
ious products increase slowly with the temperature of
the pyrolysis.
The percentage yields of ethylene, propene and
methane were observed to increase with an increasing
degree of conversion. Fig. 5 depicts the dependence
of the ethylene, propylene and methane yields on the
conversion. The yields appear to increase with con-
version in a linear manner. By contrast, in the case
of 1-hexene, the yields at higher temperatures did not
change and sometimes even decreased (Table 3).
Mathematical model
Hydrocarbon pyrolysis has long been the subject
of study. There is now a general trend towards the
development and use of detailed kinetic models for a
wide range of feedstocks. Thus, there is an increasing
need for the development of well-validated and reliable
models to represent the pyrolysis of hydrocarbons or
complex mixtures (Savage, 2000) There is no doubt
that a pyrolysis model is essential for the simulation,
control, and optimisation of an ethylene-cracking fur-
nace. The modelling method could be divided into
three types: the empirical model, the molecular re-
action kinetics model, and free-radical reaction mech-
anism model (Fabuss et al., 1964). Pant and Kunzru
(1996) constructed a molecular model from the pyrol-
ysis of heptane, consisting of an overall primary re-
action together with a set of molecular reactions be-
tween the primary products. Unlike Pant and Kunzru
(1996), the present results detected the initial selec-
tivity towards hydrogen with the value of 0.23 mole
of hydrogen formed per one mole of heptane decom-
posed. The primary reaction measurements can be
represented as:
Fig. 4. Variation of ethylene (ꢀ), propylene ( ), and methane
( ) selectivities (S, product amount in moles per 100
moles of heptane decomposed) with space time τ.
Fig. 5. Dependence of ethylene ( ), propylene ( ), and metha-
ne (ꢀ) yields Y on conversion X.
products formed per 100 moles of heptane decom-
posed, whereas the percentage yield of a product is
based on the mass of heptane fed into the reactor.
Fig. 4 shows the variation of the ethylene, propylene
and methane selectivities with space time at 720◦C.
With a constant temperature, space time had no ob-
servable effect on selectivities. The selectivities of var-
k1
C7H16 −−−−→ 0.23H2 + 0.59CH4 + 1.69C2H4 +
+ 0.15C2H6 + 0.104C3H8 + 0.44C3H6 +
+ 0.13C4H8 + 0.1C4+
(19)
C4+ represents all products heavier than C4. Ta-
ble 4 shows a comparison of the experimental results of

viii
N. Olahová et al./Chemical Papers
Fig. 8. Comparison between experimental (Yexp) and calcu-
lated (Ycalc) yields of hydrogen.
Fig. 6. Comparison between experimental (Xexp) and calcu-
lated (Xcalc) conversion values of heptane.
ethyl, or propyl) and high yields of low alkenes (ethy-
lene, propylene). Bimolecular hydrogen abstraction
between low radicals and heptane, which proceeded
to a lesser extent under the current experimental con-
ditions, led to the formation of four types of heptyl
radicals in the system. Higher 1-alkenes are formed
by heptyl radical decomposition as:
˙
˙
CH3(CH2)3CHCH2CH3 → 1-C6H12 + CH3 (20)
˙
˙
CH3(CH2)3CHCH2CH3 → 1-C4H8 +C2H4 +C H3(21)
˙
˙
CH3(CH2)2CH(CH2)2CH3 → 1-C5H10 + C2H5 (22)
As noted above, the set of molecular reactions de-
veloped by Pant and Kunzru (1996) was used as a
guide for developing the current model. The reac-
tion scheme for heptane pyrolysis is shown in Ta-
ble 1 and contains the primary reaction with the pre-
exponential factor and activation energy determined
numerically.
Fig. 7. Comparison between experimental (Yexp) and calcu-
lated (Ycalc) yields of ethylene.
the initial selectivity of products obtained with those
published in the literature. A markedly higher selec-
tivity towards methane and ethylene and lower selec-
tivity towards higher 1-alkenes (1-C4H8, 1-C5H10, 1-
C6H12) were observed in comparison with the selec-
tivity reported by previous researchers and the selec-
tivities calculated in accordance with Rice–Kossiakoff
theory (Kossiakoff & Rice, 1943). These findings can
be attributed to the low partial pressure of heptane
due to the excessive steam dilution used in the cur-
rent experiments. Under the experimental conditions,
namely high temperature and low substrate concen-
trations, the monomolecular reactions (β-scission, iso-
merisation) are favoured over the bimolecular hydro-
gen abstraction step or the bimolecular radical addi-
tion to olefin (Pant & Kunzru, 1996).
The experimental yields and conversions were com-
pared with the values predicted by the reactor model
developed and described earlier. Fig. 6 shows a com-
parison between the experimental and calculated con-
versions; it is apparent that the differences are minute.
Figs. 7–9 show comparisons between the experimental
and the calculated yields of ethylene, hydrogen, and
methane, respectively. While the formation of ethylene
and hydrogen accord well with the calculated yields,
the predicted yields of methane during the pyrolysis
at higher conversions are higher than the experimental
values, while for ethane this tendency is even stronger
(Fig. 10). The presumed reaction mechanism for the
formation of methane and ethane was satisfactory only
within the area of low conversion values. On the other
hand, when it comes to a comparison between the ex-
perimental and predicted yields of propylene, there are
significant differences. There are substantially higher
experimental yields of propylene than the yields pre-
During the initiation step of heptane pyrolysis,
the heptane molecule undergoes the fission of C—C
bonds to form two primary radicals. It is assumed,
as noted earlier, that in the propagation step these
radicals prefer to split themselves in the β-position
to the radical centre and form lower radicals (methyl,

N. Olahová et al./Chemical Papers
ix
Fig. 11. Comparison between experimental (Yexp) and calcu-
lated (Ycalc) yields of propylene.
Fig. 9. Comparison between experimental (Yexp) and calcu-
lated (Ycalc) yields of methane.
Fig. 12. Comparison between experimental (Yexp) and calcu-
lated (Ycalc) yields of 1-butene.
Fig. 10. Comparison between experimental (Yexp) and calcu-
lated (Ycalc) yields of ethane.
dicted by modelling (Fig. 11). Similarly, in the case of
1-butene and C4+, the experimental yields were higher
than the predicted yields, especially at the higher con-
version values (Figs. 12 and 13).
There exist some possible explanations. Rice and
Herzfeld (1934) clarify this type of chain reaction and
explain that, during propagation, the radical abstracts
an atom of hydrogen from the reactant molecule which
decomposes into a molecule and a new radical. The
newly formed radical further reacts by β-scission to re-
produce the abstracting radical and to form a molecule
with a double bond with the carbon atom that was a
radical centre.
The mechanism proposed by Fabuss–Smith–Sat-
terfield (Fabuss et al., 1964) is another type of mech-
anism which predominates at mild temperatures and
high concentrations of substrate. The radical β formed
by β-scission abstracts hydrogen from the substrate
rather than breaking a β-bond. This bimolecular ab-
straction has lower activation energy than the scission
of a β-bond. This mechanistic regime is encountered
Fig. 13. Comparison between experimental (Yexp) and calcu-
lated (Ycalc) yields of C4+
.
in a liquid-phase pyrolysis.
On the other hand, the Rice–Kossiakoff theory

x
N. Olahová et al./Chemical Papers
(Kossiakoff & Rice, 1943) favours β-scission and iso-
merisation (mono-molecular steps) over hydrogen ab-
straction (bi-molecular step). The concentrations of
substrate are very low and the temperatures are very
high (gas-phase). If the β-scission is faster, then the
radical eliminates the short molecules, e.g. ethylene
and propylene. If the isomerisation, which produces
stable radicals, is faster, then higher hydrocarbons and
less ethylene are formed. The bi-molecular hydrogen
abstraction step leads to the formation of four types
of heptyl radicals in the system: 1-heptyl, 2-heptyl,
3-heptyl, and 4-heptyl. Bajus et al. (1979) described
four routes for higher olefin formation from heptane.
The decomposition of four C7H15 radicals proceeds as
follows:
(pyrolysis) in microreactor configuration via a numer-
ical approach. The data on the longitudinal tempera-
ture profile of the microreactor demonstrated that the
dependence of the temperature on the reactor length
exhibited a parabolic profile. The overall decomposi-
tion of heptane in a temperature range from 680◦C to
760◦C can be represented by a first-order reaction with
a frequency factor of 3.13 × 1013 s−1 and activation
energy of 249.1 kJ mol−1
.
The pyrolysis reactor was modelled on a simple
mathematical model, using the concept of a primary
reaction of heptane decomposition followed by the
24 parallel and consecutive reactions, as proposed by
Pant and Kunzru (1996). Using the kinetic parame-
ters of the primary reaction estimated from the cur-
rent experiments and kinetic parameters of secondary
reactions, a good agreement between simulation and
experiment was achieved, especially between the ex-
perimental and calculated conversion of heptane and
yields of ethylene and hydrogen. The model could pos-
sibly be improved by estimating the kinetic parame-
ters of all the reactions from the experimental data.
As stated above (see Experimental), the laboratory
reactor had a very high internal surface-to-volume ra-
tio. Because the internal surface of the reactor tube
can play an important role in the reaction mechanism
(formation of radicals, termination of radicals, etc.)
this can influence the kinetic parameters.
˙
˙
CH3(CH2)5CH2 → 3C2H4 + CH3
(23)
˙
˙
CH3(CH2)4CHCH3 → C3H6 + C2H4 + C2H5 (24)
˙
˙
CH3(CH2)3CHCH2CH3 → 1-C6H12 + CH3 (25)
˙
˙
CH3(CH2)3CHCH2CH3 → C2H4 + 1-C4H8 + CH3(26)
˙
˙
CH3(CH2)2CH(CH2)2CH3 → 1-C5H10 + C2H5 (27)
It appears that, in the present system, reactions
(Eqs. (23) and (24)) are preferred. In a reaction sys-
tem with a high ratio of surface area to volume SA/V,
formation of the 2-heptyl radical is more likely. This
is another explanation for the higher yields of propy-
lene. Chakraborty and Kunzru (2009) investigated the
high-pressure pyrolysis of heptane in a tubular reactor.
With increasing pressure, the selectivities of propane,
butane and 1-butene also increased, but those of hy-
drogen, methane, ethylene, and propylene decreased.
This can be attributed to the fact that the inlet par-
tial pressure was high (61.8 kPa), the reaction tem-
peratures were lower and the catalytic activity of the
inner surface was minimised by the addition of CS2
(0.02 mole %) to the feed in each run. In contrast to
the above, the current experiments were carried out
under high temperatures and with steam used as an
inert diluent. The decreased formation of ethane dur-
ing the decomposition of heptane in the stainless steel
flow tubular reactor could be ascribed to the catalytic
activity of the wall. The inner surface of the stain-
less steel reactor wall exerted a catalytic effect leading
to the dehydrogenation of ethane or ethyl radicals to
ethylene (Bajus et al., 1979):
Acknowledgements. We would like to express our gratitude
to the VEGA Scientific Grant Agency of the Slovak Republic
for providing financial support for this investigation through
Research Project no. 1/0228/12.
Symbols
CA
molar concentration of heptane
mol m−3
CF,A molar concentration of heptane in
feed
mol m−3
J mol−1
mol s−1
s−1
E
activation energy
FF,A molar feed rate of heptane
kV, k1 first order reaction rate constant
NI
NR
P
R
Re
rT
number of components
number of reactions
pressure
universal gas constant
Reynolds number
reaction rate at measured
temperature
Pa
J mol−1 K−1
mol m−3 s−1
mol m−3 s−1
rT
reaction rate at reference
temperature
R
˙
˙
S
SA
T
TR
V
selectivity
surface area
temperature
reference temperature
volume
equivalent reactor volume
volumetric flow-rate
C2H5 → C2H4 + H
(28)
m2
◦C
◦C
Conclusions
m3
VR
m3
This work simulates the non-isothermal homoge-
nous decomposition of heptane during steam-cracking
m3 s−1
˙
V

N. Olahová et al./Chemical Papers
xi
˙
VF
XA
ε
ν
ξ
˙
ξV
τ
volumetric feed flow-rate
conversion of heptane
expansion factor
stoichiometric coefficient
reaction extent
m3 s−1
perature. The Journal of Physical Chemistry A, 117, 3266–
3278. DOI: 10.1021/jp311498u.
Fabuss, B. M., Smith, J. O., & Satterfield, C. N. (1964). Ther-
mal cracking of pure saturated hydrocarbons. In J. J. McK-
etta, Jr. (Ed.), Advances in petroleum chemistry and refining
(Vol. 9, pp. 156–201). New York, NY, USA: Wiley.
Hájeková, E., & Bajus, M. (2005). Recycling of low-density
polyethylene and polypropylene via copyrolysis of polyalkene
oil/waxes with naphtha: product distribution and coke for-
mation. Journal of Analytical and Applied Pyrolysis, 74,
270–281. DOI: 10.1016/j.jaap.2004.11.016.
reaction rate
space time
mol m−3 s−1
s
References
Hájeková, E., Mlynková, B., Bajus, M., & Špodová, L. (2007).
Copyrolysis of naphtha with polyalkene cracking products;
the influence of polyalkene mixtures composition on product
distribution. Journal of Analytical and Applied Pyrolysis, 79,
196–204. DOI: 10.1016/j.jaap.2006.12.022.
Albright, L. F., & Tsai, T. C. (1983). Importance of surface
reactions units. In L. F. Albright, B. L. Crynes, & W. H.
Corcoran (Eds.), Pyrolysis: Theory and industrial practice
(pp. 233–254). New York, NY, USA: Academic Press.
Allara, D. L., & Shaw, R. (1980). A compilation of kinetic pa-
rameters for the thermal degradation of n-alkane molecules.
Journal of Physical and Chemical Reference Data, 9, 523–
560. DOI: 10.1063/1.555623.
Appleby, W. G., Aver, W. H., & Meerbott, W. K. (1947). Ki-
netics and mechanism of the thermal decomposition of n-
heptane. Journal of the American Chemical Society, 69,
2279–2285. DOI: 10.1021/ja01202a012.
Aribike, D. S., & Susu, A. A. (1988a). Kinetics and mechanism
of the thermal cracking of n-heptane. Thermochimica Acta,
127, 247–258. DOI: 10.1016/0040-6031(88)87501-4.
Aribike, D. S., & Susu, A. A. (1988b). Mechanistic modeling of
the pyrolysis of n-heptane. Thermochimica Acta, 127, 259–
273. DOI: 10.1016/0040-6031(88)87502-6.
Bajus, M., & Veselý, V. (1974). CS Patent No. 175812. Prague,
Czechoslovakia: Czechoslovak Patent and Trademark Office.
Bajus, M., & Veselý, V. (1976). Hydrocarbon pyrolysis. I. Py-
rolysis of individual n-alkanes. Ropa a Uhlie, 18, 126–135.
Bajus, M., Veselý, V., Leclercq, P. A., & Rijks, J. A. (1979).
Steam cracking of hydrocarbons. 1. Pyrolysis of heptane. In-
dustrial & Engineering Chemistry Product Research and De-
velopment, 18, 30–37. DOI: 10.1021/i360069a007.
Bajus, M. (1989). Sulfur compounds in hydrocarbon pyrolysis.
Sulfur reports, 9, 25–71. DOI: 10.1080/01961778908047982.
Berreni, M., & Wang, M. H. (2011). Modelling and dynamic op-
timization of thermal cracking of propane for ethylene man-
ufacturing. Computers & Chemical Engineering, 35, 2876–
2885. DOI: 10.1016/j.compchemeng.2011.05.010.
Hougen, O. A., & Watson, K. M. (1947). Chemical process prin-
ciples (Vol. 3). New York, NY, USA: Wiley.
Jazayeri, S. M., & Karimzadeh, R. (2011). Experimental in-
vestigation of initial coke formation over stainless steel,
chromium, and iron in thermal cracking of ethane with hy-
drogen sulfide as an additive. Energy & Fuels, 25, 4235–4247.
DOI: 10.1021/ef2005173.
ꢀ
Kapur, S. (2005). ABB Lummus Global SRT cracking tech-
nology for the production of ethylene. In R. A. Meyers (Ed.),
Handbook of petrochemicals production processes (Chapter
6.1). New York, NY, USA: Mc Graw-Hill.
Karaba, A., Zamostny, P., Lederer, J., & Belohlav, Z. (2013).
Generalized model of hydrocarbons pyrolysis using auto-
mated reactions network generation. Industrial & Engineer-
ing Chemistry Research, 52, 15407–15416. DOI: 10.1021/
ie4006657.
Katta, V. R., Aggarwal, S. K., & Roquemore, W. M. (2012).
Evaluation of chemical-kinetics models for n-heptane com-
bustion using a multidimensional CFD code. Fuel, 93, 339–
350. DOI: 10.1016/j.fuel.2011.10.035.
Kopinke, F. D., Zimmermann, G., & Ondruschka, B. (1987).
Tendencies of aromatization in steam cracking of hydro-
carbons. Industrial & Engineering Chemistry Research, 26,
2393–2397. DOI: 10.1021/ie00071a037.
Kossiakoff, A., & Rice, F. O. (1943). Thermal decomposition of
hydrocarbons, resonance stabilization and isomerization of
free radicals. Journal of the American Chemical Society, 65,
590–595. DOI: 10.1021/ja01244a028.
Bounaceur, R., Warth, V., Marquaire, P. M., Scacchi, G., Dom-
iné, F., Dessort, D., & Brevart, O. (2002). Modeling of hy-
drocarbons pyrolysis at low temperature. Automatic gen-
eration of free radicals mechanisms. Journal of Analytical
and Applied Pyrolysis, 64, 103–122. DOI: 10.1016/s0165-
2370(01)00173-5.
Chakraborty, J. P., & Kunzru, D. (2009). High pressure pyroly-
sis of n-heptane. Journal of Analytical and Applied Pyrolysis,
86, 44–52. DOI: 10.1016/j.jaap.2009.04.001.
Chakraborty, J. P., & Kunzru, D. (2012). High-pressure pyrol-
ysis of n-heptane: Effect of initiators. Journal of Analytical
and Applied Pyrolysis, 95, 48–55. DOI: 10.1016/j.jaap.2012.
01.004.
Dente, M. E., & Ranzi, E. M. (1983). Mathematical modeling
of hydrocarbon pyrolysis reactions. In L. F. Albright, B. L.
Crynes, & W. H. Corcoran (Eds.), Pyrolysis: Theory and
industrial practice (pp. 133–175). New York, NY, USA: Aca-
demic Press.
Murata, M., Saito, S., Amano, A., & Maeda, S. (1973). Predic-
tion of initial product distributions from pyrolysis of normal
paraffinic hydrocarbons. Journal of Chemical Engineering of
Japan, 6, 252–258. DOI: 10.1252/jcej.6.252.
Pant, K. K., & Kunzru, D. (1996). Pyrolysis of n-heptane: ki-
netics and modeling. Journal of Analytical and Applied Py-
rolysis, 36, 103–120. DOI: 10.1016/0165-2370(95)00925-6.
Reid, R. C., Prausnitz, J. M., & Poling, B. E. (1988). The prop-
erties of gases and liquids (4th ed.). New York, NY, USA:
McGraw-Hill.
Reyniers, M. F. S. G.,
& Froment, G. F. (1995). Influ-
ence of metal surface and sulfur addition on coke de-
position in the thermal cracking of hydrocarbons. Indus-
trial & Engineering Chemistry Research, 34, 773–785. DOI:
10.1021/ie00042a009.
Rice, F. O., & Herzfeld, K. F. (1934). The thermal decom-
position of organic compounds from the standpoint of free
radicals. VI. The mechanism of some chain reactions. Jour-
nal of the American Chemical Society, 56, 284–289. DOI:
10.1021/ja01317a006.
Ding, J. X., Zhang, L., & Han, K. L. (2011). Thermal rate con-
stants of the pyrolysis of n-heptane. Combustion and Flame,
158, 2314–2324. DOI: 10.1016/j.combustflame.2011.04.015.
Ding, J. X., Zhang, L., Zhang, Y., & Han, K. (2013). A reactive
molecular dynamics study of n-heptane pyrolysis at high tem-
Savage, P. E. (2000). Mechanisms and kinetics models for hy-
drocarbon pyrolysis. Journal of Analytical and Applied Py-
rolysis, 54, 109–126. DOI: 10.1016/s0165-2370(99)00084-4.

xii
N. Olahová et al./Chemical Papers
Sundaram, K. M., & Froment, G. F. (1978). Modeling of ther-
mal cracking kinetics. 3. Radical mechanisms for the pyroly-
sis of simple paraffins, olefins, and their mixtures. Industrial
& Engineering Chemistry Fundamentals, 17, 174–182. DOI:
10.1021/i160067a006.
Yuan, T., Zhang, L. D., Zhou, Z. Y., Xie, M. F., Ye, L. L.,
& Qi, F. (2011). Pyrolysis of n-heptane: Experimental and
theoretical study. The Journal of Physical Chemistry A, 115,
1593–1601. DOI: 10.1021/jp109640z.