Laser powered homogeneous pyrolysis of butane initiated by methyl radicals in a quasi-wall-free reactor at 750-1000 K

Article abstract of DOI:10.1039/b005219l

The reaction of pure hydrocarbon systems at high temperatures (thermal pyrolysis) is important in chemical industry, combustion in power stations, waste incineration with incomplete combustion, carbon black industry, and pyrolytic hydrocarbon production. The pyrolysis of n-butane, initiated by methyl radicals was studied at 750-1000 K and 0.08-0.13 bar in the a quasi-wall-free reactor using laser heating by fast vibrational-translational energy transfer. The radial temperature distribution in the reactor was examined via stationary heat balance, optical absorption, pressure rise, and the temperature dependence of the rate of an isomerization reaction. GC-MS was used to analyze the products. Methyl radicals were formed through the fast thermal dissociation of di-tert-butyl peroxide. The reaction produced mainly C₂H₄, C₃H₆, and C₃H₈, while 1-C₄H₈, n-C₅H₁₂, and iso-C₅H₁₂ were minor components, all exhibiting a strong dependence on temperature. A kinetic model of 61 species and 164 reactions developed for the high-temperature butane and the low-temperature n-pentane oxidation was utilized to examine the product distribution and the temperature dependence. Experimental results agreed well with modeling results.

Full text of DOI:10.1039/b005219l

View Article Online / Journal Homepage / Table of Contents for this issue
Laser powered homogeneous pyrolysis of butane initiated by methyl
radicals in a quasi-wall-free reactor at 750–1000 K
Elke Goos,a Horst Hippler,*a Karlheinz Hoyermannb and Bettina Jurgesb
a Institut fur Physikalische Chemie, Universitat Karlsruhe, Fritz-Haber-W eg 4,
D-76128 Karlsruhe, Germany. E-mail: horst.hippler=chemie.uni-karlsruhe.de
b Institut fur Physikalische Chemie, Universitat Gottingen, T ammannstr. 6, D-37077 Gottingen,
Germany
Received 29th June 2000, Accepted 14th September 2000
First published as an Advance Article on the web 18th October 2000
The pyrolysis of n-butane, initiated by methyl radicals has been studied in the temperature range 750È1000 K
and at pressures 0.08È0.13 bar in a quasi-wall-free reactor using laser heating by fast vibrationalÈtranslational
(VÈT) energy transfer. This is a convenient method to study homogeneous high-temperature kinetics since the
reactor walls remain cold. The radial temperature distribution in the reactor has been investigated by four
di†erent methods: stationary heat balance, optical absorption, pressure rise, and the temperature dependence
of the rate of an isomerization reaction. Methyl radicals were produced via the fast thermal dissociation of
di-tert-butyl peroxide and product analysis was performed by the use of GC-MS. The main products of the
overall reaction of the model system (n-C H ] CH ) were C H , C H , C H , whereas 1-C H , n-C H
10 12
iso-C H were minor components, all showing a strong dependence on temperature. The product distribution
,
4
3
2
4
3
6
3
8
4
8
5
5
12
and the temperature dependence were analyzed by a kinetic model of 61 species and 164 reactions developed
for the high-temperature butane and the low-temperature n-pentane oxidation. Good agreement was found
between our experimental investigations and the modeling. However, we had to slightly adjust the rate
constants for the reactions
CH ] n-C H ] n-C H ] CH
(3)
(4)
3
4
10
4
9
4
CH ] n-C H ] sec-C H ] CH
3
4
10 4
4
9
At a temperature of 1000 K we found a larger branching ratio of k /(k ] k ) \ 1/3 compared to 1/8 as
3
3
4
extrapolated from low-temperature data. The total rate coefficient was found to be (k ] k ) \ 8 ] 109 cm3
mol~1 s~1 which is about 50% higher than the extrapolated values.
3
4
processes of pyrolysis and combustion with the inherent
problem of the interplay of chemistry, transport processes,
and Ñow, thus complicating or hindering the extraction of an
unambiguous mechanism or the reaction rate of a speciÐc
species.3h5 Static or fast Ñow reactors are often used due to
the advantages of well-controlled temperature, pressure and
the experimental combination with all sorts of analytical
methods. As in most cases heating is accomplished by heat
transfer from the wall, heterogeneous catalytic e†ects for the
homogeneous mixture have to be considered. Heating of the
reaction mixture through the wall was avoided by the applica-
tion of laser heating via a fast VÈT energy transfer, as experi-
mentally realized in laser-powered homogeneous pyrolysis
(LPHP) by Shaub and Bauer6,7 (used for dissociation/
isomerization reactions) and in laser-heated reactors by
Heymann et al.8h10 (used for the isomerization of cylcohep-
tatriene to toluene and for the study of collisional energy
transfer and its temperature dependence).
Introduction
The reaction of pure hydrocarbon systems at high tem-
peratures, often termed thermal pyrolysis, plays an important
role in chemical industry, combustion in power stations, waste
incineration with incomplete combustion, the carbon black
industry, and pyrolytic hydrocarbon production and thereby
triggers theoretical and practical studies in chemical kinetics
for its description in terms of elementary reactions. Cracking
processes of high molecular weight hydrocarbons are used in
the production line of chemicals for the polymer industry (i.e.
formation of unsaturated small hydrocarbons) as well as the
formation of unsaturated hydrocarbons from natural gas (i.e.
CH , C H ). The formation of soot, polycyclic aromatic
4
2 6
hydrocarbons and other toxic compounds in the atmospheric
environment is strongly linked to the pyrolysis mechanisms.1
In order to understand the underlying chemistry di†erent
experimental methods, increasingly coupled with modeling
studies, have been applied. In order to put the method,
described in this paper, into perspective with other methods
these will be brieÑy summarized. High-temperature shock
tube studies o†er nearly wall-free conditions at temperatures
of T [ 1500 K and reaction times \1 ms with limited species
detection and analysis, whereas the single-pulse shock tube/
chemical shock tube allow in favorable cases a nearly com-
plete product analysis for reactions with high activation
energies.2 Fuel-rich premixed or di†usion Ñames are wall-free
and cover the whole temperature range 300È2500 K of the
The objective of this experimental approach was to apply
laser heating (to realize a quasi wall-free reactor in order to
reduce catalytic e†ects), to induce
a chemical radical/
hydrocarbon chemistry (in order to initiate reactions with low
activation energy in the temperature range 750È1000 K), and
to install a powerful analysis system (qualitative and quanti-
tative detection of all hydrocarbons, especially of isomers).
Initiation by methyl radical was chosen for the following
chemical reasons: The interpretation of the alkane pyrolysis
in terms of chemical kinetics is highlighted by the RiceÈ
Herzfeld mechanism with its main features being alkane disso-
DOI: 10.1039/b005219l
Phys. Chem. Chem. Phys., 2000, 2, 5127È5132
This journal is ( The Owner Societies 2000
5127

View Article Online
ciation by CÈC bond rupture leading to alkyl radicals
(initiation process); alkyl radical reactions with the alkanes
(chain propagation); alkyl radical decomposition by b-bond-
scission or H atom loss forming small, non-decomposable and
(Coherent 201). The intensity proÐle was measured by applica-
tion of a variable aperture before and after the reactor and is
illustrated in Fig. 2. This proÐle can be represented by a
Gaussian and we deÐne a beam diameter of 4 mm given by
twice the distance from the beam center to the point, where
the intensity has fallen to 1/e of its maximum value. No dis-
tortion of the laser beam proÐle is observed by the reaction
cell, as the proÐles before and behind the reactor are identical.
(This is a necessary condition for the method (2), see below.)
The absorbed energy by the gas mixture (corrected for the
losses by the windows and the air) was varied by the partial
less reactive species like CH ; and alkyl radical combination
3
and disproportionation (chain termination). The similarity of
the pyrolysis of the whole class of alkanes is strongly linked to
the formation and destruction of small alkyl radicals, espe-
cially of the CH radical. Therefore, it seems challenging to
3
start with the essential radical CH and follow the fate of a
3
speciÐc hydrocarbon.
The objectives are as follows: (i) A wall-free, medium to
high temperature reactor is realized by laser heating, using
pressure of the SF and the incident power of the laser.
Typical operation conditions are given in the Ðgures.
6
fast VÈT energy transfer after IR-laser absorption. (ii) CH
radicals in the presence of an excess of a hydrocarbon/inert
3
GC-MS analysis
gas mixture are produced via thermal decomposition of di-
tert-butyl peroxide. (iii) An overall pyrolysis process is broken
down into sub-reaction units by applying di†erent initial con-
Samples were withdrawn with a syringe (100 ll) and analyzed
by GC-MS. Two di†erent capillary columns were used for
GC, Poraplot Q (Chrompack) at an operating temperature of
140 ¡C, and a combination of two capillary columns (FS-OV-
1-CB, 25 m ] 0.25 mm, id df 0.5 lm, and FS-SE-54-CB, 25
m ] 0.25 mm, id df 0.25 lm) at 25 ¡C. The quadrupole mass
spectrometer was operated at an ionization energy of 70 eV
leading to characteristic mass spectra. The products are iden-
tiÐed by the retention time of the GC and their mass spec-
trum. The concentrations could be determined by calibration
centration ratios of the CH radicals and a speciÐc hydrocar-
3
bon. (iv) A detailed Ðnal product analysis is accomplished by
GC-MS. (The presented results on the CH ] alkane reac-
3
tions are part of a larger project including the hydrocarbons
1-C H , 2-C H , iso-C H , C H , CH C H, CH CCH ,
4
8
4
8
4
8
2
2
3 2
2
2
C H /C H , where the CH ] alkane system is one small
2
2
2
4
3
building block in an overall reaction scheme.) (v) The mea-
sured product distributions are analyzed with respect to a
possible reaction mechanism and are modeled using a reaction
scheme with the rate coefficients from the literature. This is
used to validate/falsify the rate coefficients reported so far for
the oxidation of butane.
vs. appropriate sample mixtures. The non-reactive CF was
4
admixed in all measurements as an internal concentration ref-
erence registered at m/z \ 69. The area of the chromato-
graphic peak was obtained by integration and taken as the
concentration measure. In cases of mass interferences, integra-
tion at characteristic fragment ion peaks was applied.
Experimental
The experimental arrangement is sketched in Fig. 1, consisting
of the reactor, the laser, the analysis system and the handling
system for the chemicals.
Chemicals
Commercial grade chemicals were used without further puriÐ-
cation. Liquids were carefully degassed before use. Gaseous
reaction mixtures were prepared before a series of experiments
and stored in 2 l vessels in the dark. ((CH ) COOC(CH ) ,
DTBP: 95%; SF : 99.9%; c-C H : 99.0%; C H : 99.0%, all
Merck; CF : 99.8%; n-C H : 99.95%; iso-C H : 99.0%,
10 10
all Linde; Ar: 99.9998%, Messer-Griesheim).
Reactor
3 3
3 6
4
3 3
6
3
6
The main reactor was a stainless steel cylindrical cell with id
20 mm and length 7.8 cm. The length was varied from 1.8 to
7.8 cm and also the dead space by the insertion of the appro-
priate cylinders of stainless steel or glass. The KCl windows (5
mm thick) were Ðxed at the Brewster angle. Several air-tight
connections were installed for the supply of chemicals, pumps,
pressure transducer (measuring the pressure and the pressure
rise with time), and a port for a septum for taking samples
with a syringe, e.g. for the GC-MS analysis.
4
4
The composition of the reaction mixtures is given in Table
1 together with the experimental conditions.
Principle of operation
When a laser pulse of 50 ms duration is applied the tem-
perature of the given mixture (SF , Ar, hydrocarbon etc.) rises
due to VÈT transfer from SF to argon and to all other
species, which are present in the mixture and a stationary tem-
6
6
Laser
The CO -laser (Apollo 560) could be operated continuously
2
(CW) or in pulse mode with pulse duration of 10~4 to 1 s at a
repetition rate of 1È100 Hz. In most cases the laser was stabil-
ized at the P20 line (10.59 lm), controlled by a spectrum
analyzer (ORIEL). The laser output (up to 40 W) was mea-
sured continuously during the experiment by a power meter
Fig. 2 The radial intensity proÐle of the CO laser (*I/*F \ radiant
2
power per unit area; (=) before entering the reaction cell; (…) after
passing the reaction cell; full line: Gaussian Ðt).
Fig. 1 Experimental set-up.
5128
Phys. Chem. Chem. Phys., 2000, 2, 5127È5132

View Article Online
Table 1 Composition of the reaction mixtures, experimental conditions and relative product yields for the reaction CH ] n-C H : (p(Ar)
3
4 10
added up to p ; pulse duration: 50 ms; accumulation of 1000 to 10 000 pulses). The experiment with a total pressure of 80 mbar and 4.53 mbar
total
SF was repeated to show the reproducibility. The alternative relative C H and C H yields are due to the uncertainties in the MS sensitivities
6
2
4
3
6
relative to CF
4
T /K
765
795
945
p
/mbar
80
0.8
0.5
133
1.33
0.8
80
4.53
0.5
80
4.53
0.5
133
4.53
0.8
total
SF
6
CF
4
Educts p/mbar
DTBP
n-C H
Relative product yield (%)
C H
2.72
24.5
4.53
40.7
2.72
24.5
2.72
24.5
4.53
40.7
4
10
30.7
or 35.9
50.7
or 45.5
1.4
1.9
8.7
28.1
or 32.9
48.1
or 43.3
2.5
0.9
12.8
7.6
47.3
or 53.6
49.0
or 42.8
1.2
0.1
2.2
49.7
or 56.0
46.7
or 40.5
1.1
0.1
2.1
48.3
or 54.7
48.5
or 42.3
1.0
0.1
2.0
2
4
C H
3
6
1-C H
n-C H
C H
iso-C H
4
8
5
12
3
8
6.6
0.2
0.3
0.1
5
12
perature proÐle is reached within 1È3 ms with a maximum
temperature at the optical axis of the cell. During the 50 ms
using the temperature-dependent optical absorption coeffi-
cient of SF (similar to ref. 8); (3) by the measurement of the
6
duration of the CO laser pulse the temperature proÐle in the
pressure
rise
in
the
total
volume
(V (reaction
2
cell remains constant and decays after the laser pulse within
cylinder) ] V (dead space)) for a given absorbed energy;10 and
(4) by the measurement of the conversion of cyclopropane (c-
C H ) to propene (C H ) via the well-established
about 1È3 ms. Cycles of heating and “coolingÏ were obtained
due to the repeated application of laser pulses, hence a com-
plete mixing of the heated gas in the middle of the cell with
the cold gas near the wall is achieved through gas convection.
This leads to a complete homogenization of the gas composi-
tion between the laser pulses. The time for ordinary di†usion
from the center to the wall is of the order of 30 ms, thus rad-
icals produced in the center are removed mainly by homoge-
neous reactions. Convective Ñow into the non-isothermal
region leads to mixing with the gas of the dead space, but no
deposit was found at the reactor wall.
3
6
3 6
temperature-dependent rate of this isomerization.6
(1) The VÈT energy transfer from the SF , excited by the IR
absorption at 10.59 lm, to the excess of Ar at pressure P is
6
fast with a relaxation time q, as given by the relation log (P É q
atm~1 s~1) \ [ 8.5 ] 31.1 ] (T /K)~1@3.6 Therefore, station-
ary heat balance between the absorbed laser Ñuence and heat
conduction through a cylinder surface (2prl) can be assumed.
We solved the di†erential equation numerically taking into
account the laser intensity proÐle (see Fig. 2), the temperature-
dependent thermal conductivity11,12 of the mixture (Ar,
hydrocarbon, SF , DTBP, CF ), and the condition of a con-
The maximum temperature can be measured by using a
chemical thermometer like the isomerization reaction cyclo-
propane (c-C H ) ] propene (C H ) or thermal decomposi-
6
4
stant wall temperature of 300 K. We found stationary tem-
perature proÐles to be established within the Ðrst ms of the
3
6
3 6
tion of di-tert-butyl peroxide (DTBP). Since the rate constants
show considerable activation energies, measurable conversion
occurs only at the highest temperature in the center of the
reaction cell. We found only a low conversion of c-C H , due
rectangular CO laser pulse of 50 ms duration. Temperature
2
proÐles are shown in Fig. 3 (curve (a), (b)) for typical experi-
mental conditions.
3
6
to the low rate coefficient and long life time, respectively
(q(700 K) D 105 s; q(1000 K) \ 2.5 s).
(2) The optical absorption coefficient of SF is strongly tem-
6
perature dependent.13 By measuring the absorbed energy at
The temperature in each CH ] n-C H experiment was
di†erent radial positions in the volume, through variation of
the exit aperture diameter and applying BeerÏs law, the gas
density and thereby the temperature was deduced. This pro-
cedure is equivalent to that of ref. 8 and 9 where the density of
3
4 10
determined experimentally in the following way. The same gas
mixture in which small amounts of argon have been replaced
by c-C H was heated with identical laser power. The
3
6
resulting temperature for both experiments was deduced from
the conversion of c-C H into C H .
CS or cycloheptatriene was determined by an in situ optical
2
UV absorption method. Typical temperature proÐles from this
3
6
3 6
The thermal decomposition of DTBP, the CH radical
method are also given in Fig. 3 (curve (c), (d)) however, for
lower absorbed power than in curve (a) and (b). Good agree-
ment between the methods (1) and (2) has been found.
3
source in this study, proceeds on a ms timescale and the CH
3
radicals disappear predominately in a recombination reaction
producing ethane. The rate of the hydrocarbon reaction under
study can be adjusted by the partial pressure of the hydrocar-
bon. Here the reaction CH ] C H is shown as an example.
(3) The pressure rise is directly related to the absorbed laser
power and thus, for a given geometry, is a measure of tem-
perature proÐle in the cell. Under our condition we kept the
pressure rise in the total volume low (less than 5%) by choos-
ing the volume of the reaction cylinder to be much smaller
than the volume of the dead space, as described in ref. 10. As a
result, all experiments were performed under essentially iso-
thermal conditions. We calibrated the measured pressure
increase by methods (1) and (2) and in this way had a simple
way of checking the reproducibility of the repetitive laser
heating procedure.
3
4 10
Experiments without the methyl radical precursor DTBP did
not give any products since at temperatures below 1000 K the
decomposition of butane is too slow.
Results
Temperature proÐles in the reactor
The radial temperature proÐles in the reactor were deduced
by four di†erent methods, which have been described in prin-
ciple in the literature: (1) By measurement of the total energy
absorbed in the volume and analysis of the heat transfer
balance absorbed power/heat conduction;8,9 (2) by the mea-
surement of the radial distribution of the absorbed energy
(4) The fundamentals of using a chemical thermometer have
been discussed extensively in ref. 6 and therefore only our
experimental results are presented. The experimentally deter-
mined conversion of c-C H to C H is compared to calcu-
3
6
3 6
lated conversions by adopting the literature value for the
Phys. Chem. Chem. Phys., 2000, 2, 5127È5132 5129

View Article Online
Only at higher temperature (T \ 1750 K) were the products
CH CHO, tert-C H OH, CH COC H and CH CO detected
3
4
9
3
2
5
2
besides the main products CH COCH and C H . Therefore
3
3
2 6
the thermal peroxide dissociation is a suitable source for CH
radicals at temperatures around and below 1000 K.
3
The products of the reaction CH ] n-C H at three tem-
3
4 10
peratures are listed in Table 1. Data for two experiments
under identical conditions (p \ 80 mbar, T \ 945 K) are
total
given to show the reproducibility. No pressure dependence
was observed in the small pressure range 80 to 133 mbar. The
alternative values in Table 1 for the two main products C H
2
4
and C H were obtained using di†erent mass spectroscopic
3
6
sensitivities reÑecting the uncertainties in the calibration
factors14 and show the uncertainties in the yield determi-
nation. We conclude that all relative product yields have a
relative uncertainty of ^15%.
Besides the products from Table 1 cis-2-C H was detected
in low concentration which we attributed to the isomerization
4
8
of 1-C H to cis-2-C H as has been proven in separate
4
8
4 8
experiments.
Methane concentration could not be determined because of
superposition of the N /O peak with the CH peak.
Fig. 3 Radial temperature proÐles (T) in the reactor as deduced by
di†erent methods. (a) and (b) via stationary heat balance, method (1),
length of reactor l \ 7.8 cm, diameter: 2.0 cm, diameter of the laser
2
2
4
The formation of the main products C H , C H and C H
3
6
2
4
3 8
beam (Gauss proÐle): 4 mm, p (SF ) \ 4 mbar; (a) absorbed power:
is explained by the mechanism and the kinetic modeling given
below. Since the methyl precursor decomposes completely
within the Ðrst ms of the laser heating pulse, stable product
formation is independent of the applied reaction times of 50
ms (see below).
6
15.9 W; (b) absorbed power: 10.4 W. (c) and (d) via temperature
dependence of the absorption coefficient of SF ; l \ 3.9 cm, (c)
6
p(SF ) \ 4 mbar, absorbed power: 8.1 W; (d) p(SF ) \ 1.3 mbar,
6
6
absorbed power: 5.5 W. (e) concentration proÐle of c-C H , using
3
6
temperature proÐle (c), (f) concentration proÐle of C H formed in the
3
6
isomerization reaction c-C H ] C H .
3
6
3 6
Reaction mechanism and modeling
We analyzed the data using a kinetic model with 61 species
and 164 reactions. The mechanism consisted of a combination
of the high-temperature butane oxidation and the low-
temperature n-pentane oxidation mechanism from Warnatz
and co-workers,15 where we include the di-tert-butylperoxide
decomposition reactions (1) and (2). The rate constants for
reactions (1) and (2) were taken from ref. 16È18, respectively,
but must be extrapolated from low temperature (near 400 K)
to our conditions between 750 and 1000 K.
isomerization rate constant and the radial temperature proÐle
from the measurements using methods (1), (2), and (3). We
identify these conversions with a reaction rate averaged mean
temperature, in the following just called temperature. This
method is best suited for conditions where accurate concen-
tration measurements of c-C H can be performed. Conver-
3
6
sions vary with the number of applied laser pulses and for the
temperature of 945 K the following measured and calculated
values of *[c-C H ]/[c-C H ] of 76% (calculated: 80%),
The central part of this CH /n-C H mechanism is pre-
3
6
3 6 0
50% (55%), and 47% (48%) were found for 2000, 1000, and
800 laser pulses, respectively. The good agreement between
measured and calculated c-C H conversions supports our
3
4 10
sented in Fig. 4, where dominant reaction paths for the forma-
tion of the essential products are shown. This reaction scheme
was obtained from a sensitivity analysis of the yields of the
main products. Obviously, C H is the key product reÑecting
3
6
methods (1)È(3) for the determination of the temperature pro-
Ðles. In Fig. 3 (curve (e), (f)) we include the calculated c-C H
3 6
the intermediate sec-C H radical concentration produced
3
6
and C H concentrations as a function of the radial position
4 9
from CH abstracting a secondary H-atom from n-C H
10
The two other products C H and C H are from n-C H
.
3
6
in the reactor. This nicely illustrates that in our laser heated
reactor the reactions with an appreciable activation energy
automatically switch o† near to the cold wall. It also shows
that chemical reaction proceeds only at the highest tem-
perature and that our reaction rate averaged temperature is
close to the maximum of the temperature proÐle. However, we
claim the uncertainty of the rate averaged temperature to be
relatively large, here T \ (945 ^ 40) K, since it contains not
only the uncertainties of the simpliÐed temperature proÐle
model but also the uncertainties of the concentration mea-
surements and of the isomerization rate constant.
3
4
2
4
3
8
4 9
radicals decomposition, which are formed by methyl attacking
primary H-atoms in n-C H
.
4
10
The model calculations were performed with the
CHEMKIN Code19 assuming isothermal conditions. We
neglected the heating period since the temperature rise was
always very fast compared to the reaction time. We also disre-
Product formation by the reaction CH + n-C H
4 10
3
First the CH radical formation from di-tert-butyl peroxide
3
was investigated. At low temperature (T \ 900 K) the pro-
ducts and relative concentrations were: CH COCH (1000),
3
3
CH CHO (7), tert-C H OH (1) and C H (500). The main
3
4
9
2 6
products are explained by the following reactions.
(tert-C H O) ] 2 tert-C H O
(1)
(2)
4
9
2
4 9
tert-C H O ] CH COCH ] CH
4
9
3
3
3
CH ] CH ] C H
Fig. 4 Central part of the reaction mechanism.
3
3
2 6
5130
Phys. Chem. Chem. Phys., 2000, 2, 5127È5132

View Article Online
garded the small pressure increase because of the large dead
space of the system. The initial concentrations of the reactants
di-tert-butyl peroxide and n-C H as well as the Ar pressure
4
10
from the experiment were used as the input conditions for the
model calculation.
The temporal evolution of the reactants (n-C H and CH )
4
10
3
and of the products (n-C H , 1-C H , C H , C H and
12
C H ) is depicted in Fig. 5. Clearly a stationary state has been
5
4
8
3
8
2 4
3
6
reached for n-C H , C H and CH after a few ms while the
10
concentration of C H , C H and 1-C H show a further
4
3
8
3
3
2
4
8
4 8
slight increase during the reaction time of 50 ms. However, the
ratio of these three products remains unchanged within one
laser pulse and is independent of the irradiation time. This
result supports our experimental procedure in which we accu-
mulated the products by periodically repeating laser heating
followed by convectional cooling.
The resulting relative product ratios have been compared
with experiment with the exception of the minor product i-
C H , since this species was not included in the mechanism.
5
12
In Fig. 6 we present this comparison for the three main pro-
ducts C H , C H and C H . For clarity, the minor product
2
4
3
6
3 8
1-C H is not shown although similar agreement between
Fig. 6 Relative product yields as function of temperature (full
4
8
experiment and model calculation has been achieved. The
slight discrepancies at 765 K and 795 K between modelled
and measured product ratios of C H and C H have no
inÑuence on the branching ratio of the reaction of methyl rad-
icals with n-butane.
symbols: model calculation; open symbols: experiment from Table 1).
2
4
3 8
50% higher, which is in reasonable agreement with the predic-
tions from the literature within the experimental uncertainties
(temperature (^4%), relative product yields ( ^ 15%), and
reaction rate constants in the model).
The sensitivity analysis suggests that only the product ratio
C H /(C H ] C H ), for which we found good agreement
3
6
2
4
3 8
between model and experiment, is sensitive to the ratio of the
rate constants k /k .
3
4
Conclusion
CH ] n-C H ] n-C H ] CH
(3)
(4)
3
4
10 4
4
9
The application of laser induced heat transfer by fast V ÈT
energy transfer from SF to Ar combined with the hydrocar-
bon conversion by methyl radical initiation seems to be well
CH ] n-C H ] sec-C H ] CH
3
4
10
4
9
4
6
Both rate constants have been discussed in the literature.
suited for the study of hydrocarbon reactions in the medium
temperature range 750È1000 K. The reactive system is well
characterized with respect to the temperature determined by
di†erent procedures and the chemical species production as
measured by GC-MS analysis. This has been proven for the
model system (n-C H ] CH ) leading to an agreement
between experiment and modeling. It is interesting to note,
that the assumed main reaction scheme stems from and has
been validated by the oxidation studies of butane, thus our
Ðndings support those mechanisms. Nevertheless small devi-
ations (ratio of product formation in reactions (3) and (4),
total rate coefficient) are to be noted. In summary the com-
bination of homogeneous laser heating/radical initiation/
product analysis seems to be a powerful addition to the
methods previously available.
From experiments by Sway20 near 400 K k \ 5.5 ] 1011
4
exp([39.96 kJ mol~1/RT ) cm3 mol~1 s~1 has been proposed,
and thermochemical estimation21 leads to k \ 5.0 ] 1011
3
exp([56.90 kJ mol~1/RT ) cm3 mol~1 s~1. Through extrapo-
lation from low temperature (around 400 K) to 1000 K a
branching ratio of k /(k ] k ) ^ 1/8 is extrapolated.
4
10
3
3
3
4
However, from our model calculation we conclude this ratio
to be 1/3 and the total rate constant (k ] k ) to be about
3
4
Acknowledgements
E. G. and H. H. gratefully acknowledge Ðnancial support by
the Deutsche Forschungsgemeinschaft (SFB 551 ““Kohlensto†
aus der Gasphase: Elementarreaktionen, Strukturen,
Werksto†eÏÏ). K. H. and B. J. thank the Deutsche Forschungs-
gemeinschaft for Ðnancial support within the Sonderfors-
chungsbereich 357 (““Molekulare Mechanismen unimolekul-
arer ProzesseÏÏ). Both groups also acknowledge the Ðnancial
support given by the Fonds der Chemischen Industrie.
Appendix
~CH
CH
methyl radical
methane
tetraÑuoromethane
carbon disulÐde
ethyne
3
4
CF
CS
Fig. 5 Calculated temporal evolution of the mole fractions of the
reactants and main products (T \ 950 K; p((tert-C H O) ) \ 4.53
4
2
4
9
2
mbar; p(n-C H ) \ 40.8 mbar; p(Ar) \ 87.8 mbar).
C H
4
10
2
2
Phys. Chem. Chem. Phys., 2000, 2, 5127È5132
5131

View Article Online
4
5
N. M. Marinov, M. J. Castaldi, C. F. Melius and W. Tsang,
Combust. Sci. T echnol., 1997, 128, 295.
N. M. Marinov, W. J. Pitz, C. K. Westbrook, A. M. Vincitore, M.
J. Castaldi, S. M. Senkan and C. F. Melius, Combust. Flame,
1998, 114, 192.
C H
2
ethene
ethane
prop-1-yne
allene, propa-1,2-diene
propene
cyclopropane
propane
but-1-ene
but-2-ene
isobutene
sec-butyl radical
n-butyl radical
n-butane
isobutane
ketene
propanone, acetone
acetaldehyde, ethanal
tert-butanol
ethyl methyl ketone
sulfur hexaÑuoride
4
6
C H
2
C H , CH C H
3
4
3 2
C H , CH CCH
3
C H
3
4
2
2
6
7
W. M. Shaub and S. H. Bauer, Int. J. Chem. Kinet., 1975, 7, 509.
W. P. Horn, M. S. Sheldon and P. C. T. De Boer, J. Phys. Chem.,
1986, 90, 2541.
6
c-C H
3
6
C H
3
8
8
9
M. Heymann, H. Hippler, D. Nahr, H. J. Plach and J. Troe, J.
Phys. Chem., 1988, 92, 5507.
1-C H , CH CHCH CH
4
8
8
2
3
2
3
2-C H , CH CHCHCH
iso-C H , (CH ) CCH
3 2
sec-~C H , CH ~CHCH CH
M. Heymann, H. Hippler and J. Troe, J. Chem. Phys., 1984, 80,
4
3
1853.
4
8
2
10 H. Hippler, B. Otto and J. Troe, Ber. Bunsen-Ges. Phys. Chem.,
1989, 93, 428.
11 R. C. Reid, J. M. Prausnitz and B. E. Poling, T he Properties of
Gases and L iquids, McGraw-Hill, New York, 4th edn., 1987.
12 Encyclopedie des Gaz, DÏAir Liquide, Amsterdam, 1976.
13 A. V. Nowak and J. L. Lyman, J. Quant. Spectrosc. Radiat.
T ransfer, 1975, 15, 945.
14 A. Cornu and R. Massot, Compilation of Mass Spectral Data,
Heyden, London, 2nd edn., 1975.
15 C. Chevalier, M. Nehse and J. Warnatz, T wenty-Sixth Symposium
(International) on Combustion, The Combustion Institute, Pitts-
burgh, PA, 1996, pp. 773È780.
16 M. J. Perona and D. M. Golden, Int. J. Chem. Kinet., 1973, 5, 55.
17 M. I. Sway, J. Chem. Soc., Faraday T rans., 1991, 81, 2157.
18 C. Fittschen, H. Hippler and B. Viskolcz, Phys. Chem. Chem.
Phys., 2000, 2, 1677.
4
9
3
2
3
n-~C H , ~CH CH CH CH
4
n-C H
4
9
2
2
2
3
10
iso-C H , (CH ) CHCH
4
CH CO
2
10
3 2
3
CH COCH
3
3
CH CHO
3
tert-C H OH
4
9
CH COC H
3
2 5
SF
6
(CH ) COOC(CH ) , (tert-C H O) DTBP, di-tert-butyl
3 3
KCl
3 3
4
9
2
peroxide
potassium chloride
19 R. J. Kee, F. M. Rupley, J. A. Miller, M. E. Coltrin, J. F. Grcar,
E. Meeks, H. K. Mo†at, A. E. Lutz, G. Dixon-Lewis, M. D.
Smooke, J. Warnatz, G. H. Evans, R. S. Larson, R. E. Mitchell, L.
R. Petzold, W. C. Reynolds, M. Caracotsios, W. E. Stewart and
P. Glarborg, Chemkin Collection III, Reaction Design, Inc., San
Diego, CA, 1999.
References
1
2
3
B. S. Haynes and H. Gg. Wagner, Prog. Energy Combust. Sci.,
1981, 7, 229.
W. Tsang, in Shock W aves in Chemistry, ed. A. Lifshitz, Marcel
Dekker, New York, 1981, 59.
20 M. Sway, Indian J. Chem. Sect. A, 1990, 29, 748.
21 Yu. P. YampolÏskii, React. Kinet. Catal. L ett., 1975, 2, 449.
J. Warnatz, H. Bockhorn, A. Moser and H. W. Wenz, Nineteenth
Symposium (International) on Combustion, The Combustion Insti-
tute, Pittsburgh, PA, 1982, p. 197.
5132
Phys. Chem. Chem. Phys., 2000, 2, 5127È5132