Experimental study of a low pressure stoichiometric premixed methane,

methane/ethane, methane/ethane/propane and synthetic natural gas flames

A. Turbieza, A. El Bakalia,*, J.F. Pauwelsa, A. Ridab, P. Meunierb

aPhysicochimie des Processus de Combustion et de l’Atmosphere, Centre d’Etudes et de Recherches sur les Lasers et Applications, UMR CNRS/USTL-PC2A,

Universite des Sciences et Technologies de Lille, 59655 Villeneuve d’Ascq Cedex, FrancebGaz de France, Direction de la Recherche, B.P. 33, 93211 Saint Denis La Plaine Cedex, France

Received 26 March 2003; revised 9 October 2003; accepted 28 October 2003; available online 19 November 2003

Abstract

This study reports the structure analysis of four laminar stoichiometric premixed CH4/O2/Ar, CH4/C2H6/O2/Ar, CH4/C2H6/C3H8/O2/Ar

and CH4/C2H6/C3H8/n-C4H10/i-C4H10/n-C5H12/i-C5H12/n-C6H14/O2/Ar flames. The flames have been stabilized on a flat flame burner at low

pressure (40 Torr). Experimental mole fraction profiles of stable and reactive species have been obtained by coupling molecular-beam/mass

spectrometry (MB/MS) with gas chromatography/mass spectrometry (GC/MS) analyses. Temperature profiles were measured with a coated

thermocouple in the sampling conditions. This work provides a detailed experimental data on the nature and concentrations of the stable and

reactive species produced by oxidation of various representative natural gas mixtures. In addition to reactants and major products, the mole

fractions of stable intermediate and reactive species have been measured. The experimental results show that ethane and propane, increase

significantly intermediate species mole fractions especially olefins. On the other hand, the higher alkanes play a minor role in the oxidation of

the natural gas. It appears therefore, that the combustion of methane/ethane/propane mixture represents appropriately natural gas

combustion. A qualitative explanation of the experimental observations, based on the current literature mechanisms of alkanes combustion, is

reported.

q 2003 Elsevier Ltd. All rights reserved.

Keywords: Combustion; Kinetics; Natural gas

1. Introduction

Nowadays, gas equipments have to satisfy different more

stringent criteria: heat release control, thermal efficiency,

cleanness and compactness. Therefore, design engineers

need design tools that take into account more and more

detailed physical phenomena in burners and furnaces. These

tools should fulfill two opposite requirements: to be

sufficiently accurate and simple enough to be used in

designing equipment.

Even if methane is the dominant component in natural

gas, ethane, propane, and heavier hydrocarbons present in

this fuel may have an impact on its combustion; furthermore

there is no typical composition of natural gas. For example,

the complex chemical mechanism of NOx formation in

natural gas combustion involves reactive species like

N, H, O, OH, CHi [1], whose concentration may be affected

by natural gas composition. Furthermore, it is well known in

the gas industry that gas composition plays a role in

determining the flame height and stability [2]. The influence

of heavier hydrocarbons, like ethane, propane, butane

or pentane, on the oxidation of methane has been studied

under a wide range of experimental conditions by mea-

suring ignition delay behind reflected shock waves [2–9],

laminar burning velocities [10] and concentrations profiles

of stable species in jet-stirred reactor [11]. All results

show that methane reactivity is significantly enhanced by

higher hydrocarbons at a low level of concentration:

heavy hydrocarbon radicals, more reactive than methyl

radical and formed by larger alkanes oxidation,

initiate reactions more quickly and thereby decrease

significantly the ignition delay times or increase the burning

velocities.

The main objective of this work is to provide detailed and

accurate data on stable and reactive species mole fraction in

0016-2361/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/j.fuel.2003.10.017

Fuel 83 (2004) 933–941

www.fuelfirst.com

* Corresponding author. Tel.:þ33-3-20-43-48-04; fax:þ33-3-20-43-67-

77.

E-mail address: abderrahman.el-bakali@univ-lille1.fr (A. El Bakali).

laminar stoichiometric premixed CH4/O2/Ar, CH4/C2H6/

O2/Ar, CH4/C2H6/C3H8/O2/Ar and CH4/C2H6/C3H8/n-

C4H10/i-C4H10/n-C5H12/i-C5H12/n-C6H14/O2/Ar flames at

low-pressure. To our knowledge, there is no previous

experimental investigation of the combustion of heavy

alkanes mixtures, representative of natural gas composition

in flame conditions. In a second step, these results will be

used to validate a detailed reaction mechanism for natural

gas combustion.

2. Experimental

We have investigated at low-pressure laminar pre-

mixed stoichiometric CH4/O2/Ar, CH4/C2H6/O2/Ar,

CH4/C2H6/ C3H8/O2/Ar and CH4/C2H6/C3H8/n-C4H10/i-

C4H10/n-C5H12/i-C5H12/n-C6H14/O2/Ar flames by progress-

ively substituting small quantity of methane by ethane,

ethane/propane and finally by natural gas. The initial

conditions of the studied flames are given in Table 1. The

composition of the synthetic natural gas used is given in

Table 2.

The mole fraction profiles of stable, atomic and radical

species are obtained by coupling molecular beam-mass

spectrometry (MB/MS) with gas chromatography (GC/MS,

GC/FID/TCD) analyses. A detailed description of the

experimental procedures is given in Ref. [12] and only the

main features are repeated here.

The flames are stabilized on a movable vertical water-

cooled plug flat flame burner (7 cm in diameter). Chemical

species are sampled by a deactivated 458 quartz cone with

an orifice diameter of 100 mm and form a molecular beam

passing through three differential pumped stages delimited

by the sampling cone, the skimmer and the collimator,

respectively. The chemical species present in the molecular

beam are ionized by electron impact and analyzed by a

quadrupole mass spectrometer. The beam is modulated by a

rotating toothed wheel. A phase sensitive detection is

then used to discriminate the background signal from

the molecular beam signal. Each species is monitored in

such conditions to avoid strong fragmentation effects

perturbing the flame composition profiles and/or eventually

to discriminate the contributions of species with the same

m=e contribution.

Nevertheless in the case of O–CH4, C2H6–CH2O,

HCO–C2H5, mass interferences cannot be avoided by

reducing the electron energy. Oxygen atoms have an

ionization potential (13.6 eV) higher than the methane one

(12.7 eV); in these conditions it is only possible to establish

the O profiles in a region where CH4 is totally absent. For

the same reason, OH profiles are corrected for the

contribution of 13CH4 at 17 a.m.u. following the method

described by Lazzara et al. [13]. No attempt has been made

to separate HCO and C2H5 signals; however, it is reasonable

to expect that in the methane flame, the 29 a.m.u. signal has

been assigned to HCO because in our experimental

conditions the importance of the CH3 recombination

pathway yielding to C2 oxidation route is then lowered. In

the other flames the 29 a.m.u. signal has been assigned to

C2H5. In the same way, the signal at 30 a.m.u. has been

assigned to CH2O in the methane flame and to C2H6 in the

other flames. The mass spectrometer calibration method

[14] is performed using (i) the usual cold gas procedure for

most stable species like CH4, C2H6, O2, H2, CO, CO2, C2H2,

C2H4, C3H6 and C3H8, (ii) the conservation of the total

number of atoms for H2O, (iii) the pseudo-equilibrium

method [14] in the burnt gases for H and OH radicals and

(iv) the Relative Ionization Cross Section (RICS) method

[14] for C2H5, CH3, O, HCO and CH2O species with C2H6,

CH4, O2, O2 and O2, respectively, as reference species.

The uncertainty on flame position was estimated to

be < 100 mm. The uncertainty of mole fraction measure-

ments depends on (i) the amount of species present in the

investigated flame, (ii) the working potential used for

species detection and (iii) the calibration method. The

accuracy for species measurements is estimated to be less

Table 1

Initial conditions of laminar stoichiometric premixed CH4/O2/Ar (C1), CH4/C2H6/O2/Ar (C1C2), CH4/C2H6/C3H8/O2/Ar (C1C2C3) and CH4/C2H6/C3H8/n-

C4H10/i-C4H10/n-C5H12/i-C5H12/n-C6H14/O2/Ar (NG) flames studied in this work

C1 Flame C1C2 flame C1C2C3 flame NG flame

CH4 9.56 £ 1022 8.19 £ 1022 7.86 £ 1022 7.75 £ 1022

C2H6 – 8.251 £ 1023 8.095 £ 1023 8.35 £ 1023

C3H8 – – 1.778 £ 1023 1.61 £ 1023

i-C4H10 – – – 6.06 £ 1025

n-C4H10 – – – 7.86 £ 1025

i-C5H12 – – – 4.37 £ 1025

n-C5H12 – – – 4.32 £ 1025

n-C6H14 – – – 4.35 £ 1025

O2 18.2 £ 1022 19.27 £ 1022 19.44 £ 1022 19.43 £ 1022

Ar 72.2 £ 1022 71.71 £ 1022 71.71 £ 1022 71.8 £ 1022

Global flow (l/h) 334.3 334.3 334.3 334.3

Equivalence ratio 1.05 0.99 1 0.99

Pressure (Torr) 40 40 40 40

A. Turbiez et al. / Fuel 83 (2004) 933–941934

than ^7% for major species, ^10% for stable intermediate

species and ^20% for reactive species. More details

concerning the accuracy of the method used in this work

can be found in Ref. [12].

As the MB/MS technique does not allow to separate

isomers and can be confronted to unresolved problems of

fragmentation, the MB/MS set-up is coupled with gas

chromatography technique whose sensibility and selectiv-

ity are very well suited to access minor molecular

species. The volume sampled at low-pressure by the

MB/MS quartz cone can be directed towards a chro-

matograph by using a compression procedure [15]. The

chromatograph is equipped with thermal conductivity and

flame ionization detectors placed in series or with a

quadrupole mass spectrometer. Poraplot Q and U, and

Molecular Sieve 5A semi-capillaries or capillaries

columns allow separation of species present in the

samples. The species are calibrated by sampling suitable

mixtures of known composition. The accuracy of the

method is estimated to be ^8% for major alkane species

and ^15% for the other species.

Temperature profiles are obtained using a coated Pt/Rh

6—Pt/Rh 30% thermocouple ðf ¼ 100 mm) located

200 mm upstream the quartz cone tip. Conduction heat

losses are avoided by setting the thermocouple in a plane

perpendicular to the laminar flow. Radiative heat losses are

corrected using the electric compensation method [16].

Errors in the peak temperatures were estimated to

be ^ 100 K.

3. Results and discussion

To point out the influence of the fuel composition, the

four studied flames were stabilized in the same conditions.

As shown in the Table 1, flames were stabilized at 40 Torr

with the same dilution (72% Ar) and the same exit gas

velocity at the burner surface. The initial composition

corresponds to stoichiometric conditions; the equivalence

ratio has been calculated by taking into account the

chemical equations corresponding to each fuel present in

the natural gas:

CH4 þ 2O2 ¼ CO2 þ 2H2O

C2H6 þ7

2O2 ¼ 2CO2 þ 3H2O

C3H8 þ 5O2 ¼ 3CO2 þ 4H2O

n-C4H10 þ13

2O2 ¼ 4CO2 þ 5H2O

i-C4H10 þ13

2O2 ¼ 4CO2 þ 5H2O

n-C5H12 þ 8O2 ¼ 5CO2 þ 6H2O

i-C5H12 þ 8O2 ¼ 5CO2 þ 6H2O

n-C6H14 þ19

2O2 ¼ 6CO2 þ 7H2O

For example, the expression of the equivalence ratio for

the natural gas flame is as follows:

f¼2XCH4

XO2

þ7

2

XC2H6

XO2

þ5XC3H8

XO2

þ13

2

Xn-C4H10

XO2

þ13

2

Xi-C4H10

XO2

þ8Xn-C5H12

XO2

þ8Xi-C5H12

XO2

þ19

2

Xn-C6H14

XO2

Fig. 1 displays the experimental temperature profiles

obtained in the four flames. The temperature profiles were

,1 mm closer to the burner surface in the methane/ethane

(noted C1C2), methane/ethane/propane (noted C1C2C3)

and natural gas (noted NG) flames, indicating that the trends

in changing the fuel composition leads to accelerate the

combustion process mainly because of methane substitution

by ethane. In the burnt gases, the temperature is ,60 K

higher when methane is partly substituted; this small

difference is principally explained by the flame stabilization

closer to the burner.

Table 2

The composition of the synthetic natural gas flame used in this work

CH4 C2H6 C3H8 n-C4H10 i-C4H10 n-C5H12 i-C5H12 n-C6H14

Mole fraction 88.4 £ 1022 9.52 £ 1022 1.84 £ 1022 0.09 £ 1022 0.07 £ 1022 0.05 £ 1022 0.05 £ 1022 0.05 £ 1022

Fig. 1. Experimental temperature profiles obtained in methane (C1),

methane/ethane (C1C2), methane/ethane/propane (C1C2C3) and synthetic

natural gas (NG) flames.

A. Turbiez et al. / Fuel 83 (2004) 933–941 935

The mole fraction profiles of the various species

analyzed have been plotted in Figs. 2–5. These figures are

given in increasing order of molar masses except for

reactants (CH4, C2H6, C3H8, n-C4H10, i-C4H10, n-C5H12,

i-C5H12, n-C6H14, O2), major species (CO, CO2, H2, H2O)

and reactive species (H, O, OH, HCO, CH3, C2H5).

Fig. 2 compares the consumption of the reactants mole

fraction profiles. The methane consumption is faster in the

C1C2 and C1C2C3 flames while other reactants are totally

consumed at the same distance from the burner (,5 mm). A

very close similitude was observed for oxygen consumption

in all flames (Fig. 2).

As shown in Fig. 3, presence of ethane or/and higher

alkanes (.C3) in the initial mixture has no influence on the

mole fractions of major products (H2, H2O, CO and CO2) in

the burnt gases, except a slight increase for CO2 and a

decrease for H2 and H2O. On the contrary in the flame front,

mole fraction profiles of these species are shifted towards

the burner when methane is substituted, which increases the

mole fraction of CO and H2 close to the burner; this shift

exhibits an increase of the reactivity of the mixture when

methane is partly substituted by higher alkanes.

Marked differences concern C2 and C3 stable interme-

diate species (ethane, acetylene, ethylene, propene and

Fig. 2. Experimental mole fraction profiles of reactants (O2, CH4, C2H6, C3H8, n-C4H10, i-C4H10, n-C5H12, i-C5H12, n-C6H14) (A; methane flame, K

methane/ethane flame, X methane/ethane/propane flame, W natural gas flame).

A. Turbiez et al. / Fuel 83 (2004) 933–941936

propane) as shown in Fig. 4. Propane is in an intermediate

species in the methane (C1) and methane/ethane (C1C2)

flames. Partly substituting methane by ethane increases its

peak mole fraction by a factor of 6.

Olefins are very sensitive to the fuel composition. The

main difference concerns ethylene and propene whose peak

mole fractions increase strongly respectively in the C1C2,

C1C2C3 and NG flames.

The increase of acetylene mole fraction is less strong

than the one observed for olefins.

Concerning oxygenated stable intermediate species, only

small quantities of formaldehyde (CH2O) have been

analyzed in the methane flame (Fig. 4).

The increase of the reactivity is also obvious by the shift

of the profiles of H, O and OH reactive species (Fig. 5)

which affects the peak values of experimental mole fraction

of intermediate hydrocarbon intermediate species. In the

burnt gases, mole fraction of O, OH and H are not strongly

affected by changing the composition of the fuel. Never-

theless, one can observe a slight decrease of H atom mole

fraction in NG flame (Fig. 5).

The reactive hydrocarbon intermediate species analyzed

in this work are ethyl, methyl and formyl radicals (Fig. 5).

An increase of methyl radical peak mole fraction is

observed in C1C2C3 and NG flames. The ethyl radical

was not detected in the methane flame whereas its profiles

exhibit the same maximum peak mole fraction in NG and

C1C2C3 flames. Contrary to methyl radical, ethyl profiles

are not shifted toward the burner in C1C2, C1C2C3 and NG

flames.

Globally the experimental results show that the oxidation

of the natural gas can be described appropriately by the

oxidation of methane—ethane— propane mixture. Indeed,

the nature of the species measured in C1C2 and C1C2C3

flames are the same and present very similar concentrations.

One notes a shift of mole fraction profiles obtained in the

C1C2C3 flame toward the burner. Thus, this work confirms

the investigation of Tan et al. [11] conducted on the

oxidation of natural gas in Jet Stirred Reactor conditions.

Our experimental results can be explained, at least

qualitatively, by using detailed mechanisms of C1/C2/C3

alkanes oxidation available in the literature [11,17–21]. The

kinetic models of Tan et al. [11] and Dagaut et al. [21] were

validated by using experimental data obtained in a Jet

Stirred Reactor for oxidation of C1 to C4 hydrocarbons in

the range 900–1450 K and 1–10 bars. The Warnatz

[17–19] and Miller et al. [20] mechanisms were validated

principally by using experimental results (flame velocity

and structure of free flames, concentration and temperature

profiles in burner-stabilized flames) obtained in premixed

flames of alkanes, alkenes and acetylene in a wide range of

pressure (0.1–1 bar).

Fig. 3. Experimental mole fraction profiles of H2, H2O, CO and CO2 (A methane flame, K methane/ethane flame, X methane/ethane/propane flame, W NG

flame).

A. Turbiez et al. / Fuel 83 (2004) 933–941 937

According to these mechanisms, methane is consumed

by H, O and OH attacks with methyl radical. In our

conditions, the composition of the fuel has a small influence

on these reactions; a slight decrease of these reactions rates

can be expected according to the lower quantity of CH4

(<10 %) present in the initial mixture of C1C2, C1C2C3

and NG flames.

The peak mole fraction of methyl radical is more

important in the C1C2, C1C2C3 and NG flames mainly

due to ethane presence in the fuel composition. In the

flames, ethane is mainly consumed by H abstraction via

C2H6 þ (H, OH, O) ! C2H5 þ (H2, H2O, OH) reactions.

The resulted ethyl radical is essentially consumed via two

reactions: C2H5 þ H ¼ CH3 þ CH3 and C2H5 þ M ¼

C2H4 þ H þ M. These reactions are responsible of the

quantities of methyl radical and ethylene in the C1C2,

C1C2C3 and NG flames.

In the C1C2C3 and NG flames, the thermal degradation

of n-propyl radical by b-scission, NC3H7 ¼ C2H4 þ CH3

becomes effective and explains the high concentration of

ethylene measured in these flames. The increase of the

concentration of ethylene affects acetylene mole fraction

directly according to the following sequence:

C2H4 ! C2H3 ! C2H2.

Fig. 4. Experimental mole fraction profiles of C2H6, C2H2, C2H4, C3H6, C3H8 and CH2O (A methane flame, K methane/ethane flame, X

methane/ethane/propane flame, W natural gas flame).

A. Turbiez et al. / Fuel 83 (2004) 933–941938

As we observed, the concentration of propene is

significantly affected by the initial composition of the

fuel. This effect can be attributed mainly to the C–H

thermal decomposition of isopropyl radical i-C3H7 þ M ¼

C3H6 þ H þ M and to the recombination of CH3 and C2H3

radicals CH3 þ C2H3 ¼ C3H6. This last reaction is the

major linkage between the C1/C2 and the C3 oxidation

routes, whereas the thermal decomposition of i-C3H7

dominates in the C1C2C3 and NG flames. The consumption

of propene proceeds via H-abstraction reactions leading to

allyl radical: C3H6 ¼ AC3H5 þ H and via the reaction

C3H6 þ H ¼ C2H4 þ CH3. These reactions are more

effective when propane is a reactant. So, propane presence

as reactant enhanced both the C3 oxidation route and the

formation of C2 and C3 intermediate species.

The chemical species analyzed in C1C2C3 and NG

flames are nearly identical (Figs. 2–5). A light gap is

observed for propene and ethylene. These observa-

tions show that the higher alkanes (n-butane, isobutane,

n-pentane, isopentane and n-hexane), present in the

natural gas, play a minor role in the oxidation of the

natural gas due to the low quantity of these alkanes

(extensively lower to 0.01%). In the higher temperature

range corresponding to the flames studied in this work,

Fig. 5. Experimental mole fraction profiles of CH3, C2H5, HCO, H, O and OH (Amethane flame,Kmethane/ethane flame,Xmethane/ethane/propane flame,W

natural gas flame).

A. Turbiez et al. / Fuel 83 (2004) 933–941 939

the primary steps of oxidation are hydrogen atom

abstraction reactions with active species and thermal

decomposition of the fuel molecule. The first route which

is generally faster than the second one leads to high alkyl

radicals whose number of isomers is determined by the

distinct carbon atoms number identified in the fuel

molecule; these alkyl radicals are mainly consumed by

thermal decomposition which occur preferentially in b-

position with respect to the radical site [19,22–24]. These

reactions lead to smaller alkyl radicals and olefins. When

we apply these rules to n-butane, isobutane, n-pentane,

isopentane and n-hexane, one can obtain 3 alkyl radicals

for n-hexane and n-pentane, and four alkyl radicals for

isopentane

n-C6H14 þ X ¼ 1-C6H13=2-C6H13=3-C6H13 þ XH

ðX ¼ OH;H;O;CH3Þ

n-C5H12 þ X ¼ 1-C5H11=2-C5H11=3-C5H11 þ XH

i-C5H12 þ X ¼ i-C5H11_1=i-C5H11_2=i-C5H11_3

þ i-C5H11_4þ XH

At high temperature, these alkyl isomers react principally

by the b-scission of C–C bond producing essentially

ethylene, propene, butenes or 1-pentene

1-C6H13 ¼ C2H4 þ C4H9

2-C6H13 ¼ C3H6 þ C3H7

3-C6H13 ¼ 1-C4H8 þ C2H5

3-C6H13 ¼ 1-C5H10 þ C2H3

1-C5H11 ¼ C2H4 þ C3H7

2-C5H11 ¼ C3H6 þ C2H5

3-C5H11 ¼ 1-C4H8 þ CH3

i-C5H11_1 ¼ C2H4 þ C3H7

i-C5H11_2 ¼ 2-C4H8 þ CH3

i-C5H11_3 ¼ i-C4H8 þ CH3

i-C5H11_4 ¼ 1-C4H8 þ CH3

The oxidation of n-butane forms but-1-yl (PC4H9) and

but-2-yl (SC4H9) radicals while isobutane produces iso-

butyl (i-C4H9) and tertio-butyl (TC4H9) radicals

n-C4H10 þ X ¼ SC4H9/PC4H9 þ XH

i-C4H10 þ X ¼ TC4H9/IC4H9 þ XH

Except tertio-2-butyl radical, the formed radicals react

principally by thermal decomposition producing ethylene or

propene

PC4H9 ¼ C2H4 þ C2H5

SC4H9 ¼ C3H6 þ CH3

IC4H9 ¼ C3H6 þ C2H3

Our experimental results confirm that at high temperature,

H-atom abstraction reactions play a minor role in the

consumption of high alkyl radicals. Indeed, high olefins

(hexenes, pentenes, butenes) were not detected in this study.

Moreover, the formation of propene, ethylene, butenes

(1-butene, isobutene and isomeric forms of 2-butene) is

expected to be early in the natural gas flame than in the other

flames. Indeed, these species are produced in the first steps of

high alkanes oxidation. In our conditions, as shown in Fig. 4,

only propene could be observed early in the natural gas flame.

This observation can be attributed mainly to the thermal

decomposition of 2-C5H11 and 2-C6H14. The formation of

these radicals are probably more effective than the others.

While studying n-pentane oxidation at atmospheric pressure

and 753K, Baldwin et al. [25] observed that the formation of

the radical 2-C5H11 is more important.

It appears according to this work that superior alkanes

play a minor role in the natural gas combustion. The kinetics

of the natural gas studied in these conditions is determined

extensively by methane/ethane/propane system. It is there-

fore reasonable to consider that the natural gas can be

represented appropriately by a methane – ethane – propane

mixture as reported previously in jet stirred reactor

conditions, at lower temperature and higher pressure [11,21].

4. Conclusion

Four laminar stoichiometric premixed CH4/O2/Ar, CH4/

C2H6/O2/Ar, CH4/C2H6/C3H8/O2/Ar and CH4/C2H6/C3H8/

n-C4H10/i-C4H10/n-C5H12/i-C5H12/n-C6H14/O2/Ar flames

have been studied at low-pressure. Experimental mole

fraction profiles of stable and reactive species have been

obtained by coupling molecular-beam/mass spectrometry

(MBMS) with gas chromatography/mass spectrometry

(GCMS) analyses. Temperature profiles were measured

with a coated thermocouple in the sampling conditions. The

experimental data showed that ethane and propane, present

in the composition of natural gas, have an important impact

on its combustion, especially on intermediate species

concentrations that would affect pollutant emissions. On

the other hand, the superior alkanes (.C4) can be

disregarded to correctly describe the combustion of the

natural gas. It appears therefore that the combustion of

methane/ethane/propane mixture represents appropriately

synthetic natural gas oxidation in stoichiometric and low

pressure conditions. It is interesting to know if the

conclusion of this work remains valid when one varies the

equivalence ratio and the pressure. Indeed, these two

parameters can affect considerably the relative importance

of some reactions. This work in progress in our laboratory

will be submitted for publication next.

The second stage of this investigation will be to use these

experimental results to validate a detailed kinetic model

representative of the natural gas combustion. However, it is

hoped that these results were helpful for other modeling

groups to validate detailed reaction mechanisms of natural

gas combustion. The modeling will permit to provide a more

precise information on the evolution in the reaction pathways

A. Turbiez et al. / Fuel 83 (2004) 933–941940

induced when changing the fuel composition. However,

some qualitative explanations could be given for our main

experimental results by using existing mechanisms.

Acknowledgements

The authors gratefully acknowledge the CNRS comput-

ing center IDRIS for its support. The Centre d’Etudes et de

Recherches Lasers et Applications (FR CNRS 2416) is

supported by the Ministere Charge de la Recherche, the

Region Nord—Pas de Calais and the Fonds Europeens de

Developpement Economique des Regions (FEDER).

References

[1] Miller JA, Bowman CT. Prog Energy Combust Sci 1989;15:287.

[2] Lamoureux N. PhD thesis, Universite d’Orleans, 1995.

[3] Higgin RMR, Williams A. Proc Combust Inst 1969;12:579.

[4] Burcat A, Scheller K, Lifshitz A. Combust Flame 1971;16:29.

[5] Burcat A, Crossley RW, Scheller K, Skinner GB. Combust Flame

1972;18:115.

[6] Crosley RW, Dorko EA, Sheller K, Burcat A. Combust Flame 1972;

19:373.

[7] Eubank CS, Rabinowitz MJ, Gardiner Jr. WC, Zellner RE. Proc

Combust Inst 1981;23:471.

[8] Frenklach M, Bornside DE. Combust Flame 1984;56:1.

[9] Spadaccini LJ, Colket MB. Prog Energy Combust Sci 1994;20:431.

[10] Hill PG, Hung J. Combust Sci Technol 1988;60:7.

[11] Tan Y, Dagaut P, Cathonnet M, Boettner JC. Combust Sci Technol

1994;102:21.

[12] Crunelle B, Pauwels JF, Sochet LR. Chim Phys 1997;94:433.

[13] Lazzara CP, Biordi JC, Papp JF. Combust Flame 1973;21:371.

[14] Vandooren J, Ballakhin VP, Van Tiggelen PJ. Archivum combustio-

nis 1981;1:229.

[15] Turbiez A. PhD thesis Universite de Lille, 1998.

[16] Wagner HW, Bonne U, Grewer T. Zeischrift fur Physicaliche Chemie

1960;26–93.

[17] Warnatz J. Proc Combust Inst 1981;18:369.

[18] Warnatz J, Bockorn H, Moser A, Wenz HW. Proc Combust Inst 1982;

19:197.

[19] Warnatz J. Proc Combust Inst 1984;20:845.

[20] Miller JA, Melius CF. Combust Flame 1992;91:21.

[21] Dagaut P, Luche J, Cathonnet M. Proc Combust Inst 2000;28:

2459.

[22] Pitz WJ, Westbrook CK. Combust Flame 1986;63:113.

[23] Axelsson EI, Brezinsky K, Pitz WJ, Westbrook CK. Proc Combust

Inst 1986;22:783.

[24] El Bakali A, Delfau JL, Vovelle C. Combust Sci Technol 1998;140:

69.

[25] Baldwin RR, Bennett JP, Walker RW. Proc Combust Inst 1977;16:

1041.

A. Turbiez et al. / Fuel 83 (2004) 933–941 941