Experimental study on nano-sized particles produced in flames
burning methane
M. Commodo1, E. Bichi1, A. D’Anna1, P. Minutolo1, R. Pagliara2, C.
Allouis2
1. Dipartimento di Ingegneria Chimica - Università Federico II, Napoli - ITALY
2. Istituto di Ricerche sulla Combustione - C.N.R., Napoli - ITALY
1. Introduction
Natural Gas is one of the primary sources of energy for many of our activities; moreover
natural gas is mostly composed of methane, therefore it is of great interest to study the
combustion of methane and its byproducts in different combustion systems.
During the last few years it has been shows that primary particles with typical sizes smaller
than three nanometers are easily produced close to the flame front of rich premixed flames
and in non-premixed combustion systems [1-3]. These clusters, which are probably composed
by polymers of polycyclic aromatic hydrocarbons, may growth for effect of particle-particle
coagulation process and for the addition of molecules from the gas-phase and lead to the
formation of graphitic-like structures with typical size larger than ten nanometres, defined as
soot or elemental carbon. The characterisation of the kinetic of the formation of the particles
emitted from combustion requires the development of experimental techniques able to detect
particles in the size range between 2-50 nm accounting also for their different chemical
composition.
At this purpose, in this paper we present an experimental study of Nanoparticles of Organic
Carbon (NOC), with a typical size range of 1 – 10 nm, and soot particles, with sizes in the
range 10 – 100 nm, formed in methane flames at atmospheric pressure. An optical diagnostic
method based on the interaction of the fifth harmonic of a Nd:YAG laser with combustiongenerated aerosols was performed to measure the concentration profiles of both NOC and
soot particle with high spatial resolution [4]. The high energy of the selected harmonic
enhances fluorescence from aromatic chromophoric groups and also allows soot particles to
heat up and emit incandescent radiation [5]. Optical measurements are complemented by
Differential Mobility Analysis (DMA) measurements, in selected conditions, which allow the
measurement of the size distribution functions of the formed particles. The experimental data,
have been performed in different flame conditions: laminar premixed, laminar diffusion and
turbulent diffusion flames. In the present work we compare the amounts of particulates from
the methane combustion, with respect to other gaseous fuels such as ethylene in similar flame
conditions.
Finally, the experiments were performed on three different systems, to the aim to study the
effects of premixing, diffusion and turbulence on the particles formation in methane
combustion.
2. Experimental procedure
Methane/oxygen laminar premixed flames were stabilized on a McKenna burner. Fuel and
oxygen flows were selected to produce flames with different equivalence ratio,
(C/O)/(C/Ostoich), in the range 1 to 2,5, keeping the cold gas velocity constant at 4,74 cm/s.
A laminar diffusion methane flame was stabilized on a 12 mm diameter un-cooled vertical
tube for the fuel and a concentric tube (108 mm i.d.) for air. Details of the burner construction
1
30th Meeting on Combustion
and operation are given elsewhere [3]. The flow rate was chosen at 4 cm3/s with an air flow
rate of 666,7 cm3/s to produce the same flame studied by Smooke et al. [6].
A vertical turbulent non-premixed flame of methane has been obtained by flowing methane
from a 2.5 mm internal diameter nozzle into ambient air with a fresh gas velocity of 24.4 m/s,
corresponding to a Reynolds number of 3.5E6. Details of the burner configuration are given
in ref. [7].
Laser Induced Fluorescence (LIF) and Laser Induced Incandescence (LII) measurements have
been performed using an ultraviolet laser source corresponding to the fifth harmonics of a
Nd:YAG laser at 213 nm with a pulse duration of 8 ns FWHM operating at 10 Hz. The laser
pulse energy was kept constant at 1,5 mJ. The laser beam was focused at the centre of the
burner with a 500 mm focal lens. The signal was collected at 90° respect to the incident beam
through a 280 μm entrance slit of a spectrograph and was detected by a gated ICCD camera.
All the spectra were obtained by an accumulation of 150 shots and corrected for the
background and the wavelength-dependent sensitivity of the camera. Typical, time resolved,
emission spectra are reported elsewhere [4, 5]. Fluorescence signals are emitted for about
20 ns after the excitation. The spectra show a broadband signal in the region between 270 and
400 nm with a maximum at about 340 nm, typical of fluorescence from aromatic compounds
[8], and typical C2-emission bands at 465 and 516 nm. Incandescence signals are instead
emitted for hundreds of nanoseconds, they show a continuum with a maximum at about
600 nm, which corresponds to a black-body Planck law with a temperature of about 4000 K,
superimposed to it C2-emission bands are present.
The different temporal decay and emission spectral regions of LIF and LII signals allows to
selectively detect the two signals [9]. The fluorescence is detected in the ultraviolet acquiring
the signal synchronizes to the laser pulse with an acquisition duration of 20 ns. The
incandescence emission is detected 30 ns after the laser pulse fixing the acquisition duration
at 100 ns [5]. In the delayed spectra the signal can be entirely attributed to LII, while with the
experimental set up used for fluorescence measurements, in regions of the flame where the
soot concentrations is not to high, the contribution of incandescence at the emission signal at
300 nm is negligible.
LIF and LII calibrated with the extinction measurements allows to measure the volume
fraction of the particles [4]. In addition, scattering measurements are performed to determine
the mean size of the particles at different locations in the flame.
For the laminar premixed methane flame measurements were performed at different heights
above the burner and for different C/O ratios.
The diffusion flames, instead, were investigated at different heights above the nozzle: from
1 cm to 35 cm for the turbulent flame and from 1 cm to 6 cm for the laminar one, and at
different radius with a spatial resolution of 1 cm on the axis and 1 mm radially.
Differential Mobility Analysis (DMA) measurements have been performed in the turbulent
flame at the same locations of the optical measurements, using a TSI nano-DMA n°3936, that
allows us to obtain particle size distributions in the range 3 – 65 nm. Gases are sampled using
a 0.8 mm ID and 1.5 mm OD probe vertically positioned into the flame, for effect of the low
pressure generated by an air flow through a calibrated throat [5].
3. Results and Discussion
From fluorescence and incandescence signals, calibrated with the extinction data, we have
estimated the volume fraction of NOC and soot particles in the described flames.
At this purpose in figure 1 the volume fractions of both the particles classes, NOC (a) and
soot (b), are reported, as function of the equivalent ratio. NOC are detectable just above the
stoichiometric value and its concentration increases with the equivalent ratio. Soot particles,
2
Italian Section of the Combustion Institute
instead, are practically absent in flames with equivalence ratio ranging from 1 to 2, for these
flames the incandescence signal of soot particles was not detectable. Thereafter, soot volume
fraction quickly increases. In figure 2 the volume fraction of NOC and soot for two flames,
φ=1,76 and φ=2, are reported as function of the height above the burner, (HAB). Only in
correspondence of φ=2 soot is detectable, and its concentration increases at higher HAB. In
order to estimate the mean size of the particles, scattering measurements were performed in
the same flame conditions. Figure 3 shows, the scattering profile measured by using the fifth
harmonic (213 nm), for the flame with φ=2, at different HAB, fig. (a). On the same figure the
estimated gas contribution is reported too. It is possible observe that the Qvv signal measured
is larger than the gas one, also in flame zones where soot is absent and only NOC particles are
present.
0,25
0,4
0,35
(a)
(b)
0,2
fv soot, ppm
fv NOC, ppm
0,3
0,25
0,2
0,15
0,1
0,15
0,1
0,05
0,05
0
0
1
1,2
1,4
1,6
1,8
2
2,2
2,4
1
2,6
1,2
1,4
Fig. 1
1,6
1,8
2
2,2
2,4
2,6
(C/O)/(C/O)stoich.
(C/O)/(C/O)stoich.
NOC (a) and soot (b) volume fraction in the laminar premixed flame for different
equivalent ratios at 10 mm HAB.
0,12
0,25
(a)
0,1
(b)
0,2
NOC
NOC
fv, ppm
fv, ppm
0,08
0,06
0,15
0,1
0,04
Soot
0,05
0,02
Soot
0
0
0
5
10
15
20
25
0
5
10
HAB, mm
Fig. 2
20
25
NOC and soot volume fraction in the laminar premixed flame for two equivalent
ratios: 1,76 (a) and 2 (b) at different HAB.
12
Scattering coefficient, cm-1sr-1
1,0E-05
(a)
(b)
10
Qvv measured
D63, nm
8
1,0E-06
6
4
2
Gas contributions
0
1,0E-07
0
5
10
15
20
0
HAB, mm
Fig. 3
15
HAB, mm
5
10
15
20
HAB, mm
Scattering coefficient measured in the laminar premixed flame with equivalent ratio
2 at different HAB (a), and mean particles diameter (b).
3
30th Meeting on Combustion
From the scattering coefficients, after subtraction of the gas contribution, and from the
particles volume fraction, previously determined, the mean particles diameter, D63, was
estimated using the Rayleigh theory. Figure 3b, shows, the mean particles size which is about
3 nm in a region of the flame where only NOC particles are present, and increases at higher
HAB where soot begins to appear.
Methane/oxygen combustion in laminar premixed condition, as expected, needs very fuel-rich
condition in order to produce and therefore to emit soot particle, while NOC is formed also in
less rich conditions in the premixed flames investigated and with a higher volume fractions
respect to the soot particles. Nevertheless, methane/oxygen premixed combustion is not really
representative of the methane combustion in practical systems, therefore in order to better
understand the formation of by-products from methane combustion is of great interest
investigate diffusion flames. These flames have wider application in many industrial fields.
On the bases of the results obtained on premixed methane flames, its relevant to study their
emission in consideration of the fact that in a diffusion flame a wide range from fuel-rich to
fuel-lean conditions are presents.
At this purpose, in figure 4 the radial concentrations of NOC (fig. a), and soot (fig. b),
measured in the laminar diffusion flame, are reported. As in the case of a similar ethylene
flame [3], in methane flames soot formation is located in a narrow annular region close to the
maximum temperature zone, and its maximum volume fraction is measured just after the
decrease of organic carbon particle volume fraction. At increasing heights above the burner,
the radial position of the maximum soot volume fraction shifts towards the centerline. Instead,
NOC is preferentially formed in the lower part of the flame and closer to the flame axis. The
maximum volume fraction of organic carbon particles decreases at increasing heights in the
flame shifting toward the centerline.
1
0,18
z20
z25
z30
(a)
0,9
0,8
0,14
fv soot, ppm
0,7
fv NOC, ppm
z20
z25
z30
z35
(b)
0,16
0,6
0,5
0,4
0,3
0,12
0,1
0,08
0,06
0,2
0,04
0,1
0,02
0
0
0
1
2
3
4
5
6
r, mm
Fig. 4
0
1
2
3
4
5
6
r, mm
Radial concentration profiles of NOC (a) and soot (b) at different flame heights in
the laminar diffusion flame.
From these data it is possible to observe that methane flame produces a lower amount of both
NOC and soot respect to a similar laminar diffusion flame of ethylene [3] but the reduction of
soot is stronger than that of NOC. In particular, it is possible to observe that the ratio between
the maximum concentration value of soot to NOC in the ethylene flame is about 1.0 while in
the methane flame it is 0.2.
These results are confirmed by the profiles of light scattering reported in Fig. 5(a). From the
scattering data, after evaluation of gas contribution, and with the evaluation of the total
volume fraction of the particles, reported above, we have obtained a rough estimate of the D63
of the total particles distribution function which is reported in Fig. 5(b). Also in this case the
particles mean size has been evaluated using the Rayleigh equation for light scattering using
the optical properties for NOC and soot reported in [2].
Close to the flame centerline, where the concentration of soot is very low, the mean size of
4
Italian Section of the Combustion Institute
the particle is of the order of 2-3 nm whereas in the annular flame region where soot is
formed, the mean particle size is of the order of 10 nm.
14,0
1,0E-05
z20
(a)
z25
(b)
12,0
z20
z25
z30
z30
z35
-1
Size, nm
Qw, cm sr
-1
10,0
1,0E-06
8,0
6,0
4,0
2,0
0,0
1,0E-07
0
1
2
3
4
5
0
6
1
2
r, mm
Fig. 5
3
4
5
6
r, mm
Radial scattering profiles at different flame heights (a) and radial mean particle size
at different flame heights (b) in the laminar diffusion flame.
To complete the information about the methane combustion, LIF and LII were also performed
in a turbulent diffusion flame. Figure 6(a) reports NOC volume fractions, measured from
through ultraviolet LIF, at three heights from the nozzle: 7 cm, 10 cm and 20 cm, as a
function of the radius.
At 7 cm, NOC maximum concentration is located at 2.5 mm from the flame axis with a value
of 0.25 ppm. Also at 10 cm NOC shows also a maximum between 2.5 and 3 mm with quite
the same value measured at 7cm (0.26 ppm). At 20 cm NOC volume fractions increase up to
0.36 ppm, in this case at the flame axis.
Figure 6(b) reports soot concentrations at two heights above the nozzle, namely 10 and 20 cm
vs. the radial positions. At 10 cm soot particles shows a maximum of about 0.015 ppm at
2 mm from the flame axis. At 20 cm, which approximately corresponds to the location of the
maximum soot concentration, the profile shows a maximum at the center of the flame with a
value of about 0.035 ppm.
0,4
0,04
0,35
7 cm
10 cm
20 cm
(a)
0,03
fv soot, ppm
fv NOC, ppm
0,3
10 cm
20 cm
(b)
0,035
0,25
0,2
0,15
0,025
0,02
0,015
0,1
0,01
0,05
0,005
0
0
0
5
10
15
20
25
30
35
r, mm
Fig. 6
0
5
10
15
20
25
30
35
r, mm
Radial concentration profiles of NOC (a) and soot (b) at different flame heights in
the turbulent diffusion flame.
In turbulent diffusion flame is very difficult to estimate the contribution of gaseous species to
light scattering so it is difficult to employ light scattering measurements for the evaluation of
particles sizes. In this flame the size distribution function has been therefore measured by
DMA. Figure 7 reports the size distribution functions measured on the flame axis at 10 cm
from the jet nozzle. Particle diameters range from 3 nm, the limit of the detection system, up
to 20nm. The number concentration of the particles with sizes below 10 nm is more than one
order of magnitude higher than that of particles with sizes larger than 10 nm confirming the
5
30th Meeting on Combustion
very low concentration of soot in the analyzed flame.
dN/dLOG(DP), cm-3
1.E+06
1.E+05
1.E+04
1.E+03
1
10
100
DP, nm
Fig. 7
Size distribution function of the particles detected at 10 cm along the flame axis of
the turbulent methane flame.
4. Conclusion
NOC and soot particles have been measured in premixed and non-premixed methane flames
at atmospheric pressure by using a calibrated optical procedure based on the use of a
ultraviolet laser source. The optical results have been, then, confronted with size distribution
function measured by DMA. The experimental data, performed in different flame conditions:
laminar premixed, laminar diffusion and turbulent diffusion flames, show that methane forms
a low amount of particulates, sensibly lower than that produced by other gaseous fuels such as
ethylene in similar flame conditions. However, methane particulates comprise mainly, very
small particles with diameter in the range 3-10 nm. The concentration of the larger particles is
very low, soot is formed only in correspondence of a very high C/O ratio.
5. References
1.
2.
3.
4.
5.
6.
7.
8.
9.
Sgro L.A., Basile G., Barone A.C., D'Anna A., Minutolo P., Borghese A., D'Alessio A.,
2003, Chemosphere, 51(10), 1079.
Cecere, D., Sgro, L. A., Basile, G., D’Alessio, A., D’Anna, A., Minutolo, P., 2002,
Combust. Sci. Technol. 174(11-12), 377.
D’Anna, A., Rolando, A.,Allouis, C., Minutolo, P., D’Alessio, A.,: Proceedings of The
Combustion Institute, 30:1449-1456 (2005).
Commodo, M., Violi, A., D'Anna, A., D'Alessio, A., Allouis, C., Beretta, F., Minutolo,
P., Combust. Sci. Technol. 179:387-400 (2007)
D’Anna, A., Commodo M., Violi, S., Allouis, C., Kent, J.,: Proceedings of The
Combustion Institute, 31:621-629 (2007).
Smooke, M.D., McEnally, C.S., Pfefferle, L.D., Combust. Flame 117: 117-139, (1999).
Commodo, M., Bichi, E., D’Anna, A., Minutolo, P., Pagliara, R., Allouis, C., Third
European Combustion Meeting, Crete, Greece, April, (2007).
Minutolo, P., Gambi, G., D’Alessio, A., D’Anna, A., Combust. Sci. Technol., 101 (1/6)
309-325 (1994).
Vander Wal, R.L., Jensen, K. A., Choi, M. Y., Combustion and Flame, 109: 399-414
(1997).
6