Combustion Science and Technology
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/gcst20
Soot Formation and Growth in Toluene/
Ethylene Combustion Catalyzed by Ruthenium
Acetylacetonate
Fanggang Zhang, Cong Wang, Juan Wang, Sönke Seifert & Randall E. Winans
To cite this article: Fanggang Zhang, Cong Wang, Juan Wang, Sönke Seifert & Randall E. Winans
(2021): Soot Formation and Growth in Toluene/Ethylene Combustion Catalyzed by Ruthenium
Acetylacetonate, Combustion Science and Technology, DOI: 10.1080/00102202.2021.2013830
To link to this article: https://doi.org/10.1080/00102202.2021.2013830
Published online: 14 Dec 2021.
Submit your article to this journal
Article views: 33
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=gcst20
COMBUSTION SCIENCE AND TECHNOLOGY
https://doi.org/10.1080/00102202.2021.2013830
Soot Formation and Growth in Toluene/Ethylene Combustion
Catalyzed by Ruthenium Acetylacetonate
Fanggang Zhang
a
, Cong Wanga, Juan Wang
a
, Sönke Seifertb, and Randall E. Winansb
a
School of Energy and Power Engineering, Beihang University, Beijing. PR China; bX-ray Science Division,
Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois, USA
ABSTRACT
ARTICLE HISTORY
Ruthenium-based compounds are efficient catalysts to enhance com­
bustion performance and suppress soot emission. However, systematic
mechanism and in-situ research on reducing soot are rarely addressed.
In this study, ruthenium acetylacetonate (Ru(C5H7O2)3, Ru(acac)3) was
utilized as the catalyst precursor in different concentration conditions
to investigate its impacts on the size, volume, and morphology of the
soot particles in the ethylene flames with central toluene injection. The
soot particles were detected by in-situ small-angle X-ray scattering
(SAXS) and ex-situ transmission electron microscopy (TEM). This study
confirms that Ru(acac)3 can suppress the surface growth for the primary
soot particles, but no significant change of the morphology has been
discovered for the aggregates. Compared with the undoped flame, the
volume of the soot particles in the Ru(acac)3-doped flames is lower and
the particle size is smaller, indicating a valid inhibition effect of
Ru(acac)3 on the soot emission by affecting its surface growth and
oxidation process. Despite the remarkable distinction between the
undoped and doped flames, no obvious difference is found between
the flames with different catalyst concentrations.
Received 22 September 2021
Revised 29 November 2021
Accepted 30 November 2021
KEYWORDS
Soot; catalyst; ruthenium
acetylacetonate; surface
growth; oxidation; in-situ
small angle X-ray scattering
Introduction
Soot particles, of particular concern in the environment pollution, impact human health
negatively. Inhalation of soot particles might cause respiratory diseases, cardiovascular
diseases, pulmonary function loss, or even death (Guaita et al. 2011). They also have
a serious effect on environment and global warming (Tollefson 2013). In addition, carbon
deposition on the engine’s inner wall, fuel nozzle, and other places caused by incomplete
combustion reduces combustion efficiency and engine performance. Therefore, there is an
urgent need to solve soot emission issue efficiently.
Enhancing soot oxidation is an efficient way to suppress emission of soot particles. Metalbased catalysts exhibit a vital potentiality in soot elimination and have been widely used in
catalytic soot combustion over the past years. Many investigations have demonstrated that
noble metals and their modified catalysts (Lee et al. 2018; Oi-Uchisawa et al. 2000; Ren et al.
2019), alkaline metal (Li et al. 2012, 2016), transition metals (Howard and Kausch 1980; Hu
et al. 2017) and metal oxides (Gao et al. 2018; Neeft, Makkee, Moulijn 1996) are capable to
enhance soot oxidation. Marsh et al. (2007)indicated that ferrocene, ruthenocene, iron
CONTACT Juan Wang
100191, PR China
juanwang@buaa.edu.cn
© 2021 Taylor & Francis Group, LLC
School of Energy and Power Engineering, Beihang University,Beijing
2
F. ZHANG ET AL.
naphthenate, and methylcyclopentadienyl manganese tricarbonyl (MMT) are quite effective
in soot reduction in JP-8 lean flames. Likewise, Kim and Hahn (2016) reported that the
catalytic effect of iron compounds can greatly reduce activation energy for soot oxidation
reaction. In addition, Wei and Lee (1999) conducted a pyrolysis experiment of polystyrene
doped with MnSO4 in inert conditions, and they pointed out that MnSO4 addition decreased
the amount of Polycyclic Aromatic Hydrocarbons (PAHs) due to the transition metal
chelation oxidation mechanism. In our recent study (Tang et al. 2019), an in-situ study
was performed regarding effects of nickel acetylacetonate on soot inception and growth in an
ethylene flame with toluene injection. It was found that Ni additive reduced the primary soot
particle size and soot volume fraction significantly, and the additive also delayed the soot
surface growth, eventually leading to the primary soot particles with rough surfaces. To
explore catalysts for reducing soot emission more efficiently, investigations on a series of
catalysts and further studies of detailed catalytic mechanism are needed.
Platinum group metals behave as effective catalysts in soot reduction (Castoldi et al.
2017; Yashnik and Ismagilov 2019). Ruthenium is the cheapest metal in the platinum group
and it has been used in soot control. Castoldi et al. (2017) found that the ruthenium catalyst
promotes soot oxidation by roughly 90% and the Ru-based catalyst has higher activity in the
soot combustion than the Pt-based catalyst. Besides, there are also many studies demon­
strating that the catalytic reactivity of the metal oxide catalysts for soot oxidation was
significantly enhanced by doping with Ru nanoparticles (Aouad et al. 2007; Ledwa, Pawlyta,
Kepinski 2018; Nascimento et al. 2014). The temperature of carbon-black oxidation
decreased by 100–150°C after the addition of ruthenium to ceria as the catalyst (Aouad
et al. 2007). However, all of these experimental studies related to the impacts of Ru-based
catalysts on soot particles are ex-situ. To understand the mechanism how Ru-based catalysts
impact the soot particles and further improve the design of catalysts, it is essential to
perform experiments by in-situ measurement.
In this work, Ru(acac)3 of different concentrations was used as the organometallic catalyst
precursor in ethylene base flames with central toluene injection. Ru(acac)3 was selected as the
Ru-based catalyst precursor because it is soluble in toluene. Synchrotron small-angle X-ray
scattering (SAXS) was used to in-situ detect the information of the size, morphology, and
volume fraction of the soot particles, and the data were analyzed by fitting the scattering
intensities based on SAXS theory. Till now, many researchers have shown interest in investigat­
ing soot and fuel-soluble metal catalysts by utilizing SAXS (Braun et al. 2004; Di Stasio 2017;
Mitchell et al. 2009; Tang et al. 2019; Wang et al. 2015; Yon et al. 2018). It is worth mentioning
that this is the first study to in-situ investigate the effects of Ru-based catalysts on the soot
particles. As a complementary method, transmission electron microscopy (TEM) was used to
study the morphological and structural properties of soot, and the particle statistical size
distribution.
Experimental Methods
Experimental setup
In this study, the SAXS experiments were conducted on the beamline 12-ID-C at the
Advanced Photon Source in Argonne National Laboratory, USA. An 8 keV incident
X-ray with 0.6 mm in width and 0.1 mm in height was used. The experimental exposure
COMBUSTION SCIENCE AND TECHNOLOGY
3
time was 0.5 s and the SAXS signals were recorded by the CCD detector with a resolution of
2048 × 2048 pixels and normalized by the incident photon intensity. The 2D platinum CCD
detector was 2 m away from the flame. The transmitted X-ray beams were blocked by
a beam stop before they arrived at the detector.
A Hencken burner was used in our experiments as shown in Figure 1. As can be seen in
the enlarged figure of the burner surface, the diameter of each fuel tube is approximately
0.5 mm, while the oxygen tubes are regular hexagonal in cross-section with 0.4 mm each
side length. The total cross-section area of the fuel and oxygen flow is 25.4 × 25.4 mm,
shielded by a 6.3 mm wide nitrogen co-flow at 7.730 L/min. The flow rates of ethylene and
oxygen were 0.159 and 0.400 L/min, respectively. Another nitrogen line at 0.900 L/min was
mixed with oxygen before flowing into the burner. Different concentrations of Ru(acac)3
were dissolved in toluene (50, 100, 150 ppm by weight) and the solution was aerosolized by
an aerosol generator (SONAER, Inc., 241PG). The solution was aerosolized at a rate of
0.063 mL/min, and then carried into a 1 mm inner diameter central tube of the burner by
nitrogen at 0.114 L/min. It should be noted that a Hencken burner is to generate flat flames
over the fuel injection region in normal operation. However, in this study, the flame is more
like a laminar jet flame with the toluene solution injected from the central tube. And toluene
was used in all cases to enable the fuels’ concentrations are the same for both doped and
undoped flames.
Data analysis of SAXS
SAXS is an online noninvasive detection approach that is particularly useful in studying the
size distribution and structure of soot particles. The scattered intensity I(q) mainly relies on
the size, shape, and number density N of particles in a dilute suspension where the internal
reactions can be neglected. The basic equation for I(q) versus the scattering vector q is
given by
Figure 1. Schematic layout for the SAXS experiments.
4
F. ZHANG ET AL.
I ðqÞ ¼ N ½F ðqÞ�2
(1)
where F(q) is the coherent sum of the scattering amplitudes of the individual scattering
centers within the particle given by the Fourier transform of the electron density distribu­
tion (Brumberger 2013). The scattering vector q is defined as
q ¼ ð4π=λÞsinðθ=2Þ
(2)
where λ is the wavelength of the incident X-ray and θ is the scattering angle.
For values of qRG< 1, Guinier et al. (1955) showed
�
I ðqÞ ¼ Gexp q2 R2G =3
(3)
where G is the pre-exponential factor given by G ¼ NV2 r2e Δρ2 , V is the volume of the
particle, re is the Thomson scattering length equal to the classical electron radius
(2.818 × 10−15 m), and RG is the radius of gyration of a particle relative to its center of
gravity. The quantity Δρ ¼ ρ ρM is the contrast electron density or relative scattering
length density of the particles with respect to the medium. A lnI ðqÞ vs q2 plot, i.e., the
Guinier plot, can be obtained by taking the logarithm of Eq. (3), which is well-known used
in small-angle scattering for determining RG.
Another expression of I(q) is proposed when qRG ≫ 1,
I ðqÞ ¼ 2πre2 NΔρ2 Sq
4
(4)
where S is the particle surface area. Eq. (4) refers to the Porod region, and the Porod exponent
P that related to the fractal dimension is summarized as follows (Brumberger 2013):
8
P ¼ 4sharpinterface
>
<
3 � P < 4surfacefractal
p
I ðqÞ / q
(5)
>
P < 3massfractal
:
P � 2gaussianpolymerchain
Beaucage and Schaefer (1994) combined a Guinier and an associated Porod region into
a unified exponential scattering function as one level, which can be more generally applied
to mass fractal and surface fractal as well as those having smooth surfaces. This function is
defined as
�h �
�p
pffiffi �i3
�
2 2
I ðqÞ ¼ Gexp q RG =3 þ B erf qRG = 6 =q
(6)
where G is the Guinier pre-exponential factor as indicated above, B ¼ 2πNΔρ2 S is the Porod
pre-factor, erf is the error function, and P is the power law slope, i.e., the Porod exponent.
For two-level structures consisting of primary particles (level one) and aggregates (level
two), the above equation may be extended to (Beaucage 1995; Beaucage and Schaefer 1994):
�
�p
pffiffi �i3
�
� h �
2 2
2 2
I ðqÞ ¼ Gexp q RG =3 þ Bexp q Rsub =3
erf qRG = 6 =q
(7)
�h �
�Ps
pffiffi �i3
�
2 2
þ Gs exp q Rs =3 þ Bs erf qRs = 6 =q
COMBUSTION SCIENCE AND TECHNOLOGY
5
where Rs is the radius of gyration for small-scale particles, and these particles agglomerate to
form large-scale particles with RG. Mostly, the value of Rsub (the high-q cutoff for the power
law region) is identical to Rs. Besides, Gs,Bs and Ps are the corresponding factors for the
small-scale particles.
Moreover, the Porod invariant Q is obtained through the integral of the q2 · I(q) vs q plot
(Pilz, Glatter, Kratky 1982),
1
Q ¼ ò I ðqÞq2 dq
0
(8)
For homogeneous particles, Q is proportional to the volume fraction φV of the particles:
Q ¼ 2π2 Δρ2 φV
(9)
where Δρ is the difference of the scattering length densities between the medium air and
soot. Because the scattering length density of air is much smaller than that of carbon, thus
we took Δρ ≈ ρsoot = 15.26 × 1010 cm−2 in this study.
The scattering intensity profiles were fitted and analyzed by using Irena tool suite
(Ilavsky and Jemian 2009) in Igor Pro software package. Unified fitting method developed
by Beaucage and Schaefer (1994) are suitable for polydisperse particles with multiple
structure levels and was used to obtain RG and the Porod exponent P in this work.
Besides, the particle volume distribution can be yielded by taking the inverse Fourier
transform of the measured I(q) vs q relation in the total non-negative least-squares
(TNNLS) method (Merritt and Zhang 2005). This unified fitting procedure applies an
interior-point gradient method to solve the so-called totally nonnegative least-squares
problems. Before fitting, the SAXS intensities was needed to be calibrated by subtracting
the background, which contains noises scattered from air and instrument windows. It is
challenging to subtract a suitable background to get an accurate signal. The gas density in
the flame regime is lower than that in the surrounding air, which often lead to an over­
estimation of the background (Mitchell et al. 2009). In an attempt to compensate for this
error, we corrected the scattering curve of the ambient air by reducing the air density at the
temperature comparable to the flame, and used this corrected curve as the air background in
the flame. Besides, the measured raw intensity data was normalized by the incident photon
intensity, sample transmission, and the solid angle subtended by the detector pixels. The
scattering intensity of glassy carbon was measured as the standard sample to calibrate the
SAXS intensities of the target samples. Thus, we could obtain the absolute scattering
intensities of SAXS signals.
Following SAXS analysis of the soot particles, TEM was adopted as a complementary
characterization method to identify soot morphology and size. In this work, we collected the
soot particles by using C/Cu TEM grids. TEM images (25,000×) were obtained by a 200 kV
JEM-2200FS TEM with a Gatan digital camera in University of Science and Technology
Beijing. To determine the size distribution and the mean diameter of the primary particles,
three hundred particles at least were counted. The primary particles are merged together, so
only the particles with strong contrast and clear boundary were picked to determine the
particle size.
6
F. ZHANG ET AL.
Results and discussion
Effect of catalysts on soot morphology
TEM images were obtained for the soot particles collected at 30 and 40 mm heights above
the burner (HAB) for each flame. The comparisons of the soot morphology between the
undoped flame and Ru(acac)3-doped flames are shown in Figure 2. It is observed that the
aggregated particles are composed of near-spherical primary particles and behave as long
chains with quite loose structure. The particles in the undoped flame are generally larger
than those in the 50, 100, and 150 ppm Ru(acac)3-doped flames.
In our previous study (Tang et al. 2019), lean flame experiments containing only catalyst
particles were conducted, and sooting flames containing both soot and metal particles were
also performed. All results showed that the concentration of the metal or metal oxide
nanoparticles is very small and below the detection limit of SAXS measurement even in the
condition with 450 ppm additive, which is larger than the highest concentration (150 ppm)
we chose in this work. Therefore, the SAXS signals in these measurements are only from the
soot particles.
The scattering intensities I(q) versus scattering vector q were obtained at q range from
about 0.04 to 0.8 nm−1. Considering the correlation R ~ π/q (nm), this q range approxi­
mately corresponds to particle radius from 4 to 79 nm including primary particles and small
aggregates. Figure 3 presents the selected scattering curves at 20, 25, 30, 35 and 40 mm HAB,
respectively. It is noted that the scattered intensities are close to each other at HAB above
40 mm, which makes it difficult to compare individual scattering curves at different HABs.
Thus, only the scattered intensities below 40 mm are presented. It is obvious that I(q)
increases with increasing HAB for each flame, indicating the total volume of the soot
particles increases with increasing HAB. Detailed discussions after data fitting using
SAXS theory are presented later. Comparing the undoped flame with the doped flames in
Figure 2. TEM images (25,000× magnification) of the soot particles sampled at HAB = 30 mm, 40 mm of
the undoped flame and Ru(acac)3-doped flames. The first row is the soot sampled from 30 mm HAB, and
the second row is the soot sampled from 40 mm HAB. (a, e) Undoped flame. (b, f) 50 ppm Ru(acac)3doped flame. (c, g) 100 ppm Ru(acac)3-doped flame. (d, h) 150 ppm Ru(acac)3-doped flame.
COMBUSTION SCIENCE AND TECHNOLOGY
7
different concentration conditions, it can be seen that I(q) decreases significantly with
Ru(acac)3 dopant. The intensity uncertainties including the measured systematic errors
and errors from the data reduction are shown in Figure 3(b). It can be seen that the
uncertainties are much smaller than the differences between any two profiles. So the
uncertainties do not affect the trends of the curves at different HABs. Since the uncertainties
are so small, we do not show them in other figures.
It seems that the scattering intensities below 27 mm HAB are approximate to the
background curve, which makes it difficult to fit. Thus, only the fitting results of the
scattering curves in the range of 27―40 mm are presented. Figure 4 presents such an
example of the application of unified fitting at HAB = 35 mm. Each term appearing in Eq.
(7) is represented by different colored lines in the figure. Due to the limit of the SAXS
measurement range, only the primary particles for level 1 at high q and the power law region
of the soot aggregates for level 2 at low q can be detected in our measurement. The Guinier
region of level 2 is omitted by setting G to 0 and RG is automatically set to 1010. Therefore,
we can obtain the size and morphology information of the primary particles for level 1, and
some morphology information of the aggregates for level 2 can also be acquired in this
study. Unfortunately, large aggregates are out of the measurement range in this work .
The Porod exponent P was obtained from the two-level unified fitting to reflect shape and
surface morphology of the primary particles and the aggregates. As shown in Figure 5, the
Porod exponent of the primary particles increases with increasing HAB and finally gets close
Figure 3. I(q) of the soot particles in the undoped and doped flames at 20, 25, 30, 35 and 40 mm HAB,
respectively. (a) Comparison between the undoped and 50 ppm Ru(acac)3-doped flames. (b) Enlarged
figure of Figure 3(a) with uncertainties. (c) Comparison between the undoped and 100 ppm Ru(acac)3doped flames. (d) Comparison between the undoped and 150 ppm Ru(acac)3-doped flames.
8
F. ZHANG ET AL.
Figure 4. An example of unified fitting of SAXS scattering intensity at HAB = 35 mm.
Figure 5. The Porod exponents of the primary particles (black open circles: undoped flame, red open
triangles: 50 ppm Ru(acac)3-doped flame, green open triangles: 100 ppm Ru(acac)3-doped flame, blue
open triangles: 150 ppm Ru(acac)3-doped flame) and aggregates (black closed circles: undoped flame,
red closed triangles: 50 ppm Ru(acac)3-doped flame, green closed triangles: 100 ppm Ru(acac)3-doped
flame, blue closed triangles: 150 ppm Ru(acac)3-doped flame).
to 4. According to the previous studies, P equal to 4 represents that the particles have smooth
surfaces (Brumberger 2013), and the exponent between 3 and 4 implies rough surfaces. That
COMBUSTION SCIENCE AND TECHNOLOGY
9
is to say, in this study, the surfaces of the primary particles are flat and rough at low HABs
and grow smoother with increasing HAB. In Figure 5, it is seen that Ru(acac)3 addition
decreases the P exponent of the primary soot particles, indicating that Ru(acac)3 can depress
the surface growth of the primary particles. It could be due to the inhibition effect of the
metal additives on the growth of PAHs, which suppresses the process of PAH-soot surface
addition. This is consistent with the report of Ritrievi, Longwell, and Sarofim (1987) who
found the addition of ferrocene reduced the surface growth rate of soot particles. Similarly, in
our previous study (Tang et al. 2019), it was found that nickel additive could delay the soot
surface growth. Another possible explanation for the rough surfaces of the soot particles with
Ru(acac)3 addition is soot oxidation. Ru-based catalysts have high activity in the soot
combustion and can promote soot oxidation significantly. As is proven in an earlier study
(Over and Seitsonen 2002), the Ru–O bond is weak so that the dissociative adsorption of the
active oxygen on the Ru site is easy to occur. Thus, the oxidation of organic species on the
soot surface are facilitated in the Ru(acac)3-doped flames, eventually leading to the formation
of the primary soot particles with rough surfaces.
In addition, the Porod exponent of the aggregates in both undoped and doped flames are
between 2 and 2.5 as shown in Figure 5, and there is no significant difference found between
different conditions. A exponent of three reflects a compact spherical particle, and a lower
value indicates an aggregate with more fluffy and looser structure (Di Stasio et al. 2006). It
means that in this work, the aggregates in all flames are formed loosely and behave as
Gaussian polymer chains, which is in agreement with Figure 2. These results indicate that
Ru(acac)3 affects the morphology of the primary particles significantly, but slightly influ­
ence the aggregates.
Effect of catalysts on size and volume of soot particles
To investigate the effects of Ru(acac)3 on the soot particle size, we obtain the radius of
gyration, RG of the primary particles by unified fitting. It has been elucidated that RG can
represent the real particle radius for poly-dispersed spheres (Beaucage, Kammler, Pratsinis
2004). The results of RG are set out in Figure 6, showing that RG grows as HAB increases,
and then hits a plateau under all conditions. It means that the soot particles grow quickly in
the low flame region and then tends to keep a steady size. In this figure, we can see that the
additions of 50, 100 and 150 ppm catalysts all slightly abate RG compared with the undoped
flame. As we know, surface growth is an important process for soot particle size growing,
and the particle size decreases during the oxidation process. Thus the primary soot particle
size mainly relies on the net rate of surface reactions, i.e., competition between the oxidation
and growth on the particle surface (Frenklach 2002). From Figure 5, we found that
Ru(acac)3 reduced the rate of the surface growth. Besides, it has been confirmed that the
oxidative activity of the primary particles can be enhanced with the addition of Ru catalysts.
Thus, the primary soot particles can be oxidized into much smaller ones after the addition
of Ru(acac)3, resulting in smaller soot particles in the doped fames compared with the
undoped flame at the same HABs.
The size distribution and the mean diameter of the primary soot particles in the undoped
and doped flames were obtained by the TEM statistical analysis, and the results at different
HABs are shown in Figure 7 and Figure 8, respectively. Compared with the primary soot
particles in the undoped flame, the peak frequency of the diameter distribution shifts
10
F. ZHANG ET AL.
Figure 6. The radius of gyration with uncertainties for the primary particles as a function of HAB in the
undoped flame, 50 ppm, 100 ppm and 150 ppm Ru(acac)3-doped flame, respectively.
Figure 7. The primary particle size distributions and mean diameters of the soot particles in the undoped
flame, 50 ppm, 100 ppm and 150 ppm Ru(acac)3-doped flame at 30 mm HAB.
toward lower diameter in the doped flames. This change may come from the oxidation of
the large particles. That is to say, the large particles are prone to be oxidized into smaller
particles, which leads to the increase of the number proportion of the small particles and
decrease of the number proportion of the larger particles, thus making the distribution
curves concentrated around small diameter values. Besides, the average diameters of the
primary particles also decrease in the doped flames, which is in accord with our observa­
tions for RG from SAXS results in Figure 6.
According to the above discussion, Ru(acac)3 catalysts affect the primary soot particles
markedly in the process of the surface growth and oxidation. Figure 9 shows the proposed
reaction mechanism of the soot combustion catalyzed by Ru(acac)3. During the thermal
decomposition of Ru(acac)3 in the flames, Ru metal particles were generated initially and
COMBUSTION SCIENCE AND TECHNOLOGY
11
Figure 8. The particle size distributions and mean diameters of the primary soot particles in the undoped
flame, 50 ppm, 100 ppm and 150 ppm Ru(acac)3-doped flame at 40 mm HAB.
Figure 9. Predicted reaction mechanism of the soot combustion catalyzed by Ru(acac)3.
then they were oxidized into RuO2 when the flame temperature reached a certain value
according to the previous study (Musić et al. 2002). Villani et al. (2006), Tschamber et al.
(2007) and Jeguirim et al. (2010) demonstrated that Ru surface was a generator of active
oxygen species even at low temperatures. Chemisorption of oxygen on the sites of RuO2
metal surfaces produces weakly bonded atomic oxygen with an extremely high chemical
reactivity. The active oxygen radicals could migrate to the carbon surface, leading to a high
concentration of surface oxygenated carbon complexes and making the carbon more
susceptible to be oxidized. In addition, PAHs play a vital role in the soot surface growth.
Metal catalysts can inhibit the surface growth rate of the soot particles and then make the
soot surfaces rougher by decreasing the amount of PAHs. Based on these effects of
Ru(acac)3 on the process of surface growth and oxidation, the primary soot particles can
grow into smaller particles with rougher surface in the doped flames.
12
F. ZHANG ET AL.
The volume distribution of the soot particles at several selected HABs are obtained by
using TNNLS fitting method and shown in Figure 10. It is suggested that the peak of the
soot volume distribution shifts from 18 nm diameter to 22 nm diameter as HAB increases,
indicating the particle size grows larger downstream along the flame centerline. Besides, the
fluctuation shown at higher diameter and slight hoist shown at lower diameter values are
probably artifacts of the fitting procedure, and we can ignore these spikes. Another feature
Figure 10. The volume distributions of the soot particles by TNNLS at 30, 35 and 40 mm HABs in the
undoped flame, 50 ppm, 100 ppm, and 150 ppm Ru(acac)3-doped flame, respectively.
COMBUSTION SCIENCE AND TECHNOLOGY
13
in these figures is that the volume distributions of the soot particles in the Ru(acac)3-doped
flames are much lower than those in the undoped flame at almost each same HAB.
However, there is not a clear trend between different concentration of the catalysts.
Shown in Figure 11 are the volume fractions of the soot particles obtained by Eq. (8)
and Eq. (9) by assuming the particles are homogeneous. Thus, the uncertainties of
scattered intensity will pass to the calculated volume fractions of the soot particles, and
the uncertainties of the volume fractions are also presented in the figure. As can be seen,
the volume fraction increases with increasing HAB in both undoped and doped flames.
The volume fraction is, respectively, reduced after the addition of 50, 100, or 150 ppm
Ru(acac)3, which means that Ru(acac)3 suppresses the soot volume. The previous studies
found that Ru is an effective catalyst or catalyst promoter for soot oxidation (Castoldi
et al. 2017; Villani et al. 2006). In our work, the Porod exponent of the primary particles
apparently decrease after adding Ru(acac)3 catalyst, meaning a suppression effect in the
surface growth rate of the primary soot particles. Therefore, the size of the primary
particles decreases due to the effects of rapid oxidation and slow surface growth. All of
these effects could be the reasons to lead to a descent in the volume fraction. However, it
should be noted that the scattering length density of carbon was used as the contrast
scattering length in the data analysis, so the actual contrast length density should be
a little smaller than that of carbon, and for each soot particle, the contrast scattering
length could be different. Therefore, the actual volume fractions of the soot should be
slightly smaller than the results shown in Figure 11.
Together, these results provide some evidence on how the Ru(acac)3 catalysts affect
the soot particles. Prior study (Frenklach 2002) noted that the soot mass accumulated
is determined primarily by surface reactions, i.e., growth and oxidation. While the
surface growth can lead to an increase in the particle size and the soot volume
Figure 11. The volume fractions and their uncertainties of the soot particles generated in the undoped
flame, 50 ppm, 100 ppm, and 150 ppm Ru(acac)3-doped flame, respectively.
14
F. ZHANG ET AL.
fraction, soot oxidation plays the opposite role. In this study, Ru and RuO2 nanopar­
ticles produced by the thermal decomposition of Ru(acac)3 promote the oxidation
reactions between oxygen and soot. Molecular oxygen O2 can be easily adsorbed on
the surface of metal nanoparticles and turn into various active forms, and thus
participate in the soot oxidation reactions. Besides, Ru(acac)3 can reduce the amount
of the surface growth precursors, e.g., PAHs. And then inhibit the surface growth for
the primary soot particles. Overall, these results indicate that the primary soot particles
grow into smaller particles with rougher surfaces in the doped flames than those in the
undoped flame due to the combined effects of the surface growth and oxidation.
However, no clear trend is observed in the soot particle size and volume among the
three flame groups doped with 50, 100, or 150 ppm Ru(acac)3. A further study on
whether more intensive concentration intervals can make a difference will be
performed.
In addition, we should mention that this work only provides a fundamental investigation
on soot emission under catalytic combustion. Application of this method on soot reduction
in real combustors will require further studies and simulations possibly involving design of
specific catalytic chamber. Moreover, the metal particles should not be released into the
atmosphere maybe by some treatment in the exhaust pipe, which will be a different subject
needed for further investigations.
4. Conclusions
The present study investigated the impacts of Ru(acac)3 catalyst on the soot particles in
the growth and oxidation process. In-situ SAXS experiment was performed to get the
information of the morphology, size and volume fraction of the soot particles. Besides,
TEM was used to support our results by direct observation of the soot morphology and
statistical analysis.
Based on the foregoing results and discussion, it was found that the catalysts can suppress
the surface growth for the primary soot particles, leading to rougher surfaces, while they rarely
affect the aggregates. Thus, the primary particle size significantly decreases due to the slow
surface growth and the rapid oxidation effects in the doped flames. The maximum frequency
of the diameter distribution shifts to a lower diameter after the addition of Ru(acac)3.
Compared with the undoped flame, the volume fraction of the soot particles is reduced in the
Ru(acac)3-doped flames. As a result, Ru(acac)3 inhibits soot emission effectively. However, no
clear trend can be obtained from the concentration change of Ru(acac)3. Wider concentration
range of the catalyst may provide more information. Further research will be conducted to
explore the effect from the composition change of the catalyst and detect the intermediate
species produced in the combustion process to establish a detailed mechanism.
Acknowledgment
The authors gratefully acknowledge the National Natural Science Foundation of China (the grant
number U2032119, 91641125 and 71690245) for their financial support. We also thank for the
“IRENA” micro package provided by Dr. Jan Ilavsky for data fitting. Besides, we would like to
acknowledge Yuchen Zhu, a PhD student in University of Science and Technology Beijing, for his
advice in TEM analysis. This research was performed at the APS, a U.S. Department of Energy (DOE)
Office of Science User Facility under Contract No. DE-AC02-06CH11357.
COMBUSTION SCIENCE AND TECHNOLOGY
15
Disclosure statement
The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.
Funding
This work was supported by the National Natural Science Foundation of China [71690245,91641125,
U2032119].
ORCID
Fanggang Zhang
http://orcid.org/0000-0001-9288-9596
Juan Wang
http://orcid.org/0000-0002-8556-044X
References
Aouad, S., E. Saab, E. A. Aad, and A. Aboukaïs. 2007. Reactivity of Ru-based catalysts in the oxidation
of propene and carbon black. Catal. Today 119: pp. 273–277.
Beaucage, G., and D. Schaefer. 1994. Structural studies of complex systems using small-angle
scattering: A unified Guinier/power-law approach. J. Non-Cryst. Solids 172: pp. 797–805.
Beaucage, G., H. K. Kammler, and S. E. Pratsinis. 2004. Particle size distributions from small-angle
scattering using global scattering functions. J. Appl. Crystallogr 37: pp. 523–535.
Beaucage, G. 1995. Approximations leading to a unified exponential/power-law approach to
small-angle scattering. J. Appl. Crystallogr 28: pp. 717–728.
Braun, A., F. E. Huggins, S. Seifert, J. Ilavsky, N. Shah, K. E. Kelly, Sarofim, A., and Huffman, G. P.
2004. Size-range analysis of diesel soot with ultra-small angle X-ray scattering. Combust. Flame
137: pp. 63–72.
Castoldi, L., E. Aneggi, R. Matarrese, R. Bonzi, A. Trovarelli, and L. Lietti. 2017. Simultaneous
removal of soot and NOx over silver and ruthenium-based catalysts. Top. Catal 60: pp.
209–213.
Di Stasio, S., J. Mitchell, J.-L. LeGarrec, L. Biennier, and M. Wulff. 2006. Synchrotron SAXS< in situ>
identification of three different size modes for soot nanoparticles in a diffusion flame. Carbon 44:
pp. 1267–1279.
Di Stasio, S. 2017. Soot with 1013 cm−3 high concentration and 25 Å radius of gyration as detected by
small-angle X-ray scattering in a premixed ethylene-air flame at sooting threshold. J. Aerosol Sci
110: pp. 11–24.
Frenklach, M. 2002. Reaction mechanism of soot formation in flames. Phys. Chem. Chem. Phys 4: pp.
2028–2037.
Gao, Y., X. Wu, S. Liu, D. Weng, and R. Ran. 2018. MnOx-CeO2 mixed oxides for diesel soot
oxidation: A review. Catal. Surv. Asia 22: pp. 230–240.
Guaita, R., M. Pichiule, T. Maté, C. Linares, and J. Díaz. 2011. Short-term impact of particulate matter
(PM2.5) on respiratory mortality in Madrid. Int. J. Environ. Health Res 21: pp. 260–274.
Guinier, A., G. Fournet, C. Walker, and K. Yudowitch. 1955. Small-angle scattering of X-rays.
New York: Jahn Willey-Champan.
Howard, J. B., and J. W. J. Kausch. 1980. Soot control by fuel additives. Prog. Energy Combust. Sci 6:
pp. 263–276.
Hu, C., W. Li, Q. Lin, X. Cheng, Q. Huang, H. Zhang, and Wang, Z. Y. 2017. Effects of ferrocene on
flame temperature, formation of soot particles and growth of polycyclic aromatic hydrocarbons.
J. Energy. Inst 90: pp. 893–901.
Ilavsky, J., and P. R. Jemian. 2009. Irena: Tool suite for modeling and analysis of small-angle
scattering. J. Appl. Crystallogr 42: pp. 347–353.
16
F. ZHANG ET AL.
Jeguirim, M., K. Villani, J. F. Brilhac, and J. A. Martens. 2010. Ruthenium and platinum catalyzed
carbon oxidation: A comparative kinetic study. Appl. Catal. B-Environ 96: pp. 34–40.
Kim, K., and D. W. Hahn. 2016. Interaction between iron based compound and soot particles in
diffusion flame. Energy 116: pp. 933–941.
Ledwa, K. A., M. Pawlyta, and L. Kepinski. 2018. RuxCe1-xO2-y nanoparticles deposited on functio­
nalized γ-Al2O3 as a thermally stable oxidation catalyst. Appl. Catal. B-Environ 230: pp. 135–144.
Lee, C., Y. Jeon, T. Kim, A. Tou, J.-I. Park, H. Einaga, and Y.-G. Shul. 2018. Ag-loaded
cerium-zirconium solid solution oxide nano-fibrous webs and their catalytic activity for soot and
CO oxidation. Fuel 212: pp. 395–404.
Li, Q., X. Wang, H. Chen, Y. Xin, G. Tian, C. Lu, Zhang, Z. L., Zheng, L. R., and Zheng, L. 2016.
K-supported catalysts for diesel soot combustion: Making a balance between activity and stability.
Catal. Today 264: pp. 171–179.
Li, Z., M. Meng, F. Dai, T. Hu, Y. Xie, and J. Zhang. 2012. Performance of K and Ni substituted La1-x
KxCo1-yNiyO3-delta perovskite catalysts used for soot combustion, NOx storage and simultaneous
NOx-soot removal. Fuel 93: pp. 606–610.
Marsh, N. D., I. Preciado, E. G. Eddings, A. F. Sarofim, A. B. Palotas, and J. D. Robertson. 2007.
Evaluation of organometallic fuel additives for soot suppression. Combust. Sci. Technol 179: pp.
987–1001.
Merritt, M., and Y. Zhang. 2005. Interior-point gradient method for large-scale totally nonnegative
least squares problems. J. Optim. Theory. Appl 126: pp. 191–202.
Mitchell, J., S. Di Stasio, J.-L. LeGarrec, A. I. Florescu-Mitchell, T. Narayanan, and M. Sztucki. 2009.
Small angle x-ray scattering study of flame soot nanoparticle aggregation and restructuring. J. Appl.
Phys 105 p. 124904-1-9.
Musić, S., S. Popović, M. Maljković, K. Furić, and A. Gajović. 2002. Formation of RuO2 and Ru by
thermal decomposition of ruthenium(III)-acetylacetonate. J. Mater. Sci. Lett 21: pp. 1131–1134.
Nascimento, L. F., R. F. Martins, R. F. Silva, P. C. de Sousa, and O. A. Serra. 2014. Ru-doped
ceria-zirconia mixed oxides catalyze soot combustion. React. Kinet Mech. Catal 111: pp.
149–165.
Neeft, J. P. A., M. Makkee, and J. A. Moulijn. 1996. Metal oxides as catalysts for the oxidation of soot.
Chem. Eng. J 64: pp. 295–302.
Oi-Uchisawa, J., A. Obuchi, R. Enomoto, S. Liu, T. Nanba, and S. Kushiyama. 2000. Catalytic
performance of Pt supported on various metal oxides in the oxidation of carbon black. Appl.
Catal. B-Environ 26: pp. 17–24.
Over, H., and A. P. Seitsonen. 2002. Oxidation of metal surfaces. Science 297: pp. 2003–2005.
Porod, G., , and . 1982. Small-angle X-ray scattering. , Vol. 61 eds. Glatter, O., and Kratky, O.
(London, New York: Academic Press), pp. 18–22.
Ren, W., T. Ding, Y. Yang, L. Xing, , , Cheng, Q., Zhao, D., Zhang, Z., Li, Q., Zhang, J., Zheng, L.,
Jiang, Z., and Li, X. 2019. Identifying oxygen activation/oxidation sites for efficient soot combus­
tion over silver catalysts interacted with nanoflower-like hydrotalcite-derived CoAlO eetal oxides.
ACS Catal. 9: pp. 8772–8784.
Ritrievi, K., J. Longwell, and A. Sarofim. 1987. The effects of ferrocene addition on soot particle
inception and growth in premixed ethylene flames. Combust. Flame 70: pp. 17–31.
Schmidt, P. W. 2013. Modern aspects of small-angle scattering. , Vol. 451 ed. Brumberger, H. Series
C: Mathematical and Physical Sciences (Como, Italy: Springer Science & Business Media) , pp. 2–9.
Tang, X., C. Wang, F. Zhang, Q. Wang, J. Wang, S. Seifert, and Winans, R. E. 2019. Effect of nickel
acetylacetonate addition on soot inception and growth in an ethylene flame studied by using in situ
small-angle X-ray scattering. Combust. Flame 206: pp. 390–399.
Tollefson, J. , 2013. Soot a major contributor to climate change, https://www.nature.com/news/soota-major-contributor-to-climate-change-1.12225 Accessed15 Jan 2013
Tschamber, V., M. Jeguirim, K. Villani, J. Martens, and P. Ehrburger. 2007. Comparison of the
activity of Ru and Pt catalysts for the oxidation of carbon by NO2. Appl. Catal. B-Environ 72: pp.
299–303.
Villani, K., C. E. A. Kirschhock, D. Liang, G. Van Tendeloo, and J. A. Martens. 2006. Catalytic carbon
oxidation over ruthenium-based catalysts. Angew. Chem-Int. Edit 45: pp. 3106–3109.
COMBUSTION SCIENCE AND TECHNOLOGY
17
Wang, J., S. Seifert, R. E. Winans, E. Tolmachoff, Y. Xin, D. Chen, Wang, H., and Anderson, S. L.
2015. In situ X-ray scattering and dynamical modeling of Pd catalyst nanoparticles formed in
flames. J. Phys. Chem. C 119: pp. 19073–19082.
Wei, Y.-L., and J.-H. Lee. 1999. Manganese sulfate effect on PAH formation from polystyrene
pyrolysis. Sci. Total Environ 228: pp. 59–66.
Yashnik, S. A., and Z. R. Ismagilov. 2019. Pt-Pd/MnOx-Al2O3 oxidation catalysts: Prospects of
application for control of the soot emission with diesel exhaust gases. Kinet. Catal 60: pp. 453–464.
Yon, J., F.-X. Ouf, D. Hebert, J. B. Mitchell, N. Teuscher, J.-L. Le Garrec, Bescond, A., Baumann, W.,
Ourdani, D., Bizien, T., Perez, J. 2018. Investigation of soot oxidation by coupling LII, SAXS and
scattering measurements. Combust. Flame 190: pp. 441–453.