- Email: [email protected]
The role of CO2 dilution on soot formation and combustion characteristics in counter-flow diffusion flames of ethylene
Experimental Thermal and Fluid Science 114 (2020) 110061
Contents lists available at ScienceDirect
Experimental Thermal and Fluid Science journal homepage: www.elsevier.com/locate/etfs
The role of CO2 dilution on soot formation and combustion characteristics in counter-flow diffusion flames of ethylene M. Sirignano, A. D'Anna
T
⁎
Dipartimento Ingegneria Chimica, dei Materiali e della Produzione Industriale – Università degli Studi di Napoli Federico II, Piazzale Tecchio 80, 80125 Napoli, Italy
A R T I C LE I N FO
A B S T R A C T
Keywords: Soot Carbon nanoparticles CO2 dilution Fluorescence Counterflow flames
The effect on soot formation of carbon dioxide addition as diluent in ethylene/O2/Ar laminar counter flow diffusion flames has been examined both experimentally and with the help of a multi-sectional kinetic model. Different concentrations of CO2 have been used in the oxidizer stream to replace argon, whereas the amount of oxygen has been kept constant. Optical techniques have been adopted to measure particulate matter: laserinduced fluorescence (LIF) to detect carbon nanoparticles and laser-induced incandescence (LII) to detect soot. Significant reductions of both soot and carbon nanoparticles have been observed with the addition of CO2: by increasing the amount of CO2, the reduction of particulate matter increases. In the case of 100% CO2 as diluent, soot particles are completely depleted, whereas carbon nanoparticles are still formed but at a lower rate. Therefore, the net result of CO2 addition is the reduced emission of total particulate matter enriched in carbon nanoparticles. On the basis of the results of the kinetic model, particle reduction has been mostly attributed to a thermal effect due to the reduced adiabatic flame temperatures at increasing amounts of CO2 added and only partially to the chemical effect of CO2 in enhancing OH radical formation.
1. Introduction
formation was inhibited by CO2 addition in either the fuel and the oxidized streams. The inhibition was shown to be through a combination of dilution, thermal, and a chemical effects. Du et al. [3], Liu et al. [4], Oh et al. [5] and Gülder and Baksh [6] through experimental and numerical works in laminar diffusion flames refer the chemical reactivity of CO2 to increased production of O and OH radicals that are believed to be responsible for soot precursor oxidation. Guo and Smallwood [7] linked the chemical effect of CO2 addition to the concentration of H radical suggesting that CO2 reduces radical concentration inhibiting surface growth of soot particles but has no effect on soot oxidation. Abian et al. [8,9] performed experiments on ethylene pyrolysis in a flow reactor and found a different effect, depending on the amount of CO2 dilution: for lower CO2 concentrations, soot formation was enhanced, whereas soot was reduced at higher levels of CO2. The effect of CO2 on pollutants formation was considered to be purely a chemical effect. Finally, some differences of CO2 effect on soot formation were found in dependence on the type of fuel used. Ethylene has been largely used to study CO2 effect and generally a large decrease in soot formation has been observed. However, in CH4 diffusion flames a lighter effect has been found [10]. Recently Wu et al. [11] studied CO2 addition to the fuel on soot evolution in ethylene and propane coflowing diffusion
The most important concern in the use of fossil fuels in combustion processes is the high production of carbon dioxide that contributes to greenhouse effect. CO2 emissions could be reduced improving combustion efficiency or by its separation from the combustion exhausts and sequestration. In the last years, to achieve CO2 emission reduction, pre-combustion, post-combustion and oxy-combustion processes are under development in order to produce as output stream only water and carbon dioxide, thus favoring its capture and sequestration [1]. Oxy-combustion consists in using oxygen instead of air in combustion reactions, potentially with extreme conditions in terms of temperatures. The process is often realized by recycling oxygen from the combustion exhaust, having H2O and CO2 as diluents in order to mitigate the temperature and making the process compatible with existing infrastructures. CO2 diluted environments might influence the combustion process, particularly the formation of pollutants, products of incomplete combustion. A number of investigations have been focused on the study of soot formation by introducing diluents in the oxidizer or fuel sides of flames [2–7]. Shung et al. [2] found that the effect of carbon dioxide is predominantly thermal, due to a decrease in the flame temperature, with no appreciable chemical influence. Du et al. [3] found that soot ⁎
Corresponding author. E-mail addresses: [email protected] (M. Sirignano), [email protected] (A. D'Anna).
https://doi.org/10.1016/j.expthermflusci.2020.110061 Received 2 December 2019; Received in revised form 9 January 2020; Accepted 24 January 2020 Available online 25 January 2020 0894-1777/ © 2020 Elsevier Inc. All rights reserved.
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
flames and found that CO2 addition caused a slight decrease of the flame temperature, a suppression of both soot and PAH and more interestingly, that addition of CO2 inhibited the conversion of PAHs to soot. Zhang et al. [12] found a suppression of soot by CO2 addition in counterflow flame of methane and more recently Wang and Chung [13] extended the study in counterflow ethylene flames, confirming inhibitive effect of CO2 addition on soot formation and, based also on a numerical investigation, attributing this effect to a chemical effect more than to a thermal effect, although this one was not excluded. Therefore, an overall agreement in the scientific community is not reached yet. Most of the studies found in literature on this topic have been focused mainly on soot particles and no definitive studies have been performed on carbon nanoparticles, precursors to soot [14]. The aim of this work is to give further insights on the effect of different amount of CO2 used as diluent on the formation of particulate matter particularly on carbon nanoparticles. The study has been performed in a counter-flow ethylene diffusion flame, which approximates combustion occurring at the flame front of practical turbulent flames. Optical techniques are used to characterize carbon particles produced in the flame to understand the influence of the different inert on particulate production. Laser-induced fluorescence (LIF) and incandescence (LII) measurements have been used to measure concentration profiles in flames of carbon nanoparticle and soot, respectively.
the fourth harmonic radiation (266 nm) of a Nd:YAG laser as excitation source. Laser beam diameter at focal point of 350 μm- was focused in the center of the flame. The emitted radiation at 90° with respect to the laser beam was focused onto the 280 μm entrance slit of a spectrometer and detected by an intensified CCD camera thermoelectrically cooled down to −10 °C to reduce noise. The energy of the laser pulse was kept constant at 0.8 mJ with pulse duration of 8 ns. The chosen laser energy gave the better compromise between laser induced emission signals and species fragmentation interference. The measured spectra were corrected for the spectral response of the detection system. The emission spectra were detected using a gate time of 100 ns synchronized with the laser pulse and by summing the CCD counts over 150 scans. In this way both LIF and LII signals were contemporarily detected. This technique has been validated over time and it has been used in counterflow flames with different fuels [15–18]. The spectrum shows the scattering signal at 266 nm, a broad LIF emission in the region between 290 and 500 nm with two distinct maxima in the spectral regions 300–350 and 400–500 nm, and a continuum due to the incandescence emission of solid particles which maximizes at longer wavelengths. The continuum in the visible was fitted by black-body radiation at 4000 K, according to literature data, matching the measured emission intensity values at 550 nm. Although particle temperature is not known, for wavelengths smaller than 500 nm the black-body spectrum slightly changes with particle temperature, little affecting only the 440 nm LIF signal. Therefore, the uncertainty on the 440 nm band is of about 5% whether 4000 K or 3500 K are used. Fig. 2 reports emission spectra at different heights from the lower nozzle: the 4 mm spectrum corresponds to a flame region below the stagnation plane with temperature less to 1000 K and high fuel concentration; 5 mm is approximately the position of the stagnation plane whereas at 6 mm high temperatures occur close to the front flame. According to previous studies [16,21], LIF signals are attributed to aromatic hydrocarbons in condensed-phase nanostructures which are not able to incandesce [14,21 and reference therein]. We collected the entire spectrum and thus the swan band interference was accurately avoided. The UV LIF is assigned to condensed-phase nanostructures(up to 4 nm in size) mostly constituted by aromatic compounds with 2–3 condensed rings and generally a loose structure; visible-LIF is attributed condensed-phase nanostructures constitued by PAH with 4–5 condensed rings and a well-organized structure. LII signal is assigned to soot particles of 10–20 nm in size and larger aggregates [21].
2. Experimental procedure The counter-flow burner used in this study was the same as that used for previous works [15–18]. The burner consists of two opposite jet nozzles (ID 2.54 cm) vertically positioned. The oxidizer stream (O2/ diluent) was introduced from the upper nozzle, while the fuel stream (C2H4/Ar) from the bottom. Different flames were analyzed varying the concentration of CO2 in the oxidizer side. The fuel stream was fixed at 25% vol. C2H4 and 75% vol. Ar with an axial velocity of 13.2 cm/s. In the oxidizer stream, the percentage of O2 was fixed at 22% vol. and the axial velocity at 16.1 cm/s. In a first configuration (standard case) the 78% vol. of the oxidizer side was constituted totally by Argon then the 25%, 50% and 78% was substituted with CO2 and the remaining Argon. The distance between the two burners was maintained at 1.5 cm for all flames. With these flux velocities the stagnation plane is located around 5 mm from the lower nozzle. In all the configurations, the flame front was stabilized in the oxidizer plane. Measurements were performed at different heights along the centerline. The experimental setup is the same as that used in previous works and is sketched in Fig. 1 [15]. Some details are reported below. Measurements within the flame were performed at different locations between the opposed jet nozzles, from the fuel to the oxidizer. The measuring location was changed by moving the entire burner assembly up or down with respect to the sampling point using a translation system, with a spatial resolution of 0.1 mm. Temperature was measured by using a silica-coated Pt/Pt–13%Rh thermocouple with a large bead diameter of about 300 µm. The relatively large thermocouple bead was necessary to avoid thermocouple melting in the high temperature zones of the flame. A lower-size thermocouple bead was used in the fuel zone of the flame (temperatures below 1000 K) and gave approximately the same results after radiation correction. A rapid insertion procedure was applied to measure temperature in order to reduce soot deposition on the thermocouple bead [19]. To minimize the heat conduction along the wires, the thermocouple was positioned parallel to the nozzle surfaces. Particulate deposited on the thermocouple was burnt off by using a methane torch. Temperature measurements were corrected for gas radiations [20]. Temperature data were highly repeatable; measurement uncertainty remains within 50 K. Laser Induced Emission (LIE) measurements were performed using
3. Kinetic model The mechanism of gas phase oxidation and pyrolysis presented here is the same that has been tested for ethylene in different combustion conditions including premixed [22–24] coflow [25] and counterflow diffusion flames [26]. The mechanisms includes the principal routes for small hydrocarbon oxidation and pyrolysis and the main pathways to form benzene ring. Benzene formation pathways includes both the C4 routes with the C2H2 addition and the self-combination of propargyl radicals (C3route). For PAH formation the HACA mechanism [27–28] is considered together with the combination of resonantly-stabilized radicals. This latter includes the combination of two cyclopentadienyl radicals and the combination of benzyl and propargyl radicals [29] for the formation of naphthalene. PAH growth is followed punctually up to pyrene which is here considered the largest gas phase compound. The gas-phase kinetic mechanism consists of 460 reactions involving 120 species. Details of the reaction rates used are reported in refs [22,23]. Species with molecular weight MW > 202u (i.e., pyrene) were treated as lumped species, i.e. particle phase, and thus divided into classes. Using lumped species allows us to evaluate the collision frequency of reactions based on gas-kinetic theory. The activation energy for the reactions involving lumped species was assigned based on 2
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
Fig. 1. Lay out of the experimental system. 1.0E+6
pathways reported above. These large molecules are treated as lumped species but cannot be considered particles or particle nuclei. Inception of particles - i.e. the formation of the first particle nucleus or lumped species that has a condensed phase behavior - is supposed to occur as a coagulation event between two molecules. Molecule coagulation was considered irreversible and its reaction rate was modeled by considering a coagulation efficiency with respect to the collision frequency. The collision frequency increases with the increase of molecular mass of the molecules whereas the coagulation efficiency, γ, depends on both local flame temperature and chemistry of the colliding molecules. The chemistry of the particles was considered by evaluating the Hamaker constant [30] for the species involved in the coagulation process. The computed coagulation efficiency at 1500 K is of the order of 1E-4 for small colliding entities (the first lumped species) and increases to values of about 1 when the C number is about 1E6. These values are in agreement with experimental coagulation rates. Coagulation efficiency also increases as temperature decreases due to lower thermal rebound involved in coagulation process. Particle nuclei continue to react in the same way as the molecules; they can add molecules to increase their size, or remove H-atoms by dehydrogenation or C-atoms by OH and O2 oxidation, or they can coagulate with other molecules or other clusters.
4 mm 5 mm
LIE A.U.
8.0E+5
6 mm
6.0E+5
4.0E+5
2.0E+5
0.0E+0 250
300
350
400 450 Wavelenght [nm]
500
550
Fig. 2. Laser induced emission spectra for the flame C2H4/Ar/O2 collected at 4 mm (circles), 5 mm (squares) and 6 mm (triangles) from the fuel nozzle.
analogues reactions in gas-phase or, if not possible as for dehydrogenation [23] or fragmentation [25], chosen to fit the experimental data. Thirty-one lumped species or sections were used in a geometric series of carbon number with a ratio of two between sections; five sections were used for H/C variation; radical and neutral molecules were separately grouped. The equivalent size range of 1–800 nm was obtained with this discretization considering a density varying from 1 g/cm3 for the smallest species with the highest hydrogen content to 1.8 g/cm3 for species above 10 nm with the lowest hydrogen content. The molecular growth of aromatics starts with the formation of an aromatic radical by an H atom loss from aromatic molecules. Radical formation can occur via H or OH attack or unimolecular decomposition. Termination reactions of aromatic radicals end the growth sequence. The formation of PAH was modeled by addition of acetylene, following the HACA sequence. Alternatively, a PAH from gas phase can be added through chemical addition. The molecular growth process competed with oxidation of stable molecules and radicals by OH and O2, respectively. Details of the reaction rates used in the growth process of aromatics are reported elsewhere [25]. Large molecules can indeed be formed by chemical reaction between two gas phase PAH and can keep growing following the chemical
4. Results and discussion The flame of the base case (undiluted conditions) has been stabilized by using argon as diluent for the oxidizer stream (Ar 78%mol) and for the ethylene. Fig. 3 shows the LIF and the LII signal profiles along the distance between opposed jets. Three regions can be distinguished along the flame: fuel side, stagnation plane and oxidizer region. The flame front is located in the oxidizer side in which PAH growth is due to chemical pathways. Rapid growth due the high temperature is confirmed by the fast increase of both LIF and LII signal. The stagnation plane is located around 4.8 mm from the lower nozzle; in this region the signals maximize due to both contributions from oxidizer and fuel side. In the fuel side, the LIF and LII signals can be justified with pyrolysis reactions of the fuel and a physical mechanism of PAH growth (van der Waals interactions). Two broad peaks are distinguished in the emission spectra reported in Fig. 2: a broad one in the UV region at around 330 nm and a sharper one in the visible around 440 nm. In a previous study [21,32] we have attributed, on the basis of LIF emission wavelengths and LIF lifetimes, 3
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
8.0E+5
1.0E+6 LIF@330nm
LIF@440nm 0% CO2
LIF@440nm
8.0E+5
LIF@440nm 25% CO2
LII
6.0E+5
LIF@440nm 78% CO2
LIF A.U.
6.0E+5 LIE A.U.
LIF@440nm 50% CO2
4.0E+5
4.0E+5
2.0E+5
2.0E+5 0.0E+0
0.0E+0 2
2.5
3
3.5 4 4.5 5 5.5 Distance from fuel nozzle [mm]
6
6.5
2
7
LIF A.U.
7
6.5
LII A.U.
4.0E+5
0.0E+0 2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
7
Distance from fuel nozzle [mm]
Fig. 6. Laser induced incandescence signals as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer stream. Stagnation plane is reported as dashed line.
CO2 content increases, being still detectable when only CO2 is used as diluent. Fig. 6 shows LII signals with the addition of CO2. LII signals show the same behavior found for LIF signals. In fact, for small percentage of CO2 added, signal reduction, i.e., particle formation suppression, is more evident in the oxidizer side. As the CO2 percentage increases, the reduction of LII signal is significant thorough the flame. Finally when the 78% CO2 was added the LII signal is quite close the uncertainty of the experimental measurements, suggesting the formation of large soot particles is negligible. Fig. 7 shows the trends of ratio between laser light scattering signals and LII signals with the addition of CO2. This ratio is proportional to the third power of the mean particle size and it is not reported for the 78% CO2 flame because of the very low LII signal due to the absence of particles when 78% of CO2 is added. It is evident from scattering that increasing the amount of CO2 added, the average size of the particles decreases. The effect of CO2 additionis clearly evidenced by the percentage reduction of the different signals with respect to the 0% CO2 flame reported in Fig. 8. An almost total LII reduction is registered when 78% CO2 is added to the oxidizer stream while 70–80% LIF signal reduction is observed. The effect of CO2 addition on particle formation has been also analyzed with the help of numerical analysis. The model adopted here has been already tested and found able to reproduce the formation and evolution of nanoparticles and soot in the base case (argon dilution).
0.0E+0 6
LII 50% CO2
2.0E+5
2.0E+5
5.5
6.5
6.0E+5
4.0E+5
5
6
LII 78% CO2
LIF@330nm 78% CO2
4.5
5.5
LII 25% CO2
LIF@330nm 50% CO2
4
5
8.0E+5
LIF@330nm 25% CO2
3.5
4.5
LII 0% CO2
LIF@330nm 0% CO2
3
4
1.0E+6
8.0E+5
2.5
3.5
Fig. 5. Laser induced fluorescence signals at 440 nm as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer stream. Stagnation plane is reported as dashed line.
LIF signals to nanoparticle of organic carbon having different chemical structures and morphology. In particular, LIF in the UV was attributed to high-molecular mass cross-linked aromatic molecules with few condensed aromatic rings, whereas LIF in the visible was attributed to highly-packed, sandwich-like clusters of aromatic hydrocarbons [21,31]. These structures are considered in both cases to be in condensed phase. LII signal was attributed to solid soot particles which are able to dissipate the acquired energy by thermal emission rather than by LIF emission. In order to evaluate and subtract the contribution of the LII signal, a black-body radiation curve at 4000 K has been used. After assessing the soot formation in the base case flame, the effect of CO2 has been evaluated by substituting the Ar in the oxidizer stream. The amount of CO2 has been varied in the oxidizer stream in order to study the effect of CO2 on soot formation by performing LIF and LII measurements. Fig. 4 reports UV LIF signal profiles along the flame at varying CO2 concentration in the oxidizer stream in comparison with the flame C2H4/Ar/O2 (0% CO2 in the figure). LIF signal can be distinguished both in fuel and oxidizer side. Overall it can be noticed a strong reduction of the signals with respect to the flame C2H4/Ar/O2. LIF signal in the oxidizer side, i.e. close to the flame front, in the zone of higher temperature and larger particles formation is significantly reduced already at low percentage of CO2. Only when 78% CO2 is added, the LIF signal is strongly reduced also in the pyrolytic zone. Fig. 5 shows visible LIF signals with the addition of CO2. The LIF at 440 nm maximizes the signal in the pyrolytic zone of the flame, being related to nanostructure formed by large PAHs at low temperatures (T < 1500 K). The reduction of both UV and visible LIF increases as
2
3
Distance from fuel nozzle [mm]
Fig. 3. Laser induced emission signals as a function of the distance from the fuel nozzle for the flame C2H4/Ar/O2. Stagnation plane is reported as dashed line.
6.0E+5
2.5
7
Distance from fuel nozzle [mm]
Fig. 4. Laser induced fluorescence signals at 330 nm as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer stream. Stagnation plane is reported as dashed line. 4
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna 100
% reduction of mean particle volume
1.0E+02
QVV/LII
1.0E+01
1.0E+00
QVV/LII 0% CO2
90
80
70
60
50
modelling QVV/LII
QVV/LII 25% CO2 QVV/LII 50% CO2
40
0
1.0E-01
2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
LIF@330nm LIF@440nm LII
% Signal Reduction
60
40
20
0 40
60
80
100
% CO2
Fig. 8. Percentage reduction of laser induced emission signal as a function of the CO2 concentration in the oxidizer stream. 100 90 80
% soot reduction
40
50
60
70
80
90
100
respectively. The agreement of the model results with the experimental data for both soot (LII signal) and the mean particle sizes reduction can be considered satisfying. In particular, the model is able to reproduce the rapid disappearing of large particles as the CO2 percentage in the oxidizer stream is increased. The model analysis suggests that the effect of CO2 on soot formation is mainly a thermal effect; temperature is significantly reduced by moving from a monoatomic gas (Ar) to a polyatomic gas (CO2) as diluent. Temperature reduction is more significant closer to the flame front, i.e. in the oxidizer side, where the CO2 effect is observed also for small amount of CO2 added. Contrarily, to have a significant impact on temperature also in the pyrolytic side, the CO2 percentage has to be increased above 50%. The effect on large particles is more evident since their formation occurs mostly in the high temperature zone of the flame whereas small particles and PAH which are detected by LIF are also formed in the pyrolytic side of the flame and in general the effect on the formation of small particles seems to be less significant. The results presented here are generally in agreement with other results presented in previous works [2–7]. It is worth to note that in the investigated conditions a chemical effect of CO2 cannot be completely excluded; however, its role has to be considered of minor importance with respect to thermal effect especially when CO2 is used in high percentage as diluent.
100
20
30
Fig. 10. Percentage reduction, with respect to the 0% CO2 flame, of the measured (QVV/LII) and predicted particle volume as a function of the CO2 concentration in the oxidizer stream.
Fig. 7. Laser light scattering/LII ratio as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer). Stagnation plane is reported as dashed line.
0
20
CO2, %
Distance from fuel nozzle [mm]
80
10
7
70
5. Conclusions
60
Optical diagnostics were used to understand the effect of different amount of CO2 as diluent in a counter flow ethylene diffusion flame. Both LIF and LII were performed in order to detect both nanoparticles and large soot aggregates respectively. CO2 was added in the oxidizer stream up to 78% vol. of the total flux. No qualitative variations of LIF and LII trends were found along the centerline of the burner. The CO2 addition strongly reduces LIF and LII signals in the fuel side and in the oxidizer side. However, while LII, i.e. soot aggregates are completely depleted when only CO2 was used as diluent, LIF signal, i.e. nanoparticles, were still detected in the flame. Modelling analysis suggests that the effect of CO2 on soot formation is mainly a thermal effect in the investigated conditions.
50 40 30 20 modelling
10 0
LII
0
10
20
30
40
50
60
70
80
90
100
CO2, % Fig. 9. Percentage reduction, with respect to the 0% CO2 flame, of laser induced incandescence and predicted soot concentration as a function of the CO2 concentration in the oxidizer stream.
CRediT authorship contribution statement M. Sirignano: Methodology, Investigation, Writing - original draft. A. D’Anna: Supervision, Conceptualization, Methodology, Writing original draft, Writing - review & editing.
Figs. 9 and 10 show comparison of model results with experimental data reported as percentage reduction for soot and mean size
5
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
Declaration of Competing Interest [16]
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.
[17]
References [18] [1] T.F. Wall, Combustion processes for carbon capture, Proc. Combust. Inst. 31 (1) (2007) 31–47. [2] K.P. Schug, Y. Manheimer-Timnat, P. Yaccarino, I. Glassman, Sooting behaviour of gaseous hydrocarbon diffusion flames and the influence of additives, Combust. Sci. Technol. 22 (1980) 235–250. [3] D.X. Du, R.L. Axelbaum, C.K. Law, The influence of carbon dioxide and oxygen as additives on soot formation in diffusion flames, Proc. Combust. Inst. 23 (1990) 1501–1507. [4] F. Liu, H. Guo, G.J. Smallwood, Ö.L. Gülder, The chemical effects of carbon dioxide as an additive in an ethylene diffusion flame: implications for soot and NOx formation, Combust. Flame 125 (2001) 778–787. [5] K.C. Oh, H.D. Shin, The effect of oxygen and carbon dioxide concentration on soot formation in non-premixed flames, Fuel 85 (2006) 615–624. [6] Ö.L. Gülder, M.F. Baksh, Influence of carbon dioxide dilution on soot formation in diffusive combustion, Combust. Inst. Can. Sect., Quebec, QC (1993) pp. 40–1–40-5. [7] H. Guo, G.J. Smallwood, A numerical study on the influence of CO2 addition on soot formation in an ethylene/air diffusion flame, Combust. Sci. Technol. 180 (2008) 1695–1708. [8] M. Abián, A. Millera, R. Bilbao, M.U. Alzueta, Effect of recirculation gases on soot formed from ethylene pyrolysis, Combust. Sci. Technol. 184 (7–8) (2012) 980–994. [9] M. Abián, A. Millera, R. Bilbao, M.U. Alzueta, Experimental study on the effect of different CO2 concentrations on soot and gas products from ethylene thermal decomposition, Fuel 91 (1) (2012) 307–312. [10] M.R. Charest, Ö.L. Gülder, C.P. Groth, Numerical and experimental study of soot formation in laminar diffusion flames burning simulated biogas fuels at elevated pressures, Combust. Flame 161 (2014) 2678–2691. [11] J. Wu, L. Chen, P.E. Bengtsson, J. Zhou, J. Zhang, X. Wu, K. Cen, Effects of carbon dioxide addition to fuel on soot evolution in ethylene and propane diffusion flames, Combust. Flame 199 (2019) 85–95. [12] C. Zhang, A. Atreya, K. Lee, Sooting structure of methane counterflow diffusion flames with preheated reactants and dilution by products of combustion, Symp. Int. Combust. 24 (1) (1992) 1049–1057. [13] Y. Wang, S.H. Chung, Formation of soot in counterflow diffusion flames with carbon dioxide dilution, Combust. Sci. Technol. 188 (4–5) (2016) 805–817. [14] A. D’Anna, Combustion-formed nanoparticles, Proc. Combust. Inst. 32 (1) (2009) 593–613. [15] A. D’Anna, M. Commodo, M. Sirignano, P. Minutolo, R. Pagliara, Particle formation
[19] [20]
[21]
[22] [23]
[24] [25] [26] [27] [28] [29]
[30] [31]
[32]
6
in opposed-flow diffusion flames of ethylene: An experimental and numerical study, Proc. Combust. Inst. 32 (2009) 793–801. M. Sirignano, A. Collina, M. Commodo, P. Minutolo, A. D’Anna, Detection of aromatic hydrocarbons and incipient particles in an opposed-flow flame of ethylene by spectral and time-resolved laser induced emission spectroscopy, Combust. Flame 159 (2012) 1663–1669. M. Salamanca, M. Sirignano, M. Commodo, P. Minutolo, A. D’Anna, The effect of ethanol on the particle size distributions in ethylene premixed flames, Exp. Therm. Fluid Sci. 43 (2012) 71–75. M. Sirignano, M. Salamanca, A. D’Anna, The role of dimethyl ether as substituent to ethylene on particulate formation in premixed and counter-flow diffusion flames, Fuel 126 (2014) 256–262. J.H. Kent, H. Gg, Wagner, Why do diffusion flames emit smoke, Combust. Sci. Technol. 41 (5–6) (1984) 245–269. A. Rolando, C. Allouis, F. Beretta, A. D’Alessio, A. D’Anna, P. Minutolo, Measurement of particulate volume fraction in a co-flow diffusion flame using transient thermocouple technique, Combust. Sci. Technol. 176 (5–6) (2004) 945–958. M. Sirignano, D. Bartos, M. Conturso, M. Dunn, A. D’Anna, A.R. Masri, Detection of nanostructures and soot in laminar premixed flames, Combust. Flame 176 (2017) 299–308. A. D’Anna, Detailed kinetic modelling of particulate formation in laminar premixed flames of ethylene, Energy Fuels 22 (3) (2008) 1610–1619. M. Sirignano, J. Kent, A. D’Anna, Detailed modeling of size distribution functions and hydrogen content in combustion-formed particles, Combust. Flame 157 (6) (2010) 1211–1219. A. D’Anna, M. Sirignano, J. Kent, A model of particle nucleation in premixed ethylene flames, Combust. Flame 157 (11) (2010) 2106–2115. M. Sirignano, J. Kent, A. D’Anna, Modeling formation and oxidation of soot in nonpremixed flames, Energy Fuels 27 (4) (2013) 2303–2315. M. Sirignano, J. Kent, A. D’Anna, Further experimental and modelling evidences of soot fragmentation in flames, Proc. Combust. Inst. 35 (2) (2015) 1779–1786. M. Frenklach, H. Wang, Detailed modeling of soot particle nucleation and growth, Symp. (Int.) Combust. 23 (1991) 1559–1566. M. Frenklach, Reaction mechanism of soot formation in flames, Phys. Chem. Chem. Phys. 4 (11) (2002) 2028–2037. N.M. Marinov, W.J. Pitz, C.K. Westbrook, M.J. Castaldi, S.M. Senkan, Modeling of aromatic and polycyclic aromatic hydrocarbon formation in premixed methane and ethane flames, Combust. Sci. Technol. 116 (1–6) (1996) 211–287. H.C. Hamaker, The London-van der Waals attraction between spherical particles, Physica 4 (10) (1937) 1058–1072. A. D’Alessio, A.C. Barone, R. Cau, A. D’Anna, P. Minutolo, Surface deposition and coagulation efficiency of combustion generated nanoparticles in the size range from 1 nm to 10 nm, Proc. Combust. Inst. 30 (2) (2005) 2595–2603. D. Bartos, M. Dunn, M. Sirignano, A. D'Anna, A.R. Masri, Tracking the evolution of soot particles and precursors in turbulent flames using laser-induced emission, Proc. Combust. Inst. 36 (2) (2017) 1869–1876.
Contents lists available at ScienceDirect
Experimental Thermal and Fluid Science journal homepage: www.elsevier.com/locate/etfs
The role of CO2 dilution on soot formation and combustion characteristics in counter-flow diffusion flames of ethylene M. Sirignano, A. D'Anna
T
⁎
Dipartimento Ingegneria Chimica, dei Materiali e della Produzione Industriale – Università degli Studi di Napoli Federico II, Piazzale Tecchio 80, 80125 Napoli, Italy
A R T I C LE I N FO
A B S T R A C T
Keywords: Soot Carbon nanoparticles CO2 dilution Fluorescence Counterflow flames
The effect on soot formation of carbon dioxide addition as diluent in ethylene/O2/Ar laminar counter flow diffusion flames has been examined both experimentally and with the help of a multi-sectional kinetic model. Different concentrations of CO2 have been used in the oxidizer stream to replace argon, whereas the amount of oxygen has been kept constant. Optical techniques have been adopted to measure particulate matter: laserinduced fluorescence (LIF) to detect carbon nanoparticles and laser-induced incandescence (LII) to detect soot. Significant reductions of both soot and carbon nanoparticles have been observed with the addition of CO2: by increasing the amount of CO2, the reduction of particulate matter increases. In the case of 100% CO2 as diluent, soot particles are completely depleted, whereas carbon nanoparticles are still formed but at a lower rate. Therefore, the net result of CO2 addition is the reduced emission of total particulate matter enriched in carbon nanoparticles. On the basis of the results of the kinetic model, particle reduction has been mostly attributed to a thermal effect due to the reduced adiabatic flame temperatures at increasing amounts of CO2 added and only partially to the chemical effect of CO2 in enhancing OH radical formation.
1. Introduction
formation was inhibited by CO2 addition in either the fuel and the oxidized streams. The inhibition was shown to be through a combination of dilution, thermal, and a chemical effects. Du et al. [3], Liu et al. [4], Oh et al. [5] and Gülder and Baksh [6] through experimental and numerical works in laminar diffusion flames refer the chemical reactivity of CO2 to increased production of O and OH radicals that are believed to be responsible for soot precursor oxidation. Guo and Smallwood [7] linked the chemical effect of CO2 addition to the concentration of H radical suggesting that CO2 reduces radical concentration inhibiting surface growth of soot particles but has no effect on soot oxidation. Abian et al. [8,9] performed experiments on ethylene pyrolysis in a flow reactor and found a different effect, depending on the amount of CO2 dilution: for lower CO2 concentrations, soot formation was enhanced, whereas soot was reduced at higher levels of CO2. The effect of CO2 on pollutants formation was considered to be purely a chemical effect. Finally, some differences of CO2 effect on soot formation were found in dependence on the type of fuel used. Ethylene has been largely used to study CO2 effect and generally a large decrease in soot formation has been observed. However, in CH4 diffusion flames a lighter effect has been found [10]. Recently Wu et al. [11] studied CO2 addition to the fuel on soot evolution in ethylene and propane coflowing diffusion
The most important concern in the use of fossil fuels in combustion processes is the high production of carbon dioxide that contributes to greenhouse effect. CO2 emissions could be reduced improving combustion efficiency or by its separation from the combustion exhausts and sequestration. In the last years, to achieve CO2 emission reduction, pre-combustion, post-combustion and oxy-combustion processes are under development in order to produce as output stream only water and carbon dioxide, thus favoring its capture and sequestration [1]. Oxy-combustion consists in using oxygen instead of air in combustion reactions, potentially with extreme conditions in terms of temperatures. The process is often realized by recycling oxygen from the combustion exhaust, having H2O and CO2 as diluents in order to mitigate the temperature and making the process compatible with existing infrastructures. CO2 diluted environments might influence the combustion process, particularly the formation of pollutants, products of incomplete combustion. A number of investigations have been focused on the study of soot formation by introducing diluents in the oxidizer or fuel sides of flames [2–7]. Shung et al. [2] found that the effect of carbon dioxide is predominantly thermal, due to a decrease in the flame temperature, with no appreciable chemical influence. Du et al. [3] found that soot ⁎
Corresponding author. E-mail addresses: [email protected] (M. Sirignano), [email protected] (A. D'Anna).
https://doi.org/10.1016/j.expthermflusci.2020.110061 Received 2 December 2019; Received in revised form 9 January 2020; Accepted 24 January 2020 Available online 25 January 2020 0894-1777/ © 2020 Elsevier Inc. All rights reserved.
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
flames and found that CO2 addition caused a slight decrease of the flame temperature, a suppression of both soot and PAH and more interestingly, that addition of CO2 inhibited the conversion of PAHs to soot. Zhang et al. [12] found a suppression of soot by CO2 addition in counterflow flame of methane and more recently Wang and Chung [13] extended the study in counterflow ethylene flames, confirming inhibitive effect of CO2 addition on soot formation and, based also on a numerical investigation, attributing this effect to a chemical effect more than to a thermal effect, although this one was not excluded. Therefore, an overall agreement in the scientific community is not reached yet. Most of the studies found in literature on this topic have been focused mainly on soot particles and no definitive studies have been performed on carbon nanoparticles, precursors to soot [14]. The aim of this work is to give further insights on the effect of different amount of CO2 used as diluent on the formation of particulate matter particularly on carbon nanoparticles. The study has been performed in a counter-flow ethylene diffusion flame, which approximates combustion occurring at the flame front of practical turbulent flames. Optical techniques are used to characterize carbon particles produced in the flame to understand the influence of the different inert on particulate production. Laser-induced fluorescence (LIF) and incandescence (LII) measurements have been used to measure concentration profiles in flames of carbon nanoparticle and soot, respectively.
the fourth harmonic radiation (266 nm) of a Nd:YAG laser as excitation source. Laser beam diameter at focal point of 350 μm- was focused in the center of the flame. The emitted radiation at 90° with respect to the laser beam was focused onto the 280 μm entrance slit of a spectrometer and detected by an intensified CCD camera thermoelectrically cooled down to −10 °C to reduce noise. The energy of the laser pulse was kept constant at 0.8 mJ with pulse duration of 8 ns. The chosen laser energy gave the better compromise between laser induced emission signals and species fragmentation interference. The measured spectra were corrected for the spectral response of the detection system. The emission spectra were detected using a gate time of 100 ns synchronized with the laser pulse and by summing the CCD counts over 150 scans. In this way both LIF and LII signals were contemporarily detected. This technique has been validated over time and it has been used in counterflow flames with different fuels [15–18]. The spectrum shows the scattering signal at 266 nm, a broad LIF emission in the region between 290 and 500 nm with two distinct maxima in the spectral regions 300–350 and 400–500 nm, and a continuum due to the incandescence emission of solid particles which maximizes at longer wavelengths. The continuum in the visible was fitted by black-body radiation at 4000 K, according to literature data, matching the measured emission intensity values at 550 nm. Although particle temperature is not known, for wavelengths smaller than 500 nm the black-body spectrum slightly changes with particle temperature, little affecting only the 440 nm LIF signal. Therefore, the uncertainty on the 440 nm band is of about 5% whether 4000 K or 3500 K are used. Fig. 2 reports emission spectra at different heights from the lower nozzle: the 4 mm spectrum corresponds to a flame region below the stagnation plane with temperature less to 1000 K and high fuel concentration; 5 mm is approximately the position of the stagnation plane whereas at 6 mm high temperatures occur close to the front flame. According to previous studies [16,21], LIF signals are attributed to aromatic hydrocarbons in condensed-phase nanostructures which are not able to incandesce [14,21 and reference therein]. We collected the entire spectrum and thus the swan band interference was accurately avoided. The UV LIF is assigned to condensed-phase nanostructures(up to 4 nm in size) mostly constituted by aromatic compounds with 2–3 condensed rings and generally a loose structure; visible-LIF is attributed condensed-phase nanostructures constitued by PAH with 4–5 condensed rings and a well-organized structure. LII signal is assigned to soot particles of 10–20 nm in size and larger aggregates [21].
2. Experimental procedure The counter-flow burner used in this study was the same as that used for previous works [15–18]. The burner consists of two opposite jet nozzles (ID 2.54 cm) vertically positioned. The oxidizer stream (O2/ diluent) was introduced from the upper nozzle, while the fuel stream (C2H4/Ar) from the bottom. Different flames were analyzed varying the concentration of CO2 in the oxidizer side. The fuel stream was fixed at 25% vol. C2H4 and 75% vol. Ar with an axial velocity of 13.2 cm/s. In the oxidizer stream, the percentage of O2 was fixed at 22% vol. and the axial velocity at 16.1 cm/s. In a first configuration (standard case) the 78% vol. of the oxidizer side was constituted totally by Argon then the 25%, 50% and 78% was substituted with CO2 and the remaining Argon. The distance between the two burners was maintained at 1.5 cm for all flames. With these flux velocities the stagnation plane is located around 5 mm from the lower nozzle. In all the configurations, the flame front was stabilized in the oxidizer plane. Measurements were performed at different heights along the centerline. The experimental setup is the same as that used in previous works and is sketched in Fig. 1 [15]. Some details are reported below. Measurements within the flame were performed at different locations between the opposed jet nozzles, from the fuel to the oxidizer. The measuring location was changed by moving the entire burner assembly up or down with respect to the sampling point using a translation system, with a spatial resolution of 0.1 mm. Temperature was measured by using a silica-coated Pt/Pt–13%Rh thermocouple with a large bead diameter of about 300 µm. The relatively large thermocouple bead was necessary to avoid thermocouple melting in the high temperature zones of the flame. A lower-size thermocouple bead was used in the fuel zone of the flame (temperatures below 1000 K) and gave approximately the same results after radiation correction. A rapid insertion procedure was applied to measure temperature in order to reduce soot deposition on the thermocouple bead [19]. To minimize the heat conduction along the wires, the thermocouple was positioned parallel to the nozzle surfaces. Particulate deposited on the thermocouple was burnt off by using a methane torch. Temperature measurements were corrected for gas radiations [20]. Temperature data were highly repeatable; measurement uncertainty remains within 50 K. Laser Induced Emission (LIE) measurements were performed using
3. Kinetic model The mechanism of gas phase oxidation and pyrolysis presented here is the same that has been tested for ethylene in different combustion conditions including premixed [22–24] coflow [25] and counterflow diffusion flames [26]. The mechanisms includes the principal routes for small hydrocarbon oxidation and pyrolysis and the main pathways to form benzene ring. Benzene formation pathways includes both the C4 routes with the C2H2 addition and the self-combination of propargyl radicals (C3route). For PAH formation the HACA mechanism [27–28] is considered together with the combination of resonantly-stabilized radicals. This latter includes the combination of two cyclopentadienyl radicals and the combination of benzyl and propargyl radicals [29] for the formation of naphthalene. PAH growth is followed punctually up to pyrene which is here considered the largest gas phase compound. The gas-phase kinetic mechanism consists of 460 reactions involving 120 species. Details of the reaction rates used are reported in refs [22,23]. Species with molecular weight MW > 202u (i.e., pyrene) were treated as lumped species, i.e. particle phase, and thus divided into classes. Using lumped species allows us to evaluate the collision frequency of reactions based on gas-kinetic theory. The activation energy for the reactions involving lumped species was assigned based on 2
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
Fig. 1. Lay out of the experimental system. 1.0E+6
pathways reported above. These large molecules are treated as lumped species but cannot be considered particles or particle nuclei. Inception of particles - i.e. the formation of the first particle nucleus or lumped species that has a condensed phase behavior - is supposed to occur as a coagulation event between two molecules. Molecule coagulation was considered irreversible and its reaction rate was modeled by considering a coagulation efficiency with respect to the collision frequency. The collision frequency increases with the increase of molecular mass of the molecules whereas the coagulation efficiency, γ, depends on both local flame temperature and chemistry of the colliding molecules. The chemistry of the particles was considered by evaluating the Hamaker constant [30] for the species involved in the coagulation process. The computed coagulation efficiency at 1500 K is of the order of 1E-4 for small colliding entities (the first lumped species) and increases to values of about 1 when the C number is about 1E6. These values are in agreement with experimental coagulation rates. Coagulation efficiency also increases as temperature decreases due to lower thermal rebound involved in coagulation process. Particle nuclei continue to react in the same way as the molecules; they can add molecules to increase their size, or remove H-atoms by dehydrogenation or C-atoms by OH and O2 oxidation, or they can coagulate with other molecules or other clusters.
4 mm 5 mm
LIE A.U.
8.0E+5
6 mm
6.0E+5
4.0E+5
2.0E+5
0.0E+0 250
300
350
400 450 Wavelenght [nm]
500
550
Fig. 2. Laser induced emission spectra for the flame C2H4/Ar/O2 collected at 4 mm (circles), 5 mm (squares) and 6 mm (triangles) from the fuel nozzle.
analogues reactions in gas-phase or, if not possible as for dehydrogenation [23] or fragmentation [25], chosen to fit the experimental data. Thirty-one lumped species or sections were used in a geometric series of carbon number with a ratio of two between sections; five sections were used for H/C variation; radical and neutral molecules were separately grouped. The equivalent size range of 1–800 nm was obtained with this discretization considering a density varying from 1 g/cm3 for the smallest species with the highest hydrogen content to 1.8 g/cm3 for species above 10 nm with the lowest hydrogen content. The molecular growth of aromatics starts with the formation of an aromatic radical by an H atom loss from aromatic molecules. Radical formation can occur via H or OH attack or unimolecular decomposition. Termination reactions of aromatic radicals end the growth sequence. The formation of PAH was modeled by addition of acetylene, following the HACA sequence. Alternatively, a PAH from gas phase can be added through chemical addition. The molecular growth process competed with oxidation of stable molecules and radicals by OH and O2, respectively. Details of the reaction rates used in the growth process of aromatics are reported elsewhere [25]. Large molecules can indeed be formed by chemical reaction between two gas phase PAH and can keep growing following the chemical
4. Results and discussion The flame of the base case (undiluted conditions) has been stabilized by using argon as diluent for the oxidizer stream (Ar 78%mol) and for the ethylene. Fig. 3 shows the LIF and the LII signal profiles along the distance between opposed jets. Three regions can be distinguished along the flame: fuel side, stagnation plane and oxidizer region. The flame front is located in the oxidizer side in which PAH growth is due to chemical pathways. Rapid growth due the high temperature is confirmed by the fast increase of both LIF and LII signal. The stagnation plane is located around 4.8 mm from the lower nozzle; in this region the signals maximize due to both contributions from oxidizer and fuel side. In the fuel side, the LIF and LII signals can be justified with pyrolysis reactions of the fuel and a physical mechanism of PAH growth (van der Waals interactions). Two broad peaks are distinguished in the emission spectra reported in Fig. 2: a broad one in the UV region at around 330 nm and a sharper one in the visible around 440 nm. In a previous study [21,32] we have attributed, on the basis of LIF emission wavelengths and LIF lifetimes, 3
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
8.0E+5
1.0E+6 LIF@330nm
LIF@440nm 0% CO2
LIF@440nm
8.0E+5
LIF@440nm 25% CO2
LII
6.0E+5
LIF@440nm 78% CO2
LIF A.U.
6.0E+5 LIE A.U.
LIF@440nm 50% CO2
4.0E+5
4.0E+5
2.0E+5
2.0E+5 0.0E+0
0.0E+0 2
2.5
3
3.5 4 4.5 5 5.5 Distance from fuel nozzle [mm]
6
6.5
2
7
LIF A.U.
7
6.5
LII A.U.
4.0E+5
0.0E+0 2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
7
Distance from fuel nozzle [mm]
Fig. 6. Laser induced incandescence signals as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer stream. Stagnation plane is reported as dashed line.
CO2 content increases, being still detectable when only CO2 is used as diluent. Fig. 6 shows LII signals with the addition of CO2. LII signals show the same behavior found for LIF signals. In fact, for small percentage of CO2 added, signal reduction, i.e., particle formation suppression, is more evident in the oxidizer side. As the CO2 percentage increases, the reduction of LII signal is significant thorough the flame. Finally when the 78% CO2 was added the LII signal is quite close the uncertainty of the experimental measurements, suggesting the formation of large soot particles is negligible. Fig. 7 shows the trends of ratio between laser light scattering signals and LII signals with the addition of CO2. This ratio is proportional to the third power of the mean particle size and it is not reported for the 78% CO2 flame because of the very low LII signal due to the absence of particles when 78% of CO2 is added. It is evident from scattering that increasing the amount of CO2 added, the average size of the particles decreases. The effect of CO2 additionis clearly evidenced by the percentage reduction of the different signals with respect to the 0% CO2 flame reported in Fig. 8. An almost total LII reduction is registered when 78% CO2 is added to the oxidizer stream while 70–80% LIF signal reduction is observed. The effect of CO2 addition on particle formation has been also analyzed with the help of numerical analysis. The model adopted here has been already tested and found able to reproduce the formation and evolution of nanoparticles and soot in the base case (argon dilution).
0.0E+0 6
LII 50% CO2
2.0E+5
2.0E+5
5.5
6.5
6.0E+5
4.0E+5
5
6
LII 78% CO2
LIF@330nm 78% CO2
4.5
5.5
LII 25% CO2
LIF@330nm 50% CO2
4
5
8.0E+5
LIF@330nm 25% CO2
3.5
4.5
LII 0% CO2
LIF@330nm 0% CO2
3
4
1.0E+6
8.0E+5
2.5
3.5
Fig. 5. Laser induced fluorescence signals at 440 nm as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer stream. Stagnation plane is reported as dashed line.
LIF signals to nanoparticle of organic carbon having different chemical structures and morphology. In particular, LIF in the UV was attributed to high-molecular mass cross-linked aromatic molecules with few condensed aromatic rings, whereas LIF in the visible was attributed to highly-packed, sandwich-like clusters of aromatic hydrocarbons [21,31]. These structures are considered in both cases to be in condensed phase. LII signal was attributed to solid soot particles which are able to dissipate the acquired energy by thermal emission rather than by LIF emission. In order to evaluate and subtract the contribution of the LII signal, a black-body radiation curve at 4000 K has been used. After assessing the soot formation in the base case flame, the effect of CO2 has been evaluated by substituting the Ar in the oxidizer stream. The amount of CO2 has been varied in the oxidizer stream in order to study the effect of CO2 on soot formation by performing LIF and LII measurements. Fig. 4 reports UV LIF signal profiles along the flame at varying CO2 concentration in the oxidizer stream in comparison with the flame C2H4/Ar/O2 (0% CO2 in the figure). LIF signal can be distinguished both in fuel and oxidizer side. Overall it can be noticed a strong reduction of the signals with respect to the flame C2H4/Ar/O2. LIF signal in the oxidizer side, i.e. close to the flame front, in the zone of higher temperature and larger particles formation is significantly reduced already at low percentage of CO2. Only when 78% CO2 is added, the LIF signal is strongly reduced also in the pyrolytic zone. Fig. 5 shows visible LIF signals with the addition of CO2. The LIF at 440 nm maximizes the signal in the pyrolytic zone of the flame, being related to nanostructure formed by large PAHs at low temperatures (T < 1500 K). The reduction of both UV and visible LIF increases as
2
3
Distance from fuel nozzle [mm]
Fig. 3. Laser induced emission signals as a function of the distance from the fuel nozzle for the flame C2H4/Ar/O2. Stagnation plane is reported as dashed line.
6.0E+5
2.5
7
Distance from fuel nozzle [mm]
Fig. 4. Laser induced fluorescence signals at 330 nm as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer stream. Stagnation plane is reported as dashed line. 4
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna 100
% reduction of mean particle volume
1.0E+02
QVV/LII
1.0E+01
1.0E+00
QVV/LII 0% CO2
90
80
70
60
50
modelling QVV/LII
QVV/LII 25% CO2 QVV/LII 50% CO2
40
0
1.0E-01
2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
LIF@330nm LIF@440nm LII
% Signal Reduction
60
40
20
0 40
60
80
100
% CO2
Fig. 8. Percentage reduction of laser induced emission signal as a function of the CO2 concentration in the oxidizer stream. 100 90 80
% soot reduction
40
50
60
70
80
90
100
respectively. The agreement of the model results with the experimental data for both soot (LII signal) and the mean particle sizes reduction can be considered satisfying. In particular, the model is able to reproduce the rapid disappearing of large particles as the CO2 percentage in the oxidizer stream is increased. The model analysis suggests that the effect of CO2 on soot formation is mainly a thermal effect; temperature is significantly reduced by moving from a monoatomic gas (Ar) to a polyatomic gas (CO2) as diluent. Temperature reduction is more significant closer to the flame front, i.e. in the oxidizer side, where the CO2 effect is observed also for small amount of CO2 added. Contrarily, to have a significant impact on temperature also in the pyrolytic side, the CO2 percentage has to be increased above 50%. The effect on large particles is more evident since their formation occurs mostly in the high temperature zone of the flame whereas small particles and PAH which are detected by LIF are also formed in the pyrolytic side of the flame and in general the effect on the formation of small particles seems to be less significant. The results presented here are generally in agreement with other results presented in previous works [2–7]. It is worth to note that in the investigated conditions a chemical effect of CO2 cannot be completely excluded; however, its role has to be considered of minor importance with respect to thermal effect especially when CO2 is used in high percentage as diluent.
100
20
30
Fig. 10. Percentage reduction, with respect to the 0% CO2 flame, of the measured (QVV/LII) and predicted particle volume as a function of the CO2 concentration in the oxidizer stream.
Fig. 7. Laser light scattering/LII ratio as a function of the distance from the fuel nozzle for different concentration of CO2 in the oxidizer). Stagnation plane is reported as dashed line.
0
20
CO2, %
Distance from fuel nozzle [mm]
80
10
7
70
5. Conclusions
60
Optical diagnostics were used to understand the effect of different amount of CO2 as diluent in a counter flow ethylene diffusion flame. Both LIF and LII were performed in order to detect both nanoparticles and large soot aggregates respectively. CO2 was added in the oxidizer stream up to 78% vol. of the total flux. No qualitative variations of LIF and LII trends were found along the centerline of the burner. The CO2 addition strongly reduces LIF and LII signals in the fuel side and in the oxidizer side. However, while LII, i.e. soot aggregates are completely depleted when only CO2 was used as diluent, LIF signal, i.e. nanoparticles, were still detected in the flame. Modelling analysis suggests that the effect of CO2 on soot formation is mainly a thermal effect in the investigated conditions.
50 40 30 20 modelling
10 0
LII
0
10
20
30
40
50
60
70
80
90
100
CO2, % Fig. 9. Percentage reduction, with respect to the 0% CO2 flame, of laser induced incandescence and predicted soot concentration as a function of the CO2 concentration in the oxidizer stream.
CRediT authorship contribution statement M. Sirignano: Methodology, Investigation, Writing - original draft. A. D’Anna: Supervision, Conceptualization, Methodology, Writing original draft, Writing - review & editing.
Figs. 9 and 10 show comparison of model results with experimental data reported as percentage reduction for soot and mean size
5
Experimental Thermal and Fluid Science 114 (2020) 110061
M. Sirignano and A. D'Anna
Declaration of Competing Interest [16]
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.
[17]
References [18] [1] T.F. Wall, Combustion processes for carbon capture, Proc. Combust. Inst. 31 (1) (2007) 31–47. [2] K.P. Schug, Y. Manheimer-Timnat, P. Yaccarino, I. Glassman, Sooting behaviour of gaseous hydrocarbon diffusion flames and the influence of additives, Combust. Sci. Technol. 22 (1980) 235–250. [3] D.X. Du, R.L. Axelbaum, C.K. Law, The influence of carbon dioxide and oxygen as additives on soot formation in diffusion flames, Proc. Combust. Inst. 23 (1990) 1501–1507. [4] F. Liu, H. Guo, G.J. Smallwood, Ö.L. Gülder, The chemical effects of carbon dioxide as an additive in an ethylene diffusion flame: implications for soot and NOx formation, Combust. Flame 125 (2001) 778–787. [5] K.C. Oh, H.D. Shin, The effect of oxygen and carbon dioxide concentration on soot formation in non-premixed flames, Fuel 85 (2006) 615–624. [6] Ö.L. Gülder, M.F. Baksh, Influence of carbon dioxide dilution on soot formation in diffusive combustion, Combust. Inst. Can. Sect., Quebec, QC (1993) pp. 40–1–40-5. [7] H. Guo, G.J. Smallwood, A numerical study on the influence of CO2 addition on soot formation in an ethylene/air diffusion flame, Combust. Sci. Technol. 180 (2008) 1695–1708. [8] M. Abián, A. Millera, R. Bilbao, M.U. Alzueta, Effect of recirculation gases on soot formed from ethylene pyrolysis, Combust. Sci. Technol. 184 (7–8) (2012) 980–994. [9] M. Abián, A. Millera, R. Bilbao, M.U. Alzueta, Experimental study on the effect of different CO2 concentrations on soot and gas products from ethylene thermal decomposition, Fuel 91 (1) (2012) 307–312. [10] M.R. Charest, Ö.L. Gülder, C.P. Groth, Numerical and experimental study of soot formation in laminar diffusion flames burning simulated biogas fuels at elevated pressures, Combust. Flame 161 (2014) 2678–2691. [11] J. Wu, L. Chen, P.E. Bengtsson, J. Zhou, J. Zhang, X. Wu, K. Cen, Effects of carbon dioxide addition to fuel on soot evolution in ethylene and propane diffusion flames, Combust. Flame 199 (2019) 85–95. [12] C. Zhang, A. Atreya, K. Lee, Sooting structure of methane counterflow diffusion flames with preheated reactants and dilution by products of combustion, Symp. Int. Combust. 24 (1) (1992) 1049–1057. [13] Y. Wang, S.H. Chung, Formation of soot in counterflow diffusion flames with carbon dioxide dilution, Combust. Sci. Technol. 188 (4–5) (2016) 805–817. [14] A. D’Anna, Combustion-formed nanoparticles, Proc. Combust. Inst. 32 (1) (2009) 593–613. [15] A. D’Anna, M. Commodo, M. Sirignano, P. Minutolo, R. Pagliara, Particle formation
[19] [20]
[21]
[22] [23]
[24] [25] [26] [27] [28] [29]
[30] [31]
[32]
6
in opposed-flow diffusion flames of ethylene: An experimental and numerical study, Proc. Combust. Inst. 32 (2009) 793–801. M. Sirignano, A. Collina, M. Commodo, P. Minutolo, A. D’Anna, Detection of aromatic hydrocarbons and incipient particles in an opposed-flow flame of ethylene by spectral and time-resolved laser induced emission spectroscopy, Combust. Flame 159 (2012) 1663–1669. M. Salamanca, M. Sirignano, M. Commodo, P. Minutolo, A. D’Anna, The effect of ethanol on the particle size distributions in ethylene premixed flames, Exp. Therm. Fluid Sci. 43 (2012) 71–75. M. Sirignano, M. Salamanca, A. D’Anna, The role of dimethyl ether as substituent to ethylene on particulate formation in premixed and counter-flow diffusion flames, Fuel 126 (2014) 256–262. J.H. Kent, H. Gg, Wagner, Why do diffusion flames emit smoke, Combust. Sci. Technol. 41 (5–6) (1984) 245–269. A. Rolando, C. Allouis, F. Beretta, A. D’Alessio, A. D’Anna, P. Minutolo, Measurement of particulate volume fraction in a co-flow diffusion flame using transient thermocouple technique, Combust. Sci. Technol. 176 (5–6) (2004) 945–958. M. Sirignano, D. Bartos, M. Conturso, M. Dunn, A. D’Anna, A.R. Masri, Detection of nanostructures and soot in laminar premixed flames, Combust. Flame 176 (2017) 299–308. A. D’Anna, Detailed kinetic modelling of particulate formation in laminar premixed flames of ethylene, Energy Fuels 22 (3) (2008) 1610–1619. M. Sirignano, J. Kent, A. D’Anna, Detailed modeling of size distribution functions and hydrogen content in combustion-formed particles, Combust. Flame 157 (6) (2010) 1211–1219. A. D’Anna, M. Sirignano, J. Kent, A model of particle nucleation in premixed ethylene flames, Combust. Flame 157 (11) (2010) 2106–2115. M. Sirignano, J. Kent, A. D’Anna, Modeling formation and oxidation of soot in nonpremixed flames, Energy Fuels 27 (4) (2013) 2303–2315. M. Sirignano, J. Kent, A. D’Anna, Further experimental and modelling evidences of soot fragmentation in flames, Proc. Combust. Inst. 35 (2) (2015) 1779–1786. M. Frenklach, H. Wang, Detailed modeling of soot particle nucleation and growth, Symp. (Int.) Combust. 23 (1991) 1559–1566. M. Frenklach, Reaction mechanism of soot formation in flames, Phys. Chem. Chem. Phys. 4 (11) (2002) 2028–2037. N.M. Marinov, W.J. Pitz, C.K. Westbrook, M.J. Castaldi, S.M. Senkan, Modeling of aromatic and polycyclic aromatic hydrocarbon formation in premixed methane and ethane flames, Combust. Sci. Technol. 116 (1–6) (1996) 211–287. H.C. Hamaker, The London-van der Waals attraction between spherical particles, Physica 4 (10) (1937) 1058–1072. A. D’Alessio, A.C. Barone, R. Cau, A. D’Anna, P. Minutolo, Surface deposition and coagulation efficiency of combustion generated nanoparticles in the size range from 1 nm to 10 nm, Proc. Combust. Inst. 30 (2) (2005) 2595–2603. D. Bartos, M. Dunn, M. Sirignano, A. D'Anna, A.R. Masri, Tracking the evolution of soot particles and precursors in turbulent flames using laser-induced emission, Proc. Combust. Inst. 36 (2) (2017) 1869–1876.