Sooting structure of a premixed toluene-doped methane flame

Combustion and Flame 190 (2018) 252–259

Contents lists available at ScienceDirect

Combustion and Flame journal homepage: www.elsevier.com/locate/combustflame

Sooting structure of a premixed toluene-doped methane flame Carmela Russo∗, Lucia Giarracca, Fernando Stanzione, Barbara Apicella, Antonio Tregrossi, Anna Ciajolo Istituto di Ricerche sulla Combustione, IRC-CNR, Piazzale Tecchio, 80, Napoli, Italia

a r t i c l e

i n f o

Article history: Received 14 July 2017 Revised 5 October 2017 Accepted 5 December 2017

Keywords: Soot Aromatic fuels Toluene combustion Soot structure Spectroscopic diagnostics Toluene

a b s t r a c t The structure of a sooting premixed methane/oxygen flame doped with toluene (C/O = 0.66) was studied by means of sampling and chemical and spectroscopic analysis of gas and condensed phases. It was found that the addition of toluene in small amounts (0.8 mol%) to the methane flame (C/O = 0.6) used as reference did not affect the maximum temperature level and the distribution of the main combustion products and C1–C4 hydrocarbons. The main differences observed regarded the higher yield of soot and C5–C6 hydrocarbons as cyclopentadiene and benzene. Some insights on soot structure were obtained by means of spectroscopic tools, i.e. Raman and UV–Visible spectroscopy. Spectral features of toluene-doped soot indicated that toluene doping not only enhances soot production, but also changes its structural properties in terms of aromatic character of soot particles. These features could be of concern about the use of practical fuels rich in aromatic components possibly leading to higher environmental and health impact of aromatic-derived soot in respect to aliphatic-derived soot. © 2017 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

1. Introduction Aromatic hydrocarbons are important as additive or components of transportation fuels as gasoline, kerosene and diesel fuels because of their resistance to knocking in turn related to the aromatic ring stability and resistance to oxidation [1,2]. However, aromatics are much more prone to form soot and are considered the main precursors in soot formation [3]. Toluene is the most representative aromatic component of gasolines and a good surrogate of aromatics contained in diesel oils, thereby it is regarded as a primary reference fuel whose chemistry has to be studied in combustion conditions. It is also important to study the toluene combustion as involving the chemistry of benzyl which is an important intermediate in flames of alkylated aromatics leading to polycyclic aromatic hydrocarbons (PAH) [4], and eventually to soot. Most of the experimental work on toluene chemistry concerns pyrolysis in flow reactors and oxidation in a jet stirred reactor [5–8]. Furthermore, some works on toluene chemistry in flames have been carried out in flames of pure toluene at low pressure [4,9–13] and in diffusion flames burning toluene added to aliphatic hydrocarbons as ethylene [14,15], methane [16–18] and heptane [19]. The present work reports the study of the sooting structure of a premixed fuel-rich flame of methane doped with relatively low amount of toluene (0.8 mol%) to avoid thermal perturbations to

∗

Corresponding author. E-mail address: [email protected] (C. Russo).

the methane flame which has been previously characterized in detail [20–22]. The chemical environment in which soot is formed has been investigated by the detailed analysis of gaseous and condensed phases sampled along the axis of the toluene-doped flame. Particular attention has been given to the structural analysis of the toluene-doped soot in comparison to soot formed in methane flames to study the effects of toluene doping on soot nanostructure, which could be expected on the basis of recent works on the microstructural analysis of soot formed from aliphatic and aromatic hydrocarbon fuels [21–28]. 2. Experimental The toluene-doped premixed flame was produced on a McKenna (Holthuis & Associates) water-cooled burner by adding small amounts (0.8 mol%) of toluene prevaporized to a previously studied methane/oxygen flame [20,21]. The cold gas velocity of the toluene-doped flame (5 cm/s) was equal to that of the methane flame (54.55 CH4 /45.45 O2 mol.%), hereafter named baseline methane flame. To avoid toluene condensation, the cooling temperature of the burner was kept at 70 °C, largely above the dew point evaluated (0 °C) for the feed mixture (0.8 toluene/54.1 methane/45.1 oxygen mol%). Flame temperature was measured using a fast-response silicacoated fine wire Pt/Pt–13%Rh thermocouple: 100 μm wires are jointed to 25 μm wires, which are used to realize a very small bead size of about 50 μm. The mechanical strength of the larger wires avoided any vibrations. The temperature profile measured

https://doi.org/10.1016/j.combustflame.2017.12.004 0010-2180/© 2017 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

C. Russo et al. / Combustion and Flame 190 (2018) 252–259

with such a thermocouple system can be considered quite unperturbed and a reliable reference for adjusting the axial coordinates of intrusive sampling data. A fast-insertion procedure was used to avoid massive soot deposition on the thermocouple bead and the consequent change of the bead size and of the emissivity of the thermocouple junction. Temperature was also corrected for the radiative losses [29]. The uncertainty of the measured temperature was estimated to be as high as 100 K. By means of in-situ laser measurements, the extinction coefficient (defined as -ln(I/I0 )/L, where I0 is the incident monochromatic radiation intensity, I is the emerging radiation intensity and L is the flame thickness) was measured using a 5 mW He/Ne laser (λ = 632.8 nm). Gaseous and condensed phases were sampled from the flames by a stainless steel, water-cooled, isokinetic probe with a conical tip (i.d. 2 mm, o.d. 3 mm, the outer diameter increases from 3 to 6 mm with a cone angle of 60 °C). Stable gases (O2 , CO, CO2 , H2 , C1 –C6 hydrocarbons and toluene) were analyzed by on-line gaschromatographic analysis. C1 –C6 hydrocarbons and toluene were analyzed on a HP5890A gas chromatograph equipped with a 7515 Chrompack Al2 O3 /KCl capillary column and a flame ionization detector (FID). O2 , CO, CO2 , H2 were analyzed by using a 8700 Altech coaxial column and a thermal conductivity detector (TCD) on a HP5700A gas chromatograph. The gas sampling at low height above the burner (HAB) was performed removing the ice trap in order to avoid the condensation of unburned toluene before of the gas sampling valve. The data were obtained as average of measurements repeated thrice. Condensed phases collected in an ice-cooled trap and on a teflon filter placed along the sampling line were extracted with dichloromethane (DCM) to separate the organic carbon soluble in DCM, named DCM extract, from soot. Both DCM-extract and soot were dried and weighed to determine their concentration in flame. To get enough amounts of condensed phases and to estimate the reproducibility of the measurements, each sampling point was repeated at least three times and the samples were added together for further analysis. The uncertainty of the measured concentrations due both to the sampling and analytical procedure was about 25% at low HAB within the main reaction zone, less than 10% up to 10 mm HAB. The uncertainty increases again up to 20% after 10–12 mm due to the oscillations of the flame tail approaching the stabilizer plate. Twenty-four polycyclic aromatic hydrocarbons (PAH) with molecular masses up to 300 u (Acenaphthylene, Acenaphtene, Anthracene, Benz[a]anthracene, Benzo[b]fluoranthene, Benzo[k]fluoranthene, Benzo(ghi)fluoranthene, Benzo[ghi]perylene, Benzo[a]pyrene, Benzo[e]pyrene, Biphenyle, Biphenylene, Chrysene, 4H-Cyclopenta(def)phenantrene, Coronene, Dibenz[a,h]anthracene, Fluorene, Fluoranthene, Indeno[1,2,3-cd]pyrene, Naphthalene, Phenanthrene, 2-phenyl-naphtalene, Perylene, Pyrene) were detected and quantified by gas chromatography/mass spectrometry (GC–MS) of the DCM-extract by a HP5890 gas chromatograph coupled with a HP5975 mass spectrometer equipped with an HP-5MS crosslinked 5% PhMe siloxane 30 m × 0.25 mm × 0.25 μm film thickness column. UV–Visible absorption spectra of soot and DCM-extract suspended in N-methyl-2-pyrrolidinone (NMP) (concentration = 10 mg/l) were measured in a 1-cm quartz cell by using an UV–Visible spectrophotometer (HP8453). Raman spectra of soot were measured directly on the filter in the range of 900– 3400 cm−1 (Raman shift) by means of a Horiba XploRA Raman microscope system with an excitation wavelength of 532 nm. To minimize the possibility of structural damages due to the thermal decomposition induced by the laser, the power of the excitation laser beam was reduced to about 0.1 mW. Hydrogen and carbon content of soot and carbon standard materials were measured by a Leco CHN628 elemental analyzer. FT-IR spectra of soot were obtained with a Nicolet iS10 spectrophotometer in the range of 60 0–340 0 cm−1 on dispersions prepared by mixing and grinding

253

Fig. 1. Axial profiles of temperature (upper part) and in-situ laser extinction coefficient measured at 632 nm (lower part) measured in the toluene-doped (full symbols) and in the baseline (open symbols) methane flames.

the carbon samples in KBr pellets at a fixed concentration (0.25 wt. %) [30]. 3. Results and discussion 3.1. Flame structure The thermal field and sooting structure of the toluene-doped flame and baseline methane flame [20–22] are represented in Fig. 1 that reports the axial profiles of the temperature and in-situ laser extinction coefficient, measured in the visible (λ = 632.8 nm). It can be noticed that the temperature steeply increases to reach the same maximum value (1780 K) around 2 mm HAB in both flames, demonstrating that there is no effect of toluene addition on the thermal field in the early oxidation region. Downstream of the temperature peak, the steeper temperature decrease in the toluene-doped flame is symptomatic of a larger soot formation causing higher radiative losses and the consequent flame temperature reduction. The higher soot formation along the toluene-doped flame axis is confirmed by the much higher laser extinction coefficient values measured in the visible (lower panel of Fig. 1), which is directly associated to soot absorption. The axial concentration profiles of gaseous and condensed phases measured respectively, by on-line gaschromatography and batch analysis, give an overview of the chemical structure of the toluene-doped flame. To account for the probe interference on the thermal and fluid-dynamic flame fields, the concentration profiles have been shifted upstream of few millimeters, namely 2.4 mm, evaluated by placing the maximum of CO concentration in correspondence of the flame position (height) where the temperature reaches the maximum value [29].

254

C. Russo et al. / Combustion and Flame 190 (2018) 252–259

Fig. 3. Carbon yield profiles of cyclopentadiene (upper part) and benzene (lower part) along the toluene-doped flame (filled symbols) and baseline methane flame (empty symbols).

Fig. 2. Concentration profiles of fuel mixture (methane and toluene) components and oxygen (a and b), main combustion products, and HC (sum of C2–C6 hydrocarbons) multiplied by a factor of 2 (b), C2–C3 (c) and C4 hydrocarbons (d) measured along the toluene-doped flame.

3.2. Gas phase The concentration profiles of the fuel mixture components (methane and toluene), O2 , H2 , CO, CO2 , the sum of the C2 –C6 hydrocarbons (named HC) and individual C2 –C4 hydrocarbons, reported in Fig. 2, outline the typical fuel-rich premixed laminar flame structure. Specifically, the main oxidation region, extending few millimeters (about 3 mm) above the burner, is featured by the steep consumption of fuel components and oxygen along with the

maximum formation of the main combustion products (CO, CO2 , H2 ) (Fig. 2a and b). It is worth noting that the toluene doping does not affect CO and CO2 formation, as shown in Table 1 reporting the maximum carbon yield ((moles of carbon in the product/s)/(total moles of fed carbon/s)) of the main combustion products and the flame position where the maximum yield occurs. Concerning the hydrocarbon distribution, acetylene (C2 H2 ) is the most abundant product (Fig. 2c) as typically occurs in fuel-rich premixed flames. Acetylene as well as the other alkynes identified among the hydrocarbon products, namely propyne (C3 H4 ), 1buten-3–yne (C4 H4 ) and butadiyne (C4 H2 ), early increase attaining to an almost constant high concentration value downstream of the flame. C2 –C4 alkenes as ethylene (C2 H4 ), propene (C3 H6 ), propadiene (C3 H4 ) and butadiene (C4 H6 ), exhibit inside the main reaction region the rise-decay trend characteristic of intermediates. Among C2 –C4 , propyne, 1,3 butadiene and 1-buten-3–yne are the hydrocarbons mainly affected by toluene addition, presenting higher yields in comparison to the baseline flame (Table 1). Likewise, the formation of the most abundant components of the C5 – C6 hydrocarbon class, namely cyclopentadiene (C5 H6 ) and benzene (C6 H6 ), is significantly enhanced in the toluene-doped flame (Table 1). This is well shown in Fig. 3 contrasting their carbon yield profiles along both toluene-doped and baseline flames. The higher yield of C3 –C6 hydrocarbons could be merely attributed to the slightly higher C/O ratio of the toluene-doped flame, however, the different shape of the benzene concentration profile testifies the occurrence of a specific effect of toluene doping on hydrocarbon formation. Particularly, the benzene

C. Russo et al. / Combustion and Flame 190 (2018) 252–259

255

Table 1 Maximum carbon yields of the main combustion products and C2–C6 hydrocarbons measured in the toluene-doped and baseline methane flames. species

CO CO2 C2 H 4 C2 H 2 C3 H 6 C3 H 4 C3 H 4 C4 H 6 C4 H 4 C4 H 2 C5 H 6 C6 H 6

(ethylene) + C2 H6 (ethane) (acetylene) (propene) (propadiene) (propyne) (1,3 butadiene) (1-buten-3–yne) (butadiyne) (cyclopentadiene) (benzene)

Baseline methane flame

Toluene-doped flame

HAB, mm

Carbon yield, %

HAB, mm

Carbon yield, %

2.60 2.60 1.83 3.60 1.83 1.83 2.60 1.83 2.60 3.60 2.60 3.60

5.13E + 01 1.16E + 01 4.14E + 00 1.83E + 01 9.30E−02 1.95E−02 3.24E−01 4.10E−02 2.09E−01 5.20E−01 3.60E−02 7.30E−01

2.60 2.60 1.69 3.60 1.69 1.69 2.60 1.69 2.60 3.60 3.60 1.69

5.04E + 01 1.12E + 01 4.24E + 00 1.92E + 01 1.24E−01 2.03E−02 4.34E−01 5.00E−02 2.51E−01 4.20E−01 4.70E−02 1.12E + 00

Table 2 DCM-extract concentration and carbon yield measured in the toluene-doped and baseline methane flames. The weight percentage of GC-MS PAH identified is given in parentheses. HAB, mm

3.6 4.6 5.6 7.6 11.6

DCM-extract mg/Nl

DCM-extract carbon yield, %

Baseline methane flame

Toluene-doped flame

Baseline methane flame

Toluene-doped flame

0.67 (69.8%) 1.21 (59.2%) 1.02 (61.1%) 1.08 (39.5%) 1.44 (38.9%)

1.32 (37.2%) 1.61 (34.9%) 1.13 (42.6%) 1.18 (39.8%) 1.51 (34.3%)

0.28 0.54 0.45 0.47 0.61

0.57 0.71 0.52 0.50 0.68

Table 3 Soot concentration and carbon yield measured in the toluene-doped and baseline methane flames. HAB, mm

3.6 4.6 5.6 7.6 11.6

Soot mg/Nl

Soot carbon yield, %

Baseline methane flame

Toluene-doped flame

Baseline methane flame

Toluene-doped flame

0.21 0.97 1.35 2.29 2.37

0.51 1.21 2.15 3.02 4.48

0.09 0.44 0.61 1.03 1.04

0.22 0.54 1.01 1.32 2.12

concentration (lower part of Fig. 3) follows a rise-decay trend in form of a “spike” inside the main oxidation region of the toluenedoped flame. The methyl extraction from toluene through the reaction of toluene with the H radical, and/or through the formation of the benzyl radical (C7 H7 ), which is converted to benzene and phenyl in the reactions with HO2 [6], justifies the surplus of benzene in the main oxidation region of the toluene-doped flame. It is also noteworthy that cyclopentadiene is formed in larger amounts and later on in the toluene-doped flame, reaching the maximum just after the spike observed in the benzene profile suggesting the occurrence of the oxidation of the benzene surplus as main source of cyclopentadiene [1,31–34]. The concentration of individual C2–C6 hydrocarbons is reported in Figs. S1 and S2 in the Supplemental material. Summing up, the gas phase analysis of the toluene-doped flame so far reported demonstrates that the most evident effect of the toluene doping is the change of the shape and the higher concentration profile of benzene and cyclopentadiene. 3.3. Condensed phases The higher formation of the condensed phases, constituted of DCM-extract and soot, was expected just because of the higher benzene and cyclopentadiene formation [35–36]. The higher formation of the DCM-extract and soot are clearly shown in the Tables 2 and 3 reporting the concentration and yields of the

DCM-extract (Table 2) and soot (Table 3), measured at different HAB of the toluene-doped and baseline flames. 3.3.1. DCM-extract analysis The DCM-extract concentration trends along both flames are similar presenting a maximum at the end of the oxidation region (Table 2). The formation of PAH, which are the main components of the DCM-extract [35], is enhanced in the toluene-doped flame, as shown by the higher DCM-extract concentration. However, as also shown in Table 2, PAH up to 300 u, quantitatively measured by GC-MS analysis, account for a smaller amount (about 30 wt %) at the beginning of the toluene-doped flame in comparison to the baseline flame (about 60 wt %). More details on the individual PAH concentrations are reported elsewhere [37]. The lower PAH content in the DCM-extract produced early in the toluene-doped flame corresponds to the larger abundance of species as small aromatics (mainly styrene, indene, etc.), oxo-aromatic compounds (mainly phenol, substituted phenols, benzaldeyde and benzofuran) and substituted-PAH. All these species have been found in the oxidation region of benzene flames [36] as typical products of aromatic oxidation. Hence, their presence in the toluene-doped flame can be traced to the degradation of toluene and demonstrates the relevant effect of the toluene doping on the quality of the DCM-extract. In spite of the disappearance of light aromatic and oxo-aromatic species, the identification of PAH remains still scarce downstream of the main toluene-doped flame region,

256

C. Russo et al. / Combustion and Flame 190 (2018) 252–259

Fig. 4. H/C atomic ratio and mass absorption coefficient (measured at 500 nm) of the DCM-extract along the toluene-doped flame (filled symbols) and the baseline methane flame (empty symbols).

Fig. 5. H/C atomic ratio and mass absorption coefficient (measured at 500 nm) of soot formed along the toluene-doped flame (filled symbols) and the baseline methane flame (empty symbols).

attaining to a value, around 30 wt%, similar to that measured for the baseline flame (Table 2). It can be hypothesized that species with molecular masses above the mass detection threshold of GC-MS (> 300 u) should account for the unidentified part of the DCM-extract [35,38]. The unidentified part of the DCM-extract could be responsible for the noticeable light absorption in the visible range observed in Fig. 4 where the mass absorption coefficients (measured by UV– Vis spectroscopy at 500 nm) and the H/C ratio of the DCM-extract sampled along both flames are reported. It is worth to underline that for complex mixtures of unknown molecular mass it is common to evaluate the mass absorption coefficient from the Lambert– Beer law substituting the mass concentration to the molar concentration. In the practice, the light absorption at a given wavelength has been measured on solutions of the DCM-extract at a known concentration (10 mg/l) using a quartz cell of 1 cm length. Then, the mass absorption coefficient having units of m2 /g is calculated by dividing the absorption by the mass expressed as concentration (g/m3 ) and the cell length expressed in meters. It is noteworthy that the absorption in the visible is absent for a mixture of PAH as those detected by GC-MS, hence it could be ascribed to the presence of more aromatic and heavier species not amenable for gaschromatographic analysis. However, this attribution is not supported by the relatively high H/C ratio varying around 0.5 that is the typical H/C of standard two- to sevenring PAH. Detailed infrared spectroscopic analysis of soluble organic fractions derived from premixed flames [30,39], suggested the presence of aliphatic carbon inside and/or connecting aromatic moieties which could compensate for the decrease of H/C expected for larger aromatic species. In addition to being early and largely formed in the toluenedoped flame (Table 2), the DCM-extract exhibits also a higher absorption coefficient shifted upstream indicating that high molecular weight precursors with a higher aromatic character are associated to the earlier soot formation for effect of toluene doping. Unfortunately, the DCM-extract species are so strongly fluorescent that Raman spectroscopy, in the following applied for soot analysis, could not be used for a deeper investigation of their aromaticity features.

ticular, soot along the toluene-doped flame reaches values of concentration and yield higher by a factor of more than 2 (Table 3) compared to the baseline flame. This behavior along with the increasing trend of soot concentration is in agreement with the laser extinction trends (Fig. 1). Beside the enhancement of soot concentration/yield (Table 3), significant differences in soot properties for effect of toluene doping can be observed. The higher visible light absorption accompanied by the more effective soot dehydrogenation process of toluene-doped soot are shown in Fig. 5 which reports the comparison of the axial profiles of the mass absorption coefficient in the visible (500 nm) and H/C atomic ratio of soot particles sampled in the two flames. The mass absorption coefficient of soot has been derived from the measure of the UV–Vis spectral absorption of soot suspended in NMP (10 mg/l) in a quartz cell (length 1 cm). Moreover, Fig. 5 shows that the mass absorption coefficient of toluene-doped soot increases just after soot inception overcoming by a factor of 2 the mass absorption coefficient of baseline flame soot which instead remains rather low and constant throughout all the flame. By considering that the number and the size of sp2 mainly affect the absorption properties of soot particles clusters [40], the larger increase of the mass absorption coefficient for the toluene-doped soot indicates the occurrence during soot formation of a more significant aromatization in terms of delocalization and number of sp2 clusters. The Raman technique has been employed to give further information on the structural features of soot [41–43], particularly on the aromatic sp2 bonding [44,45]. The first order Raman region (10 0 0–180 0 cm−1 ) of soot spectra is characterized by the two typical peaks of carbon materials, around 1600 cm−1 (G or “graphite” peak) and 1350 cm−1 (D or “defect” peak) [41–43,46]. Beside the D and G bands, other minor modulations, induced by defects outside and inside the crystal lattice, occur next to the G peak and in the 110 0–130 0 cm−1 region [41]. Through an accurate spectral analysis purposely set up for soot samples [41] it is possible to make allowance of the contributions and features of the main and minor peaks evidencing differences not detectable on the raw spectra. A Lorentzian and a Breit–Wigner–Fano (BWF) curves were used to fit the D and the G peak, respectively. Two additional Lorentzian lines (D5 and D4) have been used to fit the features at 110 0–130 0 cm−1 , while a Gaussian line-shape has been chosen for the D3 band around 1500 cm−1 . Structural information can be derived from the exami-

3.3.2. Soot analysis In comparison to the DCM-extract, the effect of toluene doping appears more evident on soot formation (Table 3). In par-

C. Russo et al. / Combustion and Flame 190 (2018) 252–259

Fig. 6. I(D)/I(G) ratio (circles) and Pos(G) (diamonds) as evaluated from deconvolution of Raman spectra of soot formed along the toluene-doped flame (filled symbols) and the baseline methane flame (empty symbols) (A colour version of the figure is available online).

nation of the G peak position (Pos(G)) and I(D)/I(G) ratio reported in Fig. 6 for soot sampled along the toluene-doped and baseline flames. The first observation regards just the high value of Pos(G) which keeps well above the 1600 cm−1 value for all soot samples. The G band in the Raman spectra of carbons has been usually assigned to the C = C stretching of all pairs of sp2 atoms. Regardless the excitation wavelength, the G peak position takes a fixed value of 1580 cm−1 in perfect and infinite graphite crystals [44], upshifted to 1600 cm−1 for microcrystalline graphite, because of the effect of finite crystal size. The Pos(G) upshift in respect to microcrystalline graphite is generally attributed to the promotion and activation of the vibrational modes of sp2 groups characterized by a higher C = C bond stretching frequency. Likewise, the upshift of the Pos(G) recently found by multiwavelength Raman analysis of flame soot [41] can be ascribed to the presence of shorter C = C bonds featuring olefinic groups and/or small aromatic layers. Figure 6 clearly shows the decrease of the Pos(G) from around 1610 cm-1 toward the typical value of micrographite, i.e. 1600 cm−1 , as soot ages, in particular for the toluene-doped soot, testifying the increase of the aromatic layer size and sp2 aromatic content, i.e a more effective soot aromatization in respect to the baseline flame. This finding is also consistent with the much more evident increase of I(D)/I(G) ratio of toluene-doped soot (Fig. 6). In fact, the I(D)/I(G) ratio can be related to the size of the sp2 phase organized in ring clusters, i.e. the aromatic cluster size, La . Specifically, it has been assessed that the I(D)/I(G) decreases with increasing La for La > ∼2 nm [47], whereas it increases for La < ∼ 2 nm [45,48]. The HR-TEM analysis of whichever soot has shown that the in-plane layer length is generally much lower than 2 nm [26,49]. Actually, by means of the equation proposed by Ferrari and Robertson [45], the La estimated from the I(D)/I(G) ratio was found to modestly increase from 1.05 to 1.12 nm. Overall, both the values and trends of I(D)/I(G) ratio and Pos(G) (Fig. 6) indicate that soot aromatization occurs to a larger extent in the toluene-doped soot in agreement with the mass absorption coefficient trend (Fig. 5). To clear up the uncertainty about the effect of the higher C/O ratio on the toluene-doped soot properties, soot probed downstream of a methane flame having the same C/O ratio (C/O = 0.66) of the toluene-doped flame has been also analysed. Indeed, such flame, hereafter labeled as LT-M 0.66, has a lower temperature

257

Fig. 7. Axial profiles of the concentration of hydrogen linked to aliphatic and aromatic carbon soot along the toluene-doped (full symbols) and the baseline methane flame (empty symbols).

(Tmax = 1700 K), in comparison to the toluene-doped and baseline flames (Tmax = 1780 K), but the lower temperature should not affect the methane soot properties at least in the 160 0–180 0 K range as previously demonstrated [20]. Specifically, it has been demonstrated that the flame temperature variation in the 160 0–180 0 K temperature range does not affect neither the H/C ratio (around 0.2) nor the absorption coefficient of methane soot, which remains low (about 2 m2 /g) and rather constant along the flames [20]. The H/C and absorption coefficient of soot produced downstream (at around 12 mm HAB) of the toluene-doped flame, baseline and LTM 0.66 methane flames are reported in Table 4. It can be clearly noticed that the toluene-doped soot exhibits the higher absorption coefficient and I(D)/I(G) ratio, as well as the lower H/C in respect to both methane flames (baseline and LT-M 0.66 flames) confirming the enhancing effect of toluene doping on soot evolution, namely aromatization and dehydrogenation. Just as regards dehydrogenation, FT-IR analysis carried out in our previous work has shown that methane soot barely dehydrogenates in comparison to ethylene soot [30]. The same FT-IR methodology has been applied to the toluene-doped soot to investigate the effect of toluene doping on the hydrogen content and hence on the dehydrogenation process in respect to the baseline methane soot. The concentration of hydrogen linked to aromatic and to aliphatic carbon of the toluene-doped and the baseline soot samples, calculated multiplying the hydrogen weight fraction (measured by FT-IR analysis [30]) by soot concentration, is reported in Fig. 7. Soot hydrogen linked to aromatic carbon increases along the axis of both flames and levels off where soot formation rate decreases. The striking difference concerns the hydrogen linked to aliphatic carbon of soot that is rather abundant and scarcely depleted along the baseline methane flame whereas it is present in lower amounts and rapidly and fully consumed for soot formed in the toluene-doped flame (Fig. 7). As previously found, the FT-IR analysis shows that hydrogen linked to aliphatic carbon of soot is present mainly in form of methylene groups [30] that are the more labile hydrogen to be removed [50]. In previous work the lower soot dehydrogenation in the methane flame in comparison to an ethylene flame was related to the higher presence of molecular hydrogen and water deactivating soot and soot precursors radicals limiting the methane

258

C. Russo et al. / Combustion and Flame 190 (2018) 252–259 Table 4 Properties of soot sampled downstream of the toluene-doped, baseline methane and low temperature methane (LT-M 0.66) flames. Soot properties

Toluene-doped flame

Baseline methane flame

LT-M 0.66

Concentration, mg/Nl H/C Mass abs. coeff. @500 nm, m2 /g I(D)/I(G)

4.48 0.06 4.04 0.79

2.37 0.12 2.7 0.72

4.22 0.2 1.94 0.68

soot dehydrogenation and aromatization. This explanation does not hold for interpreting the soot structural differences found in the present work because toluene doping does not modify the concentration of molecular hydrogen that reaches a final similar value around 50% on dry basis for both flames. It can be speculated that the lower content of hydrogen linked to aliphatic carbon of soot is symptomatic of the higher reactivity of the chemical environment featuring soot inception phase in the toluene-doped flame. Specifically, as before mentioned the environment surrounding soot formation in the toluene-doped flame is richer of aromatic and oxyaromatic species and, very likely, of radicals (Table 2), which in turn could enhance soot reactivity promoting soot dehydrogenation and mass growth.

4. Final remarks The structure of a sooting premixed flame (C/O = 0.66) of methane doped with toluene (0.8 mol%) was studied by means of sampling and spectroscopic analysis and compared to that of the baseline methane flame burning at the same temperature. The overall flame structure in terms of main gas concentration and temperature was not significantly perturbed by the toluene doping. In particular, the detailed analysis of the gaseous phase demonstrated that the toluene addition to methane fuel did not significantly change the formation of the main combustion products (CO, CO2 and major light hydrocarbons), whereas it enhanced the formation of minor, but very important hydrocarbons as cyclopentadiene and benzene typically involved in the oxidation and pyrolysis of aromatic fuels and implied in the formation of higher molecular weight pollutant species as PAH and soot. Indeed, consistently with the higher amounts of gaseous C5 –C6 species, larger yields of condensed phases (DCM-extract fraction and soot) have been measured in the toluene-doped flame. Besides, the striking effect of toluene doping regarded soot structural properties investigated in comparison to soot formed in methane flames burning in similar temperature and/or C/O feed ratio. In particular, the toluene-doped soot underwent higher aromatization and dehydrogenation processes in respect to the other flames. A detailed study of soot dehydrogenation and growth by means of infrared analysis showed the rapid and full consumption of hydrogen linked to aliphatic carbon of the toluene-doped soot. The early and steep decrease of the more labile hydrogen of the aliphatic groups of toluene-doped soot could be interpreted as a signature of the higher reactivity of environment surrounding the first phase of soot formation in the toluene-doped flame which in turn caused the increase of active surface sites available for further carbon addition from tarry aromatic species and hydrocarbons. Afterwards, coagulation and thermal annealing of soot particles accompanied by a negligible dehydrogenation became the predominant phenomena downstream of the soot formation region, and the corresponding steep rise of the absorption coefficient is a signature of the increase of aromatic structures more aligned and more interconnected by conjugated sp2 bonds, that enables electron mobility.

Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.combustflame.2017.12. 004. References [1] J.L. Emdee, K. Brezinsky, I. Glassman, A kinetic model for the oxidation of toluene near 1200K, J. Chem. Phys. 96 (1992) 2151–2161. [2] S.D. Klotz, K. Brezinsky, I. Glassman, Modeling the combustion of toluene-butane blends, Symp. (Int.) Combust. 27 (1998) 337–344. [3] C.S. McEnally, L.D. Pfefferle, B. Atakan, K. Kohse-Höinghaus, Studies of aromatic hydrocarbon formation mechanisms in flames: Progress towards closing the fuel gap, Prog. Energy Combust. Sci. 32 (2006) 247–294. [4] Z. Tian, W.J. Pitz, R. Fournet, P.A. Glaude, F. Battin-Leclerc, A detailed kinetic modeling study of toluene oxidation in a premixed laminar flame, Proc. Combust. Inst. 33 (2011) 233–241. [5] R. Sivaramakrishnan, R.S. Tranter, K. Brezinsky, High pressure pyrolysis of toluene. 2. Modeling benzyl decomposition and formation of soot precursors, J. Phys. Chem. A 110 (2006) 9400–9404. [6] P. Dagaut, G. Pengloan, A. Ristori, Oxidation, ignition and combustion of toluene: Experimental and detailed chemical kinetic modeling, Phys. Chem. Chem. Phys. 4 (2002) 1846–1854. [7] W. Yuan, Y. Li, P. Dagaut, J. Yang, F. Qi, Investigation on the pyrolysis and oxidation of toluene over a wide range conditions. I. Flow reactor pyrolysis and jet stirred reactor oxidation, Combust. Flame 162 (2015) 3–21. [8] W. Yuan, Y. Li, P. Dagaut, J. Yang, F. Qi, Investigation on the pyrolysis and oxidation of toluene over a wide range conditions. II. A comprehensive kinetic modeling study, Combust. Flame 162 (2015) 22–40. [9] A. Hamins, K. Seshadri, The structure of diffusion flames burning pure, binary, and ternary solutions of methanol, heptane, and toluene, Combust. Flame 68 (1987) 295–307. [10] V. Detilleux, J. Vandooren, Experimental and kinetic modeling evidences of a C7 H6 pathway in a rich toluene flame, J. Phys. Chem. A 113 (2009) 10913–10922. [11] Y.Y. Li, L.D. Zhang, Z.Y. Tian, T. Yuan, J. Wang, B. Yang, F. Qi, Experimental study of a fuel-rich premixed toluene flame at low pressure, Energy Fuels 23 (2009) 1473–1485. [12] Y. Li, J. Cai, L. Zhang, T. Yuan, K. Zhang, F. Qi, Investigation on chemical structures of premixed toluene flames at low pressure, Proc. Combust. Inst. 33 (2011) 593–600. [13] V. Detilleux, J. Vandooren, Experimental and kinetic modeling investigation of toluene combustion in premixed, one-dimensional and laminar toluene-oxygen-argon flames, Proc. Combust. Inst. 33 (2011) 217–224. [14] S.J. Harris, A.M. Weiner, Soot particle growth in premixed toluene/ethylene flames, Combust. Sci. Technol. 38 (1984) 75–87. [15] S. Jahangirian, C.S. McEnally, A. Gomez, Experimental study of ethylene counterflow diffusion flames perturbed by trace amounts of jet fuel and jet fuel surrogates under incipiently sooting conditions, Combust. Flame 156 (2009) 1799–1809. [16] F. Carbone, A. Gomez, The structure of toluene-doped counterflow gaseous diffusion flames, Combust. Flame 159 (2012) 3040–3055. [17] C.S. McEnally, L.D. Pfefferle, Improved sooting tendency measurements for aromatic hydrocarbons and their implications for naphthalene formation pathways, Combust. Flame 148 (2007) 210–222. [18] H. Anderson, C.S. McEnally, L.D. Pfefferle, Experimental study of naphthalene formation pathways in non-premixed methane flames doped with alkylbenzenes, Proc. Combust. Inst. 28 (20 0 0) 2577–2583. [19] M.L. Botero, S. Mosbach, M. Kraft, Sooting tendency and particle size distributions of n-heptane/toluene mixtures burned in a wick-fed diffusion flame, Fuel 169 (2016) 111–119. [20] M. Alfè, B. Apicella, J.N. Rouzaud, A. Tregrossi, A. Ciajolo, The effect of temperature on soot properties in premixed methane flames, Combust. Flame 157 (2010) 1959–1965. [21] C. Russo, M. Alfè, J.-N. Rouzaud, F. Stanzione, A. Tregrossi, A. Ciajolo, Probing structures of soot formed in premixed flames of methane, ethylene and benzene, Proc. Combust. Inst. 34 (2013) 1885–1892. [22] C. Russo, A. Tregrossi, A. Ciajolo, Dehydrogenation and growth of soot in premixed flames, Proc. Combust. Inst. 35 (2015) 1803–1809. [23] R.L. Vander Wal, A.J. Tomasek, Soot oxidation: dependence upon initial nanostructure, Combust. Flame 134 (2003) 1–9.

C. Russo et al. / Combustion and Flame 190 (2018) 252–259 [24] M. Alfè, B. Apicella, R. Barbella, A. Tregrossi, A. Ciajolo, Distribution of soot molecular weight/size along premixed flames as inferred by size exclusion chromatography, Energy Fuels 21 (2007) 136–140. [25] B. Apicella, A. Ciajolo, R. Barbella, A. Tregrossi, T.J. Morgan, A.A. Herod, R. Kandiyoti, Size exclusion chromatography of particulate produced in fuel-rich combustion of different fuels, Energy Fuels 17 (2003) 565–570. [26] M. Alfè, B. Apicella, R. Barbella, J.N. Rouzaud, A. Tregrossi, A. Ciajolo, Structure-property relationship in nanostructures of young and mature soot in premixed flames, Proc. Combust. Inst. 32 (2009) 697–704. [27] K. Yehliu, R.L. Vander Wal, O. Armas, A.L. Boehman, Impact of fuel formulation on the nanostructure and reactivity of diesel soot, Combust. Flame 159 (2012) 3597–3606. [28] C. Russo, F. Stanzione, R. Barbella, A. Tregrossi, A. Ciajolo, The characteristics of soot formed in premixed flames by different fuels, Chem. Eng. Trans. 22 (2010) 41–46. [29] R.M. Fristrom, A.A. Westenberg, Flame structure, McGraw-Hill„ New York, USA, 1965, p. 151. [30] C. Russo, F. Stanzione, A. Tregrossi, A. Ciajolo, Infrared spectroscopy of some carbon-based materials relevant in combustion: Qualitative and quantitative analysis of hydrogen, Carbon 74 (2014) 127–138. [31] C. Saggese, A. Frassoldati, A. Cuoci, T. Faravelli, E. Ranzi, A wide range kinetic modeling study of pyrolysis and oxidation of benzene, Combust. Flame 160 (2013) 1168–1190. [32] M.U. Alzueta, P. Glarborg, K. Dam-Johansen, Experimental and kinetic modeling study of the oxidation of benzene, Int. J. Chem. Kinet. 32 (20 0 0) 498–522. [33] H. Richter, J.B. Howard, Formation and consumption of single-ring aromatic hydrocarbons and their precursors in premixed acetylene, ethylene and benzene flames, Phys. Chem. Chem. Phys. 4 (2002) 2038–2055. [34] K. Brezinsky, The high-temperature oxidation of aromatic hydrocarbons, Prog. Energy Combust. Sci. 12 (1986) 1–24. [35] A. Ciajolo, M. Alfè, B. Apicella, R. Barbella, A. Tregrossi, Characterization of carbon particulate matter relevant in combustion, Chem. Eng. Trans. 17 (2009) 99–104. [36] A. Tregrossi, A. Ciajolo, R. Barbella, The combustion of benzene in rich premixed flames at atmospheric pressure, Combust. Flame 117 (1999) 553–561. [37] C. Russo, L. Giarracca, F. Stanzione, B. Apicella, A. Tregrossi, A. Ciajolo, A. D’Anna, M. Sirignano, Modelling analysis of PAH and soot formation measured in a premixed toluene-doped methane flame, 10th Mediterranean Combustion Symposium (2017) paper PFC-12.

259

[38] A. Ciajolo, R. Barbella, A. Tregrossi, L. Bonfanti, Spectroscopic and compositional signatures of PAH-loaded mixtures in the soot inception region of a premixed ethylene flame, Symp. (Int.) Combust. 27 (1998) 1481–1487. [39] A. Ciajolo, B. Apicella, R. Barbella, A. Tregrossi, Correlations of the spectroscopic properties with the chemical composition of flame-formed aromatic mixtures, Combust. Sci. Technol. 153 (20 0 0) 19–32. [40] T.C. Bond, R.W Bergstrom, Light absorption by carbonaceous particles: An investigative review, Aerosol Sci. Technol. 39 (2006) 1–41. [41] C. Russo, A. Ciajolo, Effect of the flame environment on soot nanostructure inferred by Raman spectroscopy at different excitation wavelengths, Combust. Flame 162 (6) (2015) 2431–2441. [42] A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner, U. Pöschl, Raman microspectroscopy of soot and related carbonaceous materials: spectral analysis and structural information, Carbon 43 (2005) 1731–1742 40. [43] B. Dippel, H. Jander, J. Heintzenberg, NIR FT-Raman spectroscopic study of flame soot, Phys. Chem. Chem. Phys. 20 (1999) 4707–4712. [44] A.C. Ferrari, J. Robertson, Interpretation of Raman spectra of disordered and amorphous carbon, Phys. Rev. B: Condens. Matter 61 (20) (20 0 0) 14095–14107. [45] A.C. Ferrari, J. Robertson, Resonant Raman spectroscopy of disordered, amorphous, and diamond like carbon, Phys. Rev. B: Condens. Matter 64 (7) (2001) 754141–7541413. [46] M. Lapuerta, F. Oliva, J.R. Agudelo, J.P. Stitt, Optimization of Raman spectroscopy parameters for characterizing soot from different diesel fuels, Combust. Sci. Technol. 183 (2011) 1203–1220. [47] F. Tuinstra, J.L. Koenig, Raman spectrum of graphite, J. Chem. Phys 53 (1970) 1126–1130. [48] M.M. Lucchese, F. Stavale, E.H. Martins Ferreira, C. Vilani, M.V.O. Moutinho, RodrigoB. Capaz, C.A. Achete, A. Jorio, Quantifying ion-induced defects and Raman relaxation length in graphene, Carbon 48 (2010) 1592–1597. [49] B. Apicella, P. Pré, M. Alfè, A. Ciajolo, V. Gargiulo, C. Russo, A. Tregrossi, D. Deldique, J.N. Rouzaud, Soot nanostructure evolution in premixed flames by High Resolution Electron Transmission Microscopy (HRTEM), Proc. Combust. Inst. 35 (2015) 1895–1902. [50] J.B. Howard, Carbon addition and oxidation reactions in heterogeneous combustion and soot formation, Symp. (Int.) Combust. 23 (1991) 1107–1127.