Characterization of Combustion-Generated Carbonaceous

Article pubs.acs.org/JPCA

Characterization of Combustion-Generated Carbonaceous Nanoparticles by Size-Dependent Ultraviolet Laser Photoionization Mario Commodo,† Lee Anne Sgro,† Patrizia Minutolo,*,† and Andrea D’Anna‡ †

Istituto di Ricerche sulla Combustione, CNR, P.le Tecchio 80, 80125 Napoli, Italy Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale - Università degli Studi di Napoli Federico II, P.le Tecchio 80, 80125 Napoli, Italy

‡

ABSTRACT: Photoelectric charging of particles is a powerful tool for online characterization of submicrometer aerosol particles. Indeed photoionization based techniques have high sensitivity and chemical selectivity. Moreover, they yield information on electronic properties of the material and are sensitive to the state of the surface. In the present study the photoionization charging efficiency, i.e., the ratio between the generated positive ions and the corresponding neutral ones, for different classes of flame-generated carbonaceous nanoparticles was measured. The fifth harmonics of a Nd:YAG laser, 213 nm (5.82 eV), was used as an ionization source for the combustion generated nanoparticles, whereas a differential mobility analyzer (DMA) coupled to a Faraday cup electrometer was used for particle classification and detection. Carbonaceous nanoparticles in the nucleation mode, i.e., sizes ranging from 1 to 10 nm, show a photoionization charging efficiency clearly dependent on the flame conditions. In particular, we observed that the richer the flame is, i.e., the higher the equivalent ratio is, the higher the photon charging efficiency is. We hypothesized that such an increase in the photoionization propensity of the carbonaceous nanoparticles from richer flame condition is associated to the presence within the particles of larger aromatic moieties. The results clearly show that photoionization is a powerful diagnostic tool for the physical-chemical characterization of combustion aerosol, and it may lead to further insights into the soot formation mechanism.

■

optimization of flame synthesis processes for carbon and inorganic or organometallic compounds.6 Particle inception and growth in flames is a very complex topic, which involves gas-phase free radical reactions, particle nucleation through polymerization and/or clustering pathways, particle growth by both heterogeneous gas-to-solid reactions and physical coagulation and/or coalescence processes, particles annealing, and oxidation. The elusive nature of the chemical and physical processes involved in the soot formation mechanism is further amplified by the fact that all of the abovementioned phenomena are strongly dependent on a large variety of parameters related to the combustion conditions such as, for instance, the fuel chemical composition or physical status, the burning configuration, the flame temperature, and the pressure.7−9 Based on both modeling and experimental studies of different laboratory flame systems,10−14 increasing evidence has been reported that high temperature combustion reactions of hydrocarbon fuels, in fuel rich conditions, mainly produce two classes of carbonaceous nanoparticles: nanoparticles of organic carbon and soot particles. They differentiate by both

INTRODUCTION

Particle formation in flames and combustion systems is an ongoing and a very active research topic due to a variety of motivations, which span from environmental impact and health effects to energy conversion optimization and finally to flame synthesis of new nanostructured materials.1,2 Stationary and vehicular combustion of fossil fuels are the most important anthropogenic sources of carbonaceous particulate matter in the atmosphere. Furthermore, aerosols emitted from combustion systems are typically composed of a complex mixture of particles characterized by different sizes, volatilities, reactivities, chemical compositions, and optical properties. In recent years, the understanding of the mechanisms leading to the formation of incipient organic nanoparticles in flame has attracted the interest of the combustion research community not only because of their role as soot precursors but also as a possible constituent of the emitted combustion aerosol. Recent studies, for example, have shown the presence of a considerably large number of organic carbon nanoparticle, with sizes ranging from 1 to 10 nm, also at the exhaust of domestic burners for home heating burning natural gas and at the exhausts of gasoline engines.3−5 Moreover, the enormous attraction and interest toward new and nanostructured material has been the driving force for the © 2013 American Chemical Society

Received: January 30, 2013 Revised: April 5, 2013 Published: April 15, 2013 3980

dx.doi.org/10.1021/jp401061d | J. Phys. Chem. A 2013, 117, 3980−3989

The Journal of Physical Chemistry A

Article

increase when PAHs of three or more rings were adsorbed on the surface.25,26,29 Photoemission experiments with different pure PAH aerosols of different sizes were reported by Niessner30 as well as the implementation of a tunable laser source, within a spectral range of 207.5−241 nm, for photoionization of NaCl and carbon particles coated by different PAH.31 Details of theoretical principles and experimental methods related to aerosol photoemission studies can be found in two exhaustive reviews by Burtscher15 and by Wilson et al.16 Furthermore, more recently, Zhou and Zachariah32 adopted photoelectron spectroscopy to investigate the electronic properties of controlled metal nanoparticles in free flight. By means of a tunable UV light source, the authors were able to measure the size-resolved work function of such nanoparticles either as singlet spheres or as aggregates, showing that the aggregation state had little or no effect on the work function values.32 Following these experimental works on aerosol photoemission, in the present study an UV laser beam was used coupled to a differential mobility analyzer (DMA) to measure the photoionization efficiency of nanoparticles produced in flames. Specifically, ultrafine particles were sampled online from well characterized laboratory laminar flames, and size selected particles were investigated in terms of their propensity to be ionized by interaction with UV photons with energy of 5.82 eV, given by the fifth harmonic of a Nd:YAG laser at the wavelength of 213 nm. The size range of the investigated particles was 2−10 nm. Such particles are, in fact, representative of the organic nanoparticles produced in slightly sooting flames across the soot inception threshold. The aim of this work is to gain further insights into the chemical/physical properties of the organic nanoparticles involved in the soot formation process. Furthermore, this online aerosol based technique offers the advantage, over conventional off-line methods, to gain information on the pristine optical and electronic properties of the particles without interference of substrate interaction effects, condensation of molecules on the particle surface, and/ or particles coagulation-coalescence.

chemical composition and size, resulting in different optical and spectroscopic properties, physical state, chemical affinity, and coagulation efficiency. For instance, organic nanoparticles, as compared to primary soot particles, absorb much less radiation in the visible, possess a less ordered aromatic structure, are smaller, and do not form aggregates at flame temperature since coalescence follows particle−particle collision with the formation of singlet spherical particles.7,10,11 A further difference between these two particle classes, which reflect a structural difference between a compound with molecular-like composition and a solid one, is the response to an intense light beam. Indeed, several studies have shown that flame-formed organic nanoparticles when irradiated by UV radiation emit fluorescence, whereas soot particles are heated by the absorption of an intense UV, visible, or IR laser light resulting in an incandescence emission.12−14 Therefore, to control particle formation, for both emission reduction and/or particle synthesis, the full understanding of the complex chemical−physical processes occurring in the flame reactor is required. However, despite considerable research efforts in this field, a complete understanding of these mechanisms is still yet to be accomplished. Photoelectric charging of particles is a powerful tool for the online characterization of submicrometer aerosol particles.15,16 Photoionization based techniques have high sensitivity and chemical selectivity. Furthermore, photoemission studies yield information on the electronic properties of the investigated compounds, i.e., valence and conduction bands, and are very sensitive to the surface composition of aerosol particles.15,16 In the framework of chemical kinetic studies of the combustion byproduct, vacuum ultraviolet (VUV) singlephoton photoemission spectroscopy has recently attracted a good deal of interest for its potential in discriminating among different molecular isomers.17,18 Particularly relevant are some recent advances in the use of tunable-synchrotron VUV radiation combined with molecular beam mass spectrometry (MBMS) in studies related to gas-phase combustion chemistry.17−21 It is worth noting that in all of the above-mentioned studies relatively high photon energies of the UV light source, typically ranging from 8 to 10 eV, are necessary because of the large ionization potential (IP) of the small gas-phase molecules produced from the fuel-rich combustion reactions.21 Such photon energies require us to operate in high vacuum condition and are therefore suitable in MBMS experiments. By contrast, UV photons, hν = 4−6 eV, are typically sufficient to photoionize nanoparticles and combustion aerosols,15,16 so that they can be used to implement atmospheric pressure aerosol based techniques. The method based on the photoemission from small particles suspended in a gas was first developed at the ETH in Zürich in the early eighties by Schimidt-Ott et al.,22 and later it was extensively investigated and further developed to examine small aerosol particles and to characterize and monitor combustion aerosols emissions.22−32 In the first of a series of studies on aerosol photoemission, Schimidt-Ott et al.22 found a very large enhancement of the photoemission yields for nanoparticles as compared to bulk material. Aerosol photoemission phenomenon was later described as a four steps process: (1) excitation of an electron by photon absorption, (2) escape of the photoexcited electron from the surface barrier potential, (3) escape of the electron from the image charge and Coulomb potential, and (4) back diffusion and electron recapture.23 Photoactivity of the combustion aerosol was also found to

■

EXPERIMENTAL SET-UP AND PROCEDURES Atmospheric pressure ethylene/air flames were stabilized on a McKenna burner, d = 6 cm, with a cold gas velocity of the unburned premixed gases of 10 cm/s. This laboratory flame configuration allows the more suitable combustion environment for experimental and modeling purposes to study flame chemistry since it is designed in a way that all flame parameters: temperature, species concentration, and velocities can be approximated as a function of only one flame parameter, i.e., the distance from the burner surface or height above the burner (HAB), which has a direct correspondence to the flame residence time. Details of burner configurations, gas-phase species concentration, and particles sizes and number concentrations can be found elsewhere.7−11 In the present study, combustion carbonaceous nanoparticles were collected by keeping the height above the burner (HAB) constant and changing the flame fuel/air equivalence ratio, i.e., φ = (C/O)/ (C/O) stoichiometric, from 1.73 (C/O = 0.57) to 2.03 (C/O = 0.67). Changes in flame stoichiometry caused the flame luminosity to move from blue (leaner flames) to slightly yellow (richer flames) appearance. The increase in flame stoichiometry, promotes the formation of different concen3981



Article

products were sampled through a small orifice (ID = 0.3 mm, thickness = 0.5 mm) into a dilution tube probe operated with N2 as the diluent. A critical dilution ratio on the order of 1000 was used as described in more detail in early works.35,36 It is well-known that particles may acquire charges by diffusion charging involving ions formed by chemi-ionization reactions in the flame front or by thermoionization. As a consequence, particles collected from flames (point A in Figure 1) already have a charge distribution.37−39 The particle charge distribution in the flames examined here results from diffusion charging and can be described by the Boltzmann charge fraction distribution evaluated at the local flame temperature where the particles interacted with the chemi-ions.37,38 However, other effects can influence the charge fraction distribution like particle−particle coagulation38 and particle nucleation in the post flame zone, i.e., at higher residence times, where the chemi-ions concentration is drastically reduced, thus generating neutral inception particles.37,38 To perform an aerosol photoionization experiment, it is necessary to know with great accuracy the amount of neutral species at the entrance of the photoionization cell and the charged ones produced by interaction with the light source. In our experiment, neutral particles were not directly measured

trations of gas-phase carbonaceous byproducts like PAHs and, in turn, of organic nanoparticles and soot, which may also differ in terms of their chemical composition, e.g. percentage of aliphatic/aromatic functionalities.33,34 Details of the investigated flame conditions are listed in Table 1. Table 1. Laminar Premixed Ethylene/Air Flame Conditions for Combustion Carbonaceous Nanoparticles Formation and Collectiona flames C/O ratio equivalence ratio (Φ) cold gas velocity (cm/s) sampling probe position (mm) PSD*

F1

F2

F3

F4

F5

0.57 1.73 10 15

0.61 1.85 10 15

0.63 1.91 10 15

0.65 1.97 10 15

0.67 2.03 10 15

M: a

M: a

M: a

B: a, b

B: a, b′

a

Particle size distribution (PSD); (M) monomodal; (B) bimodal; a, b, and b′ refer to the different modes of the particle size distribution reported in Figure 2.

The experimental setup for the online particle sampling and photoionization measurements of size-selected particles is reported in Figure 1. To prevent particle coagulation, flame

Figure 1. Experimental lay-out for flame formed carbonaceous nanoparticles photoionization measurements. 3982



Article

but were derived from the measurement of size-selected charged particles and the known charge distribution. Therefore, to overcome possible uncertainty on charge fraction distribution, the sampled particles, suspended in the N2 flow, were first passed through a bipolar diffusion charger (Am-241 120 MBq), allowing the aerosol particles to attain the well know Fuchs’ steady-state charge distribution (point B in the Figure 1).40 Charged particles were subsequently selected by their electrical mobility in the classifier and counted by the electrometer. From this value, N−Fuchs(D), the amount of total, N(D), and neutral particles, N0Fuchs(D), were derived by using the theoretical charge fraction distribution as reported by Wiedensohler.41,42 It is worth noting that charge fraction distribution reported in the literature refers to air as carrier gas, whereas in this experiment we had to dilute the flame-generated particles in nitrogen to quench combustion reactions. Even though different carrier gas may determine changes in the properties of the charging ions used in the Fuchs’ theory, it was recently verified that the Fuchs’ charging theory is very insensitive on variations in the ion properties.43 In addition we also found no differences in the charging efficiency of flame-sampled aerosol by using different carrier gases such as N2 or CO2. Once the number of neutral particles entering the ionization chamber was evaluated, the photoionization experiment was conducted. All of the collected particles, charged and neutral, were passed through an electrostatic precipitator to allow only neutral particles to enter in the photoionization chamber (point C in Figure 1). Neutral particles entering the photoionization cell were irradiated by the UV laser beam, the fifth harmonic of a Nd:YAG laser (λ0 = 213 nm, hν = 5.82 eV photon energy). The photoionization cell was a 16 cm long, grounded, metal tube with i.d. = 19.5 mm and two quartz windows on the two sides of the cell. The laser was operated at 10 Hz, with pulse duration of 5 ns. After passing the photoionization chamber, the positively charged particles in the aerosol (point D in Figure 1) were classified based on their electrical mobility by means of a differential mobility analyzer (DMA) operated with the central electrode as the negative pole. The DMA system was a TapCon (3/150), also known as Vienna-type DMA. The DMA detector was a Faraday cup electrometer, with sensitivity as low as 1 fA; further details are reported elsewhere.11,38 Once both the number of neutral and photoionized particles were measured, the photoionization charging efficiency (CE) was obtained by the following equation:

CE =

+ NPI

N0

Figure 2. Particles size distribution (PSD) of the investigated laminar premixed ethylene/air flames.

(1)

+ where NPI is the number of the particles charged via photoionization mechanism and N0 is number of neutral particles.

molecular clusters by ion-induced nucleation in the bipolar charger.11,38,43 As a results, the presence of this peak prevents to measure an additional contribution of flame-formed particles in this size region. Therefore the following analysis only consider particles larger than 1.5 nm. The PSD is unimodal for the F1 flame, C/O = 0.57, with the maximum particle number concentration at the diameter of about 2−2.5 nm, in the following named as mode a. When the equivalence ratio is increased, the PSD is initially still unimodal, as for the F2 and F3 flames, with a larger particle number concentration and extending to slightly larger diameters. In fuel richer conditions, the PSD becomes bimodal as is evident for F4 and F5 flames, whose size distribution have a first mode still

■

RESULTS Particle size distributions (PSDs) were measured in laminar premixed flames of ethylene/air with equivalence ratio (Φ) ranging from 1.73, C/O = 0.57, (blue colored flame), up to 2.03, C/O = 0.67, (fairly yellow flame), by positioning the sampling probe at the HAB = 15 mm. The results are reported in Figure 2. It is well-known that size distribution measured using bipolar-diffusion charger presents a subnanometer peak for mobility diameter below 1.5 nm due to the building-up of 3983



Article

investigated carbonaceous nanoparticles. Other examples of the linear trend measured selecting particles at a fixed diameter from different flames, are reported in the Figure 4. For both photoionization mass spectrometry and/or aerosol based experiments, the ionization order, n, is an important and useful parameter to obtain qualitative information from the photoemission phenomenon. The ionization order can be defined as

located at the diameter of 2−2.5 nm (mode a) and a second mode with a maximum positioned in the size range from 4 (mode b) to 12 nm (mode b′), on dependence of the flame C/ O ratio. The particles with a fixed diameter were selected by keeping constant the voltage applied to the classifier of the DMA and the fraction of particles photoionized by the light beam was measured as function of the laser power density, I. A typical result is shown in Figure 3a for the particles with diameter of 2.4 nm, corresponding to the mode “a”, in the F3 flame (C/O = 0.63).

+ PPI ∝ (I )n

(2)

P+PI

where is the number of positive particles generated by photoionization, I is the laser power density, and n is the ionization order. If the photon energy is lower than the characteristic ionization potential of a species, then a multiphoton ionization process is required and the ionization order is higher than one, i.e., n > 1, whereas n = 1 is observed when the photon energy is higher or at least equal to the photothreshold of the particle. The measured particle ionization order is reported in Table 2 for each particle class. Based on the above discussion, it is possible to conclude that all of the analyzed flame-formed nanoparticles have a photothreshold, or ionization potential, lower or at most equal to the energy of the employed photon; thus, IP ≤ 5.82 eV. It is worthwhile noting that the fitting procedure for the particles in the modes b and b′, in the F5 flame, gives values lower than unity, see Table 2. In order to verify the goodness of the fitting procedure we fit the CE vs I values with both a power law function and a linear one and compared the two models by using Akaike information criterion. The result of the test indicated that both functions are equivalent in describing the experimental data so that we cannot exclude that the ionization order is one also for these conditions. Even if unlikely, deviation from n = 1 might be dependent on the higher polydispersity of the F5 flame, which may both cost interfering signals due to multiple photocharging phenomena,47 and promote electron and/or ion recombination because of the higher particle concentration in this flame condition. Beside the order and threshold of the photoionization process, the photoionization yield is a very powerful parameter which may be used as an analytical method to provide a chemical characterization of particles. Photoelectric yield, is defined as the probability of electron emission per incident photon per unit surface area,22,23 but in aerosol based experiments particle photoelectric yield, yparticle, can be also considered.32 This is the probability of electron emission per incident photon per particle and is therefore proportional to the charging efficiency CE divided by the light beam power density.

Figure 3. (a) Photoionization efficiency (CE) vs laser power density (I). (b) ln(CE) vs ln(I) for ionization order measurement. Data refer to the C/O = 0.63 flame (F3).

For low laser power density, usually below 0.5−0.6 MW/ cm2, the observed linear trend is indicative of proportionality between number of photoionized particles and number of incident photons. Such a result is clear evidence of a single photon photoionization process as further demonstrated by the value of the ionization order, which is given by the slope of the experimental data plotted as ln(CE) vs ln(I); for the data in the Figure 3b, it is n = 1.01 ± 0.04 (adjusted R2 = 0.99). At higher laser power densities, the charging efficiency starts to deviate from linearity. This phenomenon might be ascribed to the occurrence of ion recombination and/or electron recapture.44 Also, particle photofragmentation cannot be excluded at high laser power.14,45,46 Linearity of the photoionization CE as a function of the laser power density was observed for all of the

CE ∝ yparticle I

(3)

Thus, the slope of the linear plots in the Figure 4, i.e. CE/I, allows to compare changes in the yparticle of different particles. In Figure 5a, CE/I measured for all the investigated particles is plotted for comparison. As a first results, particle photoelectric yield increases with the particle size, CE/I (mode a) < CE/I (mode b) < CE/I (mode b′). However, it is worth noting that to obtain analytical information on chemical structure by direct comparison of photoelectric yield for nanoparticles with different sizes is very complex and challenging for several reasons. Aerosol photoemission is, in fact, the results of several phenomena. As discussed previously in the introduction, it is 3984



Article

Figure 4. Photoionization charging efficiency of D = 2.4 nm nanoparticles for different flame conditions as function of laser power below the occurrence of photofragmentation and/or saturation.

Table 2. Ionization Order for the Investigated Particles Obtained by Fitting ln(CE) vs ln(I)a flames particle modes

F1

F2

F3

F4

F5

a

0.91 ± 0.07 (0.97)

1.01 ± 0.06 (0.98)

1.01 ± 0.04 (0.99)

1.2 ± 0.2 (0.93) 1.1 ± 0.1 (0.89)

1.0 ± 0.1 (0.90) 0.65 ± 0.05 (0.96) 0.68 ± 0.08 (0.84)

b b′ a

Numbers in parentheses refer to the resulting adjusted R2.

■

DISCUSSION In this work, we have analyzed the photoionization process of three classes of particles that were produced in flames across the soot threshold and selected on the bases of their size. The first one, class a, consist of particles belonging to the first mode in the size distribution, such species can be reasonably considered as inception nuclei. The second class, b, is formed in richer flames (F3 → F4) as the result of growing of inception nuclei by coalescence and gives rise to a second mode in the size distribution. In richer conditions (F4 → F5), further increase in particle growth gives rise to the class b′ which can be identified as primary soot nuclei. The very good linearity between CE and I, measured in our work for all of the particles, employing the fifth harmonic of the Nd:YAG laser, gives information about the positioning of the valence band. For all three classes of investigated species, the

considered a four step process and depends on various parameters including: photon absorption, ionization potential of the material, electron ejection and recapture processes, particle shape and possible effects due to surface absorbed compound.22−31 Therefore, in the present study, we focused the investigation to the comparison of particles with the same size but collected from different flames. As shown in Figure 5 (a and b), increasing the C/O ratio, i.e., moving from flame F1 to F5, a clear increase of CE/I for all the three classes of investigated particles, D = 10.6, 4.9, and 2.4 nm is observed. Particularly interesting is the case of particles in the first mode, d = 2.4, shown in Figure 5b. Such species present a strong increase in CE/I in unimodal flames by increasing the flame C/O ratio (from F1 to F3) while it remains almost constant in bimodal flames F4 and F5. 3985



Article

Figure 5. (a) Photoionization charging efficiency/laser power density (CE/I) for three different particle sizes as function of the different flame conditions; (b) expanded view of data for particles in the mode a (2.4 nm).

Figure 6. Schematic of the flame-formed carbonaceous compounds and the relative electronic levels.

ionization potential, IP, or, in other words, the position of the highest occupied molecular orbital, HOMO level, must be lower or equal to 5.82 eV. This value for the IP of organic nanoparticles is not surprising given that such species are intermediate between gas-phase PAH molecules, whose IPs are typically of the order of 8−10 eV,17,21 and solid-state soot particles, which can be assimilated to graphite-like material, with a photothreshold of about 4.4 eV.15

The photothreshold limit obtained in this work seems to be fairly in agreement to the values previously reported by Grotheer and co-workers48,49 using photoionization mass spectrometry in similar combustion systems, i.e., laminar premixed flames. Consistently with our data, the authors found that particles with sizes smaller than 3.3 nm (900−17 000 Da) can be ionized by a single photon with energy of 6.4 eV. However, they also observed, in richer flame conditions, the 3986



Article

possibility for small particles (2500 Da) to be ionized by a second order process.49 Additional information on the particle energy levels can be derived by combining the HOMO level evaluation to the determination of the optical band gap, Eg, thus obtaining the determination of valence and conduction bands and so a complete description of the electronic properties of the investigated material. Optical band gap of combustion-formed carbonaceous particles has been determined by many authors by measuring the light absorption spectra by in situ measurements or off-line analysis.50−52 Despite differences in the methodology and flame conditions, two distinctive ranges for the optical band gap have been reported. Organic nanoparticles have Eg typically ranging between 2 and 2.5 eV, whereas soot particles are characterized by energy gap lower than about 0.5−1 eV. A sketch representing the energy levels of various combustion products is proposed in Figure 6. In the evolution from molecules to solid soot particles the energy position of HOMO level increases with number of carbon atoms, in the meanwhile the band gap reduces. While the electronic properties of PAH molecules and large soot particles are known with a rather good approximation, those of inception particles are largely unknown. Recent studies have pointed out the possibility that in premixed flames two main opposed chemical pathways are possible in the nucleation process leading to the formation of first carbonaceous nuclei with sizes of only few nanometers but chemically different. Stacks of pericondensed aromatic rings or polymeric-chain of aromatic hydrocarbon are the two proposed chemical structures for the first nuclei-particles of organic carbon formed in flame1 sketched by the two model compounds in Figure 6. Also, the extension of the aromatic moieties forming the organic nanoparticles is expected to vary during the aromatization process leading to soot formation.2,50 Photoionization study may help discriminate among these different structure and add new insights into the mechanism of soot formation. Exhaustive information can be determined by a further development of the experimental set up, for instance by the use of a tunable UV light sources, which would allow a more accurate positioning of the HOMO level, possibly leading to a correlation with differences in the particle structure. A further result of our measurements is that increasing the C/O ratio and particle size the CE/I increases. As mentioned in the results section, the ratio CE/I is proportional to the particles photoelectric yield, see eq 3. To further understand the relevance of this quantity, CE/I, it is necessary to consider that the photoionization is related to the following four step phenomena: (1) photon absorption, (2) electron escape from the surface barrier potential, (3) electron escape from the image charge/Coulomb potential, and (4) electron recapture.23 The dependence on photon absorption is evident in the expression for CE/I in terms of the photoelectric quantum yield, Y, defined as the ratio between the number of ions produced and the number of photons absorbed22,23,32

CE ∝ yparticle ∝ σabsY I

This equation is valid for macroscopic surfaces near the photothreshold, for hν − ϕ ≤ ∼ 1.5 eV and the exponent x is 2 for metals.22,23 The general expression is still valid for small particles or clusters as demonstrated by Schmidt-Ott et al.23 Our measurements indicate an increase of yparticle with size. However, for particles of different size and chemical composition the three quantities, Φ, Y, and σabs may vary considerably so that it is very challenging to interpret the results of ionization yield measurements. This is the case of the nanoparticles we selected from the different modes of the size distribution, which differ also for composition: the very small ones have a molecular-like or cluster-like composition, whereas the larger ones have solid state structure.7 It is interesting to highlight how these results on collected flame-formed nanoparticles closely resemble those previously reported by Niessner.30 The author showed, in fact, experimental evidence of photoelectric yield dependence on molecular size and chemical composition for pure PAH clusters as well as for PAH-coated graphite particles. However differences must be taken into account between these compounds and those collected from high temperature flame environment.1 Further work is necessary to fully understand the changes in particle properties which give rise to this experimental evidence. As already discussed before, a relevant result in this work is that increasing the flame C/O ratio an increase of CE/I is observed for particles with the same size, as shown in Figure 5b. Due to the higher electron delocalization, the larger the aromatic structures are, the lower is the photoionization potential, Φ. Therefore, the data showen in the Figure 5 may be consistent with the assumption that in richer flame conditions, the higher fuel concentration permits the building-up of macromolecular structures, particles in the nucleation mode, characterized by more extended aromatic functional groups. As a result, changes in the flame stoichiometry lead to the inception of organic nanoparticles characterized by different chemical and physical properties. The possibility of using aerosol-based photoionization experiments for the identification of the different chemical structures, e.g., stacks or chains of aromatic hydrocarbons, is part of our ongoing research and will be published in future works.

■

CONCLUSIONS An aerosol based experiment has been designed to measure photoionization properties of size-selected inception and grown particles formed in ethylene/air premixed flames. Particles have been selected by their electrical mobility in order to investigate species in the various modes of the size distribution. Specifically, the inception mode was centered in the 2−2.5 nm size range and the grown ones in the size range 3−6 nm and 6−12 nm. Laser light at 213 nm has been used to photoionize the flame formed nanoparticles. A single photon ionization process was observed for all of the investigated conditions. This is an indication of the fact that these organic nanoparticles possess an ionization threshold below or at least equal to the energy of the employed photons, i.e., 5.82 eV. Also, since the ionization threshold corresponds to the energy of the HOMO level of the compound, this information, if coupled to the measure of the particle optical gap, allows the determination of the electronic properties, valence and conduction levels. The second results obtained in this work is related to the observed increase of the particle photoelectric yield, or CE/I, for size selected nanoparticles with increasing C/O ratio. This

(4)

where σabs is the particle absorption cross section. The others phenomena are explicit in Y as expressed by the FowlerNordheim law

Y ∝ (hν − Φ)x

(5) 3987



Article

(11) Sgro, L. A.; Barone, A. C.; Commodo, M.; D’Alessio, A.; De Filippo, A.; Lanzuolo, G.; Minutolo, P. Measurement of Nanoparticles of Organic Carbon in Non-Sooting Flame Conditions. Proc. Comb. Inst. 2009, 32, 689−696. (12) Commodo, M.; Osler, F.; De Lisio, C.; D’Anna, A.; Minutolo, P. Size Measurements of Fluorescent Carbon Nanoparticles in a Coflowing Laminar Diffusion Flame by Time-Resolved Fluorescence Anisotropy. Combust. Sci. Technol. 2012, 184, 916−928. (13) D’Anna, A.; Commodo, M.; Violi, S.; Allouis, C.; Kent, J. Nano Organic Carbon and Soot in Turbulent Non-Premixed Ethylene Flames. Proc. Comb. Inst. 2007, 31, 621−629. (14) Commodo, M.; Violi, S.; D’Anna, A.; D’Alessio, A.; Allouis, C.; Beretta, F.; Minutolo, P. Soot and Nanoparticle Formation in Laminar and Turbulent Flames. Combust. Sci. Technol. 2007, 179, 387−400. (15) Burtscher, H. Measurement and Characteristics of Combustion Aerosols With Special Consideration of Photoelectric Charging and Charging by Flame Ions. J. Aerosol Sci. 1992, 23, 549−595. (16) Wilson, K. R.; Bluhm, H.; Ahmed, M. Aerosol Photoemission. In Fundamentals and Applications in Aerosol Spectroscopy; Signorell, R., Reid, J. P., Eds.; Taylor & Francis Group: FL, 2011; pp 367−399. (17) Li, Y. Y.; Qi, F. Recent Applications of Synchrotron VUV Photoionization Mass Spectrometry: Insight Into Combustion Chemistry. Acc. Chem. Res. 2010, 43, 68−78. (18) Qi, F. Combustion Chemistry Probed by Synchrotron VUV Photoionization Mass Spectrometry. Proc. Comb. Inst. 2013, 34, 33− 63. (19) Hansen, N.; Cool, T. A.; Westmoreland, P. R.; KohseHöinghaus, K. Recent Contributions of Flame-Sampling MolecularBeam Mass Spectrometry to a Fundamental Understanding of Combustion Chemistry. Prog. Energy Combust. Sci. 2009, 35, 168−191. (20) Taatjes, C. A.; Hansen, N.; Osborn, D. L.; Kohse-Höinghaus, K.; Cool, T. A.; Westmoreland, P. R. Imaging” Combustion Chemistry via Multiplexed Synchrotron-Photoionization Mass Spectrometry. Phys. Chem. Chem. Phys. 2008, 10, 20−34. (21) Yang, B.; Wang, J.; Cool, T. A.; Hansen, N.; Skeen, S.; Osborn, D. L. Absolute Photoionization Cross-Sections of some Combustion Intermediates. Int. J. Mass Spectrom. 2012, 309, 118−128. (22) Schmidt-Ott, A.; Schurtenberger, P.; Siegmann, H. C. Enormous Yield of Photoelectrons from Small Particles. Phys. Rev. Lett. 1980, 45, 1284−1287. (23) Burtscher, H.; Scherrer, L.; Siegmann, H. C.; Schmidt-Ott, A.; Federer, B. Probing Aerosol by Photoelectric Charging. J. Appl. Phys. 1982, 53, 3787−3791. (24) Burtscher, H.; Schmidt-Ott, A. Experiments on Small Particles in Gas Suspension. Surf. Sci. 1985, 156, 735−740. (25) Burtscher, H.; Schmidt-Ott, A. In Situ Measurement of Adsorption and Condensation of a Polyaromatic Hydrocarbon on Ultrafine C Particles by Means of Photoemission. J. Aerosol Sci. 1986, 17, 699−703. (26) Burtscher, H.; Schmidt-Ott, A.; Siegmann, H. C. Monitoring Particulate Emissions from Combustions by Photoemission. Aerosol Sci. Technol. 1988, 8, 125−132. (27) Müller, U.; Burtscher, H.; Schmidt-Ott, A. Photoemission from Small Metal Spheres: A Model Calculation Using an Enhanced ThreeStep Model. Phys. Rev. B 1988, 38, 7814−7816. (28) Burtscher, H.; Niessner, R.; Schmidt-Ott, A. In Situ Surface Analysis of Coated Particles. In Aerosols: Science, Technology, and Industrial Applications of Airborne Particles; Liu, B. Y. H., Pui, D. Y. H., Fissan, H., Eds.; Elsevier: New York, 1984; p 443. (29) Niessner, R.; Wilbring, P. Ultrafine Particles as Trace Catchers for Polycyclic Aromatic Hydrocarbons: The Photoelectric Aerosol Sensor as a Tool for in Situ Sorption and Desorption Studies. Anal. Chem. 1989, 61, 708−714. (30) Niessner, R. The Chemical Response of the Photo-Electric Aerosol Sensor (PAS) to Different Aerosol Systems. J. Aerosol Sci. 1986, 17, 705−714. (31) Niessner, R.; Roberts, W.; Wilbring, P. Laboratory Experiments on the Determination of Polycyclic Aromatic Hydrocarbon Coverage

may indicate a change in the chemical composition of the organic nanoparticles and possibly a larger extension of the aromatic island in the particles formed by richer flame conditions. Furthermore, photoionization yield results to increase by increasing particle size. However, a direct interpretation of the different yparticle for various classes of nanoparticles results to be quite challenging since the following effects should be considered: (1) different absorption cross section of the small particles as compared to larger ones; (2) deviation from the spherical approximation for the smaller particles; (3) larger compounds may have larger internal energy dissipation; (4) the increase of electron recapture for larger particles; and (5) polarization effects due to surface functionalities. Future work is certainly needed to better understand photoemission of combustion formed nanoparticles. However, the reported results reveal the photoionization as a very powerful diagnostic means to gain new information on ultrafine combustion aerosol. Further applications to investigate surface properties of organic and inorganic metal nanoparticles suspended in a gas are also crucial in terms of knowledge of their reactivity and solubility and toxicological effects.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: + 39 081 768 2963. Fax: + 39 081 593 6936. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge the financial support by the Ministero dello Sviluppo Economico within the “accordo di programma MSE-CNR gruppo tematico Carbone Pulito”, 2010. The authors also thank Gabriella Tessitore for her valuable contribution.

■

REFERENCES

(1) Bockhorn, H., D’Anna, A., Sarofim, A. F., Wang, H., Eds.; Combustion Generated Fine Carbonaceus Particles; KIT Scientific Publishing: Karlsruhe, Germany, 2009. (2) Bockhorn, H., Ed. Soot Formation in Combustion: Mechanisms and Models; Springer-Verlag, Berlin, Germany, 1994. (3) Minutolo, P.; D’Anna, A.; Commodo, M.; Pagliara, R.; Toniato, G.; Accordini, C. Emission of Ultrafine Particles from Natural Gas Domestic Burners. Environ. Eng. Sci. 2008, 25, 1357−1363. (4) Wagner, A. Y.; Livbjerg, H.; Kristensen, P. G.; Glarborg, P. Particle Emissions from Domestic Gas Cookers. Combust. Sci. Technol. 2010, 182, 1511−1527. (5) Myung, C. L.; Park, S. Exhaust Nanoparticle Emissions from Internal Combustion Engines: A Review. Int. J. Automot. Techn 2012, 13, 9−22. (6) Strobel, R.; Pratsinis, S. E. Flame Aerosol Synthesis of Smart Nanostructured Materials. J. Mater. Chem. 2007, 17, 4743−4756. (7) D’Anna, A. Combustion-Formed Nanoparticles. Proc. Comb. Inst. 2009, 32, 593−613. (8) Wang, H. Formation of Nascent Soot and Other CondensedPhase Materials in Flames. Proc. Comb. Inst. 2011, 33, 41−67. (9) Desgroux, P.; Mercier, X.; Thomson, K. A. Study of the Formation of Soot and Its Precursors in Flames Using Optical Diagnostics. Proc. Comb. Inst. 2013, 34, 1713−1738. (10) D’Alessio, A.; D’Anna, A.; Gambi, G.; Minutolo, P. The Spectroscopic Characterization of UV Absorbing Nanoparticles in Fuel Rich Soot Forming Flames. J. Aerosol Sci. 1998, 29, 397−409. 3988



Article

of Submicrometer Particles by Laser-Induced Aerosol Photoemission. Anal. Chem. 1989, 61, 320−325. (32) Zhou, L.; Zachariah, M. R. Size Resolved Particle Work Function Measurement of Free Nanoparticles: Aggregates vs. Spheres. Chem. Phys. Lett. 2012, 525−526, 77−81. (33) Minutolo, P.; Rusciano, G.; Sgro, L. A.; Pesce, G.; Sasso, A.; D’Anna, A. Surface Enhanced Raman Spectroscopy (SERS) of Particles Produced in Premixed Flame Across Soot Threshold. Proc. Comb. Inst. 2011, 33, 649−657. (34) Ciajolo, A. Condensed Phases in Soot Formation Process. In Combustion Generated Fine Carbonaceous Particles; Bockhorn, H., D’Anna, A., Sarofim, A. F., Wang, H., Eds.; KIT Scientific Publishing: Karlsruhe, Germany, 2009; pp 333−344. (35) Kasper, M.; Siegmann, K.; Sattler, K. Evaluation of an In Situ Sampling Probe for Its Accuracy in Determining Particle Size Distributions from Flames. J. Aerosol Sci. 1997, 28, 1569−1578. (36) Zhao, B.; Yang, Z.; Wang, J.; Johnston, M. V.; Wang, H. Analysis of Soot Nanoparticles in a Laminar Premixed Ethylene Flame by Scanning Mobility Particle Sizer. Aerosol Sci. Technol. 2003, 37, 611− 620. (37) Burtscher, H.; Reis, A.; Schmidt-Ott, A. Particle Charge in Combustion Aerosols. J. Aerosol Sci. 1986, 17, 47−51. (38) Sgro, L. A.; D’Anna, A.; Minutolo, P. Charge Fraction Distribution of Nucleation Mode Particles: New Insight on the Particle Formation Mechanism. Combust. Flame 2011, 158, 1418− 1425. (39) Mätzing, H.; Baumann, W.; Bockhorn, H.; Paur, H. -R.; Seifert, H. Detection of Electrically Charged Soot Particles in Laminar Premixed Flames. Combust. Flame 2012, 159, 1082−1089. (40) Reischl, G. P.; Mäkelä, J. M.; Karch, R.; Neeid, J. Bipolar Charging of Ultrafine Particles in the Size Range Below 10 nm. J. Aerosol Sci. 1996, 27, 931−949. (41) Wiedensohler, A.; Lütkemeier, E.; Feldpausch, M.; Helsper, C. Investigation of the Bipolar Charge Distribution at Various Gas Conditions. J. Aerosol Sci. 1986, 17, 413−416. (42) Wiedensohler, A. An Approximation of the Bipolar Charge Distribution for Particles in the Submicron Size Range. J. Aerosol Sci. 1988, 19, 387−389. (43) Steiner, G.; Reischl, G. P. The Effect of Carrier Gas Contaminants on the Charging Probability of Aerosols Under Bipolar Charging Conditions. J. Aerosol Sci. 2012, 54, 21−31. (44) Maisels, A.; Jordan, F.; Fissan, H. On the Effect of Charge Recombination on the Aerosol Charge Distribuition in Photocharging Systems. J. Aerosol Sci. 2003, 34, 117−132. (45) Bengtsson, P. -E.; Alden, M. Optical Investigation of LaserProduced C2 in Premixed Sooty Ethylene Flames. Combust. Flame 1990, 80, 322−328. (46) D’Alessio, A.; Gambi, G.; Minutolo, P.; Russo, S.; D’Anna, A. Optical Characterization of Rich Premixed CH4/O2 Flames Across the Soot Formation Threshold. Symp. (Int.) Combust. 1994, 25, 645− 651. (47) Jung, Th.; Burtscher, H.; Schmidt-Ott, A. Multiple Charging of Ultrafine Aerosol Particles by Aerosol Photoemission (APE). J. Aerosol Sci. 1988, 19, 485−490. (48) Happold, J.; Grotheer, H. -H.; Aigner, M. Distinction of Gaseous Soot Precursor Molecules and Soot Precursor Particles Through Photoionization Mass Spectrometry. Rapid Commun. Mass Spectrom. 2007, 21, 1247−1254. (49) Baquet, T. G.; Grotheer, H.-H.; Aigner, M. Simultaneous Detection of Two Types of Soot Precursor Particles Using Photoionization Mass Spectrometry. Rapid Commun. Mass Spectrom. 2007, 21, 4060−4064. (50) Minutolo, P.; Gambi, G.; D’Alessio, A. The Optical Band Gap Model in the Interpretation of the UV-Visible Absorption Spectra of Rich Premixed Flames. Symp. (Int.) Combust. 1996, 26, 951−957. (51) Russo, C.; Stanzione, F.; Ciajolo, A.; Tregrossi, A. Study on the Contribution of Different Molecular Weight Species to the Absorption UV−Visible Spectra of Flame-Formed Carbon Species. Proc. Comb. Inst. 2013, 34, 3661−3668.

(52) Miller, J. H.; Herdman, J. D.; Green, C. D. O.; Webster, E. M. Experimental and Computational Determinations of Optical Band Gaps for PAH and Soot in a N2-Diluted, Ethylene/Air Non-Premixed Flame. Proc. Comb. Inst. 2013, 34, 3669−3675.

3989


Characterization of Combustion-Generated Carbonaceous

Recommend Documents