Evolution of soot size distribution during soot formation and soot oxidation-fragmentation in premixed flames: experimental and modeling study

The goal of this research was to provide better insights into the evolution of particle size distributions (PSDs) during soot formation, oxidation and fragmentation in premixed flames. Soot formation was studied in premixed ethylene/air, ethylene/benzene/air and benzene/air flat flames under different flame temperatures and C/O ratio conditions. The results demonstrated the major differences in the evolution of the PSDs, both measured and modeled, of soot derived from these flames. The model included reaction pathways leading to the formation of nano-sized particles and their coagulation to larger soot particles. The PSDs for these flames, both experimental and modeled, evolved from a single particle mode to a bimodal PSD. An important distinction between these flames was the behavior of the nucleation mode which persisted at all heights above the burner (HAB) for ethylene/air whereas it decreased for ethylene/benzene/air and was greatly suppressed at greater HAB for the pure benzene/air flames. The explanation for the decreased nucleation mode at higher elevations in the benzene flames was mainly associated with the decrease of soot precursors after the main oxidation zone of the benzene flames.

EVOLUTION OF SOOT SIZE DISTRIBUTION DURING SOOT FORMATION AND SOOT OXIDATION-FRAGMENTATION IN PREMIXED FLAMES: EXPERIMENTAL AND MODELING STUDY by Carlos Andres Echavarria A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemical Engineering The University of Utah August 2010 Copyright © Carlos Andres Echavarria 2010 All Rights Reserved STATEMENT OF DISSERTATION APPROVAL The dissertation of has been approved by the following supervisory committee members: , Chair Date Approved , Member Date Approved , Member Date Approved , Member Date Approved , Member Date Approved and by , Chair of the Department of and by Charles A. Wight, Dean of The Graduate School. Carlos Andres Echavarria JoAnn S. Lighty 4/30/2010 Adel F. Sarofim 4/30/2010 Ronald Pugmire 4/30/2010 Eric Eddings 4/30/2010 David Lignell 4/30/2010 JoAnn S. Lighty Chemical Engineering ABSTRACT The goal of this research was to provide better insights into the evolution of particle size distributions (PSDs) during soot formation, oxidation and fragmentation in premixed flames. Soot formation was studied in premixed ethylene/air, ethylene/benzene/air and benzene/air flat flames under different flame temperatures and C/O ratio conditions. The results demonstrated the major differences in the evolution of the PSDs, both measured and modeled, of soot derived from these flames. The model included reaction pathways leading to the formation of nano-sized particles and their coagulation to larger soot particles. The PSDs for these flames, both experimental and modeled, evolved from a single particle mode to a bimodal PSD. An important distinction between these flames was the behavior of the nucleation mode which persisted at all heights above the burner (HAB) for ethylene/air whereas it decreased for ethylene/benzene/air and was greatly suppressed at greater HAB for the pure benzene/air flames. The explanation for the decreased nucleation mode at higher elevations in the benzene flames was mainly associated with the decrease of soot precursors after the main oxidation zone of the benzene flames. Soot oxidation via O2 and OH* studies were carried out experimentally in a two-stage burner under fuel lean and slightly rich conditions. Soot was produced in a first-stage ethylene/air flame, while in a second stage, the soot was oxidized. Results for the leanest ethylene/air flame, showed a decrease in particle mean diameter and an increase in number concentration for ultra fine-sized particles (Dp < 10 nm) with increasing HAB, which indicates fragmentation of the fine-sized particles. At higher HAB, the soot oxidation process was dominated by soot burnout, evidenced by the decrease in number and mass concentration. The fuel rich conditions did not show particle fragmentation and the PSDs were governed by soot burnout. Soot oxidation rates were calculated using experimental and modeled concentrations for O2 and OH*. Oxidation via O2 promoted fragmentation and was favored under fuel lean conditions close to the burner surface. Oxidation via OH* produced faster oxidation rates and it appeared to be the major path for soot burnout under near-stoichiometric and fuel-rich conditions. iv A mi Esposa, mis Padres y mi Hermana. CONTENTS ABSTRACT....................................................................................................................... iii NOMENCLATURE. ......................................................................................................... ix LIST OF ABBREVIATIONS........................................................................................... xii ACKNOWLEDGEMENTS............................................................................................. xiv Chapter 1 INTRODUCTION ........................................................................................................ 1 1.1 Literature Review.................................................................................................. 2 1.1.1 Soot Formation............................................................................................... 2 1.1.1.1 Soot Molecular Precursors...................................................................... 2 1.1.1.2 Particle Inception .................................................................................... 5 1.1.1.3 Particle Coagulation and Soot Growth. .................................................. 6 1.1.1.4 Particle Agglomeration .......................................................................6 1.1.1.5 Oxidation.............................................................................................7 1.2 Aliphatic Versus Aromatic Flames....................................................................... 8 1.3 Soot Oxidation and Fragmentation ..................................................................... 10 1.4 Analysis and Presentation of Size Distribution Data.......................................... 15 1.5 References........................................................................................................... 18 2 OBJECTIVES AND APPROACH............................................................................. 26 3 EXPERIMENTAL METHODS AND TECHNIQUES.............................................. 27 3.1 Premixed Flat Flame Burner............................................................................... 27 3.2 Methods and Techniques Used in the Flat Flame Burner................................... 30 3.2.1 Temperature Profile Measurements............................................................. 30 3.2.2 Particle Size and Concentration ................................................................... 30 3.2.3 Importance of Dilution During the Sampling of Nano-Particles from Flames ................................................................................................. 33 3.2.4 Chemical Analysis ....................................................................................... 34 3.2.5 Soot Morphology ......................................................................................... 36 3.3 Combustion System, Experimental Methods and Techniques for Studies on Soot Oxidation ................................................................................. 38 3.4 References........................................................................................................... 42 4 MODELING OF SOOT FORMATION AND PARTICLE SIZE DISTRIBUTIONS ...................................................................................................... 45 4.1 Introduction......................................................................................................... 45 4.2 Model of PSD Using Detailed Kinetic Models Coupled to a Discrete Sectional Approach.............................................................................. 47 4.3 References........................................................................................................... 52 5 MODELING AND MEASUREMENTS OF SIZE DISTRIBUTIONS IN PREMIXED ETHYLENE AND BENZENE FLAME............................................... 54 5. 1 Introduction........................................................................................................ 54 5.2 Results and Discussion ....................................................................................... 56 5.2.1 Temperature Profiles and Particle Size Distributions.................................. 56 5.2.2 Measurements and Model Predictions for Ethylene, C/O = 0.69 and 0.89........................................................................................................ 57 5.2.3 Measurements and Model Predictions, Benzene Flame, C/O = 0.69 and 0.89........................................................................................................ 63 5.3 Summary............................................................................................................. 66 5.4 References.......................................................................................................... 69 6 EVOLUTION OF SOOT SIZE DISTRIBUTION IN PREMIXED ETHYLENE/AIR AND ETHYLENE/BENZENE/AIR FLAMES: EXPERIMENTAL AND MODELING STUDY........................................................ 70 6.1 Introduction......................................................................................................... 70 6.2 Results................................................................................................................. 73 6.2.1 Temperature Profiles and Evolution of the Soot Size Distribution ............. 73 6.2.2 Modeling of Soot Size Distribution and Number Concentration................. 77 6.3 Discussion and Summary.................................................................................... 82 6.4 References........................................................................................................... 86 7 STUDIES OF SOOT OXIDATION AND FRAGMENTATION IN A TWO-STAGE BURNER UNDER FUEL-LEAN AND FUEL-RICH CONDITIONS ...................................................................................... 88 7.1 Introduction......................................................................................................... 88 7.2 Results and Discussion ....................................................................................... 91 7.2.1 Temperature Profiles and Evolution of the Soot Size Distribution ............. 91 7.2.2 Concentrations of H2, O2, CO, CO2 and OH*.............................................. 95 7.2.3 Experimental and Predicted Soot Oxidation Rates...................................... 97 7.3 Summary........................................................................................................... 100 7.4 References........................................................................................................ 102 vii 8 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK............. 103 8.1 Ethylene/air Versus Benzene/air and Ethylene/benzene/air Flames................. 104 8.2 Soot Oxidation via O2 and OH* in a Two-Stage Burner.................................. 106 Appendices A SOLID WORKS SCHEMES OF THE FLAT FLAME BURNER .......................... 108 B BENZENE FEEDING SYSTEM USED IN CHAPTER 5....................................... 114 C EXPERIMENTAL SETUP FOR MEASUREMENTS OF FLAME TEMPERATURE AND METHOD TO CORRECT FOR RADIATION LOSSES IN FLAMES....................................................................... 117 D DETAILS ON DILUTION SYSTEM FOR PSD MEASUREMENTS ................... 123 E CORRECTIONS ON PSDS ..................................................................................... 135 F EXAMPLES OF CHEMKIN INPUT FILES FOR FLAMES IN CHAPTERS 5, 6 AND 7 .......................................................................................... 144 G GC MEASUREMENTS OF BENZENE AND PAHS FOR ETHYLENE FLAME (C/O = 0.69) AND BENZENE FLAME (C/O = 0.89) .............................. 153 H SOOT MORPHOLOGY AND NANOSTRUCTURE FOR ETHYLENE/AIR AND BENZENE/AIR FLAMES.............................................................................. 156 I PSDS, NUMBER AND MASS CONCENTRATION, GAS-PHASE CHARACTERIZATION, SOOT OXIDATION RATES, AND SOOT MORPHOLOGY AND NANOSTRUCTURE FOR STUDIES ON SOOT OXIDATION IN THE TWO-STAGE BURNER UNDER "overall = 0.87, 0.94, AND 1.07". ......................................................................... 159 J SOOT SURFACE AREA MEASUREMENTS........................................................ 173 viii NOMENCLATURE T Temperature, [K] amu Atomic mass units A Reactive sites in the edges B Reactive sites in the basal plane W Specific oxidation rate, [g/cm2·sec] PO2 O2 partial pressure, [atm] x Fraction of reactive sites A OH* Fraction of collisions of OH* POH* OH* partial pressure, [atm] A Surface area of soot particles per unit volume of gas, [cm2/cm3] m Total soot mass per unit volume of gas, [grams/cm3] t Residence time of soot into the flame, [sec] Dp Particle size diameter, [nm] N(Dp) Number concentration of particles at the center of certain size interval, [#/cm3] Dm Mobility diameter, [nm] Vug Unburned gas velocity, [cm/sec] CP Carbon percentage as benzene, [%]  Equivalence ratio R Ration of Ws predicted using NSC and Neoh's equations O2 Molecular Oxygen OH* Hydroxyl Radical O* Oxygen Radical Arm Represents aromatic species of high molecular weight Arm* Radical specie derived from Arm Ar Argon Tg Combustion gas temperature, [K] Tj Thermocouple junction temperature, [K] j Junction/bead emissivity  Stefan-Boltzmann constant, [W/m2·K4 ] Nuj Junction Nusselt number dj Junction diameter, [m] kg0 Gas thermal conductivity, [W/m·K2] Pp Pressure drop through the pinhole (dilution probe), [Dynes/cm2] P3 Inlet differential pressure through the eductor, [psig] OD Outside diameter of the dilution probe, [mm] ID Inside diameter of the dilution probe, [mm] Qp Sampled flow through the pinhole, [lpm, (STP)] c Deposition constant dp Pinhole diameter, [cm] l Length of the pinhole orifice, [cm]  Dynamic viscosity, [P] Nt Total number concentration, [#/cm3] x Ci Raw counts provided by the SMPS, [counts] Q Sample flow trhough the SMPS, [lpm] tSMPS Sample time through the SMPS, [sec]  otal sampling efficiency SMPS SMPS efficiency probe Sampling efficiency thought he dilution probe line Sampling efficiency trough the sampling line T Sampling efficiency trough the sampling line (turbulent side) p Sampling efficiency trough the pinhole orifice He Helium Y Molar concentration in gas phase xCO2 Ratio of CO2 adsorbed to sample initial weight PCO2 CO2 partial pressure, [atm] xs Ratio of CO2 adsorbed at saturation to sample initial weight  Affinity coefficient in DP's equation, k Surface property constant in DP's equation, Ps Saturation pressure of CO2 at 297 K, [atm] SA Specific micropore surface, [m2/gsample] Patm Atmospheric pressure xi LIST OF ABBREVIATIONS PAHs Polycyclic Aromatic Hydrocarbons HACA Hydrogen Abstraction and Acetylene Addition TPD Thermocouple Particle Densitometry PSDs Particle Size Distributions DMA Differential Mobility Analyzer NSC Nagle and Strickland-Constable STP Standard Temperature (298.15 K) and Pressure (1atm) C/O Carbon to Oxygen ratio SMPS Scanning Mobility Particle Sizer UCPC Ultrafine Condensation Particle Counter DCM Dichloromethane GC Gas Chromatography MS Mass Spectrometry HP Hewlett Packard TEM Transmission Electron microscopy HR-TEM High Resolution TEM TGA Thermo gravimetric analyzer GRI Gas Research Institute HAB Height Above the Burner PID Proportional-Integral-Derivative NIST National Institute of Standards and Technology RBE Volumetric rate of benzene evaporation DR Dilution Ratio DP Dubinnin-Polayi xiii ACKNOWLEDGEMENTS First at all, I wish to express my sincere appreciation to my advisors Prof. JoAnn S. Lighty and Prof. Adel F. Sarofim for leading and helping me during this journey that started almost 5 years ago. Their guidance and support during my doctorate studies always encouraged me to do things in the best possible way. I would like to thank my committee members Prof. Ronald Pugmire, Prof. Eric Eddings and Dr. David Lignell for their willingness to evaluate this dissertation. I wish to thank Prof. Eddings again because he was the person that initially gave me the opportunity to do my doctorate studies at the University of Utah. I would like to thank Mr. Dana Overacker and Mr. Dave Wagner for the technical support provided. I am also grateful to all members of our research group who provided me with assistance and help during part of my experimental work. I would also like to express my gratitude to Prof. Andrea D'anna for giving me the opportunity to visit his lab. and learn from his modeling work. I want to express my gratitude to Prof. Fanor Mondragon and Dr. William Ciro for pointing me out toward the University of Utah to pursue my doctorate studies. I am deeply thankful to my wife "Paula", my parents and my sister, their love, help, support and understanding give me the strength that I need everyday to continue with my dreams. This work was sponsored by the Nanotechnology and Interdisciplinary Research Team (NIRT) and the Strategic Environmental Research & Development Program (SERDP). Their support was greatly appreciated. xv CHAPTER 1 INTRODUCTION The important problem associated with the emission of carbonaceous nano-particles from liquid and gaseous fuels, requires an interdisciplinary team specializing in combustion, computational chemistry, aerosol dynamics and analytical chemistry in multi-level time and particle size scales, which will increase the overall understanding of nano-particles formation and transformation, specifically addressing the following needs: development of theoretical and experimental methodologies to describe the formation of heavier hydrocarbons and particle inception; modeling and measuring the particle size distributions formed in combustion systems during soot formation and soot oxidation. The present study focused on these needs, by developing and validating experimental and modeling tools in terms of measuring and predicting soot size distribution under conditions that allowed for studying the processes of soot formation and soot oxidation independently. The experimental and computational results provided significant insights into these two processes, which will improve the existing models to predict the evolution of particle in combustion systems. 1.1 Literature Review 1.1.1 Soot Formation Soot formation during combustion continues to be a major subject of experimental and theoretical studies due to its impact on human health, environment and radiation forcing [1-3]. However, soot formation is a complex process involving a great number of chemical and physical steps, which are not completely understood. Fig. 1 illustrates the widely accepted reaction path that leads to the formation of soot in flames. These steps include the molecular precursor formation, particle inception, coagulation and soot growth, particle agglomeration of the primary particles to form chain-like aggregates and oxidation [4-8]. Even though this mechanism has been well accepted, the formation of the first aromatic ring and particle inception have received great interest, particularly, particle inception which involves the transition from gas-phase to solid. 1.1.1.1 Soot Molecular Precursors The formation of molecular precursors is the first step during the soot formation process. This step involves the oxidative pyrolysis of the initial fuel which degrades into small hydrocarbons and free radicals. The reactions between them lead to the formation of molecular species such as ions, polyacetylenes, and polycyclic aromatic hydrocarbons (PAHs) that are regarded as the soot precursors [4-15]. Ionic species were studied by Calcote [11, 12]. He proposed an ionic mechanism where chemi-ions are the precursors on which free radicals, polyacetylenes, and PAH repeatedly add through fast ion-molecule reactions. In Calcote's study, H3O+ was the dominant specie in near stoichiometric and lean flames, while C3H3 + was the dominant ion under fuel rich conditions. 2 Fig. 1. Schematic representation of the soot formation process. CH3 CO C2H3 OH* . In- In- flame processes O2 Theory: Detailed Kinetic Modeling Experiments: Flat Flames Analytical Techniques : Nano-DMA, TEM, HR-TEM, GC, etc. H* H2 Particle inception surface reaction and coagulation agglomeration oxidation Molecular Precursors PAH formation 3 Homann and Wagner [9, 10] considered that polyacetylenes play a significant role during soot precursors. Their mechanism explains the formation of large hydrocarbon molecules due to the reaction of large radicals with themselves and with polyacetylenes, forming even larger molecules. However, Cullis et al. [16] discarded this mechanism based on the fact that polyacetylenes can't growth fast enough to account for the rapid formation of soot. Recent experimental and modeling studies support the hypothesis that the formation of soot from aliphatic fuels generally proceeds through the relatively slow conversion of the aliphatic molecule to aromatic compounds (PAHs) that can rapidly undergo polymerization to soot or growth via hydrogen abstraction and acetylene addition [4, 6-8, 13-16]. For this reason, the formation of aromatic compounds and in particular the step that involves the formation of the first aromatic ring have received great attention, and they are considered as the rate-limiting steps in the reaction sequence that leads to the formation of soot. Several pathways have been proposed for the formation of the first aromatic ring. One of them involves the addition of acetylene to nC4 radicals, followed by an aromatization processes that lead to the formation of benzene or phenyl radical [7]. Miller and Mellius [17] dismissed these reactions based on the fact that nC4 radicals transform rapidly to resonantly stabilized isomers. Along with others, they proposed a mechanism that involves self-combination of propargyl radicals, followed by series of cyclization steps that lead to benzene or phenyl formation. However, McEnally et al. [18] and Sidebothan et al. [19] showed that this mechanism can be highly dependent on fuel and flame type. For instance, in premixed flames the parent fuel is broken down primarily to acetylene in the main reaction zone prior to soot 4 inception which reduce the importance of the C3/C3 mechanism. Other alternative paths for the formation of the first aromatic ring involve the reaction between C5H5 and CH3 radicals to form benzene or the combination of propargyl and acetylene to form cyclopentadienyl radical which reacts rapidly to form benzene [7, 20]. Monte Carlo studies have shown that the reaction C3H3/C2H2 might play an important role on the formation of the first aromatic ring in particular under typical flame conditions T> 1700 K [6, 21]. Following the formation of the first aromatic ring, higher-order aromatics species can be formed via the hydrogen abstraction and acetylene addition (HACA) mechanism, or kinetic pathways involving resonantly stabilized free radicals [6, 7, 13, 14, 22-24]. The HACA mechanism consists of a two step process in which the aromatic molecules are activated by hydrogen atom abstraction, followed by acetylene addition. This process promotes the growth of PAHs and the soot as well. Alternative PAH growth involves the addition of acetylene to chair-like structures via Diels-Alder reaction and migration that can occur in the edges of PAH molecules and on the particles surface [15, 25, 26]. 1.1.1.2 Particle Inception During particle inception, high molecular weight species present in the gas phase turn into solid particles. This process generates spherical-like particles that have a C/H ratio around 2. Experimental measurements suggest that this step occurs at molecular masses between 500-2000 atomic mass units (amu) [4, 27] or 300-700 amu [26]. For larger amu, PAHs can be regarded as solid particles rather than molecules. However, how this transition takes place is not completely understood. Amman et al. [28] proposed a physical condensation model where the partial pressure of the molecular precursors 5 forces the heavier PAHs to condense physically into liquid-like particles. Other theories consider this transition as a result of purely chemical growth. Under this hypothesis, once molecules of PAHs reach a certain size they stick to each other during collisions forming dimmers, which, following similar mechanisms, form trimmers, tetramers and so on. As a result of this polymerization process the first nucleus is formed [7]. 1.1.1.3 Particle Coagulation and Soot Growth. Coagulation and surface reactions takes place simultaneously and contribute to subsequent soot growth. The coagulation is usually expressed as a process where individual particles collide and stick together forming larger particles. Particles can remain together either by sharing a common outer shell generated by deposition or by interaction forces between particles. This process occurs for relatively small particles (up to 10 nm) where growth rates are relatively high [29-31]. The parallel process involves the formation of active sites due to H abstraction and subsequent C2H2 addition from gas-phase. As in the growth of PAHs, surface growth is assumed to be dominated by the HACA mechanism [3, 5-7, 13, 16, 22, 32]. 1.1.1.4 Particle Agglomeration The latest phase in the soot formation process is characterized by a decrease in surface growth and increase in the particle size, this phenomenon is known as superficial aging, and as a consequence, chain like aggregates (see Fig. 2) with a log-normal distribution are formed [4, 6, 33, 34]. Recent studies based on simulation results and mechanistic interpretations attribute this phenomenon to both the decrease in H atom concentration which forces the system to move towards the equilibrium and the decrease 6 Fig. 2. Soot aggregates formed in a premixed ethylene/air flame. in the number of active sites at the particle surface able to maintain additional particle growth [6, 7, 35-37]. 1.1.1.5 Oxidation This step occurs during the entire soot formation process. Starting with the oxidation of the initial fuel and continuing parallel to the reactions that lead to the final product. O2 and OH* radicals are considered the major species that act as soot oxidizers [4, 6, 7, 38-40]. Neoh et al. [41-43] proposed that OH* radical is the major oxidant specie under near stoichiometric and fuel rich conditions because of its higher reactivity under typical flames conditions. OH* might also suppress soot formation via oxidative destruction of precursors. By contrast, Frenklach [6] mentioned that the oxidation of 7 aromatic radicals by O2 seems to be the controlling mechanism at earlier stages in the flame where particle nucleation occurs and a significant reduction of O2 concentration is observed under fuel rich conditions. A more detailed background on soot oxidation is presented in section 1.3. 1.2 Aliphatic Versus Aromatic Flames Experimentally, the soot formation process has been mainly studied at laboratory scale conditions using premixed, diffusion and counter-flow systems [4, 11, 44-50]. Laminar premixed flat flames are the most commonly used because they provide uniformity with respect to flow field and temperature distribution across the burner surface, facilitating experimental measurements and modeling under different fuel/O2 conditions [9, 10, 51-59]. Under fuel-rich conditions, hydrocarbon premixed flat-flames exhibit a characteristic luminosity. Near to the burner surface, a blue region is visible and it is mainly due to the formation of OH* radicals in the flame front. It is followed by an almost transparent region and a yellow-orange luminosity downstream which indicates the presence of soot in the flame [4, 9, 55]. Traditionally, premixed flame properties were characterized by methods such as light scattering, UV absorption and fluorescence, thermocouple particle densitometry (TPD), transmission electron microscopy, etc., [54, 60-65]. Recent studies have added particle size distributions (PSDs) using differential mobility analyzers (DMA) [52, 53, 55, 66-69]. The use of the DMA provides spatially resolved, rapid, and online measurements of nanoparticle size. Most of these methods have focused on characterizing soot and flame properties for premixed aliphatic fuels (ethylene, acetylene) as compared to those of premixed aromatic flames. Some of these 8 studies have proposed that the formation of soot in aromatic flames takes place earlier than in aliphatic; this has been attributed to the larger concentration of aromatics precursors in the main oxidation region. Ciajolo and coworkers [60, 70] have attributed the differences in soot inception mechanism to the differences in the structure of soot obtained from aliphatic versus aromatic fuels. Their results from UV visible spectroscopy analysis showed that benzene flame had a more ordered soot structure and aromatics within a narrower size. On the other hand, complex aliphatic/aromatic structures and a lower degree of order were found for aliphatic flames. Vander Wal and Tomasek [71, 72] demonstrated a variation in the oxidation reactivity for soot derived from pyrolysis of ethanol, benzene and acetylene. Violi [24], who modeled soot particle inception in aromatic(benzene) and aliphatic (acetylene) premixed flames, studied the transformations that occur during the formation of soot precursors. Her model confirmed that fuel composition plays an important role in the pathways for the growth of incipient soot particles. Echavarria et al. [68] studied the effect of the added step from aliphatic to aromatic in the evolution of the PSD using ethylene and benzene premixed flames. The results showed that the most striking difference between PSD in ethylene vs. benzene premixed flames was the persistence of the nucleation mode with height above burner (HAB) for ethylene flames compared to its rapid decline for benzene flames with the same C/O ratios. This distinct behavior in the benzene flame was attributed to soot precursors being consumed in the oxidation zone thereby eliminating the nucleation peak in the upper regions of the flame. Other studies have suggested that high temperatures, not fuel composition, determined the bimodal versus unimodal size distributions in premixed flames [73, 74]. 9 1.3 Soot Oxidation and Fragmentation Most of the research on soot formation in combustion systems is concentrated in the earlier stages of formation. However, oxidation, which takes place parallel to soot formation, defines important physical and chemical characteristics of the final carbonaceous material formed. Due to the great number of steps involved during soot formation and soot oxidation, it is difficult to study them under typical flame conditions. A method to overcome this problem is to characterize the oxidation of soot after it is formed. The oxidation of soot under O2 rich conditions is especially relevant for diesel engines and gas turbine combustors because under normal operational conditions, nearly 90 % of the soot formed is burned out [75]. It is well accepted that the main soot oxidizer species are molecular oxygen (O2), hydroxyl radical (OH*) and oxygen radical (O*). The importance of other species such as H2O, CO, CO2 has been demonstrated to be negligible under typical flame conditions [43]. Studies on carbon oxidation via O2 can be found inflame [4, 39, 41, 44, 71, 75-84] and nonflame [38, 85] environments. Most of these studies have focused their attention on the formulation of semiempirical correlations to improve the understanding of the soot oxidation process via O2 such as those of Lee et al. [39] and Nagle and Strickland- Constable (NSC) [38]. Lee et al. reported a first order reaction expression over the temperature range of 1200 to 1700 K for O2 partial pressures from 4x10-2 to 12 x10-2 atm. NSC proposed a widely used expression for the oxidation of carbonaceous materials via O2. This expression, developed specifically for the oxidation of pyrolytic graphite over a temperature range of 1273 to 2673 K [38, 86] with O2, assumes that the reaction takes 10 place in two active sites of different reactivity (A and B). Sites A correspond to the edge sites and B to reactive sites in the basal plane. 2 2 2 2 (1 ) sec 1 A O B O z O W g k P xk P x cm k P       (1) where W is the specific oxidation rate, kA kB, kz and kT are the kinetic reaction constant, PO2 is the O2 partial pressure and x is the fraction of sites A given by:. 2 1 1 T B O x k k P         (2) A higher reactivity has been associated to the reactive sites located at the edges (A) of the carbonaceous molecules [38, 71, 87], and the lower reactivity of the basal plane sites (B) becomes significant for temperature > 2500 K. B sites can be found either in the carbon surface or as a result of thermal rearrangement of the type A sites. The relative higher reactivity of the A type sites can be significantly reduced for amorphous carbonaceous materials like soot under typical combustion conditions. The NSC equation has been found to fit experimental data over a narrow range of experimental conditions, particularly when graphite-like materials and O2 are the main components in the system [88, 89]. The NSC equation is first order at low O2 partial pressure and approaches to zero at higher pressures [88]. Under typical flame conditions and when the presence of other soot oxidizers such OH*, O* become important, the NSC equation underestimated the soot oxidation rates [41, 42, 90]. Other semiempirical expressions to account for the oxidation of carbonaceous materials with O2 have used an Arrhenius-like form for the 11 equations, and have presented first [91] and fractional [40, 92, 93] reaction orders where a fractional order has been associated to nonhomogeneous reaction surfaces with a wide distribution of adsorption and desorption energies. During soot oxidation via O2 or under a lean environment, experimental and modeling results in premixed flames [42, 43, 75, 76] and other combustion systems [37, 94-96] have shown that particles can break apart, resulting in a decrease in particle mean diameter and an increase in particle number concentration. This phenomena, known as particle fragmentation, has been observed to occur at approximately 80 % soot burnout and attributed to oxygen penetrating the porosity network of soot particles [41, 82], which in turn, cause internal burning and break up of the primary particles. This is consistent with percolation theory where particles can break up due to the loss of connectivity among the phases within the particle [42, 97, 98]. Kerstein and Niksa [97] predicted a critical porosity for fragmentation between 0.2 and 0.9 for various pore shapes and network structures. This critical porosity can be related to the change in surface area which is directly proportional to the soot burnout. At low burnout percentages, fragmentation can involve the bridges between primary particles and at high soot burnout it involves the break up of primary particles. Harris and Maricq [37] incorporated an arbitrary fragmentation rate into the population balance to study the effect of fragmentation in the evolution of the soot size distribution. Their results supported the theory that a better understanding of the fragmentation processes can help to improve the existing models for predicting size distribution and mean properties of particles as they undergo oxidation. When soot or carbonaceous materials are oxidized in rich combustion environments, 12 radicals such OH*, are major soot oxidizers [41, 76, 78]. Fennimore and Jones [78] reported higher soot oxidation rates under low O2 partial pressures and temperatures from 1530 to 1890 K. They attributed the faster oxidation rates, as compared to those predicted by Lee et al. [39] , to carbon oxidation via OH* where 10% of the collision between OH* and soot resulted in carbon removal. Typical collision efficiencies for OH* with carbonaceous materials can be found in the range from 0.1 to 0.28 [41, 78, 82-84, 99- 102]. Other studies have estimated collisions efficiencies via O* radical in the range from 0.1 to 0.5 [80, 90]. Oxidation via O* radical becomes significant for high temperature systems (> 1900 K). Neoh et al. [41] reported a well-known empirical correlation for soot oxidation via OH* based on her experiments under near stoiquiometric and slightly rich conditions in a two-stage burner. Her results supported the importance of OH* as major oxidizer under near stoiquiometric and rich conditions. Neoh's correlation is given by: * * 2 2 1.29 10 sec OH OH W g x P cm T      (3) where OH* is the fraction of collisions of OH* with soot particles that resulted in carbon removal. Neoh found a value of 0.13 for OH* which remained relatively constant with increasing height above burner. POH* is the OH* partial pressure and T is the temperature at a specific position into the flame. In her experiments, Neoh observed fragmentation of particles, as evidenced by the increase in number concentration and decrease in mean diameter for lean flames. She attributed this behavior to internal burning due to O2 penetrating the porous network of the soot. OH* was not expected to penetrate the porous system because of its higher intrinsic reactivity compared to O2. Neoh [42] verified this hypothesis by calculating the effectiveness factor for O2 and OH*. Her results showed 13 that, for soot agglomerates, the effectiveness factor was on average 0.2 for 100 nm diameter particles, which indicates very low penetration of OH* into the pores. In contrast, the effectiveness factor for O2 was on average 0.7 indicating higher penetration for the same particle size, particularity for 100 nm particles diameter with pore radius of 4 nm where the effectiveness factor was around 0.9. Besides temperature, O2 and OH* concentration, soot nanostructure has been shown to play a significant role on carbon oxidation [42, 103-105]. Vander Wal and coworkers [71, 106, 107] observed higher reactivity for soot derived from aromatic flames. They attributed this to the smaller size of graphene layers and higher degree of curvature found in benzene soot. Side edge size carbon was considered more reactive than carbon in the basal plane. Other observations on soot nanostructure showed increasing in the ordering of carbon layers [72, 103, 104, 108, 109], particle shrinkage [110], larger interplanar spacing [111], changes in morphology [95, 112], increase in surface area and microporosity [41, 113, 114], with increase in the extent of burnout during carbon oxidation. Soot oxidation has been commonly studied by measuring changes in particle size and particle concentration using methods such as light scattering and absorption techniques, electron microscopy and electrical mobility analysis. The information extracted from these experimental techniques has been used to calculate soot oxidation rates as follows: 2 1 sec W g dm cm A dt     (4) 14 where A is the surface area of soot particles per unit volume of gas, m is the total soot mass per unit volume of gas, and t is the residence time. 1.4 Analysis and Presentation of Size Distribution Data Typical aerosols derived from combustion sources show from unimodal size distributions to multimodal size distributions (see Fig. 3) [2, 115]. In premixed laminar flames it is common to find bimodal size distributions [53, 68, 116] with the lower mode corresponding to nucleation-size particles and the larger mode as a result of coagulation/agglomeration of the smaller particles. Experimental characterization of the particle size distribution requires isolating a sample, discriminating among the particles on the basis of particle size (Dp), and quantifying the particles in each sizing interval (Dp+dDp). If the number of particles at the center of certain size interval is N(Dp), then the number size distribution is give by: dN(Dp) N(Dp)*dDp (5) or in logarithmic form: ( ) ( )* log p p p dN D  N D d D (6) The logarithmic form is recommended since the particle size diameters in flames can range over several orders of magnitude. The change in particle surface area, volume and mass concentration can easily be characterized by assuming spherical particles either with constant density or accounting for the change in particle density with particle diameter [117, 118]. There are different 15 Fig. 3. Typical aerosols number distribution showing the multimodal nature of ambient aerosols. Adapted from Lighty et al. [2]. 1 10 100 1000 Number Concentration Particle Diameter, nm Fine Particles Ultrafine Particles Nanoparticles Inception Mode Accumulation Mode Dp > 1mn Coarse Mode 16 types of geometric particle size used to characterize nonspherical particles; these are expressed as "equivalent" diameters that can be obtained from several analytical techniques. The most often used are: aerodynamic, optical, diffusion, and electrical mobility diameter [115]. For our interest, the mobility diameter (Dm), provided from nano-DMA measurements, is the equivalent diameter of a spherical particle with the same migration velocity in a constant electric field as the particle of interest [66, 69, 119, 120]. Dm is basically obtained from a force balance between the electrical force of a constant electric field on the net charges of the particle and the drag force exerted over the particle. Zhao et al. [52, 121], and Fernandez de la Mora et al. [122] have presented mathematical expression to correlate the actual geometric diameter to mobility diameter. Dm has been found to be ~5 % higher for 10 nm diameter particles and ~1 % for 50 nm. 17 1.5 References [1] G. L. Borosky, Journal of Organic Chemistry 64 (1999) 7738-7744. [2] J. S. Lighty; J. M. Veranth; A. F. Sarofim, Journal of the Air & Waste Management Association 50 (2000) 1565-1618. [3] A. F. Sarofim; J. S. Lighty; E. G. Eddings, in: Fine Particles: Health Effects, Characterization, Mechanisms of Formation, and Modeling, 224th ACS National Meeting, Boston, Massachusetts, August 18-22, 2002, p. 618-621. [4] H. Bockhorn, Soot formation in combustion: mechanisms and models, Springer- Verlag, New York, N.Y., 1994. [5] M. Frenklach; D. W. Clary; T. Yuan; W. C. Gardiner; S. E. Stein, Combustion Science and Technology 50 (1986) 79 - 115. [6] M. Frenklach, Physical Chemistry Chemical Physics 4 (2002) 2028-2037. [7] M. Frenklach; H. Wang, in: Soot formation in combustion: mechanisms and models, Springer-Verlag: New York, N.Y., 1994, p. 165-190. [8] W. Bartok; A. F. Sarofim, Fossil fuel combustion: a source book, Wiley, New York, 1991. [9] K. H. Homann; H. G. Wagner, Proceedings of the Combustion Institute 11 (1967) 371-379. [10] K. H. Homann; H. G. Wagner, Proc. R. Soc. Lond. A 307 (1968) 141-152. [11] H. F. Calcote, Combustion and Flame 1 (1957) 385. [12] H. F. Calcote, Combustion and Flame 42 (1981) 215. [13] A. D'Anna; A. Violi; A. D'Alessio; A. F. Sarofim, Combustion and Flame 127 (2001) 1995-2003. [14] A. Violi; A. F. Sarofim; T. N. Truong, Combustion and Flame 126 (2001) 1506- 1515. [15] K. Siegmann; K. Sattler, Journal of Chemical Physics 112 (2000) 698-709. [16] C. F. Cullis; I. A. Read; D. L. Trimm, Proceedings of the Combustion Institute 11 (1967) 391. [17] J. A. Miller; C. F. Melius, Combustion and Flame 91 (1992) 21. 18 [18] C. S. Mcenally; L. D. Pfefferle, Combustion and Flame 123 (2000) 344-357. [19] G. W. Sidebotham; I. Glassman, Combustion and Flame 90 (1992) 269. [20] C. F. Melius; M. E. Colvin; N. M. Marinov; W. J. Pit; S. M. Senkan, Proceedings of the Combustion Institute 26 (1996) 685. [21] R. E. Winans; N. A. Tomczyk; J. E. Hunt; M. S. Solum; R. J. Pugmire; Y. J. Jiang; T. H. Fletcher, Energy & Fuels 21 (2007) 2584-2593. [22] A. D'Anna; A. Violi; A. D'Alessio, Combustion and Flame 121 (2000) 418-429. [23] A. D'Anna; A. D'Alessio; J. Kent, Combustion and Flame 125 (2001) 1196-1206. [24] A. Violi, Combustion and Flame 139 (4) (2004) 279-287. [25] K. Siegmann; H. Hepp; K. Sattler, Combustion Science and Technology 109 (1995) 165 - 181. [26] M. Frenklach; L. B. Ebert, The Journal of Physical Chemistry 92 (1988) 561-563. [27] A. D'Anna; A. Rolando; C. Allouis; P. Minutolo; A. D'Alessio, Proceedings of the Combustion Institute 30 (2005) 1449-1456. [28] C. A. Amann; D. C. Siegla, Aerosol Science and Technology 1 (1981) 73 - 101. [29] A. D'Alessio; A. D'Anna; G. Gambi; P. Minutolo, Journal of Aerosol Science 29 (1998) 397-409. [30] A. D'Alessio; A. D'Anna; P. Minutolo; L. A. Sgro; A. Violi, Proceedings of the Combustion Institute 28 (2000) 2547-2554. [31] A. D'Alessio; A. C. Barone; R. Cau; A. D'Anna; P. Minutolo, Proceedings of the Combustion Institute 30 (2005) 2595-2603. [32] J. Y. Hwang; S. H. Chung, Combustion and Flame 125 (2001) 752. [33] R. A. Dobbins; C. M. Megaridis, Langmuir 3 (1987) 254-259. [34] A. P. Palotàs; C. J. Feldermann; A. F. Sarofim; J. B. Vander Sande, Microscopy Research and Technique 33 (1996) 266-278. [35] S. J. Harris; A. M. Weiner, Combustion Science and Technology 32 (1983) 267 - 275. [36] S. J. Harris; M. M. Maricq, Journal of Aerosol Science 32 (2001) 749-764. 19 [37] S. J. Harris; M. M. Maricq, Journal of Aerosol Science 33 (2002) 935-942. [38] J. Nagle; R. F. Strickland-Constable in: Proceedings of the Fifth Carbon Conference, Pergamon Press, 1962; p. 154-164. [39] K. B. Lee; M. W. Thring; J. M. Beer, Combustion and Flame 6 (1962) 137-145 [40] R. H. Hurt; B. S. Haynes, Proceedings of the Combustion Institute, 30 (2005) 2157-2164. [41] K. G. Neoh. Soot burnout in flames, Thesis Sc.D., MIT, 1981. [42] K. G. Neoh; J. B. Howard; A. F. Sarofim, Proceedings of the Combustion Institute 20 (1985) 951-957. [43] K. G. Neoh; J. B. Howard; A. F. Sarofim, in: Particulate Carbon: Formation during Combustion, G. W. S. D. C. Siegla, (Ed.) Plenum: New York, 1981; pp 261-277. [44] J. L. A. Garo, G. Prado, Proceedings of the Combustion Institute, 21 (1986) 1023- 1031. [45] B. A. V. Bennett; C. S. McEnally; L. D. Pfefferle; M. D. Smooke, Combustion and Flame 123 (2000) 522-546. [46] M. Commodo; S. Violi; A. D'Anna; A. D'Alessio; C. Allouis; F. Beretta; P. Minutolo, Combustion Science and Technology 179 (2007) 387-400. [47] A. D'Anna; M. Commodo; M. Sirignano; P. Minutolo; R. Pagliara, Proceedings of the Combustion Institute 32 (2009) 793-801. [48] A. Santamaria; F. Mondragon; E. G. Eddings; A. F. Sarofim, Abstracts of Papers of the American Chemical Society 231 (2006). [49] A. Santamaria; N. Yang; E. Eddings; F. Mondragon, Combustion and Flame 157 (2010) 33-42. [50] A. D'Anna; J. H. Kent; R. J. Santoro, Combustion Science and Technology 179 (2007) 355-369. [51] S. R. Turns, in: An introduction to combustion: concepts and applications, McGraw-Hill, Boston, 2000. [52] B. Zhao; Z. W. Yang; M. V. Johnston; H. Wang; A. S. Wexler; M. Balthasar; M. Kraft, Combustion and Flame 133 (2003) 173-188. 20 [53] M. M. Maricq; S. J. Harris; J. J. Szente, Combustion and Flame 132 (2003) 328- 342. [54] A. D'Anna; A. Alfe; B. Apicella; A. Tregrossi; A. Ciajolo, Energy & Fuels 21 (2007) 2655-2662. [55] A. D'Anna, Proceedings of the Combustion Institute 32 (2009) 593-613. [56] A. D'Anna; M. Sirignano; M. Commodo; R. Pagliara; P. Minutolo, Combustion Science and Technology 180 (2008) 950-958. [57] A. D'Anna; M. Commodo; P. Minutolo, Combustion Science and Technology 180 (2008) 758-766. [58] A. D'Anna, Energy & Fuels 22 (2008) 1610-1619. [59] M. B. Colket; R. J. Hall, in: Soot Formation: Mechanism and Models, Springer- Verlag, New York, N.Y, 1994; p. 442-468. [60] M. Alfe; B. Apicella; R. Barbella; J. N. Rouzaud; A. Tregrossi; A. Ciajolo, Proceedings of the Combustion Institute 32 (2009) 697-704. [61] C. R. Shaddix in: Correcting thermocouple measurements for radiation loss: A critical review, Proceedings of 33rd National Heat and Mass Transfer Conference, Albuquerque, New Mexico, 1999. [62] R. L. Vander Wal, Combustion and Flame 112 (1998) 607-616. [63] R. L. Vander Wal; T. M. Ticich; A. B. Stephens, Combustion and Flame 116 (1999) 291-296. [64] C. S. McEnally; U. O. Koylu; L. D. Pfefferle; D. E. Rosner, Combustion and Flame 109 (1997) 701-720. [65] M. Kurz; A. D'Anna; A. D'Alessio; S. S. Merola; A. Borghese, Particle & Particle Systems Characterization 16 (1999) 77-84. [66] M. M. Maricq, Combustion and Flame 137 (2004) 340-350. [67] S. L. Manzello; D. B. Lenhert; A. Yozgatligil; M. T. Donovan; G. W. Mulholland; M. R. Zachariah; W. Tsang, Proceedings of the Combustion Institute 31 (2007) 675-683. [68] C. A. Echavarria; A. F. Sarofim; J. S. Lighty; A. D'Anna, Proceedings of the Combustion Institute 32 (2009) 705-711. 21 [69] D. R. Chen; D. Y. H. Pui; D. Hummes; H. Fissan; F. R. Quant; G. J. Sem, Journal of Aerosol Science 29 (1998) 497. [70] M. Alfè; B. Apicella; R. Barbella; A. Tregrossi; A. Ciajolo, Proceedings of the Combustion Institute 31 (2007) 585-591. [71] R. L. Vander Wal; A. J. Tomasek, Combustion and Flame 134 (2003) 1. [72] R. L. Vander Wal; A. J. Tomasek, Combustion and Flame 136 (2004) 129. [73] A. D. Abid; E. D. Tolmachoff; D. J. Phares; H. Wang; Y. Liu; A. Laskin, Proceedings of the Combustion Institute 32 (2009) 681-688. [74] B. Zhao; Z. W. Yang; Z. G. Li; M. V. Johnston; H. Wang, Proceedings of the Combustion Institute 30 (2005) 1441-1448. [75] J. S. Lighty; V. Romano; A. F. Sarofim in: Proceedings of the International Workshop on Combustion Generated Fine Carbon Particles, Anacapri, Italy, 2008; H. Bockhorn; A. D'Anna; H. Wang; A. Sarofim, (Eds.) Anacapri, Italy, 2008. [76] C. J. Merrill. The oxidation and fragmentation of soot in a two-stage burner, S. M. thesis, University of Utah, Salt Lake City, 2005. [77] W. Rybak; P. Chambrion; J. Lahaye, Carbon 33 (1995) 259-264. [78] C. P. Fenimore; G. W. Jones, Journal of Physical Chemistry 71 (1967) 593-597. [79] J. F. B. B. R. Stanmore, P. Gilot, Carbon 39 (2001) 2247-2268. [80] G. D.-L. D. Bradley, S. El-Din Habik, E. M. J. Mushi in: The Oxidation of Graphite Powder in Flame Reaction Zones, Proceedings of the Combustion Institute 20 (1984) 931-940. [81] R. A. Dobbins; G. J. Govatzidakis; W. Lu; A. F. Schwartzman; R. A. Fletcher, Combustion Science and Technology 121 (1996) 103-121. [82] A. Garo; G. Prado; J. Lahaye, Combustion and Flame 79 (1990) 226-233 [83] A. Garo, Lahaye, J., Prado, G., Proceedings of the Combustion Institute 21 (1988). [84] F. Xu; A. M. El-Leathy; C. H. Kim; G. M. Faeth, Combustion and Flame 132 (2003) 43-57. [85] D. E. Rosner; H. D. Allendorf, AIAA Journal 6 (1968) 650-654). 22 [86] J. R. Walls; R. F. Strickland-Constable, Carbon 1 (1964) 333-338. [87] T. R. Acharya; D. R. Olander, Carbon 11 (1973) 7-18. [88] C. Park; J. P. Appleton, Combustion and Flame 20 (1973) 369-379. [89] K. J. Higgens; K. Jung; D. B. Kittelson; J. T. Roberts; M. R. Zachariah, Journal of Physical Chemistry 106 (2002) 96-103. [90] R. Vierbaum; P. Roth, Proceedings of the Combustion Institute, 29 (2002) 2423- 2429. [91] J. Neeft; N. P. A.; T. Xander; E. Smakman; M. Makkee; J. A. Moulijn, Fuel 76 (1997). [92] A. Yezerets; N. W. Currier; D. H. Kim; H. A. Eadler; W. S. Epling; C. H. F. Peden, Applied Catalysis B: Environmental 61 (2005) 120-129. [93] B. S. Haynes; T. G. Newbury, Proceedings of the Combustion Institute 28 (2000) 2197-2204. [94] P. Z. Jiang; D. O. Lignell; K. E. Kelly; J. S. Lighty; A. F. Sarofim; C. J. Montgomery, Journal of the Air & Waste Management Association 55 (2005) 437-445. [95] H. Jung; D. B. Kittelson; M. R. Zachariah, Combustion and Flame 136 (2004) 445. [96] P. Jiang. Modeling of Aerosol Dynamics in Flames and Exhaust Plumes. PhD thesis, University of Utah, Salt Lake City, 2003. [97] A. R. Kerstein; S. Niksa Proceedings of the Combustion Institute, 20 (1985) 941- 949. [98] C. A. Sundback; J. M. Beér; A. F. Sarofim; Massachusetts Institute of Technology. Energy Laboratory., Fragmentation behavior of single coal particles in a fluidized bed, Massachusetts Institute of Technology, Energy Laboratory, 1984. [99] F. M. Page; F. Ates, Advances in Chemistry Series 166 (1978) 190-197. [100] P. Roth; O. Brandt; S. V. Gersum Proceedings of the Combustion Institute, 23 (1990) 1485-1491. [101] D. Bradley, Dixon-Lewis, G., Habik, S. El-Din, Mushi, E. M. J., Proceedings of the Combustion Institute 20 (1984). 23 [102] R. P. R. J. Santoro; K. C. Smyth, Combustion and Flame 97 (1994) 125-144. [103] A. P. Palotàs; L. C. Rainey; C. J. Feldermann; A. F. Sarofim; J. B. V. Sande, Microscopy Research and Technique 33 (1996) 266-278. [104] A. P. Palotàs; L. C. Rainey; A. F. Sarofim; J. B. V. Sande; P. Ciambelli, Energy & Fuels 10 (1996) 254-259. [105] L. C. R. A. P. Palotàs, A. F. Sarofim, J. B. Vander Sande, P. Ciambelli, Energy & Fuels 10 (1996) 254-259. [106] R. L. Vander Wal; A. J. Tomasek, Abstracts of Papers of the American Chemical Society 226 (2003) U564-U564. [107] R. L. VanderWal; A. Yezerets; N. W. Currier; D. H. Kim; C. M. Wang, Carbon 45 (2007) 70-77. [108] J. B. Donnet; J. C. Bouland, Rev. Gen. Caoutch. 41 (1964). [109] F. A. Heckman; D. E. Harling, Rubber Chem. Technology 39 (1966.) [110] R. H. Hurt; A. F. Sarofim; J. P. Longwell, Combustion and Flame 95 (1993) 430- 432. [111] C. R. Shaddix, Palotas, Arpad B., Megaridis, Constantine M., Choi, Mun Y., Yang, Nancy Y.C., International Journal of Heat and Mass Transfer 48 (2005) 3604-3614. [112] J. Zhu; K. O. Lee; A. Yozgatligil; M. Y. Choi, Proceedings of the Combustion Institute 30 (2004) 2781-2789. [113] A. W. Kandas. Structural evolution of carbon during oxidation. PhD thesis, MIT, 1997. [114] Z. Du. Kinetic modeling of carbon oxidation. PhD thesis, MIT, 1990. [115] W. C. Hinds, Aerosol technology: properties, behavior, and measurement of airborne particles, J. Wiley, New York, 1982. [116] B. Zhao; Z. W. Yang; J. J. Wang; M. V. Johnston; H. Wang, Aerosol Science and Technology 37 (2003) 611-620. [117] M. M. Maricq; X. Ning, Journal of Aerosol Science 35 (2004) 1251-1274. [118] K. Park, Kittelson, David B., Zachariah, Michael R., McMurry, Peter H., Journal of Nanoparticle Research 6 (2004) 6 24 [119] M. M. Maricq, Combustion and Flame 141 (2005) 406-416. [120] L. A. Sgro; G. Basile; A. C. Barone; A. D'Anna; P. Minutolo; A. Borghese; A. D'Alessio, Chemosphere 51 (2003) 1079-1090. [121] B. Zhao; K. Uchikawa; H. Wang, Proceedings of the Combustion Institute 31 (2007) 851-860. [122] J. Fernández de la Mora; L. de Juan; K. Liedtke; A. Schmidt-Ott, Journal of Aerosol Science 34 (2003) 79. 25 CHAPTER 2 OBJECTIVES AND APPROACH The main objectives of this dissertation were to develop and validate experimental and modeling tools for measuring and predicting soot size distributions under conditions that allowed for studying the processes of soot formation and soot oxidation independently. Experimentally, the work had the following objectives: (1) build and optimize a flat-flame burner for studying the effect of the added step from aliphatic to aromatic formation during soot formation; (2) set-up and optimize chemical and physical characterization methods for measuring soot size distribution, flame temperature, gas-phase composition, soot morphology and nanostructure in premixed ethylene/air, ethylene/benzene/air and benzene/air flames; and, (3) use a two-stage burner to study the effect of flame temperature and O2 and OH* concentration on the evolution of soot size distribution during soot oxidation and fragmentation. The modeling objectives were: (1) predict particle size distributions using a detailed kinetic model coupled to a sectional approach and compare to the experimental results; (2) predict the concentration of OH* radicals in a two-stage burner to calculate soot oxidation rates via OH* and compare to soot oxidation rates via O2. CHAPTER 3 EXPERIMENTAL METHODS AND TECHNIQUES 3.1 Premixed Flat Flame Burner Studies on the evolution of soot size distribution for ethylene/air, benzene/air and ethylene/benzene/air flames (Chapters 5 and 6) were carried out in a flat flame premixed burner (see Fig. 4). This system consisted of a stainless steel chamber (2" ID, Schedule 80, 5" long) where fuel and air were injected and mixed prior to entering the burner. A 1" thick bed of mixing beads inside the burner chamber completed the mixing. The flame was stabilized over a tube bundle (1/16" ID, 1 ¼" long) through which the mixture passed in laminar flow. To shield the premixed flame from atmospheric interference a nitrogen shroud was utilized (N2 average flow rate was in average 9 lpm, STP). A metallic mesh was placed 3.5 cm above the burner surface to stabilize the flat premixed flame. Santoro et al. [1, 2] suggested to keep the height of the stabilization plate constant to avoid changes in flame structure and temperature. Solid works schematics of the flat flame burner are presented in Appendix A. Tables 1 and 2 summarize the conditions for the experiments in Chapters 5 and 6, respectively. Experiments ranging, from lightly (C/O = 0.69) to heavily sooting conditions (C/O = 0.89) were run for atmospheric-pressure ethylene/air, benzene/air and ethylene/benzene/air flames. Air and ethylene (> 99.99%) were fed to the burner using Brooks 5850E mass flow controllers. Benzene (HPLC grade) was delivered via a bubbler Fig. 4. Schematic representations of the flat flame burner and feeding system. Heating Tape Tube Bundle Glass Beads Air Syringe Pump Vaporizer for Liquid Fuels (Benzene) Ethylene N2 Shroud Stabilization mesh 250 oC Temperature Controller 28 Table 1. Experimental conditions for pure ethylene and benzene flames studied in Chapter 5 Flame C2H4 C6H6 O2 N2 C/O *Vug, cm/s (STP) E_0.69 0.1267 0 0.1834 0.6899 0.69 7.76 E_0.89 0.1575 0.00000 0.1769 0.6656 0.89 7.76 B_0.69 0.0000 0.04620 0.2003 0.7535 0.69 7.76 B_0.89 0.0000 0.05880 0.1977 0.7436 0.89 7.76 MOLAR FRACTION * Vug = Unburned gas velocity at STP conditions Table 2. Experimental conditions for ethylene/air and ethylene/air/benzene flames studied in Chapter 6 Flame C2H4 C6H6 O2 N2 * CP % E1 0.1267 0 0.1834 0.6899 0.00 EB1 0.1197 0.00250 0.1843 0.6935 5.90 EB2 0.1060 0.00750 0.1862 0.7003 17.51 EB3 0.0720 0.02000 0.1907 0.7173 45.45 MOLAR FRACTION * CP = Carbon % as Benzene in the fuel mixture, Vug was kept constant at 7.27 cm/s for experiments in Chapter 5; the temperature of the air/benzene feeding line and below the tube bundle was controlled at 250 oC to avoid benzene condensation. Details on the bubbler conditions and setup (saturator) are given in Appendix B. For experiments in Chapter 6, the benzene feed system was optimized by using a commercial (Mesoscopic Devices Inc.) vaporizer coupled to a syringe pump and temperature control system (see Fig. 4). The temperature in the vaporizer was controlled at 250 oC. Both systems provided a uniform fuel flow to the burner. However, the system with the commercial vaporizer allowed a wider range of flow conditions. Temperature, PSDs, chemical analysis, soot morphology and nanostructure measurements were carried out as a function of the height above burner for all conditions studied. Sampling probes and measurements systems were mounted in a x-y translation stage with a relative positional accuracy of ± 0.001 cm. 29 3.2 Methods and Techniques Used in the Flat Flame Burner 3.2.1 Temperature Profile Measurements Temperature profiles along the centerline of the premixed flames were measured using a 0.008" diameter uncoated Type-B thermocouples. Comparisons to uncoated versus coated yielded a difference in temperature of approximately 40ºC which is below the typical uncertainty (±50-100oC) for this thermocouple [3-5]. The thermocouple was inserted into the flame using a fast insertion mechanism (see Appendix C). For each measurement, the transient response of the thermocouple was recorded at a sampling rate of 50 samples per second. Radiation correction was applied using the methodology developed by Rosner et al. [4]. According to Rosner and coworkers when a thermocouple is inserted into a sooty flame, particles will deposit on its junction due to the thermal gradient between the gas and the thermocouple. Thus a correlation of the transient response and thermocouple growth rate can be used not only to infer the local gas temperature but also to estimate the local soot volume fraction. This method does not require assumptions about particle size distribution, particle optical properties or gas uniformity/symmetry. Details on the radiation correction, preliminary measurements of temperature across the flat flame burner and method reproducibility are presented in Appendix C. 3.2.2 Particle Size and Concentration Particle size distributions and soot concentration were measured with a scanning mobility particle sizer (SMPS). This system consisted of a TSI 3080 classifier with a model 3085 nano-differential mobility analyzer (nano-DMA) and a 3025 ultrafine condensation particle counter (UCPC). The SMPS was optimized to operate in the 3-80 30 nm range with a sheath flow of 10 L/min and an aerosol sample flow of 1 L/min. A general overview of the SMPS and sampling system is presented in Fig. 5. In this system, particles from the flame were drawn into a horizontal sampling probe through a small orifice (0.24 mm diameter) and instantaneously diluted using a N2 stream at 30 lpm (STP). A small portion (1 lpm) of the exit stream was sent to the SMPS for particle sizing and number concentration measurements. The large portion of the exit was additionally diluted with air in an eductor and vented through a hood. The eductor not only diluted the large stream but also allowed controlling the pressure drop through the pinhole. In the DMA [6-9], an electrostatic classifier extracts a particular size fraction of particles from the polydisperse stream. The size-selected particle stream enters a Kr-85 neutralizer where the particles are engaged in frequent collisions with bipolar ions. Equilibrium is quickly reached and the particles carry a bipolar charge distribution. This charged aerosol then travels to the differential mobility analyzer. The DMA consists of two concentric cylinders: a collector rod that is maintained at a set negative voltage and an electrically grounded outer cylinder. This cylinder orientation creates an electric field. This electric field leads to the attraction of positively charged particles to the negatively charged collector rod. Particles within the user-defined range of electrical mobility exit the collector rod through a small slit and travel to an Ultrafine Condensation Particle Counter (UCPC) where the particles pass over a heated pool of alcohol and are saturated with alcohol vapor. Next, the aerosol passes into a condenser and is cooled. The alcohol condenses onto the surface of the particles and they reach a size that is optically visible. At this point, the particles can be counted. The flow schematic for the electrostatic classifier is shown in Fig. 6. 31 Fig. 5. SMPS and dilution probe for PSD measurements. Fig. 6. Flow Schematic for the Electrostatic Classifier (TSI Inc.). Polydisperse Neutralizer Aerosol Sheath Display Control Knob Impactor Inlet Filter Pump Excess Exhaust Bypass Monodisperse Filter Aerosol Out N2 30 lpm Pinhole Diameter 0.24 mm P1 P2 P3 Motive air Eductor TSI Classifier 3025 UCPC 1 lpm Vent Nano-DMA Polydisperse Stream Diameter, nm dN/dlogDp, #/cm3 32 3.2.3 Importance of Dilution During the Sampling of Nano-Particles from Flames An essential part in the study of aerosols is the ability to collect representative samples for analysis. These samples must accurately reflect the aerosol concentration and the particle size distribution. During online analysis by the SMPS in a flame, sampling issues range from particle coagulation and agglomeration to diffusive losses in the sampling probe and sampling line. The most common method to minimize these problems during sampling from flames is by diluting the soot-laden combustion gases. The use of a cold dilution gas leads to an immediate quench of particle growth chemistry and minimizes the thermophoretic deposition of soot in the sample line that occurs when high-temperature, soot-laden gases come in contact with a cold surface. Zhao et al. [10] showed that particle diffusion losses and particle coagulation can be minimized by systematically increasing the dilution ratio to a critical value where the particles size distribution function becomes independent of the dilution ratio. Experimentally, this critical dilution ratio was around 104. However, other studies have shown that this value can range from 103 to 104 depending on equivalence ratio and combustion system utilized [11-23]. In this study, we used a dilution system, similar to that of Zhao et al. [15] and Kasper et al. [19] with some changes in probe size (OD = 11mm), pinhole diameter (0.24 mm) (see Fig. 4) and dilution control system, yielding dilution ratios greater than 104, which minimized wall losses and quenched reactions or coagulation that would otherwise occur in the sampling system. Details on the probe optimization such as, critical pinhole diameter, flow conditions, dilution ratio calculations, control of the pressure drop through 33 the pinhole, method reproducibility, and effect of the dilution ratio on PSD are presented in Appendix D. Although, this system minimized diffusion losses and particle coagulation during sampling, corrections are still necessary to ensure that the final result reflects the actual evolution of the particle size in our flames. Corrections for penetration efficiency, into the probe and probe orifice, and diffusion losses during transport were mainly applied following the procedure presented by Minutolo et al. [24-27]. Corrections due to diffusion losses in the SMPS were conducted using the AIM software upgrade [27]. Details on these corrections are presented in Appendix E. Besides the sampling probe effects due to coagulation and particle losses, there are other effects associated to an intrusive technique like the one described here, these effects are related to flame cooling and changes of gas velocity which affect the particle history along the flame. Results of these effects on flame temperature are given in Appendix D. 3.2.4 Chemical Analysis Samples for chemical analysis were isokinetically taken as a function of the height above burner surface for both ethylene and benzene flames. These samples were collected in a 1m filter trap and a dichloromethane (DCM) solvent trap. The sampling system consist of a water-cooled probe (temperature inside the probe was ~ 150 oC) coupled to a filter trap (temperature of the filter was kept around 100 0C to avoid water condensation) followed by a solvent trap (see Fig. 7), the soot extractable material that condenses on filter during the sampling process correspond to organic compounds of high molecular weight also known as PAHs and the soluble material found in the solvent trap represents compounds of low molecular weight (such as benzene) that are not retained by 34 Fig. 7. Sampling system to collect heavy and light PAHs. filters. Fig. 8 shows the methodology used to extract the organic material from filters and solvent trap. Initially, samples from the filter trap were washed with DCM. The obtained solution and the filters were taken to an ultrasonic system for at least 15 minutes, where the inorganic and organic material that remained on the filter were removed. The final solution was filtered and the organic phase which passed through the filter was analyzed by gas chromatography (GC). On the other hand, the sample that was collected in the cold trap was diluted or concentrated and taken directly to GC analysis. A Hewlett Packard model 5890A gas chromatograph was used to analyze the extracts collected from the flames. A 30m x 0.25mm x 0.1μm GC column (HP5, J&W Scientific) was operated at 276 kPa with a linear velocity of 35 cm/s. The GC temperature program consisted of an initial time of 3 minutes at 40°C, a 4°C/min ramp to 300°C, and 15 minutes at 300°C. A blank of DCM was run to isolate the actual sample peaks. Observable compounds in the chromatograms were identified by comparison with standard solutions and further GC/MS analysis. 35 Fig. 8. Experimental methodology to extract and analyze the organic phase from the samples collected in the filter and cold traps. 3.2.5 Soot Morphology Samples for transmission electron microscopy (TEM) were taken using a thermophoretic probe commonly referred to as a "frog tongue" (see Fig. 9). A TEM grid holder was attached to a piston and compressed air at 60 psig was used to quickly insert the TEM grid (200 Mesh) into the flame [28] . Multiple insertions were necessary to get a representative soot sample on the grid. The grid was oriented with the face parallel to the gas flow. In this way, the disturbance of the flame was minimal. Soot deposits on the DCMM Extractable phase or sample from cold trap Sample from filter trap Nonextractable phase Extractable phase Chromatogram DCM GC Vacuum 36 Fig. 9. Thermophoretic sampling system for transmission electron microscopy analysis. N2 Shroud Grid Tube Bundle Carbon Formvar Grid 12 V Battery Fuse Timing Adjustment Switch Ignition Switch Pressure Regulator TEM Sample Holder Piston Feedback 4-Way Valve Cylinder Circuit Timer Relay 37 grid because of the thermophoretic gradient between the cold grid and the hot flame. This technique allows "freezing" some heterogeneous reactions, avoiding changes on the soot morphology after the particles have impacted upon the cold surface. TEM and HR-TEM micrographs samples were performed on transmission electron microscopes FEI, Models Tecnai F30 & F20 EFTEM. 3.3 Combustion System, Experimental Methods and Techniques for Studies on Soot Oxidation A two-stage burner (see Fig. 10), designed after the apparatus of Neoh [29], and with improvements performed by Merrill, Lighty and coworkers [30, 31] was used to generate soot in a fuel-rich premixed flame, which served as the first stage. The soot was then burned in a secondary, premixed burner. In the first-stage burner, air and fuel (ethylene >99.99 %) were added to the bottom of a 2" ID stainless steel chamber under an equivalence ratio 1 = 2.5. Complete mixing occurred over a thick bed of glass beads and the flame was stabilized over a tube bundle through which the mixture passes in laminar flow (see Fig. 4 for details). The secondary burner and pre-oxidation chamber had multiple purposes: efficient mixing of the soot-laden combustion gases with an injection port for air to control the stoichiometry and temperature of the secondary flame; generation of a uniform flow of combustion reactants to the burner; and, oxidation of the soot from the first-stage burner. In a similar way, the secondary flame was stabilized over a tube bundle and shielded from atmospheric interference using a nitrogen shroud. Temperature measurements, PSDs, and gas-phase measurements were carried out along the centerline of the top flame under five different fuel/air conditions (overall = 38 Fig. 10. Schematic of the two-stage oxidation burner. C2H4 Air 1 = 2.5 Tube Bundle Bottom Burner Tube Bundle Top Burner Secondary Air Inlet overall = 0.8, 0.87, 0.94, 1.07 and 1.14 N2 Shroud 39 0.8, 0.87, 0.94, 1.07 and 1.14) for atmospheric-pressure, ethylene-air flames. The overall equivalence ratio, overall is defined as follows:     2 4 1 2 2 4 /( ) / actual overall stoichiometric C H Air Air C H Air    (7) where Air1 and Air2 are the amounts of air injected through the first and secondary ports respectively. The velocity of the unburned gases in the bottom burner was kept constant at 3.4 cm/sec (STP) and the secondary air was changed to reach the desired overall equivalence ratio. The total velocity (based on the total flow of air and C2H4) of the unburned gases for the leanest (overall = 0.8) and richest (overall = 1.14) conditions were 9.54 and 6.82 cm/s (STP), respectively. Temperature profiles were measured using a 0.008" diameter, type-B thermocouple. Experimental setup and radiation corrections were similar to that of Rosner et al. [4] (see section 3.2.1 and Appendix C for details). PSDs were measured with a Scanning Mobility Particle Sizer (SMPS), which consisted of a TSI 3080 Classifier with a Model 3085 Nano DMA and a TSI 3025 Ultrafine Condensation Particle Counter (UCPC). Two different impactor sizes (0.071 and 0.0457 cm) were used to measure PSDs in the range from 3 to 160 nm. The SMPS sampling system, detailed in Section 3.2.3 was used with dilution ratios in the range of 103 to 104 which minimized particles losses and coagulation in the sampling system. Corrections for diffusion losses and coagulation along the sampling line were carried out following the procedure presented by Minutolo et al. [24, 25, 32]. The AIM software accounted for corrections due to diffusion losses in the SMPS. H2, O2, CO and CO2 were isokinetically sampled 40 using a water-cooled probe (0.2-cm ID) (similar to that in Fig. 7) and analyzed by online gas chromatography (VARIAN, CP-4900 Micro GC). The temperature inside the probe was kept between 150 - 200 oC to avoid water condensation. Soot samples for TEM and HR-TEM analysis were taken using the procedure in Section 3.2.5. However, the TEM probe was oriented with the face perpendicular to the flow of combustion products. It did not perturb the flame significantly and allowed to take samples in the exact position. Soot surface area measurements were carried out for the leanest flame (overall = 0.8) at 0 and 2.5 mm above burner surface. Soot samples were collected in a filter trap following the methodology previously described in Section 3.2.4. The surface area of these samples was measured gravimetrically in a thermogravimetric analyzer (Cahn TG-151) by adsorption of CO2 at 297 K [33, 34]. The main advantage of this method was that very small amounts (20-30 mg) of sample can be used due to the sensitivity of the TGA (1g). The volumetric method usually requires sample amounts > 100 mg to obtain reliable results; Details on this methodology are presented in Appendix H. 41 3.4 References [1] A. V. Menon; S.-Y. Lee; M. J. Linevsky; T. A. Litzinger; R. J. Santoro, Proceedings of the Combustion Institute 31 (2007) 593. [2] V. R. Katta; W. M. Roquemore; A. Menon; S. Y. Lee; R. J. Santoro; T. A. Litzinger, Proceedings of the Combustion Institute 32 (2009) 1343-1350. [3] C. R. Shaddix in: Correcting thermocouple measurements for radiation loss: A critical review, Proceedings of 33rd National Heat and Mass Transfer Conference, Albuquerque, New Mexico, 1999. [4] C. S. McEnally; U. O. Koylu; L. D. Pfefferle; D. E. Rosner, Combustion and Flame 109 (1997) 701-720. [5] A. Rolando; A. D'Alessio; A. D'Anna; C. Allouis; F. Beretta; P. Minutolo, Combustion Science and Technology 176 (2004) 945-958. [6] M. M. Maricq; X. Ning, Journal of Aerosol Science 35 (2004) 1251-1274. [7] D. R. Chen; D. Y. H. Pui; D. Hummes; H. Fissan; F. R. Quant; G. J. Sem, Journal of Aerosol Science 29 (1998) 497. [8] M. M. Maricq, Combustion and Flame 141 (2005) 406-416. [9] M. M. Maricq; D. H. Podsiadlik; R. E. Chase, Aerosol Science and Technology 33 (2000) 239-260. [10] B. Zhao; Z. W. Yang; M. V. Johnston; H. Wang; A. S. Wexler; M. Balthasar; M. Kraft, Combustion and Flame 133 (2003) 173-188. [11] A. D'Anna, Proceedings of the Combustion Institute 32 (2009) 593-613. [12] S. L. Manzello; D. B. Lenhert; A. Yozgatligil; M. T. Donovan; G. W. Mulholland; M. R. Zachariah; W. Tsang, Proceedings of the Combustion Institute 31 (2007) 675-683. [13] C. A. Echavarria; A. F. Sarofim; J. S. Lighty; A. D'Anna, Proceedings of the Combustion Institute 32 (2009) 705-711. [14] A. D. Abid; E. D. Tolmachoff; D. J. Phares; H. Wang; Y. Liu; A. Laskin, Proceedings of the Combustion Institute 32 (2009) 681-688. [15] B. Zhao; Z. W. Yang; J. J. Wang; M. V. Johnston; H. Wang, Aerosol Science and Technology 37 (2003) 611-620. 42 [16] J. Fernández de la Mora; L. de Juan; K. Liedtke; A. Schmidt-Ott, Journal of Aerosol Science 34 (2003) 79. [17] A. D. Abid; N. Heinz; E. D. Tolmachoff; D. J. Phares; C. S. Campbell; H. Wang, Combustion and Flame 154 (2008) 775-788. [18] A. De Filippo; M. M. Maricq, Environmental Science & Technology 42 (2008) 7957-7962. [19] M. Kasper; K. Siegmann; K. Sattler, Journal of Aerosol Science 28 (1997) 1569- 1578. [20] M. M. Maricq; R. E. Chase; N. Xu, Abstracts of Papers of the American Chemical Society 222 (2001) U478-U478. [21] M. M. Maricq; D. H. Podsiadlik; R. E. Chase, Environmental Science & Technology 33 (1999) 1618-1626. [22] M. M. Maricq; D. H. Podsiadlik; R. E. Chase, Environmental Science & Technology 33 (1999) 2007-2015. [23] P. Minutolo; A. D'Anna; M. Commodo; R. Pagliara; G. Toniato; C. Accordini, Environmental Engineering Science 25 (2008) 1357-1363. [24] P. Minutolo; A. D'Anna; A. D'Alessio, Combustion and Flame 152 (2008) 287- 292. [25] L. A. Sgro; A. Borghese; L. Speranza; A. C. Barone; P. Minutolo; A. Bruno; A. D'Anna; A. D'Alessio, Environmental Science & Technology 42 (2008) 859-863. [26] L. A. Sgro; A. Simonelli; L. Pascarella; P. Minutolo; D. Guarnieri; N. Sannolo; P. Netti; A. D'Anna, Environmental Science & Technology 43 (2009) 2608-2613. [27] TSI, In: Technical report, TSI Incorporated. (2006). [28] R. A. Dobbins; C. M. Megaridis, Langmuir 3 (1987) 254-259. [29] K. G. Neoh. Soot burnout in flames, Thesis Sc.D, MIT, 1981. [30] J. S. Lighty; V. Romano; A. F. Sarofim in: Proceedings of the International Workshop on Combustion Generated Fine Carbon Particles, Anacapri, Italy, 2008; H. Bockhorn; A. D'Anna; H. Wang; A. Sarofim, (Eds.) Anacapri, Italy, 2008. [31] C. J. Merrill. The oxidation and fragmentation of soot in a two-stage burner, S. M. thesis, University of Utah, Salt Lake City, 2005. 43 [32] L. A. Sgro; A. C. Barone; M. Commodo; A. D'Alessio; A. De Filippo; G. Lanzuolo; P. Minutolo, Proceedings of the Combustion Institute 32 (2009) 689- 696. [33] A. W. Kandas. Structural evolution of carbon during oxidation. PhD thesis, MIT, 1997. [34] Z. Du. Kinetic modeling of carbon oxidation. PhD thesis, MIT, 1990. 44 CHAPTER 4 MODELING OF SOOT FORMATION AND PARTICLE SIZE DISTRIBUTIONS 4.1 Introduction The development of new experimental techniques such as nano-DMA, which allows measurements of size down to 3 nm, has motivated interest in modeling soot particle formation. Previously, models focused on the prediction of concentration profiles for PAHs and soot mean properties. Recently, studies have been able to predict concentration profiles and the evolution of size distribution in combustion systems [1-7]. Previously, experimental PSDs were usually fitted assuming a particle size distribution function (PSDF) either by a log-normal function or the combination of two log-normal functions for unimodal and bimodal size distributions respectively [8, 9]. Afterwards, studies began to use kinetic reaction models to account for the soot mass formed but with limited capabilities to predict particulate size distributions. In these models, soot was assumed to be the mass accumulated in aromatic species above a certain molecular mass. However, the effect of particle growth and coagulation were not accounted for. More recent investigations use detailed kinetic models that include PAHs coagulation, formation of ring-ring aromatic species, and surface growth to determine the final soot concentration and size distribution [1, 10-17]. Two common numerical approaches that couple gas-phase chemistry with aromatic growth mechanism are the stochastic and the discrete sectional method. In the stochastic method [1, 3, 18], an alternative method based on a "stochastic" description of the particle ensemble is proposed to solve the population balance of soot particles as follows: Nt, k R(t) * G(t, k) W(t, k) t in       (8) with the initial condition N(0, k)  N (k)  0 o , where N(t, k) is the number density of particle size k at time t, R(t) is the rate of particle inception, * in  is the size of the smallest particles upon inception, G(t, k) is the rate of coagulation, and W(t, k) is the rate of surface reactions. This method is coupled to a detailed kinetic gas-phase mechanism to simulate soot formation and oxidation in laminar premixed flames, and the results of species concentrations are used as an input for the stochastic approach. In the discrete-sectional approach, an ensemble of aromatic compounds is divided into classes of different molecular mass and all reactions are treated in the form of common gas-phase chemistry by using the compound properties such as mass, number of carbon and hydrogen atoms averaged within each section. The molecular mass distribution of the species is obtained from the calculation and is not hypothesized a priori. Particle evolution is followed by combining the laws of reacting flows with the population balance for suspended particles. Details on this method are given in the next section since this is the methodology used for predicting the evolution of size distribution on premixed ethylene and benzene flames. In addition, this model was used to predict OH* radical concentrations and the concentration of major species such O2, CO2, H2, CO in the top burner of the two stage system used for soot oxidation studies. 46 Another approach commonly used to predict PSDF is the method of moments [10]. Although this method deals with size distributions, it requires longer computational time. 4.2 Model of PSD Using Detailed Kinetic Models Coupled to a Discrete Sectional Approach The details of the model have been reported previously [1, 4, 13, 15, 19-24]. Fig. 11 represents a general overview of the reaction paths used to model soot formation in ethylene and benzene premixed flames. The model includes mechanisms for PAH formation and reaction pathways responsible for nano-sized particle nucleation, i.e. the transition from gas-phase species to nascent particles, and their coagulation to larger soot particles. Initially, fuel pyrolysis was modeled using the well known GRI mechanism for C1 and C2 species and Miller and Mellius mechanism for C3 and C4. Aromatic (benzene) oxidation included submechanism from Emdee et al. and Frank et al., with improvements by Zhang and Mckinoon and, elementary reactions from the Wang and Frenklach oxidation mechanism. Three main reaction paths accounted for the formation of benzene and phenyl radical. Two of them were the reaction between nC4 (n-C4H3 and n-C4H5) radicals and acetylene to form phenyl (R1) and benzene (R2). The last route involved the reaction between 1-methylallennyl and propargyl to form benzyl radicals and their decomposition to benzene (R3). 4 3 2 2 6 5 nC H  C H  C H R1 nC H  C H  C H  H 4 5 2 2 6 6 R2 47 C H  C H  C H CH  H 4 5 3 3 6 5 2 R3 Fig. 11. Schematic representation of the kinetic model used to predict PSDs. The next step, which involves the formation of heavier aromatic species (naphthalene, phenanthrene, and heavier-order rings), was modeled by the HACA mechanism and reaction pathways involving resonantly stabilized free radicals. Three different routes of resonantly stabilized radicals, which are considered as important sources for PAH growth, were included: the self-combination of cyclopentadienyl radicals and the reaction between benzyl and propargyl radicals, both forming naphthalene (R4 and R5). Phenanthrene is formed via indenyl and cyclopentadienyl (R6). 5 5 5 5 10 8 C H C H C H 2H R4 6 5 2 3 3 10 8 C H CH C H C H 2H R5 H H H H H H H H H H H H H H H H H H H H H H H H HH H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H Formation of First ring cyclization polymerization coagulation soot 1 - 2 nm C100 aromatics 3 - 4 nm Formation of Nanoparticles of Organic Carbon 20 nm C10 - C20 PAH 20 - 30 nm 48 indenylC5H5C14H102H R6 Polymerization, or aromatic growth, occurred via radical-molecule addition reactions (R7-R9). Surface growth was mainly accounted for using the HACA mechanism, where acetylene is added on activated molecule sites (R10). * i i 2 Arm H  Arm H R7 * i j i j Arm Arm Arm H     R8 * * i i i j Arm Arm Arm    R9 i 2 2 i Arm C H Arm H R10 i j i j Arm Arm Arm    R11 Reactions of the type R7 to R11, which take place at the surface of particles, accounted for the increase of carbon mass (mainly through PAHs and C2H2 addition) in the system. However, these reactions occur simultaneously with oxidation where mass is removed from the particles through reactions with OH* and O2 (R12, R13). * i j Arm OH Arm HCO R12 2 2 i j Arm O  Arm  CO R13 Further increase in particle size occurred by collision of growing particles (R11). When these particles collide they form new spherical-like structures via coalescence, and subsequently agglomerate forming, chain-like structures. This model did not account for agglomeration; instead a size dependent coagulation rate was used where the sticking 49 probability of the kinetic collision rate was evaluated from the interaction potential between particles [25-27]. A discrete-sectional approach was used for the modeling of the gas-to-particle process; in this method, the ensemble of aromatic compounds is divided into classes of different molecular mass and all reactions are treated in the form of common gas-phase chemistry. Particle size distribution is defined by a range of sections (BINs), each containing a nominal hydrocarbon species in order of increasing atomic mass (see Table 3). Two bins are assigned to each particle size, one for the stable species and the other for the radical. Twenty-five size sections (x2 to allow for radicals) are used in a geometric series with a carbon number ratio of 2 between sections. The carbon number range is from 20 to 4×108 which represents a particle diameter range of 1 - 250 nm. In this approach, the molecular mass distribution of the species is obtained from the calculation and not hypothesized a priori. Premixed flame modeling was performed using the CHEMKIN software package. Examples of input files for modeling flames in Chapters 5, 6 and 7 are presented in the Appendix F. A modified version of the gas-phase interpreter was used allowing the handling of molecules with molecular masses sufficiently large to follow soot particle inception. Particles were assumed to be spherical with a density of 1.2 for PAHs to 1.8 for 20 nm particles. The change in density with mobility diameter was also taken into account [28]. Transport and thermodynamic information were obtained from the CHEMKIN database, Stein et al., Marinov et al. Unavailable data for some species were estimated using Benson's group additivity method and from Wang and Frenklach. 50 Table 3. Definitions of the discretized particle size classes BIN nC nH H/C MW* Dp,nm ** 1 24 12 0.500 300 1.000 2 48 24 0.500 600 1.259 3 96 48 0.500 1200 1.586 4 193 84 0.435 2400 1.998 5 388 144 0.371 4800 2.517 6 778 264 0.339 9600 3.171 7 1560 480 0.308 19200 3.994 8 3124 912 0.292 38400 5.031 9 6256 1728 0.276 76800 6.337 10 12528 3264 0.261 153600 7.982 11 25088 6144 0.245 307200 10.054 12 50240 11520 0.229 614400 12.665 13 100608 21504 0.214 1228800 15.953 14 201472 39936 0.198 2457600 20.095 15 403456 73728 0.183 4915200 25.312 16 807936 135168 0.167 9830400 31.884 17 1617920 245760 0.152 19660800 40.162 18 3239936 442368 0.137 39321600 50.589 19 6483968 835584 0.129 78643200 63.723 20 12972032 1622016 0.125 1.57E+08 80.267 21 25944100 2464690 0.095 3.14E+08 101.024 22 51888250 4773718 0.092 6.27E+08 127.242 23 1.04E+08 9028564 0.087 1.25E+09 160.256 24 2.08E+08 17019674 0.082 2.51E+09 201.837 25 4.15E+08 33212000 0.080 5.02E+09 254.233 * MW = molecular weight ** Dp = particle diameter 51 4.3 References [1] A. D'Anna; A. Violi; A. D'Alessio; A. F. Sarofim, Combustion and Flame 127 (2001) 1995-2003. [2] A. D'Anna, Energy & Fuels 22 (2008) 1610-1619. [3] M. B. Colket; R. J. Hall, in: Soot formation in combustion: mechanisms and models, Springer-Verlag, New York, N.Y, 1994, p. 442-468. [4] A. D'Anna; J. H. Kent, Combustion and Flame 144 (2006) 249-260. [5] M. Balthasar; M. Kraft, Combustion and Flame 133 (2003) 289. [6] J. Appel; H. Bockhorn; M. Frenklach, Combustion and Flame 121 (2000) 122- 136. [7] H. Wang; M. Frenklach, Combustion and Flame 110 (1997) 173-221. [8] B. Zhao; Z. W. Yang; M. V. Johnston; H. Wang; A. S. Wexler; M. Balthasar; M. Kraft, Combustion and Flame 133 (2003) 173-188. [9] W. C. Hinds, in: Aerosol technology: properties, behavior, and measurement of airborne particles, J. Wiley, New York, 1982. [10] M. Frenklach; H. Wang, in: Soot formation in combustion: mechanisms and models, Springer-Verlag, New York, N.Y., 1994; p. 165-190. [11] J. A. Miller; C. F. Melius, Combustion and Flame 91 (1992) 21. [12] R. E. Winans; N. A. Tomczyk; J. E. Hunt; M. S. Solum; R. J. Pugmire; Y. J. Jiang; T. H. Fletcher, Energy & Fuels 21 (2007) 2584-2593. [13] A. D'Anna; A. Violi; A. D'Alessio, Combustion and Flame 121 (2000) 418-429. [14] S. J. Harris; A. M. Weiner, Combustion Science and Technology 32 (1983) 267 - 275. [15] A. D'Anna; A. Violi, Energy & Fuels 19 (2005) 79-86. [16] M. Balthasar; M. Kraft; M. Frenklach, Abstracts of Papers of the American Chemical Society 229 (2005) U860-U860. [17] J. Singh; R. I. A. Patterson; M. Kraft; H. Wang, Combustion and Flame 145 (2006) 117-127. 52 [18] H. Richter; S. Granata; W. H. Green; J. B. Howard, Proceedings of the Combustion Institute 30 (2005) 1397-1405. [19] A. D'Anna; M. Commodo; M. Sirignano; P. Minutolo; R. Pagliara, Proceedings of the Combustion Institute 32 (2009) 793-801. [20] A. D'Anna; M. Sirignano; M. Commodo; R. Pagliara; P. Minutolo, Combustion Science and Technology 180 (2008) 950-958. [21] A. D'Anna; A. D'Alessio; J. Kent, Combustion Science and Technology 174 (2002) 279-294. [22] A. D'Anna; J. H. Kent, Combustion and Flame 132 (2003) 715-722. [23] A. D'anna; G. Mazzotti; J. Kent, Combustion Science and Technology 176 (2004) 753-767. [24] A. D'Anna; J. H. Kent, Combustion and Flame 152 (2008) 573-587. [25] A. D'Alessio; A. D'Anna; P. Minutolo; L. A. Sgro; A. Violi, Proceedings of the Combustion Institute 28 (2000) 2547-2554. [26] A. D'Alessio; A. C. Barone; R. Cau; A. D'Anna; P. Minutolo, Proceedings of the Combustion Institute 30 (2005) 2595-2603. [27] A. D'Anna, Proceedings of the Combustion Institute 32 (2009) 593-613. [28] M. M. Maricq; X. Ning, Journal of Aerosol Science 35 (2004) 1251-1274. 53 CHAPTER 5 MODELING AND MEASUREMENTS OF SIZE DISTRIBUTIONS IN PREMIXED ETHYLENE AND BENZENE FLAMES 5. 1 Introduction Soot formation during combustion continues to be a major subject of experimental and theoretical study due to the impact of soot on human health and radiation forcing. However, soot formation is a complex process involving a great number of chemical and physical steps, and is still incompletely understood. It is widely accepted that the process of soot formation can be described by the steps of molecular precursor formation, particle inception, coagulation and soot growth, particle agglomeration and soot oxidation [1-3]. It is also widely accepted that the formation of soot from aliphatic fuels generally proceeds through the relatively slow conversion of the aliphatic molecule to aromatic compounds that can rapidly undergo polymerization to soot or growth via hydrogen abstraction and acetylene addition. This study focuses on particle inception and growth in premixed flames of ethylene and benzene in order to determine the impact of the added step of aliphatic to aromatic formation on the evolution of the particle size distributions (PSDs) in these two classes of flames. Most previous studies have focused on characterizing soot properties for premixed aliphatic fuels (ethylene, acetylene) as compared to those of premixed benzene flame [4-8]. The studies support the thesis that soot forms early in flames of aromatic compounds relative to those of aliphatic compounds consistent with the added time needed to form the first aromatic ring in an aliphatic flame. In addition, the structure of the compounds formed in aromatic flames are consistent with aromers formed by polymerization of aromatic compounds whereas a greater weighting is given to the polycondensed structures formed by the HACA (Hydrogen abstraction carbon addition) mechanism in flames of aliphatic fuels [5]. Traditionally, temperature profiles, soot volume fraction, particle size distribution, morphology, and structure are commonly characterized by methods such as light scattering, UV absorption and fluorescence, thermocouple particle densitometry (TPD), transmission electron microscopy, etc., [7-10]. Recent studies have added differential mobility analyzers (DMA) of particle size to investigate soot formation in flames [10-12]. The use of DMA provides spatially resolved, rapid, and online measurements of particle size down to 3 nm. This study applies the DMA to examine the evolution of the PSD in ethylene and benzene flames and compares the results with calculations using a sectional method in order to obtain insight on the particle inception and agglomeration. It is, of course, recognized that completely comparable conditions for the ethylene and benzene flames cannot be achieved because of their different sooting tendencies. C/O ratios of 0.69 and 0.89 were selected for both the ethylene and benzene flames. Although attempts were made to obtain comparable temperatures, the greater soot concentrations and temperature gradients in the benzene flames made this difficult. However, compensations for differences in conditions, soot concentrations and temperatures, are taken into account in the model comparisons. 55 Experimental methodologies, conditions (Table 1) and model details were described in Chapters 3 and 4, respectively. Insights into the changes of soot nanostructure and morphology of soot derived from ethylene versus benzene flames is presented in Appendix H. 5.2 Results and Discussion 5.2.1 Temperature Profiles and Particle Size Distributions Fig. 12a (ethylene) and 12b (benzene) present the flame temperatures as functions of height above the burner (HAB) for C/O ratios of 0.69 and 0.89. Temperature peaks in the benzene flame were found to be higher than in the ethylene flame for the same C/O ratio. The high soot concentrations for the benzene flame for C/O = 0.89 were responsible for both the large temperature drop with the increase in C/O from 0.69 and also for the large temperature gradient resulting from radiation cooling. The temperature profiles in Fig. 12 were used as inputs for the model calculations. Fig. 12. Temperatures as a function of HAB in flames of (a) ethylene and (b) benzene for C/O ratios of 0.69 (circles) and 0.89 (squares). 400 600 800 1000 1200 1400 1600 1800 2000 400 600 800 1000 1200 1400 1600 1800 2000 0 5 10 15 C/O = 0.69 C/O = 0.89 Flame Temperature, K HAB, mm a b 0 5 10 15 C/O = 0.69 C/O = 0.89 Flame Temperature, K HAB, mm 56 The size distributions determined by a number of investigators [10-13] using a DMA in laminar premixed ethylene flames under lightly-sooting conditions showed the presence of nucleation and agglomeration modes with the relative magnitude changing with combustion conditions and height above the burner. The results for the present study for a C/O ratio of 0.69 for ethylene and benzene are presented in Fig. 13. At lower HAB, 7 and 10 mm, the size distribution for the ethylene flame is unimodal (the nucleation mode) showing a monotonic decrease in number concentration starting at the detection limit of 3 nm of the SMPS. As the height increases above 10 mm a maximum at around 4 nm is discernible in the size distribution of the nucleation mode. At heights starting at 11 mm a bimodal distribution is observed with the larger sizes corresponding to the agglomeration mode. The number of particles in the nucleation mode however remains high throughout the flame. For the benzene flame the bimodal distribution is evident starting at the lowest HAB (7mm). The number concentration at the lowest size of 3 nm of the nucleation mode decreases with HAB and is difficult to characterize at higher HAB. The decrease in the number of particles in the nucleation mode is a distinguishing feature of the benzene flame. Given that the soot volume fraction is higher for the benzene flame at the same C/O ratio, the discussion from here on will be conducted using the combined results of the model and experiments. 5.2.2 Measurements and Model Predictions for Ethylene, C/O = 0.69 and 0.89 In Fig. 14 comparisons are provided of model and experimental results for a lightly-sooting ethylene flame (soot < 5.0E-7 g/cm3). The model data were shifted downstream by 2 to 6 mm to be compared with experimental data in order to take probe 57 Fig. 13. Particle size distributions in ethylene and benzene flames with a C/O ratio of 0.69 for (a) ethylene and (b) benzene at HAB of 7mm (Δ), 8mm (○), 9mm (▲), 10mm (●), 11mm (□), 12mm (◊), 14mm (x). 108 109 1010 1011 1012 1013 1014 1 10 100 dN/dlogDp, #/cm3 Dp, nm a b 108 109 1010 1011 1012 1013 1014 1 10 100 dN/dlogDp, #/cm3 Dp, nm 58 Fig. 14. Results for ethylene flame, C/O = 0.69: a) Size distributions at HAB of 7mm (-- model, Δ data), 8mm (- - - model, ○ data), 9mm (- - - model, ▲ data), 10mm (- - - model, ● data). b) Size distributions at HAB of 11mm (-- model, □ data), 12mm (- - - model, ◊ data), and 14mm (- - - model, x data). c) Number concentration as a function of HAB for nucleation (- - - model, □ data) and agglomeration (-- model, ■ data). d) Mass concentration as a function of HAB for nucleation (- - - model, □ data) and agglomeration (-- model, ■ data). 108 109 1010 1011 1012 1013 1 10 100 dLogDp, #/cm3 Dp, nm 108 109 1010 1011 1012 1013 1 10 100 dN/dLogDp, #/cm3 Dp, nm 105 107 109 1011 1013 3 5 7 9 11 13 15 N, #/cm3 HAB, mm 10-12 10-11 10-10 10-9 10-8 10-7 10-6 3 5 7 9 11 13 15 Concentration, g/cm3 HAB, mm a b c d dN/59 effects into account (a discussion of the effects of the perturbations by the dilution probe is provided by Zhao et al. [12], and experimental evidence of the probe effects is shown in Appendix D. The amount of shift was determined by the best match between model predictions and data for the lowest HAB. Fig. 14a and 14b provide comparisons of the modeled flame results with experimental data for HAB of 7, 8, 9 ,10 mm and 11, 12, 14 mm, respectively. The nucleation mode dominated the results at the smaller HAB (Fig. 14a) whereas clear evidence of the agglomeration mode was seen higher above the flame. Fair agreement was obtained between the model results (lines) and the experimental data (symbols) although the model showed only a point of inflexion between the nucleation and agglomeration modes while the data show a clear minimum. Particles in size intervals of 3 to 10 nm and greater than 10 nm were integrated to provide rough measures of the contents of the nucleation and agglomeration modes, respectively. The value of 10 nm was picked roughly to separate the nucleation and agglomeration modes for ethylene. The number concentrations in Fig. 14c show how the concentration of particles in the nucleation mode exceeded that in the agglomeration mode for all HAB. The number of particles in the nucleation model rose rapidly and approached a relatively constant level of 1013 cm-3 particles at HAB > 5 mm. These results will be contrasted later with the behavior of particles in the nucleation mode in benzene flames. The particles in the agglomeration mode (dark symbols in Fig. 14c) showed a fairly steady growth in number concentration up to HAB of about 11 mm, and remained fairly constant at higher elevations. In Fig. 14d, the mass concentrations of the nucleation and agglomeration mode are provided as a function of HAB. The mass of particles in the nucleation mode immediately increased after the flame front and approached a constant value (the 60 predicted values passed through a peak). At the end of the flame most of the mass of the particles was in the agglomeration mode and increased with increased HAB, with satisfactory agreement between the modeled and measured values. Fig. 15 provides comparisons of the modeled and experimental results for a slightly richer, C/O = 0.89, ethylene flame, again with the displacement of the model of 2.5-4 mm for the effect of the sampling probe. Fig. 15a and 15b provide comparisons of the size distributions for HAB of 7, 8, 9 mm and 10, 11, 12 mm, respectively. Good agreement is obtained between the nucleation modes in Fig. 15a. Agreement is not as satisfactory for the results shown in Fig. 15b where the experimental data showed a broad agglomeration mode that is not captured by the model results. The model, however, represents the nucleation mode that contained most of the particle number, with the exception of the peaks in the measured distributions. The integrated particle numbers in Fig. 15c show a rapid increase in the particles in the size range 3 nm < Dp < 10 nm for HAB up to 5 mm followed by a leveling off in the number concentration, with fair agreement between data and modeled results. For particles with sizes greater than Dp of 10 nm (the agglomeration mode), the number concentration builds up rapidly for HAB between 7 and 9 mm. The modeled number concentration levels off at larger HAB but the experimental results fell off, most likely as a result of clogging of the nozzle in the sampling probe. The number concentrations in the nucleation mode were consistently higher than those in the agglomeration mode. The comparison of the measured and modeled mass concentrations as a function of HAB (Fig. 15d) shows a major deficiency in the measured mass for the accumulation mode as HAB increases, again believed to be due to a clogging of the probe. 61 Fig. 15. Results for an ethylene flame, C/O = 0.89: a) Size distributions at HAB of 7mm (-- model, Δ data), 8mm (- - - model, ○ data), 9mm (- - - model, ▲ data). b) Size distributions at HAB of 10 mm (-- model, ● data), 11 mm (- - - model, □ data), and 12 mm (- - - model, ◊ data). c) Number concentration as a function of HAB for nucleation (- - - model, □ data) and agglomeration (-- model, ■ data). d) Mass concentration as a function of HAB for nucleation mode (- - - model, □ data) and agglomeration mode (-- model, ■ data). 3 5 7 9 11 13 15 Concentration, g/HAB, mm 107 109 1011 1013 3 5 7 9 11 13 15 N, #/cm3 HAB, mm 108 109 1010 1011 1012 1013 1 10 100 dN/cm3 Dp, nm 108 109 1010 1011 1012 1013 1 10 100 dN/dlogDp, #/cm3 Dp, nm c d a b 10-9 10-8 10-7 10-6 cm3 dlogDp, #/62 5.2.3 Measurements and Model Predictions, Benzene Flame, C/O = 0.69 and 0.89 Comparisons of model predictions and data for the benzene flame with a C/O ratio of 0.69 are presented in Fig. 16. Fig. 16a and 16b provide the PSDs for HAB of 7, 8, 9 mm and 11, 12, 14 mm, respectively. A shift of 6 mm in the modeled data downstream of the burner was needed to obtain reasonable correspondence between data and predictions. In this case, a particle size of approximately 5 nm was used to determine the separation between the nucleation and agglomeration modes. In Fig. 16a, the data at 7 and 8 mm showed a very narrow nucleation mode; otherwise, most of the distribution was dominated by the agglomeration mode. Small nucleation modes are apparent in Fig. 16b, for both measurements and predictions. These results are reflected in Fig. 16c which shows a spike for particles in the nucleation mode around 7 mm HAB falling to negligible levels above 15 mm HAB. By contrast the agglomeration mode particles (bounded by the upper measurement limit of 80 nm) persist throughout the flame. Fig. 16d shows the mass of the agglomeration mode particles dominate at high HAB with fair agreement between the measured and predicted values, yielding soot levels of about 1.E- 6 g/cm3. The observation to be emphasized is the essential disappearance of the mass of the nucleation mode above 11 mm HAB. Similar results are obtained for a benzene flame with a C/O = 0.89, as summarized in Fig. 17. The experimental results for the particle size distributions are shown in Fig. 17a and 17b. The experimental results in Fig. 17a for HAB of 7, 8, and 9 mm showed a narrow nucleation mode only at 7 mm HAB. The predictions best fit the data in Fig. 17a when the modeled data were shifted downstream of the burner by 4 mm. Fig. 17b provides the experimental particle size distribution for HAB values 10, 11, 12, 63 Fig. 16. Results for benzene flame, C/O = 0.69: a) Size distributions at HAB of 7mm (-- model, Δ data), 8mm (- - - model, ○ data), 9mm (- - - model, ▲ data), and 10mm (- - - model, ● data). b) Size distributions at HAB of 11mm (- - - model, □ data), 12mm (- - - model, ◊ data), and 14mm (-- model, x data). c) Number concentration as a function of HAB for nucleation (- - - model, □ data) and agglomeration (-- model, ■ data). d) Mass concentration as a function of HAB for nucleation (- - - model, □ data) and agglomeration (-- model, ■ data). 1 10 100 dN/dLogDp, #/Dp, nm 1 10 100 dN/dLogDp, #/Dp, nm 10-10 10-9 10-8 10-7 10-6 10-5 3 5 7 9 11 13 15 Concentration, g/cm3 HAB, mm a c b d 107 108 109 1010 1011 1012 1013 cm3 107 108 109 1010 1011 1012 1013 cm3 107 109 1011 1013 1015 3 5 7 9 11 13 15 N, #/cm3 HAB, mm 64 Fig. 17. Results for benzene flame, C/O = 0.89: a) Size distributions at HAB of 7mm (-- model, Δ data), 8mm (- - - model, ○ data), and 9mm (- - - model, ▲ data). b) Size distributions at HAB of 10mm (-- model, ● data), 11mm (- - - model, □ data), 13mm (- - - model, ◊ data), and 15mm (- - - model, x data). c) Number concentration as a function of HAB for nucleation (- - - model, □ data) and agglomeration (-- model, ■ data). d) Mass concentration as a function of HAB for nucleation (- - - model, □ data) and agglomeration (-- model, ■ data). 107 108 109 1010 1011 1012 1013 1 10 100 dN/dLogDp, #/cm3 Dp, nm 106 108 1010 1012 1014 3 5 7 9 11 13 15 N, #/cm3 HAB, mm 10-12 10-10 10-8 10-6 10-4 3 5 7 9 11 13 15 Concentration, g/cm3 HAB, mm a c b d 108 109 1010 1011 1012 1013 1 10 100 dN/dLogDp, #/cm3 Dp, nm 65 and 14 with no evidence of a nucleation mode. Again, the experimental results are well fitted by the predictions when the data were shifted towards the burner by 4 mm. The predicted number and mass concentrations for the nucleation mode exceed the measured values which fall to negligible levels at about 15 mm HAB in Fig. 17c and Fig. 17d. This may be due to the loss of smaller particles in the sampling probe. Good agreements are obtained (Fig. 17c and 17d) for the agglomeration mode between measurements and predictions. 5.3 Summary The most striking difference between Figs. 14-17 is the persistence of the nucleation mode with HAB in Figs. 14 and 15 for ethylene flames compared to its rapid decline in Figs. 16 and 17 for benzene flames with the same C/O ratios. Since conditions other than fuel changed between the flames, the question still remains as to whether the difference is accountable by the fuel change or by other flame conditions. Previous studies [10, 13] have suggested that high temperatures, not fuel composition, determined the bimodal versus unimodal size distributions. In order to evaluate the role of flame chemistry the concentration profiles of the soot precursors acetylene and benzene are plotted for the ethylene (C/O=0.69) and benzene (C/O=0.89) flames (Fig. 18) since these two flames had similar temperature distributions. In the benzene flame, benzene and, to a lesser extent, acetylene showed high concentrations near the burner (< 3 mm HAB) which is consistent with the observed location of the nucleation mode in the benzene flame experiments and predictions. By contrast, in the ethylene flame, the acetylene and benzene concentrations persist at even high HAB which is again consistent with the observed persistence of the nucleation mode at larger HAB. These results suggest that 66 Fig. 18. Concentration of soot precursors as a function of HAB for (a) ethylene (C/O = 0.69) (b) benzene (C/O = 0.89). Temperature K ●, C2H4 ○, C6H6 Δ, C2H2 □. 0.5 1.5 2.5 3.5 4.5 0 5 10 15 20 400 600 800 1000 1200 1400 1600 1800 Mole Fraction, % HAB, mm x50 Flame Temperature, K 0.1 0.3 0.5 0.7 0.9 0 5 10 15 400 600 800 1000 1200 1400 1600 1800 Mole Fraction, % HAB, mm Flame Temperature, K a b 67 benzene flames have a distinct behavior because the PAH are consumed in the oxidation zone thereby eliminating the nucleation peak in the upper regions of the flame [8]. Experimental verification of the persistence and consumption of benzene and PAHs for ethylene versus benzene flames is presented in the Appendix G. Given the good agreement between the model predictions and experiments in this study, future modeling work should explore the separate roles of fuel chemistry and flame temperature on particle size distribution. For example, for the comparison in Fig. 18, the soot concentrations formed in the benzene flame were greater than that in the ethylene flame. The model can be used to explore the role of soot concentration on the nucleation mode since all conditions are difficult to control independently in the experiments. 68 5.4 References [1] H. Bockhorn, in: Soot formation in combustion: mechanisms and models, Springer-Verlag, New York, N.Y., 1994. [2] M. Frenklach, Physical Chemistry Chemical Physics 4 (2002) 2028-2037. [3] M. Frenklach; H. Wang, in: Soot formation in combustion: mechanisms and models, Springer-Verlag, New York, N.Y., 1994, p. 165-190. [4] K. H. Homann; H. G. Wagner, Proc. R. Soc. Lond. A 307 (1968) 141-152. [5] A. Violi, Combustion and Flame 139 (2004) 279-287. [6] A. D'Anna; A. Alfe; B. Apicella; A. Tregrossi; A. Ciajolo, Energy & Fuels 21 (2007) 2655-2662. [7] M. Alfe; B. Apicella; R. Barbella; J. N. Rouzaud; A. Tregrossi; A. Ciajolo, Proceedings of the Combustion Institute 32 (2009) 697-704. [8] M. Alfè; B. Apicella; R. Barbella; A. Tregrossi; A. Ciajolo, Proceedings of the Combustion Institute 31 (2007) 585-591. [9] C. S. McEnally; U. O. Koylu; L. D. Pfefferle; D. E. Rosner, Combustion and Flame 109 (1997) 701-720. [10] M. M. Maricq, Combustion and Flame 144 (2006) 730-743. [11] M. M. Maricq, Combustion and Flame 137 (2004) 340-350. [12] B. Zhao; Z. W. Yang; J. J. Wang; M. V. Johnston; H. Wang, Aerosol Science and Technology 37 (2003) 611-620. [13] B. Zhao; Z. W. Yang; Z. G. Li; M. V. Johnston; H. Wang, Proceedings of the Combustion Institute 30 (2005) 1441-1448. 69 CHAPTER 6 EVOLUTION OF SOOT SIZE DISTRIBUTION IN PREMIXED ETHYLENE/AIR AND ETHYLENE/BENZENE/AIR FLAMES: EXPERIMENTAL AND MODELING STUDY 6.1 Introduction Soot is a well-known pollutant that constitutes a serious health concern and contributes to global warming due to its radiation effects [1]. Because of this, recent studies have focused their attention on establishing how and through what intermediate steps carbonaceous particles like soot are formed during the burning of fossil fuels. Studies on the evolution of particle size distribution (PSDs) in combustion systems such as premixed flames, diesel engines, plug flow reactors etc. have provided critical information to gain a better insight into the formation, growth and oxidation of soot particles [2-4]. Measurements of the changes in number, mass and size of the particles have been facilitated by new developments of instrumentation and sampling systems such as the differential mobility analyzer (DMA) [5, 6] and dilution sampling probes [3, 6]. Theses type of equipments have allowed spatially resolved, faster and online measurement of particle sizes in the nanometer-sized range. These new techniques, particularly, the scanning mobility particle sizer (SMPS), which consists of a nano-DMA coupled to an ultrafine condensation particle counter, have resulted in a better characterization of the evolution of soot size distribution in premixed flames. Traditionally, PSDs were assumed to follow a log-normal behavior. However, PSDs in premixed flames at heights above the particle inception regime are now found to be bimodal [3, 7]. This bimodal size distribution is attributed to persistent particle inception together with particle-particle interactions which generates the second or agglomeration mode of larger particles. The magnitude of the bimodal behavior is highly dependent on the combustion conditions. Several studies have studied the experimental conditions on the evolution of PSD in premixed systems [8-11]. These studies have determined that flame temperature and benzene concentration in the initial fuel mixtures have a marked effect on the evolution of the PSDs. Zhao et al. [8] studied the effect of flame temperature on PSDs by changing the velocity of the fresh gases and adding N2 to an ethylene-O2-Ar mixture, which generated peak flame temperatures in the range from 1790 K to 1920 K. Their results showed that the features of the PSDs were strongly dependent on the flame temperature; bimodality was favored at lower flame temperatures. In contrast, high temperature flames favored unimodality. More recent studies [10, 11] have compared the PSDs in pure ethylene, ethylene/benzene and pure benzene flames. Pure benzene and ethylene flames doped with benzene tended to produce higher temperature flames. In addition, major changes in the history of the evolution of PSDs were observed under these conditions. Abid et al. [11] measured soot size distributions in premixed ethylene flames with and without benzene doping. Their results showed that PSDs shifted towards smaller sizes for high temperature flames. In addition, doping the ethylene flame with benzene had little or not effect on soot volume fraction and the bimodal behavior in the size distribution. 71 In a previous study [10], we accounted for the added step of converting, aliphatic to aromatic moities on the PSDs of soot particles by using ethylene and benzene flames with similar C/O ratios. Our results suggested that benzene had a major effect on the unimodal versus bimodal behavior of PSDs in premixed flames. For pure ethylene flames, the PSDs evolved from a unimodal distribution close to the burner surface to a bimodal behavior which persisted even at higher elevations in the flame. For pure benzene flames, the bimodal behavior appeared closer to the burner surface, and the concentration of nucleation-sized particles decreased drastically in the upper region of the flame, therefore the PSD higher in the flame consisted of a unimodal size distribution of larger particles. We associated the presence of a bimodal behavior closer to the burner surface on benzene flames to the earlier formation of soot, and attributed this behavior to the higher concentration of aromatic soot precursors in the oxidation region of the flame [69]. In addition, the drastic fall of nucleation-sized particles in the postflame region of benzene flames was correlated to the decrease in soot precursors. By contrast, in pure ethylene flames, the concentration of the aromatic soot precursors persisted even in the upper region of the flame which favoring the continued presence of a particle nucleation mode. The effect of benzene on the evolution of soot size distribution in premixed flames has been subject of discussion. The importance of the benzene arises from the fact that the formation of the first aromatic ring is considered to be the kinetic controlling step during the process of soot formation [12-18]. How benzene concentration in the initial fuel impacts the evolution of the soot size distribution has been our objective in the last couple of years. In this paper we want to extend the results of our previous publication 72 [10] to mixtures of ethylene and benzene, and make use of a detailed kinetic model to evaluate the effect of benzene when the flame temperature is kept constant, a condition that is difficult to achieve experimentally. Experimental methodologies, conditions (Table 2) and model details were described in Chapters 3 and 4, respectively. 6.2 Results 6.2.1 Temperature Profiles and Evolution of the Soot Size Distribution Temperature profiles for flames E1, EB1, EB2 and EB3 are presented in Fig. 19. Temperature increased rapidly in the oxidation region of the flames until it reaches a peak value. Downstream, temperature leveled and started to drop as a result of radiation cooling augmented by the presence of soot. Peak temperatures were observed to increase from flame E1 (carbon percentage in fuel as benzene, CP = 0 %) to flame EB3 (CP = 45.45 %). The higher soot concentrations in ethylene/benzene flames (EB1, EB2 and EB3) were responsible for the larger temperature gradient in the postflame region of those flames, particularly, for flame EB3. The evolution of the soot size distribution for flames E1, EB1, EB2 and EB3 was measured as a function of the HAB and carbon percentage as benzene. Fig. 20a, 20b, 20c and 20d present the results for the 4 conditions at HAB = 9, 11, 13 and 15 mm, respectively. The results for the pure ethylene flame (E1) showed a characteristic evolution of the soot size distribution in premixed flames [3, 7, 10]. Close to the burner surface (see Fig. 20a, "E1"), a unimodal behavior was observed, which is characteristic of nucleation-sized particles (Dp < 10 nm). Higher in the flame (Fig. 20b, 20c, 20d, "E1"), a bimodal size distribution behavior started to evolve, with the larger or agglomeration 73 Fig. 19. Experimental temperature profiles for flames E1, EB1, EB2 and EB3 as a function of HAB in flames. 1000 1200 1400 1600 1800 2 4 6 8 10 12 14 E1 EB1 EB2 EB3 Flame Temperature, K HAB, mm FLAME 74 Fig. 20. Experimental PSDs in ethylene (E1) and ethylene/benzene (EB1, EB2, EB3) flames at heights of (a) 9 mm, (b) 11 mm, (c) 13 mm and (d) 15 mm above burner surface. 1010 1014 1012 108 100 101 102 0 10 20 30 40 50 HAB = 9 mm EB3 E1 EB1 EB2 dN/dlogDp, #/cm3 Dp, nm Carbon Percentage, % a b c 1010 1014 1012 108 100 101 102 0 10 20 30 40 50 HAB = 11 mm EB3 E1 EB1 EB2 dN/dlogDp, #/cm3 Dp, nm Carbon Percentage, % 1010 1014 1012 108 100 101 102 0 10 20 30 40 50 HAB = 13 mm EB3 E1 EB1 EB2 dN/dlogDp, #/cm3 Dp, nm Carbon Percentage, % 1010 1014 1012 108 100 101 102 0 10 20 30 40 50 HAB = 15 mm EB3 E1 EB1 EB2 dN/dlogDp, #/cm3 Dp, nm Carbon Percentage, % d 75 mode corresponding to the coagulated particles. The nucleation-sized particles were always present in the flame E1, evidenced by the persistence of the lower mode over the whole range of HAB studied. As we will see later, this behavior is directly related to the relatively high concentration of soot precursors over the whole range of HAB studied. A similar behavior, where the unimodal distribution dominated close to burner surface and bimodality was present upper in the flames, was observed for the ethylene flames doped with benzene (Fig. 20a, 20b, 20c, 20d, "EB1","EB2","EB3"). However, we can distinguish major differences on the history of the evolution of soot size distribution for these flames. First, bimodality was evident to appear closer to the burner surface. This effect was more significant with increasing carbon percentages as benzene. Second, number concentration in the nucleation mode slightly decreased along the flame axis for flame EB1, and it was higher for flames EB2 and EB3. At the upper regions of these flames (HAB = 13 mm), the bimodal behavior just started to be developed for flame E1. In contrast, for flames doped with benzene this behavior was clearly seen, which is characteristic of aged-soot particles. In general, the earlier presence of the bimodal behavior and the decrease in number concentration in the nucleation mode for the benzene-doped flames was a distinguishing feature in the evolution of the size distributions and it was a clear difference with respect to the pure ethylene flame (E1). The changes in the evolution of PSD were observed to take place by increasing the carbon percentage as benzene. However, these effects occurred in parallel with an increase in peak temperature (see Fig. 19). Now, the question arises as to whether these effects on PSDs can be accounted either for the temperature change or by the presence of benzene which can promote the earlier formation of soot. To answer this question, the 76 detailed kinetic model, described in section 3, was used. The model was validated with the experimental data presented here and with the results of our previous publication [10] for pure ethylene and benzene flames. D'Anna and coworkers [17, 19] have also found an excellent agreement between experimental measurements of soot precursors and model predictions in a variety of premixed and diffusion flames. 6.2.2 Modeling of Soot Size Distribution and Number Concentration Temperature profiles, inlet fuel/air concentration and mass flow rates were used as inputs into D'Anna's kinetic model to predict the history of the particle size distribution within the flame. Fig. 21 illustrates the experimental PSDs versus model predictions at HAB = 9, 11, 13 and 15 mm for flames E1 and EB3. Model predictions were in good agreement with the experimental data if the model results were shifted downstream by 6mm. This shifting was the result of taking into account the effect of the sampling probe on the flame. The sampling probe not only cools down the flame close to the sampling probe, but also affects the velocity of the burned gases, which in turn, changes the history of the PSD along the flame. The amount of shifting has been supported by experimental measurements of flame temperature with and without the probe in the burner [10]. The literature [3] has described why this shifting should be performed to correctly compare experimental data with model predictions. PSDs in the size ranges from 3 to 10 nm and larger than 10 nm were integrated to approximate the number concentration of particles in the nucleation and agglomeration modes respectively. Fig. 22a, 22b, 22c and 22d compare experimental versus model predictions of number concentration in the nucleation and agglomeration modes for the four flames 77 Fig. 21. Model prediction versus experimental results of PSDs at HAB = 9, 11, 13 and 15 mm for flames E1 (left) and EB3 (right). dN/dLogDp, #/108 109 1010 1011 1012 1013 dN/dLogDp, #/cm3 10 100 9 mm_exp. 9 mm_model Particle Diameter, nm 11 mm_exp. 11 mm_model 15 mm_exp. 15 mm_model FLAME E1 FLAME EB3 13 mm_exp. 13 mm_model 107 108 109 1010 1011 1012 1 10 9 mm_exp. 9 mm_model cm3 Particle Diameter, nm 108 109 1010 1011 1012 11 mm_exp. 11 mm_model dN/dLogDp, #/cm3 108 109 1010 1011 1012 13 mm_exp. 13 mm_model dN/dLogDp, #/cm3 15 mm_exp. 15 mm_model 78 Fig. 22. Model predictions versus experimental results of number concentrations in the nucleation (3-10 nm) and agglomeration (>10-80 nm) modes for flames: (a) E1, (b) EB1, (c) EB2 and (d) EB3. #/100 102 104 106 108 1010 1012 1014 7 8 9 10 11 12 13 14 15 (3-10 nm)_exp. (>10-80 nm)_exp. (3-10 nm)_model (>10-80 nm)_model Number Concentration, #/cm3 HAB, mm FLAME EB2 100 102 104 106 108 1010 1012 1014 7 8 9 10 11 12 13 14 15 (3-10 nm)_exp. (>10-80 nm)_exp. (3-10 nm)_model (>10-80 nm)_model HAB, mm FLAME EB3 a b c d 100 102 104 106 108 1010 1012 1014 (3-10 nm)_exp. (>10-80 nm)_exp. (3-10 nm)_model (>10-80 nm)_model Number Concentration, cm3 FLAME E1 100 102 104 106 108 1010 1012 1014 (3-10 nm)_exp (>10-80 nm)_exp (3-10 nm)_model (>10-80 nm)_model FLAME EB1 79 studied. Number concentration in both modes are well reproduced and the model confirms the evident decrease of the nucleation-sized particles (Dp<10 nm) concentration at higher HAB with increasing benzene concentration in the fuel mixture. In the agglomeration mode (Dp>10 nm) the concentration of particles rose faster with increasing benzene concentration and leveled off higher in the flame. The increase of bigger particles close to the burner surface was also a clear indication of the earlier formation of soot. To confirm that the observed effects on the evolution of soot size distribution and changes of number concentration in the nucleation mode were mainly a result of the presence of benzene in the parent fuel, the model was run with the input temperature profile constant. Fig. 23a and 23b summarize the results of the integrated PSDs in terms of number concentration in the nucleation mode (3-10nm) as a function of the HAB for these four flames. In Fig. 23a and 23b the experimental temperature profiles for flames E1 (peak temperature = 1640 K) and EB3 (peak temperature = 1785 K) were used, respectively. As it can be seen in Fig. 23a and 23b, a similar trend of the data was obtained compared to the results in Fig. 22. Close to the burner surface, nucleation-sized particle concentration increased with increasing benzene concentration, while higher in the burner, the concentration of nanoparticles dropped. This observation was similar for both cases presented in Fig. 23a and 23b. However, the effect of the benzene was more evident for the case with the higher temperature profile (Fig. 23b). 80 Fig. 23. Model predictions of number concentration in the nucleation mode as a function of the HAB for: (a) flames E1, EB1, EB2 and EB3 using as input into the model the experimental temperature profile obtained for flame "E1" and (b) using as input into the model the experimental temperature profile obtained for flame "EB3". 109 1010 1011 1012 1013 0 5 10 15 20 E1 EB1 EB2 EB3 N-nanoparticles, #/cm3 HAB, mm 109 1010 1011 1012 1013 0 5 10 15 20 E1 EB1 EB2 EB3 N-nanoparticles, #/cm3 HAB, mm a b 81 6.3 Discussion and Summary The present study showed the major impact of benzene concentration in the initial fuel on the evolution of the PSDs, both measured and modeled, in premixed ethylene/air and ethylene/benzene/air flat flames. Experimentally, variables such as C/O ratio and unburned gas velocity were maintained constant. Temperature and PSD were measured for pure ethylene/air and ethylene/air flames doped with benzene. Carbon percentage as benzene was varied in the range from 0 to ~46 %. The experimental results showed that for carbon percentage as benzene as low as 5.9 % the