Soot formation in laminar diffusion flames

Laminar, sooting, coflow diffusion flames at atmospheric pressure have been studied experimentally and theoretically as a function of fuel dilution by inert nitrogen. The flames have been investigated with laser diagnostics. Laser extinction has been used to calibrate the experimental soot volume fractions and an improved gating method has been implemented in the laser-induced incandescence (LII) measurements resulting in differences to the soot distributions reported previously. Numerical simulations have been based on a fully coupled solution of the flow conservation equations, gas-phase species conservation equations with complex chemistry, and the dynamical equations for soot spheroid growth. The model also includes the effects of radiation reabsorption through an iterative procedure. An investigation of the computed rates of particle inception, surface growth, and oxidation, along with a residence time analysis, helps to explain the shift in the peak soot volume fraction from the centerline to the wings of the flame as the fuel fraction increases. The shift arises from changes in the relative importance of inception and surface growth combined with a significant increase in the residence time within the annular soot formation field leading to higher soot volume fractions, as the fuel fraction increases.     

Soot formation in high pressure laminar diffusion flames

The details of the chemical and physical mechanisms of the soot formation process in combustion remain uncertain due to the highly complex nature of hydrocarbon flames, and only a few principles are firmly established mostly for atmospheric conditions. In spite of the fact that most combustion devices used for transportation operate at very high pressures (e.g., aircraft gas turbines up to 40 atm, diesel engines exceeding 100 atm), our understanding of soot formation at these pressures is not at a desirable level, and there is a fundamental lack of experimental data and complementary predictive models. The focus of this review is to assess the experimental results available from laminar co-flow diffusion flames burning at elevated pressures. First, a brief review of soot formation mechanisms in diffusion flames is presented. This is followed by an assessment of soot diagnostics techniques, both intrusive and non-intrusive, most commonly used in soot experiments including the laser induced incandescence. Then the experimental results of soot measurements done at elevated pressures in diffusion flames are reviewed and critically assessed. Soot studies in shock tubes and in premixed flames are not covered. Smoke point fuel mass flow rate is revisited, and shortcomings in recent measurements are pointed. The basic requirements for tractable and comparable measurements as a function of pressure are summarized. Most recent studies at high pressures with aliphatic gaseous fuels show that the soot yield displays a unified behaviour with reduced pressure. The maximum soot yield seems to reach a plateau asymptotically as the pressure exceeds the critical pressure of the fuel. Lack of experimental data on the sensitivity of soot morphology to pressure is emphasized. A short summary of efforts in the literature on the numerical simulation of soot formation in diffusion flames at high pressures is the last section of the paper.     

Soot formation in ethanol/gasoline fuel blend diffusion flames

The aim of this paper is to examine how adding ethanol to gasoline affects soot formation. This is currently an important question with respect to particulate emissions from gasoline powered motor vehicles, but in this paper the ethanol impact is examined in co-flow diffusion flames to decouple combustion chemistry from the effects of engine operating parameters. Soot size distributions are measured as a function of height above the burner for E0, E20, E50, and E85 blends. For all fuels, the size distributions evolve from a single nucleation mode low in the flame through a bimodal distribution at mid heights and finally a single accumulation mode. The soot agglomerates in the accumulation mode, exhibit a bipolar charge. The nucleation mode initially includes charged particles, but becomes electrically neutral with increasing height in the flame. Thermodesorber measurements reveal significant hydrocarbon condensation on nucleation mode particles. This is more extensive for E0, E20, and E50 fuels as compared to E85. In other respects as well, the flames fall into two classes: (1) E85 versus (2) E0, E20, and E50. These groups of flames are visibly distinct and exhibit quantitatively different trends in terms of the size and quantity of particulate matter. The E85 flame appears similar to an ethylene diffusion flame, whereas those in the second group are more akin to a benzene flame. The results are discussed with respect to their implications regarding the effects of ethanol blends on PM emissions from gasoline engines.     

Physical and chemical comparison of soot in hydrocarbon and biodiesel fuel diffusion flames: A study of model and commercial fuels

Data are presented to compare soot formation in both surrogate and practical fatty acid methyl ester biodiesel and petroleum fuel diffusion flames. The approach here uses differential mobility analysis to follow the size distributions and electrical charge of soot particles as they evolve in the flame, and laser ablation particle mass spectrometry to elucidate their composition. Qualitatively, these soot properties exhibit a remarkably similar development along the flames. The size distributions begin as a single mode of precursor nanoparticles, evolve through a bimodal phase marking the onset of aggregate formation, and end in a self preserving mode of fractal-like particles. Both biodiesel and hydrocarbon fuels yield a common soot composition dominated by C_x ions, stabilomer PAHs, and fullerenes in the positive ion mass spectrum, and and C_2xH⁻ in the negative ion spectrum. These ion intensities initially grow with height in the diffusion flames, but then decline during later stages, consistent with soot carbonization. There are important quantitative differences between fuels. The surrogate biodiesel fuel methyl butanoate substantially reduces soot levels, but soot formation and evolution in this flame are delayed relative to both soy and petroleum fuels. In contrast, soots from soy and hexadecane flames exhibit nearly quantitative agreement in their size distribution and composition profiles with height, suggesting similar soot precursor chemistry.     

A model of particulate and species formation applied to laminar, nonpremixed flames for three aliphatic-hydrocarbon fuels

A detailed kinetic mechanism is developed that includes aromatic growth and particulate formation. The model includes reaction pathways leading to the formation of nanosized particles and their coagulation and growth to larger soot particles using a sectional approach for the particle phase. It is tested against literature data of species concentrations and particulate measurements in nonpremixed laminar flames of methane, ethylene, and butene. Reasonably good predictions of gas and particle-phase concentrations and particle sizes are obtained without any change to the kinetic scheme for the different fuels. The model predicts the low concentration of particulates in the methane flame (about 0.5 ppm) and the higher concentration of soot in the ethylene and butene flames (near 10 ppm). Model predictions show that in the methane flame small precursor particles dominate the particulate loading, whereas soot is the major component in ethylene and butene flames, in accordance with the experimental data. The driving factors in the model responsible for the quite different soot predictions in the ethylene and butene flames compared with the methane flame are benzene and acetylene concentrations, which are higher in the ethylene and butene flames. Soot loadings in the ethylene flame are sensitive to the acetylene soot growth reaction, whereas particle inception rates are linked to benzene in the model. A coagulation model is used to obtain collision efficiencies for some of the particle reactions, and tests show that the modeled results are not particularly sensitive to coagulation at the rates used in our model. Soot oxidation rates are not high enough to correctly predict burnout, and this aspect of the model needs further attention.     

Soot sheet dimensions in turbulent nonpremixed flames

A method is presented to automate the generation of the statistics of the characteristic length (L) and width (W) of the 2-D slices through 3-D soot sheets in turbulent nonpremixed flames. The method employs the simplified approximation of fitting an equivalent ellipse. The effectiveness of the method is first evaluated by comparison with two types of manual assessment. It is then applied to assess the evolution of the soot sheet dimensions in two turbulent nonpremixed propane flames. These two flames, had the same fuel flow rates but different global mixing rates, were produced by a precessing jet (PJ) and a simple jet (SJ) burner. The global mixing rate of the flames was observed to affect the shape of probability density functions (PDFs) of soot sheets L and W. The shapes of the PDFs of the flames reveal that burnout of the soot sheets in the PJ and SJ flame proceeded predominantly through the reduction of the soot sheet dimensions, and the number of soot sheets, respectively. Finally, a convenient way to estimate the actual maximum width of the single-branched soot sheets is proposed, and this value is estimated to be ∼20 mm for both flames.     

Soot formation and temperature field structure in laminar propane-air diffusion flames at elevated pressures

The effect of pressure on soot formation and the structure of the temperature field was studied in coflow propane-air laminar diffusion flames over the pressure range of 0.1 to 0.73 MPa in a high-pressure combustion chamber. The fuel flow rate was selected so that the soot was completely oxidized within the visible flame and the flame was stable at all pressures. Spectral soot emission was used to measure radially resolved soot volume fraction and soot temperature as a function of pressure. Additional soot volume fraction measurements were made at selected heights using line-of-sight light attenuation. Soot concentration values from these two techniques agreed to within 30% and both methods exhibited similar trends in the spatial distribution of soot concentration. Maximum line-of-sight soot concentration along the flame centerline scaled with pressure; the pressure exponent was about 1.4 for pressures between 0.2 and 0.73 MPa. Peak carbon conversion to soot, defined as the percentage of fuel carbon content converted to soot, also followed a power-law dependence on pressure, where the pressure exponent was near to unity for pressures between 0.2 and 0.73 MPa. Soot temperature measurements indicated that the overall temperatures decreased with increasing pressure; however, the temperature gradients increased with increasing pressure.     

Soot size distributions in rich premixed ethylene flames

Electrical mobility measurements of the soot particles generated in rich premixed ethylene/air flat flames reveal a characteristic lognormal size distribution that is distinct from the self preserving distribution expected from coagulation dominated aerosol dynamics. The distribution changes to a bimodal form as the equivalence ratio and the height above the burner increase. The soot particles are sampled using a three-stage ejector pump, with an overall dilution of the soot mole fraction by 3000, that quenches the flame chemistry and dilutes the sample for analysis by a nano-differential mobility analyzer. The measured soot volume fraction is in good agreement with optical extinction data, quantitatively reproducing the increases previously noted with respect to increasing equivalence ratio and height above the burner. The trends for mean particle diameter are also reproduced, but the mobility diameters are roughly threefold smaller than their light scattering counterparts. Particle number density is found to increase with height, whereas the optical data exhibit the opposite trend. Residence chamber experiments show that with sufficient time the quenched soot sample evolves from the lognormal shape found in the flame to the expected self preserving distribution.     

The effects of molecular structure on soot formation,I. Soot thresholds in premixed flames

Soot thresholds have been measured in atmospheric pressure, laminar premixed flames of 55 pure hydrocarbons. Soot threshold data as a function of fuel structure parameters are presented in four forms; threshold soot indices, TSIs; equivalence ratios assuming combustion to CO₂ and H₂O, f_c; equivalence ratios assuming combustion to CO and H₂O, ψ_c; and carbon to oxygen atom ratios (C/O)_c. The relative TSI scale allows comparison with previous measurements and removes apparatus-dependent variations in the data. Data were obtained for more than 40 previously unstudied fuels, including 15 alkanes, 14 alkenes, 7 alkynes, and 15 aromatics. Comparison with similar data in the literature gave good agreement. Suggested TSI values are given for all fuels which have been reported.     

Soot formation, spray characteristics, and structure of jet spray flames under high pressure

Coaxial jet spray flames of kerosene and oxygen are experimentally studied over a pressure range of 0.1–1.0 MPa to determine the relationship between flame structure, droplet behavior, and soot formation region, which varies with changes in pressure. The direct images and chemiluminescence spectra show that the spray flames have three regions: the blue flame region, which has a peak of CH* and C₂* radical chemiluminescence, luminous flame region caused by soot emission, and blue emission region caused by CO₂ emission. With increase in ambient pressure, the flame length shortens drastically, the luminous flame region envelopes the blue flame region, and the blue emission becomes more intense. The result of phase-Doppler anemometry shows that a large number of small droplets evaporate and disappear near the burner, and the evaporation of large droplets also occurs rapidly under high pressure. The result of temperature measurements shows that high-temperature regions appear near the burner. The flame temperature drastically decreases along the central axis, and a minimum temperature point appears. This point moves upstream with increase in ambient pressure because evaporation of the droplets occurs further upstream. A laser-induced incandescence measurement shows that the soot volume fraction does not monotonously increase or decrease with increase in ambient pressure. The soot volume fraction at the central axis becomes low upstream and high downstream. As pressure increases, the vertical position at which the peak of soot volume fraction appears at the central axis moves upstream.     

Soot properties of laminar jet diffusion flames in microgravity

The soot properties of round, non-buoyant, laminar jet diffusion flames are described, based on experiments carried out in microgravity conditions during three flights of the Space Shuttle Columbia (Flights STS-83, 94 and 107). Experimental conditions included ethylene- and propane-fueled flames burning in still air at an ambient temperature of 298 K and ambient pressures of 35-100 kPa. Measurements included soot volume fraction distributions using deconvolved laser extinction imaging and soot temperature distributions using deconvolved multiline emission imaging. Mixture fractions were estimated from the temperature measurements. Flow field modeling based on the work of Spalding is presented. It is shown that most of the volume of these flames is inside the dividing streamline and thus should follow residence time state relationships. Most streamlines from the fuel supply to the surroundings exhibit nearly the same maximum soot volume fraction and maximum temperature. The present work studies whether soot properties of these flames are universal functions of mixture fraction, i.e., whether they satisfy soot state relationships. Soot state relationships were observed, i.e., soot volume fraction was found to correlate reasonably well with estimated mixture fraction for each fuel/pressure selection. These results support the existence of soot property state relationships in steady non-buoyant laminar diffusion flames, and thus in a large class of practical turbulent diffusion flames through the application of the laminar flamelet concept.     

Soot precursor measurements in benzene and hexane diffusion flames

To clarify the mechanism of soot formation in diffusion flames of liquid fuels, measurements of soot and its precursors were carried out. Sooting diffusion flames formed by a small pool combustion equipment system were used for this purpose. Benzene and hexane were used as typical aromatic and paraffin fuels. A laser-induced fluorescence (LIF) method was used to obtain spatial distributions of polycyclic aromatic hydrocarbons (PAHs), which are considered as soot particles. Spatial distributions of soot in test flames were measured by a laser-induced incandescence (LII) method. Soot diameter was estimated from the temporal change of LII intensity. A region of transition from PAHs to soot was defined from the results of LIF and LII. Flame temperatures, PAH species, and soot diameters in this transition region were investigated for both benzene and hexane flames. The results show that though the flame structures of benzene and hexane were different, the temperature in the PAHs-soot transition region of the benzene flame was similar to that of the hexane flame. Furthermore, the relationship between the PAH concentrations measured by gas chromatography in both flames and the PAH distributions obtained from LIF are discussed. It was found that PAHs with smaller molecular mass, such as benzene and toluene, remained in both the PAHs-soot transition and sooting regions, and it is thought that molecules heavier than pyrene are the leading candidates for soot precursor formation.     

Soot formation in aerodynamically strained methane–air and ethylene–air diffusion flames with chloromethane addition

The effects of chloromethane (CH₃Cl) addition on soot inception in methane–air and ethylene–air counterflow diffusion flames were investigated by varying the concentrations of chloromethane and nitrogen in the fuel stream. Experiments showed a monotonic increase in the critical sooting stretch rate for methane–air flames when methane was replaced by chloromethane, while ethylene and chloromethane flames exhibited a larger sooting tendency than flames under comparable conditions and burning either ethylene or chloromethane alone. For the conditions investigated, the critical sooting stretch rates of methane–chloromethane–nitrogen flames were shown to be primarily a function of the chloromethane loading in the fuel stream. The structure of these flames was modeled using detailed chemistry and transport. Modeling results suggested that the enhancement of soot formation in ethylene–chloromethane flames may be a combined result of increased concentrations of C₂ species and chlorinated C₁ radicals (CH₂Cl and CHCl). A large rate of the reactions among these species may be the first steps in the molecular growth processes, which leads to the inception of soot particles.     

Soot formation with C1 and C2 fuels using an improved chemical mechanism for PAH growth

Recently, an improved chemical mechanism of PAH growth was developed and tested in soot computations for a laminar co-flow non-premixed ethylene–air diffusion flame. In the present work, the chemical mechanism was enhanced further to accommodate the PAH gas phase growth in methane, ethylene and ethane co-flow flames. The changes in the mechanism were tested on a methane/oxygen and two ethane/oxygen premixed flames to ensure no degradation in its application to C₂ fuels. The major soot precursors were predicted in a satisfactory matter. The robustness of the soot solution methodology was tested for different fuels by solving methane/air, ethane/air and ethylene/air co-flow laminar diffusion flames using a single solution algorithm for all three cases. The peak soot volume fractions, which varied by two orders of magnitude between fuels, were predicted within a factor of two for all flames. The computations were also able to reproduce the spatial distributions of soot and to explain the variation in soot formation pathways among the fuels. Despite a similarity in bulk properties of the flame, the soot particles in different flames exhibit significantly different growth modes. Ethylene/air flames tend to form soot earlier than methane/air flames and inception plays a bigger role in the latter.     

Effect of biodiesel fuels on diesel engine emissions

The call for the use of biofuels which is being made by most governments following international energy policies is presently finding some resistance from car and components manufacturing companies, private users and local administrations. This opposition makes it more difficult to reach the targets of increased shares of use of biofuels in internal combustion engines. One of the reasons for this resistance is a certain lack of knowledge about the effect of biofuels on engine emissions. This paper collects and analyzes the body of work written mainly in scientific journals about diesel engine emissions when using biodiesel fuels as opposed to conventional diesel fuels. Since the basis for comparison is to maintain engine performance, the first section is dedicated to the effect of biodiesel fuel on engine power, fuel consumption and thermal efficiency. The highest consensus lies in an increase in fuel consumption in approximate proportion to the loss of heating value. In the subsequent sections, the engine emissions from biodiesel and diesel fuels are compared, paying special attention to the most concerning emissions: nitric oxides and particulate matter, the latter not only in mass and composition but also in size distributions. In this case the highest consensus was found in the sharp reduction in particulate emissions.     

Free convection diffusion flames from burning solid fuels

A general analysis for free convection diffusion flames is presented for the burning of solid fuels having surface geometries which give rise to boundary layer flows. The combustion chemistry is treated at the level of a single global reaction. Thermal radiation from the flame and the burning fuel surface is included in the formulation of the problem as well as the possibility of radiation from an external source. Allowance is also made for the case in which the gaseous products of pyrolysis include an inert component in addition to the combustible fuel. Solutions are presented for a variety of cases which show the effects of varying the chemical reaction rate, the radiation parameters, the Lewis number and the inert component in the fuel. Extinction phenomena are discussed, and results are given which show how extinction limits are affected by ambient oxygen concentration, water droplet sprays and chemically active fire suppressants. A discussion of some of the numerical techniques developed to obtain and verify the solutions is included.     

Numerical and experimental study of soot formation in laminar diffusion flames burning simulated biogas fuels at elevated pressures

The effects of pressure and composition on the sooting characteristics and flame structure of laminar diffusion flames were investigated. Flames with pure methane and two different methane-based, biogas-like fuels were examined using both experimental and numerical techniques over pressures ranging from 1 to 20 atm. The two simulated biogases were mixtures of methane and carbon dioxide with either 20% or 40% carbon dioxide by volume. In all cases, the methane flow rate was held constant at 0.55 mg/s to enable a fair comparison of sooting characteristics. Measurements for the soot volume fraction and temperature within the flame envelope were obtained using the spectral soot emission technique. Computations were performed by solving the unmodified and fully-coupled equations governing reactive, compressible flows, which included complex chemistry, detailed radiation heat transfer and soot formation/oxidation. Overall, the numerical simulations correctly predicted many of the observed trends with pressure and fuel composition. For all of the fuels, increasing pressure caused the flames to narrow and soot concentrations to increase while flame height remained unaltered. All fuels exhibited a similar power-law dependence of the maximum carbon conversion on pressure that weakened as pressure was increased. Adding carbon dioxide to the methane fuel stream did not significantly effect the shape of the flame at any pressure; although, dilution decreased the diameter slightly at 1 atm. Dilution suppressed soot formation at all pressures considered, and this suppression effect varied linearly with CO₂${\text{CO}}_2$ concentration. The suppression effect was also larger at lower pressures. This observed linear relationship between soot suppression and the amount of CO₂${\text{CO}}_2$ dilution was largely attributed to the effects of dilution on chemical reaction rates, since the predicted maximum magnitudes of soot production and oxidation also varied linearly with dilution.