Flame Reaction Zone Research Articles

Flame Acceleration in Narrow Channels: Applications for Micropropulsion in Low-Gravity Environments

A laminar flame propagating towards the open end of a narrow channel filled with a gaseous combustible mixture can accelerate or oscillate, depending on the wall temperature and the channel width. The accelerating flame is able to produce a high-speed flow that has the potential to provide significant thrust. We study these phenomena using multidimensional reactive Navier-Stokes numerical simulations, and show that for adiabatic walls, the maximum flame acceleration occurs when the the channel is about five times larger than the reaction zone of a laminar flame. In three-dimensional square channels, the flame speed increases roughly two times faster than in two-dimensional channels. The accelerating flame generates weak compression waves that propagate with a local sound speed and can accelerate the unreacted material ahead of the flame to the velocities close to the sound speed without creating strong shocks. This combustion regime is of particular interest for micropropulsion because it allows an efficient use of fuel and a gradual development of the trust. I. Introduction This work is based on a micropropulsion idea generated from a relatively recent discovery concerning the behavior of laminar flames propagating in very narrow tubes. When a slow laminar flame is ignited near the closed end of a tube, the material ahead of the flame is accelerated and a boundary layer is formed along the walls in the unreacted material. Due to the presence of the boundary layer, the flow velocity changes across the channel, increasing from zero at the wall to a maximum in the middle of the tube. Previous simulations [1-3] have shown that the nonuniform flow stretches the flame, so that the shape of the flame becomes similar to the velocity profile. This increases the flame surface area, thus accelerating the energy generation and the flow. There is little change in the laminar burning rate, but the speed of the flame that propagates with the flow grows very rapidly in the laboratory frame of reference, from centimeters per second to hundreds of meters per second for adiabatic walls. This phenomenon, explained by Ott et al. [1-3], is different from the well known case of flame propagation in a channel with cold walls. When the walls are cold (the case originally considered in the classical work of Mallard and Le Chatelier [4]), the flame quenches near the walls, and the flame velocity oscillates. We consider regimes in which the flame remains laminar until the end of the tube and does not produce strong shocks or detonations. The high-speed flow created by the accelerating flame has the potential to provide a significant thrust that can be further enhanced by attaching an appropriately shaped exist nozzle. The thrust will gradually increase as the flame and flow accelerate, and reach a maximum when the flame approaches the open end of the tube or the nozzle. In contrast to the majority of other combustion-based propulsion concepts, most of the thrust here can be provided by the material accelerated ahead of the flame. Therefore, a propulsion device could be only partially filled with a combustible mixture. The rest of the material can be an inert gas or combustion products from a previous pulse. This micropropulsion concept is of particular interest for low-gravity environments, where relatively small, controlled thrusts could be used for navigational corrections. In such cases, there might not be a need for the high repetition rates required for pulse combustion and pulse detonation engines. The purpose of this paper is to confirm original results [1-3] and to carry the study further to examine the possibility of using flame acceleration in narrow tubes for micropropulsion. As a model energetic system, we consider the same stoichiometric acetylene-oxygen mixture used in [3], perform numerical simulations for two-dimensional (2D) and three-dimensional (3D) channels of different widths and lengths, and analyze effects of channel sizes on integral propulsion characteristics.

Liftoff characteristics of partially premixed flames under normal and microgravity conditions

An experimental and computational investigation on the liftoff characteristics of laminar partially premixed flames (PPFs) under normal (1-g) and microgravity (μ-g) conditions is presented. Lifted methane–air PPFs were established in axisymmetric coflowing jets using nitrogen dilution and various levels of partial premixing. The μ-g experiments were conducted in the 2.2-s drop tower at the NASA Glenn Research Center. A time-accurate, implicit algorithm that uses a detailed description of the chemistry and includes radiation effects is used for the simulations. The predictions are validated through a comparison of the flame reaction zone topologies, liftoff heights, lengths, and oscillation frequencies. The effects of equivalence ratio, gravity, jet velocity, and radiation on flame topology, liftoff height, flame length, base structure, and oscillation frequency are characterized. Both the simulations and measurements indicate that under identical conditions, a lifted μ-g PPF is stabilized closer to the burner compared with the 1-g flame, and that the liftoff heights of both 1-g and μ-g flames decrease with increasing equivalence ratio and approach their respective nonpremixed flame limits. The liftoff height also increases as the jet velocity is increased. In addition, the flame base structure transitions from a triple- to a double-flame structure as the flame liftoff height decreases. A modified flame index is developed to distinguish between the rich premixed, lean premixed, and nonpremixed reaction zones near the flame base. The 1-g lifted flames exhibit well-organized oscillations due to buoyancy-induced instability, while the corresponding μ-g flames exhibit steady-state behavior. The effect of thermal radiation is to slightly decrease the liftoff heights of both 1-g and μ-g flames under coflow conditions.

Catalyzed combustion of hydrogen–oxygen in platinum tubes for micro-propulsion applications

The present investigation addresses the need to understand the physics and chemistry involved in propellant combustion processes in micro-scale combustors for propulsion systems on micro-spacecraft. These spacecraft are planned to have a mass less than 50 kg with attitude control estimated to be in the 1–10 mN thrust class. Micro-propulsion devices behave differently than macro-scale devices because of the differences in magnitude of flow rates and heat transfer. Reducing the combustor size increases the relative surface area, increasing the heat loss, and as combustors are continuously reduced in size, they approach the quenching dimensions of the propellants. Combustors of this size are expected to significantly benefit from surface catalysis processes. A miniature flame tube apparatus is chosen for this study because microtubes can be easily fabricated from known catalyst materials, and their simplicity in geometry can be used in fundamental simulations for validation purposes. Experimentally, we investigated the role of catalytically active surfaces within 0.4 and 0.8 mm internal diameter microtubes, with special emphases on ignition processes in fuel rich gaseous hydrogen and gaseous oxygen. Calculations of flame thickness and reaction zone thickness predict that the diameters of our test apparatus are below the quenching diameter of the propellants in most atmospheric test conditions. The temperature and pressure rise in resistively heated platinum microtubes and the exit hydrogen concentration were used as an indication of exothermic reactions. Data on imposed heat flux/preheat temperature required to achieve ignition versus mass flow rate are presented. With a plug flow model, the experimental conditions were simulated with detailed gas-phase chemistry and surface kinetics. Computational results, in general, support the experimental findings.

Experimental and numerical investigation of microscale hydrogen diffusion flames

Characteristics of microscale hydrogen diffusion flames produced from sub-millimeter diameter ( d = 0.2 and 0.48 mm) tubes are investigated using non-intrusive UV Raman scattering coupled with LIPF technique. Simultaneous, temporally and spatially resolved point measurements of temperature, major species concentrations (O 2, N 2, H 2O, and H 2), and absolute hydroxyl radical concentration (OH) are made in the microflames for the first time. The probe volume is 0.02 × 0.04 × 0.04 mm 3. In addition, photographs and 2-D OH imaging techniques are employed to illustrate the flame shapes and reaction zones. Several important features are identified from the detailed measurements of microflames. Qualitative 2-D OH imaging indicates that a spherical flame is formed with a radius of about 1 mm as the tube diameter is reduced to 0.2 mm. Raman/LIPF measurements show that the coupled effect of ambient air leakage and pre-heating enhanced thermal diffusion of H 2 leads to lean-burn conditions for the flame. The calculated characteristic features and properties indicate that the buoyancy effect is minor while the flames are in the convection–diffusion controlled regime because of low Peclet number. Also, the effect of Peclet number on the flame shape is minor as the flame is in the convection–diffusion controlled regime. Comparisons between the predicted and measured data indicate that the trends of temperature, major species, and OH distributions are properly modeled. However, the code does not properly predict the air entrainment and pre-heating enhanced thermal-diffusive effects. Therefore, thermal diffusion for light species and different combustion models might need to be considered in the simulation of microflame structure.

The Necessity of Studying Chemical Reactions of the Clean Agent Heptafluoropropane in Fire Extinguishment

Halon-based products for gas protection systems against fire will be abandoned completely, probably in China by 2008. Clean agents with zero ozone depletion potential and low global warming potential now are actively being developed. One of the candidates is heptafluoropropane and total flooding systems based on it are starting to be widely used in computer rooms, switching facilities and many other facilities. However, the extinguishing mechanism of heptafluoropropane with afire is not well-understood. It is believed that physical flame cooling is the main mechanism. Heat will be absorbed efficiently from the flame reaction zone to reduce its temperature. Apart from flame cooling, whether there are chemical inhibition effects must be further investigated in order to give a better understanding on the extinguishing properties of the agent. Pointing out the importance of studying the associated gas-phase chemistry is the objective of this paper. Results in the literature will be briefly reviewed. In unfavourable cases, using a clean agent with a low capability to suppress a fire might generate more pollutants which damage the environment.

Condensation flame of acetylene decomposition

Acetylene decomposition flame propagation was numerically analyzed and was found to be the result of the condensation reaction. Condensation processes provide reaction heat and act as a driving force for C2H2 flame propagation. The kinetic model reasonably predicts the level of burning velocity of the acetylene decomposition flame. The model does not demonstrate the relatively strong positive pressure dependence of burning velocity as was observed experimentally in the work of Cummings et al. [Proc. Combust. Inst. 8 (1962) 503–510]. Heat-release kinetics demonstrates a two-stage process. The first stage corresponds to heat release due to benzene formation, and the second stage of heat release corresponds to soot inception and carbonization processes. It was demonstrated that the burning velocity is sensitive to the surface growing rate constant. The use of a simplified form of presentation of the surface growing process [P.R. Lindstedt, in: Soot Formation in Combustion: Mechanisms and Models, Springer-Verlag, Berlin/New York, 1994, pp. 417–441] represents positive thermal feedback in the heat generation in a flame reaction zone.

Toward a Flame Embedding Model for Turbulent Combustion Simulation

Combustion in turbulent flows may take the form of a thin flame wrapped around vortical structures. For this regime, the flame embedding approach seeks to decouple computations of the outer nonreacting flow and the combustion zone by discretizing the flame surface into a number of elemental flames, each incorporating the local impact of unsteady flow-flame interaction. An unsteady strained laminar flame solver, based on a boundary-layer approximation of combustion in a time-dependent stagnation-point potential flow, is proposed as an elemental flame model. To validate the concept, two-dimensional simulations of premixed flame-vortex interactions are performed for a matrix of vortex strengths and length scales, and a section of the flame is selected for comparison with the flame embedding model results. Results show that using the flame leading-edge strain rate gives reasonable agreement in the cases of low strain rate and weak strain rate gradient within the flame structure. This agreement deteriorates substantially when both are high. We propose two different schemes, both based on averaging the strain rate across the flame structure, and demonstrate that agreement between the one-dimensional model and the two-dimensional simulation greatly improves when the actual strain rate at the reaction zone of the one-dimensional flame is made to match that of the two-dimensional flame.

Effects of Halons and Halon Replacements on Hydrogen-Fueled Laminar Premixed Flames

Fundamental unstretched laminar burning velocities and flame response to stretch, as characterized by Markstein numbers, were considered experimentally and computationally. The investigation was limited to laminar premixed flames involving hydrogen burning in air with Halon 1301 (CF 3 Br), nitrogen, and carbon dioxide considered as flame suppressing diluents. Outwardly propagating spherical laminar premixed flames were observed for fuel equivalence ratios of 0.6-1.8, concentrations of oxygen in oxygen/nitrogen mixtures of 21% by volume before dilution with a suppressant, and normal temperature and pressure. Suppressants involved concentrations of the chemically active suppressant, Halon 1301, of 0-2% by volume and 'concentrations of the chemically passive (chemically inert) suppressants, nitrogen and carbon dioxide, of 0-50% by volume. The present flames were sensitive to flame stretch, yielding values of unstretched-to-stretched laminar burning velocities in the range of 0.6-1.2 for levels of flame stretch well below quenching conditions, for example, for Karlovitz numbers less than 0.15. The agreement between measured and predicted unstretched laminar burning velocities was good based on the mechanisms of Mueller et al. for H 2 /O 2 reactions and Babushok et al. for halogen reactions. The resulting flame structure predictions suggest that H and OH radical production and transport are important for flame suppression and preferential-diffusion/stretch interactions; this reflects the strong correlation between laminar burning velocities and the maximum concentrations of H and OH in the reaction zone of the present flames. It also was found that effects of flame suppression that reduced concentrations of H and OH in the reaction zone made the present flames more unstable to preferential-diffusion/stretch interactions, an effect that tends to counteract the beneficial effects of flame suppressants to some extent.

Scandium ionization mechanisms in H 2–O 2–N 2 flames supported by calculated thermodynamic data

For 11 scandium neutrals and 20 positive ions containing hydrogen and oxygen in the gas phase, standard enthalpies and free energies of formation and entropies are tabulated at three temperatures of 0, 298.15 and 2400 K using energies and structural data from density functional calculations. These data are first employed to calculate bond dissociation energies, some of which can be compared with experimental and other calculated values. They are then used to determine the composition of scandium neutral species and also ions in an experimental flame at 2400 K, and to assess possible flame-ion production reactions involving chemi-ionization and thermal (collisional) ionization. The only viable mechanism appears to be the exothermic reaction of atomic Sc with O to yield ScO + and free electrons in the flame reaction zone. Further details of the flame-ion chemistry of scandium are discussed. In particular, a theoretical calculation shows that the formation of the stable dihydroxide ion Sc(OH) 2 + from ScO + proceeds via the hydrate intermediate ScO +·H 2O through a cyclic transition state.

Chemical kinetics of scandium ionization in H 2–O 2–N 2 flames

The ionization of scandium was investigated in five premixed, fuel-rich, H 2–O 2–N 2 flames at atmospheric pressure in the temperature range 1820–2400 K. Scandium was introduced (≤10 −6 mol fraction) using an atomizer technique, and the ions were observed by sampling the flame gas through a nozzle into a mass spectrometer. From calculations based on theoretical thermodynamic data, neutral scandium species were present primarily as Sc(OH) 3 with a small amount of OScOH, and the dominant ionic species were Sc(OH) 2 + and ScO +. This is in good agreement with experimental observations of the ions as members of the oxide ion series ScO +· nH 2O ( n=0–3). The ionization was observed to receive an initial boost, evidently from the atomic chemi-ionization reaction of Sc with O to form ScO + and free electrons in the flame reaction zone. No other such reaction was calculated to be exothermic. Thermal (collisional) ionization of ScO is negligible. When a trace of methane was added to produce H 3O + as a chemical ionization (CI) source, scandium ions were formed by proton transfer with a global rate constant given by k PT=(9.58±0.76)×10 −23 T 3.92±0.62 cm 3 molecule −1 s −1; it exhibits a positive temperature dependence with an activation energy barrier of 16.2 kcal mol −1. The global recombination coefficient for scandium ions with electrons is given by k Sc=(3.72±0.22)×10 −4 T −1.03±0.26 cm 3 molecule −1 s −1 with the expected negative temperature dependence.

Asymptotic structure of rich methane-air flames

The asymptotic structure of unstrained, laminar, fuel-rich, premixed methane flame is analyzed by using a reduced chemical-kinetic mechanism made up of three global steps. Analysis is carried out for values of equivalence ratio greater than 1.3. The flame structure is presumed to comprise three zones: an inert preheat zone, a thin reaction zone, and a post-flame zone. In contrast to previous asymptotic analyses of lean flames and moderately rich flames, where the reaction zone of these flames was presumed to be made up of two layers, for rich flames analyzed here all chemical reaction are presumed to take place in one layer. The structure of the reaction zone of rich flames is obtained by integrating two second order ordinary differential equations, one giving the consumption of fuel and the other the consumption of oxygen. For values of equivalence ratio greater than 1.3, burning velocities obtained from the asymptotic analysis are found to agree reasonably well with those obtained using a chemical-kinetic mechanism made up of elementary reactions.

Current–voltage characteristics in a flame plasma: analysis for positive and negative ions, with applications

A flat, cylindrical, laminar, H 2–O 2–N 2 flame in plug flow with velocity v f = 19.8 m s −1 and cross-sectional area A = 1.05 × 10 −4 m 2, at 1 atm and 2400 K, was doped with appropriate additives to give a weak, continuum, quasi-neutral plasma involving Cs +/ e − or H 3O +/ e − or H 3O +/Cl −/ e −. The flame impinged on a planar, normal, water-cooled, conducting electrode designated N (for nozzle) located a variable distance z downstream from the flame’s luminous reaction zone which is separated by a dark space or gap δ ≃ 1 mm from a water-cooled metallic burner B, the second electrode. Neither electrode has the properties of a Langmuir probe. Two types of data were measured: i– V characteristics with the current i(N) collected by the nozzle versus the applied voltage Δϕ between N and B at a fixed value of z (e.g. 20 mm); and profiles of i(N) versus z at a fixed value of Δϕ. For Δϕ < 0 (e.g. −50 V; N is negative), the electrons are stopped in the flame and the constant saturation current is controlled by the convective flow of positive ions to the nozzle: i +(N) = eAn +(N) v f . This provides a wonderfully simple and accurate measurement of the absolute density n +(N). Alternatively, if a weak solution of a Cs salt (e.g. 10 −4 M CsCl) is sprayed using a pneumatic atomizer such that Cs is completely ionized, the delivery factor of the atomizer can be calibrated. Furthermore, if the delivery factor giving n +(N) is known, v f can be determined. With Δϕ fixed (e.g. −50 V), a profile of n +(N) versus z can be obtained throughout the flame gas downstream; the density distribution is not affected by the application of the applied voltage. For Δϕ > 0 (e.g. +50 V, N is positive), the current through the flame i +(B) = − i e (N) − i −(N) (the latter term is included if negative ions are present), is controlled by the flow of positive ions of mobility μ + to the burner across the potential gradient in the burner gap g: i +(B) = eAn +(B)μ +∇ϕ g. When Δϕ is sufficiently positive to achieve a constant saturation current, n +(B) can be determined; it represents the total ion production in the flame reaction zone. When negative ions are present replacing even a large fraction of the electrons, the effect on the i– V characteristic is relatively minor; it does not appear possible to provide a separate analysis for n e and n −. However, profiles through the flame plasma clearly show the effects of negative ion processes such as ion–ion recombination, for example. For a wide range of the applied voltage Δϕ, both positive and negative, it is possible to calculate the potential distribution in the burner gap, the bulk flame plasma and in the positive ion sheath at the nozzle. This provides a quantitative understanding of the ion and electron behaviour throughout the flame.

A numerical and experimental investigation of “inverse” triple flames

Tribrachial or triple flames represent a class of partially premixed flames that generally contain three spatially distinct but synergistically coupled reaction zones, namely a rich premixed, a lean premixed, and a nonpremixed reaction zone. The generally considered flow arrangement for burner-stabilized triple flames involves a rich mixture issuing from a central port and a lean mixture from two outer ports, which we call a reference configuration (RC). Herein, we examine an inverse configuration (IC) in which a fuel-lean stream is flanked by two fuel-rich streams. The reaction zone topology in this configuration is richer and more complex compared to that in a RC flame. A numerical-experimental investigation is conducted to characterize the fundamental differences and similitude between the RC and IC methane–air triple flames in both spatial and mixture fraction based coordinates. The detailed structure of IC triple flames and their response to variations in the rich and lean equivalence ratios are examined. Finally, the transient behavior of the flames in both configurations is described. The IC and RC flames have markedly different spatial structures. In the inverse configuration, the global flame contains five reaction zones. The predicted and measured topologies of the various reaction zones are in excellent agreement. The modified mixture fraction (ξ) is found to be effective in characterizing the structure of both RC and IC flames. The scalar profiles in terms of ξ clearly illustrate the similitude between the two flames. The three reaction zones in the RC flame have a structure that is similar to that of the corresponding five reaction zones of the IC flame. The two nonpremixed (or the two rich premixed reaction zones) in the IC flames are not differentiated in terms of ξ. Both the flames are subjected to a buoyancy-induced instability at normal gravity that generates large vortex structures, which cause the reaction zones to flicker. The flame–vortex interaction is initiated in the nonpremixed reaction zone for the IC flame and in the lean premixed reaction zone for the RC flame. Consequently, the IC flame flickers with higher amplitude but lower frequency as compared to the RC flame. There is good agreement between the measured and predicted flickering frequencies for the two flames. As φrich is increased, the height of the two rich premixed reaction zones in the IC flames increases, their tips open, and the chemical activity in these zones decreases. While the oscillation frequency is essentially constant, the oscillation amplitude increases as φrich is increased. The effect of increasing φlean is to enhance chemical activity in the lean premixed zone. However, the two rich premixed zones and the outer nonpremixed zone are relatively unaffected by variations in φlean. The oscillation amplitude and frequency also remain unaffected by variations in φlean.

Anomalous behaviour of the non-selective ionisation signal pulse under laser ablation into a flame

When applying to the probing electrode the potentials with the different polarity one can notice that the non-selective and selective signals behave quite differently. In case of the positive potential compared with the negative one the non-selective signal shows anomalous behavior: it changes the shape and polarity and its amplitude increases dramatically while the selective one only changes the polarity that can be explained by the electrical properties of the flame reaction zone, which effectively eliminates the charges from laser plasma by the recombination.

Recombination coefficients for H 3O + ions with electrons e − and with Cl −,Br − and I − at flame temperatures 1820–2400 K

Rate coefficients were measured for gas-phase electron–ion recombination of H 3O + and for ion–ion recombination of H 3O + with Cl −,Br − and I − in five H 2–O 2–N 2 flames spanning a temperature range 1820–2400 K. A novel, simple and accurate technique was employed involving total positive ion collection at an electrode located a variable distance downstream of the flame reaction zone; a bias voltage was applied between the collection electrode and the metallic burner. The measured values of the recombination coefficients in cm 3 molecule −1 s −1 units are: (0.0132±0.0004) T −1.37±0.05 for H 3O +/e −;(20.8±0.7) T −2.52±0.16 for H 3O +/Cl −;(1.01±0.04) T −2.14±0.07 for H 3O +/Br −; and (8.82±0.35) T −2.40±0.12 for H 3O +/I −.

Partially Premixed Combustion in Lifted Turbulent Jets

This article reports observations of structures consistent with triple flame reaction zones in the stabilization region of turbulent nonpremixed jet flames. Previous studies of laminar mixing layers and nonpremixed jets show Hi brachial structures, however reports of triple flame structures in turbulent flowfields are sparse, The asymmetric coflow, which facilitates visualizing the luminous flame structure, is described and the observed double and triple flame structures are discussed. Flame luminosity data is presented and the relevance to CH planar laser-induced fluorescence (PLIF) studies of leading edge flame structures is discussed. Furthermore, connection is drawn to select theoretical studies of triple flames and partially premixed combustion.

Pressure change and transport process on flames formed in a stretched, rotating flow

Flame characteristics in a stretched, rotating flow have been investigated by numerical simulation of tubular laminar flames for lean hydrogen, methane, and propane/air mixtures. Twin planar flames in counterflow have been also simulated for comparison. A fixed inlet velocity at the porous wall of the burner was assumed in all cases, and the cylindrical containing tube (radius R = 9.5 mm) was either maintained stationary or rotated. Results showed that, within the range studied, the flame temperatures always increase monotonically with increasing fuel concentration, and at the same time the reaction zones move outwards. However, while the introduction of rotation also causes a monotonic temperature increase of hydrogen and methane/air mixtures, that of a propane/air mixture decreases. The temperature change with rotation becomes smaller with an increase of the fuel concentration. As a consequence of the centrifugal force, ρ ν θ 2/r, induced by the rotation, a pressure gradient is formed in the cylindrical containing tube, with low pressure along the axis. The pressure gradient at the outer, unburnt edge of the flame reaction zone becomes smaller as the fuel concentration increases. The resultant decreased mass transport by pressure diffusion provides an explanation for part of the above-mentioned temperature change associated with rotation. The remainder of the effect is associated with changed stretch characteristics of the flames.

A study of the chemical and physical processes governing CO 2 laser-induced pyrolysis and combustion of RDX

The flame structure of 1,3,5-trinitro-1,3,5-triazacyclohexane (RDX) propellants under laser-assisted combustion was studied to better understand related chemical and physical processes in the gas phase. Experiments were conducted from 0.1 to 3 ATM in pressure with heat fluxes of 50 to 600 W/cm 2. Gaseous products were extracted through the use of quartz microprobes and analyzed by a triple quadrupole mass spectrometer (TQMS). Temperature profiles were measured using micro-thermocouple techniques to investigate reaction zones in RDX flames. Flame behavior was observed using a high-magnification video system. Major species in RDX flames were identified as H 2, H 2O, HCN, H 2CO, NO, HNCO, N 2O, and NO 2 at low masses ( m/ z ≤ 47). In addition to these species, H 2CNH with m/ z = 29 was found to exist in the near-surface reaction zone as an important minor species. Higher molecular weight species were found at m/ z values of 47, 54, 56, 70, 81, and 97; with the daughter mode operation of TQMS, they were identified as HONO, C 2H 2N 2, C 2H 4N 2, C 2H 2N 2O, C 3H 3N 3, and C 3H 3N 3O, respectively. Increasing heat flux and decreasing pressure stretched out the reaction zones and were useful for investigating reactions near the deflagrating surface. However, the conditions appeared to have no effect on major reaction pathways. Two-stage chemical reaction pathways in the gas phase were explicitly identified from the major species profiles at all experimental conditions. Also, the reactions of minor high-mass species occurred in the primary reaction zone. The decomposition of RDX at the surface showed evidence of the two competing branch reactions into H 2CO + N 2O and HCN + HONO, as well as two subsequent reactions: H 2CO + N 2O → H 2O + CO + N 2 and 2HONO → H 2O + NO + NO 2. With the consideration of the previous four reactions, the branching ratio for the two decomposition pathways of RDX was estimated to be about 2:1. For all experimental conditions, temperature profiles had a near-surface region where temperature increased very slowly; the extent of this zone increased as the near-surface reaction zones expanded. After this region, the temperature profiles increased to final flame temperatures without any dark zone temperature plateau. Based on comparisons of species and temperature profiles, this near-surface region is believed to be related to the consumption of NO 2, production of NO and H 2O, and production and consumption of high-mass species.

Properties of carbonaceous nanoparticles in flat premixed C 2H 4/air flames with C/O ranging from 0.4 to soot appearance limit

The presence of very fine carbonaceous particles in the ambient causes concern either for their toxic properties or for their influence on atmospheric chemistry and physics. Previous work has shown that carbonaceous nanoparticles with a typical size of about 2–3 nm are formed in the reaction zone of rich premixed sooting flames and may act as soot precursors. This paper ascertains the presence of nanoparticles in rich laminar C 2 H 4 /air premixed flames with C/O ratio ranging over a wide interval below the soot limit. Ultraviolet (UV)-visible absorption starting from 200-nm, laser-induced fluorescence (LIF), and laser light scattering (LLS) methods were used for the in situ measurements and for the characterization of the material collected from the flames as hydrosol in water traps. It was found that the absorption spectra from 200 to 300 nm were due both to hot CO 2 , whose absorption cross sections are orders of magnitude higher than those found at room temperature, and to carbonaceous high molecular mass structures. The absorption spectra of those structures, obtained by subtracting the CO 2 contribution, and the fluorescence spectra show the same behavior as those measured on the hydrosol suspension. The monochromatic absorption coefficient in the UV due to high molecular mass structures have their onset at C/O=0.4, whereas the fluorescence excited at λ 0 =266 nm appears around C/O=0.5, where also light scattering coefficients, in the UV and visible, in excess with respect to those due to burned gases, start to be measurable. These results are interpreted in terms of a carbonaceous network of one-and two-ring aromatic functionalities, connected by aliphatic bonds, with typical size around 2 nm. The high-temperature formation of these carbonaceous nanoparticles below soot formation limit is considered an important route for explaining the presence of organic carbon aerosols in the atmosphere.

Improvement of two-photon laser induced fluorescence measurements of H- and O-atoms in premixed methane/air flames

photodissociation is shown to be an important photolytic source of H atoms in the reaction zone of the methane flames. At 226 nm, an efficient energy transfer between O(3P) and N2 is established from the observation of O-resonant emissions from the second positive system of N2. The subsequent rate of O(3P) depletion appears to be essentially “controlled” by the O(3P) concentration and is probably only minimally temperature dependent. Finally quenching rate coefficients for O(3P) by water and nitrogen are deduced from quenching rates measurements performed in CH4/O2/N2 and CH4/O2/Ar flames.