1. Introduction

For lack of special conditions the processes of gas-phase combustion have non-steady character; these processes occur under non-stationary currents, density and pressure fluctuations ^[1]. For instance, L. D. Landau theoretically proved ^[2] that a flat flame front is unstable from the hydrodynamic point of view. This work is focused on the consideration of gaseous combustion in non-stationary and unstable regimes. These unstable modes can be classified as thermal diffusion, hydrodynamic, thermo acoustic and heterogeneous ones; the latter mode is first detected in this work. Thermal diffusion instability is often observed in flames, in which Le≠1 (Lewis's number Le = D/α, where D is the diffusivity of the component determining the rate of combustion, α - is the heat diffusivity). This instability results in the existence of cellular flames in lean hydrogen mixtures ^[3].

Hydrodynamic instability relates to a difference between the density of the burned-down and initial flammable gas caused by thermal expansion of products; under certain conditions this can lead to occurrence of cellular structures on a flame front (FF) ^[4].

Thermo acoustic instability is caused by a feedback between non-stationary mode of combustion and acoustic modes of the reactor. The criterion of occurrence of this instability has been obtained by Rayleigh and consists in the fact that acoustic fluctuations are sustained if the maximum heat is transferred to the fluctuating gas at the time of its maximum compression ^[5].

The knowledge of features of chemical transformation in the course of unstable combustion is an urgent problem both for the combustion theory and for practical applications, namely combustion intensification and explosion safety. Experimentally, this problem is solved for a case of thermal diffusion instability ^[6] under the conditions of zero gravity in the International Space Station. It was shown in ^[6] for the first time that under conditions of zero gravity a separate isolated motionless combustion cells arise in lean hydrogen-oxygen mixtures, i.e. the separate "chemical reactors" in the combustible environment occur. Ya.B.Zeldovich predicted theoretically ^[7] that the steady heat and mass conservation equations admit a solution corresponding to a stationary spherical flame, though the same equations in planar geometry provide a solution in the form of a propagating wave.

Direct experimental verification of the hydrodynamic instability of the flame front ^[2] was performed in ^[8]. The flat flame front shape was kept by means of imposing an acoustic field. The growth rate of two-dimensional perturbations in time was observed after switching off the acoustic field. It should be noted that the experiment shows the very important interrelation between the major factors causing flame instability, namely, hydrodynamic and thermo acoustic ones ^[5].

Flame instability may be also induced by heterogeneous factors associated with the occurrence of chemically active disperse solid particles (aerosol) in a gas phase. It should be noted that the advancement in catalytically stabilized (CS) combustion technology requires the development of catalysts with improved activity (desired light-off temperature is less than 450^oC) and thermal stability. The onset of homogeneous ignition within the catalytic reactor is detrimental to the catalyst integrity (it can cause catalyst meltdown) and understanding the nature of such an event is of prime interest in the CS reactor design. The gas-phase ignition is strongly influenced by the hetero/homogeneous coupling (catalytic fuel depletion, adsorption/desorption reactions of combustion intermediates). Therefore, the understanding the ignition mechanism in the presence of a catalyst is essential for accurate prediction of the behavior of the CS reactor. The concept of hydrogen-assisted CS combustion is of particular interest in natural gas-fueled turbines ^[9]. The addition of small amounts of H₂ to natural gas reduces the catalyst light-off temperature and improves the combustion stability by damping flame pulsations. Hence, the knowledge of hydrogen CS combustion appears to be an important step in understanding the hydrogen-assisted CS combustion.

Some experimental facts relating to the reactions between platinum (one of the most effective combustion catalysts) and oxygen at temperatures up to the melting point were considered by Chaston ^[10]. A thin film of feebly stable, solid platinum oxide (probably PtO₂ or PtO ^[11]) is formed on platinum surfaces in the air or oxygen at room temperature ^[12] and thickens as the temperature increases to about 500°C, then it decomposes ^[13]. The loss of weight of platinum at higher temperatures is attributed to the formation of gaseous platinum oxide, and deposition of platinum on cooler surfaces (above about 500°C) - to its disproportionation. In the Figure presented in the Ref. 10 it is shown that Pt deposits on a furnace brick removed after long service in such a way that black oxide-containing deposits can be observed on the cooler edges and crystalline platinum deposits on hotter surfaces ^[10].

This means that the molecules or clusters of both platinum oxide and platinum metal exist in gaseous phase at temperatures over 500⁰ C. Therefore, Pt containing particles extending by diffusion into the volume containing combustible gas (e.g. H₂ – air mixture) in the process of heating of a Pt wire become the catalytic centers, where hydrogen ignition can take place in the course of FF propagation. Therefore, one can expect the occurrence of instabilities of the flame front of hydrogen combustion initiated with the Pt wire caused by catalytic centers distributed in the gas phase. This instability has to be observed under conditions in which thermal diffusion instability is missing (near stoichiometric mixtures ^{[14, 15, 16]}). The detection of that regime is one of the objectives of this work.

Comparably long delay times of 40% H₂ + air ignition at 1 atm were first observed over a Pt foil ^[11]. It was established that the process of ignition of H₂–air mixtures at atmospheric pressure begins with a primary center occurrence on the most chemically active site of the surface that initiates flame propagation ^[11]. In addition, it was shown that the introduction of a Pt wire into a reactor eliminates the phenomenon of negative temperature coefficient in combustion of stoichiometric n-pentane-air mixtures; however, the Pt wire has no effect on the ignition delay time of thermal ignition at lower temperatures ^[14].

In the work, both spatial propagation of unstable FF and the features of chemical transformation in the flame in a bomb of constant volume are investigated by methods of optical 4D spectroscopy and color high-speed cinematography by the example of combustion of n-pentane - air mixtures. The experiments were carried out under conditions when a spherical shape of the flame front transforms into unstable flat front after contact of FF with reactor walls. The work is also focused on the detection of instabilities of the spatial development of 40% H₂–air flame over Pt as well as to the establishment of the temperature dependence of delay times of ignition in the heated reactor at 1 atm by means of a quick gas transfer in the presence of platinum foil or a Pt wire.

Thus, the work consists of two parts; the first one is aimed at investigation of instabilities in a homogeneous medium, and the second deals with heterogeneous instabilities in gaseous flames.

2. Experimental

Experiments were carried out in the stainless steel cylindrical reactors I and II 25 cm long and 12 cm in diameter, supplied with removable covers and an optical window at an end face (Figure 1, I,II) ^[17].

Experiments in the reactor I were performed at atmospheric pressure with previously prepared stoichiometric mixtures of n-pentane with air and carbon dioxide (CO₂). CO₂ additives made up 10% and CCl₄ additives made up to 2%. Previously prepared gas mixture was admitted into the reactor up to necessary pressure; then a spark discharge (1.5 J) was initiated. The electrodes of spark ignition were placed in the center of the reactor I. Registration of ignition and FF propagation was performed by means of an optical 4D spectrometer (hyper spectrometer) and a color high-speed camera Casio Exilim F1 Pro (frequency of shots – 600 s^-1). A video file was stored in computer memory and its time-lapse processing was performed ^[18]. 4D spectrometer ^[18] registered both a narrow strip (a red line 3 in Figure 1. Ia) (a spatial coordinate) and a spectral wavelength with a two-dimensional optical detector array. The data from the optical detector array was stored in computer memory with a 300 Hz frequency. The speed video filming of combustion and 4D spectrophotometry was carried out simultaneously. In Figure 2, the process of hyper spectral recording of a combustion process is shown ^[18].

Before each experiment, the reactor was pumped out with a fore vacuum pump 2NVR-5D. Pressure in the reactor was controlled with the vacuum gauge. Gases n-pentane (n-C₅H₁₂) “Merck”, CO₂, CCl₄ were chemically pure. Carbon tetrachloride CCl₄ was used as the inhibitor of hydrocarbon combustion ^[3].

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window

View next figure

Figure 1. I: a- experimental setup (1) – stainless steel reactor, (2) - quartz optical glass, (3) – a strip on which 4D spectral recording was carried out. The width of the strip is about 1 mm; b- the scheme of experimental installation: (1) - reactor, (2) - electric heater, (3) - thermal insulation, (4) - valves, (5) - mixer, (6) - optical window, (7) - digital movie camera, (8) - 4D spectrometer (hyper spectrometer), (9) - pressure gauge, (10) - registration system, (11) - digital millivoltmeter, (12) - spark igniter. II: a - experimental setup; b,c - its schematic diagram: (1) - reactor, (2) - electromagnetic valve, (3) - buffer volume, (4) - cylinder with gas mixture, (5) - hyper spectrometer, (6) - digital video camera, (7) - rotary mirror, (8) - internal thermal isolation, (9) - heater, (10) - external thermal isolation, (11) - optical window, (12) - pressure transducer and spark ignition circuit, (13) - ADC–computer based data acquisition system, (14) - millivolt meter and the thermocouple, (15) - aluminum ring for prevention of gas swirling. The blue circle notes the place of tangential input of gas into the reactor. The line on which 4D-spectral shooting was carried out is specified red. The width of this line is about 1 mm.

The scheme of paths of light beams in the 4D spectrometer is presented in Figure 2a. The field of view of the spectrometer provides the registration of a narrow strip along a window (the red line 3 in Figure 1). In Figure 2b, hyperspectral data (hypercube) in RGB pseudo-colors are shown (across - spatial coordinates of the strip, and down – the time coordinate). In Figure 2c, the emission spectrum of one of the hypercube points is shown. The spectrum evidently depends on the place both on the spatial coordinate and on time.

The experiments over Pt surface were performed in the reactor II with gas mixtures of 40% H₂ + 60% air at 270 – 350 C without gas swirling (an Al ring was inserted into the reactor). A heated cylindrical stainless steel reactor II 25 cm in length and 12 cm in diameter, equipped with a tangential gas input, demountable covers and an optical quartz window in one of the covers was used (Figure 1, II) ^{[19, 20, 21]}. The accuracy of temperature measurements was 0.3 K. Registration of the ignition and flame propagation was performed by means of an optical 4D spectrometer (hyper spectrometer) and a color high-speed camera Casio Exilim F1 Pro. The pumped and heated reactor II was quickly filled with a gas mixture under investigation from a high-pressure buffer volume to a necessary pressure. An electromagnetic valve was used to open and close gas communications. Because of a sharp pressure difference in the buffer volume and the reactor, there arose a gas swirl in the reactor that led to the reduction of time of establishment of uniform temperature distribution ^[14]. To prevent gas swirling the Al ring focused normally to a gas flow was inserted into the reactor.

Notice that the direct measurements of the dynamics of temperature change in the center of the reactor I by means of thin thermocouples were performed under almost similar conditions in ^[15]. It was shown that the time of warming up of the gas mixture was no longer than 0.3 s. In this case, the formula relating to only conductive heat exchange gives a considerably greater value of the order of tens of seconds ^[16]. A pressure transducer recorded pressure in the course of bleeding-in and combustion. A light-emitting diode was turned on at the moment of the valve opening, and its flash was recorded by the camera. It allowed determining a delay time of ignition τ from a frame sequence independently for each separate ignition. Pt foil 12×6 cm and 0.3 cm thick or Pt wire 15 cm long and 0.3 cm in diameter was placed in the reactor. Before each experiment, the reactor was pumped down to 0.1 Torr. For diagnostics of dust structures, the particles emitted by a platinum wire under heating in atmospheric air, were illuminated with a wide, flattened laser beam “laser sheet” at λ = 532 nm ^[22]. Total pressure in the reactor was monitored with a vacuum gauge, and the pressure in the buffer volume was controlled with a manometer. Chemically pure gases and 99.99% Pt were used.

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window
• PREV
  View previous figure

View next figure

Figure 2. Deriving 4D spectra a – scheme of paths of light beams in a hyper spectrometer (1) – entrance lens, (2) – diaphragm with a slit, (3) – collimator, (4) – dispersing element, (5) – projective lens, (6) – two-dimensional optical detector array, (7) – the line of the image (the red line in Figure 1) formed by a narrow slit of the hyperspectrometer slit; b – hyper spectral data (hypercube) in RGB pseudo-colors (across – relative spatial coordinate x of a red window strip, down – relative time coordinates y of the image of a slit); c –emission spectrum of stoichiometric mixture of n-pentane with air + 10% CO₂ + 1% CCl₄ at 1 atm recorded in a given pixel of the hypercube.

3. Results and Discussion

In Figure 3, typical frames of high-speed filming of FF propagation in n-pentane-air mixtures in the presence of 10% CO₂ and 1% CCl₄ are presented. In the frames, after the moment of initiation the stationary spherical FF propagation until a contact with the inner surface of the reactor is observed. Then FF propagates in the cylindrical part of the reactor in the direction to end faces. Upon transition to the combustion in the cylinder, cellular combustion takes place. Really, at a stage of spherical FF propagation, the FF radius grows so quickly that instabilities do not have time to develop; FF is unperturbed ^[16]. At a propagation stage along the cylinder, unstable flat flame arises, as the theory predicts ^[2]. As is seen from Figure 3, this instability manifests itself in the formation of cellular structures on FF. Formation of cells is characteristic for a certain extent of dilution with inert additive – in fast-burning mixtures (not diluted with inert gas) flame cells do not occur. In the flames of the mixtures diluted with argon, cellular structures do not move in space, thus, the cell size slowly grows. As the extent of dilution of the stoichiometric mixture increases by adding CO₂ and CCl₄ instead of argon, cellular structures move in the direction of gravity (Figure 3).

We will emphasize that for stoichiometric flames one cannot expect occurrence of both thermal diffusion instabilities (there is no significant distinction in transfer coefficients) and thermo acoustic ones (the velocity of FF is comparably small, and sound vibrations are missing ^[23]). Thus, the mode of cellular combustion is caused by the gas dynamic instability peculiar to flat flames ^[2]. Features of combustion in each separate cell were established experimentally with the use of 4D spectroscopy, which allows detecting spectra emitted both from the border between cells and from the internal area of a flame cell.

The typical frame characterizing cellular combustion of n-pentane is presented in Figure 4a (see also Figure 3). In Figure 4b, the hypercube for this image in Blue pseudo color along a vertical axis (see Figure 1) is given, and in Figure 4c, a fragment of this hypercube is shown, on which the points of spectral analysis are specified.

The spectra recorded both from the border and from the internal area of one of flame cells are given in Figure 5.

Since the mixture contains an inhibitor additive (CCl₄), the intensity of combustion is lower, than without an inhibitor; respectively the heat release is comparably small, therefore "hot" lines of Na and K atoms, which are usually observed in flame emission spectra, are missing ^[3]. The spectra obtained agree with known data and contain the bands of active intermediates and products of hydrocarbon oxidation ^[3]. These are CH (A¹Δ–X² Π) at 431 nm, C₂ (A³P_g –X³P_u) (transitions 1-0, 0-0, 0-1) in the range of 470 to 570 nm ^[24] and emission bands of water vapor near 950 nm (for example, (1, 2, 0), (3, 0, 0) ^[25]). Notice that CH and C₂ bands belong to the zone of intensive chemical transformation (FF zone) ^[3]; and emission bands of water vapor belong to the zone of emission of combustion products.

Therefore, by the ratio of intensity of the bands of C₂ radical to those of water molecules in the spectra it is possible to make the qualitative conclusion about what zone of combustion characterizes the spectrum: FF zone (the zone of intensive chemical transformation) or the zone of reaction products. If the relative intensity of C₂ bands considerably exceeds the relative intensity of water vapor bands, the emission spectrum essentially corresponds to FF zone; and if the ratio of intensities is inverse one, the spectrum belongs to reaction products.

The scanning of emission spectra of the flammable mixture along a window axis (see Figure 1), namely, along the coordinate from top to down in the marked interval in Figure 4 is shown in Figure 5. As is seen from Figure 5, intensities of spectral bands along a window axis change in opposite directions: while the relative intensity of C₂ bands has a maximum at x=20 and x=180 (x is the relative value of the spatial coordinate), intensity of H₂O bands at the same x values has a minimum. This means, firstly, that the combustion is non-uniform in space, otherwise intensities of C₂ and H₂O spectral bands would both increase or decrease smoothly; i.e. 4D spectroscopy allows observing combustion cells in the same way that by high-speed filming (Figure 3). Secondly, the fact that the intensity of C₂ bands has a maximum at the same values (x=20, x=180) at which the intensity of H₂O bands has a minimum, means that at these x values the light is emitted mainly from the FF zone. At the x values, at which the ratio of intensities of C₂ and H₂O bands is inverted, light is emitted from the zone of reaction products. From previously mentioned it follows that the use of 4D spectroscopy allows establishing that combustion cells observed in Figure 4 represent in essence separate "chemical reactors"; in each of the “reactors” the process of complete chemical transformation occurs.

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window
• PREV
  View previous figure

View next figure

Figure 3. Result of speed filming of process of propagation of FF in stoichiometric mixture of n-pentane with air + 10% CO₂ + 1% CCl₄ at 1 atm in stainless steel reactor with a frequency of shots 600 s^-1. The figure on a shot corresponds to a number of a frame when shooting

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window
• PREV
  View previous figure

View next figure

Figure 4. a) The video frame of cellular combustion of a stoichiometric mixture of n-pentane with air + 10% CO₂ + 1% CCl₄ at 1 atm, the spatial interval x = 1-234, in which the spectra were recorded is marked with a white vertical line, b) the hyper spectral image (hyper cube) in pseudo-color B for better allocation of flame cells, c) the allocated fragment of Figure 4b with the interval, in which the spectra were recorded.

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window
• PREV
  View previous figure

View next figure

Figure 5. Spectra of combustion of n-pentane-air mixture +10% CO₂ + 1% CCl₄ at 1 atm along a window axis (along the vertical line in Figure 4c), t=3 ms from the initiation

We should recall that it was experimentally shown ^[6] that thermal diffusion instability in lean hydrogen-oxygen mixtures under conditions of zero gravity manifests itself in the occurrence of separate, isolated motionless combustion cells i.e. separate "chemical reactors" in the combustible environment. In that paper, by means of both 4D optical spectroscopy and color high-speed filming, the features of combustion in flame cells caused by hydrodynamic instability are experimentally established for the first time. Besides, in the paper ^[8] aimed at direct experimental verification of L.D. Landau hypothesis on hydrodynamic instability of the flat FF the interrelation of the major factors causing instability of flames – hydrodynamic and acoustic ones - was worked out ^[5].

This means that in a combustion cell caused by the instability of any nature (thermal diffusion, hydrodynamic, thermo acoustic ones) the complete cycle of chemical transformations, being characteristic of the given combustion process, takes place.

In the following experiments, the spatial development of ignition of H₂–air mixtures in the presence of Pt at 1atm was investigated. Notice that the temperature of ignition of H₂–air mixtures at 1 atm in the reactor containing Pt foil ^[11] is about 170 K less than that in the stainless steel reactor. Note also that the changeover across a critical condition of ignition is accompanied by a substantial growth of the delay period τ only over a Pt catalytic surface, though τ of ignition over stainless steel changes stepwise in a very narrow temperature interval ~ 1°. Induction periods in the mixtures can reach tens of seconds both at temperatures less than 260°C and in the case of the “fresh” state of Pt surface. That state is realized in the very first experiment, in which Pt surface is not changed by the active centers of ignition. In the Figure 6, the sequences of video images of development of the initial centers of ignition are shown.

As is seen from Figure 6a, the smooth, uniform FF is observed for spark-initiated ignition in 40% H₂ + air mixture at room temperature of the reactor walls for a stainless steel surface. As is shown in Figure 6, if Pt foil is placed in the stainless steel reactor, FF is almost uniform. It agrees with earlier experimental observation ^{[10, 12, 13]} that an oxide layer on the bulk Pt is thinner than on a Pt wire, so the amount of Pt containing particles in the volume is too small to influence on a flame front structure.

The portions of video recording of the process of heating of a Pt wire by the current J = 2A are presented in Figure 6c. The wire was illuminated with a laser sheet. As is seen in Figure 6c, the ultra-disperse particles of Pt oxide are emitted under heating. As platinum oxide emission from the wire is registered with a frequency of 60 frames per second, it is obvious that in experiments on initiation of hydrogen ignition by a wire in the heated reactor, during a delay period (under our conditions 3 ÷ 70 s) ultra-disperse Pt oxide can spread about all of the reactor volume before ignition. However, in the presence of Pt wire (Figure 6d) the cellular structure of FF is observed; the Pt wire becomes red-hot before and after ignition due to catalytic reactions on the Pt surface. Addition of 15% CO₂ to the combustible mixture provides complete suppression of the cellular regime of combustion; thus, 15% additive of helium does not influence on the cellular combustion regime.

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window
• PREV
  View previous figure

View next figure

Figure 6. Sequences of video images of spatial development of the flame front for P = 1 atm, the numbers in each frame correspond to a consecutive number of the video image: a - spark initiated ignition in 40% H₂ + 60% air at the temperature of reactor walls 20 C; 600 shots s^–1; P = 1 atm; the material of the wall is stainless steel; b - 40% H₂ + 60% air mixture at the reactor wall temperature of 247 C; Pt foil is placed in the stainless steel reactor. It is clearly seen in the shot 3; c - the heating of Pt wire (J=2А). The wire is lighted with a vertical green laser sheet, 60 frames/s; d - 40% H₂ + 60% air mixture at the reactor wall temperature of 316 C; Pt wire is placed in the stainless steel reactor. It is clearly seen in shots 1, 61. As is seen from shots 1, 61, Pt wire is heated before and after the explosion due to catalytic reactions; e - 85% (40% H₂ + 60% air mixture) + 15% CO₂ at the reactor wall temperature of 320 C; f - 85% (40% H₂ + 60% air mixture) + 15% He at the reactor wall temperature of 309 C; Pt wire is placed in the stainless steel reactor. It is clearly seen in shots 1, 61. As is seen from shots 1, 62, Pt wire is heated before and after the explosion due to catalytic reactions; g - The dynamics of an increase in the radius R of the flame fronts estimated from the data of Figure 6: experiments a; d; e. P₀ = 1 atm, 600 frames/s.

In this part of the paper, the cellular combustion regime is detected in combustion of near stoichiometric H₂ – air mixtures initiated with a Pt wire. In the absence of the Pt wire, the flame front of hydrogen combustion in similar conditions is smooth and uniform ^{[11, 26]}. According to previously mentioned, at the temperatures > 500°C the molecules or clusters of both platinum oxide and platinum metal are formed in the gaseous phase.

Thus, the instability regime is caused by the Pt – containing particles extending into the reactor volume by diffusion and acting as catalytic centers, on which hydrogen ignition takes place in the course of FF propagation. Unlike to He additives (Figure 6f), in the presence of CO₂ the diffusivity of catalytic particles decreases and these probably do not influence on the shape of a propagating flame; therefore, FF is uniform (Figure 6e).

In Figure 6g, the estimates of normal flame velocities from the change of a visible flame radius by the equations given in ^{[19, 26]} are presented. As is seen in Figure 6g, at spark initiation in the mixture diluted with carbon dioxide, the constant flame velocity is reached in a certain period corresponding to the time of formation of the steady FF ^[26]. However, in the presence of the platinum catalyst (Figure 6g), the constant FF velocity (within an experimental error) is reached practically at once, i.e. the catalytic effect of Pt leads to the sharp reduction of the time of the steady FF formation. Besides, as is seen in Figure 6g, the normal velocity of FF in the presence of a catalytic surface is much higher (≈ 2.6 m/s, d), as compared to the velocity without the catalyst (at spark initiation ≈ 1.9 m/s, a, in the presence of 15% of CO₂ ≈ 1.8 m/s, e).

The nonuniformity of FF light emission, caused by catalytic instability, is not only detected by the method of high-speed filming (Figure 6), but also with the hyper spectrometer (Figure 7, the experiment under conditions of Figure 6d). This nonuniformity is seen directly in a hyper cube (Figure 7).

As is seen in Figure 7, light spots (hot spots) corresponding to the combustion cell movement in time, are observed in the hypercube of the mixture of 40% H₂ – air. These spots are the images of the single cell at consecutive time points. The main feature of these "spots" is that the emission spectra intensity dependencies both on coordinate and on time have a maximum in the spots. The emission spectra of the spots are close to a gray body spectrum, i.e. the radiation of the spots (cells) really corresponds to the emission of heated catalyst particles.

The ignition delay time (τ) is one of the most important macro kinetic characteristics of chain thermal ignition, which can be measured in a relatively simple way. According to Ref. 27 and 28 by Saytzev et al., and Livengood et al correspondingly, in the shock tube and rapid compression machine, thermal ignition is of a kernel nature. We have recently shown ^{[11, 14]} that the ignition of both hydrogen and n-pentane–air mixtures in a rapid mixture injection static reactor at 1 atm begins with the appearance of an initial center at the most chemically active site of the surface; thus, thermal ignition includes the stages of warming up, local ignition and flame propagation. In all likelihood, the kernel ignition is a rule rather than the exception. Figure 8 presents the temperature dependence of ignition delay times for 40% H₂–air mixture in the reactor in the presence and absence of gas flow over the catalytic surface (Pt foil or a Pt wire) in Arrhenius coordinates assuming that the effective rate constant value is k ≈1/τ ^[14].

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window
• PREV
  View previous figure

View next figure

Figure 7. The hypercube of combustion of 40% H₂ – air mixture, initiated by a platinum wire obtained with the hyper spectrometer, T₀ = 320 C, P₀ =1 atm, a spectral interval over 550 - 650 nm.

Download as

• PPT
  PowerPoint Slide
• PNG
  Larger image(png format)

Veiw figureFigures index

•  NEW
  View larger figure in new window
• PREV
  View previous figure

Figure 8. Temperature dependencies of delay times of thermal ignition of 40% of H₂ - air mixture in the presence and in the absence of a gas stream at P₀ = 1 atm. Black triangles - the ignition initiated with platinum foil, motionless gas, R= 0.987; black squares - the ignition initiated with platinum foil, circulating gas, R= 0.989; empty squares - the ignition initiated with a platinum wire, motionless gas, R = 0.981

In each experiment, Pt foil or the Pt wire were treated with 6 preliminary ignitions of 40% H₂–air mixture at 1 atm to stabilize the surface state of Pt. The data were processed with the use of the program package Statistica 9 (Statsoft). Correlation coefficients R are presented in the caption to Figure 8. As is evident from Figure 8, the effective activation energy E is the same for Pt foil or wire as well as for the presence or absence of a gas flow. The experimental value of E makes 19±3 kcal/mole and is close to that of the branching chain step of hydrogen oxidation H + O₂ → OH + O (16.7 kcal/mole ^[29]). It means, in agreement with literature data, that the activated chain branching is the slowest step in the whole consequence of reactions leading to flame propagation.

As is known ^[3] the intercept of straight lines in Figure 8 is roughly the reciprocal of collision frequency. Really, the surface area of a Pt wire is less than that of the Pt foil; therefore, the frequency of the encounters with the surfaces is less for the Pt wire. The frequency of the encounters with the surfaces for Pt foil in motionless gas is less than that for the same foil in swirling gas. All of the preceding is illustrated in Figure 8.

4. Conclusions

The features of combustion in homogeneous flame cells caused by hydrodynamic instability are experimentally established for the first time by means of both 4D optical spectroscopy and color high-speed filming. It is shown that in a combustion cell caused by the instability of any nature (thermal diffusion, hydrodynamic, thermo acoustic) the complete cycle of chemical transformations, being characteristic of the given combustion process, takes place.

The cellular combustion regime of 40% H₂-air mixture in the presence of Pt wire within the interval 270-350°C was observed for the first time. It is established that the regime is caused by the catalytic action of Pt containing particles formed by decomposition of volatile platinum oxide in the gas phase. It is shown that the effective activation energy of the temperature dependence of delay times of thermal ignition over Pt makes 19±3 kcal/mole and is close to that of the branching chain step of hydrogen oxidation.

References

[1]	Markstein G 1964. Nonsteady flame propagation (Oxford London: Pergamon Press).
	In article
[2]	Landau L 1944. On the theory of slow combustion Acta Phys.-Chim. URSS 19 77.
	In article
[3]	Lewis B and Von Elbe G 1987. Combustion, Explosions and Flame in Gases (New York, London: Acad.Press).
	In article
[4]	Clavin P and Williams F A 1982. Effects of molecular diffusion and of thermal expansion on the structure and dynamics of premixed flames in turbulent flows of large scale and low intensity J. Fluid Mech. 116 251.
	In article		View Article
[5]	Kampen J F 2006. Acoustic pressure oscillations induced by confined turbulent premixed natural gas flames, PhD thesis (University of Twente, Enschede, The Netherlands, Printed by Febodruk BV).
	In article
[6]	Ronney P D 2001. Premixed-Gas Flames in: Microgravity Combustion: Fires in Free Fall (Academic Press, London: H. Ross, Ed.) p 35.
	In article
[7]	Zel’dovich Ya B 1984. Selected works Chemical Physics and Hydrodynamics Moscow Nauka: ed. Chariton Yu A) (in Russian).
	In article
[8]	Clanet C and Searby G. 1998. First experimental study of the Darrieus-Landau instability Phys. Rev. Lett. 27 3867.
	In article		View Article
[9]	Appel C Mantsaras J Schaeren R Bombach R and Inauen A. 2004. Catalytic combustion of hydrogen – air mixtures over platinum: validation of hetero-homogeneous reaction schemes Clean Air 5 21.
	In article		View Article
[10]	Chaston J C. 1964. Reactions of Oxygen with the Platinum Metals Platinum Metals Rev. 8. 50.
	In article
[11]	Rubtsov N M Seplyarskii B S Troshin K Ya Chernysh V I and Tsvetkov G I. 2012. Investigation into Spontaneous Ignition of Hydrogen–air Mixtures in a Heated Reactor at Atmospheric Pressure by High-speed Cinematography Mendeleev Commun. 22 222.
	In article
[12]	Perry D L. 1995. Handbook of Inorganic Compounds (N.-Y.:CRC Press).
	In article
[13]	Lagowski J J. 2004. Chemistry Foundations and Applications (London:Thomson Gale).
	In article		PubMed
[14]	Borisov A A Rubtsov N M Skachkov G I and Troshin K Ya. 2012. Gas-phase spontaneous ignition of hydrocarbons Russ. J. Phys. Chem. B 6 517.
	In article
[15]	Borisov A A Knorre V G Kudrjashova E L and Troshin K Ya. 1998. On temperature measurement in an induction period of the ignition of homogeneous gas mixtures in rapid mixture injection static setup Chem. Phys. Rep. B 17 105.
	In article
[16]	Zel’dovich Ya B Barenblatt G I Librovich V B and Machviladze G M. 1980. Mathematical Theory of Combustion and Explosion (Moscow: Nauka) (in Russian).
	In article
[17]	Rubtsov N M Seplyarskii B S Troshin K Ya Chernysh V I and Tsvetkov G I. 2011. Initiation and propagation of laminar spherical flames at atmospheric pressure Mendeleev Comm. 21 218.
	In article
[18]	Kalinin A P Orlov A G Rodionov A I and Troshin K Ya. Demonstration of the probability of the investigation of combustion and explosion by hyperspectral methods Physicochemical kinetics in gasdynamics www.chemphys.edu.ru/ pdf/2009-06-18-001.pdf (in Russian).
	In article
[19]	Naboko I M Rubtsov N M Seplyarskii B S Troshin K Ya Tsvetkov G I and Chernysh V I. 2011. High-speed colour cinematography of the spontaneous ignition of propane–air and n-pentane–air mixtures Mendeleev Commun. 21 31.
	In article		View Article
[20]	Khalil A E and E Gupta A K. 2013. Fuel Flexible Distributed Combustion With Swirl For Gas Turbine Applications Applied Energy 109 2749.
	In article		View Article
[21]	Khalil A E and E Gupta A K. 2014. Swirling Flowfield for Colorless Distributed Combustion Applied Energy 113 208.
	In article
[22]	Rubtsov N M. 2016. The modes of gaseous combustion (Switzerland: Springer International Publishing)
	In article		View Article
[23]	Rubtsov N M Seplyarskii B S Naboko I M Chernysh V I Tsvetkov G.I. and Troshin K.Ya., 2015. Acoustic instabilities in hydrogen-air mixtures in the closed reactor at the central spark initiation PCAIJ 10 073.
	In article
[24]	Herzberg G. 1950. Molecular Spectra and Molecular Structure Vol. 1 Spectra of Diatomic Molecules (New York:Van Nostrand).
	In article
[25]	Coheur P F Bernath P F Carleer M and Colin R. 2005. A 3000 K Laboratory Emission Spectrum of Water The Journal of Chemical Physics 122 074307.
	In article
[26]	Rubtsov N M Seplyarskii B S Troshin K Ya Chernysh V I and Tsvetkov G I. 2011. Initiation and propagation of laminar spherical flames at atmospheric pressure Mendeleev Commun. 21 218.
	In article		View Article
[27]	Saytzev S G and Soloukhin R I. 1962. Proceedings of the 8th symposium (International) on combustion (The Combust. Inst., Pittsburgh, PA) p 2771.
	In article
[28]	Livengood J C and Leary W A. 1951. Autoignition by rapid compression Industrial and Engin. Chem. 43 2797.
	In article
[29]	Germann T C and Miller W H. 1997. Quantum mechanical pressure dependent reaction and recombination rates for OH + O → O2 + H J. Phys. Chem. A: 101 6358.
	In article		View Article