Oscillatory Ignition Research Articles

Isothermal interpretations of oscillatory ignition during hydrogen oxidation in an open system. I. Analytical predictions and experimental measurements of periodicity

The oxidation of hydrogen under well stirred, flowing conditions is the natural prototype of branched-chain reactions in open, gaseous systems. Experimentally, it exhibits the classical forms for the (p-Ta) first and second ignition limits in the flow system. The constant reactant supply ensures that stationary states exist at subcritical conditions whereas the supercritical reaction is a repetitive, oscillatory sequence of events. In a linked pair of papers, we investigate isothermal criticality at the second limit in terms of changing nature of the singularities for the mass conservation equations and derive kinetic relationships that explain the oscillatory features. In this paper the origins of oscillatory ignition are traced analytically to chain-branching via H atoms coupled to the changing third-body efficiency in the elementary process H + O2+ M → HO2+ M, when water formed during ignition is displaced by the inflow of fresh hydrogen and oxygen. Analytical predictions are made of the periods between successive ignitions and of the conditions at which oscillatory reaction is transformed to a stationary state. A composition limit for the existence of oscillatory ignition in the (lean) mixture H2+ 15O2+ 14N2is located experimentally and explained in terms of the analytical inter­pretations presented here.

Oscillations, glow and ignition in carbon monoxide oxidation in an open system II. Theory of the oscillatory ignition limit in the c. s. t. r

The oxidation of carbon monoxide in the presence of hydrogen can produce a single ignition pulse in a closed vessel and repetitive, i. e. oscillatory, ignition in an open system. It is possible to predict the locus of critical conditions on a map of reactant pressure, p , against vessel temperature, T a , in a flow system by a treatment based on the change in local stability of the stationary state. Even the very simplest kinetic model for the CO + H 2 + O 2 reaction allows satisfactory predictions of the dependence of the critical pressure on T a , and of the displacement of such p – T a peninsulae as the mixture composition (CO : H 2 ratio) is varied. Many of the results can be obtained in terms of simple algebraic expressions. The relation between this approach and classical treatments of criticality based on the unbounded growth of the steady-state radical concentration or on tangency conditions (chain–thermal theory) is investigated. Oscill­atory periods (the interval between successive ignition pulses) are calcu­lated, and the variation in the mean residence time arising from the change in the number of moles during reaction and the accompanying self-heating is discussed.

Oscillations, glow and ignition in carbon monoxide oxidation in an open system. I. Experimental studies of the ignition diagram and the effects of added hydrogen

The oxidation of carbon monoxide in equimolar mixtures (CO + O 2 ) has been studied in a well-stirred open system (0.5 dm 3 ) at vessel temperatures in the range 700-840 K, and reactant pressures up to 100 Torr ( ca . 13.3 kPa) at a mean residence time of 8.5 s. Stationary states are established and oscillatory states sustained indefinitely in this system. The effect of small quantities of added hydrogen is studied by a carefully controlled, continuous supplement to the principal reactants. Four different modes of reaction (I-IV) have been characterized, and conditions for their occurrence mapped on a reactant pressure-vessel temperature ( p - T a ) ignition diagram. Most boundaries are quite sharp, and some show evidence of hysteresis. Close to the axes, reaction is slow, non-luminous and non-oscillatory (I). Within a first broad promontory (II) reaction is accompanied by steady luminescence. Crossing the boundary is not accompanied by a step change in reaction rate, but there is a change in character from stable node (in I) to stable focus (in II). Auto-oscillatory luminescence occurs in a closed region (III) wholly within the promontory II. The effects of adding hydrogen on all these modes is to increase the reaction rates markedly and to make them non-isothermal; the boundaries between I, II and III are not as greatly affected. However, systems to which more than 0.10% H 2 have been added also display a new mode, of oscillatory ignition. This appears at first in a region (IV) of high temperatures and pressures but as more H 2 is increased its realm expands and it eventually dominates the ignition diagram, invading the region of luminescence and soon obliterating the oscillatory part completely.

Oscillatory ignition of the H 2+O 2 and CO+H 2+O 2 reactions in open systems

In closed vessels, branched-chain processes such as the oxidation of H2 display either slow, steady reaction or a single explosion. These two forms of behaviour are separated by a well-defined p-Ta ignition limit. In open systems, the constant inflow of fresh reactants and matching outflow of products allows ignition to be a repetitive process. Again there are pressure-temperature (p-Ta) limits for this oscillatory ignition. We may derive a simple expression for the location of the upper, or second limit for H2+O2 mixtures in a well-stirred, flow reactor (c.s.t.r.). The condition for the onset of explosion is clearly related to the classical result for a closed vessel (pcr/RT)=(2k2/k4) with a correction term proportional to the flow-rate or mean residence time in the reactor. Good agreement is found between the predicted limit and that observed experimentally. This provides a solid foundation for the interpretation of the effects on the p-Ta limit location of replacing H2 by CO. Initial replacement enhances ignition, expanding the explosion peninsula. The upper limit for a system containing 10% H2 and 40% CO lies at lower temperatures than that for an equimolar H2+O2 mixture. If more H2 is replaced by CO, so that the system contains less than 1% H2, ignition is inhibited. The peninsula moves to higher temperatures. For mixtures containing less than 0.1% H2, the limit lies beyond 840 K. These two effects, enhancement and inhibition, can be simply correlated in terms of the H2+O2 branching cycle of reactions and two additional elementary steps. The expulsion of the oscillatory ignition peninsula leads to an uncovering of p-Ta regions of two extra patterns of reaction: steady and oscillatory chemiluminescence (glow).

Spontaneous ignition of hydrocarbon and related fuels: A fundamental study of thermokinetic interactions

A numerical interpretation of thermokinetic interactions leading to oscillatory cool flames and complex, multiple-stage ignitions in acetaldehyde oxidation is presented. The results are well matched to quantitative, experimental measurements in a well-stirred flow reactor; they are the first numerical predictions of oscillatory ignition phenomena in a non-isothermal system with chain branching. The kinetic foundation comprises 25 reactions fulfilling three principal roles. These are: initiation and “low temperature” degenerate branching; thermokinetic interaction leading to an overall negative temperature-coefficient of rate and oscillatory cool flames; chain branching and ignition instability due to reactions of the cool flame products. The cool flame phenomena originate from methyl radical oxidation and degenerate branching via methyl hydroperoxide. Formaldehyde is the principal pre-cursor of hot ignition, but chain branching is augmented by hydrogen peroxide decomposition. Very substantial heat release in the final stages is due to carbon monoxide oxidation by hydroxyl radicals. These are the first clear indications of how spontaneous, two-stage ignition may develop in the end-gas during the spark-ignition engine cycle, and thus they provide a foundation from which a general thermokinetic model to interpret engine knock may be developed.

Ignitions, extinctions and thermokinetic oscillations accompanying the oxidation of ethane in an open system (continuously stirred tank reactor)

This paper is concerned with the spontaneous oxidation of ethane and oxygen at constant pressure (1 atm) in a well stirred, continuous flow reactor (0.5 dm 3 ) over the range of vessel temperatures from 500 to 800 K and at mean residence times of 15, 30, 60 and 120 s. The reactant compositions are varied in the stages C 2 H 6 :O 2 of 8:1, 6:1, 4:1, 2:1 and 1:1, and these are chosen to match the compositions at which cool flame phenomena were observed previously by Knox & Norrish (1954) in closed unstirred vessels and by J. A. Gray (1953) in a laminar flow tube. The work is concerned principally with the thermokinetic modes that occur and the nature of transitions between them, when the vessel temperature is varied at each of the compositions and residence times. Four distinct reaction modes are found; two stationary states and two oscillatory states. The oscillatory states are associated only with the leanest compositions (C 2 H 6 :O 2 < 2:1) and fastest flow rates ( t res = 15 s). One mode is characteristic of oscillatory cool flames, the other is an oscillatory ignition. Both stationary reaction modes are found in all compositions at all residence times, one occurring at higher vessel temperatures than the other and distinguishable by its great reactivity and an accompanying stationary luminescence (CH 2 O*). In the most reactive systems (low C 2 H 6 :O 2 , low t res ) the luminescent state is entered discontinuously and although the temperature rise associated with it is not great (∆ T < 60 K), it is found to be consistent with that predicted for complete combustion of the rich ethane and oxygen in a non-adiabatic flow system. This study reveals some features of ethane oxidation that match the predictions of quite simple thermal theories for combustion and others that are a product of complex thermokinetic interactions. Quantitative tests are made possible by the assessment of heat release rates from temperature measurements by using fine-wire thermocouples and from overall stoichiometries obtained by chemical analysis.

Branched-chain reactions in open systems: theory of the oscillatory ignition limit for the hydrogen+ oxygen reaction in a continuous-flow stirred-tank reactor

The oxidation of hydrogen is the classic example of an ‘isothermal’, branched-chain reaction, and it is studied here from experimental and theoretical standpoints as the natural prototype of branched-chain reactions in open systems. With an inflow of reactants and a matching out-­flow of products, ignition now occurs as a repetitive, oscillatory sequence of events. By identifying the critical conditions as the marginal loss of stability of the stationary-state reaction, a simple criterion for ignition can be derived. This criterion is seen to be a generalization of the elementary treatments, going over to the classical results for closed vessels in the limit of zero flow-rate; it illustrates the stabilizing effect of opening the system. The experimental location of thep─Talimit for an equimolar H2+O2mixture in a continuous-flow, stirred-tank reactor (c. s. t. r.) reported here, shows good agreement with the new predictions (from a simple isothermal kinetic model). Extensive measurements of extents of reactant consumption and of the (small) degree of self-heating are also presented. These lead to rates of reaction and rates of heat release. We show how these are related under conditions of steady-state (non-explosive) reaction and, hence, how accurate measurements of the small temperature-excess can be used to give measurements of the reaction rate.

Oscillatory ignitions and cool flames accompanying the non-isothermal oxidation of acetaldehyde in a well stirred, flow reactor

Oxidation, cool flames and ignition in equimolar mixtures of acetaldehyde and oxygen have been studied in a well stirred, continuous-flow reactor (an open system), over temperatures from 450 to 625 K and pressures from 5 to 25 kN m -2 . The reactor is a 500 cm 3 , spherical, glass vessel, ca . 10 cm in diameter, and is stirred mechanically at a revolution rate of up to 1200 min -1 . Residence times can be varied down to a few seconds; our work relates to 10, 7, 5 and (mainly) 3 s. These conditions broadly resemble those in which transient cool-flame phenomena can be seen in closed vessels, and for which steady-state calculations have been made for open systems. Open systems, however, allow unlimited numbers of oscillations to be observed, and stationary states to be maintained indefinitely. The emphasis in the present work is on establishing the variety of behaviour, and on characterizing the new modes of reaction possible in open systems by quantitative and continuous measurements. Concentrations of reactants, products and molecular intermediates, and hence rates of reaction, are monitored continuously by a mass spectrometer; light emissions are monitored instrumentally; self-heating and hence rates of heat evolution are detected and recorded by measuring excess temperatures with fine-wire thermocouples. An unprecedented variety of behaviour has been encountered. Nine chemically and physically distinct, stable modes of reaction have been observed. There are stable oscillations (limit cycles) of seven clearly differentiated forms, and two stationary states. The conditions for the occurrence of each mode have been mapped for various residence times on a traditional ignition diagram of reactant pressure and vessel temperature. They occupy five regions (representative, overlapping reactortemperature ranges in parentheses): I, steady reaction without light emission (up to 550 K); II, oscillatory ignition (500-520 K); III, five modes of oscillatory ignition interspersed with cool flames (520-540 K); IV, oscillatory cool flames (500-600 K); V, steady reaction with chemiluminescence (580-650 K). At the various boundaries between the five regions, sharp jumps occur from one kind of behaviour to another. At three segments of the boundary, there is hysteresis, the jumps occurring at different temperatures during heating (I -> II, III or IV) and cooling (II, III or IV -> I) traverses. There are thus regions of bistability, where identical external conditions - vessel temperatures, reactant pressures and flow rates - can give rise to alternative states inside the reactor. The two non-oscillatory, stationary states have different characters: I is a stable node and V is a stable focus. In region I, the reaction rate increases with temperature; but in region V, both reaction rate and extent of self-heating show a near-zero or negative temperature-coefficient.