ebook img

Combustion and Diffusion Flames at High Pressures to 2000 bar PDF

6 Pages·1988·1.392 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Combustion and Diffusion Flames at High Pressures to 2000 bar

W. Schilling and E. U. Franck: Combustion and Diffusion Flames at High Pressures to 2000bar 631 (Poltz and Jugel [19], Riedel [20], Tufeu et al. [21], Rastorguev [12] I. F. Golubev and T. Vasilkovskaya, Teploenergetika 16,77 and Ganiev [22], Filippov [23] and Vargaftik [24]) were taken (1969)(in Russian). with steady stateinstrumentsand the deviationsextendto ±10%. [13] G. M.Malian,M.S.Michaelian,and F.J.Lockhart,J.Chern. and Engng. Data 17,412(1972). Thework describedinthispaperwaspartiallyfinanced byE.E.C. [14] J.Taborek,in:D.T.Jamieson,J.B.Irving,and 1.S.Tudhope, GrantST2J-0030(StimulationAction),whichisgratefullyacknowl "Liquid Thermal Conductivity. A Data Survey to 1973", edged. HMSO, Edinburg 1975. References [15] W.Jobst, Int. J. Heat Mass Transfer 7,752(1964). [1] C.A.Nietode Castro,S.F. Y.Li,A.Nagashima, R.D.Tren [16] R. Mauch, Ph. D.Thesis, ETH Zurich 1959. gove,and W.A.Wakeham,J.Phys. Chern. Ref.Data15,1073 [17] J. E. S. Venart and C. Krishnamurty, Proc. 7th Conf. on (1986). Thermal Conductivity, Gaithesburg, Maryland 1967. [2] E.Charitidou, M. Dix, M. J. Assael, C. A.Nieto de Castro, [18] R. W. Powell and H. Groot, Int. J. Heat Mass Transfer 15, and W. A.Wakeham, Int. J.Thermophys.8,511 (1987). 360(1972). [3] M.J.Assael, E.Charitidou,C.A.Nietode Castro,and W.A. [19] H. Poltz and R. Jugel, Int. J. Heat Mass Transfer 10,1075 Wakeham, Int. J.Thermophys. 8,663(1987). (1967), [4] E. Charitidou, R. Molidou, and M. J. Assael, Int. J. Ther- [20] L. Riedel,Chern. Ingr. Tech. 23,321(1951)(inGerman). mophys. 9,37(1988). [21] R.Tufeu, B. Le Neindre, and P. Johannin, C. R. Acad. Sci. [5] J.J.deGroot,J.Healy,andJ.Kestin, Physica92A,102(1970). Paris, Ser. B.262,229(1966). [6] J. Menashe, Ph. D.Thesis, Imperial College, London 1980. [22] Yu. L. Rastorguev and Yu. A.Ganiev, Inzh-fiz,Zh. 14,689 [7] Y.NagasakaandA.Nagashima, Rev.Sci.Inst. 52,788(1981). (1968)(in Russian). [8] D. A.Vermilyea, Acta Met. 1,282(1953). [23] L. P. Filippov, Vestnik. Mosk. gos. Univ. Ser.3,Fiz.Asrton. [9] W.A.Wakeham and M. Zalaf, Fluid Phase Equil. (in press). 2,43 (1960)(in Russian). [10] J. V. Sengers, J. T. R. Watson, R. S. Basu, and B. Kamgar [24] N. B.Vargaftik, Proc. Conf. on Thermodynamik and Trans Parsi, J. Phys. Chern. Ref.Data 13,893(1984). port Properties ofFluids, London 1957. [11] C. A. Nieto de Castro, S.F. Y. Li,A.Nagashima, and R. D. Trengove,J. Phys. Chern. Ref.Data 15,1073(1986). (Eingegangenam 11.August 1987) E 6704 Combustion and Diffusion Flames at High Pressures to 2000 bar w. Schilling andE. U. Franck Institutfur PhysikalischeChemie und Elektrochemie, Universitat Karlsruhe, KaiserstraBe 12,7500Karlsruhe Chemical Kinetics / Combustion / Flames / High Pressure / Supercritical Fluids The combustion of methane with oxygen in supercritical homogeneous aqueous fluids has been investigated and stationary diffusion flamesgenerated to pressuresof2000bar. Areaction autoclave with sapphirewindowscontainshigh pressure homogeneous mixturesof waterandmethaneto 500°C. Atypical mixturecompositionis70to 30molepercentofH20 and CH4•Aquasi-circularfluidflowpermits the steady injection of oxygen through a 0.5 mm nozzle at rates of1-10mrrr'S-1 at constant pressures. - Above 400°C spontaneous ignition offlames occurred. The flames were observed and recorded with microscope and video camera. Emission spectra in the UV region wereobtainedandsamples could betakenforanalysis. Belowabout400°Cflame-lessoxidationisdetected. Thestationarydiffusion flames are cone-shapedand typically about 3mm high. Flameexamples for pressures between 300and 2000bar are shown. Preliminary temperatures derived from OH-spectra are close to 3200K. - Watercan be replaced by argon. I. Introduction phases are homogeneous and the components are miscible Recent experimentalresults on thermodynamicproperties at all densities. ofhigh pressuresupercriticalfluidshave opened up the pos Itcould be expected, that combustion reactionsand pos sibility to study combustion and flames at very high pres sibly flames can be produced in such dense supercritical sures and in unusual environments. Stationary diffusion mixtures.Technicalaspectsof"hydrothermal" oxydationat flames have been produced up to 2000barindense aqueous moderate pressures have already been tested and discussed mixed fluid phases. [7,8]. The study of combustion and flames in supercritical Investigations of phase diagrams of several binary phases offers several possibilities: 1.The variation of pres aqueous systems [1] have provided knowledge of phase sure over wide ranges shouldinfluence reactionmechanisms equilibrium surfaces, critical curves and supercritical ho andflamecharacteristicsbecausethedensitycan bechanged mogeneousphasesto temperaturesof400°Cand aboveand fromlow,gas-like, tohigh,liquid-like,values.2.The variable to pressureshigherthan 2000bar. Among these systems are temperature of the dense, fluid environment can have an H [2], H [3], H [4], H [5] influence on reactionsandflames.3.Thechemicalandphys 20-H2 20-N2 20-02 20-CH4 and H [6]. These systems have critical curves, ical character of this environment can be varied consider 20-C02 which begin at the critical point of pure water (374°C, 221 ably, for example by using supercritical water as the major bar) and proceed to higher pressures in the three-dimen component, as in the present experiments. Certainly, the sional pressure-temperature-composition (PTx) space. At knowledge of transport coefficients of gases involved is de temperatures higher than those along these curves the fluid sirable. For water the viscosity has been determined to Ber.Bunsenges. Phys. Chern. 92,631-636(1988) - © VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1988. 0005-9021/88/0505-0631 $02.50/0 632 W.Schillingand E. U.Franck: Combustion and DiffusionFlames at High Pressures to 2000bar 500°C and 5000 bar [9] and the thermal conductivity to 300°C and 3000 bar [10]. NMR-relaxation times were measured to 700"C and 1500 bar [11]. Thus estimates of diffusion coefficientscould be feasible. Various kinds of information can be expected from the high pressure combustion and flameexperiments: Reaction kinetics data forconditions ofveryhigh collision rates. Re sults about combustion products obtained at high density and with the quenching action of supercritical water, with out or with flame formation. Flame ignition temperatures in the high pressure aqueous phases and the ranges of sta bility can be determined as well as flame size, shape and perhapstemperature. Stationarydiffusionflamesat elevated pressures to 10 bar and to 40 bar are described in the lit erature [12-14]. The aim of the present work was to design and operate an apparatus in which stationary combustion and flames can beproducedand sustained to pressures of2000bar and with environmental temperatures up to 500°C. Visual ob servation of the interior of the reaction vessel should be possible.Arrangements had to bemade bywhicha gasflow of only a few microlitres per second could be fed steadily Fig.lb into thereactionvesselat pressures to twokilobar. Asimilar Schematic cross section ofa burner, to be introduced into the re provisionwasnecessarytoextractsmallsamplesforproduct action cell.A:Threadedscrew.B:Cone to fitinto the reaction cell analysis at constant conditions. The principle ofdesignand opening. C: Injector nozzle with two concentric tubes. D: Twin operation will be described. First results will be given for valvefor the inlet oftwo gases experiments with oxygen introduced into supercritical water-methane mixtures. fluidsisinsignificant.Two highpressure windowsat opposite ends permit observations of the interior. Fig. 1a shows the main parts ofthe reaction cell. II.Apparatus The cylindrical body [1] with80mmo.d.and 30mmi.d.isofa The main part oftheapparatus isareaction cellofabout 30crrr' highstrength nickel-base superalloy1). It isclosedat both ends by sample volumewhichcan holdasample pressure ofupto2000bar sealingcones[2] madefrom a similaralloy.The conesare pressed at 500°Cand higher for periods ofhours. Corrosion withaqueous into position by threaded screws[3]. The tube-like inner parts of theconeshaveplane,polishedsurfacesverticaltothecellaxis.Onto thesesurfacesthe polished,flatsurfacesofthe two cylindricalsyn theticsapphire windows [4] of15mm diameter and 10mmthick nessare pressed.Stainlesssteelcaps keepthe windowsin position. Thefreeapertureis8mm.Thereactioncellishorizontally mounted. The burner (seeFig. 1b)is introduced through the lower opening [5]. Three otheropenings (one[6] shown in Fig. 1a)areinletsfor asheathed chromel-alumel thermocoupleandfortwostainlesssteel capillaries to extract samples and to connect the reaction cellwith thefeedautoclaves(seebelow).Thecellhasexternalelectricheating elements,additional thermocouples and thermal insulation. The burner isshown in Fig. 1b. Its screw(A)and cone (B)fit into the cellopening [5] (seeFig. 1a).The injector nozzle(C) has 5- 6 two concentric tubes to inject twogasessimultaneously, ifdesired. The outer tube isofstainless steelwith3mm o.d.and 1.5mmi.d. The inner capillary ofthe same material, which protrudesslightly, --~ has0.8mm o.d.and 0.5mm i.d.Bothconcentric tubes can receive --1 gasseparately fromthe twin valve(D)below,whichcan beheated -2 to 50°Cto prevent formation ofsolidgas hydrates. Thefollowingprocedure servestoproduceaslow,controlledand steady gas flow through the burner nozzle into the reaction cell: Two additional cylindrical "feedautoclaves" fromanon-corrosive high strength steel alloy, each with 80 em"internal volume and inletsat both ends are used.Bothcontain stainlesssteelbellowsof 30crrr'capacity, connected with one ofthe inlets.The bellowscan befilledwithmethane,oxygenoranyother gastopressuresof2000 bar, provided that the space outside the bellowsisfilledthrough the second inlets with water and brought to the same pressure. Fig.la These water-filledspaces of the feedautoclaves can be connected Reaction cell.1:Cylindrical body with 80mmo.d.and 30mm i.d. with the interior of the reaction cell,when this is filledwith the 2:Sealingcones.3:Threadedscrews.4:Sapphire windows.5:Open ing for the introduction of a burner. 6: Openings for a sheathed thermocouple and two capillaries ATS340W.Nr. 24969. 1) W.Schillingand E. U. Franck: Combustion and Diffusion Flames at High Pressures to 2000bar 633 reaction fluid. The closed ends of the bellows are connected with compressionrates inthe room temperature oxygenfeedau polished stainless steel rods, which lead through a high pressure toclave. The actual volume and linear flow rates into the packing to the outside of the feed autoclaves. With special gears, hot reaction cellhad to becalculatedusingcelltemperature, hightorqueelectric motorsand electronic controlsthe bellowscan nozzlediameter and oxygen equation ofstate data [15]. be compressed slowly and steadily. A corresponding amount of pressurized gas isled from the bellows through the burner nozzle At temperatures below 400°C no flames were observed. into the sample in the reaction cell at almost isobaric conditions. Athigher temperaturesstationaryflameswereformedat the Anequivalentamount offluidisdriven into the water-filledspaces tip of the oxygen nozzle, depending on the pressure of the ofone or both ofthe feedautoclaves. Thus the fluidin the system reaction cell.No electric spark or other means were neces performsacircularmotion: Gas from the bellowsinto the reaction celland aqueous fluidfrom this cellback into the feedautoclaves. sary. The flame ignition started spontaneously. The flame A very slow, steady gas flow at very high pressures is achieved and the combustion space can be illuminated from behind without pumps inthefluidstream. When the bellowshavereached with a simple lamp giving diffuselight. maximum compression, they have to be refilledand extended. Ex Flame sizescan be determined with ocular scales of the periments with steady flowup to one hour are possible. microscope. One ofthe microscope tubes carries either the In order to extract small fluid samples from the reaction cellat high temperatures, a capillary with a tip of sintered alumina, can ordinary or the video camera connected to monitor and be introduced sidewise into the center of the reaction cell. The recorder. Typically a 20-foldenlargement was used for the samples pass into a section ofstainless steelcapillary between two flame supervision. Samples from the reaction cellcould be high pressure valves mounted on a heated metal block. The cap extracted either with a thin alumina capillary close to the illary section serves as a pipette from which liquid and gaseous components can be extracted for analysis. Another method is,to flameor from an outlet at the top of the inner space ofthe usea small spindle press. Byturning it back veryslowly,samples reaction cell. - The samples, usually of0.5to 1cnr' volume can besucked out steadily from the reaction cell. wereexpanded and cooled with the valveand capillary set For visual observation ofthe cellinterior through the sapphire described above. Gaseous and liquid parts could be sepa windows a lamp mounted behind one end is used. A mirror and rated and analyzed by gaschromatography. stereo microscope2)at theotherend facilitate the observation.The microscope isequipped witha normalcameraor a videocamera3). Normally the phenomenawithinthecellarecontinuouslyobserved and controlled with video camera and colour monitor. A video IV.Observations andResults recorder serves for documentation, for inspection of short time Within the frame ofthe present firstseriesofexperiments processesand forthe productionofstandingflamepictures forsize and shape determination. Instead ofthe microscope a Jarrell-Ash it wasalmost always oxygen whichwasinjected into super .diodearray rapid scan spectrometercan beattached to the cellto critical water-methanemixtures. There wereseveralreasons obtain flamespectra in the visibleand UV-regions. for this first choice. One of these was the desire, to study richflamesand their possibleproductsfirst.Often thewater III.Procedure to methane mole fraction ratio was0.7to 0.3.But mixtures For a first series of investigations an initial fillingof the down to a methane mole fraction of0.1 were also used. It reaction cell consisted of supercritical water-methane mix was possible, however, to inject oxygen and methane si tures, typically of about 70 mole percent of water and 30 multaneously into the supercritical water and produce a flame. Not possible was the production of true premixed molepercent ofmethane. Atroom temperaturea calculated flames.After a retraction of the thin inner nozzle capillary amount ofwater was introduced into the cell.The cellwas closed and heated, for example to 450°C and brought to a ofthe burner(seeFig. 1b)the twogasescould bemixedjust predetermined pressure. Subsequently the methane was below the reaction cell, but the flame reaction proceeded pressedinto the hot cellby using one ofthe feedautoclaves fromthenozzletipinthecellback towardsthismixingpoint immediately. described above. Atpressures ofseveral hundred baror one kilobar, theintroducedmethanedid not distributeitselfvery With the injection ofoxygeninto thewater-methanefluid quicklyinthedensesupercriticalwater.At500°Cthemixing aflameappearedspontaneouslyat 1000bar at temperatures process took about 10 minutes. At 400°C between 30 and as low as 400°C. At 500bar ignition occured at 405°C and 60 minutes were necessary to observe visually a homoge at 200 bar at 420°C. Since ignition temperatures reported neous fluid phase. - If small amounts of water from the for thermal ignition ofmethane-oxygen mixtures at normal reaction cell penetrated into the adjacent capillaries and pressures are above 550°C, depending on the conditions valves,the formation of solid gas hydrates obstructed the [16,17], the ignition temperatures observed here are com gasflow.Gas hydrateformation in peculiar shapes wasalso parativelylow.The highpressuresdecreasetheignition tem seen in the reaction cell, when methane or oxygen were peratures in spite of the quenching effectof the dense su introducedinto highpressure water near room temperature. percritical water. Preliminary ignition experiments without Although the concentric nozzle(seeFig. 1b)and the twin water inthesame pressure range indicatedignition tempera set offeedautoclaves permitted the simultaneous introduc tures about 15- 20°Clowerthan those givenabove.Amore detailed discussion of ignition phenomena shall be given tion of two gases into the reaction cell,during the present elsewhere. experiments only one gas wasintroduced at a time.In most The figures 2 to 5 show photographs of typical flames cases this was oxygen into a supercritical water-methane mixture. The oxygen flow was adjusted to constant values produced byoxygen injection into water-methane phases at between 1 and 6 mrrr' s determined from the bellows 450°C and pressures of 300,600, 1000and 2000 bar. The -1, flowvelocityasgivenbytheoxygenfeedautoclave'sbellows Leitz-Wild, M3. compression at these pressures and room temperature was 2) 3) Bauer, VCE 25. about 3 mrrr' S-l for the first three pressures and about 634 W. Schilling and E. U. Franck: Combustion and Diffusion Flames at High Pressures to 2000 bar Fig. 2 Fig. 4 300 bar flame, 1.2 mm high and 0.5 mm wide at base. Oxygen 1000 bar flame, 3.2 mm high and 0.5 mm wide at base. Oxygen injected at 3 mrrr's:I (at room temp.) into a homogeneous 7:3 injected at 3 mrrr' S-1 (at room temp.) into a homogeneous 7:3 water-methane mixture at 773 K water-methane mixture at 773K Fig. 3 Fig. 5 600 bar flame, 2.5 mm high and 0.5 mm wide at base. Oxygen 2000 barflame, 3.7high, 0.5mm wide at base. Oxygen injected at injected at 3 mm/ S-1 (at room temp.) into a homogeneous 7:3 2rnm!S-1(at roomtemp.) intoahomogeneous7:3water-methane water-methane mixtureat 773 K mixture at 723 K 2 mrrr' s for 2000 bars. Flickers and eddies along the Schumann[19] withthemodification ofJost [20] theheight -1 flamesare not observed. No significant appearance ofsolid h ofthe flamecone is carbon isseen.The flamesare nearly conical in shape. The gas funnelabove the flamesiscaused by hot, lessdensegas V h= emanating from the flame.Differencesin refraction indices 2nD are relatively high and schlieren patterns distinct. It must be remembered, that no phase separations occur at these with V the volume flowand D an effective diffusioncoeffi supercritical temperatures. The color ofthe flamesis white. cient. Sincethe volume flowfor the first three flames(Figs. The flow of oxygen through the inner capillary of the 2,3,4)isaboutequal and that ofthe2000bar flame(Fig.5) burner(Fig. 1b)islaminar. Theestimated Reynoldsnumber is similar, it is understandable, that the flame height in inthis regionfor1000bar (Fig.4)isabout 200,much below creaseswithpressuretoanextent,whichqualitatively agrees the critical number for turbulence. This is also true for the with the expected decrease ofD with pressure. other pressures investigated. The flamescan clearly becon The effective temperature in the reaction zone isnot suf sidered as diffusion flames. Because of their conical shape ficientlyknown. The environment temperature around the the conventional simplified treatment of laminar diffusion flames in the reaction vessel was 773 and 723 K for the flames can be applied [16-18]. According to Burke and above shown flames. For a test of the above equation a W.SchillingandE.U.Franck:Combustion andDiffusion Flames at HighPressures to2000 bar 635 somewhathigher temperatureof900 K isused.This means, thatthe oxygenvolume flowrate isapproximately9 x 10-3 ~ WAVELENGTH em' s-1.Iftheflameheight at 1000baris0.3em,oneobtains an effectivediffusion coefficient of 3 x 10-3em' S-I. This appears to be a reasonable number. Only a very approxi mate comparison with a calculated diffusion coefficient is possible,because ofthe obvious uncertaintiesaboutspecies, density and temperature in the reaction zone. Basic data were taken from Hirschfelder et al. [21]. An average diffu ~~. sion coefficientfor one atmosphere, room temperature and gases like H20, O2 and CH4 can be taken as 0.22em?s-1. ~li~1A1'. It is assumed, that this number increases with temperature proportional to T1.75 at 1bar of pressure. At moderate gas pressures the diffusion coefficient decreases inversely with the increasing pressure. At high density modifications have to be applied which can bederivedfrom the Enskog theory 300 310 320 nm for dense gases and corresponding state considerations of Fig. 6 thegasviscosity.The result isanestimatedeffectivediffusion OH-cmission spectrum ofamethane-oxygen flame at 100bar coefficientat 900 K and 1bar of 1.5em?S-1 and at 900 K and 1000 bar of 1.0 x 10-3 em?S-I. Considering all the uncertainties, this value compares reasonably wellwith the above value of 3.0 x 10-3em?S-I, derived from the flame size.It is easy to raise the calculated value to 3.0 x 10-3 em?s-1 by application of an effectivetemperature for the reaction zone of 2000 to 2500 K. Although the knowledge of species and temperature in the reaction zone is as yet insufficient, the above estimates permit a qualitative con sistent descriptionofthe flameshapesat veryhigh pressure. A profound quantitative discussion on the basis of the ex isting knowledge for counterflow diffusion flames [32] has to await further results. Foranumberofthe highpressureflamesemission spectra havebeen recorded.AJarrellAshspectrometerwith adiode array was used. Scanning time was about 1s.The spectra WAVELENGTH - were recorded by multichannel analyzer, disk and monitor. 300 310 320 nm The part of the spectra in the near UV-region between 300 and 330 nm is mainly caused by emission of OH-radicals Fig. 7 and can be used for temperaturedetermination. Fig.6gives OH-emissionspectrumofamethane-oxygenflame at1000bar(see a recorded emission spectrum in this region for a methane Fig.4) oxygen flameat a pressureof100bar. Fig. 7givesthe same kind ofspectrum at a total pressure of 1000 bar. The pres sure broadening is very pronounced. Instead of many in on conditions, such as flow rate, environment and overall dividual lines only two main wings can be observed. The pressure. spectrum is produced by the vibrational 0- 0 band of the It is uncertain to what extent thermal equilibria are electronic 2~ - 2n transition and its six main and six sat achieved in different parts of the flames. - A number of ellite rotational branches. procedures are (in principle) available to determine flame Anumberofanalyticdeterminationsofreactionproducts temperatures: The immediate measurement,for example by bytakingsamples as described above have been made. The thermocouples,thethermochemicalcalculation,linereversal results show, that the oxygen consumption in the flame is methods for electronic excitation temperatures, determina nearly complete, with CO and CO as major products. At tion of vibrational or rotational temperatures. In addition 2 temperatures below flame ignition, "cold" combustion was more recent methods likeadvanced Raman techniques may also detected with a similar range of products. be applied. Inthe present caseonly the determinationofa rotational temperature based on the OH-radical spectrum has been V.Discussion and Outlook used. The fundamental data of all necessary OH-bands in A particularly important characteristic of a flame is the the UV-region are given by Dieke and Crosswhite [23]. temperature. It is not always possible, however, to define a Methodsofthe temperaturecalculationare described inthe temperature sufficiently. For the high pressure flames dis literature by several authors [24- 26]. Here the procedure cussedhere,temperaturesmay bequite differentinlocations of Eisenreich and Schneider [27] has been applied. At first within the flameand the temperatures willcertainlydepend the rotational line intensities of the six main and satellite 636 W. Schilling and E. U. Franck: Combustion and Diffusion Flamesat High Pressures to 2000 bar branches are calculated with the available data for energy References levelsand transition probabilities [19]. Every line is given [1] E.U. Franck,IUPAC Rossini Lecture,J.Chern.Thermodyn. a Gauss profile.The total intensity in dependence of wave 19,225(1987). length can then be calculated as a function of temperature [2] T.M.SewardandE.U.Franck,Ber.Bunsenges. Phys.Chern. and a certainselectedconstant halfwidth ofthe Gauss pro 85, 2(1981). files. These two quantities: temperature and half width are [3] M. L.Japas and E.U. Franck, Ber. Bunsenges. Phys. Chern. 89,793 (1985). the two parameters to beadjusted to the experimental high [4] M. L. Japas and E. U. Franck, Ber. Bunsenges. Phys. Chern. pressure spectrum. The procedure was successful to deter 89, 1268(1985). mine flame temperatures of solid fuels.In the present case [5] H. Welsch, "Die Systeme Xe-H 0 and CH 0 bei hohen 2 4-H2 the best fit to a spectrum of 700 bar was obtained with a Driickenund Temperaturen".Thesis,lnst.forPhysicalChern., University ofKarlsruhe 1973. ~A temperature of 3400 K and a Gauss half width of [6] K. Todheide and E. U. Franck, Z. Phys. Chern. Neue Folge . 37,387(1963). 0.32 nm. The same treatment of the 100 bar spectrum re [7] M. Modell, G. C. Gaudet, M. Simson, G. T. Hong, and K. ~A Bieman, Solid Wastes Management,August(1982).S.H.Tim sulted in T = 3100 K and = 0.125nm. A preliminary berlake,G. T.Hong, M.Simson,and M. Modell, SAETechn. Pap. Ser. 1982,No. 820872. thermochemical calculation ofthe adiabatic temperature of [8] R.K. HellingandJ.W.Tester,J. Energy Fuels, 1,417(1987). the 700bar flame,burning in an 70- 30water-methaneen [9] K. H. Dudziak and E. U. Franck, Ber. Bunsenges. Phys. vironment of400°C gave T ~ 3100 K. These calculations, Chern. 70,1120(1966). [10] F.J. Dietz, J.J.de Groot,and E.U. Franck, Ber.Bunsenges. however, are based on as yet insufficientdata. Phys. Chern. 85, 1005(1981). The spectrum of Fig. 7 was taken from an area ofabout [11] W.J. Lamb,G. A.Hoffman,andJ.Jonas, J.Chern. Phys. 74, 0.1mm squared in the middle part ofthe 3mm flame.It is 6875(1981). expected, that it will be possible to obtain spectra at three [12] J. Diederichsen and H. G. Wolfhard, Proc. Roy. Soc.236,89 (1956). or four other positions and discuss temperature variations. [13] W. L. Flower and C. T. Bowman, 21. Int. Symposium on Application of thermocouples will not be possible because Combustion, No. 157,1986. of the high temperature and the small flame size. Other [14] H. Eberius, Th. Just, and Th. Kieck, Bericht JB 442-86/1, methods, however, willbe attempted, like line reversal and Institutfur PhysikalischeChemieder Verbrennung, DFVLR, improved thermochemical calculations. For the latter pur Stuttgart 1986. [15] N. B. Vargaftik: "Tables on the Thermophysical Properties pose extended analysis of reaction products and transport of Liquidsand Gases", John Wiley, New York 1975. coefficientsare necessarywhichare underway. - Anumber [16] A.G. Gaydon and H. G. Wolfhard: "Flames", 4. Ed. Chap of variations of these diffusion flames are obviously inter man and Hall, London 1979. esting and desirable. Lean flamescan be made by injecting [17] R.A.Strehlow: "Combustion Fundamentals", McGraw-Hill, New York 1984. methaneinstead ofthe oxygeninto the hydrothermalphase. [18] J. A. Barnard and J. N. Bradley: "Flame and Combustion", Methane can be replaced by otherfuels.The extraordinary 2.Ed. Chapman and Hall, London 1985. miscibilityofthe dense supercriticalwater with many other [19] S.P. Burkeand T.E.W.Schumann, lndustr. Eng. Chern.20, gases and liquids suggests a number of different combina 998(1928). tions. Some ofthese possibilities are already being pursued. [20] W.Jost, "Explosions-und VerbrennungsvorgangeinGasen", Springer, Berlin 1939. Atentativeseriesofexperimentshave been made with pres [21] J. O. Hirschfelder, C. F. Curtiss, and R. B.Bird: "Molecular surized argon instead of water. Flames have been obtained Theory ofGasesand Liquids",John Wiley,New York 1954. with similar appearance as those in the hydrothermal fluid. [22] H. Tsuji, Progr. Energy Combust. Sci.,8,93(1982). [23] G. H. Dieke and H. M.Crosswhite,J.Quant.Spectrosc. Rad. We are indebted to the Fraunhofer-Institut fiir Treib- und Ex Transf. 2,97(1962). plosivstolTe,Berghausen/Karlsruhe, in particular to Dr. N. Eisen [24] R. Mavrodineanu and H. Boiteux, "Flame Spectroscopy", reich and Dr. H.Schneider,forinstrumentalsupportand foradvice John Wiley, New York 1965. and assistance with the spcctrophotmetric temperature determi [25] D.B.Vaidya,J.J. Horvath,andA.E.S.Green,Appl. Optics, nation.Wethank Mr.W.BaltzandthestalToftheshopfor excellent 21,3357(1982). work. Financial support from the Stiftung Volkswagenwerk and [26] D. H. Campbell, S.Hulsizer,T. Edwards, and D. P. Weaver, the Fondsder Chemischen lndustrie isgratefully acknowledged. J. Propulsion, 2,414(1986). [27] N.EisenreichandH.Schneider:"Temperaturbestimmungvon FesttreibstolT-Flammen durch Berechnung der OH(O-O) Bande", Bericht 10/86,Fraunhofer-InstitutfurTreib-und Ex plosivstolTe,D-7507 Berghausen/Karlsruhe 1986. (Eingegangen am 21.Januar 1988) E6722

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.