NASA Technical Reports Server (NTRS) 20000091021: Raman Gas Species Measurements in Hydrocarbon-Fueled Rocket Engine Injector Flows

8 Pages·0.46 MB·English

by  NASA Technical Reports Server (NTRS)

AIAA 2000-3391
Raman Gas Species Measurements in Hydrocarbon-
Fueled Rocket Engine Injector Flows
J. Wehrmeyer
Mechanical Engineering Department
Vanderbilt University
Nashville, TN 37235
R. Hartfield
Aerospace Engineering Department
Auburn University
Auburn, AL 36849
H. Trinh, C. Dobson, and R. Eskridge
Space Transportation Directorate
NASA-Marshall Space Flight Center
Huntsville, AL 35812
36 th AIAA/ASME/SAE/ASEE Joint Propulsion
Conference & Exhibit
17-19 July 2000
Huntsville, Alabama
Raman Gas Species Measurements in Hydrocarbon-Fueled Rocket Engine Injector Flows
Joseph A. Wehrmeyer*
Mechanical Engineering Department, Vanderbilt University, Nashville, TN 37235
Roy J. Hartfield, Jr]
Aerospace Engineering Department, Auburn University, Auburn, AL 36849
Huu P. Trinh, _Chris C. Dobson, _and Richard H. Eskridge _
Space Transportation Directorate, NASA-Marshall Space Flight Center, Huntsville, AL 35812
Abstract Hardware associated with this effort includes
Rocket engine propellent injector an optically-accessible high pressure combustion
development at NASA-Marshall includes experimental chamber sized for single-element fuel injectors
analysis using optical techniques, such as Raman, (Unielement Test Articlel and a newer, larger Modular
fluorescence, or Mie scattering. For the application of Combustion Test Article (MCTA_ that can
spontaneous Raman scattering to hydrocarbon-fueled accommodate multi-element fuel injector
flows a technique needs to be developed to remove the configurations and that is also optically-accessible.
intert'ering polycyclic aromatic hydrocarbon Optical accessibility allows laser-based methods, such
fluorescence from the relatively weak Raman signals. as laser Mie scattering, fluorescence, and spontaneous
A current application of such a technique is to the Raman scattering, to be applied. Raman spectroscopy
analysis of the mixing and combustion performance of has been used to analyze an oxygen-rich gaseous
multijet, impinging-jet candidate fuel injectors lbr the hydrogen/liquid oxygen (GH2/LOx) injector flow in the
baseline Mars ascent engine, which will burn methane unielement test article (1) and most recently is being
and liquid oxygen produced in-situ on Mars to reduce considered lbr use in the MCTA to analyze the mixing
the propellent mass transported to Mars for future and combustion performance of multijet, impinging-jet
manned Mars missions. The Raman technique takes candidate fuel injectors for the baseline Mars ascent
advantage of the strongly polarized nature of Raman engine (2). This engine will burn methane (CH4) and
scattering. It is shown to be discernable from LOx produced in-situ on Mars to reduce the propellent
unpolarized fluorescence interference by subtracting mass transported to Mars for future manned Mars
one polarized image from another. Both of these missions (3).
polarized images are obtained from a single laser pulse
by using a polarization-separating calcite rhomb Application of Raman spectroscopy to
mounted in the imaging spectrograph. A hydrocarbon-fueled combustion brings the issue of
demonstration in a propane-air flame is presented, as interference of the weak Raman scattering light signals
well as a high pressure demonstration in the NASA- with strong laser-induced fluorescence. The
Marshall Modular Combustion Test Artice, using the fluorescence interference can be from polycyclic
liquid methane-liquid oxygen propellant system. aromatic hydrocarbons (PAH's) which are excited and
fluoresce across the ultraviolet and visible spectrum (4)
Introduction and when using ultraviolet lasers can be due to the
Technology development associated with hydroxyl radical (OH) and to vibrationally-excited 02,
advanced space transportation propulsion systems both of which are present in combustion reaction zones
includes design and analysis of new types of propellent (5). By taking advantage of the polarization properties
injectors. An effort exists at NASA-Marshall to of Raman scattering, it is possible to discriminate the
include experimentally-obtained reactant/product Raman signal from interfering fluorescence
mixing/combustion intbrmation as part of the injector interference and hence apply Raman scattering to the
analysis. analysis of hydrocarbon-fueled combustion. This can
be done by obtaining two Raman images, one using a
*Research Associate Prof., Senior Member, AIAA vertically polarized laser and one using a horizontally-
Associate Prof., Senior Member, AIAA polarized laser, and then subtracting the intensity of
- Engineer. Senior Member. AIAA one image from another to obtain a net Raman signal
Physicist, Propulsion Research Center (6). A second method that lends itself to single-pulse
IEngineer. Propulsion Research Center (and hence instantaneous_ measurements is to obtain
Cop)right © 2000 by J. A. Wehrmeyer. Published by two simultaneous Raman images created from the same
the American Institute of Aeronautics and laser pulse. One of these images is the vertically-
Astronautics, Inc. with permission. polarized signal and one is for the horizontally-
polarized signal. The difference between these images
providesthenetRamansignal,freeof fluorescence fluorescence-free UV Raman measurements have been
interferenceT.histechniquheasbeendemonstratfeodr obtained in an iso-octane fueled engine -'_by placing a
a time-averageadpplication(7) andthis paper Glan prism at the focal plane of the imaging
describeassingle-pulsaepplicatioonfthetechnique. spectrograph, and then refocusing the two polarized
spectra onto two separate ICCD cameras. Another
Single-Pulse Polarized UV Raman Measurements method uses a calcite rhomb located at the
While visible lasers may be best at providing spectrograph entrance slit to displace the horizontally-
interference-free Raman data for steady, laminar polarized light from the vertically polarized light and
has been used to obtain time-averaged PAH
flames, it may still be desirable to use a UV laser, when
fluorescence-free UV Raman spectra in C3Hs-fueled
probing turbulent flames, in order to obtain sufficient
flames. 2s This latter method is described here, and
signal strength for single-pulse measurements. By
single-pulse UV Raman images in a C_Hs-air flame are
exploiting the differences in polarization properties
presented.
bet_een Raman and PAH fluorescence, it is possible :o
Figure 14 sho,xs the UV Raman system used
obtain UV Raman signals in PAH-ridden
for polarization-resolved measurements. It represents a
environments.
typical UV Raman system except /'or one unique
In addition to laser wavelength, the signal
feature: a calcite rhomb inserted just behind the
strength of Raman scattering depends on two molecular
entrance slit of the spectrograph. This optical element
invariants, the square of the mean polarizability (a')-"
displaces the horizontally polarized Raman image
and the square of the anisotropic polarizability ('/'V.
about 5 mm from the vertically polarized Raman
These are used to determine a constant _, which can be
image, which travels directly straight through the
considered as the Raman scattering cross section for a rhomb. The length of the entrance slit is limited to 4
single molecule, not including the laser wavelength mm to keep the two polarized images from
contribution. The value of qb also depends on the overlapping. Figure 15 shows a typical single-pulse
polarization of the incident laser beam, the polarization image obtained from the experimental system, and this
of the detected Raman signal, and the angle of detected image shows the relative strength of vertically
Raman signal with respect to incoming laser beam.
polarized signal compared to horizontally polarized
For a 90°collection angle and for the Raman and laser signal. The H20 Raman signal (from air humidity) is
beams both in a horizontal plane, the value of ¢_ for almost completely polarized, while the N, signal is
vertically polarized Raman scattering is:22 slightly depolarized and the O, signal is more
depolarized than N,. This corresponds to others"
O0= (a'}'- + 1/45 (7'_: experimental or._servat-m-ns. 8.26
Spatially-integrated single-pulse Raman
and for horizontally polarized Raman scattering ¢_is: spectra are shown in Fig. 16, obtained from a slightly
premixed C3Hs-air bunsen flame. One of these spectra
(b = t/60 (7')2 is the vertically polarized Raman-fluorescence signal.
The horizontally polarized signal shows essentially
Usually a'-" and y': are similar in magnitude only the fluorescence signal, that has two contributions.
and the vertically polarized light is almost two orders A broadband component is caused by PAH
of magnitude greater than the horizontally polarized fluorescence, extending from below 255 nm to above
light. Thus the Raman scattering signal essentially 275 nm. A second contribution to the fluorescence
retains the polarization of the incoming laser beam. background is the OH fluorescence from about 265 to
Ho_e',er a fluorescence light emission process does 270 nm. caused by tuning the laser slightly onto a
not retain the laser polarization because of the strong OH transition. This is done to demonstrate the
relatively long time the molecule exists in the excited ability of the technique to simultaneously measure
state before fluorescing. During this time the molecule Raman spectra and OH fluorescence. Before
rotates, eliminating any correlation between the subtracting the horizontally polarized signal from the
polarization of absorbed and emitted photons. vertical it is first multiplied by a factor of 2.22 to
By providing two laser pulses, one account for the ratio in transmission efficiency (for the
horizontally polarized and the other vertically spectrometer/calcite rhomb) between the vertically and
polarized, two signal spectra can be obtained, with one horizontally polarized signals. The net signal shows a
containing both Raman and fluorescence and one fluorescence-flee Raman spectrum that reveals the
containing essentially only fluorescence. This method simultaneous occurrence of CO2 and H_,O (products of
has been used to obtain fluorescence-tree, time- combustion). CO and H: (intermediate products), and
averaged UV Raman measurements in hydrocarbon unburned C_Hs. Information about the spatial structure
flames, s-'3 Single-pulse, polarization-resolved. of the flame can be revealed in the single-pulse image
ofFig.17.ThisimagsehowusnburneCd3Ho8ccurring
operational sequence is to introduce CI-I4 without 02.
nearthe0mmposition.Thecoolerd,enseurnburned This shows as a strong CH4 Raman signal occurring in
gasmixturealsoprovideasstrongeNr_R, amasnignal
the image for 1.8 sec. Alter injection of CI-L comes
nearthatlocationA. t~1.5mmtheC3Hp8yrolizeisnto injection of 02 to commence main stage combustion.
otherhydrocarboninsc,ludingPAH'sthatcausaestrip
The images obtained during this part of the sequence
ofbroadbanfdluorescentcoeappeaartthislocation. are all similar to the lower right one in Fig. 2 in that
Fartherintotheflamechemicarleactionisnvolving
there is significant broadband emission occurring
oxidationtakeplace,creatingtheOHintermediate.
across most of the spectrum, and the Raman signals for
ThisshowsupinFig.17bythereplacemeonftPAH
the reactants and products have all diminished because
fluorescenwceithOHfluorescenacteflamepositions of the increased temperature and reduced density
greatetrhan-2.5mm.Bysubtractinthgescaleudpper
during main stage combustion.
partof Fig.17(horizontallpyolarizedfluorescence
signal)fromthebottompart(verticallypolarized The question as to whether the broadband
signal)a,ndb?summintghenetRamasnignafloreach
emission is caused by laser-induced fluorescence, by
location(wavelengitnhtegrationa)qualitativpeicture
direct flame emission, or by a combination was
ofthestructuroeftheflamecanbediscerneads,inFig.
answered by the results of one of the other tests, the
18.ThisfigureshowtshedropinC3Hssignaalndthe
one wich did not have a laser beam entering the
concurrenitncreasein H20signal,showingthe
combustion chamber. The images obtained during that
formationofthatproducotccurringclosetothefuel
test had about one half of the emission intensity as
zone.Inthesameregionthe
H2 and CO signals are from Test 25, indicating that constant flame emission
generally higher than in the fuel zone or in the OH contributes about half of the interference background.
reaction zone past 2.5 ram. The exposure time used for the camera was 200 nsec
and could be reduced down to 20 nsec, the time span of
Hot-fire tests from Modular Combustion Test the laser pulse. Doing so would reduce the detected
Article (MCTA) burning liquid CH4&.__&_O, flame luminosity by a factor of ten.
The polarization-resolved UV Raman
spectroscopy system has been applied to the analysis of Other parts of the hot fire test sequence are
liquid-CH4/LOx combustion in the modular shown in the remaining images of Fig. 3. These are for
combustion test article (MCTA) at Test Stand 115 of the following events: LOx valve closing, LOx valve
NASA-Marshall.. completely off, and both propellents off with only the
N2 purge remaining at the end of the test.
Three hot fire tests were performed that
involved use of the Raman system. Of these three, The utility of the polarization resolved
only one provided Raman data in conjunction with a imaging becomes important when trying to detect
hot fire test. One test resulted in an early cut and thus Raman signals among the unpolarized interference
never provided main stage operation. One test was a background. However when the vertical polarized
successful run of the MCTA, but the Raman system imaged is scaled (to account for transmission efficiency
laser beam did not enter the test article during the test differences in the spectrogaph and calcite crystal) and
because a mirror directing the beam into the test article subtracted from the horizontally polarized imaged, no
was wrongly oriented. However the images obtained net Raman signal is left for the main stage images, one
during this test did provide some useful intbrmation, as of which is shown in Fig. 4. The problem occurs
mentioned below. because there is approximately the same amount of
Raman signal occurring in the vertically polarized
Figures 2 and 3 show selected Raman images image as for the horizontally polarized image, for both
from the successful test. These images are taken from N,_ and for CHa, and probably ['or the other species"
a group of I(X), which were obtained at 5 Hz over a Raman signals. This is completely unexpected, since
total time of 20 sec. The Raman images were initiated the Raman images of Ref. 1, and Raman theory, both
at the beginning of the test, however the Raman images show that significantly less Raman signal should exist
can be correlated to the known events occurring during at one polarization compared to the other.
the test sequence. The first Raman image at time 0 sec
{an arbitrary ret'erence timel shows the N_,purge as the The absence of a net Raman signal for the
only gas present just before ignition. The next image main stage Raman images is a result of at least two
in Fig. 2, at 0._ see.. shoves combustion of H_,and O,_ problems that must be surmounted. The first one
occurring as the igniter becomes operational. This is involves the presence of Raman signals in both
shown by the presence of their Raman signals in polarization images. This problem must be solved
addition to the purge Raman signal. Next in the because the fundamental way that Raman signals are
extractedfrominterferingbackgrounids by their
Thermal and Fluids Analysis Workshop,
highl?polarizecdharacteristicTsh.eseconpdroblem sponsored by NASA-Marshall Space Flight
isthattheRamansignalsareweakcomparetdothe
Center, Huntsville, AL, Sept. 13-17.
back_oundR. amasnignasltrengtmhustbeimproved 3. Nandula, S. P., Brown, T. M., Pitz, R. W., and
relativetobackgrounindterferencdeuringahotfire
DeBarber, P. A., "Single-Pulse, Simultaneous
test.Onewaytodothisistoreducteheexposurtieme
Multipoint Multispecies Raman Measurements in
ofthecamerdaowntothelaseprulselengthw,hichhas
Turbulent Nonpremixed Jet Flames," Optics
alreadybeenmentionedA. lsothesetestswereall
Letters, Vol. 19, No. 6, 1994, pp. 414-416.
performewdithanunfocuseladsebreamw, hichhasa
4. Reckers, W., Hiiwel, L., GrtineMd, G., and
crosssectionof I cmby2 cm. Thebeamwasleft
Andresen, P., "'Spatially Resolved Multispecies
unfocusetodeliminatethepossibilitoyflasedramage
and Temperature Analysis in Hydrogen Flames,"
totheMCTAopticawl indowsw,hichwasaproblem
Applied Optics, Vol. 32, No. 6, 1993, pp. 907-918.
whentheRamasnystemwasapplietdotheUnielement
5. Chen, Y.-C., and Mansour, M. S., "'Measurements
TestArticle(Ref.1t. Bylightlyfocusindgownto,say
acrosssectionof2 mmx 1mm,theRamansignal of the Detailed Flame Structure in Turbulent H,_-
Air Jet Diffusion Flames with Line-
strengtwhillbeimproverdelativetothebackgrounbdy
asignificanatmounat,tleasat factoorftenbecausthee Raman/Rayleigh/LIPF-OH Technique," Twenty-
Ramansignalwill occurovera narrowbandof Sixth Symposium (International) on Combustion,
wavelengthinssteadof a broadwavelengtrhegion. Combustion Inst., Pittsburgh, PA, 1996, pp. 97-
TheRamadnetectiosnystemcanalsobeplacedcloser 103.
to theMCTAopticalwindowto increaseit light 6. Tacke, M. M., Cheng, T. C., Hassel, E. P., and
gatherinpgower.Thisshouldresulitnabouatfactoorf Janicka. J., "'Study of Swirling Recirculating
four increase in signal since it was approximately 20 Hydrogen Diffusion Flame Using UV Raman
inches away and the lens system is designed for a Spectroscopy," Twent3.,-Sixth Symposium
minimum distance of 10 inches between the sample (International) on Combustion, Combustion Inst..
volume and the first light collection lens. Pittsburgh, PA, 1996, pp. 169-175.
7. Frank, J. H., Lyons, K. M., Marran, D. F., Long,
Conclusions and Future Work M. B., Stgrner, S. H., and Bilger, R. W., "Mixture
The critical issue still unresolved, after the Fraction Imaging in Turbulent Nonpremixed
first application of polarization-resolved UV Raman to Hydrocarbon Flames," Twenty-Fifth Symposium
high pressure combustion, is the polarization of the (International) on Combustion, Combustion Inst.,
Raman signal. Do the polarization characteristics of Pittsburgh, PA, 1995, pp. 1159-1164.
Raman scattering change as pressure is increased'? In 8. Schefer, R. W., Namazian, M., and Kelley, J.,
order to determine the effects of pressure on Raman •"CH, OH, and CH4 Concentration Measurements
scattering polarization, a high pressure sample cell is
in a Lifted Turbulent-Jet Flame," Twent3"-Third
being constructed that can be used to both determine
Symposium (International) on Combustion,
the influence of gas pressure and glass properties upon
Combustion Inst., Pittsburgh, PA, 1990, pp. 669-
detected Raman signal polarization.
676.
9. Yeralan, S., Pal, S., and Santoro, R. J., "Major
Acknowled2ments
Species and Temperature Profiles of LOx/GH:
This research is funded by NASA-Marshall
Combustion," AIAA paper 97-2940, 1997.
through a 1999 Summer Faculty Research Fellowship
10. Grtinefeld, G.. Beuhausen, V., and Andresen, P.,
and through Contract # H35139D.
"'Interference-Free UV-Laser-Induced Raman and
References Rayleigh Measurements in Hydrocarbon
Combustion Using Polarization Properties,"
I. Wehrmeyer, J. A.. J. M. Cramer. R. H. Eskridge,
C. C. Dobson. 1997. UV Raman Diagnostics for Applied Physics B, Vol. 61, 1995, pp. 473-478.
Rocket Engine Injector Development. AIAA II. Masri, A. R., Bilger, R. W., and Dibble, R. W.,
Paper 97-2843 presented at the AIAA 33rd Joint "'Fluorescence" Interference with Raman
Propulsion Conference, Seattle, WA, July 6-9 Measurements in Nonpremixed Flames of
Methane," Combustion and Flame, Vol. 68, 1987,
V_ehrmeyer, J. A., and H. P, Trinh. 1999. Raman pp. 109-119.
Spectroscopy fi)r Instantaneous Multipoint. 12. Rabenstein, F., and Leipertz, A., "'One-
Multispecies Gas Concentration and Temperature Dimensional, Time-Resolved Raman
Measurements in Rocket Engine Propellent Measurements in a Sooting Flame Made with 355-
Injector Flows. Paper presented at the 10a'
nmExcitation,"
Applied Optics, Vol. 37, No. 21, Volume Fraction," Applied Physics B, Vol. 59,
1998, pp. 4937-4943. 1994, pp. 1994.
13. Takagi, Y., "'A New Era in Spark-Ignition Engines 21. Will, S., Schraml, S.. and Leipertz, A., "Two-
Featuring High-Pressure Direct Injection," Dimensional Soot-Particle Sizing by Time-
Twenty-Seventh Symposium (International) on Resolved Laser Induced Incandescence." Optics
Combustion, Combustion Inst., Pittsburgh, PA, Letters, Vol. 20, 1995, pp. 2342-2344.
1998, pp. 2055-2068. 22. Wehrmeyer, J. A., Yeralan, S., and Tecu, K. S.,
14. Kee, R. J., Rupley, F. M., Miller, J. A., Coltrin, M. "'Multispecies Raman Imaging in Flames by Use
E., Grcar, J. F., Meeks. E., Moffat, H. K., Lutz, A. of an Unintensified Charge-Couple Device,"
E., Dixon-Lewis, G., Smooke, M. D., Warnatz, J., Optics Letters, Vol. 20. No. 8, 1995, pp. 934-936.
Evans, G. H., Larson, R. S., Mitchell, R. E., 23. Miles, P. C., "Raman Line Imaging for Spatially
Petzold, L. R., Reynolds, W. C., Caracotsios, M., and Temporally Resolved Mole Fraction
Stewart, W. E., and Glarborg, P., "'CHEMKIN Measurements in Internal Combustion Engines,"
Collection," Release 3.5, Reaction Design, Inc., Applied Optics, Vol. 38 No. 9, 1999, pp. 1714-
San Diego, CA, 1999. 1732.
15. Osborne, R. J., "Application of Ultraviolet Raman 24. Long, D. A,, "Raman Spectroscopy," McGraw-
Spectroscopy to Methane-Air vs. Air Counterflow Hill: New York, 1977.
Flames," Master's Thesis, Department of 25. Griinefeld, G., Beushausen, V., Andresen, P., and
Mechanical Engineering, Vanderbilt University, Hentschel, W., "'Spatially Resolved Raman
1999. Scattering for Multi-Species and Temperature
16. Tanoff. M. A., Smooke, M. D., Osborne, R. J., Analysis in Technically Applied Combustion
Brown, T. M., and Pitz, R. W., "'The Sensitive Systems: Spray Flame and Four-Cylinder In-Line
Structure of Partially Premixed Methane-Air vs. Engine," Applied Physics B, Vol. 58, No. 4, 1994,
Air Counterflow Flames," Twenty-Sixth pp. 333-342.
Symposium (International) on Combustion, 26. Knapp, M., Luczak, A.. Beuhausen, V., Hentschel,
Combustion Inst., Pittsburgh, PA, 1996, pp. 1121- W., Manz, P., and Andresen, P., "'Polarization
1128. Separated Spatially Resolved Single Laser Shot
17. Petarca, L., and Marconi, F., "Fluorescence Multispecies Analysis in the Combustion Chamber
Spectra and Polycyclic Aromatic Species in a N- of a Realistic SI Engine with a Tunable KrF
Heptane Diffusion Flame," Combustion and Excimer Laser." Twenty-Sixth Symposium
Flame Vol. 78. 1989. pp. 308-325. (International) on Combustion, Combustion Inst.,
18. Beretta, F.. Cincotti. V., D'Alessio. A., and Pittsburgh, PA, 1996. pp. 2589-2596.
Menna. P.. "'Ultraviolet and Visible Fluorescence 27. Hartfield, R., Dobson. C., Eskridge. R., and
in the Fuel Pyrolysis Regions of Gaseous Wehrmeyer. J., "'Development of a Technique for
Diffusion Flames," Combustion and Flame Vol. Separating Raman Scattering Signals from
61, 1985, pp. 211-218. Back_ound Emission with Single-Shot
19. Miller, J. H., Mallard, W. G., and Smyth, K. C., Measurement Potential," AIAA Paper 97-3357,
"'The Observation of Laser-Induced Visible 1997.
Fluorescence in Sooting Diffusion Flames," 28. Penney, C. M., Goldman, L. M., and Lapp, M.,
Combustion and Flame Vo[. 47, 1982, pp. 205- "'Raman Scattering Cross Sections." Nature
214. PhvsicalScience, Vol. 235. 1971, pp. I10-112.
20. Vander Wal, R. L., and Weiland, K. J., "'Laser-
Induced Incandescence: Development and
Characterization Towards a Measurement of Soot-
o_G_"rA_-CAME'RA W_.TH
/ GATED IHTEN'_F_ER
mFOC&L LENGTH LENS
puLSE9 U_
C_.LCITE RHOMB _l.v_ I_'T_
j J"
j1_. z./
,,=,
N
a,.
--I
<
I--N
o
255 WAVELENGTH, nm 277
-ifN_l@
o t 2 3
_mm
704e *
l
_P&IVlBJEkIOT_ nm