I

AD-759 386

iLASER RAMAN PROBE FOR FLAME TEMPERATURE

Marshall Lapp, et al

Purdue University

Prepared for:

Office of Naval Research (

April !973

DISTRIBUTED BY:

Nftioa To h I-a lufsmts SeU. S. DRPARTMENT OF COiMERCEL55 Port Royal Ro. $pdnjfeMW Va. 22151

PROJECT SQUIDTECHNICAL REPORT GE] -PU

LASER RAMAN PROBE FOR FLAME TEMPERATURE

BY

MARSHALL LAPPICARL M. PENNEY and RICHARD L. ST.% ERS C

GENERAL ELECr 11*0 CO.5CORPORATE RESEARCH4 ANI) DEVELOPMENTI

SCHENECTADY NEW YORK

PROJECT SQUID HEADQUARTERSTHERMAL SCIENCE AND PROPULSION CENTER

PURDUE VN'IVEP-SITY

WEST LAFAYETTCE. 'NO1ANA

APRIL 1973

Project SQUID is a cooperative program of basic research relating to Jet Propul.sian. It is sponsored by the Office of Niaval Research and is administ'pted byPurdue University through Contract N00014-67-A.0226-0005, NR.O98.938.

This documnt has beern approved for pubiic release and sale;

its distribution is unlimi~edJ

______________NATIONAL TECrHNICALIN O!Q,-ATION SERVCE-

-~ * -All

UnclassifiedSecuriw Ciassification-

DOCUMENT CONTROL .ATA R &D~'Stuuty lasiieafonof title, body of abn tract and indexinj4 ai-norafiot, nu.t be' entered when the *ort. rep~ort is elas hued)

. 0f131.ATING ACTIVITY (C10potate rhwr) Ia.. REPORT SECURIT CLASSIFICATIONProject SQUID Headquarters, Thermal Science & PropulA Unclassifiedsion Center, Purdue University, West Lafayette, i ub. Goup

Indiana 47907 N/A3. R.PORT TITLE

LASER RAMAN PROBE FOR FLAME TEMPERATURE

4. DESCRIPTIVE NOTIS (Tk- of report and Inefwri dates)

S AU T"ORISI (Fits; na o. Alidre initial. ls name)

MarshalI LappCarl M. PenneyRichard L. St.Peters

G. OMPRT CAT 7e. OTAL NO. or PAYJs 7b.No. o pirs

Api 1973 4.0*. CONTR CT ON GRAJT.O. 94. O:GlINATOR*S RIOIT NUMBE tRS

NODOl 4-67-A-0226-0005b. P-*,rT &so. GE-1-PU

NR-098-038_C. OT0E4C RIEOCRT HOS, (Any O43hE? Mber Vtat Asr be AS*iSWed

JNone10. OI5TFII34UT|QH 2TATEC*NT

This docunent has been approved for public release and sale;its distribution is unlimited.

i!. SUtPP £MLCNTAV NOTtS 12. SPONSORING MILITARY ACTIVITY

-eis of il,.,erotions in Office of Naval Research, Poier Program,N/A this docm.~t n iav be bettor Code 473, Department of the Navy,

studied on mkcho Arlington, Virginia 22217I). £ !TXIACT

Experimental spectra of Raman Scattering Stokes fundamental vibrationalQ-tranches for nitrogen and hydrogen in hydrogen-air flames have been fit totheoretically-caiculated profiles in order to determine temperature. Theten-perature measured from a least square fit of the fuVl nitrogen profileagreed to within about k% of the value c~lculated from the ratio of the peakintensity of the fi,-st upper state band (i.eo, "hot" band) to that of theground state band. Agreement with the temperature estimated from fine-wirether.ocouple measurements was approximately 2%. Additional theoreticalprofile calculations are given from nitrogen and hydrogen over the widerange of temperatures.

DD 'NO ,1473 (PAGE 1 UnclassifiedS^. 010 t.to.66e0 i security zias;ificatieu

KE XSLINK A LIt4K 3 IN

R3Mn Scattering

Flame ISensorsLasers

FI

DDAE# ;f~~ 73- eACV) 1/Unclassified(PGEsecuity Csification

UASER RAW PROBE XT. PLANE TEDPERATUME

General Electric Co., Corporate Research and Development2 1Schenectady, Now YorkI Subcontcact No. 4965-38

flarshail Lipp, Principal Investigator

Carl M. Penney, PhysicistRichard L. St.Peters, Physicist

Introduction

This workx is directed toward the meaeurement of vibrational

Raman scattering signatures for flme. gases, with a primary view

toward temperature measurements and a cncounitant goal of identifi-

cation and measurement of minor flame species. The development of an

optical probe for these purposes is highly desirable, since advanced

combustion system are utilizing pressures and temperatures such

that physical probes canrnot survive. The use of a Raiman scattering

probe, in particular, offers a variety of adantages over other

optical probes, along with some limitations which will be explored

during the course of this work.

During this project reporting period, experimental effort has

been focussed mainly on: (1) the observation and fitting of FUan

Stokes vibrational Q-branch profiles for N2 and H2 obtained through

use of U -air flames produced on porous plug burners, (2) explora-2

tion of the accuracy of temperature Measurement for N through ex-2

tended thermocouple calibrations and computer data-fitting tecm biques,

and (3) construwtion and preliminary use of apparatus designed for

use in these and other phascs of this research project.

The theoreticel effort during this project reporting period has

been concentrated upon analytical calculations of Raman Stokes vibra-

tional Q-branch profiles, suitably convoluted by experimentally-

determined monochromator slit functions or interference filter ba-nl-

passes. The main effort has involved analyticil procedures necess;ry

to determine the temperature, and has included both least-square data

fitting of entire profiles as well as intensity ratios cf suitably

chosen spectral regions.

I;

I. Experimental Equipment

Vhe basic double mcnochromator exper mental apparatus has been

1described previously. (Ref. I is included here as Appendix 1.) New

additions have been made to the cozmbustion, spectroscopic, tempperature-

measurement and comijuter data handling capabilities in connection with

several parallel programs in our laboratory. Those which pertain

directly to the present research effort are described next in outline

fashion:

(1) A horizontal hydrogen-oxygen burner system utilizing a Meker

burner has been assembled, which will permit flames to be produced

up to ca. 30360 K. This apparatus does not disturb the basic geometry

of our double-monochromator system, since it permits use of a vertical

laser beam passing throujh the flame 1r, the same fashion as forz our

porous plug burner assembly used at ca. 1300-1700 0K. The high-

temperature burner is desiLgned for production of "minor" flame species

of high technological interest, such as OH. Preliminary design studies

have also been carrieci out for am 1liavry optics for u.I.zation with

a vertically burning flame. In this configuration, a line image

(i.e., the scattering zone in the flame) must be rotated through 90

degrees. Design studies include bothi use of two additioanal mirrors

and use& of a Dri 'e prism.

(2) Accurate &f'low metering techniques have been install~ed for

the production of reproducible and clearly-defined fflame conditions.

The flows are now monitored and made steady by criticai flow ozifices

and regulators, u~e being made of precision high-pressure gauges for

accurate control of the fiow rates. The critical flow orifices have

been calibr-ated In our laboratory through use of basic volu~edisp13ce-

ment techniques.

(3) Fine wire thermocouples have been made in our laboratory

for independent measurements of the f la~aa tmiperature by a stenfard

2method. These thermo ,ouples were made of 0.0005 inch diameter wires of

Pt - Pt 10% Rh, coated with quartz so preveGnt catalytic heating. The

thermocouples were nnved throughout. the flame with an accurate vernier

manipulator using, as a ref'trence Vo.ition locator;. a finely-machined

metal cone which could be placed xeproduciblv on the burnez head.

When the burner assembly was then placed in the test position in

front of the spectrometer, the burner could be accurately located in

this same reference position bY placing it so that the laser beam

just touched the cone tip. (By olbserving slight attenuatiofi of the

laser beam with a power meter at the position of the laser dump,j

this positioning could be accomplished with high sensitivity.)

In principle, the ther.icouple-measurement posi ~ior; and laser

Raman scatteriAg position could be made coincident by nmaging the

laser beam on the thermo-ouple junction. However, it is experimentally

dif..icult to accomplish this. In adidition, the therrTtocouples are

relatively fragile, and ft was therefore found ti be advisable to

calibrate the flame before embarking cn the scat :ering measurements.

AddiLiona'L temperature nieasurements have been made with a com-

mercially-available 0.001 inch diameter Pt - Pt 10% Rh thermocouple

with a bead-welded end, str -tched out to the si *e linear geometry as

was used f:-r the 0.005 inch diameter thernocouples described above.

tSee Fic. 1 for schematics of these two t~pen of thermocouples.) This

thermocou-ple was not quartz-clad, but ha. longer leads. The basic

idea was to test the sensitivity of the7:mcouple measurement of tem-

perature to thermocouple geometry, without embarking upon a major

4

FIE

Ii =0 1* 4

JJ4 u 4 a

j4E. 4to u -.

14 -.4

3 0E

u a

Ljii 0 ci

C3 w E 0;)

a 0 tU

41

A( 4:C)$

5A

N--

diversion of efforts The result was that a modest but not very sgr--ft-

cant differencc existed. The smaller thermooouple gave an uncorrected

temperature of 17050 K in a carefully controlled stoichiometric hydrogen-

air flame, while the larger gave 178. These values are the average

of many measurements with only a small variation between measurements.

Since the correction for radiative losses (the 1rgest correction here)

to be applied to rhekmocouple-measured temperatures is ca. 4 times

larger than th s difference of values, the difference is not believed

to be highly significant.

(4) Data logging via paper tape for intensity and wavelength

has been installed in the double monochzrmator system. This apparatus

permits data to be accumulated in a far more accurate and -onve:.ifnt

fashion, making full use of our computer facilities. in Fig. 2 is

shown the electronic detection schematic for this apparatus. The

wavelength data is obtained through use of an optica± incremental

encoder installed on the double monochromator. This systema, no4

operational, is currently being improved by location of a new optical

encoder element directly on the wavelength screw (rather than in a

more remote mechanical locaLion, as is presently the case). The en-

coding of accurate wav_2.ength data is consideret! important because use

of an inaccurate (non-linear, etc.) wavelength axis in fitting experi-

mental Raman vibrational Q-branch profiles to theoretical shapes

results in distortion which, in turn, lead to inaccurate ccnputer fits

and therefore inaccurate temperatures.

(5) An interference filter-test cell apparatus has been constructed

which is capable of operation witn a porous plug burnier. A photograph

of the apparatus appears as Fig. 3, while a schematic is showi in Fig. 4.

6

a o-

0 4

41~ .a ca

3 W

Ir 0 O43

ac 0~ 0

ISM ! 1 0-

cma

m 0

-4 0

C- 0P4 sr ;441-4- 12OfI

e- MI

IRI

1 WATT0 ARGON LASER4880 A

INTERFERENCE

DISCRIMNTR ECORDERIBLOCKING FILTER

MOICG~RIs JL

Fig. 4 Schematic of interference filter-test cell apprtu a esigne4 forphoton-counting operation.

9J

Preliminary operation of this test cell with several hydrogen-oxygen

flames has identified areas of needed additional development work.

Particular attenUon is now being given to cooling of the interference

filters, since the filter passbands shift with the temperature, and

do not return to their initial positions when the original temperature

4s restored.

II. Experimental Conditio, s

All experiments discussed here utilized the double monochromator

and horizontally-burning hydrogen-air flames on a 2.5 cm-diameter water

cooled porous plug burner which was placed 2.0 cm away from another

water-cooled porous plug connected to a vacuum line. A continuous

wave argon ion laser source of about 1 watt at the flame position at

48n nm was used for all the results reported here. The temperatures

[ of the flames 4ere measured as a ftction of position by means of

the fine thermocouples described in the previous section.

Two different flames were studied. First, F or nitrogen data,

a steady si:ichiometric hydrogen-air flime was utilized (37.5 cc/sec.

H2 and 88.8 cc/sec. air) for which 65% of the product gases was2I

nitrogen. 7he flame was found to be about 15750 K at its center point

(i.e., halfway between the burner head and the vacuum plug), including

a rough 50°K correction for thermocouple radiative losses. Since the

image of the monochromator entrance slits at the flame scattering

position was about 5 mm high, an estimate was made of the temperature

variation along this zone. This was found to be about 16'K. The re-

producibility of the thermocouple data was about ±

10

Second, for hydrogen data, a less steady fuel-rich (four times

stoichiomet,'ic) hydrogen-air flame was used (79.3 cc/sec. H and2

47.0 cc/sec. air) for which about 511 of the product gases was

hyd :ogen. This flame was found to be about 13900 K at its center

point, including a rough 40°K correction for radiaLive losses.

Here, the variation of temperature with posItion along the slit

image was much more severe, being rought~v 1100 K over a 5 mm

vertical zon,.. Furthermore, the reproducibility was significantly

poorer, b. 'i:g roughly ± 30 K. This flam.., colored red from the3

emission of water vapor vibratien-rotation bands, was subject to

significantly more diffusion by the ambient atmosphere than the pre-

ceeding flame, which undcubtedly contributed to its less reproducible

characteristics. It had, however, the virtue of a high hydrogen content.

Ill. Theoretical Predictions of Band Profiles

For the diatomic molecules consider-ed in these experiments,

Eq. (3) of Ref. 1 (see Appendix 1) can be used for the calculation

of the Stokes Q-branch fundamental series (',+ *v) profiles. This

intensity relation neglects the small depolarized contribution for

our cases. The profiles calculated in Ref. I in this fashion were

used to fit experimental profiles in order to determine the scatter-

ing-gas temperature. Therefore, als calculated profiles were nor-

malized to the experimental curvp eak. Such normalized profiles

arc also used here in computer fits for temperature determination

(see Fig. 13), but for other purposes, it is desired to calculate

prcfilee for a given molecule at various temperatures while main-

taining the spectral intensity differences at these temperatures

(i.e., not normalizing each profile at the peak intensity). This

is the case, for example, if it is desirr.d to calculate the relative

intensities at different temperatures obtained through use of a

monochromator or filter desioned to iso -e a spectral portion of

the 2-branch. We point out here that Eq. '3) of Ref. 1 can be used

for thes.e calculations if the vibrational partition function Qib

given in Ref. I folowing Eq. (3),

Ql-x - a - 1 ,

is multiplied by the factor exp [-(hc/kTG(O,O)J, where G(O,O) is

the zero-point energy. Here:

G(O.) = 1/2)c - (1/4)w x + (1/8)weee e e

where c , ue x , and wy are vibrational constants defined by Eq. (1)e ee ee

of Ref. 1. The exponential factor compensates for the fact that the

term jalue C(vJ) defined by Fq. (1) of Ref. 1 relates to the energy

above zero rather than above the zer--point, ensr:y.

F:',thermore, substitution of w. for w in Eq. (1) for Q , wheree I

i0 w~ -w x + (3/4) yo e ee ey

results in a slightly more accurate calculation of 2 vib" Here, wo

is the coefficient of the linear v term in the term value expression

2G (v) = w v - constant x(v ) +io oFinally, if cozparisons of Q-branch intensities are to be made

between different molecules, account must be taken of the absolute

12

value of the nuclear spin statistical weight gI" This may be ac-r,-

plished by subs titution of gI for the relative factor n in Eq. (3)

of Ref. 1, and, for Z states of homoniclear molecales, multiplication

of the value of Qrot % kT. 2hcBe given in Ref. I following Eq. (3),

2by the factor (2141) 2 . Here, I is the nuclear spin quantum number.

In analoqy with the previous conment ,n Qv.b, substitution of

B for Be in the relation for Qrot results in a slightly more accurate

calculation of Qrot" Here,

B0 Be - e /2

where ae is the coefficient of the vibration-rotation interaction term

in the term value G(v,J) given ".n rq. (1) of Ref. 1.

In Fig. 5(A) are shown the calculated nitrogen rotational lines

of the vibrational Q-branch at 1000 K, different .symbols denoting

the various fundamental bands. The nitrogen zpectrum has alternately

':strong" and "weak" lines Lecause of nuclear spin degeneracy, but only

the "strcng" lines are shown hora for clarity. In Fig. S(B) is shawn

the spectrometer slit-convoluted intensity appropriate for the instru-

,ent used in our work, viz., a triangular si- t function of full width

at half maximum (F-iM) A 1 .63.R corresponding to 300uia entrance and

exit :;.lits on a Spex 1400-1 I double spectrometer. The same type of

calculations for 2000oK and 3000o.0 are shown in Figs. 6 and 7, re-

spectively.

Use of the type of data shown in Figs. 5-7 permits estimation of

appropriate spectral regions for measuring the various rotational and

vibrational excitation temperatures possible. In general, vibrational

temDeratures are proportional to the integral of intensity for partic-

ular bands (i.e., the ground st-., band ox any specific u.per state

13

T x 1000Klox 4880A

0.4.1

eri

•0 IT

0-

: O'(B) A 1.63A

Of O

i .i 0

542 5480 6 5504

WAVELENGTH()

Fig. 5 Calculated Stokes Q-branch fundamental intensLty .'t 1000OK fornitxogen. (A) Alternate "strong" line intensitie.-. The squaredata points correspond to the ground state band an, the circularpoints to the first upper tate band. (B) Trianzgi-3r slit funcrion-convoluted profili, where A is the spectral slit w., I MF1M).

14

T 2000'K ?1.*4880~ 0 0

0.4- 0

0

0

0-at 0 # ~ 0

0000 0 0

a41- A 1.63A

542 58 54846 50

WAVELENGTH (A)

FIg. 6 Calculated Stokes Q-branch fundamental intensity at 200&K fornitrogen. (A) Alternate "strong" line intensities. The squaredata points correspond to the ground state band, the circularpoints to the first upper state band, the open triangular pointsto the second upper st- e band, etc. (B) 2Itiangular slit functionconvoluted profile, where A is the spectral slit width (-WHIM).

yi

(A) N2T- 3000"K

xoz 48801000.2k -0000o

0 3

0 :300 0 0

006 0 00 i

am 0 0___'

0

•

,- A1163 A0.2L

0 S I I

5472 5480 5488 5496 5504

WAVELENGTH (A)

Fig. 7 Calculated Stokes Q-branch fundamental intensit at 300) fornit-rogen. (A) Alternate "strong" line intensiti =. Thea squaredata points correspond to the ground state band, the cimi-lar

points to the-first upper state band, the open triangular points tothe second upper state band, etc. (B) Triangular slit functionconvoluted profile, where A is the sL-.tral slit width (FIiM)

16

band), while rotational temperatures are proportional to the profile

on the short wavelengthj side of each band via the influence of rhe

vibration-rotation interaction. Thus, in principle, it is po.;ible to

determine vibrational e~citation temperdtures foi any pair of vibratiornal

levels, and rotational excitation temperatures associated with any vibra-

tional level.

For general comparative puroses, the nitrogen profile has been

calculated from 300 0K to 3500°K in Fig. 8 for A = 1.5. Here, the

rc latively Iroad spectral width at elevated temperatures becomes quite

apparent.

The profiles for hydrogen are very different from those for nitrogen,

since the individual vibration-rotation lines of the Q-branch for light

molecules are spread far apart because of their very large vibration-

rotation interaction constant, ae. (For hydrogen, a. is over -0 times

la-rger than the value for nitrogen.) In Fig. 9 are showm calculated

shapes for hydrogen from 300°K to 1500eK, while in Fig. 10 are shown

the profiles for 1900OK to 35000 K. We note that the first vibration-

rotation line of the first upper state band in hydrogen does not appear

for a longer spectral interval [starting from the (0,0) position, where

the pakenthetic notation corresponds to the lower level quantum numbers

(v,J) ] than that corresponding to the entire wavelength scale of all

the nitrogen data plotted in Fig. 8.

IV. Experimental Results for Nitrogen

The profile of nitrogen observed from the stoichiometric flame at

a thexrmk.otxple-measured tem:rerature of 15750 K (15256 indicated temperature,

plu.e an estimated 500 K correction for radiative losses) is shcwn in Fig. 11

"7

.O&AA 300K0.6t A=S

0.4-

02 r

LLL0k±,11.....

0 .. .. . . "-...'

Or .. .....

0 4 540 576 50 9r 02rK 11o3K P

S0.

02- 27IOL

310

from 300K to3506"k

i i-I.O

O : r

0.41-

SL ,

_ 700AlLj i

Ii l~ ' *' , , A

2t02

~Fig. 9 Calculated Stokes Q-Jranch fundamentdl intensitides for hydrogenfroa 300 0K to I500°K. TIhe vertical lines along the wavelengthaxis indicate the positions of the various (v,J) -v;b'ation-rotation

' : lines, wihere tne parenthetic notation corresponds to the lower:-level quantim rnz~s.

01 A

04F AI.63 !900K

021-

_ ii

270(r

o__ __c___-.- ._. , ,6 ^,i iL - A

t' . 1 1 iA -

OD21 23 0K L1 A

-. 0WELENGTH ()

Fiv-- 10 C-11culated Stckes Q-branch fundamental intensities for hydrogenfrom 1900°K to 3500K. -e ertical lines along the wavlength

axis indicate the positions of the various (v,J} vibration-rcta-tioa lines, where the Parenthetic notation corresponds to thelower level quantum numbers.<

20A

I

2H -02+3.73 N, FLAMEEC 100- r157".

E 801 An1.63AB-F

~so

DARK CURRENT

S"5490 5495 5500 5505 5510

z WAVELENGTH (A)

Fiq. 11 Analog x-y recorder trace of Stokes vibrational Q-branch for nitrogen.

21

as recorded by an analog x-y recorder. The same data as recorded by

paper tape data logging appears in Fig. 12. Here, the number of

counth was recorded every ten seconds (i.e., every 0.02 rm for the

spectrometer scan speed used).

In order to theoretically fit the exper'm'-ntai profile and the e-

by determine the nitrogen temperature, the experi s-ntal data wave-Ilength axis was first made coincident with the proper theoretical

wavelength axis by manually overlaying the ejcperimental datz on a

normalized (i.e., at the ground state band peak) set of theoretical

profiles. (See.Fig. 13) These profiles all have very similar long-

wavelength ed _s, determined over dhis temperature range alimost entirely

by the monochromator slit function shape. This long-waveiengt:1 edge

wa therefore useful in determining the proper absolute wavelength

values fc the experirontal data. in future work, it is contemplated

to perfar= this axis adjustment ,required slight backlash in the

monochro-mator scanning mechanism) automatically thr-u use of a

computer-fit of the long-wavelength edge utilizing an initiaily

assu:med zpproximate temperature.

The next step in the determination of teqperature involved a

calculation based upn-n the ratio of intensities recorded by the mono-

chromator in the vicinity of the peaks of the first upper state band

and the ground state band. (See Fig. 14) Each of t:ese bandPaSses

was 0.18 nn wide, and contained ten data points. The theoretical

ratio slown in Fig. 14 was stored as a data file in the ccc=uter, 3nd

the computer-determined peak ratio for the ex._zrimental data could

then be compared te this data file, resulti.w in a determination of

teiperature. The theoreticai ratio of peak bandrass intensities is

22

- - - _ _ -- , -

SO 840

II

0

I0'41

ol_JO

o A.lISN3J.NI 3AI.LV!3

L 23

1 0

LOo 0

00011 .1 4401 la

0 .

,

I-1j 4

2c431

4

00 Q

1V 0

C4 -l -C -- 4 4 0~.

AJSON3LNI 3AI.lYI38 0 Vn'-4

24

...... ..

0

I~ ~ 00

0~ a

.it4

~0

-0-

I L 0 4

I 34

31C 0 5.

A IVdIX)V~ IJVwAJN3N

lie IL U5

also shown in Fig. 15 over a smaller temperature range applicable to

porous plug burner experiments. This ratio is almost exactly linear

over this range of temperature. For the flame studi.ed here, the

temperature deterained by this procedure was 1 54 6uK.

The final step in temperature measurement involved a least-square

computer curve-fitting treatment of the experimental data. The initial

assumed temperature for this procedure was that determined from the

band peak ratio methoI just described. The minimum least-square devia-

tion was then searched for by the computer as a function of tc.rperature,

and the temperature corresponding to this minimum deviation determined

thereby to the nearest 10K.

This methed of temperature determination has a basic sh'ortcoming

in that it is based upon use of raw monochromator data, to which the

peak of each trial thecretical profile must be normalized. Thus, any

noise "spike" of other random inaccuracy in the grOnvid state band peak

intensity can cause a substantial distortion -f the curve fitting pro-

cedure by producing a 'false rnormalization, with subsequent vertical

stretching or squeezing of the profile. This proLleM car. be circtmvented

by the averaging of zjacent data points, which car be performed by

the computer to produce a new 'szioothed" experimental data file. A

program has been written tc acco.plish this smoothing by averaging

over any odd nl-,r of data points in an equall-j-weighted fashion.

Thus, for a three-point data average at vavelengt-h A, with 6 equal

to the spectral interval between data points, the new intemsity at

A corresponds to (1/31 times the originally-encoded intensities at

k-6, A, and A+6.

26

r4 E4

4A 4

o 0 a 104 54

101> E

ZE~ -.421cA

4 X C

0414 a9~>

0 l

L -. 4 A-4

-4 1iLa410

217

In turn, this method of data smoothing has a clear shortctning

in that it also distorts the overall Profile in the vicinity ot sharp,

nonlinear charges of intensity. 7.1hus, a compromise approach as dictated,

in which data smoothing is accomplished over an optimized spectral

interval. This has been done for the data shown in Fig. 12, for which

three--point, five-point, and seven-point data averages were taken. in

Fig. 16 is shown the three-point averaged diata, while the five-paint

averaged data is shown in Fig. 17 along with a theoretically-calculated

profile comp-ted at T= 15469K (the temperature determined by the peak

ratio method described previously) and kiormalized to the peak of the

data-averaged experimental curve.

In the table contained in Fig. 18 is shok.n the results of the

least square computer fi-tting procedure for the raw mo.nochrom-ator

data, and the three-point. five-point, and seven-point averaged data.

-te temperature correspondin- to the minimum deviation [i.e., T(min),

corresponding to the mini--= value of - (deviations) for each treat-

ment of the data increases here as the amount of data averaging in-

creases. It is easily seen that as the data averaging is increased

excessively, the spectral profile is "flattened out", resulting in an

appearance closer to that corresponding to higher te=peratures.

As a working criterion for determining the optimum amount of data

averaging, the procedure chosen txilized the s-allest vain-u value

of (deviations)". As may be seen in the table contained in Fig. 18,

the smalest value occu-rred for the five-point data average -nd,

accordingly,. this was chosen as the a~tmroriate treatment for the

data. The graph showra in Fia. 18 illustrates the variation of

1(deviations) with temperature for the fi-epoint data average, and

indicates resultant best fit at a value of T(min) = 153&K.

28

--

0

goo34

LLI"

93 0.

I*

44 -.4

CL4.7

*00

!i

AI.ISN31NI 3AlYI38J

29

14C

G 0-

- of-C3

% 0

-at

it 4

9V.-

i6. I44. Ir >

0 (01 34 r,

ALISNINI 3

343

__2 2~~-~-~-JE___ _MVATOS ITU-EVM -ON

E 1DEIAIOST !(EITO )T(MIN)*1153M* FOR5-POINT DATA AVERAGE

10I

4 2

DATA8 1505%at 3P0 DA/11

'a' j5-PIIT A io' .211_AVERAGE t

SIAVERAGE j. 44310 150 54 15

TEMPERAT'JRE (*K)

Fig - 18 -,able. Siaryo Tce rAs ?!r-ii.) correspondizm_ i, th=3n"== -alue Of l (deviat..os) for the least sqzre profie-fitting proceulure. ce.-d for the zraw momoc!.c%zoi , data and4For three cures cat data &vegrq iq~re. -Varationi f1e-iatioMS) as & func--iom of temper-atue fior th.e five--point

dataa~er~eiprofile.

31

Surwiarizing our findings for the accurate determination of flame

temperature for nitrogen:

I. Sensitivity for measurement by averaged band peak ratic-s is

shlwoun in Figs. 14 aid 15.

2. Sensitivity for curve fittig is illistrated by the set of

nornalized curves sl.omn in Fig. 13.

3. I-e tevperature measu-ed by the band peal ratio was 1546°K.

4. ne tesperatuwe measured by the be-t (five-point) data-

averaged computer profile fit was 1534°K.

5. The temerature indicated 1y the fl-ne -wre tern.oa-ple wass

1525PK plus an approximate 50K radiative correctzw. for ea

estated fla-e te--,erature of 1575K

Cur%-enL agreement Between 6-a peak inte-ity ratio method aod

-the curve fitting metho4 is about %. It is intended to _-zue these

techniques to determine their Iwiting acizr- 4 es, vith an ----

an variations of the metis wiich produce -;ood accracy with a

of ,.oplexity in the data handLing. Thu. the trend is toward utilizing

the f ul Profile fit as a cali'atic of the s le b M ra10 Me td_.

This latter ethod can be ade more a---ate b%, uzil n more tha-n the

;;reset-1;m- ___tw band-passes.

7he full p€rfiie-fitti= wethod v-I retain its utili ty for investi-

gatices of no-thema eqjir_; signas~ures. it will tue Farticule-1

usefui when neither vibrational nor -otat-4oal ".

a drterniz-atcc o r2ztive pouations of the vario°s Vilrnio;ai

ando rotzcti a;al le-wel.-

32

It is Gtress, d here that Ram~an scattering ziatures are direct

maeasures of the relative populations of the mlu r itr nlice

aad, for equiilibritsz situati-ans, these relative populations ccxrrespandI to -.he fundamental deft-ipion of teeperature. Thtus, it is contevpla tedthat this forma of temerature diagnostics has the po)tential for beccingIthe nos-, fzndamentally accurate schmme for -n -perturbing, three-dinmension-a

v. nr~x.-tT Result3 for Mtdro en

The nitrogem data d scimsed in Section V were takers vith a v'iew

toward hicL--ac-_zacy temperature diagqstics.!. he bydrogen_ data' dis-

cussed in th~is soctio were taken in order to investigate the Eifferent

type of Raman vibcrational signaturze prodaced by a very l-ight wilecale,

arA vre not intemded for a.-cuxate tenjoertuxe r-asure=_eats.

7bhe pro!le of hydragen ebtained frm a fo"'-tiines-stoll ic~etric

h-~r~e-ai Ila at a tbermuple-zmeasured te~pex-_ture of about

l3~30 K(crrete ro~hy for ra itive I -ses)i shown in- Fig. 29

as cbserved t-.c-*gh mre ti an analog x-y re-trJn-r. Yhe first fobur

Votaticnal Iin-e- of the Stokes vibratioz-al 2-b-tw~ -- re ideatified-

%.-r parposz of =*ariso, see Fig- 9 fbe a theor.. -icall.1 a-sd

CurVe Of the entire Q-br.-exh at 150 E-) 1. :?Iq. 200 is shcar the

saze daza as cbtained fr=2 the pap-r taze data loer, wiere tbe trace-

zon as the solid --=-.e ontains the R~nan scattering iLat-. a.-4 the

~isnd .. rv s a 3set4rn m~ission ctriof this is

12 Mae.Te enisszco swctzv,-- has beem f~r~dfr=~ the ~s~n

ps-scaring ~~~ inf 7rig._ 2t. and a eertcycluae

casied rve zd&d for hvydS--roemn at 1 AWPE. 3Kaen-ixrq he .j

33

8 H (+3.73 N2 FLAME11mw /1390*o

-1 I62 I 0A

m04 J:3 2 1m 0 3 62 s

WAVELENGTH (A)

g-

404

Ito ~ 90 3

+364

-404

W DU. .

4

La "4 0

a-A -.44

-4 05

0 E-

co~ - a co

C0~

ci av &a 0 .2%xC 0 c.0

)L.UN3O BUM

t -.- 1 35

t0 0~~~fA0 * &4 - W 0 -

-4 p.

344

40 V 46A LZ QU Ic

0JmL W C

00

c AW 0

-436 4J

calculated peak intensity . )rmalized to the experimental peak intensity,

profiles were also calculated for other temperatures. The peak values

for J= O, 1, and 2 at 13000 K, 14000 K, and 1OO°K aze indicated by the

appropriate horizontal lines in this figure. The accuracy of temperature

measurements for the hydrogen data shown here is not good for two main

reasons: (1) the flame is somewhat unsteady and non-isothermal, and

(2) ratios of the vibration-rotation line intensities shown here are

not particularly sensitive to the temperature over this temperature range.

Other ratios utilizing higher rotational lines are more sensitive for

this range. However, the profile presented here is indicative of the

type of data and the required treatmen- for temperature estimates

utilizing light molecules. The relatively wide spectral intervals

bet.ean vibration-rotation lines for these molecules suggests that,

with prcer choice of bandpass, interference filters could eventually

be used for temperature determinations with greater ease than would

be the case for heavier molecules.

VI. Conclusion

The vibrational Ramtn signati-s for nitrogen and hydrogen

have been studied for hydrogen-air flames produced on a water-

cooled porous plug burner. Accurate determinations of temperature

have been pexformed ntilizing the nitrogun data from a band ratio

method and from a total profile-fitting procedure. These deter-

minations as well as various other theoretical predictions havc

made wide use of computer calculation techn-ques. The tem:peratures

found from the Raman methods agreed with each other to within - percent,

and agreed with an independently measured temperiture utilizing a fine

37

wire thermocouple (only rough .' corrected for radiative losses) to

within 2%.

The hydrogen signature has been fitted to theoretical predictions

for low-lying rotational lines, and exhibits a spread-out structure

which may be particularly useful for temperature diagnostics. Addi-

tional equipment has been assembled for improvements iii the spectrc-

scopic, combustion, temperatur.-measurement, and data acquisition and

reduction aspects of the experimental program. These will be used in

further study of laser Paman probes for combustion diagnostics.

Acknowlegements

The author is grateful to Dr. John Moore for tharmocouple fabrica-

tion, to Mr. Harry Horton for his contributions te thermocouple temperature

and gas flow measurements, and to Mr. Frank Haller for computer prograning.

Some of the prelim i:ary material reported here was presented by M. Lapp at

the Third international Conference on Raman Spectroscopy, 4 Reims, France,

1972.

References

1. M. Lapp, L.M. Go]dzmn, and C.M. Penney, Science 175, 1112 (1972).

2. W.E. Kaskan, Sixth Symposium (International) on Combustion (Reinhold,New York, 1957), pp. 134-143.

3. A.G. Gaydon, Spectroscopy of Flames (Chapman and Hall, Lendon, 197),

pp. 79, 90, and 241.

4. The conference proceedings will ba published as Advances in LaserRY.man Spectroscopy, Vol. 1 ia 1973 by Heyden and Son Ltd., London.

38

Appendix 1

Raman Scattering from Flames

M- Lapp, L.M. Goldma~n, and C.m. Penney

Sci~ence 175 1112 (1972)

39

- - ~ rdig

ONF

-- 4

- IA

Em .Io ~m imn oizotlyo oospu uc(1 nItm i- race)I ouv m a spsii eical trihte ai o aaumIvn xenns e a

112 S .Wica ~pkUnt eIasEvtkC. op tRcerhadEwopee.Shaclay i Yr)4

Ramam Scaffet h Flom in the scattering zorm was about 100 :2p-m. and the hrisht frcm which thescattered radiation was accepted (a,% _z,

Abstract. La Ranun smtering do& for nkroger-, oxyen, and wa, vaptw termined by the 1-cm slit heightA", -hrhave .-en obtained from h>ydr-,W.n-r and W-d ogrn-oxygen jkl z. The endtbl image sagncation f ctur ofgrour4-tatcand upper- state ibrationad bands exhibit rrong asynmetnea 6road- about 5 mu. The roochromA iening Experimental spectral profils hate been fitted theoretically to gvie a new tance and exit slits Wer- .' to 300-measureintuit technique for the determination of rotatonal and vibraionad ex- Im. for *'h.kh the spwc'r't sli- wii.ehciraton temperdtures. was measured to be in var .io-e agrc--

ment with the vaue -ac, ted fromWe report here obhsevatiohNs of vi- with studies of species in oven. at the instrument dispersion curve. T-4

brational R.man scaturing Mm flame temperatures up to 1000'C (3), and (Rayleigh and Mie szatte:ing) xbategasc . One motivation for these obser- with a low-pressure electric discharge ut the laser beam at the evlrance slitsvations is that Raman scattering can (4). We have bwa unable to find any (as viewed by a periscope attachmentprovide spatially resolved neasurements earlier publications cncernios Raman behind the slits) &,hoied no changeof the concentration and the vibration- szattcrig in ftams or in avy syictas when :he flame was ignfied.al and roticnal excitation ei;era- at temperatures ij, excess of 10001C. The flames !tatcd *ere pro. ,Zcedtures of flame constituents. This capt- Our initial observations were cm- on a watercooled porcu plug burnerbility should prove to be of substantial fined to Stokes bands arising from (diameter. 2.5 cra) (5) operated hori-use in the dizgnostics of noneqilio- 4S80-A incident radiation from an ar- zontally nnd burned into anz te, water-rium as ve1l as equilibrium phenomena. gon ion lase (Cohebat Ridiation cooL-d poro s plug (of lrger diamet:er)

The work presented her is focased model 5281 operated for most da:a at placed about 1.5 cm away which Was.upon the ob rvation of tenperature- 1.5 watts. The scatzered ligh was ana- in turn. connected to a rough 'acuumdependent ecmts in the spectral distui- y1red by a dotble monochromator (Spcx line. In this fashion, a stable hoizontalbution of the Stokes Q-branch vibra- 1400-11) aith 5000-A blazd gratings, flame at atmospheric pressure wastional scattering. These effect arise The detector was a cooled pholtmulti- produced which possessed thz advan-predominantly from the vibratio-o- plier (RCA C3100E Quantacon) op- tare of offering a scattering test zoneration interaction and from significant crated in tN pulse-counting merde with of unif,?rm conditions (that is, at apopulations of excited vibrzzional ev- dark ctemt levels of about 1$ counts constant distance from the flat flameel. From thes excited levels origin:.te per second for this wo&r. front) for a laser beam passing in thetapper-state bands (1) which are us- The cverall =xperimental arrange- vertical direction. 3catlering data foruaily shiftei Foward the blue region ment was designed to have the laser HO arid O, %-re obtained from leanbit, the spectrum, beam tgaveling along :he direction of H.-0 thmes, wheTreas data for N.

In Fig. I we show the types :f fun- the entrance slits (that is, vertically) was obtained from a igan H2 -air flame.dameal vr,-ratinnal Raman scattering and focused at a position about 0.3 m Because of the low luminosity of theseetts that may bo ob--ed in tam from the entrncee slit. The .Raman- flames in the spectra. regions of in-Eanter Ram=n se.ttwint exper*iments scatteredl r',4iaiion was collected by a terest., no increase in background wasat elevated tempgaures have "et multickment Ian with a focal length observed Alwahe the flarres were ignited.with la heatinj 4f a vapor (2). of 75 rm- The width of the laser beam Precise 11,; data were not taken, nor

C..pyriril- lot 1, r Awiir-nss itsaHbr the? :tdmriirrnt o.f UZPiepc

42

were accurate ui dpenden temperature aone involve a much smaller spectral corrop6idg to the Qbranch do notmeasurements made. since the major rane than with the corripoodisS overlap each other exaly. There is agO lof this portion of our fame Raman matreawes of Stokes/saoi&Sokes prressive sft to showr wav lengthsscaterag investigation was the e. ratios. Thus, in the former cae it is 0caused by the vibration-rct tioo 5nter-ploration-of general tmpeatfutb4e i- easier to correct for the spectral vwa- action Senn of the energy levels for a64ve features of scattered bands. The tioa of background, absoqrd. &M molecule. To discuss the spertemperatures actuay &eermubd from spectromeer respomep and poition of each -band we-considerthe scattoAing data ae reonable values 'he asymmetry of the vibrational the term value G(vJ) fora al d

fort namts usid. bands s evident even at room temptra- tomic molecule (7). includia bi-fAlthough vibrational- temperatures ture for the molecules coosiderd I , tions from harmonic a anharmnic

have been deter:,ined from the Stags/ and particularly so for H2O. In Fig. osclator legns, rigid and anharmoncanti-Slkes rtios of vibrational as2t- 21 the MO vibrational groundtate roierx-etms, and vibratioa-rotatki in-Win ten(2, 45), the same informatia is b.,d is shown for scattering from room teraciio terms (8).

accessible from the- Stakes scattering temperatue ambient HO. The greatlyalone. Our initial attention to delet- increased broadening toward the blue G(J) E .fJ)Ikc=--.fv+%)-ter is due -in part to the geater semi- region of the spectrum under flame .r.(r + +)' + 3..(v + %r +tivity oi our spectrometer and detector cooritiocs is sholwi in Fig. 2A. ThIU .1)( + 1) - _to the Stokes scattering; However. strong asymmetry of the N2j and Os D + p.12^JJ1 )"-there is also a potentl advantage to vi'rAtional bands under flame CoD&-tis approach -which uisis from the tions can be sem in Figs. 3 and 4.-Thisfact that temperature m _easremmnts "bue asymzfit" is explained by the Here E(vj) is the eon of the levelfrom Stokes (or amU-Stokes) sc&st fact tat all te =O rotational lines (vJ); ris -Flauck's constant cis the

Stokes (av1+I)Vibr. trmsgons

~C A

- i( (0 - l ll l ., - " a. . .. . .. . . .

L (a) (b)

Wavelength in I-A intervals -

I11/ *HP0 (.3t5rn' shift

a5, lit.|

S_.__.___ _ _____ -A intervals -

1!Fl I d:f). Schemati of tom molecular transi.ions whch cotriue to fund&, ntai vibr tonRm sctei (.W±=__ 1).i "l Stolkes ,;pper-rate hand (c rded numeral "one," at right) is associate with,;he moleculr viraioa tranidon v _.- to r. 2.i - For this tranStont tL scattered photon en~erg as slighly lpeter than that for the ground s ~ (ibM is, r = ( to r= ),r,;' O-branch , irclcd tetter "G) laeo -l aarot . th pe-tebadppg t• lgtl

shorter mvle tl, i 2 • flt.a,)Tme., taa .~oa fralanH- Thepma?egtSnent above cue-f t trace is a peOoo of another measome: shown to indicate t remrrblity olr "he p~rm L'atures at

=postions (a) stad (b). T c ;atte is as uppr-sute bead; d is discued at the end of tids tepo@L Alhog feamture (-) approuimnately Conie with a wea Ar lin ,t 5925 A. euperinieta mea,'urcn fz: ed to show sufficient scattering of i aito

S to cause the obserred ()Tese , am nt to 2$K. tonr). The |curve is he , ltfucro. These cve corspond to 30-g entrance and extitsfo hde slit wdh.i.62 A t tis wavelength. The wavelength am of these curv-s are indicated in a reislive fasionu only, sic wavelengthlibraio

s sigtly to the right of the peak of (B) (that is. at slightly longer wv,- le:.m). As am easmnol of a eas fo- whc exact tat-culatior ar rot complicst,4 the peak of the 295"K s -tomvcled pr J for ?1 is at a waeleagth about 0.3 A less thun i

34

7

Ai01 A -

S i44 4 492 S49 15 o S SW S OWe $2L S26 6524 5268 5272 5276 5250 5284

les 5%063162 a* sv-cwsb*)

F~~Z B .0,1tA

U_ S50.81.

0 01a

5254 5260 S2" 5268 5272 5276 520 52s4

wavlength (A) -Wauelent

Fig. 3 (lWt). Ramnan scattering fmvm N. in a !ean H: air flame. (A) The experimntal recording ona abich hatg mensperini-posed the sptctromneter slit fwtmi for the 300-pin entrance and exit slits msed. The spectral sli u3*h .1 is 1.&3 A. The lak~ug G.1. and 2 corresponds. respcctivey. to the ground-taie Stokes vibratiastal Q-bwmmch. the upper-suwe Qbtanch (r = to r- = 2). andthe uppcr-statc Q-branch (r =2 to r =3). (3) 17e solid Umn is a trAcing of the -arizean curve ,(A). So fad-hutte conipiisonwith the intensity Calculated ZIdscm Wjiwingts f m Eq 3 cau'resb t, 10" . 16W J1 2n I *K (xL

scattering from 0. in a lean HL-Q_ flame. The geccml conunKut for Fit. 3 also applv heme. Howeer. in his case a 4itior-lVN-per-state band (labeled 3) is seen. Furthenaere. a tracing of a s~ubs eat nseasureme-s uvder ietial cnditins is ssoom asthe da4hed curve in 1B) These tu-o curves ti~c an estimate of the spriAd En the C6 dats. and are to be coanpeed wMt the theo-retically predictcd shapacs. -An ettirate of rourW1-7WK is obtained frole these dima

speed of light. @, . ,r.. and co~j. ame whcre k is Botmann's coristant and for which iy=0 (that k. the linesvibiationa! constants such that m y.. -C wr is the wxre number of the Rama corresponding so even values of I areas~(<a.,. B, and D. ame. rcsectively. fundamtental line (11): missing).the rotstional constants for- rigid and -. ;~ .1-r)It is . vident from Eq. 3 that thenimrigid rotation in the equilbrium r+ I )shapet of cachs iarcula-banWW (that is,irt*erUCkaT position; and r, and 8Pe for which is *-he wave number of the a givcn r) of the furkdatenal series(a, <8., and fl, 4 Dj. -epreseWt Ni- i:Ocaing (laser) -Ddiatimo Here the wuill be dependent upmo the rotationalbration-rotation interctions. depol.rized cont-ibution %a been neg- temerature and that a proper fi: #.o

The Ramnan shift for a fundamental lee. mvd the -Wars associzted W~~ an~ etperUimenMA prowil car then servevibrational band (that is for W =O the cross section which are not ex- to delenflif this USe rau'zur All ofand 1v = 1) is. from Eq. 1. plicitly written out 2re denoced by r, the brands of the fandamental sees

.IG(r + I J.-J) = &.- 2&-.r.(r -4 1) + The rotitial (12) and vibraizkal will haVe somewhat similar shapes.

ryf3r2 +4 6v + 13/4) - u-Jof + 1) - Partition fUnction' QI- and Q,* Xre Here-, for the purposes of 1110 J100.~LP+ ~ ~ ) c consider fth grond--atc band. rs.e

Q-, - k12hcN.shape of this band (that is. theS

,Aheie the vibraticm-rotation inc. .- kT/2hc. - pofie) may be calculatedtion. corresponding to the last two Q.,. -It - e'P ( lCMAI1 horn Eq. 3 wh ch. for a fixed temperaterrns of Eq. 2. leads to the blue asTt and the factor Y, accounts for the erea tu &ur.-bcomesgictry for the bamds Here. P1. ma)Y be of nuclear vinf. For N. (1=1M the S(CJ) cr (21- wXneglected. since P < , -i -

The radiant flux S for a rotational llat'to she nriona levels coraon sud[sc#J(;+I)

th a..-t-. . - ..- - o[s zp-. .kU Q+• .+7

fine of the fundaental series (r+ I-r) to e vn valus of , fot- which !; = I - withStokes Riaman-wcanered Q-branch is . The s tosd- m;F M e U7-ic levr CAr-given hy (9, 10) rcqxA ,to oddvalu s of .L frwiticd W " 1. - 40u, 2 It.? + )

fi1 of)f )W. V=% o .I ) thesymetric what nly hom tri .pn natIiwsa rn cc 4.. Q.-. levels cocrespond to cdd vaars of I. for sil .icanc, which contribute to the

L p !-- G(4an 1 which = 1. whereas the antaayometric relative hand shape have barn retaine.-. kT j ' k l.vels crrespod to ev e s of e o . The rtuition fou wsc u povv e only a3

44

t~P~~uC-CP~dWScale factor for ncAxtion( 3 v..m to denote the vibra GXW' tA~mV~r c u&L u.Eq4 3. and so are m of to~ o iWqw ubmfrl & 1% 0. S L Ow t fhr.tim.;%= ro Iak,pr~ fttl fundibna vibtationa smodet. 1I Owk fi is L4Oma~aWmO

The experiments reporsed bet cor- band comrepnieg to the transatwin A~dm '-w be. fti a xiromm- zvedresOndcd iuidt reaotuble crtaincy to (I. .O)mi iO~s the most likey to be -- L awn-i Q-c obn -s

Mutations for ft vr ohmed. sznc h ares froml thw popn- L W- HOLM. W. F. US:b~ 5L 1. SOOSS&Kt5Iona and rotafuni cne1ty kids. bion of 2ie lowest excized- ii&azia ak-- 5x'~ 3" M~ f ~ ~ ~ ~ Z ~ ~. __S.. $0- d a Mm. L A. cAim *md G. A.frum is " eected. th epr- a) b w~ato [hX betpom tg 5, 6 kw Secs taiteo 'W- mawse~turt factors appqvin;- in the ez~oaient upjie-stale bad and the roamd-staae ~ ~m .wie..Op.Sof Ehj3 can be d:o dasvaheesof band (for no roosii) is -x 2 When A 7e!91 L. VY OiIdmu. :I-W

of T4 &mciazed with aftCIJ. x.._ is Ok- coefficieng of the (r, + ) ,. CAOM. P#10 ;A. theMLTa energY levrKs and the v~ii a i+;A-) trun in te vwtij kml Ca S. Smt kc eamk W. IL M&. is MR&and rotazjomg-exciutiom ien natures pressaon For H-0. x._k -2o cmO IROO %CCWV~b V431 MT t34CiidcOrrcqpoodi~ to tgcsc detrees of free- amd 51 cmrespoods to a blue Shift Of ik G. F. W~ a S- I. e~ir. A-u. U.idorm can therefore be de ined.m In 7.05 A for licident 4880-A radkltm 7. 7b. Cco" gtuf amm.P .4 o mec aPeneral, the ftW6% peak 'vaai or. mae- This Abift closely comacdles with the Ammm mWit- au t am -a oft& aw

grAted inlemity v; o*ach successife band p.'euls of fcatm (b) in Fjg. 2A (since 4 Wm"O mwae& zg-r-t-y=czmdcatca the ubrafioeai tensperazawe. ; is kcaledi siijbil so the right of dle -S"'3-

aheea-ic shap (that iblasa peak at the lankan siatteru' ow n 1- .Irkg ~~ ~~ jLd -Stczv " I- $oe" agZ Lie:..metRY) of each band deimime the Fi. 28). Thms vie have esiikace of an Wa N-m" t'iffiar. NJ.-tota'io.ral empemaure. Thus, if nom- arleciAb popauat m oi the I UP V-= &Z m bk IL c e= Ceq'Jilibriusn is swspccled separateEft I'lesl for the flame smsdieid. More ins- ohm ..ft t*& 3to the shape -of ech band shoul Ui poctatly. "h resul decwuuvne the L A W&Wh. 3L Qu.o 2i)"a:lmadc. In this fahiom. 2di~reUt vibam potenialW aithy of Raz s~iet-g IHL A- SiiGAmiL fS. 0*00 **WYOWtonal 31W rataioal eciation kmr- OS UreftentSs for polrAtosde mlculeS I.j.3 Tsm serv4~ qw e. 34P"alWes can be associated laith ecxh In IMaS-rpr uv bave cevsrw 19 lu 1 "to

i~itrati mod. Ths r d Ibmwl un~ Blp~~e t5 5 ow- ILI mis eftkmm %%c P 4 3 im- .

companca ~ 01 th A-,,u~do mis. Oer". zbstc is aho substantiai imterest 4'Sb IL 2. 64C Sl - a vc-sttio tcJ ~ tx for th in the- ume of Ramans scattering as a 1= J i. &or sMNIAM

mesmnn of eteitatmo ft~a cotisttaecy probe in flames and othet WA6~ WOOtx.. Ilk teuma" mt fit11c tcbm &-"atutes, vbidi have bc applied -v *Vcvm%-~ One ifflhortaid po~ int i soma ei C too mzW-f Cli

fAvnes ;aih supoms (13); cmverpJy ?io; such mmsemts is j1lUsttrtd Mar. Uhib is 4We hd=e i boaA.# 0 1 lbk&.gU_Isneziiowd ightl serve to jeemic the foregoing mesuts. For systems m e

anhar'%owc esm for 4m'cs tint? have mhtllt2w3I ifhim2 and m tO..~ rwm.S - tor Vwwvnot )v: been sell sdied (14). taial C6tiws Occurs. tcsepatre. SA&*cWtWC~. 11ea&M S-us tiW0L C-01.

-For the cawt she.ni equilibium cv: #kicnent caku!~ions such as the Ones fl L- SOC-S 34 C 3. -4imf Jb PMA Eeisms reative vaiaes, of s(rJ) ca b see nave disussed. am evmatv (a 11141L20 r'nT.M C~rlcaluculd as a fumtion of waekr addition to ihc basic crn-N section) in S A

!tm eAL Sic Pr-Z~lx "S ZI..Ovr th nir pc C-lr Of ft Cr&r to Telate !be scatlesing joleasity MS ju Ciitnvt vdem i,=, di~catvrcd b~.- for s.aiar tempera- c-vr anye V=L'wd bandwidth to 3 mJ 4 wea -beae fromi fm0*

go St.A WAAW ite Sq. 3 ui de-Ve:urL-, andi tile rcsiisvxci -,0 eape- conmtatic-sn of a particulr comstin mwmAQ *WC -Wffzeomue SIX *&mT=ment;il curmes This has bendn u f 1Jj) tn- therais: sipifiant eyToi in co. t- Awebogf to gh ussfor N- aW Q0in Fis 3 and . wit coniratio arsa4scs may the made. par- ua i is Wt *C %L. 1be-~ Cuctm ftcv

- - -c-mad am ac'C am of a Oftlent. The:he rc ".t tb the .. daftcely & ti=ibihy in %icw of the fact that the Sa fua' was ton tob ~ xci1600eK and :he 0- data rncuily fi WWra Q-brac SeasciniOCWX r a~e ith tul k Eflii imF4l730O"K. Thesc values arc reasorai~e wibraioaajl cu-!tat-vft Nease Of 01C 2 A. a 4A mu=*and W lwP**ofor the flams t ud We be~kr tlrlfcorw ir uib tk pMe-slate band dw tmo-a bmi T*Wo~ U*

=~~ is ft M WW atttrig - a vw%"na 3b V kD e

-this Ram~an scattering nxho for'& th N "f* 31detzerminatior of tame :esnpcraturcr, as L M. Oow"3mv Fmv ezzmi.. K Nj ~ ftesw ~o

C. M- pvim OWc -. .SUof Olher cmperimcrd amd prwiza .een E7lttl~r COMWpY 49 at oa. - babe WC bzf u bd

waes 0.es mmSf -i t 44 Wehi~tSFor H.O. tie ana neag results, ame -F -"md N O rk 12M0 UOAm am~ysiwC- eew-

Cocsidraby4 ... c Complicated sinc -I4 hivt. Ibis 0-ftowe u-dd. OW.uvu

rhe c:alzlatios :nrolve =asv stmic top -u.ecqer 6veit (16). Hrcacs-me if we I.I mvttt i~~~ L i. .Hwis E&Ez Srms"az W.&

Cos'ider ony the puru-ly siraioa "Smu W~s ~~~* ami e Cm pe. # h.em eif.I' Ev.e 7-Zto the ew; ekvels I itrm. r4 p..Pmr,cj. :SU L 7# . Lconin~~~~ ~ rtit d- 0-tC V S4a ew woo~e um .. Jo k-T.vam

molecule (17), wet c=n inavestigale in a e '*D n- SM "N& 69 OL tf. U8.peb i 0 .5&w wquaitive tas1bou the klos: pm~bk th 'd~ w~rwf .- w to v wcewitA~c fO rd

ftiv 1, 1 .. . we Uaa mh.upersut ans -er w u te ew.&we1 dew u baw ftr IL %L t SMtWmbf r971 .r..4 P muai,1

45

-- 7

a. S. QP1T WNCKIII

Cetral latelliss P- Aguy USA Ann NAsearch C"WT 1 S. Navul ce.*ance tabratery CM*drMmaselwo. ff. Cr-, 20Me Owlt aeef AertYetf cs Vhi*8 late a-A?": CI#*AMIcat1.r. $bit !"o 27-1 Silver S;rAp. Nu'Y1 MIDS

- ~~Waftt 17901, CCIffoamia SOW16 Lb~rlastigtet for Dfense Analyses AM. W.* foam V. tralms

4W0 .Jirm-Uaj Drie N aval arame System cmooAatimyaft. Vi rginia- 2220 NASA, Langley MesarcA Center Depa"Nmt *f the MMwATM Dr. No"~ S. ilfmri Mamtao. Urglitia 23315 USAi tC.b . C. zomO

Sa. Staff 131p: Dr. Nowrt S. Levie ATl. M* M33wait Suep 213

3*fins4 lbcitatieft Center NMIa ftscgramte Sdoo1Cmron Station NASA Lewis fiesaarh ZWV Derjtat of Aeroautics. Cod 57Altauudria. Viraia 22314 21UM V-"k&pw 11W. kleww. Callfacgtt iI390

ClevZ1654, 001 44135 AIN: Dr. Allen E. FvhsEPA Itechical Cotter AI1M: 0. rd

fiseercb Tria" 1 e IPerk Pk Step m~-, Naval - tstva~te schel ioprtatometNart -Ca~v'Ia 2733 ~ere~ Iaifc~ea SZO

£113: Dr. 6: I*rge. P-fl MA Lewis naewrck Casttr &Ml.: Library (COD& 21Z4)1W~ypeic troewlifaf Section

Arnold 45r Force Statism mail So 6-1 U. S. Nova] ftstyrda~te Schol

£313V: AMI (MW) Cletu, ck;i 44135 AM U lirary. "~ 12;2--ATTN. Or. Louis A. fteuImtlt

kAnod Air F9rce SUaMPs - Office of metal NaseArc. Jkrase& OfffteissdSA Xorshall -Spac Il icK Ceattr 1030 East- Wras Street

Wk.-: k. L. Smith Jr.. Chief SI47At5sudeva, Calfferafa S1W-T-CeIlc Viva&e Moivetile. . ie 1541t A-TN: cr. na~blcft .. RArcu-S[csiseTs51Facility £113X: fir. eitk Chdlew

Mlx-?of Naval NAti-pchAir Forcc Casters Test bagb National Stltwct Foamtelo - Sk~ Trnciszo irwt OIflice

M)l3 C~mermgEfe.1petiks ir rei strtetPLZ~1ck Vr force 3mi Eagiae.e Wfust" S Framciscs. Cat "u~ 192Florida -IM$ kasai-tow. 0. Zi. 20M5ATN: MUM~ Tectaica! tibrar ATMl .cearse *ae off ice of Naval i*.*anh l3room office

53F- S. ClVark StreetA:r farce Mfice of Scientific 1116aeCh National science Fevdattaa Gkiao. Illimots towCs

iWO VItwo boIevae Efsimerin ipu4qrtcics MIXw: DimreaArvSgta. Virgizis 2220 togaiaev DIVt1.iooEI.T: or, 321eh F. Nast idshl."too. V. C. 2-tS0 Office af xaval 0*1ar-d

Pa'.: Dr. ff. 0jlu NW vat Aces OfficeAir Frt* At"o Pruovlsion iabesiia 207 W. 24t% Street

Mri~tt-patenw. M3. Ohio 4S433 Nettoal 5siamw* Fcatlt. am fort. 5eW TOM Well1

or- Kervya fta "agtediw C~ivr~sseft Office of Ub~al &tsi-irc Sr*~N OfficeVi g, 0. C. 2,)55O 435 Sonr Stret

Air Vz-e Pt'i Phop.sta* Laboratwry All%: 'r. I!.3dl lbstebach ftste. ftssci~tts 0C.Cwirt-pttfro-aa AFl.7Ckiz AV5?3 £113: Directorafla: Frazcis i. Os~ek hatioaal Te0a-scoIfp Servfte

Vrw7rtvaat isf t--rve Of icl of ft-j.1 steArchXie RFirce rooel, Pvtasica Labaratav. 524 1;rt mRoyo Kad StAWEr 3ar"%. tjc~i 473

ONertwmt cf Geftsic Urlogfieldg. Vlrqrlkta ;n, 5 of tn &vjoaardt Art. Ckltorota 93523 AUXh: Cklef. .iPzt sct'vei kAp)Xor-vi "~~to 2217

ATTN L16Pk el. &- bava! fir [jv1.e:(tawe 4ff Ic. of %&.,I 9ezearchU. S.-yAir ftillty *I:. ,arch and CaIra:1z Oftxctr (A2-5) F14nd tw~rics azsis~

Devlopezt Laborater7 Joiasville, Pecszvei 4& 5MEWQSU .0irectorate %AM- U£~Lbrar5 onart of ift NARY

rcit Eutis. Vigirto 0. C.A~i~: Frc~ls~oo Ofuiste ~Sival Asr Pr--clsiom Test Center (06 .13 . ro' ceroreusti. Trg~iIa (=au4u-a') retco. Ne .erser -.zt2lb

A'; .P, 1 NAt rtifta UA Pael -'c Lob. " Az ArtsI Lota Ctyo-

Deeft x Ageecy Saw*. Air srters (vwo AA!Ion . A

Ak1lE: cms~adifg officr ONn.? .M

01. Levy Nssil C. Maa4 AR31 WshiitcG%. 0. E. 23Miktstoc Ara, Alar.e 3% 'A5N Ti~sical. Div..tcaAT--'.: %AV&)2 *.AirSl Syite'.a (~

~. ~.N Oeartrwa of tott %svy Natal ;tzeartft Labratzry stefttotse fitc lafortwtlea -:x~ter ATN: Proiftsor Tetiteolcly Ask,% A"V: jI Iryr C&Je 2SC Z 0t

k !Atoes*'em.al, AMabiwt I"~ MR 330Us,tr'. A-systeft Qvwc Dwa Dovislon

USA~ E~w;-f.~ ofcua t*c!i* Wja ip~sary,185d C*1tI it CU

Eta 0 ivrctcr. Avz" 6 ,ta-salopi~sc" Wnhsts . E.- t. r7ie6 AlMt xibrary. Code Q2~4alti1cUf. I.of K. ATM; -!ecftsi .l I brtry sly-412e-

olf le of Aduac*4 esleartho & TeQtlgu AMP 6V4 w4 SHaP 1S~ss 9PMrWdI - Li::Nr. ~tbO C. Z66EtlZt 1o t$-IAy#.A:1it. WrM.tr El 2054 are6to* v

biS CO (0 uV.S. NMI Veep Csow ~ ~ WWSI mOki" Lo*.~ C:;Iftmua CotGAM~i M~tvtw Notg ~im. s1Orono Coden 6w_ Chim Law," Clfia 16*1:Loa 1N -agisuetUgg~ s~toAM-X -W. E~wd V. ti~C* mud-

Cisw, Divsion 0. S.IM Haew's Labeathay CMM". Ca(ULbe l Waila "3S 1gm~Irgiti 2B Svr. owmJrsa ff3AMlS T.0116S.-*~ eckical AM-my*13 '!cbtcui Af'tasLalCode (on

V%104.tke. Wifmtmoa 13 AIM AWs3 ist-a asrssATM., I~.cis LIWWaY

*11. 0. cca 80fAb~ C1.2 SIIIcificc C Aitht rorest a now~u

MAllt Dr. ftbo ks".afrsei Wwff. Woi~ 41

Q.n Sonca~ 9115Arn~~~e~sipact Cecatmbrwa--

Pvwshl .. ~iis&p. Ent g Z1131Est~h1~Stiit 11al.1 bas ML, SOW. JP.a UPint cATec Tuuucsocs Iplfmala 5111: fr Csqew 13ibL -hr. C-U_ P.~il 0. bou

W-. ~ ez ftrP.au 0.r imr w T e __j : b4!*

becatcl11i M fGff &I*. a Y.*q 1440is _ffttet". aw"17 SLUD C413 zcbtcal Dtbui a w~a Elct

ASTS:ti Dre- WW CeartElec ra.fWaiit qbb 4 t-3 .Bin46 7:£lezaaE*~~~~~P Oa igsa234b N ~L Ma J. rUTI: LlrarluiStell srwi:. Cgpip 15

Aeboatlc C)arm e rpea t~ A#air A hwj.c. gnwis c% Pa1eoffi .v Calif Tris.iss

£211M.Newe *raswA : or. Scme befr 'M a"11 7. 1 e ea

ATU: ~ ~ ~ ~ ~ ~ ~ ~ e Wert. Teas1 3616~r .0 ox25P 0 sAenveT09cu.3, file. ATTN ToCNSca 14221

Fesims S AM Tec tawlag M115: 6- r0 Ia~ E. 3ecritt. C*!ATW ~ et1E Lerwatig AM Te§Act 1094 CrmwibasTc.tew

Shile ragiayoie.,al tame sedev se 0?0ra AnarhCwtectmre as.g it a lwrsteDAf esC O i ta.t Viurii a~rte 22314VIIl m. 1 r(nea tts!~ a~a tVIX: Librarian -sslsue kass7MAMJ

£.~ ~yii~~ Cw~r~m elrM~d ~ I Otista ral ElectcSp a Ce Sern em-_tsun stret Fatmtudule. ~ Y1wt 1e35 Kwwp S., %ece TOM CawmA;1w Pri. imnisi U3. 0. ISMiez-

~~~~~~~. 0. an. ZvP.~ei1L ir~Dr.~ataO Ameer Fo110. 7610 -f adaui. t1*0"o ~ t "M~~T M:w1 Grd 1q. wyeia

~ S r iv~. 47vi .wkodP . o 3 P .o 9

0=0c14e Iacaqerefo fmat .1 b RsaCh wil KA borseatlesAltogbw SelisticsIarau 9ti f Ntcietcel ZqfentlesMgi msst SWUMlc br DivisiafIP. 0. lea 210 mtseal ft". Ottoe tarsn. ftw k'sog US?~CednmL Maiylczd 215K ~ (kearto, Caneftlk 0iA ATTN: Cateeerisg MhrurMat *S. t~ows& S. ha AIIN1: Dr. It. S. Wipt Bizo-

8sOrorte ftuu& Anti. 11441 Mills.F tadp. 90 ha fl Itfl Tim. California 94940 ektwd, Masbita P2Utah. IfM 4044t- Dr:2. Jack U. Xiacie AItI: liceu A, Croudle7C.: Liblrar* 1006-a

A/S tocjlery Vaaatari.totter coasy. it tmn. DiVISia4Ar X& s&*tdru3-S-I. *. s:,h Sqei& nAig-rsa .mcaGT ubo Svs cAYo. Uer-26 6633 Cvc~ge Anew301 Ae. Z. utay ATOCh: Or . g n lgark. California 91306

Lift 2. altyAIM: TehOcal lfoatiga CaterSesieor erofleasctst trtnga(crcrstc

111F buwt t raaeucs (cnaa.f 15)6 Awr4to Cwe. Bolevrt cepeOwsg

P0. t6o. 5,0; ATMh: fttcica) 1trAII 63 Cnea Ante* AM#1tF.jasn C. lttritct CA"e'a Parta elk)i:..A~w n~ iecrst 9 .is04

Swnfeeent EiM! - eaatg .tzct~eflGreeia COWeseX F. 0- &!3 2S

DeptE. 72-47. Zcet Zblc kO') ZCle. nyn r graticzlXntu ezt ?4 AI: ILtrj.t~ Mkz- eascaz!c~.

CIbeIA Tnta. f I Mfis aM!42Lactw4 S~lvesS aSpame CMep(agead w~ltc ts"s.~,piS3Mf Woover Strut:, 29 hArto de us Diz Lt.RwPAS Alto. Coemcl flW 12 (heeL 311o S= tqmmn0. Trace See' tatnsJOS: Palo Ales Library s2-S2 AIM: VS-. n. 3arrerep ." t

Lit Crtalifaunlatztua4 re- 1s (**,jimcoo ~ OV ATTN: sr.. :n KaniQ Div. *A1

Sc~otific *C4 ?ucsatcal Utory £Errejt &*a Nrcpwisti-P.- 0. Sft .QT 21Avoe Or -& Dinisisa trcbre a tcros

C.CItra9W33 2- 0atk csSgn.Fae r. o.* ic

Never.Ar toloranol Llb01 1t.4CI f1 . 11LOS~?lo~r CUMSpC~tz-attoQraor

Xass. lio Ce fl1&ttC AGS-- 1u Tecuitl ftqgtt Stte tasa oila CeU.M tendae)

ytse' E-ray ue-3u-

a--e. elsb trqtennr Ese-IM. Clf- 2131 :.1~b ^ 4wseTs,: Mn 3661 [nlMrtr.Centct(1tRct, tCal 2o':w

Eastft Wartfor4 Qnectlc- Te1 ar: >tInl CrtSnIM 516 A7ee: Pr.Lee. WVrIr

AV.s 0 r. Villi S C. resttrsa __ - " StvrkeSt xui. lsmv 3r att Rarter CPAircut Eas0 tt~t MrtfrS ali.ctlot 0SamATM:Arstrch -imefil ;.trar Al-r owr. Aul u ft.144016h AIIMv: Pfr.i.a.rem] ltI4

Prjetw.I=s- 63410m 4-M S. -ft Ct-vtf n1 fast~ iuore Cpacu ectI. LesteP.Th Z. 5rrt I.S Stmsng p..s.-

S*.4-M t~.t liii-. 611-96Balsa AflZM Mr. WC!.s Sn 2 htr;%Cftti A~iiraf st- teoeery

S-- i-41S. -XISSMhi 63r. ttcsr !o-C. cs'.- - 80 Cly, trn 7522Te.2ogPtlt dWlsyUIrf elzwqqPoas

UIMRITI[S MO INSTITUTES

ftlytecMtc In2titute of trools Cornell Aereatical Laboatory. Inc. Plasa zsts Institute of TedouagDepartat of Aerospace Wngatering 4455 Ge.1isee StreetI Dept. of Omcal Enmea

and Appllied PlechmiMcs Bufalo. New YorK 14M2 Cmridge. razat 23srocklya. Usi fork ATTN: Head Libraria ATTN: Dr. Jack 1. MowardAiM Dr. S&%al Lodems

Cornell University "assachusetis irwtte of Tec"TogtVVwA UIMisty 9wprtnt, of Otnistry "eirtmeect of Mswirtry. how 6-123

Division of Ealae. Ing iUhaca. X2 York 1485 co-fdge, Nassaximsetts 02139tax 0 ATTN: Prof. Sin Ni. Saver ATTN: Drz. Job lossFrowlddice. Anil Island 02912ATTN: Dr. A. A. DobbiLs TeciMical Wnivevsity of Denark JUT Libraries 12m 14 E-210

fluid Mecakl Departnt 17 Massachmstts Aver-meCaliforsn & tastitate of Techmologr Buio llbn 404 29W L ~ Caridge. masuchas 02139

Deprtet of 0hmincal [ng~meeria OK-Denrk ATTM: Techoical "tnrtsPasadebz California 91109 AnnM: Professor K. lufslwmdATTM: ?rof. V. ff. Corcsraa a;% tsWiueof*zW V

Frasklis lsst~tut-, ftsarch Laboratorls a*= 10 *t~California Icstitat'. Gf Toctoology Philadelphia. Pennsyvania 19103 Cab~de ?4%sacksetts 02139

jet P oul'le Laboratory' AIIM: Dr. G. P. iUacttli £111: Eagizeri'g 7veMical L-porti;4800 Oak Crowe birvePasadena. CA.0roa 9110) Fysisch Laboratorias KnOSSzet5s InsttuW of Ix.9alogATTN: I !irary fijksiamlve~'sitet. Utreckt Dept. Of VAVchaocal EM91ft*.;Al

Sneleaa Uo A 3-246Lbiversity of CaHiOM&i. Ska Diego Ut~tcht. TW_ xttferlbw Cmridge. Ma1ssacsetts CZ139

Dept. of AerospaeA ad 1kchanical AIM: Or-. F. vva, der Va1Vt A'111: Professor Jms fayEagiraeeting

La Jolla. C*Iltrn Ckorgia Institute #'f 70clolog Mssacbusetts Instit:.e of TdnlovMTNM: Professor FTo Libby Atlanta. Georqia M33 Dept. of ftowaical Lugieerial

ATTN: Price 'ilbert MOCIal Library Cmridge. FAssachusetts 02139Uiveiity of Califora. San Diego AIIM: Prof. abert E_ Stfcksey

Dept. Of [nsiaeries Pftl3'cs Georgia Institute of Technologyp. 0: So. 109 Scool of Deraspact Cogitffeibg kmvrsity of M1wg!La Jolla. Cuiiforma 923? Atlanta. Georgia 30M Dpartnt of Aerospae EaglariagAT: Professor S. S. kL*tr PALl: Dr- Se" i. Van Amn Arbor. Micbigaim 43105

Unvriyo aionaProfessor ATTN. Prof. T. C. Admom. Jr.

Sczool rf [ngteering sa u~nirsity of Illiot kidoest tsearcm InstituteAplied Science Caenaet of Zweiy Wag riuig 42n Wifer Soulevard

M53 boelter Hall1 Box 47M, sansas city: !iissow-f 64100Los ?tgeles. California 9OC44 Oticafo. lIMMIS f" AI,uu Dr. I_ A. flneATTN: Engioeerlaq "prts Growo ATI: Professor Paul M. 0"i~

RttglieJ des Vorirtac*k der Tried.Wt-isity af California mWversity of IllinoisgrwGW

Lawrence Pjoation labor-story College of Cogineer4ng 43 [ssen. Altradc-ferstai M3P. 0. Box EDDp. of EakMg [ntasering Ger"MyjLivermore. California 550 Chicago. 1II,.is ATTN: Professor Dr.-Ing.AIE: Tecinical Inforuuoion 5tt. Aiz. Dr. 00. S_ ;iscta Wilh1A 'wttvtng

liperial College New fork Jatfite olf TedhnlogUivtersity of California Lodon. Cftgland Wecatley 00;

Geftrii Libraty AIM: Professor caydon 01d kiestbury. new 1O47 11548Seiieey, California 9472 AM1 Dr. FatXVVM. ZrCx Ots De~rtrZat 4weria* Colle-pe of Sctmne--e A 'Iechuzology

Deartot 3f Mechanical Eiqlneerfaq u~Iwrt-ty of moftr* amCase Ut*stern Ars4v LaiversitS7 Exhibittof A314 College of Eagi~er4-ing

#090~0 Euclid Aveni* L">An. S.V.I. England Notre Da~m. ri 46556Cie"ela4, iCa 44106 MOR: Professor W. ftwgitrord As-on: S.csart T. rocCoas. Asst. timaATM Seers Library - fpcrt$ !*-t. for Reterch & smeial Projects

The 'ttn Ho"Ins IhiversityCase 6estern Fieserve Onltn-rsty Appli~d "s Laboratory (0*io State Lnirowsity

Dttislo of Flui ITrIl and 8(21 Ge-orgla Avtrnue Sept. of Che;cal Wngeerialkwos..eze scienes Wiver S-ngr. "1&04an 20910 14D West 19th Ae-enve

rClevelazd. chic 44106 A;;*: Oo.cical Prpo.sto- Colulfts. OhioAITRh Ivof,-s-or Ell Restto tnforwtflbo Agelcy ATIM: Dr. M,;erz S. broakey

coloraodo statt bIftWsity The 3011z4 hop~ias lh-Ilitisity 71W Prsylvania State tveirsityEoi-riel ResearO~ CeAte ftlsea Paysh4 Laboratory fto 23? 014 Put* 1181gFort Collins. Cogorado 8052162 ~ Z eogIV 4eto tbivtr"Ity Pai. Pe"AsFIVInA1 1IM0AT: Mr. 7- X. Ssjbiorr Silver spring. 1'aryliand 021~0 AIIM: Office of vice PuesIdelt

DoIM: imen~t Librarianfo ercTshe t~tvrsij zff Eawmecticlat

Drjiart-*a* of :*rt 5;&ce Engtcurial The 3Jins r4*kias Uavt, MIY P~YlVaIM1 State Unive-rsitysr. ;;e:;c~t 06,268 APlied ";-Sict Lab.ratory CJt;* 00 zoglreesing

ATTS. pr. P. Frir. Pkodttt UZI GWoSg;4 Avau wUiversity Part. Pennsylvapia 1680Associate 9yofinsr SIZVer SzOi -3 '~lal:4O 20910 ATTN: Dr. .4tis E. Lancaster

At%- Dt. -4. A- Voetberg Prof. of (ngineering E~Catfon

SzftaI of UngireefIM;geer Sc~I''ce Cuttersity of Leeds Imstitute Roietwi actc041ccvoer Sqwe Lef4s. 7 .910-4 UMldad Profesietal de zecateucese c'.Irk. Asw Terk AlfK: Vfs r ioq-teuis wexco 14. 0. V. ftsift

AI~ Dr. Villaz* thlatz .41mI: !ng. Kaftel Zorrilla*&sorat* Professor zf WE Virectcr &rftrall

PrcgIoA Uilversity 1m;versly of t* eiverslwy @f Virgimia-Oept. of Isa an . leparsmt of eicl Eglieering tP!. of Arospace Ealgota lotjms Forrestal Camus Rvhester. N York 14627 School of Er. ad Ap l ed SciePrincet-o % .ney 0840 ATTN: Dr. JM 1. Frron Chrlotsvi e. Virginia 22901ATTN: Dr. Plortin Swamer'eld RomAers: Dr. JoM E. Sctt

SPrinceton U ersity a rPrtcltn U!.llrliY ¥a grad 23 Univerity J f iriiaForrestal C&4,j-s Library 00133 ftet. Italy 'o n C5o.g, oSoiefcefTectiaoloej lifora:-. CnP. 0. Boa 710 ATTW: Prof. Gatao Salvatore Charlottesvllle, Virginsa 2290?Friactt. fs Jeise C6mo ATTI: dr1hArd 1. &At|nATTW: V. A. StmOskO. Librarian Sophia thiver wty

Sciexa and £nirneerlg Faculty Virginia Polytecdimc Institute andPurdue Untversity lol 7 To yo-htiywa Jazan 10Z State UniversitySctscl of micl Elgineeri g ATTN: Professor H. SuzukDl0t. Of Aerospace EagimeeriagLWfayette. bdiase MY MacklhfI Ygi ia 2406tATM: Prof. V. W. Goldschmidt Stanford 11serc Institute ATT: Dr. Geroge l lagerOpt. of Cherical Dysaics

Pe University lo "k, California I045 Vir2lnal folyt*Vic Institate andSchool Of Mlecanical Enginetrim; AITN: Dr. Nr) Vise State WIjitrsityala t tte. lllna 47907 Sta Cagiep.tnizai~l Eniieer.l Dept.ATII: Profc;sor S. L. K. Wittfq Stanford University Wltck;b"j Vfrgital 24061

€*t. of Aeronautics a4 Astronautics ATTN: Pr. Walter S. O3,tgen. Jr.Purdue Untwrtity Stanford. California 94305Shool of Atroue4ics, Astronautics, ATTN: Ew. Halter u. vincnti The eoro e WAshintatn Unive:sityam n21inleri sciences Mashtl , D. C. 2MLafayettk. Inilam 47907 Stanford University AM: Peports SectionATTm: Library Dept. of lie-hanical Englueria-Stamford. California 94305 Yale University

Muee Nary CoAleTe TIN: Professor A. L. Loon ieintlerlog and Applied ScienceDept. of lchaical Eritweri-v Kaon LaboratoryThtle Ed Fad Steve"s klserch ;nstitute 10 am . C o.actc t 065.10Lo d nEl , Enl ad Castle Pl ft, sta tion AT M"N Dr. Jon aI. F * M

AI- Professor P. s. Tsorirg H . N J 07030licellvtitiATTN: P'rof. adbert F. ft Alev, III To!* UniversityRie niesiy omatton La Erector FAUNm Labrstory

Welc Ptrofess or of Ch $t r7 40 T o wle St ret

ATn : r. jse* O L . ra kl in Lu.ot. Of h t o n ( h tstr y lA- : r f. Flee r P. w p m .eeF3culty of EngineeilogSunkyo-kvToryo. 48 113

ATTh- professor T. Hikits

50