>

/\) /9-cSIJ GK - / 7 ~ "772

A Reproduced Copy OF

f\J2C - ~4IS_3

~~~~HL;U1V ~~1~y

N.r~~1 1 /! 1(",/ .. :· ... : '0 Ii i 'c,: f:iC')

.L.ANGLcy F{ESc.A.1~SH CE~'~T:::t1

LiBRARY, NASA

kt/~M.e.T~QtJ-z. YJRGiN!l\

Reproduced for NASA

by the

NASA-CR-176972 19860019681

NASA Scientific and Technical Information Facility

111111111111111111111111111111111111111111111

FFNo 672 Aug 65 NF01696

https://ntrs.nasa.gov/search.jsp?R=19860019681 2020-03-20T14:50:10+00:00Z

~ , -

. -~

,.

'>"

-~~--

I

I !

I t i f t ,. > ~. t

~ • ~ ~ , f ( l' '·t

r: ;

.' t.:~·

'f , . ,"" t :~ ['. k'< ,'I fl,' "''1

,-;-------' -1

,1 ; J.;i:.:p/ /J /?: ~

" "

t~ ~P C}V ~'

;j()SSd-P>

FINAL REPORT

Covering Period March 1981 - September 1984

MEASUREMENT OF TEMPERATURE AND DENSITY FLUCTUATIONS

IN TURBULENCE USl9G AN ULTRAVIOLET LASER

:NASA-CR-176972) ~EASUREM~Nl CF lEftPEEATURE !N~ DENSllY FlOCTUA~ICNS IN lU~EUL1NCZ USING A~ UL1RAVIOlE~ LASER final Eefcrt, Udr.

N86-29153

i9dl - Sep. 1584 (San Ciego -StatE. UDi;'., Uncla~: Calif.) 30 r CSCi.. 20D G3/3 1+ 43306

Grant NAG 2-104, NASA-Ames Research Center

The NA~\ Tarhnlcal Officer for this gran- is

Dr. Robert L. McKenzie

~rinclpal Investigator:G. A. Hassey

\ ...

, ~ ,

Institution: San Diego State University San Diego, CA. 92132-0190 Department of ~lectrical and Computer Engineering -.-., ~. ~7:~ . ~~~~ \ .... ~

./.',~"\ \.' /':~--""'I; ~~ - ~ -<:~~?~ /-.\ '" "-''i. /':-,\ !J't..,c. '_~\

~{ r., ;;; ~ • 1.5'8., ;:.~;*

'.,:--- '( 11-- 1"\ ~ .... '. ,~ , J!. '1..;;-, '_~I

,y-. 17,.;.1r>,. .. ... , v, A~/) (.Oil '. >"'p $11 ., . !':'vL ~, ",~. -/ /;."", ~r j ", AQ"I'\' '4;,ZIj'v'!::>_->lf 'C "'v flt... <'" "".i . ,- ~. <;;"'" .,., '>''; t..- " 1, v/~ '\"''-i: ., . \:1 <:t; .:"-.-

i!<,-:'')/'(};;()vv/.' ~~,

':

AI (7J-~ a '9 !..5:3.:q. / V_ .. o ~ .. _ ,_ ~_H' __ "'-"-' __ ''''''"''''." ~ _. , •• ,.,

-' ,.

..

I i

I

<

I 1

i I

I I

I I L

~. ,

I. Introduction

This report summarizes research under grant NAG 2-104 from

spring 1981 through summer 1984 on the topic of noninvasive measur€-

ment of density and temperature fluctuations in turbulent air flow.

The a~proach that has been investigated uses fluorescence of oxygen

molecules which are selectively excited by a tunable vacuum

ultraviolet laser beam • The strength of the fluorescence signal and

its dependence on laser wavelength vary with the density and tempera-

ture of the air in the laser beam. Because fluorescence can be

o detected at 90 from the beam propagati0n direction, spatial

resolution in three dimensions, rather than path-integrated measurc-

ments, can be achieved. With spatial resolutions of the order of a

millimeter and at supersonic air velocities it is necGssnry to

perform each mea&urem~nt in a time of the order of A microsecond;

this is po::sible by using laser pulses of ten nanosecond duration.

In this method atmospheric O~ is excited by th~ emission of 3 4

tunable ArF excimer laser, and the fluorescence, which spans the

210 - 420 range, is detected by an ultraviolet phototube. The ArF

laser is tunable over the 193-194 om range, which includes several

rotational features of the Schumann-Runge absorption system of O~. , Figures 1 and Z show the spectroscopic details of interest here.

Since the ground st3te rotational level fractloDel populations are

affected by temperature, and~he total number of 02 molecules in the

observation volume depends on the density of the air, there is the

possibility of simultaneous dAtermlnatlon of both temperature and

density. Originally, we planned to use non-specific excit3tion with

..:~

.:

.. .,;'

l"

:>. Q) '-Q) c W

I 193nm

Absorptio~ I <,

3,_ 8 LU

8- X fluorescence 220-400nm ---

.~~~~J.~

./~- v 3 'C""'-, A,Lg

Nuclear Separation ,~

FIGURE 1. Energy levels and transitiOtiS of interest for rhe 02 molecule •

~ ~; \ . \ .' ...

f· . I-\ [

r~

F , , , ,.. ~ r t f t t ;" \. t' , t· f

~

~

~ {o.:-,.' ", ~ ~

f

t i;-

t ~

t

,'. ~ .. '.'

i (\ ~:""":

t"

f'., ~ r

~···i r/! ~c rl t .:

~~ , ... ~

~ en z ill l-z

7 6.:

fiI> •

• 0 lit ,

Ii «II

• • e • • o

l 5::·08

• &" G e Go It

I 0 Ii

4~- : It! 0: 01j1 12

. :0"; 11( 2· 41! s , ! H, ~6

~ i Ii '& ~ ~. ): :3 i 20 Ifj)c&~m to "I l~ 51 ~* ~ (; iJ Il' l . if> i !\l @ "e e til (';lJo Ql . ~ ~i· If'"

T_UHUl~'\\"" .. 1.1\. .... l" l I ! I I l ! I Iff I L-J

200 240 280 320 360 400 440

WAVELENGTH, NANOMETERS

HeUKE 2. Spectr~m of 02 fluorescence produced by Schurn.:lnn-Runge absorption near 193._, nm.

d: ---1

.-'~

,d

, <

r t· }

~' , ; ! I

t·

! < I i

! .

, t r-,

r' ~~

t 1

~ I K

..

~

~~

. -, ~

;.

~ "

{i !,>,;

r ",. ~

f~ r i

~ ,

~

~--

a wavelength-dispersive optical receiver to sense the relative

populations in the excited rotational states. However, further

investigation revealed, that predissociatton of the ex~!ted B(Y' 4)

level occurs within a few picoseconds, implying that those states do

not equilibrate. For the same reasons, the fluorescence is very

-5 weak, (the quantum yield is 4xlO ). and the losses associated with

the di~persive receiver optics would seriously reduce the signal to

noise ratio. The predissoclation results in formation of two 03

molecules for each absorbed photon at 193 nm, and 03 has a very

large, broad absorption near 260 nm. Therefore, some atten~ion was

given to the possibility of observing the 03 produced by a tuned ArF

laser excitation of the 02 rotational states. Howeve~, the maximum

14 -3 03 concentration produced is of the order of 10 cm • and this must

be detected in a 1 mm 3 volume (i.e., the optical path is 1 rom),

giving an absorption of only 0.01i.. A low pressure mercury riischarge

lamp can provide a reasonable S!N for detection near 254 nrn, but even

with parallel differential path detection we found that low level

refractive variations in the laboratory air made it difficult to

detect the presence of the 03 in a 1 mm path. It appeared unlikely

that the additional two orders of magnitude sensitivity needed for

quantitative turbulence measurements could be realized with that

approach. Similarly, fluorescence from 03 or other products of the

02 dissociation occur with very low yield and are more difficult to

monitor than 02 fluorescence. For the~e reasons we aLe led to

consider the system in which rotational states of the ground

x (v" 0) level are probed selectively with the ArF laser. with

02 fluorescence providing a signal proportional to the ground state

;; .. ," :~~ ".

. .

R

~

! ~ g t j'j

? f ? , . ... [ C r: i' ~i .. 'I f; ~< ~,

U . , H ;'1 , '! :i n ;-<.

q !~

U .;\ t~ ~.! £t

~ r;

II 1; ~

i; n ;; ~'l ,U ~t

" ~1 P f ~ :

f, t ~ .. ., I

;; , . ~:~

tf ;-':

".

L r: ; ~

" f""~

~ r ft II ~ t'R' .~""'"

populations before excitation.

Given that the above method is chosen, there re~ain some system

alternatives. In one configuration the same air volume is

illuminated by two different ArF wavelengths in rapid 3uccession. and

the time-resolved fluorescence ratio permits separation of tempera-

ture information from density variations • Alternatively, one can use

a single ArF wavelength and measure both fl~orescence (primarily

220-260 nm) and Rayleigh scattered 193 nm light. The Rayleigh signal

depends on the gas density while the fluorescence depends both on

temperature and density. so that the two effects are easily

separated. This latter method is less complicated 3nd more accurate

than the dual wavelength technique, but it cannot be employed with

air samples heaVily loaded with particulate matter. We have

investigated both techniques experimentally, and the results ara

deccrtbed below'. Details of the 02 spectroc~opy have been published

in Rcfer~nce 1 and in papers cited therein, and the 03 experiments

have been described in interim progress reports (Grant NAG 2-104

September 1981-March 1982, April-Septe~ber 1982, and October-Jun~

1983). This report will, therefore, document other findings of this

study.

2. A r F 1.a s e r

The Schumann-Runge absorptions of interest ~re the P and R

branch overlapping pairs of lines:

4 ......... -~' ... "~

.l:::t-

~, ~~

P13/R15 192.9 nm

PIS/Rl1 193.1 Dm

P17/R19 193.3 nm

P1S/R21 -193.5 nlll

P2!/R23 193.7 nm

P23/R25 193.9 nm

The ArF laser output peaks near 193_5 bro, but with prismatic tuning

optics it can supply sufficient energy {-1mJ} near 193.1 nm and

-1 193.9 nm In linewidths about 0.1 nm (20 cm ). This provides for

rotationally specific excitation, although the ~fficiency is not

-1 ideal because the 02 absorption lines are only 3 cm _ wide. The

primary ArF laser characteristics of interest-are the high single

pass gain and sh0rt gain duration (-10 ns). The optical cavity used

here i& more than one Qeter long. ~o that light completes only about

-one round trip during the gain illterval. Yavelengths selection

techniques are less effective in this case than they wDul~ be in a

d~vice with lower gain a~d longer buildu~ time. The laser developed

for this work employs a I-meter discharge gain channel enclosed by

CaF2

~indows, with two Br.ewster angle fused silica prisms and a total

reflector at one end an0 a slit and uncoated fused silica etalon at

the opposite end. The combina~lon of prisms and slit proves

wavel~ngth selectivity; tuning Is accomplished by adjusting the arlle

of the totally reflectirg mirror. The coufiguration is shown in

Figure 3. The tuning dp~ice cannot be very lossy (as a grating would

be), because there is al~ays nondispersed radiation generated in the

gain channel traveling In the direction of the output etalon, and

~

Aperture Slit

ArF Laser I Lens Tuning .'

. ~ x>

Ml

M2 Lenses

PM Tube

Slit ~ c::::J Filter

Output'

'.

FIGURE 3. Schematic diagram of the optical syster.l of the ArF laser used for fluorescence excitation. Tuning is accomp1i5hed by tilting the fully reflecting mirror Hi'

,,_::~" ~·''1'4

this hroadband light com~etes with the energy returning f~om the

tuning optics.

.- The laser is disLharge pumped with ultraviolet preionization

", furnished by a spark plug array. Voltage is supplied by a

Hipotronics Model 850-20 power supply which stores energy in a

50 11F capacitor. One end of this capacitor is swttched to ground by

a Tachisto Model 510 triggered spark gap, while the other is

connected to a 20 nF array of barium titanate capacitors mounted

directly above the cathode rail of the laser (see Figure 4). This

provides a voltage that reaches -25kV in approximately 70 ns and then

discharges into the channel in about 15 ns. The mechanical arrange-

ment of these componentD is critical in achieving the necessary fast

discharge. Four of the bartum titanate capacitors retur~ current to

ground through an array of spark plug preicnizers along one side of

t 1"1 e dis c h a r"g e c han n e 1 . Bee a use t his 1 a G (-! r i s to b e use d 1.1 i 1: h P r Ism

and ~lit tuning optics the electrodes are narrower (-1.2 em wide)

than conventional lasers designed to produce a square gain region.

With 30kV on the storage capaCitor and 2 atmosphetes totalgaa

-1 pressur.e the device emits 2-6 mJ in a 20 cm line~idth. The gas

mixture is approximately 10% Ar. 10% mixture of 5% F2 in He, aad 80%

pure He, separately metered and monitored in a flowing system

providing about 2 liter per minute longitudinal flow through the

laser. The gas mixture was found to ~e crliical for attaining high

gain. After the flow rates were optimized experi~entally, even a 10%

change in one of the Inlet components produced a large drop in

output. The system flow was made as large ~s possible consistent

with the laser body p~essure and resistance in the 6mm stainless

,,~ ".~ -.'-.... -~-; .. ,,'

c, ;/~' ..

rc~'''''''''''' "'~ :::":W' fr":"i~-~r<;.,..,=-·~~'r-r'"'-'~;""""'-':""';r -~~~ Y··r"·~·'·"~""'~T·t·~;'>.c''1'~CI''\;·'''·-''''''i';''·''7'''·'''''''1Tr'7",.~~~~,.".~",~~.",~~~7't;77~~~=~ .,

,.

r=voltage I'owe' + LUPI,ty ,

to other laser 20 uF array. BaTiOl

mQ

/;0 ;I

780 pF capacitors connected 21 plugs

50 turns un 3 cm di~. form.

r -' - 1 'Il~H-H-LJ.. IS. ' "b; ,.~ 000 n

__ :J~~tl . ~.. r Triggered Spark. Gap (Pressurized)

Laser Budy

FICURE 4. Laser electrical s~llCmatic diagram. '['1,0 lasers can be run in parallel I"ith independent triggering.

·0

~ ~!

~~\ r • ~ ~~ :

r:

f [ ~, t

t· r ~ ~ i;' ,

,~ f-

r , f: ~'

~

;~

l , {-F

;~.

~ i t t; ~~.'_oF

~ ""* ~'

k~: r t

~"~' :

L. "

r-- . i~~ t~~" b.~:}

steel tubing used to exhaust the gas.

A,large number of experiments were c&rried out eailler in this

program to improve the laser performance. Briefly, the results may

be summarized as follows:

1. A 60 em discharge chann~l produced much less (2 to 10 times) output energy and gain than the 100 cm channel under the best conditions.

2~ Mesh electrodes, segmented and ballasted electrodes. and electrodes with imbedded corona wire preionization produced 1ess output ttdn smooth single-piece electrodes.

3. It was not possible to Increase the duration of gain in the channel.

4. The optlmu~ prelonizatlon source was a spark gap array {modified Champion L78 spark plugs} on one side of the discharge channel, driven by only a small fraction of the discharge current. Techniques .,-:hich were successful with U'e:N2 lasers proveG quite unsuccessful with ArJ.'.

Mechanically the laser body was constructed of a single piece of

alu!O,'inun uith a Teflon covered 'acrylic plastic lid 'sealed with a

large Viton a-ring. The anode electrode rail and holes for the spark

preionlze~ pl~gs were machined into the aluminum body. The ca t;1ode

rail was Juspended from the plastic lid and attached to the capacitor

array by a row of feedthroughs. No recirculating blo~er was used to

move the gas in the body, although inclusion of such a device

probably would have improved the la~er performance.

Tvo identical laser bodies were constructed for the dual wave-

length experiments. We foun! that they could be changed at rep-ti-

tion rates up to a few Hz from a single DC ~ource witi ballast

resistance of about 2HQ between tbe source and the energy storage

capacitors. Trigger jitter was much less than one microsecond with

the spark gap pressure adjusted properly. The entire ~aser and high

_ .. -.-.- -~.,-- . ,-_._-'" -.. -~--.

.'.~.:' ..

/

I . ------1

!

.-

I . 1

I

I I I

i

~~p~

voltage system was encl~sed in an aluminum box for shielding.

Additional boxes with feedthroughs -at each end permitted us to

evacuate and backfill the optical paths with N2 • A purged tube

carried the lader beams to the experimental table. The laser

electronic syste~ was isolated from the laboratory voltage source by

a larg~ isolation transformer to reduce interference ~n ihe receiver

and computer •

Optical Re~eiver

Detection of the O2 fluorescence, which is concentrated in the

220-260 nm region, as well as monitoring of scattered 193 um energy,

i& cosily done using bialkali photomultiplier tubes with fused silica

windows. Ue used Hamamatsu type R292 and R760 detectors In this

work. For fluoresce~ce measurements it is necessary· to eliminate

193 nm radiation (without introducing additional fluor8scence) while

passing 220 nm and longer wavelengths. We found that filters of this

type are difficult to obtain. Several commerciaJly available f{Iter

glasses were tested, and we found that their short wavelength

characteristics were not as indicated in the literature. Alk,':!li

halide crystals were also evaluated. Samples of 6 mm thick KBr

transmitted 0.1% at 193 om, 45% at 220 nm, 60% at 240 nrn, and above

70% beyond 270 nm. Sodium chloride passed ~% of the 193 nm signal,

• more than 70% at 225 nm, and about 80% beyond 250 nm. Both materials

are easily degraded by laboratory humidity. The most successful

filte~s we have found are crystal calcite (CaeO) and multilayer

dielectric mirrors (narrowband 193 urn high reflectors) deposited on

.,­,.

.-

/

~ .~, 1

(' ~ ~ .:

~

f'

~ .

t

i·

r· ,

,"

;. ~

~. ,: r! :;~ oJ

~ C:;

both sides o~ ~ fused silica sub~trate. The multilayer [ilt~r passed

about 0.1% of the -193 nm energy and transmitted mvr~ t~an 70% '>f the

fluorescence. Our calcite sampl~ with a 1 cm thickness blockea

essentially all the 193 om radiation and a~out half t~e fluorescence.

This sample generated very weak fluo:escence tiear 600 nm, while the

mirrors shew slightly greater amounts of bi_e-while fluoresceu~e from

the coatings and the substrat~. In general th- coated filter is more

appropriate wheri there Is not -too much 193 nm back~rouad and the

fluorescence is weak. while the calcite wcrks ·better in sitJa~ions

where attenuation at 193 nm is most !m~ortant.

In this ~ork we employed a fused s:lica collecting lens of 25 m,

aperture &nd 50 mm Norklig distance. The fluorescence ~as imaged

cnto a slit in front of the filter and pho~otu~e so th~t a 1 ~m pa~h

length along the laser beam axis was selectedw For experiments

requiring a greater ~orkiTIg distance or a spatial re:olution better

than about a millimeter~ this approach probably is not practical.

For large working distances an all-re~lective optical syste- similar

to the Schwartzchild microscope objective can be used (see Figur~ 5).

In this system the spherical aberrations of concave and C0nVfX

mirrors compensate each ot~er. Early in this work we designed (see

Reference 2) and constructed such a collector wit~ a 30 cm ~orklng

distance, F/3 collection angle, and sample resolution better than 45

m i c r 0 r. ., t (! r s • Although such systems are bulky they provide a way to

collec~ broadband UV ~fficiently. and with the great pdvantage that

the masking aperture alignment and focusing can be done exactly with

visible light. Car~ must be taken to obtain hiJhly reflecting UV

coatings on substrates more than a few em in diame~er.

---~-----'I\

'!

~ • j I

IJ;~?~~~:~JlSr~~~:?:~~~)L~~I:z],:;~~;~:~:!.:!~~t~~~B.:~~~?:I!tr:l:?;;~:1,",,:RjryYnr1~~w.t7;m\~~mm~:::~un~$.t'&i~JT~~~~~~~~A~~'''~

Sourc.e

FlGUJ8 ).

Field Stop

~ -------~~~------- -,. --=E

~

Reflecting collector for use wIth large working distances. Both mirror surfaces ore s0lcrical. Humerical apertures up to 0,6 are practical.

'. -­/

'-

...

"

--

(II&lh~""il"""f'71..."'rY'_»n·'i".'_Qi· ..... ~~~ •• .,}.:-~ ~ "-" -.-'" .. --:--...----.~--'"

The number of dynodes used in the photot~be$ was s~lected to

provide an output current within the maximum rating for the tube,

I.e., a few-tens of microamperes. For th~ 10 ns optical pulse

duration this corresponds to an output ~h&rge slightly greater than

-13 10 coulomb. Typically~ the cathode charge wss in the

10- 14 -IS 10 coulomb range. To avoid problems with electromaen~tic

interference the output charge was stored cnthe coaxial cable

(-100 pF) connecting the phototube to its video amplifier. A

-3 potential step of about Q/C = 10 volts is therefore presented to

the amplifier input. Since poterttials in th~ one-volt range are

needed for signal processing by the digitizer and ~omputer,the

~equired amplifier voltage gain is about 1000. In experiments with

two lasers at diffcren t H8veleng ths) the pulses· mus t be time

resolved. This requires that the risetimes of the voltaga steps

should be about a micros~cond. The cir.cuit developed for this

purpose is shown in Figure 6.

The output voltages were displayed and processed in either of

two ways. The most direct way, useful when Duly one laser is

employed, is to apply the fluorescence detector voltage to the Y axis

and the laser power or Rayleigh &catteringvoltBge to the X axis of a

storage oscilloscope. with beam parameters correctly ~djusted each

laser pulse produces a small dot on the display~ the angle of the dot

positlon relative to t1.e axes representing the efficiency of the

fluorescence. As the ~Bmple density or temperature changes and more

laser samples ere coll~cted, a pattern Is formed on the screen. This

is easily recorded in a photograph for further processing. A single

streak of dots represents constant sample conditions with the laser

", I.

~.

... ------r---() + 15 volta I -L SO jJF

~ I Inputo----~ ';;..../ 10k

ION ~. 0 Output

t.s~ ;;- 56.2 k*

I..-"'-~ 100

10k

1 • 0 -15 volts

~, He tal f ilrn

111 - Lll0032CG (National Semiconductor) ISO ilF

i!2 Utl18H

..t-FlCUltE Ii. Vldt~O amplIfier circ:ult schematic diagwm. The ':"15 volt' source is a battery package

moun teo illH ide the Llwpllf ier cnd.4giur<:! to rcd!Jcc laHer interference.

"

.,"~ ..... , '\1

i j. I

I ! j I I I

! ~ri ,.

,~,

i r~ ,: ~i t; r; r~

II II ~l ~ P'<i ~ ~ !;; ~ r:; [j t~ ~1 t: ~~~ ~J ti k~ .;'" N ~1 fi u ,j i{ r' to! " H ~~

~ ;~ ~};

i. ~'i;.

~ ~,

~ ~ ~

I ~

i

I I ~~

power fluctuating. A two dimensional pattern indicates in polar form

any sample changes or noise (angle) or laser fluctuations ~radius)

from the X-Y origin. Figure 7 shows an example of such a data set.

~or dual wavelength experiments with two ls~ers a more complex

procensor is needed. For this purpose we developed stUD channel,

7 8-bit digitizer which samples the waveforms_ at 10 per second, stores

the values in a memory, and displays the sampled waveform. The

voltages are double steps, one for each laser in each detector

channel, and the two channels permit fluorescence and a second

variable (e.g., laser power or Rayleigh s~atter!ng) to be monitored

simultaneously. The system uses pairs of TRW Model TDe l007J AID

converters Bnd TOe 1030 first-in first-out memories for storing 64

pairs of 8-bit numbers. The sample =ate is adjustablo and typically

was set at 10 MHz. The sample values are transferred to a Cnm~odore

64 computer which reconstructs the w~veform on its monitor screen.

A movable cursor permits specific points on the wnveform to b~

evaluated numerically and printed. Ratios, diffprences, Bud othrr

functions are easily taken, and the waveforms themselves can be

recorded in hard copy form (see Figure 8).

Although this data collection procedure is slower than

collecting points automatically on the storage scope screen, it does

permit the collection of data corrupted by interference from the

lacer, and it facilitates statistical proce~slngw such as compuca-

tion of standard deviations.

./

t' t

ii~ ~:"~~~'~lf)~'\(' ~.'~ , ,~~ . ., ;~'t.-;; -\'-~ '-' .. -... ,~ ~ '~". "1oji."~~.~-:-,-"":~",,.;,..... ... ~._ ... ~-Wl.. .... l._:;:::""~_.' __ _

ORIGINAL P/.:~Z r~ Of.: POOR QUALITY

FIGURE 7. Oxygen fh;;orescence (vertical) vs. Rayleigh scattered laser energy tuned to P19/R21 for air at 29.50 C (lower li~e) and at 58.SoC (upper line). Each pulse is represented by one dot.

,- 'I

. ' .....

I~ ~

-'i" • .>?

~.~

CHAt'lI'JELl CHAN

(P d

NELl CHANNEL2 ~ CHANNEl2 ,..-~

.----------------.---------------.-...... --..------..... --------------- .'t!'-

.. ..",.--'-"."''''--''--'

( .

. ' ......... - .. - ................... _ ......... _. __ .. ........ _ ..... __ ..........

FIGURE 8,

' . . .

' .. .. ·.·.t .. _ (A) (B)

Output waveforms from the digitizer system with single pulse excitation (A) and double pulf;(~ excitation (B). Note the IHes~nce of interference on the Channel 2 waveform in (B). Increasing signal is dowllward."nd zero signal is the dark line at th~ top.

..,.--.' ....

.. ,,' i a" ...

~ ;

~;fl "\I"'~ n ( ;';

n t 'I­rtf /.1 l . :/'

"

...,,/

'~ .,~

/

//

//-"

, ' ,,:

3. Experimental Results

The apparatus used in the experiments is shown schematically In

Figure 9. A photoacoustic cell which could be filled with 02' H2 , or

air was used to verify the nature of the laser absorption spectrum

independently of optical fluorescence methods. Also a Sth order

diffraction grating and fluorescerit scr~en displayed -the position of

the laser wavelengths relative to the atmospheric absorption lines.

The reference signal could be derived .~ither from 193 um scattering

from the gas sample (Rayleigh reference) or from the laser power

scattered from one of the lens surfaces (laser reference), simply by

aiming the unfiltered phototube assembly in different directions. A

Scientech Model 360 power meter was used to ensure that the laser

energy was in the 1-2 mJ range when tuned to line cen tel:'. whlc't is

near t~~ P19/R2l absorption featur_ of O2

•

From basic =onsiderations ue expect that at temperature T the

fr~ction of 02 molecules in the J'th rotational grou~d state shoull

be

n(J)/no = (2J+1)(a/T) exp[-J(J+1) a/Tl (1)

where a = hc B /k = 2.0686 degrees Kelvin. v We also knov that the

total number density nO is proportional to p/kT, where D pressure

and k is Boltzmann's constant. In addition, since P and R branch

transitions occur in pairs. we can associata the Lumber J with the P

branch alone and write the detected fluorescence intensity as

"~"

~~"~~r~~~ ~~~ ~ '";"'\~ ~ :~~ 7"--:.--'" -.~ ," rlll .. g~':._t>::-~'t ... :!~" .. ~,.~:~.,..,! . .':".:: .... , '.

I I

:. '"-- ........ ~, .. -----.. -

r- ---, I ~j--------------------------I

l!

I I I

I J' I I

I

//

/

Pulse Generator

Delays H-' r~1

Tune \ ----?>l

IS I II ~n-----~ Gas p ! Out :...i '·1 I

Las:1Ji ,Lase~""'<-;- r~ne He, Ar. F?

Gas ~~---------------,

I P;:o'2 ] ~ ~ <} t

I _.J

Opto­isolator

1--~_--1 r-­I I

Detectors ana 1 +,: I .'\mpl Hiers L __ ..1

. ( ,t\IiHt[.l -~ Al.r II', ~*~ Sample \. ~ \l IU :&,

L./ . . ~,ot~, :~~usticrn n ~_

Spectrum \. ~ ~ Display J ~_ .' '_ . .:1.. I_~

. AID

[prioter~ 1/ :IOOitOr I FlFO

T ._--------.---------, Parallel Bus

Commodore 64

FIGtiRE 9. Experimental apraratus.

" .... /-"

r---'~--:'

ti

r';-'-." ~, i: i;

i

, .. ;

i l' .

~,;l

I~,,:.i 'I ,:'1

i:~,;,i, ~J _1

"

F',' : p-' f; f~<

-~! 1::,

~! r (::,

~:­It f.',:, ~., ..

_/-f; r f ~; f , f·; r '

t': : ,!"t.

U -- F ~

C I~ ,

. t-: !i',

~S:; . ,

:.::~

~ ~, ~t .,;

~~ iii

2 I(J, T, p, E) = (AEp/T ) {(2J+l) exp [-J(J+l} a/T}

+ (2J+3) exp [-(J+l)(J+2) a/T]} (2)

where-! Is the laser energy e:clting the P(J)/R(J+2) transition and A

Is a constant of proportionality~

If the fluctuations in T and n produce small changes in It then

the appropriate relation is

flI/l ... (llT/T) {[J(J+l) + (J+2) {J+3}1 (a/2T) - I} + fln/no (3)

For the special case of air at constant pressure we have

(l, 1/1) p (6 TIT) {[ J (J + 1) + (J -I- 2 )( J + 3 ) 1 ( a /2 T ) - 2 } (4)

There! is also the inve!:se problem. in \.hich we measure iJ. I/I at 2

laser wavelengths (2 J values) and we want to recover flT/T and

fln/no. In that case a direct application of Cramer's Rule gives

~T!T [C1(t.I/I)2 - Cz (flI/r)

11/(C1 C4 - C2 C3 ) (5)

~n/no [C 4 (flI/I)1 - C3(tI/I}21/(CIC4 - C2 C3 ) (6)

where C1 C2 1 in this case and

C3

(J1(J1+l) + (J 1+2)(J 1+3)] (a/2T) - 1 (7)

''''~

.:~ '-i I

.~~.~ ~.

t, t, r, J~~ : r!

J

" ~ ..

(""

: i

f ~ [" .r

./ ~ - l ~ t, ~ t ~.

~. ~ , .. ~ ~ k • t ~. t (

~ f ~ r .. ,· : l-~ ~ ~ ~ ~ f ~.~ ~ . '

t~~ .. ; f . -~' . . . ~. ~0

o '~A

i g~

C4 ~ [J 2 (J 2+1) + (J 2+2)(J 2+3)1 (a/2T) - 1 (8)

The goal of the experiments is to verify the basic relations

above and to determine the accuracy with which the measurements can

be cade under practical conditions.

A fuudamental question is the sensitivity of the system to

changes .In the fluorescence, after the laser power fluctuations are

n~r.malized out. With a single ArF laser tuned to the P19/R21 line of

room temperature laboratory air we have estimated the accuracy of the

X-Y display method with fluorescence pulses plotted against laser

power. A typical result is Figure 10, for which the peak spread is

~2%for a large number of sa~ples. Alternatively, we have used the

digitizer to sample the fluorescence/laser energy ratio f.or. the same

line with the air temperature and density held constant. The

computed statt1ard deviation for this ratir is 1.8%. These numbet's

are rc,'sonable since each detector channel 4 produced abuut 10 ,~ath~-Ie

electrons per pulse in these experiments.

Using a hot air b-Iower and thermistor' the, :)lometer to move alr at

constant pressure through the laser beam, we were able to verify the

tecperature coefficients for the fluorescence over the 20°C to 60°C

range. P~r example, by comparing the fluorescence signal to Rayleigh

scattered signal for air at one at~osphere pressure at two

t~mperatures. l2SoK and 296 o K, we obtain the following:

Line (Fluorescence/Ravlci~h) at 325K ( Fluores cence I RayTef~1l)a t 296K

Experiment

P17/R19 1. 128

P23/a25 1.302

Theory

1. 114

1.302

.: :, q >"l' \~( . ,~." . ,,; ..

" , ~ .~{ ,

: . ". fti~e~~~~!;h;,\ ~~.- ~:~~~?~. 'J"~ / \~>r,~,?"/~:""'~~"'~j":!)""::~"':""'J""'I.~?

i'l

'~;r:~~~v;,~~~~'~: tJ~?,?~~~~·W:i ., ·~,..:P.r:;sk4::::''''':''~~'> , .·t~;r,~P;:r:~f,"."

,. t

II>

\lU

FIGURE 10.

',' , ":'

a ii," R WU

o ~\i,· ••

W··" .:"

" o en UJ

~ l.L

'~ \'",

~I •

'~\I

1:1,;

-.,....===~

LA~ EN

", I,

--> ,,";j

> • .1,',

... '

"1'

"

.. ~(

",1<. ~

,;::::wt.s:mn:;Z£J:::m::::zc===~

y XY display of uxygen fluorescence vs. la:.;er enf-'!rgy tuned to the P19/R21 absorption wavelength. Vertical t,:hickness of the trace indicates the noise level.

t:~

00 "11 :tI "OQ O ~~ d:i .. 0'1':' :r:.d'1 C· ~~j c.: >,~. :'';') r,,'; !:' 1;1 -1 r.' .-<1JJ

\/,,'-,.

I''', , ~i

,'" ~ 'r,'

• .1','1 ;'~~lllH~: q,~!~, " 'I;";J'::':.~:;,. ttl

~, I.~';~'~· .,,1

,'~ij •.. ~ .~? 11" 1,-iJ

f'.-:;r., r. "

v~ !~;il I~ /1 ,- n i'p f,:~l (:.lj ··1

t,'

/1 ./ ;:

f,,; ,/ :'

I , i \ 1 I I I r I.'

· End of Document

__ --"TI""F"r