laser remote sensing of hydrazine, mmh, and udmh using a differential-absorption co_2 lidar

Laser remote sensing of hydrazine, MMH, and UDMH using a

differential-absorption CO2 lidar

N. Menyuk, D. K. Killinger, and W. E. DeFeo

A dual mini-TEA CO2 laser differential-absorption lidar system has been used to test the remote sensing of

hydrazine, unsymmetrical dimethylhydrazine (UDMH), and monomethylhydrazine (MMH) in atmospheric

conditions. Average concentrations of these compounds were measured using backscattered laser radiationfrom a target located at a range of 2.7 km. The experimental results indicate that average atmospheric con-

centration levels of the hydrazine compounds of the order of 40-100 ppb can be detected over ranges be-

tween 0.5 and 5 km. The level of concentration sensitivity over this interval was found to be limited primar-

ily by atmospheric fluctuations. An investigation of the effect of these fluctuations on measurement uncer-

tainties indicated that the fluctuations reduce the benefits of signal averaging over N pulses significantly

below the expected square root of N improvement. It is also shown that uncertainties due to long-term at-

mospheric drifts can be effectively reduced through use of dual-laser lidar return ratios.

1. Introduction

The single-ended differential-absorption lidar(DIAL) technique has proved to be a sensitive methodfor the long-range remote sensing of molecular con-stituents in the ambient atmosphere. DIAL systemsoperating in the infrared have been used for monitoringthe presence of major atmospheric constituents such asCO,1' 2 as well as trace constituents such as C2H4,3 ,

4 NO,5

SO2,6 ozone,7 and others.8 In this paper we report onthe use of an infrared DIAL system to test the long-range remote sensing of hydrazine, monomethylhy-drazine (MMH), and unsymmetrical dimethylhydrazine(UDMH) in atmospheric conditions. These com-pounds, which are used as rocket fuels among otherapplications, are known to be highly toxic.9 To ourknowledge these measurements represent the first ex-perimental test of long-range laser remote sensing oftoxic hydrocarbon vapors in an atmospheric environ-ment.

The possibility of accidental release of these com-pounds into the atmosphere, coupled with their acutetoxicity, makes the ability to sensitively monitor theirpresence in the atmosphere extremely important.

The authors are with MIT Lincoln Laboratory, Lexington, Mas-sachusetts 02173.

Received 14 January 1982.

0003-6935/82/122275-12$01.00/0.) 1982 Optical Society of America.

These considerations led Loper and his colleagues10 tomeasure the vapor-phase absorption cross sections ofthese compounds at a large number of CO2 laserfrequencies, where the hydrazines have moderatelystrong absorption bands. They used the results to es-timate achievable detection sensitivities for the hy-drazines and experimentally confirmed their estimatefor UDMH in a laboratory experiment using a pho-toacoustic detection system.

In this paper an experimental study is presentedwhich includes an extension of the Loper et al. 10 mea-surements as well as an investigation of the capabilitiesand limitations of the DIAL technique as a rapid andsensitive means for remotely measuring the presenceof hydrazine compounds in the atmosphere. Our ex-periments were carried out using a dual-laser CO2 DIALsystem in conjunction with a large optical tank whichcontained the toxic hydrazine compounds. Lidar re-turns were obtained from laser beams which passedthrough the tank and were reflected from a topographictarget at a range of 2.7 km. The experimental resultsestablished the capability of the DIAL system to re-motely sense the presence of the hydrazine compoundswith a path-averaged detection sensitivity of 40-100 ppbat this range, with the limitation set by atmosphericfluctuation. In addition the results showed the dual-laser DIAL system to be capable of performing thesemeasurements in real time despite the known rapidreactivity of the hydrazine compounds in air.11 12

The remote sensing measurements and an analysisof the experimental results form the principal subjectsof this paper. A description of the DIAL system usedfor these measurements is given in Sec. II. The basisfor the choice of CO2 laser transitions used for each of

15 June 1982 / Vol. 21, No. 12 / APPLIED OPTICS 2275

TARGET

I,

,1-17 REMOVABLE/' / ABSORPTIONL j TANK

STEERING ~ M ~ DTCO

BEAM MIRROR ok -$

LS- BS E 2NO B> S - ABSORPTION -- _PR

TIMING CAMAC~COMPUTERDELA UIED P ARDO (L.Croy 3500)

LASER~~~~~DSPA

Fig. 1. Schematic of dual-laser lidar system used for remote sensingof hydrazine and its derivatives.

the molecules investigated is discussed in Sec. III. Theeffects of an air atmosphere on the temporal variationof the hydrazine compounds and the ability of the DIALsystem to follow these changes were studied in thelaboratory, with the results described in Sec. IV. Theremote sensing experimental results are given in Sec.V, and these results are analyzed in Sec. VI, where it isshown that atmospheric fluctuations are the majorsource of measurement uncertainty. To achieve abetter understanding of these fluctuations and of theimprovements achievable through signal averaging,additional experiments were carried out to study ex-plicitly the accuracy and sensitivity limitations imposedon remote sensing measurements by the combinationof atmospheric and temporal effects. The results ofthese experiments are given in Sec. VII.

II. Experimental Apparatus

The dual-laser DIAL system used in these experi-ments is shown schematically in Fig. 1 and has beendescribed previously.5 13 Two line-tunable mini-TEACO2 lasers14 provided the pulsed transmitted radiation.Laser 1 was normally tuned to the low-absorptiontransition frequency, and laser 2 was tuned to thehigh-absorption frequency of the hydrazine, UDMH,or MMH species. The two lasers were separatelytriggered, and the time delay between the firing of thetwo lasers was maintained at 50 ,usec. For this short adelay time the atmosphere is essentially frozen betweenfirings.13 The beam paths from the two lasers werejoined through use of a 50/50 beam splitter. A portionof each beam was then directed to a pyroelectric de-tector to normalize the output of each pulse, and an-other portion was sent through a Pyrex absorption cell(33 cm long X 2.54-cm diam with BaF 2 windows) to asecond pyroelectric detector. The cell served to cali-brate absorption levels of the hydrazine compounds atthe frequencies emitted by our lasers. The major part

of the laser radiation was sent through a 1oX beam ex-pander and directed toward a topographic target locatedat a range of 2.7 km. The lidar returns and pyroelectricsignals were recorded and analyzed in a computerizeddata acquisition system, which determined the nor-malized differential-absorption of both the laser ra-diation through the laboratory absorption cell and thelidar returns. To measure absorption of the hydrazinecompounds along the lidar path a large enclosedchemically inert polypropylene tank 104 cm long X62.5-cm diam with 150 slanted polyethylene windowswas placed outside the laboratory in the path of the laserbeam. A moving polypropylene flapper was placedinside the tank to reduce stratification of the hydrazinespecies. The use of the tank as an enclosure for thehydrazine compounds was necessary in view of theknown toxicity9 of these species.

Ill. Choice of Laser Frequencies for Differential-Absorption Measurements of Hydrazine, UDMH,and MMH

The choice of laser frequencies for the accuratereal-time remote sensing measurements of hydrazinecompounds in the atmosphere must provide for a suf-ficient difference in absorption levels to yield a mean-ingful differential absorption in the laser return signals.In addition atmospheric transmission and possible in-terference effects from other atmospheric species mustbe considered. The choice of frequencies is limited tothose laser transitions available with our mini-TEA CO2lasers. Single-line radiation has been obtained withthese lasers from the P(8) to P(36) and the R(10) toR(34) lines of the (0001)-(100,02o0)i transition near10.6 um and from the P(10) to P(34) and the R(10) toR(34) lines of the (001)-(10O0,02o0)1i transitions near9.4 m.14 The differential-absorption coefficients ofhydrazine, UDMH, and MMH at these CO2 lasertransition frequencies have been measured by Loper etal. 10 using a tunable low-pressure cw CO2 laser.

The atmospheric transmission and possible inter-ferences of other molecular species are obtained fromthe AFGL tapes of McClatchey et al.'5 Among thepotential interfering species, ethylene and ammonia areof particular importance. Ethylene exists as a tracemolecule in the atmosphere and is present in vehicleexhausts. Average concentration levels in excess of 60ppb have been observed over an air base.4 Ammoniais both an atmospheric trace molecule with a normalconcentrations between 2 and 20 ppb and a reactionproduct of hydrazine decomposition. We have used theabsorption coefficients of ammonia and ethylene ob-tained at the CO2 laser frequencies by Patty et al. 17

On the basis of the above considerations, CO2 laserfrequency pairs were chosen which yielded as large adifferential absorption as possible consistent withminimal interference effects from either ethylene orammonia. It should be noted that it is desirable tochoose the frequency pairs close together to maximizethe mutual coherence'8"19 of the two laser beams; how-ever, this proved to be impractical for both UDMH andMMH.

2276 APPLIED OPTICS / Vol. 21, No. 12 / 15 June 1982

A. Hydrazine (N2H4)

Significant absorption of hydrazine' 0 occurs only inthe 10.6-gm band of CO2 laser transitions. CO2 laserwavelengths longer than the P(30) line are eliminatedby the extremely high NH3 absorption. The mostsuitable laser transition pair is the P(22) transition at10.611 ttm as the high-absorption line and the P(28)transition at 10.674 Am as the low-absorption line. Thecorresponding absorption coefficients'0 are 5.41 and2.17 (cm atm)-1, respectively, yielding a differentialabsorption of A = 3.24 (cm atm)-'. With this choicethe frequencies are closely spaced, the differential-absorption levels of both C2H4 and NH3 are smallerthan that of hydrazine by over an order of magnitude,and the atmospheric transmittance level is acceptableat both frequencies. The corresponding absorptionvalues are given in Table I along with the backgroundatmospheric attenuation 3 for a U.S. Standard Atmo-sphere.1 5

B. UDMH (CH3)2N2 H2

The choices of laser transition pairs for DIAL mea-surements of UDMH are extremely limited. The ab-sorption coefficient of UDMH in the 9.4-,um CO2 laserband is -1.5 (cm atm)-1, with very little variationwithin the R and P branches.10 Significant absorptionoccurs in the 10.6-,m band only for the longer wave-lengths of the P branch, rising from -1 to near 3 (cmatm)-1 between the P(28) and P(36) laser transitionsof the 10.6-gm band. There is virtually no absorptionof UDMH within the R branch of the 10.6-gm band [o

0.05-0.2 (cm atm)-1].Absorption interferences occur in the 10.6-gm P

branch due to NH3 absorption lines. Between theP(32) and P(36) transitions, NH3 absorption is a factorof 3-7 times greater than that of UDMH. AlthoughNH3 is only a minor product of UDMH decomposi-tion,12 this large a relative absorption level coupled withthe possible presence of even small amounts of ambientNH3 effectively allows only the P(28) or P(30) to beused for the high-absorption transitions. The UDMHabsorption coefficients for these lines are 1.11 and 1.45(cm atm)-1, respectively.10

For the low-absorption transition no other line couldbe found within the P branch of the 10.6-,um CO2 laserband, which, in conjunction with either the P(28) orP(30) transition, would yield a differential absorptionlevel which was of the order of unity and significantlydiscriminated against both NH3 and C2H4. To do soit was necessary to choose the low-absorption line fromwithin the R branch of the 10.6-gm band. Thus themost suitable laser transition pair appears to be theP(30) line as the on-resonance transition and the R(10)line as the off-resonance transition, both in the 10.6-gmband. The relevant absorption values are given inTable I.

C. MMH (CH3N2H 3)

It is difficult to find a suitable pair of CO2 lasertransitions for the remote sensing of MMH. There islittle MMH absorption anywhere within the 9.4-gm

Table 1. Relevant Absorption Parameters for the Remote Sensing of theHydrazines

AtmosphericCO2 laser Wavelength Absorption coefficients attenuation

Transition (Mm) (cm atm)-1 0(km)-l

b c d d aHydrazine NH3 C2 H4

P(22) 10.611 4.77 5.41 0.045 1.09 0.1142P(28) 10.675 2.06 2.17 0.36 1.30 0.0976Differential 2.71 3.24 -0.315 -0.21absorption

UDMH NH3 C2H4P(30) 10.696 2.22 1.45 0.86 1.63 0.0907R(10) 10.318 0.18 0.05 0.78 1.51 0.1142Differential 2.04 1.40 0.08 0.12absorption

MMH NH3 C2 H4R(30) 10.182 1.69 1.36 0.029 0.56 0.1137R(18) 9.282 0.31 0.23 0.13 0.61 0.1418Differential 1.38 1.13 -0.10 -0.05absorption

a Values given for U.S. Standard Atmosphere.b This work.c Ref. 10.d Ref. 17.

band, while absorption value variations within theavailable transition lines of the 10.6-gm band are ex-tremely small. The single exception is the R (8) line forwhich a- = 3.5 (cm atm)-' compared with values nearunity for neighboring transitions. 1 0 However, the CO2mini-TEA lasers do not operate easily on this line; inaddition the absorption coefficient of NH3 at the fre-quency of this transition is extremely large, namely, 25.8(cm atm)-1. Therefore, it was necessary to considertransition pairs involving CO2 laser lines in differentbands. It was found that of all the 9.4-,gm band tran-sitions only the R (18) transition will yield a sufficientlylow C2H4 differential-absorption level to serve as asuitable off-resonance line. Further consideration es-tablished the R(30) line in the 10.6-gm band as thepreferred on-resonance MMH absorption transition.The absorption values for this transition pair are alsogiven in Table I.

Although the absorption coefficients given by Loperet al. 10 were used to establish the preferred transitionsfor differential absorption, the difference in the natureof their CO2 laser and the CO2 lasers used in our DIALsystem requires that we separately determine thecoefficients in conjunction with our differential-ab-sorption measurements. Loper's results were obtainedwith a low-pressure cw CO2 laser, which generally hasa narrow-gain bandwidth (-100 MHz) capable of sup-porting only a single longitudinal mode located at ornear line center. However, to achieve the high peaklaser power needed for long-range remote sensing ourcoherent radiation was provided by pulsed mini-TEAlasers operating at atmospheric pressure. These lasershave collision-broadened gain bandwidths of the orderof 1 GHz and generally operate on more than one lon-gitudinal mode within a single transition line. Each


mode operates at a slightly different frequency, andsome tunability exists within each line. Therefore,there may well be differences between the absorptioncoefficient of a given molecule for a given CO2 lasertransition as measured by a TEA laser and by a low-pressure laser. This problem is overcome in our ex-periments by measuring the absorption coefficient ofthe molecule in the absorption cell simultaneously withthe remote sensing measurement. This was done inconjunction with the remote sensing experiment de-scribed in Sec. V; our resultant absorption coefficientvalues are included in Table I.

IV. Laboratory Measurements of HydrazineCompounds in Air

Measurements of the differential absorption of thehydrazine compounds in an air atmosphere were carriedout within the laboratory using the laboratory portionof the system shown in Fig. 1. In view of the reactivityof the hydrazine compounds in air, these measurementswere necessary to establish the ability of our dual-laserDIAL system to follow the instantaneous changes inconcentration of the compounds in real time and todetermine the extent to which this can be accomplishedwhile averaging over a sufficiently large number ofpulses to significantly reduce measurement uncer-tainties.

The measurements consisted of firing laser 1 on thelow-absorption transition line followed by laser 2 firing50 gsec later on the high-absorption transition. Eachbeam was split, with one portion going to pyroelectricdetector 1 to serve as the normalizing beam. The re-mainder of the beams passed through the Pyrex ab-sorption cell, which initially contained air at atmo-spheric pressure and to which a known amount of ahydrazine compound (in liquid form) was added. Thebeam transmitted through the cell was recorded bypyroelectric detector 2. The detector outputs alongwith gating circuitry then went to the computer systemwhere each individual pulse was normalized to the laserenergy in each pulse on a pulse-to-pulse basis. Thenormalized pulses from each laser were averaged overa preset number of pulses. The computer system wasable to perform the normalization and averaging func-tions in real time, with the lasers operating at a pulserepetition frequency of slightly under 20 Hz.

The 100% transmittance level through the absorptioncell was established by recording the values of the nor-malized laser beams averaged over a thousand pulsesafter passage through the absorption cell containing airat atmospheric pressure. A known volume of a hydra-zine compound in liquid form was then inserted byhypodermic syringe into the cell, and the laser beamtransmittance was observed as a function of time.Initial values of the average relative transmittance weretaken for every 100 pulses; after the rate of change de-clined sufficiently, values were taken of the average of500 pulses from each laser. The system was able toaverage over this number of pulses and still follow thechanging concentration levels without difficulty. Asthe following results will show the temporal change in

zhi

C,)Z

W

Id

04

5 toTIME (min)

15

Fig. 2. Time variation of relative transmittance of 10.6 -Am P(22)and 10.7-pm P(28) radiation through a 33-cm long laboratory ab-sorption cell containing air after insertion of 0.9 and 7.6 liter of liquid

hydrazine.

the concentration of the three compounds differedmarkedly from each other.

A. Hydrazine

The change in the relative transmittance of the P(22)and P(28) CO2 laser transitions of the 10.6-gm bandthrough the air-filled laboratory cell after introducing0.9 and 7.6 gliter of hydrazine is shown as a function oftime in Fig. 2. The results obtained using 0.9 gliter ofhydrazine indicate a maximum differential absorptionalmost an order of magnitude lower than the value thatwould be expected on the basis of published absorptioncoefficients'0 if one assumed all the hydrazine waspresent in vapor form. This large reduction is due toa combination of two effects; first, hydrazine has a slowvaporization rate due to its low vapor pressure (14Torr at 298 K),9 and, second, the hydrazine vapor hasa high oxidation rate. Hence this result is not partic-ularly surprising. It was observed that even after theabsorption levels started to decrease, which occurred-12 min after the introduction of hydrazine, some liq-uid was still present in the cell.

Figure 2 also shows the results after introducing 7.6Aliter of hydrazine, an amount exceeding that requiredto produce a saturated vapor pressure (-4.4 gliter). Inthis case the relative transmittance reached a minimumvalue in <4 min. and returned to a near-normal valueafter 15 min. The maximum effective differentialabsorption achieved in these conditions was a factor of-4 below that predicted for saturated hydrazine vaporin the cell.

It is apparent from Fig. 2 that the DIAL system iscapable of following relatively rapid changes in hydra-zine concentration while averaging over large numbers(-100-500) of pulses.

B. UDMH

The time variation of transmittance through theair-containing absorption cell after the insertion of 9gliter of UDMH is shown in Fig. 3. These results arequite different than those shown in Fig. 2 for hydrazine.


' ' ' | | | ..- '-.-'-.-,-' ,- . ' I, r

HYDRAZINE7 ice < 8/ I N AN

(LaboratoryAbsorption Cell) -

0.9 MICROLITERS; _\ Spot +~~ P (22) -

3 ,e" X~~~ P (28)

7.6 MICROLITERSo P (22)* P(28)

0.7

06

W 08

0 07 UDMH IN AIR(Laboratory Absorption Cell)

C/ 06z

I-0.5

. 04 (P30)

Of 03 -

02-

01-

,lll ll, ,,I,,I~~~~~~~I I1t,,

0 5 L 15 20 2

TIME (min)

Fig. 3. Time variation of relative transmittance of 10.7-gm P(30)and 10.3-,m R(10) radiation through a laboratory absorption cell

containing air after inserting 9 ,liter of liquid UDMH.

Io

10

I-

06 - R (30) MMH IN AIR

Z:Q (Laboratory Absorption Cell)

> 04

-J 03__03

02-

0.1

0 5 10 15 20 25

TIME (min)

Fig. 4. Time variation of relative transmittance of 10.2-um R(30)

and 9.3-gm R(18) through a laboratory absorption cell containing air

after inserting 9 gliter of liquid MMH.

The vapor pressure of UDMH at room temperature(-160 Torr)9 is an order of magnitude greater than thatof hydrazine, and as seen in Fig. 3 the relative trans-mittance of both the P(30) and the R(10) radiationreaches a minimum within a minute after the insertionof the UDMH into the cell. The transmittance levelthen remains virtually constant, in agreement withstudies1220 which indicate that UDMH is highly stableeven in an oxygen atmosphere.

C. MMH

Monomethylhydrazine has a vapor pressure of -50Torr at room temperature9 and is much less volatilethan UDMH. The variation with time of the relativetransmittance of the 9.28-gm R (18) and 10.18-gm R (30)transitions after inserting 9 gliter of MMH into theair-filled absorption cell is shown in Fig. 4. While therelative transmittance of both transitions decreasedrapidly for the first 3 to 4 min, the relative transmit-

tance of the 10.18-,gm R(18) transition reached a min-imum and then increased significantly as the MMH wasoxidized, while that of the 9.28-,gm R(18) radiationcontinued to decrease slowly. Apparently the oxidationproduct of MMH is more absorbing at the R(18) fre-quency than MMH itself. As a result the differentialabsorption of the two lines reduced to zero after -14min and then reversed sign.

To ascertain if ammonia, which is a known minoroxidation product of MMH,12'2 1 might be the source ofthe continued absorption of the 9.28-gm R (18) line inFig. 4, a similar experiment was performed with the9.29-gm R (16) line substituted for the R (18) transition;the ammonia absorption coefficient at the R(16) tran-sition frequency is over 2 orders of magnitude greaterthan at the R(18) frequency. The effective absorptionof the MMH product at the R (16) frequency obtained30 min after insertion of MMH into the absorption cellwas found to be only 10% greater than the value ob-tained with the R(18) frequency. Ammonia cantherefore be eliminated as the molecule primarily re-sponsible for the continued decrease of the R(18)transmittance.

Methanol is another known product of MMH de-composition, with absorption in the 9-10-gm region ofthe spectrum. However, this molecule can also beeliminated as its absorption coefficient is greater at the10.18-,gm R(30) frequency than at the 9.28-gm R(18)frequency. It is possible that the particular moleculecausing the effect may be one of the formaldehyde hy-drazones which form the bulk of the MMH oxidationproducts.1 2

V. Laser Remote Sensing Measurements ofHydrazine, UDMH, and MMH

Laser remote sensing measurements of hydrazine,UDMH, and MMH were carried out using the systemshown in Fig. 1, with the polypropylene tank2 2 set be-tween the coaxial transmitter/detector and a 1- X 1-mflame-sprayed aluminum target which was located 2.7km distant and served as a diffuse target.

For these experiments both the tank and the labo-ratory cell were initially filled with nitrogen gas, andsimultaneous measurements were made of the nor-malized lidar returns and of the laser radiation trans-mitted through the cell. After averaging 1000 pulsesto establish the 100% transmittance level a knownamount of a hydrazine compound was inserted into thecell, and, as previously described, average relativetransmittance levels were taken after a preset numberof pulses from each laser. These measurements werecarried out with the hydrazine compounds in an inertnitrogen atmosphere; therefore, the results could be andwere used to determine the absorption coefficients ofthe compounds at the measurement frequencies.

Shortly after inserting a hydrazine compound into thecell a larger amount of the same compound was insertedinto the nitrogen-filled tank. The results of the si-multaneous cell transmittance and lidar measurementsare discussed below for each of the hydrazine com-pounds in turn.


A. Hydrazine

The relative transmittance of the 10.61-grm P(22) andthe 10.68-gAm P(28) radiation through the nitrogen-filled cell after inserting 1.6-gAliter hydrazine is shownas a function of time in Fig. 5(a). The normalized lidarreturns obtained simultaneously are given in Fig. 5(b).The lidar data shown were obtained after the insertionof 0.8-mliter of hydrazine into the nitrogen-filled tank.The lengthiness of the run was necessitated by the slowevaporation rate of the hydrazine in the tank; someliquid was present until almost 90 min after insertion.Each point shown during the initial period representsthe average value of 500 pulses; this was increased to a1500-pulse average and finally to a 3000-pulse average.The minimum relative transmittance values shown inFig. 5(a) correspond to absorption coefficients of 4.77and 2.06 (cm atm)-' for the P(22) and P(28) transitionfrequencies, respectively. These results are in rea-sonable agreement with the respective published val-ues' 0 of 5.41 and 2.17 (cm atm)-'.

The most notable feature of the experimental resultsare seen in Fig. 5 is the marked contrast between thesmooth variation of transmitted radiation through thelaboratory cell and the large amount of scatter in thelidar returns. This effect was observed with all thehydrazine compounds. The deviations from a smoothline of the points shown in Fig. 5(a) of relative trans-mittance through the laboratory cell were generally <1%and approached the 0.1% digital quantization error ofthe data acquisition system.

The normalized standard deviation of the lidar 1P(28) return from the smooth line shown in Fig. 5(b)was 5%, while the corresponding value for the lidar 2P(22) return was 2.5%. The difference may have beendue to electronic effects or to differences in the pointingaccuracy of the two laser systems.

It should be noted that the corresponding points forthe relative transmittance through the laboratory cellin Fig. 5(a) and the lidar returns through the large tankgiven in Fig. 5(b) are based on the same set of laserpulses and identical normalization factors. Therefore,the larger deviations observed from the lidar returnsmust arise from effects occurring outside the confinesof the laboratory. These are presumably due to acombination of laser beam properties and atmosphericeffects. These deviations are the major factor limitingthe sensitivity of our concentration measurements.The nature of these limitations and a general discussionof error sources are given more fully in Sec. VI.

B. UDMH

Similar sets of measurements obtained after inserting5.55-gAliter UDMH into the cell and 1.95 mliter into thetank are given in Figs. 6(a) and 6(b). The initial set ofpoints are based on 200-pulse averages followed by 500-and 1000-pulse averages as indicated. The temporalbehavior of the transmittance curve through the ni-trogen-filled cell is seen to be similar to that shown inFig. 3, where an air-filled cell was used. The stable levelof relative transmittance obtained with the nitrogen-filled cell corresponds to absorption coefficients of 2.22

HYDRAZINE IN NITROGEN

0 O8

It 026oMQ2

0EQZ

C ZO 40 60 80

TIME (min)

'500 1500

100 120

3000PULSE AVERAGES

Fig. 5. Simultaneous differential-absorption measurements of hy-drazine in a nitrogen-filled laboratory absorption cell and tank: (a)time variation of relative transmittance through a cell after inserting1.6 liter of hydrazine; (b) lidar returns from laser pulses passingthrough tank and reflected from topographic target after inserting0.8 mliter of hydrazine. The number of pulses averaged for each point

is given below.

UDMH IN NITROGEN

I

WD

b0

(

I

0

a

oWa:

0 5 10 15 20 25 3C

TIME 10in)!O -I~~~~~~~~~

200 ' 500 1000 -I

PULSE AVERAGES

Fig. 6. Simultaneous differential-absorption measurements ofUDMH in a nitrogen-filled laboratory absorption cell and tank: (a)time variation of relative transmittance through a cell after inserting5.5 g1iter of UDMH; (b) lidar returns from laser pulses passingthrough tank and reflected from topographic target after inserting

1.95 mliter of UDMH.


| l.0 T I | fi I I I I I

L\(a ) LABORATORY CELLis

16 w ~P28)

k4 _ (22

k.2

0 . I _ I I , I I , I I

bI EXTERNAL TANK

; C<FoO art j (28)

1 'i., _ ~ -. P . 2 .- so>>>gO~P1(221)

I . . .e n I I I I . . . . I ,

1.0 (al LABORATORY CELL

lB R (10)

'.6

.4

t2

a , , , . I . . . . I . . . . I . . . .

0 , ,,-* ,,,,,, I, ETERNAL TANK

B 54 ,,,, R (10) * _

-U-K

° -o*°wo o P1 30)ood °

4

2

IA1 I 4 .___ * 1 4 __ __ b~~~~~~~~~~~-

M. C

a:

,,, C

!a�4 CI:

.

I

I

and 0.18 (cm atm)-' for the 10.7-,um P(30) and 10.3-gmR(10) CO2 laser transitions, respectively. These valuesare in excellent agreement with the results obtainedwith the air-filled cell as given in Fig. 3. However, theyare significantly higher than the respective values of 1.45and 0.05 (cm atm)-' obtained by Loper et al. 10

The lidar returns did not show an abrupt change afterinsertion of UDMH into the tank such as was seen in thecell. This was due to the longer time required toevaporate the larger quantity of UDMH liquid and forthe vapors to be dispersed throughout the tank volume.The scatter in these lidar returns was also of the orderof 5%.

C. MMH

The corresponding curves for monomethylhydrazineobtained after inserting 6.3-giliter MMH into the celland 1.55 mliter into the tank are given in Fig. 7. Theabsorption coefficients based on the results shown inFig. 7(a) are 1.69 and 0.31 for the 10.18-gtm R(30) andthe 9.28-gim R (18) transitions, respectively, which areslightly higher than the corresponding values of 1.36 and0.23 given by Loper et al. '0 The lidar return scatter isagain seen to be large, with an average 4-5% deviationfrom a smooth curve.

The experimental results shown in Figs. 5(b), 6(b),and 7(b) clearly demonstrate the ability of the dual-laserDIAL system to monitor the presence of hydrazine andits derivative compounds by remote sensing over a rangeof 2.7 km. Using the absorption coefficients in thelaboratory absorption cell simultaneously with the lidarmeasurements, the lidar results indicated maximumvapor levels in the tank corresponding to concentrationsof -1500 ppm of hydrazine and UDMH and a concen-tration of 3000-ppm MMH. These values are -20% lowfor hydrazine and UDMH and -40% high for MMHrelative to the amounts predicted on the basis of theliquid volume inserted. These results are not unrea-sonable in view of the uncertainties in the absolutevapor concentration within the tank. The uncertaintiesinclude possible error in the syringe reading, the pos-sible presence of oxygen in the tank due to incompleteevacuation of air from the tank prior to insertion of theliquid and to leakage thereafter, and stratification of thevapor after evaporation. The latter effect was almostcertainly the major source of the large excess absorptionobserved for MMH as the flapper used to preventstratification was not operational during the MMHexperiment.

VI. Analysis of DIAL Sensitivity of the HydrazineCompounds

The ability of the dual-laser DIAL system to detectthe presence of the hydrazine compounds for a 2.7-kmlidar range was established by the experimental resultsdescribed in the preceding section. These results canbe used to determine the limits of sensitivity with which*such measurements can be made at other detectionranges, that is, to establish the minimum path-averageddetectable concentration nmin of a trace species as afunction of detection range R.

MMH IN NITROGEN

- 0

z O62I

I 04

1 02

Z2,

Li

211LI2;d6

1.0 - * (b) EXTERNAL TANK

06 -

024

Q2 _

0 l D0 D0 40 50 60 70

TIME (min)

PULSE AVERAGES

Fig. 7. Simultaneous differential-absorption measurements ofMMH in a nitrogen-filled laboratory absorption cell and tank: (a)time variation of relative transmittance through a cell after inserting6.3 gliter of MMH; (b) lidar returns from laser pulses passing throughtank and reflected from topographic target after inserting 1.55 mliter

of MMH.

There are two common approaches taken to evaluatenmin- At long ranges the limitation is generally estab-lished by setting the difference in the backscatteredreturn at the two frequencies equal to the noise of thedetector. At shorter ranges a more restrictive limitationmay occur due to the inability of the signal processingsystem to distinguish changes in the normalized returnsignals APr = Pr (X1) - Pr (X2) at the two frequencies XAand X2 below some predetermined value. The effectsof both limitations on the detection sensitivity of mo-lecular pollutants have been discussed previously.4 Forthe case in which the difference in the return signals APrdue to the differential absorption of the absorbingspecies is set equal to the noise equivalent power (NEP)of the detector it was shown that

(NEP)'zrRnmin 2KpAPo(Au) exp(-23R) (1)

where K is the receiver efficiency, p is the reflectivityof the topographic target, A is the area of the receivingtelescope, Po is the output power of the laser beam,exp(-23R) is the atmospheric attenuation, and Au'r isthe difference in the absorption coefficients of the pol-lutant at the two wavelengths [Au = a(X2) -(X)]-

The values taken to evaluate Eq. (1) for our DIAL sys-tem are Po = 105 W, A = 600 cm2 , (NEP) = 2 X 10-8 W,K = 0.1, and p = 0.1. Then expressing nmin in parts perbillion,

nmin(Ppb) 5.2R 2

(A) exp(-2/3R) (2)

where the units are R(km), 3(km-l), and Acr(cmatm)-l.


- I I | I t , I , I r r .

a) LABORATORY CELL

R (18)

X^> ~~/R (30)

. O fI A a I . I . I I I A

-0 so5o ,000

The values of nmin as a function of R have beenevaluated on the basis of Eq. (2) for hydrazine, UDMH,and MMH using our experimentally determined valuesof Ao and the U.S. Standard Atmosphere absorptioncoefficients of : as given in Table I. The results areshown by the curves of Figs. 8, 9, and 10 labeled NEPLIMITED. They represent the ultimate sensitivity ofthe instrumentation and determine the maximumachievable range for a given pulse energy level.

At shorter ranges a more restrictive limitation thanthat given in Eq. (2) may occur due to the inability of themeasurement system to distinguish between the frac-tional change in the lidar return signal (APr/Pr) due toreal variations of the species concentrations and randomfluctuations caused by atmospheric turbulence andother perturbations. For this case it may be shown4

that5 X 103 (APr/P)

nmin = (Au)R (3)

The smallest distinguishable value of APr/Pr in Eq.(3) must be at least as great as the fluctuations of thelidar returns around a best-fit smooth curve. A mea-sure of these fluctuations is their standard deviation j,which may be taken as the lower limit of the value ofAPr/Pr in Eq. (3). On the basis of the laser remotesensing data presented in Sec. V a pulse-averaged valuefor ul of -0.05 was obtained at a range of 2.7 km. Asseen in Figs. 5, 6, and 7, little significant improvementin ul was observed as the number of pulses averaged wasincreased from a few hundred to 3000 pulses.

In general the value of APr/Pr in Eq. (3) may be afunction of both the DIAL system parameters and at-mospheric effects. It is, therefore, reasonable to assumethat APr/Pr will contain both a constant term and arange-dependent term. The value of the constant termcontained within APr/Pr may be ascertained fromshort-range DIAL measurements as well as those ob-tained from laboratory measurements, which representthe limiting case of R 0. Our measurements indicatethat a conservative estimate of the value of the constantterm for our present DIAL system, (APr/Pr)min, is of theorder of 0.01. The limitation due to this term alonecorresponds to the dashed line labeled 1% in Figs. 8, 9,and 10.

The range-dependent contribution to APr/Pr arisesfrom atmospheric turbulence effects. Although theatmosphere is effectively frozen during the 50-usecdelay between laser firings of our dual-laser DIAL sys-tem, the influence of atmospheric turbulence is signif-icantly reduced only when the temporal correlationcoefficient of these signals on a pulse-to-pulse basis isequal to or near unity. It has been found13 that systemnoise caused by electronic noise, pulse-to-pulse direc-tional shifts of the laser beams, and changes in beamquality can give rise to a decorrelation effect comparableto the turbulence-induced correlation for lidar returnsfrom a diffuse target. This is also the case in our hy-drazine experiments, where we have obtained temporalcorrelation values of <0.5 from the lidar signal returnsfrom lasers 1 and 2 on a pulse-to-pulse basis. In view

_ LIMITED \ -100 -

: \ / ~~~~~~~~~LIMITED10

001 005 01 05 1 5 10RANGE (km)

Fig. 8. Minimum detectable path-averaged hydrazine concentrationby topographic reflection as a function of range. The normalizationpoint shown is taken from the experimental results and corresponds

to AP/P = 0.05 at a range of 2.7 km.

to2

I

.5E

RANGE (km)

Fig. 9. Minimum detectable path-averaged UDMH concentrationby topographic reflection as a function of range.

looc t ' E "' lll ,,] lll MMH

F AP/P)

100 LIMITED

100

q6 LIMITED

.E ~ ~ ~ ~ ~ ~ ~ ~~~E

0.01 0.05 0.1 0.5 1 5 10

RANGE (km)

Fig. 10. Minimum detectable path-averaged MMH concentrationby topographic reflection as a function of range.

of the above decorrelating processes the effects of at-mospheric turbulence can be reduced but not elimi-nated.

The range dependence of atmospheric turbulenceeffects is quite strong. According to the Rytov formu-la,23 atmospheric turbulence leads to a value for thelog-amplitude variance -2, which is strongly range de-pendent, as

or = 0.124C' k7/6 R"16, (4)


where C' is the refractive-index structure parameter,k is the wave vector 27r/X, and the numerical factor isdependent on the nature of the wave source. For alognormal distribution the measured normalized vari-ance (2of the lidar returns is related to 2 according tothe equation (I = exp(4o4 - 1. For the moderateturbulence levels indicated by the measured variancevalues in our experiment, Us << 1, the range dependenceof (I to the lowest order will be the same as that of a;that is, a(4 R11/6. For these conditions, if the mini-mum range-dependent contribution of (APr/Pr) in Eq.(3) is taken as due totally to atmospheric fluctuationsand hence proportional to oJ, the contribution to nminhas the form

R 11/12 R-1/1 2

nlmin~ c.~ (5)(AOjR Au'

Equation (5) can be expected to overemphasize theeffect of atmospheric turbulence on the range depen-dence of statistical deviations in our measurementssince it ignores reductions due to saturation and aper-ture-averaging effects. However, pending a direct ex-perimental determination of the range dependence ofAPr/Pr, a reasonable first approximation assumes thatnmin is simply the sum of the range-independent APr/Prterm of the form of Eq. (3) and a range-dependentAPr/Pr term of the form of Eq. (5). Then

= 1 5 X 103 (APr/Pr)min + CR-1/12 (6)

where C is a constant to be determined experimentally,and as noted previously (APr/Pr)min is the minimumvalue of (APr/Pr), independent of range and taken equalto 0.01. To evaluate C the value for nmin in Eq. (3) for(APr/Pr) = 0.05 and R = 2.7 km is taken as a normali-zation point on the basis of the experimental measure-ments given in Sec. V. Using this value for nmin in Eq.(6) one obtains C = 80.5. This value was then used.inEq. (6) to generate the curves labeled Eq. (6) in Figs. 8,9, and 10. It is seen that at the shortest ranges nminasymptotically approaches the ultimate system sensi-tivity, which is assumed equal to 1% in our calculations.As the range increases the contribution of the secondterm on the right-hand side of Eq. (6), which is almostindependent of range, assumes greater relative impor-tance.

The resultant predictions shown in Figs. 8,9, and 10indicate that path-averaged concentrations of the orderof 35,45, and 70 ppb over a 2.7-km range should be ob-servable for hydrazine, UDMH, and MMH, respec-tively, in the conditions of our experiment.

VII. Analysis of Effects of Atmospheric Fluctuationsand Signal Averaging on DIAL Sensitivity

The experimental and theoretical results presentedin Secs. V and VI indicate that fluctuations due to at-mospheric effects play a dominant role in limiting thesensitivity of our DIAL system for the remote sensingof the hydrazine compounds at intermediate ranges (0.5< R < 3 km). The results in Figs. 5(b), 6(b), and 7(b)further indicated that increasing the number of pulsesaveraged above a few hundred yielded relatively small

reduction in the magnitude of the fluctuations, whichindicates that these fluctuations are due primarily toatmospheric variations occurring over the time requiredto obtain the additional pulse returns.

In view of the importance of these atmospheric fluc-tuations in limiting the sensitivity of our measurements,a study was made to measure directly the improvementachievable by averaging over increasing numbers ofpulses and to establish the role of long- vs short-termatmospheric fluctuations in setting the limits of mea-surement accuracy.

To accomplish this laser 1 was fired on the P(28) line,and 50 gsec later laser 2 was fired on the P(22) line, aswas the case for our hydrazine absorption measure-ments. The target remained the flame-sprayed alu-minum plate at R = 2.7 km. However, the large tankwas removed from the laser beam path; therefore, onlyatmospheric absorption was involved in the measure-ments.

A total of 22,528 normalized lidar return pulses fromeach of the dual lasers was recorded for later statisticalanalysis. The entire process, including the time re-quired to print out a preliminary real-time analysis afterevery 2048 pulses, took 40 min, corresponding to anoverall pulse repetition rate of slightly under 10 Hz.The real-time analysis also established that the pulse-pair temporal correlations of the normalized lidar re-turns from the corresponding pulses of lasers 1 to 2 was-0.3. Due to computational constraints, analysis waslimited to blocks of 12,288 pulses from each laser. Thefollowing discussion will analyze the data from both thefirst block of 12,288 pulses and the final block of 12,288pulses, as they exhibited somewhat differing be-havior.

The analyses included determination of the averagevalues and the full block of pulses and of smaller sec-tions of that block and a determination of the normal-ized standard deviation of the lidar returns as afunction of N, the number of pulses being averaged ineach section.

The standard deviation was measured as a functionof N using a segmental-averaging approach. As anexample, for N = 8 the segments averaged would bepulses 1-8, 9-16, 17-24, ... , 12281-12288, and themeasured standard deviation is based on the resultant12,288/N average values. This was the method used inour real-time determinations of standard deviation.The results obtained for the block containing the first12,288 pulses from both laser 1, L1 , and laser 2, L2 , aregiven in Table II, which shows the percentage standarddeviation of the normalized lidar returns as N is in-creased from 1 to 2048 by factors of 2. Table II alsoshows the standard deviation using the ratio of the re-turns from the two lasers on a pulse-pair basis. Thereis no observable reduction in the standard deviationusing the ratios, which is consistent with the fact thatthe measured temporal correlation coefficients betweenthe individual laser pulses were <0.5.4,24

The standard deviation of the lidar returns obtainedfor N = 2048 was fully an order of magnitude greaterthan the standard deviation values of 0.22 and 0.25%


Table II. Percentage Standard Deviation of Normalized Lidar Returns forthe Initial Block of Pulses for Laser 1, Laser 2, and the Ratio of Returns vs

Number of Pulses Averaged

Standard deviation (%)Laser 1 Laser 2 Ratio

N P(28) P(22) L1/L2

1 20.5 17.7 22.32 16.7 14.6 17.04 13.8 12.2 13.38 11.3 10.1 10.8

16 9.0 8.3 8.732 7.3 7.0 7.664 6.0 6.1 6.6

128 4.8 5.2 5.8256 3.8 4.5 4.9512 3.0 3.9 4.0

1024 2.7 3.3 3.32048 2.2 2.8 2.4

F

0

0

0IC

4)

U,z0:

W(12W a:

4<< 0

'W

N-

0z

5 _ a

5 lo 50 100 500 1000

N (Number of pulses averaged )

TIME (min)

Fig. 11. Statistical analysis of segmentally averaged initial set of12,288 normalized pulses from lasers 1 and 2: (a) standard deviationof the segment averages as a function of N, the number of pulses av-eraged per segment; (b) variation of the average value of the individual

segments with time for N = 512.

obtained from the normalized laser 1 and 2 pulses, re-spectively, after passage through the laboratory cellcontaining air at atmospheric pressure. The standarddeviation vs N was also calculated using a running av-erage approach. Very little difference between the twoaveraging methods was observed for all values of N, withthe results essentially identical for N < 256.

The variation of the segmental standard deviationsof the returns from lasers 1 and 2 as a function of N isshown graphically in Fig. 11(a). It is seen that the slopeis significantly smaller than the predicted N-112 de-pendence. This is presumably because the atmosphericabsorption being measured is varying significantly overtime periods of the order of or shorter than the mea-surement period. These changes in atmospheric ab-sorption also provide a lower limit to the standard de-viation achievable for large values of N.

0e

0

'0

cc

Z

Table Ill. Percentage Standard Deviation of Normalized Lidar Returns forthe Final Block of Pulses for Laser 1, Laser 2, and the Ratio of Returns vs

Number of Pulses Averaged

Standard deviation (%)Laser 1 Laser 2 Ratio

N P(28) P(22) L1/L2

1 21.7 18.3 22.22 18.0 15.2 16.84 15.1 12.7 12.88 12.8 10.7 10.0

16 10.8 9.1 7.932 9.5 7.8 6.464 8.5 7.1 5.3

128 7.9 6.6 4.5256 7.5 6.0 3.9512 7.3 5.8 3.4

1024 7.0 5.2 2.82048 6.8 4.8 2.4

(a)

LASER I

10

10 -B L A SrLAS R2 -

N-1/2

L/L2

I 5 10 50 100 500 1000

N (Number of pulses averaged}

112 STATE TTT | I ~~~-T|| ,1.12

I.08

104 ~~~~~~~LASER 2/1' 1.04-

100

096 _ aft 2 '| N-~~~~~~~N512

0.92

O E I ,

TIM E (mi )

Fig. 12. Statistical analysis of segmentally averaged final set of12,288 normalized pulses from lasers 1 and 2: (a) standard deviationof the segment averages of each of the lasers and of their ratio as afunction of N, the number of pulses averaged per segment; (b) vari-ation of the average value of the individual segments with time for

N = 512.

To illustrate the extent of the variation in averagevalue from segment to segment, the average value(normalized to unity over the complete block) is shownin Fig. 11(b) for the twenty-four N = 512 pulse seg-ments. Each segment represents the average value overa time interval of slightly under 1 min. From Table IIthe results correspond to standard deviations of -3.0and 3.9% for the laser 1 and 2 returns, respectively.Individual deviations from the mean as great as 10% areobserved, but no long-term and large departures fromthe mean are observed.

The above results are in marked contrast with resultsobtained with the block of data containing the final12,288 pulses from both lasers. The results from thelatter block are shown in Table III and Fig. 12. Figure


u

Za:

W

a:

aN �-1

2o: (

2

MW (1:ZO

12(a) shows that the variation of the standard deviationwith N is again much slower than N-112 for low N andbecomes almost flat for N > 256. The results are ingeneral accord with the results of the initial data block.However, the standard deviation values for N = 2048of almost 7 and 5% for lasers 1 and 2, respectively, aresignificantly higher than those obtained from the firstblock. The reason for the increase is apparent afterconsidering the normalized segmental-average valuesfor this block as shown in Fig. 12(b) for N = 512. It isseen that there is a slow but almost constant increasein the average value of the normalized lidar returnsegments from both lasers throughout the period en-compassed by this block, corresponding to a decreasein atmospheric absorption over the 2.7-km interval atboth the P(28) and P(22) CO2 laser frequencies. Thatthis is truly an atmospheric effect rather than an ex-perimental artifact is indicated by the standard devia-tion values of 0.38 and 0.34% obtained from the corre-sponding normalized pulses from lasers 1 and 2, re-spectively, in our simultaneous laboratory measurementthrough the air-filled cell.

A particularly important feature of the results shownin Table III and Fig. 12(a) is that the standard deviationof the ratios of the pulse-pair laser returns, 0 L1/L2 as afunction of N, closely resembles the values of theequivalent set as given in Table II. This is in spite ofthe difference in the values for the individual lasers aL1

and 0L2 and despite the small value (<0.5) of the tem-

poral correlation coefficient as measured on a pulse-to-pulse basis.

The reduction in the standard deviation values usingthe ratio L 1/L2 relative to the values for the individuallasers occurs because the observed decrease in the at-mospheric background absorption is slow comparedwith the measurement time of the individual segments.In these circumstances one can consider the temporalcorrelation of the segmental measurements relative tothe average value of the full set. This is illustrated forthe case N = 512 in Fig. 12(b), where the measurementtime was slightly under 1 min. For the values shown inFig. 12(b) one obtains Ps = 0.9, where ps is the correla-tion coefficient of the corresponding segments of thereturns from lasers 1 and 2. Using this value and thevalues of aL1 and 0

L2 given in Table III for N = 512 inthe equation24

U2 L1/L2 = c2L1 + 2 L2 - 2PSCL1L2 (7)

leads to the calculated value 0L1/L2 (calc) = 3.25%,

which is in good agreement with the measured valuecL1/L2 (meas) = 3.45% as given in Table III. It is ap-parent that the use of the lidar return ratios fromequivalent segments of a dual-laser system can effec-tively eliminate much of the uncertainties due to slowdrifts in atmospheric absorption when, as in this case,the background absorption of the atmosphere is similarat the two laser frequencies.

The above analysis indicates that increased accuracycan sometimes be achieved using lidar return ratios evenwhen the pulse-to-pulse correlation coefficient is <0.5.In this regard a distinction between short- and long-

term atmospheric fluctuations must be made since theuse of ratios can only be effective when the measure-ment time is short relative to the temporal variation ofthe atmosphere. The short-term fluctuations occur inthe order of 5-100 msec,13 and significant improvementsrelative to these atmospheric changes occur only if thepulse-to-pulse correlation coefficient p of the normal-ized lidar returns from the closely spaced (50-gsec)dual-laser outputs is >0.5.24 For topographic targetreflection p is frequently near or below this value13; thiswas the case in our measurements, where p 0.3. Thus,the use of lidar return ratios was not effective in elimi-nating the effect of short-term variations. However,our experiments have shown that when long-term at-mospheric shifts are superimposed on short-term ef-fects, the use of dual-laser lidar return ratios can ef-fectively reduce the standard deviation to values ap-proaching those due to the rapid atmospheric fluctua-tion effects. In the conditions of our experiments asgiven in Table III, the use of ratios led to a reduction ofthe standard deviation from over 6% to -2.5% afteraveraging over the order of 2000 pulses in a 3.5-min.interval.

It should be noted that the above analysis should berelevant to the hydrazine detection results presentedin Sec. VI since the level of fluctuations observed withand without the hydrazine tank was essentially thesame. It then follows that using a dual-laser DIALsystem for the remote sensing of hydrazine compoundscan be effective in reducing measurement uncertaintiesdue to long-term atmospheric drifts.

Vill. Conclusions

In this paper we have demonstrated experimentallythe capability of a dual-CO2 laser differential-absorp-tion lidar system to detect hydrazine, UDMH, andMMH over distances approaching 3 km. In addition,the ability of the dual-laser DIAL system to follow thechanges in the concentration of these species was alsoshown.

The sensitivity with which the concentration of thesecompounds could be measured was found to be limitedby the minimum change in the differential-absorptionlidar return, APr/Pr, which could be distinguished frombackground fluctuations due to atmospheric turbulenceeffects. Therefore, APr/Pr is itself range dependent.An equation expressing this dependence was derivedand used to predict the minimum concentrations of thehydrazine molecules which our DIAL system couldmeasure as a function of range. Our results indicatethat even with conservative assumptions the presenceof hydrazine, UDMH, and MMH can be remotely de-tected using lidar returns from a topographic target witha sensitivity of the order of 100 ppb or better between0.5 and 5 km.

This work was supported by the Air Force Engi-neering and Services Center.


References

1. T. Henningsen, M. Garbuny, and R. L. Byer, Appl. Phys. Lett.24, 242 (1974).

2. D. K. Killinger, N. Menyuk, and W. E. DeFeo, Appl. Phys. Lett.36, 402 (1980).

3. E. R. Murray and J. E. van der Laan, Appl. Opt. 17, 814(1978).

4. D. K. Killinger and N. Menyuk, IEEE J. Quantum Electron.QE-17, 1917 (1981).

5. N. Menyuk, D. K. Killinger, and W. E. DeFeo, Appl. Opt. 19,3282(1980).

6. R. A. Baumgartner and R. L. Byer, Opt. Lett. 2, 163 (1978).7. K. Asai, T. Itabe, and T. Igarashi, Appl. Phys. Lett. 35, 60

(1979).8. E. R. Murray, Opt. Eng. 17, 30 (1978).9. H. W. Schiessl, "Hydrazine and Its Derivatives," in Kirk-Othmer:

Encyclopedia of Chemical Technology, Vol. 12 (Wiley, NewYork, 1980).

10. G. L. Loper, A. R. Calloway, M. A. Stamps, and J. A. Gelbwachs,Appl. Opt. 19, 2726 (1980).

11. J. N. Pitts, Jr. et al., "Atmospheric Chemistry of Hydrazines: GasPhase Kinetics and Mechanistic Studies," ESL-TR-80-39 (Aug.1980).

12. D. A. Stone, "The Autoxidation of Hydrazine Vapor," ReportCEEDO-TR-78-17 (Jan. 1978), and "The Autoxidation of Mo-nomethylhydrazine Vapor," CEEDO-TR-79-10 (Apr. 1979).

13. N. Menyuk and D. K. Killinger, Opt. Lett. 6, 301 (1981).

14. N. Menyuk and P. F. Moulton, Rev. Sci. Instrum. 51, 216(1980).

15. R. A. McClatchey, R. W. Fenn, J. E. A. Selby, F. E. Volz, and J.S. Garing, "Optical Properties of the Atmosphere (Third Edi-tion), " Report AFCRL-72-0497, Environmental Research PaperNo. 411 (1972).

16. R. J. Brewer and C. W. Bruce, Appl. Opt. 17, 3746 (1978).17. R. R. Patty, G. M. Russwurm, W. A. McClenny, and D. R. Mor-

gan, Appl. Opt. 13, 2850 (1974).18. S. T. Hong and A. Ishimaru, Radio Sci. 11, 551 (1976).19. A. G. Kjelaas, P. E. Nordal, and A. Bjerkestrand, Appl. Opt. 17,

277 (1978).20. G. L. Loper, "Gas Phase Kinetic Study of Air Oxidation of

UDMH," in Proceedings, Conference on EnvironmentalChemistry of Hydrazine Fuels, Tyndall Air Force Base,CEEDO-TR-78-14 (13 Sept. 1977), p. 129.

21. R. A. Saunders and J. T. Larkins, "Detection and Monitoring ofHydrazine, Monomethylhydrazine and Their DecompositionProducts," Naval Research Laboratory Memorandum 3313/AD-A027966 (1976).

22. To avoid confusion the large container used for remote sensingwill be referred to as a tank. The term cell will be reserved forthe smaller Pyrex container used in the laboratory absorptionmeasurements.

23. R. E. Hufnagel, "Propagation Through Atmospheric Turbulence,"in The Infrared Handbook, W. L. Wolfe and G. J. Zissis, Eds.(Office of Naval Research, Washington, D.C., 1978), Chap. 6.

24. D. K. Killinger and N. Menyuk, Appl. Phys. Lett. 38, 968(1981).

Meetings Schedule

OPTICAL SOCIETY OF AMERICA

1816 Jefferson Place N.W.

Washington, D.C. 20036

16-18 June 1982 PICOSECOND PHENOMENA, THIRD TOPI-CAL MEETING, Garmish-Partenkirchen Information:Meetings Department at OSA or CIRCLE NO. 45 ONREADER SERVICE CARD

22-25 June 1982 QUANTUM ELECTRONICS, TWELFTH IN-TERNATIONAL CONFERENCE, Munich Information:Meetings Department at OSA or CIRCLE NO. 51 ONREADER SERVICE CARD

18-22 October 1982 ANNUAL MEETING OPTICAL SOCIETYOF AMERICA, Tucson, Ariz. Information: Meetings De-partment at OSA or CIRCLE NO. 58 ON READER SERVICECARD

13-15 December 1982 OPTICAL FABRICATION ANDTESTING WORKSHOP, Palo Alto Information: MeetingsDepartment at OSA or CIRCLE NO. 49 ON READER SER-VICE CARD

10-12 January 1983 EXCIMER LASERS TOPICAL MEETING,Lake Tahoe Information: Meetings Department at OSAor CIRCLE NO. 50 ON READER SERVICE CARD

10-12 January 1983 OPTICAL TECHNIQUES FOR REMOTEPROBING OF THE ATMOSPHERE TOPICAL MEETING, LakeTahoe Information: Meetings Department at OSA orCIRCLE NO. 47 ON READER SERVICE CARD

12-14 January 1983 SIGNAL RECOVERY AND SYNTHESISWITH INCOMPLETE DATA OR PARTIAL CONSTRAINTSTOPICAL MEETING, Lake Tahoe Information: MeetingsDepartment at OSA or CIRCLE NO. 48 ON READER SER-VICE CARD

12-14 January 1983 METEOROLOGICAL OPTICS TOPICALMEETING, Lake Tahoe Information: Meetings Departmentat OSA or CIRCLE NO. 57 ON READER SERVICE CARD

17-20 January 1983 OPTICAL STORAGE OF DIGITAL DATATOPICAL MEETING, Lake Tahoe Information: MeetingsDepartment at OSA or CIRCLE NO. 59 ON READER SER-VICE CARD

28 February-2 March 1983 OPTICAL FIBER COMMUNICA-TION, SIXTH TOPICAL MEETING, New Orleans Informa-tion: Meetings Department at OSA or CIRCLE NO. 46 ONREADER SERVICE CARD

17-19 May 1983 CLEO 83, CONFERENCE ON LASERS ANDELECTROOPTICS, Baltimore Information: MeetingsDepartment at OSA or CIRCLE NO. 52 ON READER SER-VICE CARD

17-21 October 1983 ANNUAL MEETING OPTICAL SOCIETYOF AMERICA, New Orleans Information: Meetings De-partment at OSA or CIRCLE NO.53 ON READER SERVICECARD

23-25 January 1984 OPTICAL FIBER COMMUNICATION,SEVENTH TOPICAL MEETING, New Orleans Information:Meetings Department at OSA or CIRCLE NO. 54 ONREADER SERVICE CARD

10-12 April 1984 CLEO 84, CONFERENCE ON LASERS ANDELECTROOPTICS, Anaheim Information: Meetings De-partment at OSA or CIRCLE NO.55 ON READER SERVICECARD


laser remote sensing of hydrazine, mmh, and udmh using a differential-absorption co_2 lidar

Documents