JOURNAL OF THE OPTICAL SOCIETY OF AMERICA VLM 3 UBR3MRH17

Amplified laser absorption: detection of nitric oxide

C. Chackerian, Jr. and M. F. Weisbach*NASA Ames Research Center, Moffett Field, California 94035

(Received 7 September 1972)

Experimental results are reported for the power loss of a carbon monoxide gas laser due to the absorptionof small amounts of nitric oxide placed in an intra-laser-cavity absorption cell. It was found, for theparticular experimental conditions employed, that the absorption coefficient of nitric oxide at 1900.04 cm-,is enhanced more than two orders of magnitude when it is in the intracavity cell. Finally, the experimentalresults are compared with theoretical calculations.

Index Headings: Laser; Absorption.

Recently the first reports of the enhancement of ab-sorption spectra of certain atomic and molecular speciesinserted in dye-laser cavities have appeared.'-' We havestudied the same kind of phenomenon with a carbonmonoxide gas laser and nitric oxide, namely, the ab-sorption enhancement of certain nitric oxide vibration-rotation lines. This is possible because there are anumber of coincidences between nitric oxide vibration-rotation absorption lines and carbon monoxide gas laserlines. We report here laser power-loss measurements as afunction of intracavity nitric oxide concentration, andquantitative comparison between theory and experi-ment using gain-saturation theory. Two nitric oxide-carbon monoxide line pairs were used in this study andthe pertinent spectroscopic information is summarizedin Table I. Calculations showed that the line pair near1900, cm-' would give the greatest sensitivity in thepower-loss measurements. However, measurements werealso made for the line pair near 1935.5 cm'7 in order toprovide a further check on the theoretical calculation ofpower loss.

EXPERIMENTAL

The experimental setup is shown in Fig. 1. The lasercavity was formed by a 10-m radius-of-curvature 10%transmitting mirror and an original grating (blazed for5.4 Mtm). The laser had an interelectrode separation of

TABLIE I. NO absorption coefficients at CO laser wavelengths.

NO line CO line NO linePosition" Position Strengthba (cm-') a' (cm-') (cm-' atm-'l) ae (cm-' atm-') C

1900.0795 1900.043d 2.99 10.0-40.7mn= 15/2, 2J1,/2 V== 9--+ 8

J= 8 -- 91935.4916 1935.4834, 0.29 4.7-0.3mn= 39/2, 2113 /2 V= 7 6

J=l12- 13

a From Ref. 4.b From Ref. 7.Vaa Voigt profile for NO with -y (He-NO, rn= 15/2) = 0.044c -, y (He-NO, in =39/2) = 0.04 cm-' and total pressure equal

0.38 atm; A-type doubling ignored.d From Ref. 5.0 From Ref. 6.

119 cm and was immersed in a liquid nitrogen bath over115 cm of its length. The laser was run in the cw mode,nominally at 10 kV and 15 mA, and the laser as well asthe 15-cm absorption cell were fitted with calciumfluoride Brewster-angle windows. About 10% of thelaser output beam was directed through a chopper onto asanded piece of calcium fluoride placed at the entranceslit of a Perkin-Elmer model No. E-14 spectrometer.The signal was detected with a liquid-nitrogen-cooledcopper-doped germanium detector, amplified and de-modulated with a PAR lock-in amplifier, and displayedwith a strip-chart recorder. The nominal cw outputpower of the laser was 0.1-0.15 W.

The gas-handling vacuum manifold (not shown inFig. 1) was connected directly to the intracavity ab-sorption cell via a Pyrex transfer line. The nitric oxidesamples were purified by fractional distillation. Heliumwas used to pressure broaden the nitric oxide absorptionlines. Originally, we tried nitrogen as the pressure-broadening agent but an unknown impurity in thenitrogen caused some of the NO to be converted to NO 2 ,as evidenced by a brown color in the gas-storage bulb.For an experimental run, laser power-loss measurementswere made on successive NO-He mixtures of increasingdilution. Each successive mixture was prepared (af terthe laser power-loss measurement) by expanding a smallpart of the preceding sample and then diluting withhelium to about 0.4 atm total pressure. (We werelimited to a total pressure of about 0.45 atm by themercury manometer in our vacuum system.) Each gassample remained in the absorption cell about 5 min;

C MD

FIG. 1. Schematic of apparatus. G, grating; D, diaphragmaperture; L, laser tube; A, absorption cell; V, to vacuum manifold;M, 10% transmissive 10-in radius-of-curvature mirror; B, calciumfluoride plate; C, chopper; SP, sanded piece of calcium fluoride;S, Perkin-Elmer model No. E-14 spectrometer; I, copper-dopedgermanium photoconductive detector.

342

VOLUME 63, NUMBER 3 MARCH 1973

March 1973 AMPLIFIED LASER ABSORPTION:

during that time the production of thermal gradients(and therefore density gradients) due to heating by theabsorption of laser power could be neglected. Thus,thermal-lens effects could also be neglected.

DATA REDUCTION

An example of the experimental data is shown inFig. 2, which is a record of laser output power vs time.The periodic variation of laser power is caused byexpansion of the laser cavity I-beam mount' (thesame amplitude in the power-tuning curve could beobtained by translating one of the mirrors along thelaser axis). At point (a) the evacuated intracavityabsorption cell was filled with 290 torr of pure helium;the laser output increased because the cavity wasbrought into better alignment (the laser output hadpreviously been optimized with the absorption cellfilled with 290 torr of pure helium). Introducing heliuminto the evacuated absorption cell also causes theoptical length of the laser cavity to increase, and theeffect on the laser output is the same as that producedby expansion of the I beam, except that it occurs in thetime required to fill the absorption cell. The number ofcycles of the power-tuning curve due to filling theabsorption cell is

-} = 2PXo-(n -1) '(1

where X is the length of the intracavity absorption cell(cm), n and P, respectively, the index of refraction (atatmospheric pressure) and pressure (atm) of the gas inthe cell, and o- the wave number (cm-') of the laserline. Gas was introduced into the absorption cell at a

100 r-

90 H

80 h

70

60 F-

50

40d

104

E 10316

(02

x

Ld 102

(I)(I)

a-1

10 lo

10°0 .5

I/Io-1.0

FIG. 3. Intracavity nitric oxide partial pressure vs normalizedlaser output for the nitric oxide line at 1935.48 cm-'. (i Experi-ment. Theory: L= 115 cm, a= 0.2, A= 0.093; go= 0.0026 cm -l-,go=0.0017 cm-l---

point on the power-tuning curve such that the laseroutput power would advance to the subsequent maxi-mum of the power-tuning curve. Use of Eq. (1) showsthat the contribution to the phase shift on the power-tuning curve from nitric oxide in the nitric oxide-helium mixtures can be neglected. Continuing now withFig. 2: The absorption cell was evacuated (b), refilledwith the nitric oxide-helium mixture (c), and thenevacuated (d).

For comparison with theory, the ratio of the attenu-ated laser output (NO-He mixture in cell) to un-attenuated laser output (pure He in the cell), I/lo, wasdetermined as (Dd/D,)- (Da/Db), where the Di aretaken as the appropriate chart deflections.

COMPARISON BETWEEN THEORY ANDEXPERIMENT

An experimental run consisted of measuring I/Io(defined in the previous section) for a succession ofNO-He mixtures of increasing NO dilution. The ex-perimental values of I/1o are plotted in Figs. 3 and 4as a function of NO partial pressure. The theoreticalcurves in Figs. 3 and 4 were obtained using an expressionfor the saturated-gain coefficient derived by Smith,9

ag(O) =goL-W(iy),

y

*- TIME

FIG. 2. Strip-chart record of typical data (1900.04-cm-' linepair). a, pure helium introduced into absorption cell; b, absorptioncell evacuated; c, helium-nitric oxide mixture (290 torr, 0.044%nitric oxide) introduced into absorption cell; d, absorption cellevacuated.

(2)

where g(O) is the saturated-gain coefficient at linecenter, go the small-signal-gain coefficient per unitlength, L the active laser length, and a the Voigtparameter [the ratio of the pressure and Doppler-broadened line half-widths multiplied by (In2)f]. W(iy)is defined by

W(iy) =exp(y 2)[1+erf (-y)]

DETECTION OF NO 343

r

C. CHACKERIAN, JR. AND M. F. WEISBACH

104

E103 H

(0 0 EOoC

L102 C

lIt 0ZC)lo,- 0

10°0 .5 1.0

I/IoFIG. 4. Intracavity nitric oxide partial pressure vs normalized

laser output for the nitric oxide line at 1900.08 cm-'. Experiment:low-gain run, tailed circles with dots; high-gain runs, circles withdots, squares with dots, triangles with dots. Theory: -, L = 115cm, a= 0.2, A = 0.0924, go = 0.00 26 cm-'.

(see Abramowitz and Stegun'0). The parameter y isdefined by

y=a(1+2w/wo) 1 , (3)

where w is the irradiance in the cavity and wo thesaturation parameter, which is a constant. 9 Equation (2)assumes that the laser is operating on a single transversemode located at the central frequency of the transition."

The laser output (I/Io) as a function of intracavityloss due to NO was calculated in the following manner.First, it was assumed that the saturated gain is equal tohalf the round-trip cavity losses,

g(0) =0+A =F(a,go,w/wo), (4)

where 0 represents half the sum of the grating reflectionand mirror transmission losses (0.092), and A the single-pass loss due to nitric oxide in the intracavity absorptioncell. Then the normalized laser output was obtainedfrom

I(AF 50) w(A#D 0)/wo

I(A=0) w(A=0)/wo

The quantities in the right-hand side of Eq. (5) werecalculated from the solutions of Eq. (4): The sum 0+Awas calculated for a series of values for w/wo with thevalue of w (A=O)/wo determined by the condition A=0and 0 = 0.092. It should be noted that the actual value ofwo is never needed, because it cancels out in Eq. (5). TheNO pressures that correspond to particular values of Awere calculated from

1 1p=-In A (6)

aX 1-A

where p is the NO partial pressure in atm, a the ap-propriate absorption coefficient taken from Table I, andX the intracavity absorption-cell length. To obtain acrude estimate of the gain of the transitions in question,for our laser operating conditions, we have assumed thatgain is proportional to laser efficiency. Bhaumik et al.'2

have measured small-signal-gain coefficients for the1900- and 1935-cm-' carbon monoxide laser lines (goequals 0.008 and 0.005 cm-', respectively). Because ourlaser was about 10% efficient and that used by Bhaumikabout 30% efficient, we estimate values for the small-signal gains in our laser, for the two transitions inquestion, to be 0.0026 and 0.0017 cm-', respectively. Ina practical situation, the laser-power loss should becalibrated with known samples of nitric oxide, becauseit is difficult to know a priori the laser-gain parameter.

The agreement between theory and experiment forthe 1935-cm-' line pair shown in Fig. 3 is excellent,especially considering that the parameters used in thecalculations were not chosen to give best agreementbetween theory and experiment but were obtained fromindependent considerations. Theoretical curves shownfor two values of the gain parameter indicate enhancedsensitivity at the lower value for go. The comparisonbetween theory and experiment for the 1900-cm-' linepair is shown in Fig. 4. The theoretical curve is about50% higher than the experimental over most of therange of the data (I/Io values greater than 0.3).

The theoretical calculations indicate that the ordinatein Figs. 3 and 4 is proportional to go. Therefore, thesensitivity of the laser spoiling to intracavity nitricoxide can be enhanced by running the laser at thelowest possible gain. The data for Figs. 3 and 4 wereobtained with the laser running at a pressure of 5 torr.In a preliminary experiment, the CO laser was operatedat 3.4 torr (at the lower pressure, presumably, the lasergain is lower) and the curve labeled low gain in Fig. 4was obtained.

DISCUSSION

An absorption coefficient for the 1900-cm-' CO-NOline pair can be calculated from the data in Fig. 4. If thelast low-gain data point (4.5 ppm, 17% absorption) isused in Beer's absorption law, an absorption coefficientof 2515 cm-' atm-' is obtained. The absorption in thiscase has therefore been amplified by 250 times thatshown in Table I.

Because these measurements were made with helium-nitric oxide mixtures (total pressure of 380 torr), theyshould be repeated with pure nitrogen at total pressuresof 760 torr. In the latter case, the sensitivity of the laseroutput to nitric oxide should be further enhanced be-cause of greater pressure broadening of NO lines due tonitrogen at 1 atm. Also, the increased sensitivity atlower gains should be investigated further. We estimatethat nitric oxide concentrations on the order of 100 partsper billion can be easily determined with an intracavitylength of 50 cm with the gas pressure broadened to

344 Vol. 63

March 1973 AMPLIFIED LASER ABSORPTION: DETECTION OF NO

1 atm of nitrogen and with the laser operating nearthreshold. It is easy to show [using Eq. (3) in Smith's 9

paper] that the sensitivity of the power loss, as ex-pressed in Eq. (5), to A becomes infinite as goL ap-proaches O+A. For the present experiments goL-0.3and O+A1z0.1. Perhaps the only practical limit on theattainable sensitivity will be due to the laser's outputstability when it is operated very near threshold.

The technique of using intracavity absorption for thedetection of small amounts of nitric oxide appears to becompetitive with two more-elaborate laser techniquesthat use carbon monoxide lasers: namely, one thatuses a spin-flip Raman laser with opto-acoustic detec-tion, as described by Kreuzer and Patel,"3 and anotherthat employs Zeeman modulation of the nitric oxideabsorption; the latter method is described by Kaldorand co-workers.'4 Another device, the solid-state tunable-laser diodes,"5 may be tuned to the center of an absorp-tion line. The calculated absorption coefficient at thecenter of the m = 15/2, 2II1/2 nitric oxide line is about70 cm-' atm-', which is considerably less than theamplified absorption coefficient of 2515 cm-' atm-';however, based on data from these experiments, thetunable-diode absorption should become competitivewith intracavity absorption for air path lengths ap-proximately 36 times as long as the intracavity cell.

CONCLUSIONS

A new experimental technique for the detection ofnitric oxide in the PPM range has been demonstratedexperimentally. In theory, the method is capable of

greater sensitivity; further experimental work shouldestablish the practical limits of the technique.

REFERENCES

*NRC Postdoctoral Fellow.'Robert J. Thrash, H. von Weyssenhoff, and James S. Shirk, J.

Chem. Phys. 55, 4659 (1971).2N. C. Peterson, M. J. Kurylo, W. Braun, A. M. Bass, and R.

A. Keller, J. Opt. Soc. Am. 61, 746 (1971).3R. A. Keller, E. F. Zalewski, and N. C. Peterson, J. Opt. Soc.

Am. 62, 319 (1972).4D. B. Keck, Ph.D. thesis, Michigan State University (1967)

(University Microfilms, Inc., Ann Arbor, Mich., ordernumber 68-7914).

5A. W. Mantz, James K. G. Watson, K. Narahari Rao, D. L.Albritton, A. L. Schmeltekopf, and R. N. Zare, J. Mol.Spectrosc. 39, 180 (1971).

6D. R. Sokoloff, A. Sanchez, R. M. Osgood, and A. Javan,Appl. Phys. Lett. 17, 257 (1970).

7L. L. Abels and J. H. Shaw, J. Mol. Spectrosc. 20, 11 (1966).'Since the laser was originally designed for millisecond-duration

shock-wave experiments, a piezoelectric stabilizer was notbuilt into it.

9P. W. Smith, IEEE J. Quantum Electron. 2, 62 (1966)."Handbook of Mathematical Functions, edited by M,

Abramowitz and I. S. Stegun, Appl. Math. Ser. 55 (U. S.Government Printing Office, Washington, D. C., 1964;Dover, New York, 1965).

"With the laser operating at 5-torr pressure, we estimate that90% of the laser power is contained in the central transversemode.

"2Mani L. Bhaumik, W. B. Lacina, and Michael M. Mann,IEEE J. Quantum Electron. 8, 150 (1972).

13L. B. Kreuzer and C. K. N. Patel, Science 173, 45 (1971)."4A. Kaldor, W. B. Olson, and A. G. Maki, Science 176, 508

(1972).15E. D. Hinkley and P. L. Kelly, Science 171, 635 (1971).

COPYRIGHT AND PERMISSION

This Journal is fully copyrighted, for the protection of the authors andtheir sponsors. Permission is hereby granted to any other authors to quotefrom this journal, provided that they make acknowledgment, includingthe authors' names, the Journal name, volume, page, and year. Reproduc-tion of figures and tables is likewise permitted in other articles and books,provided that the same information is printed with them. The best andmost economical way for the author and his sponsor to obtain copies is toorder the full number of reprints needed, at the time the article is printed,before the type is destroyed. However, the author, his organization, orhis government sponsor are hereby granted permission to reproduce partor all of his material. Other reproduction of this Journal in whole or inpart, or copying in commercially published books, periodicals, or leafletsrequires permission of the editor.

345