Repetitively Pulsed, Tunable Ruby Laser withSolid Etalon Mode Control

W. B. Tiffany

A repetitively pulsed, tunable ruby laser has been designed and constructed as a light source for high reso-lution absorption spectroscopy and selective photochemical excitation. The frequency is tuned over arange of approximately 20 cm-' by changing the temperature of the ruby. The use of a solid sapphireetalon for mode control results in linewidth and stability within 0.04 cm-' over durations exceeding 25,000consecutive pulses. Repetition rates of up to 2 pulses/sec at output energies of 0.5 J/pulse are main-tained. Design and performance features of the laser are presented, and examples of its application areshown.

I. IntroductionWhile many proposed practical applications of laser

light depend upon monochromaticity and tunability,most lasers are unfortunately neither monochromaticnor tunable.' Many techniques have been developedto provide these features, and have been used to alimited extent with both pulsed and continuous lasers.For example, the use of multiple reflecting surfaces haspermitted the single mode operation of pulsed lasersfor stimulated scattering applications 2 and holography.3Also, the temperature shift of the ruby fluorescencewavelength has been used for thermally tuning theruby laser.4

This communication describes a ruby laser whichcombines both mode control and temperature tuning,along with repetitive pulsing and reliable long termfrequency stability. The design of this device wasdictated by the application for which it was intended.The laser light was to be used to probe the fine struc-ture of bromine molecular electronic transitions in the14,400-cm-' region,5 in order to determine the natureand dynamics of a gaseous photochemical reaction.The close spacing of the individual lines of this finestructure necessitated mode control and frequencystability. The slow rate of the reaction required longterm operation at a reasonably high pulse repetitionrate. The results of this selective photocatalysis experi-ment show that the laser successfully met these require-ments.

The author was with the Physics Department, Stanford Uni-versity, Stanford, California 94305, when this work was done;he is now with Sylvania Electronic Systems-WesternlDivision,Mountain View, California 94040.

Received 21 July 1967.

11. Laser Design and ConstructionA schematic diagram of the essential features of the

laser is shown in Fig. 1. The ruby rod and the flashlamp are located inside a 7.6-cm long elliptical cylinder.The laser end reflectors consist of a totally reflectingBrewster angle input roof prism (manufactured by E.and W. Optical Co., Minneapolis, Minnesota) and apartially reflecting uncoated sapphire optical flat. Theends of the rod itself are antireflection coated. Thesecomponents are fastened to an optical bench on in-dividual adjustable mounts. The entire apparatus ispurged with nitrogen gas to eliminate dust and moisturecondensation from the optical surfaces. The principalcomponents of the laser to be described are the energydischarge and control circuits, the rod cooling and fre-quency tuning system, and the mode controlling solidetalon.

A. Energy Discharge and Control Circuits

The high repetition rate is achieved mainly by usinga water-cooled flash lamp (Model XE14-C-3, PEKLabs, Sunnyvale, California), with a maximum ratedinput energy per pulse of 1000 J. At the more con-servative input energy of 600 J, some of these lampshave survived more than 70,000 pulses without failureor noticeable deterioration. The lamp is pulsed by thedischarge of two 240-,uF, 5-kV capacitors in parallelthrough a 175 MH inductor. Initial ionization of thelamp to trigger the discharge is accomplished by afaster, higher voltage pulse, which is introduced bywinding part of the main discharge circuit as the second-ary of a trigger transformer,' and discharging a I-pFcapacitor at 2 kY through two xenon thyratrons(C3J-A) in series with the primary. The thyratronsin turn are activated by applying a positive pulse to thegrid of one of them, either with a manual button or

January 1968 / Vol. 7, No. 1 / APPLIED OPTICS 67

N2 ` .. I

EXHAUST -EXHAUST - - - ___-3mm SAPPHIRE

ROOF ETALON 7PRISM- - - --------- - CLAD- - __ __ _ _ -_| ARUBY ROD OVEN

r , , | ~~~~ENCLOSURE|

XENON FLASH LAMP

l PUMP CAVITY l

H20 1 ~~~~~~~~H0COOLING

Fig. 1. Schematic diagram of laser.

with a relay in a variable divider circuit which serves asa voltage discriminator. This latter method permitsthe laser to be fired repetitively at a rate determined bythe charging time of the capacitors. Another relayautomatically turns off the high voltage and dischargesthe capacitors through a bleeder resistor in case theflash lamp fails to fire and the voltage continues toclimb. A block diagram of the discharge and controlcircuits is shown in Fig. 2.

B. Temperature Tuning System

The cooling system for the ruby rod performs twofunctions. The first is simply to remove the heatwhich would otherwise build up in the rod witheach successive pulse. The second is to control thefrequency of the laser output by making use of thetemperature dependence of the ruby fluorescence wave-length.4 This shifts from a value of 6943 at 2950Kto 6934 A at 770 K. The cooling is accomplished by acontinuous stream of nitrogen gas which is bubbledthrough a 50-liter dewar of liquid nitrogen and passedover the rod through a single-walled fused silica coolingjacket. The use of a sapphire-clad ruby rod8 with a6.35-mm diam core and an outer diameter of 11.23 mmprovides additional thermal ballast and cooling surfacearea, as well as lowering the input energy at thethreshold of laser oscillation. The temperature, andhence the wavelength, can be controlled reasonably wellby adjusting the nitrogen flow rate so that the averagerod temperature at the time of firing remains constant.This method, however, is inconvenient and also cannotcompensate for slow fluctuations in the average tem-perature. The situation is greatly improved by in-creasing the flow rate slightly and placing 'a heaterconsisting of a meter or so of coiled nichrome wire in thenitrogen stream just ahead of the rod. The temperaturesensed by a copper-constantan thermocouple in closecontact with the rod is displayed on an indicatingpotentiometer (Speed-O-Max, Leeds & Northrup Co.).Geared to the indicator dial is a variable resistor form-ing part of a bridge circuit. The current from thebridge is amplified and used to operate a relay whichcontrols the heater. The sensitivity of the system issuch that the temperature difference between opening

and closing of the relay is about 10 C. A second variableresistor in the bridge circuit allows the operating tem-perature to be selected anywhere over the range of theindicator dial. The control system operates so that ifthe temperature tends to drift downward, the heaterremains on during a longer proportion of each pulsecycle to combat the drift. It should be noted that thetemperature sensed by the thermocouple is somewhatlower than the true rod temperature at the moment ofthe pulse. This is partly due to the fact that thethermocouple is in thermal contact with the coolernitrogen stream as well as with the rod, and partly to thefinite thermal conductivity of the rod and the responsetime of the instrument. This makes the indicated tem-perature a function of operating conditions such aspulse rate and energy, as well as a function of true rodtemperature. Therefore, although temperature sensingcan be used as a means for frequency control under agiven set of operating conditions, it does not providean absolute frequency calibration. The latter is accom-plished by using a scanning spectrometer, as describedin Sec. III.

C. Mode Control

Mode control is achieved by replacing the outputmirror of the laser with an uncoated sapphire opticalflat. The interference between the light reflected atthe two parallel faces of this flat leads to a frequencydependent over-all reflectivity, with maxima spaced atperiodic frequency intervals APe = 1/2L cm-', whereL is the optical thickness of the etalon in centimeters.9

The result of this periodic frequency dependent re-flectivity is to suppress the oscillation of most of theaxial modes of the laser cavity, as is illustrated in Fig.3. The use of such a technique to restrict the numberof oscillating laser modes has been discussed in theliterature. 10

The resonant frequencies of a solid etalon can beshifted readily by varying the temperature." Thisshift can be calculated in the following way. Expressedin wavenumbers, the frequencies of the etalon reflectiv-ity maxima are given by e = m/(2nd), where m is aninteger, n is the refractive index, and d is the actualthickness of the plate. In general, a change in tempera-ture can produce both a change in d proportional to thecoefficient of linear expansion, and a change in n.

Fig. 2. Block diagram of energy discharge and control circuits.

68 APPLIED OPTICS / Vol. 7, No. 1 / January 1968

FLUORESCENCE7GAIN LINEWI C/2L

,( Y \ I CAVITY

FREE-RUNNING LASER OUTPUT SPECTRUM

C/2Le

MODE-SELECTING ETALON REFLECTANCE

.,1 1'ILASER OUTPUT SPECTRUMWITH MODE SELECTION

FREQUENCY-

Fig. 3. Effect of etalon mode selector on spectral output of laser.

These effects combine to shift each resonant frequencyPe by the amount:

SPVe _ v. d ave anAT ad T b n a'T

m Ad mn An2nd' ST 2n0d aT

I d I n\ed d'1' n AT.

Since the quantity (1/d)(3d/T) is just the coefficientof linear expansion, the entire temperature effect isindependent of the thickness of the etalon.

Sapphire was chosen because of its excellent thermalproperties, and because its peak reflectivity is suffi-ciently high to provide mode discrimination over aconvenient range of pump energies. It has a refrac-tive index of 1.76 with a temperature coefficient of 1.3 X10-5/°C, and an expansion coefficient of 6.76 X 10-6/°C(see Ref. 12). From these values, a peak reflectivity of26% and a temperature shift of the etalon resonances of-0.204 cm-'/ 0C are obtained for ruby laser light at14,400 cm-'. The flat is 3 mm thick, which makes Ave,the interval between etalon resonances, equal to 0.95cm-'. This entire free spectral range is covered in atemperature range of approximately 5C. By com-bination of temperature control of the rod and theetalon, it is possible to place only one etalon resonancewithin the gain profile of the fluorescence line. As longas this condition is maintained, the etalon effectivelyreduces the spectral width of the laser output. Anadditional advantage in using the etalon is that thelaser continues to oscillate at the frequency determinedby the etalon resonance in spite of small shifts in the fre-quency of the ruby fluorescence line. This results in agreat improvement in spectral stability, since thermalstabilization of the passive etalon is inherently much less

difficult than thermal stabilization of the active rubyrod, which is continually subjected to violent pulses offlash lamp radiation.

The etalon is housed in an insulated oven enclosure,with temperature regulation provided by a proportionalcontroller (Model V1523, Reeves-Hoffman Division,Dynamics Corporation of America, Carlisle, Pennsyl-vania) with a range of 4 10'C and a published stabilitywithin 0.010C. The laser light enters and leaves theoven enclosure through low scatter fused silica Brew-ster angle windows. Since the sapphire etalon is cut atzero degrees, its reflectivity is independent of the polar-ization of the laser beam.

111. Laser Performance

A. Tuning

To measure the frequency shift of the laser output as afunction of rod temperature, a portion of the beam,suitably attenuated with reflectance filters, was ad-mitted through the entrance slit of a Jarrell-Ash 1-mEbert type high resolution scanning spectrometer.The signal from a photomultiplier at the exit slit wasamplified and displayed on a chart recorder. Thespectrum was scanned slowly while the laser was repeti-tively pulsed, with the result that the laser pulses weredetected and recorded at their appropriate wavelength.Two methods of wavelength calibration were employed.The first consisted of using the emission lines of anordinary neon standard lamp. In the second method,light from a tungsten projection lamp was passedthrough a 1-m tube of bromine gas and into the spec-trometer simultaneously with the pulses of laser light.

8

7

t6

-5E

1I4

W3a12

n,

0-80 -60 -40

TEMPERATURE (C) -

Fig. 4. Measured shift of laser frequency as a function of indi-cated rod temperature for various pulse repetition rates and input

energies.

January 1968 / Vol. 7, No. 1 / APPLIED OPTICS 69

Fig. 5. Fabry-Perot fringes on the left-hand side produced bysuperposition of 100 consecutive pulses of the ruby laser; fringeson the right produced by Spectra-Physics model 119 single fre-quency helium-neon laser to check the finesse of the analyzingetalon. Distance between adjacent fringes corresponds to 0.25

cm.-'

Thus, the absorption spectrum of bromine gas wasregistered, with the laser pulses superimposed at theappropriate wavelength as the spectrum was scanned.Bromine absorption lines thereby provided referencewavelength marks. The second method was especiallyconvenient for experiments in which the laser light wasused for selective excitation of bromine molecules.Results of runs made for indicated temperatures of-40 0 C, -60 0C, and -80"C, and for several values ofpulse repetition rate and input energy, are plotted inFig. 4. The fact that the frequency depends upon otherparameters besides the indicated temperature resultsfrom the previously mentioned time lag of the latterwith respect to the actual rod temperature.

B. Mode Control and Frequency Stability

The spectral width and stability of the laser output,as well as the thermal tuning properties of the sapphiremode selecting etalon, were measured by using a secondanalyzing Fabry-Perot etalon. The result of such ameasurement is shown in Fig. 5. The spectral rangebetween fringes is 0.25 cm-'. The left-hand side is thefringe pattern produced by the superposition of 100 con-secutive ruby laser pulses. The fringe width indicatesa spectral width and stability within about 0.05 cm-'.Fringes resulting from more than this number of pulses,or from single pulses, indicated substantially the samelinewidth, so long as the mode selecting etalon wasmaintained at a constant temperature. This stabilitypersisted even when the indicated temperature of theruby rod drifted over a 5 range. Rod temperaturechanges greater than this amount, however, causedlaser oscillation to occur at two resonances of the etalonseparated by about 1 cm-', and still larger tempera-ture changes shifted the laser frequency entirely to the

next resonance. On the other hand, the frequency ofthe laser output could be tuned smoothly by changingthe etalon temperature. It was possible to achieveoscillation at a single etalon resonance even when thelaser was pumped at energies greater than twice thres-hold.

The right-hand side of Fig. 5 shows the fringe patternfrom a highly monochromatic and stable helium-neonlaser (Spectra-Physics model No. 119), using the sameanalyzing Fabry-Perot etalon. The width of thesefringes is due almost entirely to the limited finesse of theanalyzing etalon itself. Comparison with the rubylaser fringes suggests that instrumental resolution con-tributes significantly to their width as well. Thus, theactual spectral width of the ruby laser output is some-what less than 0.04 cm-l.

From the dimensions of the laser the spacing of theindividual axial cavity modes was calculated to be about0.01 cm-'. It was not possible to decide on the basis ofthe fringe patterns alone if oscillation occurred in morethan one of these during a pulse. However, the ob-servation of spikling in the time-resolved output in-dicated that this was the case. With a conventionalmirror instead of the mode selecting etalon, the laseroscillation occurred over a width of about 0.35 cm-'.Hence, the upper limit of 0.04 cm-' obtained withmode control represents a substantial spectral narrowingof the laser output.

C. General Performance

Under typical operating conditions, the laser de-livered an output energy per pulse of approximately 0.5J with an input energy of 450 J. Repetition rates up to2 pulses/sec were maintained over durations of 104 ormore consecutive pulses without component failure.

IV. Applications

Examples are given of two applications of the tunableruby laser. The first of these is the use of the laser tomeasure the absorption in individual bromine lines.The second is the selective photocatalysis of a gas phasereaction between bromine and fluorinated butenes, asthe laser was tuned across the bromine absorption spec-trum.

RELATIVE TTABS ORPTION

I I I I I I i l I

0 0.2 0.4 0.6 0 0.2 0.4 0.6 0.8LINE A LINE B

RELATIVE FREQUENCY Icm')

Fig. 6. Optical absorption in two lines of Br2 spectrum obtainedby tuning the ruby laser. Solid curves show spectra obtained by

conventional high resolution spectrometer.

70 APPLIED OPTICS / Vol. 7, No. 1 / January 1968

_7 -

500-800 cm-' below that required for dissociation intoatoms. However, all photochemical reactions ofbromine had previously been attributed to free radicalchain processes, which require dissociated atoms fortheir initiation." In earlier experiments, all theseatoms were produced by direct dissociation of moleculesupon absorbing light in the continuum. With the useof the tunable laser, a photochemical reaction of brominewhich depends upon individual lines, and not con-tinuous absorption, was observed for the first time.The study of this phenomenon showed that the excitedbromine molecules acquire further energy from colli-sions, and dissociate into atoms which initiate the chainreaction. From the selectively excited reaction data,it was also possible to obtain information not formerlyaccessible concerning collisional energy transfer pro-cesses in gaseous bromine.'5

2 NJ U \ /VwV V. Summary- oW This paper has described a pulsed ruby laser which

0 | I X is tunable over a range of a few angstroms, highly2.0 1.5 1.0 0.5 0 monochromatic, and capable of repetitive operation for

A () long durations. As such, it is a useful instrument forhigh rsolutinn a.bsorntion sn~etroeonv and lpetive.

Fig. 7. Dependence of reaction rate on fine structure of Br2 ab-sorption spectrum in laser-induced photochemical addition of

bromine to fluorinated butenes in the gas phase.

A. Laser Absorption Spectroscopy of BromineGas

Experiments were performed in which the absorptionof laser light by a column of bromine was measured asthe laser was tuned stepwise across two of the prominentabsorption lines of natural Br2. The purpose of theseexperiments was to explore the feasibility of this tech-nique for absorption spectroscopy, as well as to checkthe linewidth and tuning accuracy of the laser. Thelaser beam was split, and one portion was monitoreddirectly with a photocell. A second photocell detectedthe other portion of the beam after it had passed threetimes through the 1-m absorption cell containingbromine. The signals from the two photocells werethen integrated using a Tektronix Type 0 operationalamplifier, displayed simultaneously on a Tektronix 555Dual Beam oscilloscope, and recorded on Polaroid film.The ratio of the heights of the two traces, normalizedto the conditions with the bromine removed, indicatedthe amount of absorption. These data are summarizedin Fig. 6, along with the profiles of the respective linesobtained by conventional spectroscopy. Agreementbetween the two methods is satisfactory.

B. Selective Photocatalysis of Bromine Reactions

The laser was used as a light source to catalyze an ad-dition reaction of bromine gas with fluorinated butenes.Enhancement of chemical reactivity was observed whenthe laser was tuned to individual lines in the bromineabsorption spectrum.6 ," This is shown in Fig. 7.Spectroscopic evidence shows that these transitionsproduce stable excited molecules, whose energy is about

photocatalysis. It is also potentially useful for studiesof selectively excited fluorescence and of collisionalenergy transfer. Results obtained with it so far haveled to the observation of new photochemical phenomena.

The author wishes to thank A. L. Schawlow for sug-gesting and supervising this research, and for criticallyreading the manuscript. Many helpful discussionswith J. L. Emmett are also acknowledged, as is partialfinancial support from Sylvania Electronic Systems.

This work was supported by a National Aeronauticsand Space Administration grant.

References1. See for example, C. L. Tang, H. Statz, and G. deMars, J.

Appl. Phys. 34, 2289 (1963); B. J. McMurtry, Appl. Opt.2, 767 (1963).

2. F. J. McClung and D. Weiner, IEEE J. Quantum Electron.QE-1, 94 (1965).

3. A. D. Jacobson and F. J. McClung, Appl. Opt. 4, 1509(1965).

4. I. D. Abella and H. Z. Cummins, J. Appl. Phys. 32, 117(1961).

5. 0. Darbyshire, Proc. Roy. Soc. London A159, 93 (1937).6. W. B. Tiffany, H. W. Moos, and A. L. Schawlow, Science

157,40 (1967).7. Design based on private communication from J. P. Markie-

wicz and J. L. Emmett; see for example, Catalog 2-64,PEK Labs, Inc., Sunnyvale, California (1964).

8. G. E. Devlin, J. McKenna, A. D. May, and A. L. Schawlow,Appl. Opt. 1, 11 (1962). (These rods are unfortunately nolonger being manufactured.)

9. See for example, M. Born and E. Wolf, Principles of Optics(Pergamon Press, Inc., New York, 1959), pp. 322, 328.

10. P. Grivet and N. Bloembergen, Eds. Quantum Electronics III(Columbia University Press, New York, 1964), pp. 999-1019, 1187-1202; also D. Rss, Proc. IEEE 52, 196(1964).

January 1968 / Vol. 7, No. 1 / APPLIED OPTICS 71

RELATIVE REACTION RATE =MEASURED RATEI, (Br2 )2

A DIRECT LASER BEAM

0 ATTENUATED TO 20 %

RELATIVE ABSORPTION

4-

2 F

toI-2

a

C

10

B

6

4

11. See for example, D. G. Peterson and A. Yariv, Appl. Opt. 5,985 (1966).

12. I. H. Malitson, J. Opt. Soc. Am. 52, 1377 (1962).13. W. B. Tiffany and A. L. Schawlow, Bull. Am. Phys. Soc. 11,

898 (1966).

14. C. B. Kistiakowsky and J. C. Sternberg, J. Chem. Phys. 21,2218 (1953).

15. W. B. Tiffany, Dynamics of Selective Molecular Excitation:Laser Photocatalysis of Bromine Reactions, Ph.D. Disserta-tion, Stanford University, Stanford, California (May 1967).

Meeting Reports continued from page 66

Harry Keegan Clemson University (left), L. Sloan Johns Hopkins, Norma Miller Technology Inc., J. P. Weiss E. I. du Pont de Ne-mours, and Tarry Sperling University of Texas, photographed by Carl Leistner Ultra Carbon.

pointed out that the human body radiates as a blackbody irrespec-tive of pigment. After disclaiming medical authority, he showedan excellent series of slides illustrating the use of the thermographin diagnosing circulatory and metabolic disorders.

At the Thursday evening banquet, Edwin Land Polaroid Corp.received the Frederic Ives Medal from President-elect A. F.Turner, acting on behalf of President John A. Sanderson, absentbecause of illness. Also regrettably absent for a similar reasonwas Executive Secretary Mary Warga. The speaker of the eve-ning, Harlan H. Hatcher, President of the University of Michigan,was introduced by R. A. Sawyer, Dean Emeritus of the Universityof Michigan. President Hatcher's talk on the advancement ofscience in the Midwest stressed the need for integration of hithertolargely uncoordinated disciplines of science into a unified effort tosolve the problems arising from population growth and urbancrowding.

The Ives Medal Address by Dr. Land the next day was a typicalLand-OSA-meeting presentation: a carefully prepared paper onan aspect of human vision, well illustrated, accompanied by strik-ing demonstrations, giving rise to subsequent extensive discussion.

In addition to the formal sessions, the meeting comprised Tech-nical Group meetings distributed from Wednesday eveningthrough Friday afternoon. Laboratory visits were scheduled tothe University of Michigan, Wayne State University, Bendix Re-search and Ford Motor Scientific Laboratoris. Tours were alsoprovided of General Motors Technical Center and the Universityof Michigan Willow Run Center and North Campus. Theformer included a movie of GM research activities, a walk throughthe Center, and visits to two of the laboratories. A highlight ofthe latter was the holography work at the Radar and OpticsLaboratory of the Institute of Science and Technology. Thestrike against Ford Motor Company unfortunately forced cancel-lation of visits to the Dearborn Assembly plant.

Mechanical arrangements for the meeting appeared commend-ably smooth in operation, thanks to Dave Fry General Motors andhis well-chosen committee, despite a few obstacles. D. J. Lovell

University of Michigan pointed out a threat to the meeting whichwas averted only at the last moment and, he says, by dint of heroic

efforts on the part of the Committee, viz., a World Series inDetroit. One hurdle which proved insurmountable was the prep-aration of a satisfactory noncommercial exhibit. The few in-dividual items were pleasing, but it is to be hoped that no visitorgot his total impression of the scope of science in the Midwest fromthe misfortune-ridden remnant of the original plan.

Many excellent papers have necessarily gone unmentioned inthis summary, as doubtless have a number of interesting side-lights. These omissions reflect the fact that this meeting, likeOSA meetings generally, provided such a wealth of stimulatingactivities as to make one wish for ubiquity.

First International Symposium on the Physics ofSelenium and Tellurium, Montreal,12-13 October 1967

Reported by A. Vasko, Czechoslovak Academy of Sciences

The Symposium was sponsored by the Selenium-TelluriumDevelopment Association. (Members of this Association areeight industrial companies in the United States and Canada, andthe European Selenium-Tellurium Committee. The Associa-tion's objective is to stimulate interest in selenium and telluriumand to promote new applications primarily through sponsoredresearch.) This Symposium is a natural outgrowth of the Se-lenium Physics Symposium which was held in London in June 1964under the auspices of the European Selenium-Tellurium Commit-tee. Its Chairman was W. Charles Cooper, Director of Researchat the Noranda Research Centre, Pointe Claire, Quebec, whoseInstitute was responsible for the organization of the meeting.

Recent progress in the physics of selenium and tellurium wasreviewed by the keynote speaker, J. Stuke University of Marburgand the subsequent technical program was divided into foursessions: Band structure and electrical properties (Chairman:F. T. Hedgcock McGill University), Crystal growth and character-ization (Chairman: Charles Wood Northern Illinois University),Optical properties (Chairman: W. W. Harvey Kennecott Copper

continued on page 82

72T APPLIED OPTICS / Vol. 7, No. 1 / January 1968