A lidar system for measuring atmospheric pressure and temperature profilesGeary K. Schwemmer, Mark Dombrowski, C. Laurence Korb, Jeffry Milrod, Harvey Walden, and Robert H.Kagann Citation: Review of Scientific Instruments 58, 2226 (1987); doi: 10.1063/1.1139327 View online: http://dx.doi.org/10.1063/1.1139327 View Table of Contents: http://scitation.aip.org/content/aip/journal/rsi/58/12?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Comparison of the simulated and LIDAR measured atmospheric density and temperature in lower thermosphere AIP Conf. Proc. 1628, 1337 (2014); 10.1063/1.4902746 Vertical Profiling of Atmospheric Backscatter with a RamanAerosol Lidar AIP Conf. Proc. 1203, 388 (2010); 10.1063/1.3322473 Temperature measurement of an atmospheric-pressure plasma torch Rev. Sci. Instrum. 70, 3032 (1999); 10.1063/1.1149864 Atmospheric water vapor measurements with an infrared (10μm) differentialabsorption lidar system Appl. Phys. Lett. 28, 542 (1976); 10.1063/1.88815 Vibrator sensors for atmospheric pressure, temperature, and humidity measurements J. Acoust. Soc. Am. 56, 1644 (1974); 10.1121/1.1903491

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

A lidar system for measuring atmospheric pressure and temperature profiles

Geary K. Schwemmer, Mark Dombrowski, C. Laurence Korb, Jeffry Milrod, a)

Harvey Walden, and Robert H. Kagannb)

Laboratory for Atmospheres, NASA/Goddard Space Flight Center, Greenbelt, Maryland 20771

(Received 26 May 1987; accepted for publication 11 August 1987)

The design and operation of a differential absorption lidar (Light Detection And Ranging) system capable of remotely measuring the vertical structure of tropospheric pressure and temperature are described. The measurements are based on the absorption by atmospheric oxygen of the spectrally narrowband output of two pulsed alexandrite lasers. Detailed laser output spectral characteristics, which are critical to successfullidar measurements, are presented. Spectrallinewidths of 0.026 and 0.018 cm -I for the lasers were measured with over 99.99% of the energy contained in three longitudinal modes.

INTRODUCTION

This article describes the design and operation of a differen­tial absorption lidar (DIAL) system developed for remotely measuring atmospheric pressure and temperature fields. In addition, experimental results of system tests and calibra­tions are discussed, including detailed measurements of alex­anddte laser spectral properties.

Pressure and temperature profiles are fundamental pa­rameters for specifying the state of the atmosphere and pre­dicting the weather. Accurate, high-resolution measure­ments of surface pressure and temperature profiles have been specified as observational requirements for forecasting and modeling of weather and climate by the Global Atmo­spheric Research Program I and the NASA Climate Science Working Group. The required accuracies specified were 0.3% for pressure and 1 K for temperature, with 2-km verti­cal resolution and 5OD-km horizontal resolution. Current state-of-the-art passive sounders are limited to a vertical re­solution of about 6-10 km for statistically independent mea­surements of temperature. Other Hdar methods of measur­ing tropospheric temperature profiles have been proposed or achieved based on rotational Raman scattering,2-5 DIAL techniques using two molecular absorption lines, 0 taking the vertical derivative of a lidar measured pressure profile/'S and by measuring the thermal Doppler broadening of Ray­leigh scattered laser light.9 DIAL measurements of surface pressure had been proposed as early as 1968,10 and more recently using a multi wavelength laser approach based on the refractive index dispersion of the atmosphere. I! How­ever, no remote measurements of the atmospheric pressure profile existed prior to those 12-15 made with the system de­scribed in this article. Lidar techniques can meet or exceed the above accuracy and resolution requirements for tem­perature profiling and, in addition, provide high-accuracy pressure versus height profiles.

The technique for measuring the pressure profile, sur­face pressure, and doud-top pressure height developed by Korb et al. 1

6--18 utilizes the pressure-dependent absorption of trough regions in the wings of two molecular resonant absorption lines of a uniformly mixed gas in the atmosphere.

The temperature measurements sense the line center absorp­tion of highly temperature-sensitive lines of uniformly mixed gases. 18

--20 Both the pressure and temperature mea­

surements can be made using absorption in the oxygen A band near 760 nm. In previous experiments using contin­uous wave lasers and a fixed reflective target, we have suc­cessfully demonstrated the basic technique and have accu­rately measured the average temperature and pressure over a I-km horizontal atmospheric path near ground level at the Goddard Optical Research Facility.2! The lidar system de­scribed here is part of an evolutionary series of experiments aimed at developing a satellite capability for global !idar measurements of atmospheric temperature and pressure. This system has been used to make the first remote measure­ments of the atmospheric pressure profile1Z

-14 from our

ground-based lidar laboratory at the NASA/Goddard Space Flight Center. More recently, the system was integrat­ed into the Goddard Airborne Lidar Facility on the NASA/W al10ps Lockheed Electra aircraft, and successfully made pressure profile measurements in a nadir-viewing mode. 15

I. THEORY OF OPERATION

We present a brief treatment of the DIAL technique for measuring atmospheric absorption, followed by a simplified discussion of the pressure and temperature methods. A thor­ough treatment of DIAL may be found in Ref. 22 and de­tailed descriptions of the pressure and temperature tech­niques can be found in Refs. 17 and 20, respectively.

A. DIAL technique

A transmitter directs a pulsed laser beam through the atmosphere and a portion of the transmitted beam is scat­tered in all directions by the molecules and aerosols in the atmosphere. The energy backscattered to a telescope receiv­er is collected and converted to an electrical signal using a photomultiplier (or other suitable detector). The signal is digitized and stored for later analysis. The altitude from which a given portion of the return signal was scattered is

2226 Rev. Sci.lnstrum. 58 (12), December 1987 0034-6748/87/122226-12$01.30 @ 1987 American Institute of Physics 2226

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

calculated from the round trip travel time of the light pulse. The range resolution is ultimately limited by the length of the laser pulse and is given as

tlR>ctlt /2, (1)

where c is the speed oflight and Ilt is the laser pulse temporal width. However, other factors can degrade this resolution such as the photomultiplier response time, the digitizing rate, and the electrical bandwidth. To reduce high-frequen­cy noise and avoid aliasing, an electronic filter is used to limit the frequency response to a bandwidth (at - 3 dB) equal to half the digitizing sample rate. The stored data are digitally averaged in order to improve the signal-to-noise ratio of the return signals by averaging over altitude and by multipulse (time) averaging.

The power received by the detector as a function of the range r is given by the lidar equation

Per) = &1] ~{3 (r) . 2 r 1T

xexp( - 2f[,8(R) +Kc(R) +K(R)]dR ).

(2)

where E is the energy in the transmitted laser pulse, 'YJ is the receiver efficiency, A is the receiver area, and ,8rr (r) is the sum of the molecular (Rayleigh) and aerosol backscatter coefficients. The exponential factor represents the integrated extinction due to an scattering ,8(r), continuum absorption Kc (r), and resonant absorption K(r), processes. We have assumed monochromatic laser emission for clarity. For highest accuracy, the finite spectral bandwidth of the laser, as well as the Doppler broadening of molecular backscatter, must be treated. The errors incurred because of finite spec­tral bandwidth have been evaluated in Refs. 22-25, and a method of removing the effects of nonmonochromicity has been developed in Refs. 17 and 20. To measure the resonant absorption, two laser pulses with slightly different wave­lengths are transmitted, one with weak or negligible absorp­tion (off-line) and the other with significant resonant ab­sorption (on-line) by the oxygen in the atmosphere. If the two wavelengths are spectrally dose enough so that the scat­tering and nonresonant absorption properties of the atmo­sphere are essentially the same at the two wavelengths, then the ratio of the on-line signal to the off-line signal is

Pon(r) Eon (2i'rK R K R dR) --=--exp - l 0"( ) - old)]· Porr (r) EoJf 0

E (i'- \ = ~ exp - 2 K(R)dR J, Eotl' 0

(3)

where K(r) is the differential absorption coefficient. This may also be expressed as a function of altitude (z) as

Pon (z) = Eon exp( _ 2izX(Z)dZ), (4) Potf (z) EatT z"

where Zo is the altitude of the Iidar system.

B. Pressure profile

For measuring the pressure profile, we utilize the pres­sure broadening of oxygen absorption lines and measure the

2227 Rev. SCi.lnstrum., Vol. 58, No. 12, December 1987

.' ., •••• -.-.-. ' •••••• ;' •••••••••• ·._., •••••••• 7 •••••• ,. ••• _ ••• _._ ••• '.'. • •••••••••• ~._.! .•••••• ',. •••••••.• :.:.:-: •••••••• , •••• '" •••.•.• _.

path-integrated absorption directly:

rX\Z)dZ= _.!In(POIl(Z)/POff(Z)). (5) ..Iz" 2 Eon/Eoff

Under atmospheric conditions, the oxygen absorption line­shape may be approximated by the Voigt lineshape. Using the first term in a power series expansion of this lineshape about line center, it can be shown that to a good approxima­tion the integrated absorption in the far line wing varies as the difference in the squares of the pressures at the laser altitude 20 and the measurement altitude z. 17

IZX(Z)dZ= Clp2(Z) -p~1 , (6) ::0 (VOll - 'lie)

where C is a constant which includes the line parameters, p(z) and Po are the pressures at z and zo, Von is the laser frequency, and 'lie is the oxygen absorption line-center fre­quency. For a trough region between two absorption lines, Eq, (6) becomes

K (Z) dZ = 11 0 + 2 0 , , l z - C !p2(Z) _p2j C [P2(Z) _p21

"" (Von - '1'1)2 (Von - V 2 )2

(7)

where VI and '112 are the two line-center frequencies. In prac­tice, the constant C * is determined from an experimental calibration, and higher-order terms may be included. This equation is solved for pressure p (z), which is then computed at each altitude using the locally measured pressure Po, and the measured integrated absorption.

C. Temperature profile

For measuring the temperature profile, the absorption coefficient as a function of altitude is calculated from the derivative of the natural logarithm of the return signal ratio

K(Z) = ~lr _ ~ In(POn (Z)/POff(Z))] . (8) dZ 2 Eon/Eoff

To determine the temperature, the absorption coefficient is expressed as a function of the absorption line strength and Voigt lineshape, both of which are functions of temperature. The temperature dependence of the line strength for lines with high ground-state energy is strong, due to the Boltz­mann distribution of rotational states, and is what gives the absorption coefficient its high-temperature sensitivity. The exact form of the resulting relationship is intractable in terms of a solution for temperature; hence, an iterative tech­nique is used. From Eq. (10) in Ref. 20;

E/k 7;+ I =. 2 (9)

In[Cqjj eVe - vc )] -In(KTJp)

The temperature (at each altitude) is calculated on the i + 1 iteration using the prior temperature T; and the measured absorption coefficient K; E is the ground state energy of the absorption transiti.on, k is Boltzmann's constant, C is a con­stant, q is the oxygen mixing ratio./; (ve - vc ) is the value of the Voigt lineshape function for T; and for an effective laser frequency 11 e' 26 The pressure profile p (Z) can be calculated for each iteration using the hypsometric equation with the

Lidar system 2227

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

TRANSMITTER OPTICS

-- HIGH RESOtuTION LASER

-- lOW RESOLUTION LASER

AID TRANSIENT RECORDERS TIMING CONTROL

~--~--~~------~

FIG. I. Block diagram of ground-based lidar system. Dashed arrows indicate optical paths and solid arrows indicate electrical paths.

OPERATOR'S CONSOLE

CAMAC BUS

previous iteration's temperature profile and a locally mea­sured pressure.

There are several publications addressing the calcula­tion of absorption or extinction from the lidar return signals which are sampled on finite intervals, and we refer the reader to these references27

,28 for a detailed study ofthe digital tech­niques involved. Multipulse and range averaging of signals to derive measured profiles also require· special consider­ations29

,30 which are not addressed here.

II. LIDAR SYSTEM DESCRIPTION

Our lidar system is designed to remotely measure atmo­spheric pressure and temperature profiles from the ground looking up and from an aircraft looking down to the surface. Major parts of the system are used in both configurations; only the supportive framework and telescope are not shared.

The !idar system consists of three major subsystems (Fig. 1): a transmitter subsystem containing the lasers with their power supplies and cooling systems, beam steering op­tics, and laser output monitoring instruments; a receiver

MAGNETIC TAPE

RECORDER

subsystem comprised of a telescope, gated photomultiplier, and optical and electronic filters; and the control and data­acquisition subsystem consisting of electronic modules for digitizing the transmitted and received signals, system clocks and timing circuitry, a microprocessor controller, data handling and storage devices, and a video terminal for operator interface.

Figure 2 shows the major system components as they are configured in the NASA Lockheed Electra aircraft. The transmitter and receiver optical tables are mounted to a sin­gle rigid aluminum framework which is shock-mounted to the aircraft floor tracks. The telescope is also attached to this framework and can be oriented pointing down as well as up. This system is similar to a water vapor DIAL system devel­oped by Browell et at. 3

! and utilizes the same aircraft; how­ever, it shares no components with that system.

Ao Laser transmitter

The transmitter subsystem is comprised of two narrow­band tunable pulsed alexandrite lasers, beam steering and

80 D n LASER

Pow"-~A-R S-;u-Rp-P-lY-~ COOl

R'" t5~

PRINTER

¢OIRECTION DATA

OF SYSTEM

FLIGHT

TRANSMITTER R TABLE LJ ~

CHAIRS LASER

CONTROL ELECTRONICS

2228 Rev. Sci.lnstrum., Vol. 58, No. 12, December 1987 lidar system

FIG. 2. Top view layout oflidar system com­ponents inside NASA Lockheed Electra air­craft. A 40-cm window in the aircraft belly lies directly below the telescope.

2228

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

O.5m SPECTROMETER

2·M's

OPTICS MODULE

collimating optics, photodiodes for measuring laser energy and marking time, a O.5·m focal length spectrometer for an approximate determination oflaser wavelength, and a multi­pass gas cell for spectral calibrations and diagnostics (de­scribed in Sec. IV). Figure 3 is a block diagram of the trans­mitter optical bench showing the relative location of the components. Not shown is the gas cell, attached to a shelf underneath the transmitter optics.

The lidar system currently utilizes two solid-state alex­andrite lasers which we will refer to as Alex I and Alex H. Ruby-pumped liquid dye lasers were initially used in the transmitter18 and at that time were the only commercially available tunable high-resolution lasers for this spectral re­gion. A single alexandrite laser (Alex I) and ruby-pumped dye laser were used to make the first remote measurements of the atmospheric pressure profile. 12

·14 The repetition rate

of the ruby-pumped dye laser system was limited to 0.5 Hz, whereas with two alexandrite lasers, the system rep rate is 5 Hz, limited by the data systemo Each alexandrite laser is tunable from approximately 725 to 780 nm. To obtain suffi­cient measurement accuracy, the spectral bandwidth (FWHM) of the on-line Jaser needs to be no greater than 0.02 cm - 1 with a line-center frequency stability of ± 00002 em I during the course of a set of measurements (typically 30 s-30 min). The spectral specifications for the off-line la­ser are less stringent, namely, 0.2-200 cm- 1 FWHM and ± 0.05 cm -1 stability, depending on the exact wavelength

a·SWITCH CONTROL

ALEXANDRITE LASER

ALEXANDRITE LASER

FIG. 3. Transmitter schematic diagram showing relative location of components which are mounted to an 84X244-cm optical bencho Key: M-mirror. GAP­glan air prism, A. /4--quarter-wave cir­cular polarizer, BS--beam splitter, PD-photodiode, and ND-neutral density filter.

chosen. The development of solid-state tunable alexandrite lasers32 has resulted in a more compact and efficient system and holds promise for developing a space-quali.fied Udar sys­tem capable of carrying out measurements of atmospheric temperature and pressure from satellites. Our alexandrite lasers have become the model for the design used in an auto­mated high-altitude DIAL system being built by NASA, called the Lidar Atmospheric Sensing Experiment (LASE).33

The Alex I laser is capable of 100 mJ of narrowband output in a 130-ns pulse at a lO-Hz repetition rate. It has an 80X5-mm-diam alexandrite rod pumped by two linear tlashlamps in a flooded double-ellipse pump chamber. A so­lution of sodium nitrite is used for cooling the rod and lamps and acts as a UV absorber. A small amount of sodium hy­droxide is added to buffer the pH which helps keep the sodi­um nitrite in solution. The solution is temperature stabilized to 50 "C. The optical layout of the laser head is shown in Fig. 40 The high-reflectance mirror at one end of the laser cavity has a 4-m radius of curvature, and the 70% reflectance out­put coupler is flat. The cavity is acousto-optically Q switchedo Spectral narrowing and tuning is accomplished manually using three tuning elements: a three-plate bi.refrin­gent filter (BRF), a I-mm tilt-tuned solid etalon, and a 4.23-mm solid etalon that can be either tilt or temperature tuned. For off-line operation, only the BRF is used in the cavity, which gives a spectrallinewidth of 1.5 cm - lor, with the

PULSE FORMING NETWORK #1 J ,.....-

II-

......

r- PULSE FORMING NETWORK #2

~ =----r

t I

----.-:;~ __ t_ n. __ n __ .f'.OUTPUT ~~v---r---t1--TI--'VBEAM

PUMPING CHAMBER ! THIN iALON 7~~ OUTPUT COUPLER

I

BIREFRINGENT TUNER

2229 Rev. Scl-Instrum., Vol. 58, No. 12, December 1987

THICK ETALON

lidar system

FIG. 4. Alexandriie laser optical resonator cavity configuration. The cavity has an optical path length of about 55 em. (Figure, courtesy R Sam, Allied Corporation. )

2229

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

addition of the 1-mm etalon, a linewidth 0[0.08 em- J • With all three tuning elements installed, the spectrallinewidth is about 0.026 cm -I.

Our second alexandrite laser (Alex II), which can be electronically scanned in frequency, is presently used as the on-line laser. It has an output of up to 150 mJ of narrowband energy iIi a lOO-ns pulse at a lO-Hz repetition rate. Both lasers have temporally Gaussian-shaped output pulses. This laser also has three tuning elements, but they are electroni­cally controlled. The optical layout is almost identical to the first laser, but certain specifications differ as follows: the la­ser rod is l00X 5 mm diameter, and it is kept at 80°C with heated distilled water. The flash lamps are on a separate cooling loop using distilled water. A glass flow tube around the rod separates the two loops and provides UV absorption. This allows us to optimize the laser rod gain using the high temperature without subjecting the lamps to the stress of this temperature. The high reflector has a 5-m radius of curva­ture, and the flat output coupler has a reflectivity of 60%. The five-plate birefringent filter is controlled with a dc torque motor and uses an optical encoder for position feed­back. The thin etalon is 0.74 mm thick and is tilt tuned. The high-resolution etalon is a piezoelectrically scanned, 1-em air space Fabry-Perot 6talon. The laser spectral linewidth with all three elements installed is about 0.018 cm -1. The continuous tuning range without etalon order jumping is 3.0 cm- I

. For a larger change in frequency, the thin etalon and BRF must be set to a different order number of the thick etalon before another 3.0-cm - 1 scan is possible. In this man­ner, piecewise continuous tuning over the entire fluores­cence band of the laser (725-780 nm) is possible without any gaps.

The two laser beams are combined onto a single axis using a Glan-Air prism (see Fig. 3) . The beams are orthogo­nally polarized so that one will be transmitted straight through the prism while the other, entering the prism from the side, will be nearly totally reflected at the Brewster cut in the prism. A Pockels cell is used as a circular polarizer on the two orthogonally polarized beams, producing equal polar­ization components in each laser wavelength. This is to mini­mize any systematic discrepancies in the measurements of outgoing energy or backscattered signal due to preferential reflection of either polarization in any of the other optics. The beam splitter that samples the outgoing beam for the energy monitor is placed near normal incidence to further minimize preferential reflection of either laser wavelength. The voltage on the Pockels cell is adjusted so that the ratio of the two laser energies transmitted by the system, as mea­sured by an independent photodiode placed in the outgoing beam, equals the ratio measured by the energy monitor. This effectively calibrates the energy monitor, and is double checked by exchanging detectors and digitizers.

The two lasers are tuned to within 0.5 cm -1 of the de­sired frequencies by monitoring their wavelengths with a O.S-m Czerny-Turner spectrometer. This instrument has an X-Y CRT display of its focal plane array detector output, including a cursor with pixel number readout. The laser out­puts are directed to the atmosphere or to a multipass gas cell, and a real-time plot of atmospheric (or gas cell) transmis-

2230 Rev. ScLlnstrum., Vol. 58, No. 12, December 1987

[-;4~iRZ;;:-j-;;-1- - ,-- ,--,- -T~2~~;-;~~-

00:05:50 ELAPSED TIME 115 DISPLAY COuNT

r r

I \~,-,j \" j~'.\ l

l1 L2 81 B2 ', .. - j I 8 7 7 ENERGY 7 50 RANGE '-1M ITS

28 28 28 28 GAS CEll 0 0 Li, L2 RATIOS 1:' 1~~~ 1:"i~:"~R.:f:... __ 1- __ , ___ I __ JO~ ~~~~S~O..::'

ATMOSPH1::RIC ABSORPTION DISPLAY· orr EII"'iilIMI CHANNEL '2

FIG. 5. Console monitor display showing real-time plot of atmospheric ab­sorption as the laser frequency is scanned across the two oxygen lines near 13 154 em - '. This is the (two-way) transmission across an 8S0-m horizon­tal path near the ground to a hard target. The numerals indicate photodiode readings and housekeeping information.

sion is displayed on the system console terminal (Fig. 5). From the terminal, we select the altitude and the number of range bins and laser pulses to be averaged for the transmis­sion display calculations. The on-line laser is then electroni­cally tuned with a constant scan rate while the display is continuously updated~ generating an absorpiion spectrum which is displayed on the terminal screen. Figure 6 shows a laser frequency scan of the transmission through a multipass gas cell showing the low-pressure oxygen lines at 13153.4 and 13154.2 cm - 1 • By comparing the absorption line spac­ings and relative strengths with laboratory data on these line parameters, the laser etalon position is calibrated to the known oxygen absorption line frequencies. The on-line laser is then manually tuned to the desired absorption feature, either an absorption trough for pressure measurements or an absorption peak for temperature measurements. The mea­surements oflaser signal backscattered from the atmosphere versus altitude for the on-line and off-line frequencies are then recorded on digital magnetic tape for postmeasurement analysis.

A silicon photodiode is used to monitor the output ener­gy of each laser to normalize the lidar return signals for shot-

I "'''~''''''''''''''''.'''''''''''''''''

~ BO r " <t

f:: ~ .~ ~ 60 i Vl z

~ 40[ f ;:;

20 ~.. t,

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

L V ~ o L __ .. _ . .J ___ ..L ___ 1_._._ .J. _._._.L _._.-L-13,153.0 13,153.4 1:t~53.8 13,1!}4.)

f=REQUENCY {err: 1)

FIG. 6. Playback of previously recorded gas cel! transmission measured as the laser frequency was scanned across the two oxygen lines near 13 154 em· I. The O2 pressure in the cell was 48 Torr and total path length was 149 m.

Lidar system 2230

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

to-shot energy variations and for the ratio of the on-line and off-line laser energies needed for measuring pressure [see Eq. (5) ]. A small portion of the outgoing beam is directed to

the energy monitor. An opal glass diffuser in front of the energy monitor eliminates effects due to spatial nonuniform­ities in the laser beam and the detector surface. Neutral den­sity filters are used to attenuate the laser radiation, and a 10-nm interference filter reduces background and flash-lamp radiation. This signal is fed to a 9-bit gated charge digitizer. Each laser pulse is digitized and recorded on tape with the lidar return signals.

The system includes a lidar return simulator which can be used to perform system tests without actually transmit­ting any laser energy into the atmosphere. It consists of a LED placed in front of the photomultiplier, and pulsed with an exponentially decaying current to simulate actual laser radiation scattered off the atmosphere. It simulates the dif­ferent atmospheric absorptions for the on-line and off-line laser frequencies by pUlsing the LED twice, with different decay time constants to represent the two different laser re­turn signals.

B. Receiver

The major components of the receiver are a telescope, focal plane optics, a photomultiplier, optical and electrical filters, power supplies, amplifiers, and gating electronics.

The telescope used for aircraft measurements is a 40-cm-diamf /12 Cassegrain with aluminum-coated optics and a Nasmyth focus. The ground-based system telescope is a 44-cm-diamfI4.5 Newtonian with a gold-coated primary and aluminum-coated secondary mirror. This telescope rests on the floor and is directed to the zenith through a hatch in the roof. The aircraft system telescope can be directed toward the zenith as well as the nadir, through either of a pair of 40-cm windows in the aircraft. The windows are mounted at a 7.so angle with respect to the optical bed and floor of the aircraft to prevent reflections of the transmitted laser beam from entering the receiver. The laser beams are transmitted coaxially to the telescope, directed off a mirror mounted be­hind the telescope's secondary mirror. Baffle tubes enclose the beams from the transmitter optical table to the aircraft's window in order to prevent stray scattered light or reflec­tions from entering the receiver. A low-pass optical filter with a nO-nm cutoff is used in front of the PMT to reduce background light for night measurements. The PMT's pho­tocathode is located at the focal plane. For daytime measure­ments, we use a collimating lens and narrowband (1.0--2.0 nm) interference filter to reduce background light. The pho­tomultiplier is an RCA 8852 which has a multialkali photo­cathode with about 4% quantum efficiency at 760 nm. The detector is electronically gated on for about 100 ps, starting several microseconds after the laser fires by switching the voltage on the second dynode. When looking down from the aircraft, the ground or ocean surface reflection produces a very strong transient in the signal. We notch out this tran­sient by gating the tube off for several microseconds while the ground reflection occurs. This greatly reduces a small bias in the tube's output following such a strong transient. Background measurements are made on each shot about 50

2231 Rev. Sci.lnsirum., Vol. 58, No. 12, December 1987

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

p.s after the ground reflection. The PMT output is amplified with a low-noise transimpedance amplifier providing 3.3 kV / A gain. The signal is then filtered at the Nyquist band­width with a single pole passiveLC filter before input to a 10-bit (more recently, a 12-bit) digitizer. The digitized record of the on-line and off-line return signals is transferred from the digitizer memory to magnetic tape after each pulse pair, along with data from the system clock, counters, and energy monitor. The system was built with two receiver channels for using two detectors, one with high gain and one with low gain, to increase receiver dynamic range. To date, only one channel has been utilized in Ii dar measurements.

c. Controller and data acquisition

The data-acquisition system converts the analog electri­cal signals to digital information and records that informa­tion on magnetic tape. The same microprocessor that con­trols the lasers also controls the data-acquisition equipment. The CAMAC Standard (ANSI/IEEE 583-1975) was cho­sen as the format for data-acquisition equipment. The modu­larity helps reduce the system cost by tailoring the system to meet only necessary or desired functions, and reduces obso­lescence (i.e., improvements to anyone function require re­placing only one module). Each module plugs into a crate containing power supplies and backplane for the data bus (CAMAC data-way). Contained in the crate are a con­troller, two transient digitizers, two programmable clocks, a charge digitizer, and the Edar control module which con­tains circuits to control timing functions, such as the firing of the lasers and the gating of the photomultipliers and tran­sient digitizers. The microprocessor used for system control and the coordination of all I/O is an LSI-ll/23. The com­puter system includes a dual floppy disk drive, a video graphics terminal, a graphics printer, and a ~-in., nine-track digital tape drive. The computer has floating point hardware and 2 Mbytes of RAM. All operator information is input via the video terminal, and all data are stored on the nine-track magnetic tape. The CAMAC crate controller connects di­rectly to the LSI Q-Bus and makes an the CAMAC modules appear as individually addressable peripherals to the micro­processor. A 12-bit 20-MHz CAMAC transient digitizer (Transiac model 2012) is used to digitize the output of the PMT (Fig. 7). It is capable of storing 8000 12-bit samples, although typically less than 500 samples per laser shot are taken. The digital samples are read out by the computer after each pair of shots. The digitizer has an oscilloscope display output which drives an X - Y CRT display monitor. A modi­fied LeCroy model 8501 programmable three-speed clock generator is used to control the sampling speed of the tran­sient digitizer. This allows us to control the digitizer sam­pling rate during a measurement cycle, as discussed in detail below. The clock module contains counters which we use to record the on-line and off-line laser pulse separation and the number of samples digitized.

The altitude registry on the return signals is accom­plished by triggering a clock and counter circuit with the leading edge of a saturated signal from a photo diode which detects when each laser pulse was transmitted (laser sync). The clock pulses are used to trigger the receiver signal digi-

Lidar system 2231

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

AMPLIFIER

.... ------.. .c SIGNAL

12-BIT DIGITIZER

SIGNAL DISPLAY

TRIG.

elK.

o o o o [IIJ

lID IID an lID lIDAR CONTROL DISPLAY PANEL

PROGRAMMABLE CLOCK

(MASTER)

40MH. eLK

DIG. GATE LlDAR

CONTROL & DATA LINES

b...------CI PMT GATE CONTROL t:-:>-----'i-.... MODULE 0..----1-.

LASER SYNC P.D.I-------------Ir.-d SYNC IN-,

NIMGATE...J

ENERGY MONITOR

GATED INTEGRATOR

GAS CELL I _________ ~~~~~~~~ MONITOR F.D·r PROGRAMMABLE

CLOCK (SLAVE)

FOR ADDITIONAL

PMT CHANNEL CAMAC

MODULES

FIG. 7. Data-acquisition electronics diagram. One receiver channel is shown for clarity.

tizer, and the counter ticks off a preset delay which deter­mines when digitizing commences relative to the laser out­put. The maximum clock rate to the digitizer is 20 MHz, which is derived from an 80-MHz oscillator. Each laser sync signal is used to restart the delay counters. A problem oc­curred during the data analysis when we ratioed the on-line and off-line return signals, due to inadequate accuracy in the altitude registry between the two signals. Since the laser pulse can occur randomly anywhere within the digitizer clock period, the average error in the clock phase (relative to the laser pulse) is i the clock period, or ±! the clock period. Structure in the return signals due to inhomogeneities in the atmospheric aerosols produced a ringing effect in the ratio of the return signals, particularly where large, rapid changes in the signals occurred at aerosol layer boundaries. The data were adjusted computationally to correct for the relative er­ror in altitude registry between on-line and off-line returns, and it was found that the ratio was sensitive to misregistries as small as one tenth of a clock period, even though the signal was bandwidth limited to less than half the clock frequency. Subsequently, the clock electronics were redesigned so that the 80-MHz oscillator is now rephased to each laser pulse with a residual error of a few hundred picoseconds. In addi­tion, it was determined that fluctuations in the laser output energy caused the leading edge detector to undergo signifi­cant (tens of nanoseconds) jitter with respect to the laser energy pulse peak. The leading edge detection electronics are being replaced with a constant fraction discriminator, and

2232 Rev. Sci.lnstrum., Vol. 58, No. 12, December 1981

the photodiode operated in the linear region. These have been tested and the residual jitter due to this effect is a few hundred picoseconds.

The !idar control module (LeM) is the interface between various parts of the lidar system (Fig. 7). It pro­vides trigger pulses to the lasers, gating signals to the photo­multipliers, a gate to the laser energy digitizer, and timing control pulses to the programmable clocks based on the

TRIGGER LASER 11

TRIGGER LASER #2

r-I _.-:- --~- -" -typo 300,",5 -- - ---.i

: : _IL __ ____ __ rv 1 oo~s

LASER o---.l..i /--~-;f-----L~/ SYNC PD. !-=- :_ ~~;t ---~ l-----n---lf 1-PMTll SIGNAL

CLOCK Ace #1

PMT#2 GATE

PMT12 SIGNAL

CLOCK ADen

- -K~----N--j'vL ___ llillilllll·-l._s--{I-L-.-l-um~

FIG. 8. Timing diagram i1lustrating the data-acquisition sequence. Two la­sers are triggered sequentially and a photodiode detects the laser pulses in order to synchronize the gating and digitizing of two photomultipliers, each of which measures part of the atmospheric backscatter signal from each laser pulse.

Udarsyslem 2232

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

PMT gate pulse settings. Figure 8 illustrates the timing se­quence for a laser pulse pair data-acquisition sequence. The following describes each of the LCM functions in detail. A "fire" command to the LCM, initiated via a computer in­struction or a manual pushbutton, causes the laser timing circuit to initiate a firing sequence. A separate trigger is gen­erated for each of the two lasers, and each laser emits a pulse about 100 flS later. The lasers are triggered in sequence with a time separation determined by a thumbwheel switch on the control panel.

In order to suppress ground loop currents and noise on the laser control lines from entering the data system and Edar signal cables, optically isolated logic interfaces have been installed on the laser trigger lines. The logical "OR" of both trigger pulses can be sent to the Udar return simulator (LRS) . However, either the lasers or the LRS is in operation at any given time, as determined by a toggle switch on the control panel. One or both lasers can be fired, depending on the state of a programmable flip-flop. A toggle switch on the control panel can be used to interrupt automatic (comput~ er) operation of the system to allow manual firing of the lasers. The laser and LRS triggers are suppressed when the system is programmed into the calibration mode, which al~ lows background readings to be taken on all detectors. Thumbwheel switches on the display module front panel are used to set each PMT gating delay and width. The delay is measured from the detection of the laser output by the sync photodiode until the PMT is gated on. Each digitizer is clocked at a high rate (5, 10, or 20 MHz) while the PMT is gated on. The clock is slowed to 100 kHz between the end of the PMT gate for the first lidar return and the beginning of the gate for the second lidar return. The dock then resumes its previous high rate for digitizing the waveform from the second lidar return.

An example of the lidar returns made from the Electra aircraft at an altitude of 3 km is shown in Fig. 9. The large reflection from the ocean surface was not gated out in these measurements. Each curve represents the average of 20 re­turn signals, each digitized at 5 MHz and bandwidth limited to 2 MHz, corresponding to 75-m resolution with 30-m sam~ pIing intervals. No other smoothing was performed. Figure

2500

~ UJ 0 :::l 1500 ~

f= ...J «

500

~"'" OFF-LINE /

.t I

1 2 3

RELATIVE INTENSITY

FIG. 9. Typical example of atmospheric backscatter signals from the on-line and off-line laser pulses. Structure in the signals is due to layers of atmo­spheric aerosols.

',2233 Rell.Sci. lnstrum., Vol. 58, No. 12, December 1987

".-••••• , ••••••• ; ••.•. ,~ •• ; •••••. ,:.:.: •.•••••••••••• '.' •.• ' •.•••.•.• -•.••• -.; •• ; •••.•••. ';.. • ............ ;.;0;.:.;;;.:.;" ••••••••• -' ••• -.-.-.-. 0;' ., •••••••••••••••••••••• - ••• ~ •• ,-••

1600~

PRESSURE (mber)

FIG. 10. A lidar measured atmospheric pressure profile (duts) compared with radiosonde-derived pressure (X 's) and least-squares fit to radiosonde data. 100 shot average, 45-m resolution.

lOis a plot of the pressure profile (dots) calculated from upward-looking lidar signals using Eqs. (5) and (7). These data were taken at 20:30 EST on 25 March 1983, at Goddard Space Flight Center (lat. 39.00° N, long. 76.85° W) using the absorption trough at 13153.8 cm '. The signals from which pressure was derived were averaged over 100 shots and 45-m altitude resolution to improve the signal to noise. The X's are radiosonde pressure data taken 1.5 h earlier at Dulles Airport Oat. 38.98°N, long. 77.46°W) for comparison. The solid line is a linear least-squares fit to the radiosonde data, and used to calculate the average deviation between the two measurements, which was 0.3%. The calibration constant in Eq. (5) was determined by a fit of the Iidar data to the solid line.

1110 SYSTEM SOFTWARE AND OPERATION

The lidar system is controlled by the system micropro­cessor with operator input via the console keyboard. There is a separate control panel to adjust the detector gating param­eters and laser separation, and a laser tuning control panel to operate the on-line laser. All data system functions are con­troned by the system software with the console keyboard acting as a control paneL

The software is interactive and provides real-time dis~ plays on the video terminal to assist the operator. System operation begins with the display of a menu of adjustable system parameters on the screen (see Table 1). The operator may select from among the following options: one or two laser transmissions, active (laser firing) or passive (calibra­tion) modes, one or two receiver channels (and, if one, which channel), output to be displayed on the screen only or also recorded onto tape, digitizer sampling rates of 5, 10, or 20 MHz, automatic background subtraction mode enabled or not, and an upper limit on the number of shots, if desired. For each parameter, the default value is highlighted on the screen, and to change the current value, a unique single code letter (generally, the initial letter of the parameter name, as indicated by underscoring on the screen) is input on the operator console. A command interpreter in the software is used to toggle between the permitted options for each pa­rameter. When the choice of parameter values is completed, an exit code is pressed on the console to begin laser oper-

Lid,n system 2233

····.··-.·,··-·-··-······N.·.--=.·.·.·.-;::,-.·.-.-.•.•.•. ' .•.•...•...• r." •••••.• - •. ~.-.-.-.-.,.,", ••••••••••••.•• :!'."

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

TABLE I. Lidar data-acquisition system menu of parameters available for selection prior to data acquisitiono The system operator toggles between op­tions by pressing the corresponding code letter on the console keyboardo

Parameter name Code letter Options (default value underscored)

Transmitter T One laser, Two lasers Mode M Active (laser firing),

Passive (calibrate) Receiver R Channell, Channel 2,

Both c.hannels Output 0 Display, Record on tape Sample rate/ S 20 MHz, 10 MHz, 5MHz lidar range Background B No, Yes subtraction Number of shots N ~ Enter value" Exit from menu E

• A nonzero positive integer (less than 32 K) may be entered to indicate requested number of shotso

ation, and the selected parameter values will remain invar­iant throughout the laser firing sessiono

The lidar control module then triggers the lasers, and the program acquires and records the data, ifthis option was selected. A real-time parameter menu appears on the termi­nal screen containing interactive options which may be modified during the firing of the lasers (see Table II)o As­suming the Edar data are being recorded, input ofthe appro­priate code from the console will cause data recording on the magnetic tape to pause until the code is pressed again, there­by resuming recording. A different code allows the opening or closing of a new file for storing data on the tape. Another parameter code affects the choice of method for laser fre­quency control, viz., either open loop tuning or, in the fu­ture, locking to a wavemeter which is being integrated into the system. Two additional codes permit real-time displays on the terminal screen (replacing the parameter menu) of either gas cell absorption or atmospheric absorption, as de­scribed below. Repeating the display codes returns the pa­rameter menu to the screen with all current options indicat­ed. When data taking has been completed, an exit code input from the console terminates the lidar data-acquisition pro­gram.

TABLE 110 Lidar data-acquisition real-time parameter menu. Options are selected by pressing the corresponding code letter on the cOllsole keyboard while data acquisition is in progresso

Parameter name Code letter

File F Tape T Gas cell absorption G display Atmospheric A absorption display Laser frequency L control Exit from program E

Options (default value underscored)

Close current file. Open new file Pause (leave file open), Continue On, Off

Off, Channell, Channel 2

Open loop tuning, Lock to wavemeter

2234 Rev. Sci.lnstrum., Vol. 58, 1110.12, December 1981

The gas cell or atmospheric transmission display (see Fig. 5) shows the real-time transmission for each laser shot versus shot numbec Up to 512 shots can be displayed across the screen. A linear recursive digital filter with adjustable weighting is used for averaging the transmission data. Input­ting the numeral n on the console keypad will produce the weighting coefficient (1 - 2 - n) to be associated with the previous averaged value and the weighting coefficient 2 - n

to be associated with the current parameter value, where n is in the range 0-90 The digital filter used for averaging is of the form:

( 10)

for n = 0,1,2"00,9, wherey are the averages and x the current data values. Thus higher values of n result in more "smooth­ing" of the grapho

Various diagnostic and informational data are also shown and updated on the screen in real time. The current time and the elapsed time from the initiation of measure­ments are given as well as the total number of shots fired in the session and the number of shots currently displayed on the screen, which is refreshed after every 512 shots. The des­ignations, "Ll," "L2," "HI," and "H2" refer to the respec­tive channel of laser and background data from the energy monitor. Average values of the energy monitor, the gas cell monitor, and the gas cell reference photodiode signals are displayed using the same recursive filter as for the transmis­sion graph. The range limits indicate the range elements over which the on-line and off-line signals are averaged for the purpose of plotting atmospheric (rather than gas cell) trans­mission. These limits, nominally one to 50 range bins, can be adjusted using the cursor control arrow keys on the terminal consok Also computed are the gas cell output to input ratio for each laser, and the final transmission values in percent. The information on the display may be recorded onto hard­copy by the terminal printer at any time by pressing the code letter P on the consok

The ground-based !idar system also can be used in a horizontal measurement configuration which facilitates la­ser alignment and divergence measurements, and allows horizontal path absorption measurements to be made using the large signal scattered from a hard target The horizontal hardware consists of two major parts: one part is a flat 30-cm-sq mirror mounted on a small tower which can be placed over the roof hatch to direct the transmitted laser beam hori­zontally, and which also directs backscattered laser light into the telescopeo The second part is a 13-m-high tower with a 1.5-m-sq target covered with retrorefiective encapsulated glass bead sheetingo The target tower is located 850 m from the laboratory, and the laser beam traverses a level horizon­tal path about 705 m above the highest ground along the patho A television camera with a sma111-m .focal length tele­scope is used to boresight the target and facilitates alignment of the laser beams with one another and the receiver tele­scope. Preliminary alignment of the two laser beams through the system is made by steering He-Ne laser beams, pre­aligned with each pulsed laser, through two iris apertures centered on the system optical axis. The combined beams are then aligned to the telescope optical axis using a large corner

Lidar system 2234

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

cube temporarily placed on top of the receiver telescope, and an aluminum disk, with cross hairs scribed on its surface and painted white. is placed in the telescope focal plane. The corner cube is used to direct the laser beams into the tele­scope parallel to the direction of transmission. Beam steering mirrors following the beam expander in Fig. 3 are then used to steer the combined laser beams onto the telescope optical axis by centering the beam on the cross hairs scribed on the target disk placed at the telescope focus. This apparatus fa­cilitates alignment of the laser beams to within 0.25 mrad of the telescope optical axis. Even finer alignment is achieved using the horizontal targeting system by observing the pulsed laser beam spots on the outdoor horizontal target with the system boresight camera. This procedure anows us to align the beams to within 50-100 Itrad of each other and the telescope optical axis.

tv. LASER SPECTRAL CHARACTERISTICS

Atmospheric pressure is obtained by measuring the ab­sorption of the on -line laser in a broad trough region between strong absorption lines in the 760-nm band, whereas atmo­spheric temperature is obtained by measuring the absorption of the on-line laser centered on a narrow absorption line. Thus for the temperature measurements, the laser band­width and frequency jitter must be sufficiently small to pre­vent large errors caused by some or an of the laser radiation occurring far enough from the line center to experience low­er than predicted absorption. The width (FWHM) ofthese lines at sea level atmospheric pressure 1s -0.1 cm-- I

• With increasing altitude (decreasing pressure), the linewidth de­creases until it approaches the Doppler limit of 0.03 cm --1

FWHM at zero pressure. Thus for ground-based and low­altitude measurements, the laser bandwidth should be small with respect to the linewidth, e.g., =0.02 cm- \ and the frequency stability must be approximately a factor of 10 smaner than the FWHM, e.g., =0.003 - 0.005 cm-- 1

• For higher altitude measurements, the requirements are even more stringent. In addition, measurements with strong ab~ sorption are extremely sensitive to any laser radiation out­side the nominal laser bandwidth (e.g., due to amplified spontaneous emission), since this radiation would not be absorbed by the oxygen line. For example, if 1 % of the laser energy is broadband, the actual backscattered signal would be twice the expected backscatter signal (100% error) for a measurement altitude for which 99% of the narrowband en­ergy has been absorbed by the oxygen absorption feature.

Because of the potentially large errors caused by the spectral properties of the lasers,23.24 we performed a series of measurements to characterize their spectral structure. We have made precise measurements of the spectral purity, linewidth, line shape, and shot-to-shot frequency jitter of the lasers, using a high-resolution Fizeau interferometer and a long path mUltipass absorption gas cell.

Laser linewidth and spectral profiles are measured with the experimental apparatus shown in Fig. 11. A collimator is used to expand the laser beam and provide the 0.25 mrad or better collimated beam required for the high-finesse 5-cm interferometer. The interferometer plates are aligned with a small angular deviation from paranel to generate several or-

2235 Rev. Sci.lnstrum., Vol. 58, No. 12, December 1987

1024 ELEMENT PHOTODlOOE ARRAY

~ __ ~ INTERFEROMETER

f',.! 0- it - - -11-- ,--- 47~ - LASER ~.1 _~-_- -u-- - _ -JcoLLIMAT<lRJ BEAM

\

CYLINDRICAL \l, 5cm --.J LENS ~

~ ~ ---~------

SIGNAL AVERAGER FIZEAU

FRINGES

FIG. 11. Interferometer spectrum analyzer. Collimated laser light enters a fixed length interferometer with a small angle between its plates to produce parallel Fizeau fringes. The vertically aligned fringes are focused into spots onto a photodiode array with a cylindrical lens, and the spectrum displayed on a storage oscilloscope_

ders of straight-line Fizeau fringes. A cylindrical lens is used to focus the fringe pattern in the direction perpendicular to the fringes in order to increase the intensity of light on a 1024-element linear array detector. The array signal is dis­played on a digital storage oscilloscope and photographed (Fig. 12). The Alex I laser has a three-mode output with a 0.027~cm- 1 spacing between the two outermost modes. Note in Fig. 12 that the modes are not equally spaced. This is probably due to higher-order transverse modes associated with one or more of the longitudinal modes. For a 55-em cavity, adjacent longitudi.nal modes are separated by 0.009 em-t. The individual mode widths of 0.0043 em --I are indi­cative of the interferometer resolution (finesse of 23) and not the width of the laser modes. While this apparatus is good for making qualitative assessments of the detailed laser spectral characteristics, we found a number of problems in­herent in making good quantitative measurements. The most serious problem is spatial inhomogeneities in the laser beam, due to the mUltiple transverse modes of the alexan­dtite laser. This causes a misrepresentation of the relative energy in each laser mode seen by the detector, and in some cases could even cause a mode to be absent in one of the interferometer orders. It is also suspected that various parts of the beam cross section will have a diverse mix offrequen­des if different transverse modes are associated with assort-

_0043 FWHM

FIG. 12. Single pulse output spectrum of alexandrite laser as measured with the interferometer apparatus in Fig. 1 L Two free spectral ranges of the in­strument are visible, each showing three laser longitudinal modes.

Lidar system 2235

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

100

80

60

40

i .!

20 $ ,01 % TRANSMISSiON

FIG. 13. Plot of measured gas cell trans­mission vs frequency made with 761 Torr of 0 0 and a 115-mpathlength. The smooth line is the calculated transmis­sion using line strength and width data from Burch and Gryvnak:" The low value of transmission measured at line center attests to the spectral purity of the alexandrite laser outpu!. , (I.E. lOW SPECTRAL IMPURITY)

~ OOL-____ ~ __ -J ____ ~ ____ -L~~~~~~----~----J-----~----~

-.75 13,142.57

FREQUENCY (cm-1 )

ed longitudinal modes. A way is needed to scramble uni­formly the transverse laser modes and present a uniform or Gaussian intensity distribution to the interferometer. Modi­fying the laser to produce a single transverse or longitudinal mode output reduces the laser efficiency prohibitively. Spa­tially filtering the input to the interferometer using a pinhole would undoubtedly alter the spectrum of the light being sampled,

We are developing a laser wavemeter34 with a high-re­solution spectral analyzer in collaboration with personnel at the University of Maryland. The problem of the spectral content of the laser modes described above is of particular interest to us. Attempts at a solution have used a multimode optical fiber to scramble the laser modes, while also serving as a "pinhole" for the collimator. We also investigated using an integrating sphere as a diffuser, where the exit aperture serves as a "pinhole." However, speckle patterns from both of these devices create additional spatial non uniformities in the beam. Weare currently investigating the use of a larger photodiode array and a diffuser to overcome this problem.

We measured the spectral purity of the on-line laser by using a very strong but narrow oxygen-absorption feature as an extremely narrowband notch filter. This was done in our laboratory using a multipass confocal configuration gas cell constructed by the University of Maryland and specifically designed for use with lasers. The cell has a 1-m base path-

TABLE III. Measurements of Alex II spectral impurity inferred from mea­sured transmission through a 149-m path in a multipass gas cell at a fre­quency of 13 142.57 cm- I compared with calculated values.

Pressure (Torr) Transmission (%)

O2 Nz Measured Calculated

305 120 0.1 0,042 340 120 0.05 0.030 380 120 0.033 0.022 500 120 0.013 0.010 620 120 0.007 0.0063

2236 Rev. Sci.lnstrum., Vol. 58, No. 12, December 1987

+,75

length. The beam enters through a small hole in one mirror and exits through a slit which bisects the second mirror. The number of passes is determined by the angle of incidence of the input beam and an angular adjustment of the bisected mirror. We were able to achieve up to 149 passes in the cell. This 149-m path length was verified by a photon propaga­tion time measurement. The transmission of the radiation through the cell was measured using two photodiodes, one monitoring the input beam by way of a beam splitter, and the other monitoring the cell output. We made measurements on the strong oxygen absorption line at 13 142.57 em -I at oxygen partial pressures from 305 to 620 Torr. We scanned the laser through the oxygen line (see Fig. 13 for an example of a line scan), then centered the laser frequency on the line and increased the amplifier gain on the oscilloscope with which we measured signals from the two photodiodes. At the higher oxygen pressures, the absorption line strongly at­tenuates the narrowband radiation while allowing any broadband component to be transmitted relatively unatten­uated. There may be some broadband radiation from the laser that is not well collimated. However, this uncollimated radiation is of little consequence in the !idar application where the field of view of the receiver is narrow. We have found that a large portion of the amplified spontaneous emis­sion from multiamplifier dye lasers can be highly collimated.

In Table III, a comparison of measured values of gas cell transmission with calculated values is given for various pres­sures of oxygen with the cell adjusted for a 11S-m path length. At the highest pressure, the measured transmission implies a maximum possible value for the total broadband energy which may be transmitted through the gas cell. The calculated transmission at that pressure is less than 0.01 % over a bandwidth of 0.027 em - 1 about line center. From this we conclude that the total laser emission outside of the three longitudinal modes observed with the spectrum analyzer is less than 0.01 % of the total laser output energy. At low pres­sures, shot-to-shot instabilities of the laser's mean frequency contribute to values of transmission which are larger than the calculated values, because the absorption linewidth of 0.034 cm -) (FWHM) is comparable to the laser linewidth.

Lidar system 2236

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44

Laser frequency stability was accurately measured by positioning the laser frequency on the steep side of a Doppler broadened oxygen line and measuring gas cell transmission as described above. Any frequency modulation on the side of an absorption line will result in a modulation of the transmis­sion. The frequency stability was determined to have a short­term noise or random component with a standard deviation of ± 0.005 cm - 1. Frequency drifts of much larger magni­tudes were observed over periods of many minutes due to drift in the high-resolution intracavity etalon which is not stabilized for temperature, mechanical drift, piezoelectric creep, or drift in the piezo drive signal which is used to tune the laser. It is believed much of the short-term jitter is due to noise observed on the drive voltage, and we are presently redesigning the laser electronics to correct this problem and to reduce the long-term drift so that longer measurement periods may be utilized to accurately observe changes in the atmospheric temperature or pressure profile.

ACKNOWLEDGMENTS

The authors wish to express their gratitude to the fol­lowing people who contributed to the success of the lidar development: Dr. John Theon and Dr. Robert Curran for their continued support; Leonid Roytblat for constructing much of the electronics; Lou Caudill, Bm Schaffer, Dr. Har­vey Melfi, and Dr. Ed Browell for their learned advice; Gun­ner Sundstrom and Pete Leone for designing the telescope optic mounts, laser beam expanders, and numerous other mechanical designs; Ken Kirks and Ray Di Silvestre for con­tributing their craftsmanship to all the machining; Fred Huegel for the Udar control module and program manage­ment; Lisa Steffens for helping to characterize the PMT, and Hienan Landa for computer programming; Scott Strobel for the multipass gas cell; Max Strange, Jay Smith, Bob Farmer, Tom Ashton, and the late John Semyan for the new laser electronics; Dr. Tom Wilkerson and Leo Cotnoir for devel­oping the laser wavemeter/spectrometer; Arlen Carter, Robert Allen, and Dr. John Degnan for advice on the timing electronics; Roger Navarro and the aircraft crews and sup­port staff at Wallops Flight Facility for helping make our flight program a success; and Patrice Ortiz Cogswell for pa­tiently typing this manuscript.

a) Present address: BOOL: Allen and Hamilton, Inc., Bethesda, Maryland 20814,

b) Present address: RCA Astra Space Diyision, Princeton, New Jersey 08543,

1 The First GARP Global Experiment-Objectives and Plans, GARP Pub!. Ser, No. 11 (World Meteorological Organization, Geneva, Switzerland,

2231 Rev. ScI. (nstrum., Vol. 58, No. 12, December 1981

1973), p. II. 2J. Cooney, J. App], Meteoro!. 11, 108 (1972). 3A. Cohen, J. A. Cooney, and K. N. Geller, Appl. Opt. 15, 2896 (1976). "R, Gill, K, Geller, J. Farina, J. Cooney, and A, Cohen, J, App!. Meteorol. 18, 225 (! 979) ,

5J. A. Cooney, Opt. Eng. 22, 292 (1983). "J. B. Mason, App!. Opt. 14, 76 (1975). 7W. L. Smith and e. M. R. Platt, NOAA Tech. Memo. NESS89, 1977. "I. J. Barton and 1. F. LeMarshall, Opt. Lett. 4, 78 (1979). of{, Shimizu, K. Noguchi, and C,Y. She, App!. Opt. 25,1460 (1986). lOS. F. Singer, App!. Opt. 7,1125 (1968). "e. S, Gardner, AppJ. Opt. 18,3184 (1979). 12G. K. Schwemmer, e. L. Korb, M. Dombrowski, and R. Kagann, 12th

International Laser Radar Conference, Aix,cn-Provence, France, Ab­stracts, 157 (Etablissement d'Etude et de Recherche MHeorologique, Paris, 1984 l.

uG. K. Schwemmer, C. L. Korb, M. Dombrowski, and R. H. Kagann, Topical Meeting on Optical Remote Sensing of the Atmosphere, Incline Village, NV, Tech. Dig. TuC29 (1985).

14C.1.. Korb, G. K. Schwemmer, M. Dombrowski, and R. Kagann, in Tun­able Solid State Lasers for Remote Sensing, edited by R. L. Dyer, E. K. Gustafson, and R. Trebino (Splinger, New York, 1985), p. 35.

L'e. L. Korh, G. K. Schwemmer, M. Dombrowski, J. Milrod, and H. Wal, den, 13th Intemational Laser Radar Conference, Toronto, Ontario, Can­ada, Abst.racts, NASA Conf. Publ. 2431. 52, 1986.

1be. L. Korb, 8th International Laser Radar Conference, Drexel Universi, ty, Philadelphia, PA, Abstracts, 1977.

17c. L. Korb and e. Y. Weng. App\. Opt. 22, 3759 (1983). '"C. L. Korb, G. Schwemmer, M. Dombrowski, and C. Y. Weng, in Optical

and Laser Remote Sensing, Springer Series in Optical Sciences, Vol. 39, edited by D, K. Killinger and A. Mooradian (Springer, Berlin, 1983), p. 120.

19C. L. Korb and e. Y. Weng, 9th International Laser Radar Conference, Munich, Federal Republic of Gerrnany, Abstracts, H!S, 1979.

20e. L. Korh and e. Y. Weng, J. App], MeteoroL 21, 1346 (1982). 21J. E. Kalshovcn, Jr., e. L. Karh, G. K. Schwemmer, and M. Dombrowski,

App!. Opt. 20,1967 (1981). 22R. M. Measures, Laser Remote Sensing Fundamentals and Applications

(Wiley, New York, 1984). 2,C, Cahen and G. Megic, J. Quant. Spectrosc. Radiat. Transfer 25, 15]

(1981). 24G, Megie, Appl. Opt. 19, 34 (1980), 25 A. Ansmann, App!. Opt. 24,3476 (1985). 2~hc effective frequency is that monochromatic frequency which would

have the same absorption coefficient as the spectrally integrated laser out­put. Reference 17 gives an example for a "top-hat" laser spectral profile.

27R, M. Schotland, J. App!. Meteorol. 13,71 (1974). 2"B. p, Ivanenko and 1. E. Naats. Opt. Lett. 6, 305 (1981). 29W, Staehr, W, Lahmann, and C. Weitkamp, Appl. Opt. 24,1950 (1985), 3()N. Mcnyuk, D. K, Ki!linger. and e. R. Menyuk, App!. Opt, 24, 118

(1985), "E. V. Browe!!, A, F'. Carter, S. T, Shipley, R. J. Allen, C. F. Butler, M. N.

Mayo, J. H. Siviter, Jr .. and W, M. Hall, Appl. Opt. 22,522 (1983). 32J, C, Walling. O. G. Peterson, H. P. Jenssen, R. e. Morris, and E. W.

O'Dell, IEEE J. Quantum Electron, QE-16, 1302 (1980). "E. V. Browell, W. R. Vaughan, W. M. Hall, J. J. Degnan, R. D. Averill. J.

G, Wells, D, E. Hinton, and J. H. Goad, 13th International Laser Radar Conference, Toronto, Ontario, Canada, Abstracts, NASA Conf. Pub!. 24:11,6,1986.

341,. J. Cotnoir, T. D, Wilkerson, M, Dombrowski, R, H, Kagann, C. L. Korb, G. K. Schwemmer, and H. Walden, Topical Meeting on Tunable Solid-State Lasers, Arlington, VA, Tech, Dig. FC5, 1985.

"D. E. Burch and D. A. Gryvnak, Appl. Opt. 8,1493 (1969).

Lidar system 2231

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP:

146.189.194.69 On: Mon, 22 Dec 2014 16:18:44