Solar Spectral Irradiance Measured from 11.58 km with a LeissMonochromator and a Photoelectric Filter Radiometer

J. J. Webb, C. H. Duncan, R. McIntosh, and D. Lester

Solar spectral irradiance measurements were made from NASA's Convair 990 research aircraft using aLeiss double prism monochromator and a GSFC modified Eppley Mark V radiometer. Six flights weremade over the Pacific and western United States at an altitude of 11.58 km. Excellent agreement isnoted between the two instruments through a wavelength range of 0.3-1.1 u, even though they areoptically and electronically dissimilar. Instrument calibrations were performed in flight using an NBStype quartz-iodine standard of spectral irradiance. Extrapolation to zero air mass was facilitated bythe fact that at 11.58 km the aircraft was above 80% of the permanent gases of the atmosphere and moreimportantly above 99.9% of the water vapor and all the atmospheric pollutants. The resulting solarspectral curves differ significantly from Johnson's in several regions. Solar constant measurements madewith other flight instruments resulted in a value of 0.1351 W cm-2 , which is about 3.3% less than John-son's, but is in good agreement with recently published values of Drummond, Laue, and others.

Introduction

The project to measure the solar spectral irradianceand the solar constant from a research aircraft has beendiscussed in detail elsewhere.' This paper concernsitself with two of the instruments that were a part ofthat project, a Leiss monochromator and a modifiedE]ppley Mark V photoelectric filter radiometer, thatwere the responsibility of Goddard's Spacecraft Tech-nology Division.

Prior to launch, satellites are put through rigorousenvironmental tests, one of which is solar simulation.However, an accurate simulation test is impossible toachieve because of the lack of knowledge of the truevalues of the sun's spectral distribution and total energyoutside the earth's atmosphere. In fact, investigatorshave disagreed by as much as 40% in certain spectralregions. One major cause for this disagreement is thatpractically all the measurements to date have beenmade from ground based stations and thus through thehighly absorbent atmosphere and its various pollutants.It was felt that by making these same measurements ata high altitude this problem could be eliminated and asignificant improvement in the present accuracy couldbe made. At an altitude of 11.58 km, which was thepractical ceiling of the aircraft, the instruments were

The authors are with the NASA Goddard Space Flight Center,Greenbelt, Maryland 20771.

Received 22 July 1969.

above nearly 80% of the permanent gases of theatmosphere, and more important, above nearly all thewater vapor, smoke, dust, haze, etc., so that only theozone remained as a major absorber.

During August of 1967, six flights were made aboardthe NASA owned Convair 990 research aircraft, usingMoffett Field, California (Ames Research Center), as abase of operations. Measurements were made fromspecial observation windows which had been cut intothe fuselage at an angle of 650 from the horizontal.The ports used by the Leiss and filter radiometer werefitted with 2.5-cm thick plates of Dynasil 4000 quartz.

The flight paths are shown in Fig. 1. Their positionsand times were chosen so as to give as wide a variationin air mass as possible for accurate extrapolation to thezero air mass values, i.e., the values in the absence of theearth's atmosphere. They ranged from 1.05 on thetransit flight to 2.6 on the morning flight.

Figure 2 shows a view inside the aircraft, looking upthe aisle. The nearest instrument, on the double rack,is the Leiss monochromator. The middle rack containsinstrumentation and a total radiation detector, and thefurthest rack contains the filter radiometer.

Leiss Monochromator

The Leiss monochromator is a double prism instru-ment designed to provide high dispersion with a mini-mum amount of stray light. The spectral rangecovered was from 0.3 A to 1.6 A, which contains about90% of the sun's energy. Ozone absorption preventedmeasurements below 0.3 y, while the energy above 1.6 piwas attenuated by the integrating sphere.

February 1970 / Vol. 9, No. 2 / APPLIED OPTICS 345

0zV)

00

50'N

40'N

30'N

20'N

130'W 120'W ]10°W 100'W 90'WSTEWART BC 50'N

.10 \CANADA\-- AUGUST 10

.EATTE '

SEATTLE l 1106am SIOUXPORTLAND FALLS 40°N

1 30'p P m o f \ OMAHA 90'W

L SAN I , AUGUST 169:43 FRANCISCO UNITED STATESam. MOFFETT FIELD t DALLAS15:0 6 50 pm,p OmW \5:02'p m-LOS ANGELES - - 30'N

7:40 p.m .4.. C A rk *

MEXICO

AUqULST 3440 p m GUADALAJARA 20N

UST 14 100°WALGUST 19 PACIFIC OCEAN

5.16 p m.2.31 p m 232 p m

130'W 120'W 1 1 W

LONGITUDE DEGREES WEST

Fig. 1. A chart showing the paths of the six flights.

Fig. 2. A view inside the aircraft showing the special mountingsrequired for the instruments.

As can be seen in Fig. 2, the entire instrument wasenclosed in an aluminum housing, which was mountedin a cradle arrangement so that it could be swung aboutits longitudinal axis by means of a crank and gearingmechanism. By simply turning the crank the sun wasmanually tracked as long as the plane was flying at rightangles to the solar vector.

Figure 3 shows the integrating sphere that was usedto provide a diffuse source of illumination for the Leiss.

It was coated with a magnesium oxide paint, developedby Goddard, and then lightly (1-2 mm) smoked withivigO. The sphere was mounted immediately behindthe observation window where it intercepted the directsolar beam, and thus eliminated the need for anyreflecting optics. It was rotatable by means of theknob and gears visible in the photo so that it could beaimed at the sun or turned to view the standard lampshown mounted directly to the main frame. Attachedto the sphere is a sighting tube which has a pinhole inone end and a target in the other, and by maintainingan image of the sun in the target area, the experimenterwas assured that the sphere diaphragm was properlyaligned with the sun.

The detector housing was mounted at the rear of theLeiss and contained two detectors-an EMI 9558photomultiplier tube and a Kodak Ektron lead sulfidecell. They were carefully mechanically stopped so that,while quickly interchangeable, they were always posi-tioned at the same relative spot on the detector surfaces.Also, the lead sulfide cell was thermoelectrically cooledto a constant C for stability.

In-flight calibrations were performed using a 000-WNBS type standard of spectral irradiance.3 Thesecalibrations were made before, after, and in betweensolar runs so that several calibration curves were ob-tained for each flight. An Eppley certified standardlamp was used to calibrate the on-board lamps and thiswas done on the ground in between flights. Also, wave-length markers were produced on the strip charts andcalibrated by the Fraunhofer lines of the sun.

A total of forty solar spectral runs were made duringthe six flights. However, because of cloud cover andinstrumentation problems on particular runs, onlyeighteen of these were used to produce a final curve.

Fig. 3. A close-up view of the Leiss showing the arrangement ofthe rotatable integrating sphere, sighting tube, and standard lamp.

346 APPLIED OPTICS / Vol. 9, No. 2 / February 1970

the filter rotation mechanism was automated, and newdetectors were added. A block optical diagram isshown in Fig. 4. As can be seen, the instrument con-sisted simply of automatically rotating narrow bandfilters in front of either an RCA 917 or a 935 phototubeand amplifying and recording the result on a stripchart recorder. Resolution was naturally not as goodas the Leiss because of the narrow band filters.

In-flight calibrations were used exclusively in thedata reduction. There were no heating effects noticedand therefore no normalization techniques were needed.The data were handled in a manner similar to that usedby Stair in analyzing his Mauna Loa results.8

.22

.18

Fig. 4. A block optical diagram that shows the modified Eppleyfilter radiometer and its associated electronics.

This was done by correcting the individual runs foratmospheric attenuation, quartz window transmission,and solar mean distance, and combining the resultantvalues to produce a final relative spectral distribution.A more detailed description of this procedure may befound elsewhere.4

Unfortunately, in the analysis of the calibration datait was found that they could not be used to establishabsolute values because of a pronounced drift with time.This was apparently caused by a differential heatingeffect in the Leiss due to the close proximity of thestandard lamp. This was very disappointing since itwas felt that in-flight calibrations were an importantpart of the project. However, there was little choicebut to use a normalization technique, and the value thatwas chosen was the total solar energy as measured bythe four total radiation detectors on board the aircraft.There were two Eppley Angstr6m pyrheliometers, anHy-Cal radiometer, and a cone radiometer, and theiraverage value was computed' to be 135.1 mW cm-2 .These detectors were calibrated after the flights withthe Eppley primary instrument, which is frequentlycalibrated in Davos, Switzerland,6 so there is a highdegree of confidence in the result.

Since this value represents the total energy of thesun and the Leiss covered only the 0.3-1. 6-p range, itwas necessary to normalize the data to the energy inthat region only. For this purpose the values of Gast7

were used for determining the energies outside thesewavelengths, 2.46 mW cm-2 below 0.3 p and 14.2 mWcm-2 above 1.6 p. The energy in the 0.3-1.6-A rangewas therefore 118.4 mW cm-2 , and this was the valueused for the normalization.

Filter RadiometerThe other instrument that was used was a photo-

electric filter wheel radiometer which covered a spectralrange from 0.3 p to 1.1 p using narrow band filters.

The basic mechanism of an Eppley Mark V radiom-eter was used in the design of the instrument, in which

zC L

U EZ<~ I~C U

C=-

.14 _

10

06 _

.021_

20 .40

I I

-LEISS

---- FILTER RAD.

'I

I I I I I I - - I

.60 .80 1.0 1.2

WAVELENGTH - MICRONS

1.4 1.6 1 8

Fig. 5. A graph which compares the results of the Leiss and thefilter radiometer.

ResultsThe resulting spectral curves from the two instru-

ments are shown in Fig. 5. The agreement between thetwo is within the estimated 5% experimental error,despite the difference in optics and data reductiontechniques. The two humps in the Leiss curve atabout 0.9 pu and 1.3 p had not been found in any pre-viously reported data. They seem to be real, althoughthe magnitude is questionable. Their presence seemsto indicate that the spectrum in this region is perhapsnot a blackbody as had been assumed by previous ex-perimenters. 9

Figure 6 shows a comparison of the Leiss curve toJohnson's widely accepted NRL curve.' 0 The differ-

za:In

0

ZI

U)

25 I I

.20 -LES .20 ' hi_ =JOHNSON

.15 - I

.10

.05

20 .400 .0 A0 1.0 1 1.0 1 A .0 0 A.400 .600 .800 1.000 1W200 1.400 1600

WAVELENGTH-MICRONS

1.800

Fig. 6. A graph which compares the results of the Leiss andJohnson's 1954 NRL curve.

February 1970 / Vol. 9, No. 2 / APPLIED OPTICS 347

LIMITING-DIAPHRAGMS

- .200

Table I. Solar Spectral Irradiance from theLeiss Monochromatora

0.3100.3150.3200.3250.3300.3350.3400.3450.3500.3550.3600.3650.3700.3750.3800.3850.3900.3950.4000.4050.4100.4150.4200.4250.4300.5600.5650.5700.5750.5800.5850.5900.5950.6000.6050.6100.6150.6200.6250.6300.6350.6400.6450.6500.6550.6600.6650.6700.6750.680

PX

0.08290.11610.10750.10790.11390.10490.09930. 10010.10610.10530.10830.11980.11410.11140.10430. 10010.10470.11720.15240.16860.16970.17140.16780.15950.15330.17050.16920.16880.17010.17160.17090.16850.16610.16380.16140.15890.15640.15470.15330.15160.14960.14740.14510.14300.14220.14280.14300.14210.14080.1393

DX

1.8102.2363.6703.0323.4693.8754.2464.6114.9885.3975.7686.1996.6557.0457.4807.8178.2228.5939.0909.722

10.33910.97911.60812.22212.78930.04730.67631.30031.92632.56033.19733.82634.44535.05735.65836.25236.83537.41037.98138.54639.10439.65440.19540.72841.25541.78242.31242.84143.36543.884

X

0.4350.4400.4450.4500.4550.4600.4650.4700.4750.4800.4850.4900.4950.5000.5050.5100.5150.5200.5250.5300.5350.5400.5450.5500.5550.6850.6900.6950.7000.7050.7100.7150.7200.7250.7300.7350.7400.7450.7500.8000.8500.9000.9501.0001.0501.1001.2001.3001.4001.5001.600

PX

0.16120.17460.18500.19180.19400.19370.19140.18990.19140.18820.18210.18220.18340.18110.18000.17800.17170.16970.17240.17390.17380.17240.17180.17190.17150.13790.13700.13700.13730.13760.13760.13720.13670.13600.13470.13230.12950.12720.12610.11560.10150.09630.09140.07880.06690.05880.05120.04500.03510.02810.0245

DA

13.35813.98414.65215.35416.07316.79117.50818.20918.91419.62620.30820.97521.65822.33422.99923.66724.31724.93925.57426.21626.86227.50428.13928.77629.41344.39744.90545.41245.9246.42946.93947.44847.95548.46148.96349.45849.94250.41750.88555.46859.44562.99866.57469.75572.40974.71078.65282.30285.31687.49589.484

a X is the wavelength in microns, PX is the irradiance averagedover 100- bandwidth centered at X in W cm-', and D,is the percentage.

ences are expected since his total integrated energy is140 mW cm- 2 , while our normalized value is 135.1mW cm- 2 . However, Fig. 6 agrees very well with re-cent experimenters such as Laue of JPL and his X-15measurements (136.1),"1 Drummond of Eppley Labora-tories and his B-57 measurements (136.1),12 and several

workers at the University of Denver3 who reported aneven lower value (133.4) from balloon measurements.

Table I gives the values of the Leiss data, using 100-A intervals for the photomultiplier range and wider in-tervals for the lead sulfide region. The filter radi-ometer data were not included since the resolution wasnot sufficient to allow for averaging.

Discussion

Throughout the project, great care was taken intrying to avoid the problems associated with measure-ments of this type; e.g., no reflecting mirrors were used,calibrations were made in-flight, integrating sphereswere used for diffusion, standard lamp currents wereprecisely controlled, and all the electronics were wellregulated. Nevertheless, the estimated error for thespectral instruments is approximately 5%. One of themain reasons for this is the limitation imposed by theinaccuracy of the present standards of spectral ir-radiance.' 4 It is hoped that better results can be ob-tained once NBS completes work on its new standard 5

now being developed.However, these values are being used in the Goddard

simulator facilities at the present time. In fact, theRAE satellite (Explorer 38), launched in the summer of1968, was tested using simulated solar energy based onthese measurements and has attained temperatures inorbit within a few degrees of those predicted on thebasis of the test. 16

The authors wish to thank Ralph Stair and SegunPark for their contributions to the experimental set-ups, data collection, and reduction, William Cruick-shank for his mechanical designs and stress analyses,and the personnel of the Airborne Sciences Office ofAmes Research Center for their excellent cooperationbefore and during the flights.

References

1. M. P. Thekaekara, R. Kruger, and C. Duncan, Appl. Opt.8, 1713 (1969).

2. P. R. Gast, Handbook of Geophysics (Macmillan Co., NewYork, 1960), pp. 16-14.

3. R. Stair, W. E. Schneider, and J. J. Jackson, Appl. Opt. 2,1151 (1963).

4. See Ref. 1.

5. See Ref. 1.

6. A. J. Drumnmond, ISA Trans. 7, 194 (1968).

7. P. R. Gast, in Handbook of Geophysics and Space Environ-ments, S. Valley, Ed. (McGraw-Hill Book Co., New York,1965), pp. 16-3.

8. R. Stair and H. T. Ellis, J. Appl. Meterol. 7, 639 (1968).9. M. P. Thekaekara, "Survey of the Literature on the Solar

Constant and the Spectral Distribution of Solar RadiantFlux," NASA SP-74 (1965).

10. F. S. Johnson, J. Meteorol. 11, 431 (1954).

348 APPLIED OPTICS / Vol. 9, No. 2 / February 1970

11. A. J. Drummond, J. H. Hickey, W. J. Scholes, and E. G.Laue, Nature 218, 259 (1968).

12. See Ref. 11.13. D. G. Murcray, T. G. Kyle and J. J. Kosters, "The Measure-

ments of the Solar Constant from High Altitude Balloons,"AFCRL-68-0452, University of Denver (1968).

14. H. J. Kostkowski and A. T. Hattenburg, "Development of a

Stable Ultraviolet Source and Techniques for AccurateRadiometry," NBS Rept. 9894, 63 (1968).

15. See Ref. 14.16. C. H. Duncan, "Performance, Systems Designs, Modifica-

tions, Radiometry, Operational Problems, and CorrectiveProcedures Associated with an 1.75 Meter Off-Axis SolarSimulator," NASA-GSFC X-713-69-148 (April, 1969), p. iii.

,3-H,

D. E. Williamson (chairman) CordisJ. J. Ball NBSR. L. Barns BTLP. Baumeister Institute of OpticsL. S. Birks NRLP. D. Call Texas InstrumentsD. E. Carter Eastman KodakG. W. Cleek NBSW. G. Driscoll WorcesterE. F. Du Pr6 NRLL. Eisner Barnes EngineeringV. E. Hamilton Douglas AircraftD. C. Harper Xerox

Reviewing panel

R. L. Hilliard University of ArizonaB. J. Howell Sperry Rand CorporationR. S. Hunter Hunter AssociatesW. L. Hyde New York UniversityS. F. Jacobs University of ArizonaN. S. Kapany Optics TechnologyR. Kingslake Eastman KodakM. Laikin Los AngelesM. S. Lipsett Perkin-ElmerD. J. Lovell Mass. Coll. OptometryE. Manring North Carolina StateT. K. McCubbin Penn State UniversityJ. R. Meyer-Arendt Pacific University

C. F. Mooney RochesterE. J.W. Moore IBMJ. W. Nielsen AirtronE. D. Palik NRLC. B. Rubinstein BTLH. Shenker NRLC. Shepard (consultant) RochesterJ. J. J. Staunton Coleman InstrumentsS. K. Tchejeyan AC ElectronicsL. 0. Vargady Hughes Research LabsP. R. Wakeling (ex officio) WINCW. Wojcik IBMD. L. Wood BTL

The patent number, date of issuance by the U. S. Patent Office, and classification are shown on the first line.The filing date is that of the first application. Any opinions expressed are those of the reviewers and do not re-flect the opinions of the organizations with which they are associated, the Optical Society of America, or thisjournal. Printed copies of these patents may be ordered from the Commissioner of Patents, Washington, D. C.

3,193,676 6 July 1965 (Cl. 250-43.5)Infra-red gas analysers.J. SMART. Assigned to Sir Howard Grubb Parsons & CompanyLtd. Filed 26 Nov. 1962.

This patent describes an ir gas analyzer with the primary featurebeing the incorporation of a method for establishing a reference zerowithout interrupting the flow of the sample gas. This is accomplishedby introducing a zero-setting gas into both beams of the analyzer, ina separate cell, which absorbs all wave lengths absorbed by the sam-ple gas but none of those absorbed by foreign gas also in the system.This zero-setting gas could presumably be the same as that understudy only in higher concentration. R.L.H.

3,316,088 25 Apr. 1967 (Cl. 96-1.5)Process of electrophotography based on electrophotolytic re-actions and element therefor.R. M. SCHAFFERT. Assigned to International Business MachinesCorp. Filed 11 Feb. 1963.

A photoconductor and a reactive layer are sandwiched between atransparent conductor and an electrode. A visible image is formed inone step in the reactive layer, which is decomposed by an electricfield and, in some materials, also by intense illumination. D.C.H.

3,357,830 12 Dec. 1967 (Cl. 96-1.2)Dyed image xerography.W. E. BIXBY. Assigned to Xerox Corp. (mesne). Filed 3 Aug.1961.

A color print is produced by imaging the original through a colorfilter onto a photoconductor, developing, transferring to a dye recep-tive surface, fusing, dyeing, and dissolving the xerographic image.The process is repeated for the other two filters. D.C.H.

3,415,587 10 Dec. 1968 (Cl. 350-3.5)Local reference beam generation for holography.W. T. CATHEY, JR. Assigned to North American Rockwell Corp.Filed 8 Dec. 1965.

A method is described for providing a reference beam for generatinga hologram from a reflected wave from an object. The reflected wavefrom the object is divided into two portions, and one portion is spa-tially filtered by focusing onto a pinhole aperture. This latter portionis used as the reference beam. The requirement for a mirror near theobject and the problem of the coherence length of the source as wellas compensating for the relative motion between the objects and themirror are eliminated. C.B.R.

3,421,011 7 Jan. 1969 (Cl. 250-231)Incremental transducer comprising grating traversed by light raya plurality of times.F. HocK. Assigned to Ernst Leitz G.m.b.H. Filed 6 July 1967.

This invention relates to a measuring device which comprises alight source, an optical system, a grating which is displaceable rela-tive to the optical system, and photoelectric receivers. A techniqueis disclosed for suppressing the dc portions of the position definingsignals by generating a four-phase electrical rotatory field. This isaccomplished by placing a beam splitter behind a beam-doublingelement in the direction of light rays. The additional signal outputsthat are generated can also be used to stabilize the brightness of thelight source. C.B.R.

3,423,522 21 Jan. 1969 (Cl. 178-6)Moving light source generator employing an electromagneticfield.R. ZwicK. Filed 14 Feb. 1966.

An electrically powered light generator which provides electricallycontrolled moving points or sources of light is described. Twoelongated electrodes are used to enable an electric spark to be estab-lished along a substantial length of the electrodes. A moving mag-netic field is used to drive the spark along the electrodes. An exampleof its possible use in video broadcast system is contained in the dis-closure. C.B.R.

3,431,411 4 Mar. 1969 (Cl. 250-43.5)Infrared spectra of powders by means of internal reflectionspectroscopy.N. J. HARRICK. Assigned to North American Philips Co., Inc.Filed 28 May 1964.

This patent is based on a discovery that small solid particles, ifproperly deposited on a light pipe, can be studied using the methodsof internal reflection spectroscopy without serious scattering. Themethod described here is to pass the particles through an electrostaticprecipitator and then attract them to the surface of a charged semi-conducting light pipe which transmits the infrared. R.L.H.

February 1970 / Vol. 9, No. 2 / APPLIED OPTICS 349

Emery