solar irradiance measurements from a research aircraft

Solar Irradiance Measurements from a Research Aircraft

M. P. Thekaekara, R. Kruger, and C. H. Duncan

Measurements of the solar constant and solar spectrum were made from a research aircraft flying at11.58 km, above almost all of the highly variable and absorbing constituents of the atmosphere. Awide range of solar zenith angles was covered during six flights for over 14 h of observation. Results are

presented from nine different instruments which complemented each other in measuring techniquesand wavelength range and were calibrated and operated by different experimenters. A new value of thesolar constant, 135.1 mW cm-', has been derived, as well as a revised solar spectral irradiance curve forzero air mass.

1. Introduction

A group of experimenters from Goddard SpaceFlight Center, National Aeronautics and Space Ad-ministration, undertook a project in August 1967 tomeasure the solar constant and solar spectral irradiancefrom a research airplane flying at 11.58 km (38 000ft). The flights were made from Moffett Field, Cali-fornia. The airplane was fitted out as a high-altitudesolar irradiance laboratory, with twelve different in-struments, seven for spectral irradiance and five fortotal irradiance. The instruments represented a widevariety of measuring techniques and incorporated someof the best available features of precision radiometry.

The solar constant is the amount of total solarenergy of all wavelengths received per unit time perunit area exposed normally to the sun's rays at theaverage distance of the earth in the absence of theearth's atmosphere. The solar spectral irradiance isthe distribution of the same energy as a function ofwavelength. Our present knowledge of these physicalquantities is based mainly on ground-based measure-ments.

Most authors agree that the irradiance of the sunat the average sun-earth distance is in itself highlyconstant. But there is little agreement as to what thatvalue is. Some of the more significant revised valueswhich have been proposed during the last forty yearsare presented in Table I (Refs. 1-9). Most authorsgive the value in units of calories per min per cm2 .Conversion to units of mW cm-2 has been made on theassumption of 4.186 J/cal for the mechanical equivalentof heat. Table I shows that the solar constant hasfrequently been revised and that up to the advent of

The authors (as are the additional authors named in Sec. III)are with Goddard Space Flight Center, NASA, Greenbelt, Mary-land, 20771.

Received 30 September 1968.

space-borne instrumentation most revisions yieldedvalues higher than those previously accepted. Thedisagreements between authors are due largely to thedifficulties of malting measurements through the highlyvariable smoke, dust, haze, and cloud cover of theearth's atmosphere.

The discrepancies between different investigatorsare even greater for the spectral distribution of thesolar energy. Dunkelman and Scolnik'0 have pub-lished charts which show these differences in a strikingmanner. A detailed report on the state of our knowl-edge of the solar constant and solar spectral irradiancehas been published by M. P. Thekaekara."1 In thewavelength range beyond 1.2 A where the atmosphericwater vapor bands are strong, most authors, followingP. Moon, have assumed the solar spectrum to be that ofa 6000 K blackbody. The wide differences, as muchas 40%, between authors in the uv and visible ranges,and the almost complete lack of experimental data inthe longer wavelength range strongly indicated theneed of a new method which is free of the uncertaintiesof ground-based measurements.

II. Description of the High-Altitude SolarIrradiance Observatory

At the flight altitude of 11.58 km the observers wereabove nearly 80% of the permanent gases of the at-mosphere, and more importantly above nearly 99.9%of the water vapor and all the dust and smoke whichform the highly variable and absorbent constituents ofthe atmosphere. The combination of 12 different in-struments complementing each other in wavelengthrange, spectral resolution, and measuring techniques,operated by different experimenters, calibrated withreference to several independent standards and carriedaloft in an airborne laboratory at 11.58 km, practicallyfree of atmospheric absorption, is considered a valuablefeature in enhancing the level of confidence in the finalresult.

August 1969 / Vol. 8, No. 8 / APPLIED OPTICS 1713

Table I. Significant Revisions of the Solar Constant

Solar constantAuthor Year (mW cm-2)

P. Moon' 1940 132.3L. B. Aldrich and G. C. Abbot 2 1948 132.6W. Schuepp3 1949 136.7-141.6C. W. Allen4

1950 137.4L. B. Aldrich and W. H. Hoover5 1952 134.9F. S. Johnson 6

1954 139.5R. Stair and R. G. Johnston7 1956 143.0C. W. Allen 1958 138.0E. G. Laue and A. J. Drummond9 1968 136.1M. P. Thekaekara, R. Kruger, and

C. H. Duncan 1969 135.1

The aircraft was a four-engine jet, NASA 711, ap-propriately named Galileo, commercially known underits model name Convair CV-990. The team for thesolar irradiance measurements consisted of 18 ex-perimenters, scientists, and technicians to operate theinstruments and nine aircraft personnel. An in-flightview of the aircraft is shown in Fig. 1. Prominentamong the special features which make it an airbornescience laboratory are the large extra windows on theroof of the aircraft and the replacement of some of thepassenger windows by optical quality material. Otherspecial features are electrical power outlets at eachexperimenter's station, defrosters for observation win-dows, time code generator, and precise navigationalequipment for the determination of latitude, longitude,and solar zenith angle.

A listing of the instruments in the order in whichthey were located, starting from the front of the air-plane, is given in Table II, as also some basic summarydata on each. The five total irradiance instrumentsare indicated by the word "total" in Column 3, typeof instrument. The names given in the last columnare of those who manned the stations during the flights,though the contributions of many of them extended toseveral other instruments as well, especially during thepreparation for the flights and the data analysis afterthe flights. These instruments were selected inpreference to others, partly because they had all beentested and believed to be sufficiently reliable at levelsof solar irradiance, and partly because they could bereadily mounted and operated on an airplane almostas well as in a laboratory.

The instruments were mounted in the aircraft onaluminum frames specially designed to meet the ratherstrict aircraft specifications for loading and rigidity.For all but two of the instruments the mounting fixtureswere such that the instruments could rotate about ahorizontal axis parallel to the length of the aircraft,and thus track the sun in elevation. The two excep-tions were the Perkin-Elmer monochromator and theCary 14 monochromator for which movable externaloptics were provided to direct the sunlight to the en-trance slit. Tracking the sun in azimuth was ensuredby the aircraft itself flying along a path computed priorto the flights, such that the solar rays were constantlyat right angles to the length of the aircraft.

Six flights were made between 3 August and 19August, 1967. The plan followed for each flight wasto start from Mfoffett Field, go outward to a distantpoint where observations were scheduled to begin,make a U-turn, and return along a flight path, with allthe instruments constantly tracking the sun. The out-bound leg of the flight gave the experimenters time totest the equipment and to make calibration scans whereneeded. It also enabled the aircraft to use up part ofthe fuel and become sufficiently light to maintain a con-stant high altitude for the observations. The altitudeof 11.58 km (38 000 ft) was chosen as a compromisebetween being as high in the atmosphere as possible and

Fig. 1. NASA 711 Galileo, Convair 990A, the airborne labora-tory in flight.

130'W 120W 101W 1OO'W 90W

5_'N

AU.GAS 10 CANADA

5O°N ____1 \ a \r 9°4AUGUS16 \ SIOUX F40

= 5 93 p mod °Z p m f \ UNITED STATES

o FRANCISCO~~~~~~~. IT K14- ~ ~ ~ f

SUU~~~~~~.UA'~~~~~~~~I K~3'

LONGITUDE DEGREES WEST

Fig. 2. Flight paths of NASA 711 Galileo, 3-19 August 1967.

1714 APPLIED OPTICS / Vol. 8, No. 8 / August 1969

Table II. Data on Instruments Used for Solar Irradiance Measurements Onboard NASA 711 Galileo, August 1967

AircraftType of window Wavelength

Instruments Energy detector instrument material range Experimenters

I Hy-Cal pyrheli- Thermoelectric emf Total Infrasil 0.3-4 p McNutt, Rileyometer

2 Angstrom pyr- Resistance strip Total Infrasil 0.3-4 u McNutt, Rileyheliometer 6618

3 Eppley normal Thermoelectric emf Total Infrasil 0.3-4 p McNutt, Rileyincidence radi-ometer

4 P4 interfer- 1P28 or R136 PbS Soleil prism Infrasil 0.3-0.7 A Rogers, Wardometer tube Soleil prism Infrasil 0.7-2.5 p

5 14 interfer- Thermistor Michelson's Irtran 4 2.6-15 p Ward, Rogersometer bolometer bi-mirror

6 Perkin-Elmer 1P28 tube thermo- LiF prism Sapphire 0.3-0.7 a Thekaekara,monochromator couple LiF prism Sapphire 0.7-4 p Winker, Stair

7 Cone radiometer Resistance wire Total Infrasil 0.3-4 p Kruger, Winker,Ward

8 Filter radiometer Phototube Dielectric thin Dynasil 0. 3-1 .2 p Lesterfilm filters

9 Angstrom pyr- Resistance strip Total Dynasil 0.3-4 u Duncanheliometer 7635

10 Electronic scan- Image dissector Grating Dynasil 0.3-0.48 p Webbning spec-trometer

11 Leiss mono- EMI 9558 Q.A. tube Quartz double Dynasil 0.3-0.7 a McIntosh,chromator PbS tube prism 0. 7-1. 6 A Park

12 Cary 14 mono- 1P28 tube Pbs tube Grating and Dynasil 0.3-0.7 p Arvesen, Griffin,chromator prism 0.7-2.5 a Kinney

making observations for as long a time interval as pos-sible.

During each flight, at intervals of 10 min, the flightnavigator made measurements of the latitude andlongitude of the aircraft location. Figure' 2 is a mapof the geographical area covered in the six flights, withthe curved lines showing the route taken during thedata taking portion of each flight. Each flight pathis indicated by the date and the time in Pacific daylightsaving time of the start and finish of data acquisition.

The position data provided by the navigator usingLoran A and doppler radar are considered accurate towithin 10 nautical miles. These data were used in acomputer program to calculate the zenith angle of thesun and the air mass above the aircraft for 1-minintervals. Air mass data are needed to determine thesolar irradiance for zero air mass, that is, in the absenceof the earth's atmosphere. Air mass is defined as theratio of the distance traveled by the rays of the sun inthe atmosphere for a given zenith angle of the sun andaltitude of the observer to the distance for the samealtitude and zenith angle zero. Air mass is unity whenthe sun is vertically above the observer. For otherpositions of the sun, air mass is the secant of the zenithangle in the first degree of approximation. Twoformulas are available for obtaining a more accuratevalue of the air mass for ground-based measurements,one due to Bemporad 2 published in 1907 and a more re-cent one due to Kasten 3 published in 1964. However,for measurements at a high altitude, above nearly 80%of the atmosphere, Kasten's formula is overcorrected.

The value for air mass at 11.58 km which we have usedas being most satisfactory is a weighted average of thesecant of the zenith angle and the value given byKasten's formula, with relative weights in the ratio of4 to 1. Figure 3 shows the variation of air mass withtime on the six days of the flight.

Another factor which has to be considered as a cor-rection factor common to all the instruments, is thedistance of the earth from the sun. The solar constantand spectral irradiance are defined for the mean sun-earth distance, which is the semimajor axis of the earth'sorbit or one astronomical unit (a.u.). During the

Fig. 3. Variation of

20O NOON 2:00 P...

PACIFIC DAYLIGHT SAVING TIME

air mass with timeNASA 711 Galileo.

for the flight paths of


- WINDOW/ N, LI N

WIRE WOUNDCONICAL DETECTOR

~ 0 -- 20.3 cm _ I

Fig. 4. Drawing of the Cone radiometer.

days of the flight, the ratio of the sun-earth distanceto 1 a.u. had the following values: 3 August: 1.0146;8 August: 1.0139; 10 August: 1.0136; 14 August:1.0128; 16 August: 1.0125; 19 August: 1.0119. Itis seen that since the sun was at a distance greater thanaverage during the flight days, the values obtainedfrom our observations have to be increased by per-centages varying between 2.9 and 2.4 to give the solarconstant and spectral irradiance.

A more detailed description of the NASA 711 Galileoproject has been published by the Goddard experi-menters. 4 Copies of the report may be obtained bywriting to the authors of this paper. The report dis-cusses several features of the airborne solar irradianceproject, which for the sake of brevity have been omittedhere: the window material for each instrument, mount-ing of the instruments in the aircraft, flight paths andrelated problems of water vapor above the aircraft,computation of solar zenith angle and air mass, residualgas analysis of water vapor and other absorbents in theaircraft, etc.

Brief discussions of each of the experiments on boardand their results are given below in Sec. III. The sub-sections of Sec. III have been contributed by the ex-perimenters who undertook the main responsibility forpreflight preparation, in-flight measurement, and dataanalysis of each experiment. Of the twelve experimentslisted in Table II, only nine have been chosen for dis-cussion here. The Eppley normal incidence pyrheliom-eter was used only for a short time during one of theflights and the data seemed questionable due to calibra-tion difficulties. The electronic scanning spectrometerwas used on all the flights and yielded extensive data;here again difficulties in calibration made the resultsunreliable. The Cary 14 monochromator was flownby a group of NASA experimenters from Ames Re-search Center, MXoffett Field, California, and forms thetopic of a separate publication.

Ill. The Solar Irradiance Experiments:Instrumentation and Data Analysis

A. Cone Radiometer (R. Kruger)

Description of the Equipment

The cone radiometer is designed to operate as anabsolute total radiation detector using an electricalpower substitution method. It is therefore referencedto the scale of absolute electrical units. The radiom-

eter is designed to operate in a vacuum environmentwhere gaseous conduction and convection may beneglected.

The theory behind the operation is relatively simple.If an electric current is passed through a wire wrappedin the shape of a cone with the base open, the energydissipated will be radiated out through the base of thecone and through the slant conical outer surface. Ifthe slant surface is enclosed in a housing whose tempera-ture is maintained constant, the energy incident fromthis enclosure to the slant surface may be consideredconstant. The energy incident on the base of the coneis a function of the environment viewed by the base.If, after stable conditions are reached, the energyreaching the base is changed, the wire temperature andresistance will change. If these are returned to theiroriginal values by varying the current through the wire,the change in electrical power is then a measure of thenet change in the radiant energy falling on the base ofthe cone.

The detector used in this program was made bywrapping a fine wire (0.002-in. diam insulated nickelwire) on a 30° apex angle conical mandrel. The wirewas then thinly coated with an aluminum-filled epoxy,removed from the mandrel, and painted inside and outwith Parsons' black matte lacquer. The unit had amean base diameter of 9.65 mm. The effective ab-sorbance of the detector was taken as 0.9945 based onthe work of Gouff. 15 This value is open to questionbut is, however, used in this paper. A value of 0.99for a cone would result from the method of Campaneroand Ricolfi.'6 The actual value of the area times theabsorbance used in this report is 0.7364 which includescorrections for the area of the rim of the cone and itsabsorbance (0.9) and a 1.9% correction which is dis-cussed later. If the value of 0.99 is chosen for theconical absorbance and 0.96 for rim absorbance, thearea times absorbance is less than 0.1% different thanthe value used for this report.

The detector is mounted by fine wires at four pointsinside a housing or block which is cooled by liquidnitrogen. The block is made of copper to improveisothermality. Figure 4 is a drawing of the radiometerconfiguration. The block is shaped so as to preventreflection from the front surface into the detector. The

Fig. . Electrical diagram of the Cone radiometer.


2.1' 21302.11 22.10, 125

2.09 " ax

208 !120

2 07 1 t

2.06 h11

2.05 -6

A 1 o

2.04- ALEGEND

2.03 8/3/67 SUNSET 7 POINTS

8/8/67 SUNSET 182.02 8/10/67 SUNRISE 16 105

0 8/14/67 EARLY PM 3

2.01 o 8/16/67 NOONDAY 98/19/67 EARLY PM 14 |

2.00 67 100

1.99 ' I ' I ! . | 0 1 2 3 4 5 6 7 a

AIR MASS -AT 11.58 km ALTITUDE

Fig. 6. Data from the Cone radiometer.

interior is conically shaped to decrease to a negligibleamount that light which passes through the annulusbetween the detector and the front of the block and isreflected onto the slant surface of the cone. The in-terior of the cavity is coated with 3M Velvet flat blackpaint.

The block and detector are enclosed in a vacuumtight chamber which has a sapphire window. Thechamber itself is constructed of stainless steel withViton 0-rings used on the sealing surfaces and Teflonused for both thermal and electrical insulation.

The aperture system is designed in accordance withthe method of Angstr6m.17 A 520 aperture was used.

Figure 5 is a schematic representation of the elec-trical system. The conical detector forms one leg ofthe Wheatstone bridge. Resistor R, is varied accordingto the energy level to be detected. For this series oftests the cone resistance was about 206 Ql which happensto correspond to a temperature of 230 C. A voltageto current converter provides the bridge voltage anddetects the null. Changes off null are sensed by theconverter and the bridge voltage is varied to returnthe system to null.

The data for this paper were hand recorded. A Danadigital voltmeter, model 5600, was used to measure thevoltage across the cone and across the 10-Q precisionresistor to determine current. Data were also re-corded from the Burr-Brown multiplier/divider unit,model 1661, which yields a direct measurement ofelectrical power by supplying the product of voltageand current.

2. Value of the Solar Constants

The data are presented in Fig. 6 which is plotted inthe usual manner of a graph of logioP, where P is thetotal irradiance vs air mass. The derivation of the airmass is presented earlier. Corrections for distance tothe sun have been applied.

Transmittance of the windows was computed the-oretically from the index of refraction adjusted by ex-

perimental data in the absorption bands. Variationof transmittance with angle was also computed. Bothwere compared to laboratory tests and found to bewithin 0.5%.

The method of extrapolation to zero air mass usedherein is believed more valid than passing a least-

A squares fit straight line through the data points.Basically, the curve passed through the points was ofthe form

n

logloP = logo E FriAi-,i=i (1)

where P is the total irradiance in mW cm-2 , i an intervalof bandwidth of the solar spectrum, F the fraction of thesolar energy in interval i, the transmittance of opticalelement in the interval i, A the atmospheric attenuationin the interval i, and m the air mass.

For this analysis, twenty intervals were used. Cor-rections were applied in the first and last terms of thesummation for the part of the solar spectrum unob-served in the uv below 0.295 g due to ozone and in their beyond 4 u due to quartz. These account for 2%of the energy. The result of using an equation of thisform is a concave-up curve which was plotted on trans-parent paper and slid over the data points until avisual best fit was obtained. A better fit might be ob-tained by more elegant mathematical means but it wasnot felt necessary. The y-intercept of the curve isthen the logarithm of the solar constant.

Since a family of curves may be built up by varyingthe value of A in any interval, it was decided to varyA in the interval between 1.2 p and 1.6 ,4 in order toobtain a curve which best passed through all the data.Using this approach, a value of 136.4 mW cm- 2 wasfound as the value of P, the total irradiance.

A second approach was used which took advantage ofthe other available data. In this technique, the valueof A was varied in those three intervals containing the1.4-,g, 1.9-/i, and 2.7-,g water absorption bands. Thedata from the Perkin-Elmer monochromator in thosebands were reviewed for five periods, three on 10 August1967, one on 14 August 1967, and one on 16 August1967. The values were 131.9 mW cm-2 , 134.0 mWcm-2, 133.8 mW cm-2, 135.6 mW cm-2 , and 136.0mW cm-2, respectively. It is of interest to note thatthe readings are arranged in order of decreasing air massgoing, in order, 5.0, 3.6, 2.25, 2.13, and 1.13 and also in-creasing amounts of precipitable water going, in order,9 tu 12 /, 18 bt, 22 ,, and 28 ,u.

3. Sources of Error

The following is a tabulation of sources of error andan estimate of their magnitude; the values are referredto the value of P, the total irradiance. 1.0%-un-certainty in window transmission; 1.0-value of Aused in extrapolation; 0.2-effect of error in time ofday of measurement on air mass; varies from 0.7 atair mass 8 to less than 0.1 at air mass 1.5; the value0.2 was chosen as including most data; 0.1-positionover the ground as determined by Loran-A and dopplerradar; 0.8-zero reading shift due to pressure rise in


,.I"I I I I I -

I

herein is believed more valid than passing a least-GIE3.1

squares fit straight line through the data points.

Basically, the curve passed through the points was of'I5 the form

evacuated enclosure while taking a reading derivedfrom laboratory tests; 0.5-area times cone absorb-ance; 0.5-precision derived from laboratory testsapplying inverse square relationship to energy sources;0.5-light reflected from front face of cooled blockback to aperture tube and then to detector analyticallyderived but could not be measured in the laboratory;0.1-error in measurement (digital voltmeter checkedat Ames Research Center to be within 0.01%).

A number of other sources of error exist which areless than 0.1% and have been neglected. These in-clude the effect of being off-null and the effect of lightpassing through the annulus between the cone and thecooled block. The latter was shown analytically tocause an error in the order of 0.02%.

Applying the method of the square root of the sumof the squares, one arrives at an over-all error in theorder of 1.8%.

One error which has been corrected was a bias of1.9% due to the existence of a slightly rounded edge onthe front face of the cooled block. This reflected lightinto the detector and was noticed during laboratorytests after the flight. The 1.9% figure was arrived atby conducting laboratory tests with the shiny edgepainted black.

A major uncertainty is the absolute accuracy of thecone. For comparison purposes, tests were run with astandard 1-kW lamp source (Eppley Laboratories,Rhode Island, lamp serial ETK-6704). In order tohave sufficient signal at the detector to give reliablereadings, the unit was placed at 100 cm, 75 cm, and 50cm from the source and the calibrated value of 1726-,uWcm- 2 at 200 cm was adjusted by the inverse square re-lation to 6.904 mW cm-2, 12.274 mW cm-2 , and27.616 mW cm-2 at the three distances, respectively.The radiometer indicated 7.070 mW cm-2 , 12.581 mWcm- 2 , and 28.399 mW cm- 2 at the three distances orerrors of +2.4%, +2.5%, and +2.8%. A review ofthe errors does not indicate differences of this order.

4. Results

Since five values have been indicated for the solarconstant in the previous section, a weighting has beenapplied to arrive at a single value. This weighting ishighly subjective. The value of 136.4 mW cm- 2 isthe result of a best fit curve through all the data points.Sixty percent of the final value is derived from thisnumber. The remaining 40% is derived by accountingfor the air mass, so that the proportion of the numbermaking up the final value is inversely proportional tothe air mass. That is, those values derived from thehigher air mass curve fits are given lower weighting.The value of air mass 5 has been deleted because of thelarge extrapolation. Tabulating the data results in thefollowing:

136.4 mW cm-2 X 60% = 81.8 mW cm-2

136.0 mW cm-2 X 17% = 23.1 mW cm-2

135.0 mW cm-2 X 9% = 12.1 mW cm- 2

133.8 mW cm-2 X 9% = 12.1 mW cm-2

134.0 mW cm-2 X 5% = 6.7 mW cm-2

The final value of the solar constant would then be135.8 mW cm-2 2.4 mW cm- 2 or 1.8%o based on theerror analysis.

B. Hy-Cal Normal Incidence Pyrheliometer(A. McNutt and T. A. Riley)

1. Description of EquipmentThe Hy-Cal Engineering normal incidence pyrheliom-

eter consists of a detector at the end of a tube 35.56cm long and 6.99 cm in diam with circular apertures.The acceptance angle is about 2° half-angle. The de-tector surface is a thin metal disk coated with Hy-Calhigh-emissivity graphite coating. The disk is mountedin a cylindrical copper block, 3.81-cm diam and 1.91cm length, which has a quartz window. The spectralrange of the window is 0.2-3.5,u. The output from thedetector is a dc millivolt signal which varies linearlywith incident irradiance up to 2 solar constants. Acalibration curve of millivolt output vs icident radia-tion is supplied with the pyrheliometer. The pyrheli-ometer was calibrated by Hy-Cal Engineering by com-parison to a standard pyrheliometer calibrated withreference to a blackbody cavity.

The output of the Hy-Cal pyrheliometer was mon-itored in two ways. First it was read directly fromthe instrument using a Dana digital voltmeter, model5600, about three or four times during each flight.The rest of the time it was amplified by an operationalamplifier and recorded on one of the channels of theAmpex CP-100 tape recorder. The time, synchronizedto WWV, was recorded on another channel of the sametape, so that a continuous record of signal vs time wasavailable.

2. Data AnalysisIrradiance (mW cm-2) is computed by multiplying

the signal in millivolts by 27.36, a calibration constantbased on the manufacturer's data. The analog flighttapes were sampled five times a second by a computerand averaged over a 10-sec period. The 10-sec averageswere printed out giving a table of signal in millivoltsvs time at 10-sec intervals. The analog flight tapeswere also played into a stripchart recorder giving a con-tinuous graph of the analog signal as a function of time.The stripchart was used to select periods of time whenthe instrument was not sighted on the sun or the signalwas erratic for other reasons. The 10-sec averagesduring these times were omitted. A total of 170 ir-radiance values were obtained this wav. The valuesobtained from the computer print-out agree with thevalues taken at the corresponding time on the planewith the Dana digital voltmeter to within 0.5%.The 170 irradiance values were plotted as a function oftime and showed a scatter of 1.0%. The irradiancevalues were then corrected for window transmissionand the mean distance of the earth from the sun in thesame way corrections were made for the Angstr6mpyrheliometer. The loglo of the corrected irradiancevalues were plotted as a function of the air mass (seeFig. 7). Theoretical curves were fitted to the data


2.13 \

-11302.11 __________________

2.10-125

2.09 - E

2 08 *o' 120 3

2.07 i L

2.06 ---- 115

2.05

2.04 2110 '

2.03 LEGEND "E\Ei8/3/67 41 POINTS

2.02 8/8/67 46 -,1058/10/67 39

2.01 <' 8/19/67 44170

2.00 100

1.99 10 1 2 3 4 5 6 7 8

AIR MASS-AT 11.58 k ALTITUDE

Fig. 7. Data from the Hy-Cal pyrheliometer.

as discussed earlier in the section on the cone radiom-eter.

3. Results

The data are presented in Fig. 7. The interceptfor zero air mass is 2.131 which corresponds to 135.2mW cm-2 . The error is 1.6% or 42.2 mW cm-2 .

C. Angstr6m Compensation Pyrheliometer(A. McNutt and T. A. Riley)

L Description of EquipmentAnother of the total radiation detectors was an

Angstr6m compensation pyrheliometer 6618. It con-sists of two ribbon detecting surfaces located at therear of a tube, 22.86 cm long and 5.08 cm in diam, con-taining rectangular apertures and a shutter. Theshutter allows the left, or right, or both ribbons to beexposed to the incident radiation. The ribbon is lo-cated at about 12.7 cm behind the front of the instru-ment. The front aperture is such that each ribbon hasplane angles of 4.20 and 10.60, as measured from thecenter of the ribbon to the edges of the aperture whichadmits radiation to it. The detector ribbons aremanganin sprayed with Parsons matte black opticallacquer. Each strip has a copper-constantan thermo-couple in thermal but not electrical contact with itand can be heated electrically by passing currentthrough it. Measurements are made by exposing theleft ribbon to the incident radiation and heating theright ribbon electrically until both are the same tem-perature. The current necessary to heat the ribbonis measured. The shutter position is then changed sothat the right ribbon is exposed and the left is electri-cally heated and the current is again measured. Theaverage of the two measured currents and a calibrationconstant gives the value of the incident radiation. Thecalibration constant is furnished with the instrumentand is determined by comparison to a standard pyrheli-

ometer. The auxiliary equipment consists of a Kippand Zonen galvanometer, model AL-1, which indicateswhen the ribbons are at the same temperature and abattery power supply and meter for measuring thecurrent.

2. Experimental ProcedureA measurement was made about every 2-4 min

during a flight and took about 1 min to get both aleft and right ribbon exposure. The time of measure-ment was accurate to the nearest minute. The timeand the current measurements were recorded by hand.

3. Data AnalysisThe irradiance values measured by the Angstrom

pyrheliometer were calculated using

P = Ka2, (2)

where P is the irradiance in mW cm-2 , K the calibra-tion constant, and a the average of left and right valuesof current in milliamps. The calibration constant Kwas checked by A. J. Drummond of Eppley Laboratoryat Table Mountain, California, on 25, 26 August afterthe series of flights, and was found to be 6.08, a 1.0%ochange from the original value of 6.02. The valueK = 6.08 was used in the reduction of the flight data.The current was measured with an accuracy of :1 0.5%,but a plot of irradiance vs time showed a scatter ofi 1.5%/, so the accuracy of the raw data was assumed tobe about 4 1.5%o. The irradiance values were correctedfor transmission of the aircraft window and the meandistance of the earth from the sun. One correctionfactor for the mean distance was used per flight foreach of the six flights. One value for the windowtransmission was used for all flights and all angles ofincidence for both upper and lower windows. Thiswas done since the transmittance changed only 0.5%over the full range of angles of incidence. The methodfor obtaining the window transmission is described inthe cone radiometer data analysis section of this paper.The values of log10 of the irradiance were plotted as a

214 ' - - V 2 1 3 --- . 135

~2\2 10 -1_\_-__.___________ 1_ 3130

210 A _ .-125

2 08 1 - - l20 E

* a .0 a - A206- - - --- .- -- .- -_. -_ _--- _-- -___ H0S

.

20- __ _ .. _- - _- _ ---- 110 D

LEGEND

2.03 8/3/67 19 POINTS

8/8/67 331 _

2.02 8/10/67 32 x1058/16/67 31

2.01 8/19/67 39

154

200-

1.99 _2 3 4 5

AIR MASS-AT 11.58 k ALTITUDE

Fig. 8. Data from the Angstr6m No. 6618.


6 7 8

2.14

2 13

2.12

D

0.1

To

ToCJ

2.11

2.10

2.09

2 08

2.07

2.06

2.05

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

AIR MASS - AT 11.58 km ALTITUDE

Fig. 9. Data from the Xngstrom No. 7635.

function of the air mass. The result is shown in Fig.8.

Instead of fitting a straight line to the data by themethod of least squares, a family of theoretical curveswere generated to see which best fit the data as de-scribed in the cone radiometer data analysis. Theuncertainties in the time and position of the plane whichaffect the air mass value are negligible compared to theuncertainty in the irradiance values. The uncertaintyin the window tranmittance is of the order of 1.0%.There is also an uncertainty in choosing which curveand where to fit it.

4. Results

The best fitting curve is shown in Fig. 8. Theintercept for zero air mass is 2.128 which correspondsto 134.3 mW cm- 2 . The error is 1.9% or 2.6 mWcm-2

D. Angstrom Compensation Pyrheliometer 7635(C. H. Duncan and J. J. Webb)

1. Description of the Equipment

A description of this type of instrument is givenunder the preceding section concerning Angstrom6618. This unit was originally not among the plannedinstrumentation. It was adopted after the first flightand data were recorded for the remaining flights.

1. Data AnalysisAfter the flight program, Angstr6m 7635 was cali-

brated both at Table Mountain and later at the EppleyLaborator'es. The instrument was not stable during

the two days at Table Mountain and A. J. Drummondof Eppley, who was present during the calibration

135 measurements, recommended an average value of 6.17be used for the calibration factor.

A total of 42 measurements were made during thefive flights but the readings of 14 August were dis-

130 carded because of intermittent cloud conditions. TheE remaining data were corrected and the data wereE analyzed in the same way as for the other total radia-

tion instruments.125 J

Z 3. Results

The distribution of the data points is shown in Fig.=- 9. The resulting value of total irradiance was 134.9

120 w rnW cm2

, Which is in good agreement with the otherAngstrom. The errors in the system (particularly inview of the calibration) are on the order of 3% or± 4.0 mW cm-2 .

E. Perkin-Elmer Monochromator(M. P. Thekaekara, R. Stair, A. Winker)


A Perkin-Elmer monochromator model 112, with asapphire window for the aircraft and a lithium fluorideprism as the dispersing element, was one of the in-struments used on board NASA 711 to measure thespectral irradiance of the sun.

A few modifications were introduced to make theinstrument more suitable for the solar irradiance mea-surements. A helical potentiometer with a voltagedivider was attached to the drum to generate a voltagewhich changed linearly with the drum. Anothervoltage divider was connected to the gain controlswitch to generate a step voltage which changed bypreset increments as the gain setting was increased.The output signals from both these as also from the de-tector amplifier circuit were recorded on three of thechannels of a tape recorder. A block diagram of thedetection amplification system is shown in Fig. 10.

| SOURCE ------- CO -- T

* ~~~~~~~~~~~~~~~I i

---- OPTICAL PATH- ELECTRICAL DRIVE------- MECHANICAL DRIVE

Fig. 10. Block diagram of the Perkin-Elmer data system.


° 30 e321E = = a

_ 351 0

30 3iIli 0 2.0 2 20 2 40 2.60 20 330 30 3340 3320 33 2.14 0 4.20 4.4 0 2 326 0 5 0 5 20 5340

Fig. 11. LogD8 deflection of Perkin-Elmer vs air mass.

Modifications introduced at GSFC for recording thedata on magnetic tape are enclosed in heavy lines.

Sunlight falls on a diffuse mirror which is in frontof the slit. The diffuse mirror was prepared by GSFCby evaporating aluminum on a ground glass surface.Tests had previously been made on two other types ofdiffusing surfaces, magnesium carbonate block whichmany observers have used in previous solar irradiancemeasurements,' 0 and gold evaporated on frosted glass.Aluminum was found to have a higher reflectance overthe whole wavelength range of the LiF prism. Sometype of diffusing surface is essential for the solarmeasurements since the conventional method inspectroscopy of focusing the source on the entrance slitwould have limited the observation to a very narrowstrip on the solar disk and our interest is in the ir-radiance due to the whole disk. A plane diffuser waspreferred to the integrating sphere because it increasesthe energy about ten-fold;8 in the 2-4 Au wavelengthrange where solar energy is quite weak, low signal-to-noise ratio would have made the integrating spherewholly unacceptable. A rectangular diaphragm wasmounted a short distance in front of the entrance slit.This enabled a limited and totally illuminated area ofthe diffuse mirror to be viewed by the monochromatorduring the solar scans, regardless of the small pitchand roll of the aircraft, as also during the calibrationscans with reference to the standard lamp. Sunlightentered the aircraft through the upper or side windowaccording to solar elevation. It was reflected to thediffuse mirror by an upper or lower aluminized planemirror carried on a pair of slant rails. The positionand angle of the plane mirror were adjusted betweensuccessive solar scans to center the solar image on thediffuse mirror.

The calibration with reference to the standard lampwas made several times during flight when the aircraftwas going toward the location, and in the hangar.The calibration scans made in the aircraft were sub-ject to a small error owing to light from the roof of theaircraft. This error was corrected later by comparisonwith scans made at GSFC. The transfer function ofthe instrument was highly constant in the photo-multiplier range, but in the thermocouple range, smallfluctuations had apparently occurred owing to a mis-alignment in the instrument.

2. Data Analysis and ResultsThe conventional method of determining irradiance

from spectrum charts is to compare the chart of the

unknown source with that of a standard of spectralirradiance. If the external optics are the same forboth charts, the detector output at any given wave-length is proportional to the irradiance, and the follow-ing basic equation can be applied:

P A.G..P".=AN0G2 (3)

where the subscripts x and s refer to the unknown sourceand standard source, respectively, P is the spectral ir-radiance, A the amplitude (deflection of the pen orsignal on the tape), and G is a multiplication factor toconvert amplitude at a given setting to that at astandard gain setting of 20.

Equation (3) is valid only if the measurements of thesun and of the standard lamps are made under the sameconditions, with the same external optics, and that isnot the case on board the aircraft. The standardlamp shines directly on the diffuse mirror. The sun-light undergoes attenuation by passing through theatmosphere above the aircraft and through the sap-phire window of the aircraft and by being reflected bythe specular mirror. Hence, a correction factor isnecessary to determine what the signal A_ would bein the absence of such attenuation. Thus, A. in Eq.(3) has to be replaced by AzF, where F is the correc-tion factor. F is the product of three factors, F0 forthe atmosphere, F for the sapphire window, and Fmfor the aluminum mirror. F = A -m, where A is theratio of solar spectral irradiance at 11.58 km for airmass unity to that for zero air mass, and mn is theaverage air mass during the measurement scan. F =(N2 + 1)/2N, where N is the refractive index of sap-phire. Fa is a rather complicated function which weshall not write down in full here since it is available instandard literature.'9 It is dependent on n and k, thereal and imaginary parts of the complex index of re-fraction, n + ik, of aluminum, and on the angle ofincidence 0 of sunlight on the specular mirror. Theparameters A, N, n, and k are dependent on wave-length, but are the same for all solar scans; the pa-rameters m and 4 change from one scan to another.

100

30 .35 .4_4____ 6

30

020 -

10

.30 .35 .48 .45 .50 .53 .60

WAVELENGTH- MICRONS

Fig. 12. Atmospheric transmittance vs wavelength based onPerkin-Elmer data.


WAVELENGTH -MICRONS

Fig. 13. Perkin-Elmer data in the 0.3-0.4 2-/A range.

The accuracy of the theoretical values of reflectanceof aluminum and transmittance of sapphire was checkedby actual measurements made on the aircraft and inthe laboratory. During one of the flights, a fewmeasurements were made on the aircraft using thefilter radiometer with and without the plate of sap-phire in front of it. Measurements on the sapphireplate were also made in the laboratory with a BeckmanDK-2A spectrophotometer and the Perkin-Elmermonochromator. The reflectance of the specular mir-ror for various angles of incidence was measured in the0.3-2.5 bt range using a DK-2A with the Gier Dunkleattachment.

Computed values of the attenuation due to theatmosphere are available in literature.2 0 These valuesare valid for the U.S. standard atmosphere. Thirteenof the charts of the flight of 10 August in the photo-multiplier range were analyzed using the manual datareduction method to check the validity of the theoreti-cally computed atmospheric attenuation coefficients for77 selected wavelengths. The values were plottedmanually in terms of logo of deflection vs air mass, andthe points were joined by the best-fitting straight linefor each wavelength. The atmospheric attenuationand the signal for zero air mass were determined fromthe slope and zero intercept, respectively, of thesegraphs. These Langley plots for 8 of the wavelengthsare shown in Fig. 11. Figure 12 gives the values ofatmospheric attenuation at the selected wavelengths.The y-axis in Fig. 12 is the percentage ratio of the spec-tral energy transmitted normal to the altitude of theaircraft to that incident above the atmosphere. Theresults of these computations of atmospheric attenua-tion were used to prepare a revised table which wasused in the computerized data analysis of all the solarscans made by the Perkin-Elmer and Leiss mono-chromators.

The manual data reduction method was used also toobtain the solar spectral irradiance at the 77 selectedwavelengths, based on the Perkin-Elmer scans of 10

August. The detector signals for zero air mass ob-tained from the Langley plots were corrected for re-flection of the specular mirror and transmittance of theaircraft window. The spectral irradiance was com-puted at these wavelengths using Eq. (3) with the neces-sary correction for solar distance.

The manual data reduction method is obviously tootime consuming, and further does not take into accountthe finer details of the Fraunhofer structure of the solarspectrum. Hence, a computer program was preparedfor handing the data recorded on the magnetic tapesand calculating from each spectral scan the solarspectral irradiance for zero air mass. The correctionfactors for the atmosphere, window, and mirror arecomputed at each wavelength, as also the irradianceand signal of the standard lamp by interpolation fromtables supplied to the computer. Spectral irradianceat each wavelength is calculated from Eq. (3). Theintegrated values of the irradiance over bands of desig-nated width are computed. For most of the scans thebandwidth chosen was 100 A for the photomultiplierrange and 1000 A for the thermocouple range. For afew selected scans and over limited wavelength ranges,the integrated area under the spectral curve was com-puted also for bandwidths of 10 A, 5.2 A, and 1 A todisplay more clearly the Fraunhofer structure. Theprogram provides for a print-out of the energy in eachband, the integrated value of the energy up to the endof that band, and for normalization with reference to astandard solar irradiance curve. The standard chosenwas the Johnson curve.6

A total of 71 spectral scans of the sun were thus re-duced. The data from some 30% of the scans couldnot be used for the final average for various reasonssuch as frosting of the aircraft window, cirrus cloudsabove the aircraft, maladjustment of the mirror system,roll of the aircraft, or changing of magnetic tapes duringa scan. The data from the other scans were in fairlyclose agreement.

The final results are presented in Figs. 13, 14, and15. The computer-generated graphs and tables werethe main source for the preparation of these figures.Some of the fine structure was added by comparison

I6 ' I

WAVELENGTH-MICRONS

Fig. 14. Perkin-Elmer data in the 0.4-0.61-u range.


.8 9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 20 2.1 2.2

WAVELENGTH-MICRONS

2.3 2.4 25

Fig. 15. Data from the Perkin-Elmer monochromator and theP-4 interferometer in the 0.7- 2.5-u range.

with the Leeds and Northrup charts. Figure 13 coversthe 0.3-0.42 range, where the Perkin-Elmer hasmaximum wavelength resolution, and Fig. 14 coversthe rest of the photomultiplier range. The data fromthe thermocouple range are shown by Fig. 15 by thecircles with centered dots. The black dots refer to theresults from the P-4 interferometer (discussed later)and the crosses refer to F. S. Johnson's curve.' (Theresults in the range of wavelength longer than 2.3 A areincluded in Table III and in Figs. 27 and 28 whichrepresent a weighted average of all the spectral ir-radiance instruments.)

The wavelength resolution of the Perkin-Elmersystem was determined by two independent methods,first by measuring the half-widths of very narrow emis-sion lines of mercury with different slitwidths andsecondly by comparing the half-widths of the H and Klines of calcium as seen on our Leeds and Northrupcharts and in the Utrecht Atlas.2 ' The two valueswere in close agreement. The half-width of the linesincreases from relatively small values of 2.7 A at 3000A and 4.4 A at 3500 A to 11 A at 5000 A. The wave-length resolution, </AX, where AX is the observed half-width or instrumental profile, decreases from 1100at 3000 A to 455 at 5000 A. In the thermocouplerange, where the slitwidth is twenty times greater, theresolution is considerably less; it increases from 11at 7000 A to 90 at 4 A.

The major sources of error are scattered light fromthe roof of the aircraft reflected onto the diffuse mirrorand the small degree of instability in the positioningof the mirror directly in front of the thermocoupledetector. Detailed analyses of the solar and calibra-tion scans made at different times, both in the labora-tory and in the aircraft, were made to determine themagnitude of this error. It is estimated that the ac-curacy of the spectroradiometric measurements isabout 5%.

F. Leiss Monochromator (C. H. Duncan,R. McIntosh, S. Park, and J. J. Webb)


The following two sections discuss the spectral ir-radiance instruments which were operated by theThermophysics Branch, Spacecraft Technology Divi-sion, GSFC. They are a Carl Leiss double prismmonochromator and an Eppley Filter Radiometer.Each instrument was designed for observing the solarbeam through the upper aircraft windows for solaraltitudes ranging from about 20° to 70°. No auxiliarymirrors were employed; each instrument was mountedon a pseudopolar axis in such a manner that an integrat-ing sphere intercepted a direct beam of solar radiationwhich, after diffusion, completely filled the opticalaperture of each spectral instrument.

The Carl Leiss double prism monochromator wasused for a total of forty solar scans, each from 0.3 to1.6 Au. The relative spectral distributions were in goodagreement with published values and with the filterradiometer discussed later. However, calibration dif-ficulties necessitated a normalization to the energy valueobtained from the total irradiance measurements, 135mW cm-2 .

The Leiss monochromator is a double prism instru-ment designed to provide high dispersion and minimalstray light. The instrument used in this experimentwas equipped with two ultrasil quartz prisms. The

Table Ill. Solar Spectral Irradiance (Based onMeasurements Onboard NASA-711 Galileo at 11.58 km)'

I P Dl

0.140 0.0000048 00000.150 0.0000076 0.000590.600.000059 0.000870.170 0.00015 0.001640.310 0.00035 0.00349

0.190 0.00076 0.007600.200 0.00130 0.01520.205 0.00167 0.02070.210 0.00269 0.02880.215 0.00445 0.0420

0.220 0.00575 .06090.225 0.00049 0.08350.230 0.00667 0.10790.235 0.0059: 0.3132

0.240 0.00650 0.1534

0.245 0.00723 0.17880.250 0.00704 0.20530.255 0.0104 0.23750.260 0.0 30 0.2808

0.2G5 0.0185 0.3391

0.270 0.0232 0.4130.275 0.0204 0.49800.28 0.0222 0.57580.2 5 50.0505 0.67520.290 0.04-2 0.8225

0.295 0.0584 1.0200.300 0.0514 1.2230.305 0.0602 1.4300.310 0.0606 1.6680.315 0.0757 1.935

0.320 0.019 2.2270.325 0.0958 2.5550.330 0.037 2.9250.355 0,1057 :0.:1120.340 0.1050 :.702

0.345:0.1047 4.09010.350'0.,074 4.4830.355,0.1067 4.8790.36010.0055 5.2700.3G5 0.1122 5.674

0.370 0. 173 6,0990.375,0.152 6,.5290.380 10.117 6.9490.380 0.1097 7.3590.59 00.,099 7.765

0.395 0.1191 I.I19 00.400 0.1455 0.675 00.405 0.105I 9.245 00.410 0.1759 9.076 0

0.415 0.178:1 10.5 0

0.420 0.1758 13.39 00.425 0.1705 11.03 00.430 0.1015 12.45 0

0.435 0.1075 13.06 00.440 0.182 13.70 6

0.445 0.1936 14.40 00.450 0.2020 15.14 C0.455 0.2070 15.90 00.460 0.2000 106.66 00.465 0.2060 07.43 6

0.470 0.2045 108.019 0.475 0.2055 18.95 60.480 0.2085 19.720.405 0.0906 20.47 10.490 0.959 21.20 1

0.4950.0966 21.920.500 .1946 22.650.505 0.1922 2.:16 10.510 .1882 24.07 1

0.515 0.1833 24.76 1

0.520 0. 1 25.43 10.525 0.1852 2.12 10.530 0.0842 26.I0 00.535 0.181 27.480.540 0.1783 2814

0.545 0.1754 2I.80 20.550 0.1725 29.44

0.551 .720 0.00.560 0.11i0 5 10.700.5015 0.1700 11,014

0.570 0.0705 31.070.575 0.1710 :12.60 0.580 0.1705 :1:1.2:30.55 0.1700 :03.06.0.590 (1,1005 34.4:)

0.595 0.0665 :5.110.000 0.0 41 :0.5.720.605 0.0620 If6:0:10.6I2 01I.170 :00.110.,.20 1.57 11.1 :

I A

.630 0.1542

.640 0.1517

.650 0.1487

.6,60 0.1468

.670 0.1443

.680 0.1418O.G9. L13918.700 0.0569'.710 0.13144'.720 0.1:014

,.750 0.1290.740 0.12GO0.750 0.12350.800 0.11070.850 '0908

0.900 0.08890.50 0.0055.000 0.074631.1 0.05921.2 0.0404

0.3 0.00:0961.4 0.0336.5 0.02I7I.6 0.02440.7 0.0202

0.0 0.01591.9 0.00260.0 0,01032.1 0.00902.2 0.0079

0,5 0.00602.4 0.00642.5 0.00542.6 :0.004 82.7 0.0045

2.0 0.00392.9 0.00350.0 0.0011.1 0.00261.2 0.00226

:0: 0.00102

1,4 0.0010000:1.5 1.01046

:017 1:. 00155

59.2640.9

41.5042.0045.67

44.7345.7S46.0047.8040.79

59.9559.50

69.4274.37

00.60

814.3 2

90.5990.24

91.,S[92.6:1

94.199 4 R

3.9~4.,)

'4.14.2

4.3

4.44.51

4. 7

4.14.95.1)

'i. )

X .0

8.1)91 10

12.()

14,1)

1;. 0

2,). t)

P. | D_ _ _ _ _

0.0011 01 98.9020.00103 98.9821).00095 99.055

.090O70 99. 1 2

0.000780[ 99.1250.000715[ 99.238 7

0.00048 99.484

0.00045 99.440.0005 99.400i0.000485| 99.509

0.00015 99.78

0.00009 99.807

0.00060 99.097611.000175 99.912

0.000025 99.9120.0000 |7 99.95 |

0.000002 99.9620.00000507 99.969

0.0000170 99.95000000120 99.90070.0000050 99.90450.00004 998

0.0 I03 99 9170.00000341 99.9643

0.01000! 99.99930.0000001 99.9095

95.3f

9o..

1)7. 37 1

97.579 i 97.82,198.03 11J8.2141 i

1)8.(il1;i9,.7211

9J. I I I

0X-Wavelength in microns; P-solar spectral irradianceaveraged over small bandwidth centered at X, in W cm- 2 ,rA'; Dxpercentage of the solar constant associated with wavelengthsshorter than wavelength X. Solar constant 0.13510 Wcm-2 .

August 1969 '/ Vol. 8, No. 8 / APPLIED OPTICS 1723

ZI.

caE

cI

0CC:

.60 ---

.50 . P-4 INTERFEROMETER140 _ AX:! oPERKIN-ELMER

.40- I MF.S. JOHNSON DATA

.30-

120 , I__ .____ ,110 .x I

100 _ ..9_

080 .. ]

- - -- I I ~ ~ ~ ~ r I I I I I I I I I

.070

.060

.050

.040

.030

.020

010

.00 I.6 7

._D__

.1

.1

.1

.1

.1

LIMITING

Fig. 16. Block diagram of the instrumental setup of the Leissmonochromator.

three slits (entrance, intermediate, and exit) were setwide enough to provide sufficient energy while main-taining good resolution.

The Leiss was designed as a laboratory instrument.Therefore, some modification was necessary to make itsuitable for an airborne experiment. The mirror andprism mounts were stiffened and the entire instrumentwas enclosed in a rugged, light tight aluminum boxhaving sufficient strength to provide for mountingin a cradle which could be swung by a gearing mech-anism on its longitudinal axis to allow solar tracking.The housing also provided a heat sink and air currentinsulation. Provision was made for thermostatictemperature control through the use of heating coilsmounted inside the housing.

Two detectors were used-an EI 9558 QA photo-multiplier tube for the uv and visible regions, and aKodak Ektron lead sulfide cell for the ir. These de-tectors were mounted an an auxiliary light-tight boxsecured to the primary instrument housing. The de-tectors were mounted so that they could be quicklyinterchanged by rotating a lever. For maximumstability, the PbS cell was thermoelectrically cooled toa constant temperature of approx 0C. The EAItube was chosen for its low noise and dark current, ex-tended red response, and high sensitivity.

The 7.6-cm-diam integrating sphere was coated witha GSFC-developed magnesium oxide (IgO) paint andthen lightly smoked (1-2 mm) with magnesium oxide.The paint was used instead of a thick smoked coatingbecause it was felt that the shock and vibration withinthe aircraft might damage a smoked coating becauseof its very fragile nature. The sphere was positionedso as to provide a diffuse source of illumination for theentrance slit of the monochromator. It was mountedon an axis perpendicular to the plane of the entrance

slit and could be rotated by means of a hand-drivenworm gear to observe either the sun or the standard ofspectral irradiance. This adjustment was at rightangles to that of the main elevating mechanism andprovided the second degree of freedom for tracking thesun. The use of the sphere also minimized the effectsof roll and yaw of the aircraft.

The sphere was positioned immediately behind theaircraft window where it intercepted the direct solarbeam, thus eliminating the need for any mirrors. Thedetermination of the reflectance characteristics of suchmirrors as a function of angle of incidence is oftendifficult and tedious, and the additional frame designrequirements necessary to directly view the sun werefelt justified.

The alignment of the monochromator on the sun wasaccomplished by using a 15.24-cm-long tube with a pinhole and graduated target. This tube was attacheddirectly to the sphere. The operator manually ad-justed the elevation and azimuth angles of the mono-chromator.

The wavelength drum which rotates the prisms wasoperated by a small ac synchronous motor and geardrive at 1 rpm. This made it possible to make onecomplete measurement (0.3-1.6 ) in 5 min (includingthe changing of detectors). Wavelength markers wereproduced on the strip chart recordings by means of acam and relay attached to the wavelength drum.

2. Electronics

The electronic instrumentation for the Leiss consistedof the following components: a Brower model 129synchronous amplifier tuned to 33 Hz, a specially builtchopper and electronics module mounted inside theLeiss following the intermediate slit, a Power Designsmodel 5165 high-voltage power supply, a Leeds North-rup Type G Speedomax recorder, an EG & G thermo-electric module power supply, and various control andmonitoring circuits. The high-voltage power supplyused to drive the photomultiplier tube has a voltagestability of 0.001% resulting in a tube constancy ofbetter than 1%d. Figure 16 shows a block diagram ofthe entire instrumentation system.

Fig. 17. Comparison of Leiss, Perkin-Elmer, and filter data


-.

E3

QI

.200

2N

I~ ~~~~~~~~~~~~~~~~~~ES

_____1, .

.10 I I~~~~~~~86 .00 0.0

.S00 1.000 1.C00

WAVELENGTH - MICRONS

Fig. 18. Comparison of Leiss and Stair(Mauna Loa) data.

S. Calibration

The calibration of the Leiss was accomplishedthrough the use of a 1000-W NBS type of standard ofspectral irradiance.2 2 A certified standard (EppleyNo. 1145) was employed for the basic calibration.Another lamp of the same type was used as a workingstandard and mounted in an aluminum box on theLeiss frame. The lamp current was monitored with acalibrated shunt and digital voltmeter and could beeasily maintained to within 0.01 A. Since the workingstandard was mounted inside an aluminum box, itsirradiance was somewhat higher than that of theprimary standard because of the reflectance of alu-minum and the slightly higher operating temperatureof the lamp. This was very helpful in the uv, wherethe irradiance of these lamps falls off rapidly. Mono-chromator calibrations were performed in flight as wellas on the ground.

4. Data AnalysisData were taken for each of the flights-a total of

forty measurements. However, only data from flightstwo through five have been analyzed. After the firstflight the slit settings were changed significantly toimprove resolution. Therefore, these data are not in-cluded here. The data from the final flight were notused because of an error in flight calibration. Severalother runs were discarded because of cloud cover andinstrument difficulties. The results from the remainingnineteen measurements are presented here.

During the flights, the calibration signal appearedto drift especially in the uv region. All efforts to cor-rect this difficulty in the field were futile. After theexpedition, a large number of measurements were madeto determine the exact cause of this drift, but thus far ithas not been discovered. Since the repeatability of theworking standard was verified, the drift was almostcertainly caused by some heating effect in the Leissoptical system. It is interesting that the drift wasquite reproducible (within 1%) as a function of time.Fortunately, no such drift could be observed in thesolar data, or in any of the data from the Filter Radi-ometer which used similar instrumentation.

Because of these difficulties it was impossible to ob-tain absolute measurements of the solar spectral ir-radiance from these data. Therefore, all of the dataobtained by the Leiss were normalized to the value ofthe total irradiance of the sun as measured by the totalirradiance instruments, 135.1 mW cm-2. Since thevalue represents the total energy of the sun and theLeiss covered only the 0.3-1.6-Iu wavelength range, itwas necessary to normalize the data to the energy inthat range only. For this purpose the values of Gast2 3

were used for the energies above and below. Theseare 2.44 mW cm-2 and 14.2 mW cm-2 . The energyin the 0.3 -1.6-u range is therefore 118.4 mW cm-2 .The technique used for reducing the data was essentiallythe same as for the Perkin-Elmer monochromator.Incidentally, it should be noted that the normalizationvalue 118.4 mW cm-2 is only 0.3% lower than theenergy in the same wavelength range as given in TableIII which is based on a weighted average of all our in-struments.

5. Results

The Leiss results are depicted in Figs. 17, 18, and19. Figure 17 is a comparison of the Leiss, filterradiometer, and Perkin-Elmer. As can be seen, theagreement is good between the Leiss and the filterradiometer. Figure 18 compares the Leiss data to themost recent ground-based measurements of Stair2 4 atMauna Loa, Hawaii. The agreement here is againgood in the uv region. Figure 19 compares the Leissdata to that of Johnson.6 The fact that Johnson'svisible spectrum is higher is due partly to his highertotal irradiance value of 139.5 mW cm-2. The 4.4-mWcm-2 difference occurs primarily in the visible region.

Two interesting features appear on the Leiss data.These are the two humps in the curve at about 0.9 Auand 1.3 A. None of the previously reported data showsthis. However, very little data are available for thisregion of the solar spectrum. Attempts have beenmade to account for these anomalies on the basis of apossible system error, but after careful examinationthis does not seem to be the case.

,25

LEISS

20 _

15

E 0_ _ _

WAVELENGTH-MICRONS

Fig. 19. Comparison of Leiss and Johnson data.


.

200 .400 .600 --0 160 ~U

LIM ITINGDIAPHRAGMS

Fig. 20. Block diagram of the instrumental setup of the filterradiometer.

An estimate of the errors involved in the measure-ments is as follows:

Error i total irradiance (from Angstr6m Pyr.) ±3%Error in irradiance outside 0.3-1.6-/i wavelengthinterval ±t4%Error due to window transmittance + 1%Instrumentation error ±t2%

The total RAIS error is 5.5%.

G. Filter Radiometer (R. Stair, J. J. Webb,D. Lester, and C. H. Duncan)

1. Description of Equipment

A GSFC modified Eppley XIark V filter radiometerwas employed in the design of a photoelectric filterwheel radiometer to make solar spectral measurementsfrom 0.3 to 1.1 g. No commercial instrument of therequired type was available. The thermoelectric de-tector, shutter and shutter drive, and other featuresnot associated with the filter rotation mechanism wereremoved or made nonactive. The filter wheel rota-tion mechanism provided for the step rotation of twofilter wheels, each containing 13 filter positions. Allow-ing for a zero position and for the interposition of twowide band filters or radiation (or isolation) screens, atotal of 22 narrow band filters could be installed on thetwo wheels at one time. Such filters were set up ontwo wheels for use with an RCA type 935 phototube(S-5 response) to cover the uv (11 filters) and visible(11 filters) within the 3100-6000-A range. On a thirdwheel another set of 11 filters was arranged to cover the6000-11 000-A spectral range. This latter set, to-gether with those for the visible range, was used with anRCA type 917 phototube (S-1 response). Thus, toshift the instrument coverage from the uv and visibleand near ir required the changing of one filter wheel andthe interchange of detectors. Since this procedure re-quired a significant amount of time, it was decided touse only one combination during any particular flight.This procedure was followed except that during the

flight of 16 August, the constant air mass flight, bothranges were covered by making the change at midflight.

The narrow band interference filters supplied byEppley were chosen to cover the spectral range men-tioned above. In the uv, each filter has a half-bandwidth of approx 100 A and is centered at an even wave-length, e.g., 3100 A, 3200 A, 3300 A, etc., to 3900 A.Those at longer wavelengths have wider half-bandwidths and in general, are also centered on even wave-lengths spaced at 500 A or 1000 A in the visible and ir.Thus, each filter (except for several wide band filters)covers a portion of the solar spectrum extending ineither direction from its central transmission peak suchthat, ideally, where the transmittance of one filterdrops to one-half its value, the next has increased to one-half its value. The transmittance of each filter wasmeasured on a Beckman DK-2A spectrophotometer.

2. Instrument ModificationsThe Eppley mechanism of the filter wheel was de-

signed for manual change of filters or for automaticchange at very short time intervals of less than a sec-ond. This was modified by adding to each filter wheela timing motor and an automatic switching mechanismso that the sun was viewed through each filter in suc-cession for about 8 sec.

The instrument was mounted in the aircraft in aframe similar to that used for the Leiss, that is, apseudopolar axis mount. As with the Leiss, a manualelevating mechanism provided instrument adjustmentfor solar altitude and a worm gear mechanism rotatedthe integrating sphere. The sphere was of the samesize and used the same coating as the Leiss. A blockdiagram of the instrumental arrangement is shown inFig. 20.

3. Electronics

The primary electronics used with the filter radiom-eter consisted of a Keithley model 610 R Electrom-eter and a Moseley model 7100B strip chart recorder.The current from the phototube was amplified by theelectrometer and directly recorded by the Moseley.No modification of either instrument was required aseach has several sensitivity settings.

4. CalibrationCalibration was accomplished in the same way as

with the Leiss with reference to an NBS certifiedstandard lamp.

5. ResultsThe analysis involved in reducing the filter radiom-

eter data is detailed and somewhat complex and isnot described here. The procedures and techniquesused can be found in a recent Goddard contract report."The results presented here are also taken from this re-port.

The zero air mass solar spectral curve is shown inFig. 17, along with the results of the Leiss and Perkin-Elmer measurements. The comparison with the Leiss


,goR: POLARIZER 2

BEAMSPLTTER

POLARIZER PM r/W 3OL~WAR op \ DETECTOR

COMPENSATOR

Fig. 21. P-4 polarization interferometer schematic diagram.

is very good in all but the regions around 3000 A and6500 A. It is believed that these discrepancies aredue to calibration uncertainties. The fact that theresults of the two instruments agree closely despitetheir differences in optical characteristics seems tostrengthen their validity. Also, the energy scale forthe filter radiometer was obtained from the in-flightcalibrations, not from normalization. Thus it seemsthat the normalization procedure was justified for theLeiss.

An uncertainty of approx 5% is estimated, due mainlyto the inaccuracies of the standard lamp and windowand filter transmittance values.

H. P-4 Interferometer (P. Ward andM. P. Thekaekara)

1. Description of Equipment

The P4 (made by Block Engineering, Inc., Cam-bridge, Massachusetts) interferometer operates in the0.3 -2 .6 -A wavelength where the proper alignment of aMichelson mirror system is difficult to maintain.Relative path retardation between the two componentsis introduced by a Soleil prism. An optical schematicof the instrument is shown in Fig. 21. The incidentbeam is separated by a polarizer into two componentsof equal amplitudes in mutually perpendicular planes.The wave then passes the birefringent Soleil compensa-tor that has a different index of refraction for eachwave component; thus the two components traversethe compensator at different speeds introducing anoptical path difference, B, between the two perpen-dicular components given by

B = d(no - n), (4)

where d is the thickness of the compensator andnon, are the indices of refraction in the 0 and E planes.The path difference is varied by moving the wedge-shaped section as shown in Fig. 21, thus performingthe function of the moving mirror in the Michelson-type interferometer. The radiation is recombined atthe second polarizer and then separated at the beamsplitter with approx 90% of the energy transmittedto the lead sulfide detector and 10% sent to the photo-multiplier tube. The resulting interferograms are

processed by digital computers in order to obtain theamplitude spectrum.

The mounting for the P-4 interferometer was verysimilar to that of the total irradiance instruments. TheP-4 interferometer viewed the sun through an inte-grating sphere and an aperture arrangement which re-stricted the field of view to 5°. The port in the air-craft was made of infrasil quartz.

Data were recorded on analog magnetic tape duringthe flights.

2. P-4 Data Analysis and ResultsSelected portions of the P-4 interferograms have been

reduced.For the analysis of the data from the lead sulfide

detector, fourteen spectral scans of the sun and sevenspectral scans of the standard lamp were used. Eachscan was the result of co-adding about forty interfero-grams and obtaining the Fourier transform of this sum-mation. Of the calibration runs, one was from inter-ferograms made during the flight and the rest fromthose made later at GSFC. The air mass for the solarspectra varied from 1.13 to 5.10. The print out dataof the computerized Fourier transform give for eachfrequency value of the spectrum a value of the spectralamplitude S, and of the logarithm to the base 10 ofS. It was found that the log S values were more con-venient for mathematical computations.

The spectral irradiance P is related to the spectralamplitude S by the equation

logP0 = logPL + logS - logSL, (5)

where the subscripts S and L denote the sun and thestandard lamp, respectively. The values are availableon a relative scale only, since calibration runs weremade at different distance of the lamp from the P-4,and the factors due to angular aperture, etc., could notbe determined with sufficient accuracy. Comparisonof log S_ values from different spectra showed thatwhile the absolute values for any given X varied over awide range (1.8 to 0 on the log scale or a factor of oversixty), the differences between corresponding values ofany two spectra were highly constant, independent offrequency, thus establishing the validity of Eq. (5).The wavelength resolution varies from 15 A at 0.72 uto 170 A at 2.2 ,. For each of the selected wavelengths,the average log SL from the seven calibration scansand the average log S, from the fourteen solar scanswere computed. Since log S, varies linearly with airmass, the average log S, is the value corresponding tothe average air mass, namely 2.719. To determine logS, for zero air mass, the values of log S, for each spectralscan and wavelength were plotted as a function of airmass. Although the points showed a wide scatter, duepartly to the process of normalization, there was a clearindication of the slope of the line of log S.; vs air massincreasing to a maximum at each of the well known H20absorption bands. A significant result is that even inthe so-called atmospheric windows, far from the ab-sorption bands, the slope does not become zero; thereis a continuum absorption which for unit air mass at


~_ 1. . X jl :

~97 c: I II I

o98-

9 5

188 8 . 10 ' 1al 2 13 1 IS 1 17 1 19 2

94

93 2s

92 ku

89

.7 8 .9 IS 0 ~1 3l12 0 3040501601 08092 0WAVELENGTH-MICRONS

Fig. 22. Transmittance of the atmosphere for air mass down to11.6 km in the 0.7-2.-yo range based on P-4 data.

11.58 km is about 0.5%. Figure 22 shows the resultsof these studies. The slope of the lines were plottedas a function of wavelength and a smooth curve wasdrawn through the observed points. The y-axis wasscaled to show the transmittance of the atmospheredown to the aircraft altitude of 11.58 km for normallyincident solar radiation, that is, the ratio of solarspectral irradiance at 11.58 km to that for zero airmass. This curve represents the average effect of the14 P-4 interferometer scans which were analyzed.

3. Analysis of Water Vapor AbsorbanceA comparison of Fig. 22 with the spectrum of the sun

as obtained by ground-based monochromators showsmany similar features. The absorption dips occur atthe same wavelengths. Except for one absorptionband due to O2, the rest are due to atmospheric watervapor. The high wavelength resolution of the P-4shows considerably more detail. At the two wave-lengths near 1.4 ji and 1.9 bt, where the transmittanceat the atmosphere down to ground level is zero, thetransmittance down to the aircraft altitude is about94%. There is also a great difference in the width andtransparency of the atmospheric windows as observedfrom the ground and from the aircraft. The absorb-ance of about 0.570 in these windows as shown in Fig.22 is due to the wings of the H20 bands but more sig-nificantly to the Rayleigh and aerosol scattering.

As stated earlier the transmittance curve of Fig. 22represents an average of several flights. Too few scansof the P-4 have been analyzed to yield a more detaileddata of the variations in the H20 content above the air-craft for different dates and locations. The Perkin-Elmer data anlyzed by the computer and integratedover wide wavelength bands also failed to show thefiner details of H20 absorption. As shown earlier,total irradiance as measured by the cone radiometergave clear evidence of increased atmospheric absorp-tion at certain locations. A more careful study of theLeeds and Northrup charts of the Perkin-Elmer showed

a correlation with the cone data. In order to obtainquantitative data on the amount of precipitable watervapor above the aircraft, a series of measurements weremade in the laboratory to determine the absorptionprofiles of the water bands in the 1.2-3.4-t4 range. ThePerkin-Elmer monochromator was used with the 2-mmslitwidth as for the solar measurements on board NASA711. The source was a 1000-W quartz-iodine lamp.The beam was rendered parallel by a concave mirror,traveled a measured distance in clean, dust-free air ofknown relative humidity, and then was focused on theentrance slit of the monochromator. Measurementswere made for several pathlengths and the absorbancecurves were normalized to 28 /. of precipitable watervapor (NTP). The results are presented in Fig. 23.The temperature and relative humidity in the labora-tory were such that the amount of precipitable watervapor per meter of path length was 7 .

The absorption band near 1.8 ,u was measured on alarge number of the Leeds and Northrup charts madewith the Perkin-Elmer on board NASA 711. The2.7-u band had been covered only on the flight of 16August. Computations based on these measurementsand the laboratory data of Fig. 23 showed that, ingeneral, the amount of precipitable water vapor above11.58 km varied with latitude, not with longitude.Along the path of the transit flight of 16 August, thewater vapor was 28 ,u, while an increase from 10 tonear 28 ,u was observed during the sunrise flight of 10August. On the other hand, very high values of watervapor, above 40 /u, were observed during the early partof the flights of 14 and 19 August.

Cr

0

CL

0.

i:

0s

Z

0

z

0.2

z

0

z

"I

E7

U:

'C

0-

100 _ _ _

94

88 . _ _

82 __ _

---- I -

76Al__

64 _ . A l58 1_ _ _ _ ~ I_ _

1.2 1.6 2.0 2.4WAVELENGTH-MICRONS

2.8 3.2

Fig,. 23. Transmittance of 28 ,u of precipitable water vapor basedon laboratory measurements with a Perkin-Elmer monochro-

mator.


0 1000 2000 3000 4000 5000 6000 7000 8000

I I I I I I I WAVENUMBER - --14 0 8 60 5 40 3.0 2.7

WAELENGTH - MICRONS

Fig. 24. Interferogram and spectrum from the 1-4 interometerpointed toward the sun.

4. Spectral Irradiance from P-4The average value of log S (signal from the sun)

as obtained from the P-4 was tabulated for 120 wave-lengths in the 0.7236-2.2 02-A spectral range; it wascorrected for the attenuation due to air mass 2.719,and by applying Eq. (5), log P, (solar irradiance) wascomputed. These values are on a relative scale sincea limiting diaphragm had been mounted in front of theintegrating sphere of the P-4, and that made an absolutecalibration on board NASA 711 with reference to thestandard lamp impossible. The diaphragm preventedthe P-4 from viewing all of the standard lamp. The rel-ative scale of the solar spectral irradiance was changedto an absolute scale by two methods. The Perkin-Elmer data in the thermocouple range were analyzedusing a manual method at a large number of wave-lengths with due corrections for scattered light fromthe roof of the plane and atmospheric attenuation.The energy in the range of the P-4 PbS detector wasmade to agree with that given by the Perkin-Elmerdata. A further slight adjustment was made when acomposite curve of solar spectral irradiance based onall the spectral measurements was obtained and thearea under the curve was made to agree with theweighted value of the solar constant as given by thetotal irradiance instruments. The normalized valuesof the spectral irradiances at each of the 120 selectedwavelengths are indicated by the black dots in Fig. 15.

1. 1-4 Interferometer (J. F. Rogers and P. Ward)

1. Description of InstrumentThe I-4 interferometer (manufactured by Block

Engineering Inc., Cambridge, Massachusetts) is of theA'Iichaelson type and covers the 2.5-15-A region. Themoving mirror of the Michaelson cube is actuated by avoice coil driven by a ramp function generator so thatits velocity is practically constant through the dataacquisition stroke. The data are in the form of aninterferogram which is subsequently processed by adigital computer in order to obtain the amplitudespectrum.

The instrument viewed the sun directly through anIrtran 4 window in the aircraft and it was mounted onthe same stand as the P-4 interferometer. Outputsignals were recorded on analog magnetic tape duringthe flights.

2. 1-4 Data Processing and AnalysisThe analog signals of the interferograms were sampled

at constant intervals of path retardation and digitizedfor computer analysis. One hundred interferograms insuccession were co-added to improve the signal-to-noiseratio for each spectrum. The Fourier transform andphase relationship of each spectral component werethen found. A typical amplitude spectrum with theassociated phase plot is shown in Fig. 24 in terms ofwavenumber and corresponding wavelength.

At long wavelengths the phase becomes importantowing to detector radiation. The field of view of theinstrument covers an angle of 12 deg. while the sunsubtends an angle of only '2 deg. The ther-mostatically controlled thermistor bolometer operatesat 370 C and therefore radiates energy to the cold skyabout the sun and the window frame. The detectorradiation becomes an increasingly large fraction of thenet energy interchange as we approach the longerwavelengths. At wavelengths longer than 10 thenet energy is directed out, that is, the detector radiatesmore energy than it receives from the sun.

To account for the detector radiation, a series ofmeasurements were taken while the instrument waspointed away from the sun (see Fig. 25). The energycalibration was obtained with a blackbody sourcemanufactured by Infrared Industries, Santa Barbara,California. It was run at a temperature of 1200 Kwith a 1.53-cm diam aperture at a distance of 87 cmfrom the I-4. The effect of detector radiation wasnegligible during the calibration runs since the frontplate of the source had a temperature of 350C.

The transmittance of the Irtran 4 window as afunction of wavelength and angle of incidence was

^2 INTERFEROGRAM

HA 00 68.27 158.53 20480 275.07 340.33 408.60PATH DIFFERENCE 0X101 0

WAVENUMER - -`

14 10 8.0 6.0 5.0 4.0

WAVELENGTH - MICRONS3.0 2.7

Fig. 25. Interferogram and spectrum from the I-4 interferometerpointed away from the sun.


l l

E _ H~20 j 60001E

E _ ~~~NASA 7110 0.2 0.1 ~ ~ ~ ~ (~2 2

_'_Cl l CH

0.01Z

0.0012.6 2.8 3.0 4.0 5.0 7.0

WAVELENGTH-MICRONS

Fig. 26. Solar spectrum at air mass 1.13 from theeter.

I-4 interferom-

absorption bands. The spectral irradiance of the sun,assuming it to be a 6000 K blackbody, is also plottedfor comparison. It can be seen that at long wave-lengths the solar spectral irradiance curve falls underthat of the 6000 K curve. This is of interest because inestimating the solar constant, most earlier observershad assumed the solar spectral curve in the rangebeyond 1.2 to be that of a 6000 K blackbody. It isunfortunate that data at large air masses are notavailable owing to saturation of the instrument whenoperated with a lower attenuation. This problemillustrates the major drawback of Fourier spectros-copy-the results are not immediately available intime to detect and correct such difficulties. However,in the ir region of the spectrum beyond 3 gu, the trans-mittance of the atmospheric windows is nearly 100%,so that an envelope of the reduced curve can be con-sidered a zero air mass curve.

IV. Conclusion (M. P. Thekaekara)

The following are the values of the solar constant asmeasured by four total irradiance instruments on boardNASA 711 Galileo at 11.58 km in August 1967:

Cone radiometerHy-Cal pyrheliometerAngstr6m pyrheliometer (6618)Angstr6m pyrheliometer (7635)

135.8 4t 2.4 mW cm-2135.2 i 2.2 mW cm-2134.3 i 2.6 mW cm-2134.9 4t 4.0 mW cm-2

obtained by viewing the blackbody source with theI-4 through the window. All the data was incorporatedinto a computer program that solved the equation:

[S(BKG)X i S(meas)X1 X I(BB)X 6

Ta,x X S(BB)x (6)

where S(BKG)X is the signal obtained from the back-ground run at wavelength X, S(meas)x the signal whileviewing the sun, I(BB) the irradiance on the aperture ofthe I-4 due to the calibration source, S(BB) the signaldue to I(BB), and Tx is the transmittance of theIrtran 4 window as a function of angle of incidence a,and wavelength X.

The plus or minus sign is used according as the netenergy transfer is directed into the instrument or out ofit, as is shown by the phase. Ideally the phase shouldbe constant except for the 1800 phase reversals;but noise in the signal, slight errors in locating thecenter of the interferogram, and dispersion in one of theoptical paths, will produce departures from constancy.Since the data were automatically reduced, a phase-referencing system was adopted whereby the phase at3.6 /u (in the short wavelength region and removed frommajor absorption bands) was used as a reference. Ifthe phase at this point lay in the region 0-90-deg or270-360-deg, the plus sign was used for all spectralelements that had a phase value in the same range andthe minus sign was used for all elements whose phaselay outside the region.

3. Results of I-4 DataThe reduced data for an air mass of 1.13 is presented

in Fig. 26 showing some of the major atmospheric

These values are in agreement, within the range of theestimated errors of the respective instruments. Anarithmetical average yields 135.05 mW cm-2 . Aweighted average, assuming for each instrument aweight inversely proportional to the estimated error,yields 135.08 mW cm-2 . We accept 135.1 mW cm- 2

as the value for the solar constant. The estimatederror is 4- 2.8 mW cm-2 . Assuming for the mechanicalequivalent of heat 4.186 J/cal, the solar constant is1.936 0.041 cal cm-2 min'. This is 3.3% lowerthan Johnson's value, 139.5 mW cm-2 , but is inter-mediate between two recently published values, 136.1mW cm-2 obtained by Laue and Drummond9 fromX-15 measurements and 133.4 mW cm-2 obtained byMurcray et al.26 from balloon measurements at 31-kmaltitude.

Table III lists values of spectral irradiance of thesun for zero air mass and for the average sun-earthdistance; a solar spectral irradiance curve (see Figs. 27

N 2ASA-711, GALILEO X.20 - .r--- ~ -

E __ N PAF. S. JOHNSON

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.8 2.2 2.4 2.8WAVELENGTH - MICRONS

Fig. 27. Solar spectrum for zero air mass from NASA 711Galileo, 0.2-2.6 ,u, compared with F. S. Johnson curve.


IE

a:

zc'C

.006

.006

NASA-711, GALILEO

.004 F-"%F. S. JOHNSON

.002 1'

2.5 3 0 3.5 4.0 4.0 6.0 8.0 10.0 12.0 14.8

WAVELENGTH - MICRONS

Fig. 28. Solar spectrum for zero air mass from NASA 711Galileo, 2.5-14 A, compared with F. S. Johnson curve.

and 28) displays graphically the values of Table III.The wavelength range covered by our measurementsis 0.3-15 , but for the sake of completeness and forobtaining an integrated value of the solar constant, thetable of values has been extended to 0.14 in the uvand to 20 in the ir. The F. S. Johnson curve isshown by dashed lines. The values given by a weightedaverage of our spectral irradiance instruments werescaled down bv about 0.1%, so that the integratedarea under the curve is 135.1 mW cm-2 as given by thetotal irradiance instruments. The agreement within0.1% between the total and spectral measurementswas, we believe, more fortuitous than warranted by theestimated error of each of the nine experiments.

In the 0.3-2.2-u wavelength range, which accountsfor all but 6.4%/-o of the energy of the sun, our values arebased on four instruments, the two monochromators,Perkin-Elmer and Leiss, the filter radiometer and theP-4 interferometer, though not all of them cover thewhole range. The agreement between the instrumentsis within the estimated errors of each. The differencesare small except in the range 0.4-0.7 jo, where the ratioof the spectral irradiance of the sun to that of thestandard lamp is very high. In this range the Perkin-Elmer values are higher than those of the filter radi-ometer. The normalized values of the Leiss are closeto the latter, but the unnormalized values are close tothe former. In the 0.3-0.75-/, range the value givenfor each wavelength is the average irradiance for a 100-A bandwidth centered at that wavelength. This givesa solar irradiance independent of the detailedFraunhofer structure which each instrument displaysin a different way according to its wavelength resolu-tion. In the range beyond 0.7 A where the Fraunhoferstructure is small and our wavelength resolutionbecomes less, wider bandwidths are used for theaveraging. In the 1.0-5.0-/ range each irradiancevalue is the average over 1000 A. The 2.2-15-/hwavelength range is covered by the Perkin-Elmer upto 4 , and by the I-4 from 2.6 p to 15 /.

It would seem that Table III and Figs. 27 and 28present the first direct and detailed measurements ofsolar irradiance in the wavelength range longer than1.2 . In this range the widely accepted Johnsoncurve, following that of P. Moon, is as stated earlier,that of a 6000 K blackbody. Our data indicate an

effective blackbody temperature which decreases asthe wavelength increases. This is in agreement withmeasurements made at three wavelengths between1.0 g and 2.5 A by Peytauraux, 2 7 at 11 u by Saiedy andGoody,2 8 and at 6 mm by Whitehurst et al.2 9 Ourextrapolation of the curve in the 15-20-,4 range is basedon this decrease of effective blackbody temperature.Representative values of the effective blackbody tem-perature of the sun are 5360 K at 5 u, 5000 K at 15 4and 4950 K at 20 u. In the wavelength range below0.3 A, where the ozone above the aircraft is an almostcomplete absorber, we have used the rocket data of theNaval Research Laboratory as given by Detwileret al.30 for the range 0.14-0.26 /i and by Johnson6 forthe 0.2 6 -0.3 0-M range. Since our observed values in the0.3-0.4 -u4 range are 0.927 times those of Johnson, andDetwiler et al. had adjusted their irradiance values forwavelength less 0.26 to Johnson's scale, we havescaled the NRL values down by 7.3%.

In column 3 of Table III, we have given Dx, the per-centage of the total solar energy associated withwavelengths from 0 to X. The value for the firstentry, 0.140 , is based on Detwiler et al. who give theirradiance down to 0.085 u.

A point by point comparison of our spectral irradiancevalues with those of Johnson shows that our values arelower almost everywhere, especially in the uv andvisible. In two short wavelength ranges in the ir,0.9-1.05 , and 1.35-1.9 A, our values are higher. Thedisagreement between our curve and that of Johnsonis most pronounced in the 0.52-0.70-h range, ours beinglower by as much as 12% at 0.55 u.

Our final table is based on a detailed analysis ofmany sets of data from a variety of instruments.Instrumental errors arise from different sources for eachinstrument and intercomparison of data permits to agreater extent the minimizing of their effects on thefinal weighted average. Hence we estimate that ourspectral irradiance values have an accuracy of + 5%.

A large number of people were instrumental inbringing this project to fruition and must certainly begiven credit. First we wish to acknowledge the effortsof C. Mook of NASA Headquarters and D. C.Kennard of GSFC, who encouraged the project andprovided the necessary administrative assistance, seeingthe project move from concept through approval andthrough a myriad of problems arising during itsexecution. Invaluable members of the GSFC teamwere, besides the co-authors of Sec. III, R. Maichle whohandled the electronic problems and A. Villasenor whoprovided the computer programming assistance.Particular mention should be made of R. Stair, whosesignificant contributions to preflight development of thefilter radiometer and the Leiss and to postflight dataanalysis, were all the more valuable because of hisunique experience in radiometry and solar measure-ments. Special thanks for excellent support goes tothe men of Ames Research Center, M. Bader, Head ofthe Airborne Sciences Office; R. Cameron, Test Co-ordinator; J. Kroupa, Flight Navigator; and the manyothers who helped in every way possible.


l l l

.0008

.0006 -

F. S. JOHNSON

.0004 - V

0002 ~\\\ NASA.711 GALILEO.0002 I

References1. P. Moon, J. Franklin Inst. 230, 583 (1940).2. L. B. Aldrich and G. C. Abbot, Smithsonian Inst. M\isc.

Collections 110, No. 5 (1948).3. W. Schuepp, dissertation; Basel, 1949.4. C. W. Allen, Observatory 70, 154 (1950).;. L. B. Aldrich and W. I. Hoover, Science 116, 3 (19;52).6. F. S. Johnson, J. Meteorol. 11, 431 (1954).7. R. Stair and R. G. Johnston, J. Res. Nat. Bur. Std. 57, 205

(1956).

8. C. W. Allen, Quart. J. Roy. Meteorol. Soc. 84, 362, 307(1958).

9. E. G. Lane and A. J. Drummond, Science 161, 888 (1968).10. L. Dunkelman and R. Scolnik, J. Opt. Soc. Amer. 49, 356

(1959).11. M. P. Thekaekara, SP-74 (NASA, Washington, D. C., 1965);

M. P. Thekaekara, Solar Energy 9, 7 (1965).12. A. Bemporad, Meteorol. Z. 24, 306 (1907).13. F. Kasten, Tech. Rept. 136 (AD610554) (Cold Regions Re-

search and Engineering Laboratory, Hanover, New Hamp-shire, 1964).

14. M. P. Thekaekara et al., Rept. X-322-66-304 (GoddardSpace Flight Center, Greenbelt, Maryland, August 1968).

15. A. Gouff6, Rev. Opt. 24, Nos. 1-3 (1945).16. P. Campanaro and T. Ricolfi, J. Opt. Soc. Amer. 57, 1 (1967).17. A. Angstr6m, Tellus 3, 425 (1961).18. R. Stair, W. E. Schneider, W. R. Waters, J. J. Jackson, and

R. E. Brown, NASA CR-201 (NASA, Washington, D. C.,1965), p. 44.

19. T. S. Moss, Optical Properties of Semiconductors (Butter-worth's Scientific Publications Ltd., London, 1959), p. 7.

20. L. Elterman, in Handbook of Geophysics and Space Environ-ments, S. L. Valley, Ed. (McGraw-Hill Book Co., Inc., NewYork, 1965), pp. 7-14 to 7-35.

21. M. Minnaert, G. F. W. Mulders, and J. Houtgast, Photo-metric Atlas of the Solar Spectrum (D. Schnabel, Amsterdam,1940).

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

23. P. R. Gast, in Handbook of Geophysics, C. F. Campen, Jr.,et al. Eds. (The Macmillan Co., New York, 1960), pp. 16-14to 16-30.

24. R. Stair and H. T. Ellis, J. Appl. Meterol. 7, 635 (1968).23. R. Stair, Goddard Contract NAS 5-9488 (NASA, Washing-

ton, D. C., 1968).26. D. G. Murcray, T. G. Kyle, J. J. Kosters, and P. R. Gast,

"The Measurements of the Solar Constant from High Alti-tude Balloons" (University of Denver, Denver, Colorado,August 1968).

27. R. Peytauraux, Ann. Astrophys. 15, 302 (1952).28. F. Saiedy and R. M. Goody, Monthly Notices Roy. Astron.

Soc. 119, 17 (1959).29. R. N. Whitehurst, J. Copeland, and F. H. Mitchell, J. Appl.

Phys. 28, 295 (1957).30. C. R. Detwiler, D. L. Garrett, J. D. Purcell, and R. Tousey,

Ann. Geophys. 17, 9 (1961).

Letters to the Editor

Letters to the Editor should be addressed to the Editor,APPLIED OPTICS, AFCRL, Bedford Mass. 01730

Dynamic Orientation of Spin 1 Nuclei. Il

J. M. Midha, S. K. Gupta, and M. L. Narchal

Physics Department, Punjabi University, Patiala, India.Received 24 March 1969.

The purpose of this note is to report the results of further theo-retical calculations in continuation with the work reported in twoearlier papers"'2 hereafter referred to as I and II. It has beenpointed out by Jeffries3 that the direct nuclear spin lattice relaxa-tion is as important as others in samples where the hyperfineinteraction is strong enough to lead to appreciable admixture ofstates or where the quadrupolar interaction competes with thehyperfine or nuclear zeeman interaction.

Using the same notations as in I and II, we give in Table I theexpression for nuclear polarization under suitable approximations.Similar expressions for the other orientation parameters and allthe relevant plots are omitted due to the limitatiom of space.

Table I on facing page

The shape and general features of curves relating to variousparameters remain unaltered in the presence of nuclear relaxa-tions. The nuclear relaxations: (1) reduce the polarization onthe average by about 33% in the range A < 1000; (2) do notmuch affect the alignment to a great degree, signal power, en-hancement in the high temperature limit; (3) reduce the align-ment by about 15% in the low temperature limit; (4) reduce theenhancement to about 25-33% of its value in the low tempera-ture limit; (5) increase the Jeffries-Abragam enhancement withnuclear Zeeman interaction dominating, to about three times itsvalue in the absence of nuclear relaxations; (6) reduce the signalpower by about 50% in the low temperature limit.

One of us (S. K. G.) acknowledges the financial assistance givenby C.S.I.R. India and the hospitality of the Physics Departmentof Punjabi University.

References1. K. L. Bhatia and M. L. Narchal, Appl. Opt. 4, 175, (1965).2. K. L. Bhatia and M. L. Narchal Appl. Opt. 5, 1075 (1966).3. Dynamic Nuclear Orientation (Interscience, New York, 1963).


0

solar irradiance measurements from a research aircraft

Documents