Download - Calibration Changes in EUV Solar Satellite Instruments

Calibration Changes in EUV Solar Satellite Instruments

E. M. Reeves and W. H. Parkinson

This paper reviews the problem of absolute photometric calibration in the extreme uv range with particu-lar reference to a solar satellite instrument. EUV transfer standards, the use of predispersing spectrom-eters, and polarization effects at near normal incidence are discussed. Changes in preflight calibrationassociated with the general problems of contamination are given as the background to the main discussionrelating to changes in photometric calibration during orbital operation. Conclusions relating to ade-quate photometric measurements in orbit are drawn, with a short list of the "best" solar flux measure-ments for reference. Finally, the importance of rocket flights for photometric calibration of satelliteinstruments is indicated.

1. Introduction

A precise knowledge of absolute photometric calibra-tion in the extreme uv (EUV) spectral range is vital forthe detailed interpretation of measurements of plasmadiagnostic parameters in solar, stellar, and laboratorylight sources. A special need arises in the case ofsounding rockets and satellites, because of the infre-quency and cost of observations. For our purposes wehere refer to the EUY range as that extending upwardfrom approximately 300 A, determined by the lowerlimit cutoff of normal incidence optical systems, to ap-proximately 1300-1400 A, which overlaps the lowerlimit of regions accessible with transmitting optics andenclosed photomultipliers. The general problem ofphotometric calibration for solar rocket or satelliteinstruments divides logically into two parts. First, oneneeds to have a knowledge of the absolute response ofthe instrument, consisting usually of mirrors, gratings,detectors, and geometrical parameters. The secondproblem is to maintain the precision of the absoluteresponse through periods of instrument storage and test,launch, and stabilization in orbit. We shall discussthese two phases of activity primarily in relation to theHarvard College Observatory spectrometer carried onOSO-IV, launched 18 October 1967, we shall comparethe calibration changes to analogous changes underlaboratory conditions, and finally we shall discuss therelevant conclusions.

The authors are with the Harvard College Observatory, Cam-bridge, Massachusetts 02138.

Received 6 October 1969.

II. General Background

In the EUV wavelength range, solar instrumentationdivides into two main divisions, namely, normal in-cidence optical systems characterized by angles of in-cidence generally less than 10° and grazing incidenceoptical systems, where the glancing angles are used toincrease the efficiency of the system at the lowest wave-lengths. The latter geometry has been used down to20 A, while the former has been limited to a lower wave-length characterized by Fe xv (284 A) because of thehigh rate at which reflectance in normal incidence sys-tems falls as the wavelength is decreased below 500 A.The problems of in-orbit calibration changes in thesetwo types of systems are markedly different.

Several grazing incidence spectrometers have beenflown for solar observations. These are characterizedby the instruments on Orbiting Solar Observatories,particularly OSO-III and -IV. Of special interest aretwo experiments which viewed the sun continuouslyduring the daylight portion of the orbit and were pro-vided by Hintereggerl and by Neupert et al.2 Both in-struments and results have been described in more de-tail elsewhere.

These two instruments both used Bendix magneticelectron multipliers (different types) and a grating asthe single optical element. The Air Force CambridgeResearch Laboratory instrument "unfortunately ...suffered from a slow but progressive deterioration ofabsolute efficiency, the exact time-dependence andwavelength dependence of which could be assessed onlyapproximately." 3 This steady decrease can readilybe explained in terms of a steady loss of the gain of theelectron multiplier which would cause the pulse heightdistribution of emergent pulses to shift and submerge

May 1970 / Vol. 9, No. 5 / APPLIED OPTICS 1201

below the fixed threshold of the amplification and count-ing system. An order of magnitude drop in sensitivitywas observed over a period of several months, but nochange could be specifically attributed to changes in thegrating efficiency in this trend. By itself this lattercomment would be inconclusive. However, the God-dard Space Flight Center grazing incidence spectrom-eter on the same spacecraft experienced no change ineither grating efficiency or overall sensitivity and after ayear was "functioning without deterioration." Asimilar long term maintenance of sensitivity has beenobserved by the University College London grazingincidence monochromator in the wheel portion of OSO-IV. A single Bendix magnetic electron multiplieraccommodates the positions of He ii ( 304) and L-a(X 1216) and records the integrated solar disk intensityevery few seconds. Rocket flights were used to assessthe maintenance of sensitivity.5

From the in-flight performance of these instruments itappears that optical systems at grazing incidence canmaintain a stable efficiency over long periods of time inan ambient vacuum of 10- Torr, providing the gains ofthe electron multiplier and amplification systems re-main high enough to detect the major fraction of thephotomultiplier pulses.

Normal incidence optical systems for solar EUVobservations have shown significant variations of sensi-tivity during prolonged periods of laboratory storageand in orbit operation. Instrumentation is herecharacterized by the Harvard College Observatoryspectrometer-spectroheliometer (to be described ingreater detail in Sec. III) and the Naval ResearchLaboratory instrument of R. Tousey and his collabora-tors on OSO-V. The NRL instrument consisted of anobjective grating with a series of six Mullard channelmultipliers spanning the 284-1216 A wavelengthrange. The instrument carried a source to check thesensitivity of the detectors separately from the efficiencyof the grating. All of the detection systems showedchanges in sensitivity of a factor of two or more in aboutfive months,6 which can be attributed to a combinationof changes in grating efficiency and a time-dependentgain of the electron multipliers. The sensitivitychanges were different for each of the six wavelengths,which makes it impossible to separate the optical andelectronic effects. The separation of the optical andelectronic effects can be observed in the Harvard OSO-IV experiment data, since stable electronic sensitivitywas achieved in the presence of large changes in opticalefficiency. Since these optical changes are the mainsubject of the present discussion, the instrument datawill be presented in some detail.

Ill. Instrument Description

The Harvard instrument has been described else-where,7'8 together with some preliminary results. Afirst surface concave mirror with platinum coatingforms an image of the sun at the entrance slit of thespectrometer which selects a region of 1 min of arc.The light from this region passes to the diffraction grat-ing, an original Bausch Lomb 1800-line/mm grating

ruled in gold with a first-order blaze at 800 A. Thedispersed radiation passes through an exit slit to atungsten photocathode Bendix magnetic electron multi-plier. Output pulses from the multiplier are amplifiedand counted in a binary mode for transmission to theground in real time or storage on board the space-craft tape recorders for later transmission. Thespectrum at the center of the sun could be scanned bymoving the diffraction grating in rotational increments(in an off-Rowland circle mounting) in units corre-sponding to 0.1 A at the exit slit over the 300-1400-Arange. The capability was also provided to stop themotion of the diffraction grating at any point in the scanto allow a desired wavelength to be positioned on theexit slit. The sail portion of the spacecraft was thencommanded into a raster mode to construct spectro-heliograms within a field of view of 36.5 min of arccentered on the sun.

Figure 1 shows a spectral scan of the center of thesun from 300 A to 1400 A with the more prominent linesidentified. The upper portion of the picture displayssimulated image representations of the data for fourlines at different heights in the solar atmosphere fromthe Lyman continuum at a temperature of approxi-mately 10,000 K, through the transition region, illus-trated by 0 vi 1032 A at a temperature of 300,000 Kand into the corona, with Mg x 625 A at 1.4 millionK and Si xii 499 A at a temperature of 2.5 million K.

IV. Photometric Calibration

To calibrate the Harvard experiment the reflectanceof the primary mirror and certain geometrical param-eters were first measured in separate experiments overthe 300-1600-A wavelength range. The reflectivecoating was a semitransparent platinum coating on aquartz substrate, kindly provided by G. Haas and W.Hunter at Fort Belvoir and the Naval Research Labora-tories, Washington, D.C. A 2.54-cm platinum-coatedreference flat, coated at the same time as the primarymirror, was carried within the instrument at all stagesof handling on the ground prior to launch. The mirrorwas removed and measured at intervals to assess anychanges in the photometric efficiency during the periodbetween instrument delivery and launch. Only a rela-tively small amount of optical contamination was ob-served over several months (see Sec. V).

The spectrometer portion of the instrument, consist-ing of the grating, photomultiplier, amplifier, counter,etc., was then calibrated by illuminating the entranceslit (with the telescope removed) with a monochromaticbeam and measuring the intensity of the light whichpassed through the entrance slit, by means of a tungstenphotodiode. The detection system and calibration arediscussed in greater detail elsewhere.9 As in the calibra-tion of other EUV spectrometers the use of a tungstenphotodiode constitutes the main transfer standard forphotometric calibration. The response of the diodewas here measured with a gas absorption chamber up to1000 A and with a sodium salicylated photomultiplierto 1400 A by J. A. Sampson of the Geophysics Corpora-

1202 APPLIED OPTICS / Vol. 9, No. 5 / May 1970

P a X - _; s , a --n o O iD hi S ., ,,S, ,.E~~

_ I I

. LI n V) 2 L- 2 o Q s

E3M M= N ~ -. ~ . .. . . -. 6

Fig. l. Lower portion: A spectral scan from 300 A to 1400 i at the center of the sun on a compressed scale. Intensities on the ordinatescale are in recorded detection system counts per counting interval. Upper portion: Four simulated solar images in lines formed at

selected heights in the solar atmosphere (see text).

tion of America. R. P. Madden and his colleagueskindly provided measurements of the photoelectricyield of the tungsten diode made by comparison with athermopile at selected wavelengths. The agreementbetween these independent methods was quite good.

In the photometric calibration of all normal incidencegrating spectrometers it is particularly important toconsider with great care the effects of spatial inhomo-geneity in gratings. In general, the emergent beamfrom a grating is not spatially uniform, and, furthermore,the nonuniformities are a function of wavelength. Thisnonuniformity results from the change in shape of thegroove structure over the surface and the change indiffraction properties as the angles of incidence anddiffraction are changed with wavelength. Uniformillumination of a monochromator predispersing gratingby a light source can usually be achieved without muchdifficulty. However, the instrument grating to beevaluated also has a nonuniform emergent responsecharacteristic when illuminated with a uniform beam ofradiation. This response characteristic is the param-eter to be determined, since in orbit a telescope mirror

will produce a spatially uniform illumination of theinstrument grating. Unfortunately, during laboratorycalibration the inherently nonuniform beam that resultsfrom a predispersing monochromator is usually used tofeed the instrument grating, and the resulting responseis the product of the true instrument grating efficiencyin uniform illumination and the convolution of the non-uniform effects of the beams. The second aspect canbe large and is extremely difficult to assess. A mini-mized effect can be achieved if we use the property thatthe efficiency along the length of the grating groove isless variable than that at right angles to the directionof the groove. For this instrument the predisperserand instrument gratings were used in cross dispersion.The focus, as well as angular settings in two mutuallyperpendicular directions, could be remotely controlledfor the predispersing grating. Calibration effects of afactor of two or greater have been measured when thespatial properties of the two gratings were incompatible,either by having their dispersions coplanar, or by usinga grating with marked spatial nonuniformities.

The effect of polarization in normal incidence satel-lite instruments has not previously been reported.


3

1

Io.2

'21

1

0

-1

0

go

Measurements were made on a similar flight instrumentat wavelengths between 500 A and 1300 A and werefound to be less than 3%. These measurements weremade by introducing a three-plate polarizer into theentrance beam of the predisperser grating, remotelyrotating the polarizer through 900, and measuring thecombined response of the predisperser and instrumentgratings. Since it is unlikely that the effect of twogratings in crossed dispersions will be such as to cancelthe polarization effects at three widely separated wave-lengths, the response to polarization can be safelyassumed to be negligible at these near normal incidences(less than 50). If a large response to polarization of themonochromator illuminating beam had been observed,then clearly it would have been impossible to assign apolarization effect to the flight instrument proper.Three-plate polarizers in both the entrance and exitbeams of the predisperser grating would allow such adetermination, although in the wavelength range below1200 A reflectances quickly reduce signal levels whenmultiple optical elements are employed.

The simple expedient often reported in the literaturefor measuring the polarization effects in an instrument,namely, to rotate one instrument through 900 in theemergent beam of a similar instrument, can lead toinvalid results, since the effects of the nonuniform beamprofiles can greatly outweigh the polarization effects.Such effects were observed in the early stages of thecalibration of the Harvard instrument.

V. Calibration Changes Prior to Launch

Although a specific effort is always made to reducethe effects of optical contamination in rocket and satel-lite instruments during the long periods of instrumentintegration and test prior to launch, some changes incalibration are frequently observed. A long series oflaboratory experiments has shown that the reflectance ofmirrors at normal incidence in the 300-1600-A wave-length range is particularly sensitive to small amounts ofcontamination on the surface. The contamination canusually be removed by chemical cleaning in the earlystages, but in later stages, the reflectance change is non-recoverable. A decrease by a factor of two or morecan be observed in the most sensitive region, around1200 A. The monitoring of mirror reflectance at asingle wavelength between 1100 A and 1300 A, com-monly simplest at L-a 1216 A or krypton 1236-1243 A,can be used to assess the degree of optical contamina-tion of mirrors of this type. Since the reflectancechanges are a strong function of the wavelength, whichprobably differs for differing types of contamination,no single wavelength can be used to determine preciselythe change in reflectance over a wide wavelength range.

Contamination of optical surfaces has been the sub-ject of many discussions, including those at a recentconference, Optical Contamination in Space, held atAspen, Colorado, in August 1969. Major areas of con-cern are: (1) particulate contamination; (2) vaporcontamination in vacuum systems; (3) vapor contam-ination from instrument materials; (4) vapor contam-

ination from external sources such as other parts of thespacecraft or spacecraft systems in orbit.

Particulate contamination can be reduced by thecommon procedures of cleanrooms and proper handling.This type of contamination is particularly importantfor coronagraph experiments and less important for uvspectrometers, except for the physical blocking ofspectrometer slits which frequently are of the order of10-100 it or even less. Nevertheless, careful attentionto normal cleanroom procedures produces a significantby-product, namely, ensuring a general level of careand respect for the environment in which the instrumentis placed as well as achieving the principal objective ofreducing the number of residual particles.

Contamination of optical surfaces in vacuum systemshas also been studied in the visible and uv portions ofthe spectrum. With proper attention to monitormirrors throughout the vacuum system, the monitoringof the vacuum properties with residual gas analyzers,and proper vacuum system design through careful atten-tion to backstreaming, the contamination of optical sur-faces in vacuum systems can be reduced to an accept-able level.

The selection of materials for the fabrication of satel-lite instruments is well discussed in NASA publications,and many studies have been done on acceptable mate-rials. Unfortunately, the common criterion for accept-ability of materials for space applications is weight lossas a function of time. From the standpoint of opticalcontamination the parameter of weight loss is rathergross, since reflectance losses result from monolayers ofmaterial on the mirror surface, and even the so-calledacceptable materials exhibit some weight loss with time.Selection of materials, potting procedures, surfacetreatment of metals, trapped gas, lubricants in movingparts, etc., have been variously considered in individualstudies. A good rule of thumb for optical space instru-ments is to place the maximum amount of electronics,etc., outside of the optical cavity so that any outgassingproducts cannot migrate in a direction toward theoptical system, but preferably are directed away fromthe instrument entirely.

In the category of external contamination sourcesin orbit, the problem of normal spacecraft outgassingcontaminations, which can under some circumstancesdrift into the optical instrument, is compounded by theuse of reaction jets to control spacecraft orientationand, in the case of manned experiments planned for thenear future, by waste disposal in orbit. With a properchoice of gases for reaction jets and care in their location,and planned procedures of human waste disposal, theseorbital effects can undoubtedly be reduced. Experi-ments are planned in the Apollo program to evaluatethese residual effects through the flight and recovery ofmirror samples.


Figure 2 shows the reflectance of a platinum-coatedtelescope mirror from the OSO-IV instrument as afunction of wavelength. Curve A shows the reflectancesoon after the mirror was coated, and the crosses showthe reflectance determined from the monitor mirrorsused to assess contamination during prelaunch testingby changes in reflectance. Curve B of Fig. 2 showsthe estimated reflectance after correction for the accu-mulated contamination. Figure 3 shows the ratio of themonitor mirror reflectance at instrument delivery to thereflectance of the mirror just after it was removed fromthe instrument on the launch tower, only hours priorto launch, as a function of wavelength, and thus in-dicates the accumulated reflectance change during theprelaunch test and storage period. The resultantchange is consistent with experience in laboratory ex-periments on optical contamination.

VI. Optical Efficiency Changes in Orbit

The Harvard OSO-IV instrument was launched 18October 1967, and operated successfully for six weeks.During this time marked changes were observed in theintensities of solar spectral lines, attributed to changes inthe optical efficiency of the instrument in orbit. Theexperiment was in orbit for approximately one week be-fore the high voltage detection system was fully opera-tional. Therefore, the instrument had experiencedthe launch process and a week's immersion in an orbitalvacuum of approximately 10-9 Torr with full uv solarillumination.

Two methods were used to assess the relative changein orbital calibration. First, large raster spectro-heliograms including the entire solar disk were con-structed in, solar emission lines of relatively low ion-ization energy at frequent intervals during the instru-ment lifetime. The radiation from low stages of ioniza-tion averaged over the entire disk is not sensitive tolarge fluctuations with solar conditions, particularlyover the short interval of six weeks and in the absenceof solar flares.3' 5 Solar activity primarily affects ionsof quite high stages of excitation,2 such as Si xii andFe xv, xvi, producing very strong enhancementslocally over active regions (Fig. 1).

A second method for assessing changes in orbitalcalibration stemmed from an observing program inwhich the spectrum from the center of the solar disk wasscanned at frequent intervals, normally at least onceper calendar day. In the present solar cycle the twonear equatorial belts of activity are separated from thecenter of the disk, the northerly belt being the moreactive. Hence, except at periods when solar activeregions are drifting across the sun and produce en-hancements at the geometrical center, the center of thesun can be used as a "quiet sun" standard of calibra-tion. This assumption of a quiet sun center must notbe pushed too far, since we have found that fluctua-tions of a factor of two in the intensities of even lowstages of ionization are observed in quiet regions in day-to-day spectral scans. Approximately ten readings ofthe total integrated intensity under individual lineswere used for each spectral scan to determine the in-

300 400 300 600 200 800 900 1000 100 200 300 400 I00

WAVELENGTH (A)

Fig. 2. Reflectance of the platinum-coated telescope mirror.Curve A: the reflectance measured soon after the mirror wasreceived. The two crosses (+) indicate independent measure-ments of a reference flat coated in the same operation. Curve B:the reflectance estimated for the time of launch from changes ob-

observed in a separate monitor mirror (Fig. 3).

W VE 90 E 1.GT 1

WAVELENGTH (A)

Fig. 3. Change in reflectance during the two-month period be-tween final mirror installation and launch. The ordinate givestheratio of the measured reflectance of the monitor mirror removedfrom the instrument just prior to launch to the reflectance when

the monitor and telescope mirrors were installed.

tensity of the quiet sun. However, from day to daythese intensities, which showed statistical fluctuationswithin a given single wavelength scan, varied by afactor of two, as presumably the pointing selecteddifferent locations within the quiet solar regions.However, since a great many complete wavelengthscans were obtained, a smooth curve could be drawnthrough the scattering of points to produce a resultantcalibration change. Approximately forty-five wave-lengths over the 300-1400-A instrument range wereused to follow the calibration changes in orbit.

After the instrument had been approximately sixweeks in orbit the calibration at all wavelengths wasobserved to have stabilized. Figure 4 shows theintensity changes recorded from spectral scans at thecenter of the sun as a function of time for three selectedwavelengths, C II 1334 A, L-e 938 A, and He i 537 A.These lines represent three different types of wavelengtheffect observed over the instrument range. Eachplotted point of Fig. 4 is the average of at least tenreadings of the total integrated intensity under thesolar emission lines, and on given orbits two spectralscans frequently are obtained. The ratio of the in-


2 F

1' -� ___ - -�.-..

IIIII

I21

loo I~o -30 4 I .. ".500 600 700 81

tensity at each orbit to the final stabilized intensity wasplotted as a function of time for thirty-one wavelengthsover the 300-1400-A spectral range. Again, lines ofrelatively low stages of ionization were used, and quitevariant shapes were observed in the curves, although arelatively smooth progression in shape change could befollowed. To follow the wavelength-dependent cali-bration changes accurately with time, it is necessary tofollow standard wavelengths at least at 100-A inter-vals, and preferably at 50-A intervals. The use of avery few wavelengths, such as result from certainstandard light sources, would not have been adequateto follow the observed changes. Since no appreciabletime-dependent change in signal could be determinedfor wavelengths between 850 A and 970 A, we take thisto be a valid indication that the sensitivity of the elec-tronic system, including photomultiplier gain, counterefficiency, and electronic noise, must have remainedstable. The derived internal calibration changes wereestimated to be internally consistent and valid toapproximately 20% in the 525-1350 wavelengthinterval and to <50% outside this interval.

The changes in sensitivity at wavelengths longer than970 A and shorter than 850 A were observed to have anopposite sense. That is, while the instrument sensi-tivity is increasing at the longer wavelengths, it isfrequently decreasing at the shorter wavelengths. Forthe shortest wavelengths, below 500 A, the changes be-come complex indeed, showing several reversals. Aparticularly strong change in character of variation wasobserved between 465 A and 304 A, a relatively shortwavelength increment.

VII. Interpretation of OrbitalCalibration Changes

In the satellite instrument, the changes in photo-metric response must be a combination of changes inthe optical efficiencies of the mirror and grating andthe photoelectric yield changes of the tungsten photo-cathode. Madden' 0 has shown that for tungsten photo-cathodes cleaned and in ultrahigh vacuum, the photo-electric yield changes strongly, even when monolayersof such pure gases as nitrogen and oxygen are added.Moreover, the sense of the change is a function of wave-length over our 300-1400-A instrument range. For thepurposes of these discussions, we must consider that theinstrument reflectance and tungsten photoelectricefficiency at the time of instrument calibration were notthose of clean surfaces as expected, but were ratherin a stable contaminated condition resulting from chem-ical cleaning. For mirrors this state is very close tothe reflectance measured in situ immediately after coat-ing and before contamination can occur. For thephotocathode, on the other hand, adden's measure-ments have shown that heating a tungsten photodiodeto moderate temperatures of several hundred degrees invacua of 10-1 Torr produces large changes in the photo-electric yield, with a decreasing sense near 580 A formost metals. A similar cleaning effect would beexpected from subjecting a sample to ultrahigh vacuasuch as occur in orbit for extended periods of time. We

~ ~~0- l L I- .. - - - -- -- - . - - - - '

' 0 0 -. - - - 4 4 4 4 4 ~ ~ ~ ~ ~ ~ . - - t

50- -

40 - _

30 -- -0_ 20

20

CI ° ° = 1

5 0

_ 2 0 _ _. _ -,_

1*450-

2100

HJ 50Z 40

30

20

10.

-- - i II_

_ --; 0 ~ f 0: .He 1 537 - = --t~~~ __._.4~ ____ e a r 44~ - Y-g ~

7 14 21 28 35 42TIME (DAYS)

Fig. 4. Changes in instrument sensitivity with time in orbit forthree selected wavelengths, 1334 , 938 A, and 537 A. Theordinate is the measured signal from the quiet center of the solar

disk plotted against days after launch.

have already determined from laboratory experimentsthat cleaning of optical surfaces can be accomplishedin this manner. Madden's experiments, showing thechanges in photoelectric yield after the addition of gaslayers, demonstrates that nitrogen and oxygen havedifferent effects. Presumably different curves wouldresult for other contaminating species. Since thesatellite instrument must contain a host of differentcontaminants in varying amounts, predicting the ob-served effects in terms of specific contaminants would beunjustified.

It does not seem to be possible to disentangle thechanges in reflectance and efficiency of the telescopemirror and grating from the changes in the photo-electric yield, as this would require a knowledge ofmore parameters than those available. Since othersatellite experiments using tungsten photocathodes andmagnetic multipliers have shown no similar sensitivitychanges at the lower wavelength region in common(namely, He ii X 304) we are inclined to attribute theobserved changes in the greater part to reflectance andefficiency changes in the optical system.

In passing, reference should be made to the use ofradioactive sources as a means of assessing any changesin the efficiency of the photomultiplier as a whole.Since the released energy and number of photoelectronsis different for the very energetic f rays than for themuch less energetic photons (which also vary as thewavelength changes by a factor of four), the responseof the multiplier as a whole will be different in thetwo cases. The response of these Bendix magneticmultipliers is very dependent on the energy of the firstphotoelectron. 9 It is probably valid to use a source


LyE 938 -_4, -- __ ----4-

4 A4 ----zjZ -z2Y iI I

i

to assess the multiplier gain stability (particularly atshorter wavelengths), but the results should not beapplied to the photometric efficiency of the entirephotomultiplier (particularly at longer wavelengths).

Vil. Absolute Photometric Calibration in Orbit

Experience with rocket experiments at the HarvardCollege Observatory and elsewhere has shown thatabsolute photometric calibration over extended wave-length ranges can be maintained for the short durationof a rocket preparation and flight. The results of theprevious section show that the periods of changes re-sulting from in-flight high vacuum cleaning are muchlonger than the several minutes involved in a ballisticrocket experiment. We have concluded that a reliablephotometric calibration in orbit can be achieved byintegrating the intensity from the entire solar disk inlines of low stages of ionization and then normalizingthese to data from rocket flights. An equally validapproach would be to use the quiet areas on the solardisk as a reference. However, the rocket measure-ments made to date, primarily by Hinteregger, employlight from the whole solar disk, and the effects of limbdarkening, limb brightening, and active regions canplay an important role in the integrated intensities ofcertain solar emission lines and continua. Our ownevaluation of the best values from published"- 5 re-sults is shown in Table I. We are indebted to Hinter-egger and Hall of Air Force Cambridge Labs for makingtheir unpublished data available to us. It is generallyagreed that absolute calibration in this wavelengthrange is probably not accurate to better than withina factor of two to three. Hence, one might naively as-sume that the observed orbital calibration changes areperhaps insignificant. However, the uncertainty ofabsolute calibration in this wavelength range must bereduced to 10% or 20% for proper interpretation for uvresults in terms of solar structure and mechanisms.This problem is certainly aggravated by the use of nor-mal incidence optical systems as witnessed by the pres-ent data and those of the Naval Research Laboratory. 6

However, achieving spatial resolution on the solar diskis quite difficult with grazing incidence optical systemsin the EUV, where gratings are required to achieve spec-tral resolution of about 1 A or better necessary to iso-late lines from different regions of the chromosphereand corona or to separate closely spaced lines of differ-ing stages of ionization. An experiment of the lattertype is first scheduled to fly on OSO-H in 1970/1971,prepared by Neupert of Goddard Space Flight Center.

The spectral lines of the chromosphere and transi-tion zone fall mainly in the spectral region most con-veniently accessible with normal incidence systems,i.e., about 280-1400 A. Normal incidence systemsprovide a reasonable insensitivity to problems arisingfrom multiple overlapping grating orders leading to lossof spectral purity in the image. To achieve a preciseand reliable photometric calibration over prolongedperiods with any EUV satellite instrument is a neces-

Table . Absolute Fluxes Used for In-Flight Calibration

Flux X 10-9Line Identification (photons cm-2 sec-1)

Cont. 1400a 0.44Si IV 1394 2.1C II 1334 6.9C 1176 2.4O VI 1032 2.5L-,3 1026 3.5C III 977 4.9L,,,, 897a 0.48L<> 800a 0.061O IV 791 0.026O III 702 0.24O v 630 1.1He I 584 0.89O IV 553 0.31He,, 504a 0.12He II 304 7.5

a Integrated over 3.16-A continuum.

sary objective. To achieve this objective instrumentsmust either be free of long term calibration changes orthe calibration changes tracked at frequent wavelengthintervals with normalization to an absolute standard.At the present time, available standard sources areneither sufficiently reliable nor contain sufficiently wellspaced wavelengths to serve as an absolute standard tothe desired accuracy. It would appear, however, thatthe goal can be achieved by launching a series of rocketsprobably using grazing incidence systems during thelifetime of an orbiting satellite instrument to achievepoints of absolute calibration by observing the quiet sunor the entire solar disk. This approach will be at-tempted in later experiments.

The work reported here represents a part of the datafrom the Harvard instrument on OSO-IV, the accumula-tion and reduction of which has been the result of thecombined efforts of many people. We particularly wishto acknowledge G. L. Withbroe and R. W. Noyes, andthe programming and compilation assistance of J.Flagg and E. Levin.* The engineers and techniciansof the Solar Satellite Project merit special commenda-tion for their years of effort in preparing this instrumentfor launch. N. L. Hazen, F. Kazynski, J. Rechavi, S.Diamond, and J. Crawford are among those who de-serve special mention. We also acknowledge the co-operation of H. Hinteregger of the Air Force CambridgeResearch Laboratories, G. Timothy of University CollegeLondon, and R. Tousey of the Naval Research Labora-tory for many helpful discussions. This work hasbeen supported by the National Aeronautics and SpaceAdministration under Contract NASw-184.

* M. C. E. Huber made the measurements on polarization.


References

1. H. E. Hinteregger, Space Astrophysics, W. Liller, Ed.(AcGraw-Hill Book Company, New York, 1961), p. 34.

2. W. M. Neupert, W. A. White, W. J. Gates, M. Swartz, andR. M. Young, Solar Phys. 6, 183 (1969).

3. H. E. Hinteregger and L. A. Hall, Solar Phys. 6, 175 (1969).4. L. A. Hall, Air Force Cambridge Research Laboratories,

private communication.5. G. Timothy, University College, London, private communi-

cation.6. D. M. Packer, Naval Research Laboratory, private communi-

cation.7. Sky & Telescope 34, 362 (1967).8. L. Goldberg, R. W. Noyes, W. H. Parkinson, E. M. Reeves,

and G. L. Withbroe, Science 162, 95 (1968).9. P. J. Macar, J. Rechavi, M. C. E. Huber, and E. M. Reeves,

"Solar Blind Photoelectric Detection Systems for Satellite

Applications," Harvard College Observatory Solar SatelliteProject Rept. No. TR-9 (1969).

10. R. P. Madden, "Calibration Methods in the Ultraviolet andX-Ray Regions," ESRO SP-33, Munich (1968), p. 111.

11. J. A. Bowles, W. M. Glencross, R. J. Speer, A. F. Timothy,J. G. Timothy, and A. P. Willmore, Astron. J. 73, S56 (1968).

12. L. A. Hall, K. R. Damon, and H. E. Hinteregger, Space Re-search III, Proc. Third Intern. Space Sci. Symp., W. Priester,Ed. (North-Holland Publ. Co., Amsterdam, 1963), p. 745.

13. H. E. Hinteregger, L. A. Hall, and G. Schmidtke, SpaceResearch V, Proc. Fifth Intern. Space Sci. Symp., D. S. King-Hele, P. Muller, and G. Righini, Eds. (North-Holland Publ.Co., Amsterdam, 1965), p. 1175.

14. C. R. Detwiler, D. L. Garret, J. D. Purcell, and R. Tousey,Ann. Geophys. 17, 263 (1961).

15. W. H. Parkinson and E. M. Reeves, Solar Phys. 10, 342(1969).

0 Loren P. Thompson of Mark Systems,Inc.

James R. Durig of the University ofSouth Carolina, recipient of the 1970Coblentz Memorial Prize in Molecular

Spectroscopy.


Download - Calibration Changes in EUV Solar Satellite Instruments

Top Related