the solar corona at totality · 2016-10-21 · the solar corona at totality served for all of...

THE SOLAR CORONA AT TOTALITY

Airborne experiments probe thedynamic conditions of the sun’satmosphere during a total solar eclipse.

o n February 16, 1980, themoon’s shadow raced acrossour planet from the Atlantic

Ocean to China at about 1600 kilo-meters per hour (Fig. 1). Thousands ofspectators and scientists scattered alongthe shadow’s path lifted their heads toglimpse the pearly glow of the solarcorona, the sun’s outermost atmosphere.Near the midpoint of this path wheresun, moon, and earth were aligned forthe longest time, a high-altitude UnitedStates Air Force jet made its carefullyprogrammed rendezvous with the eclipseshadow. Aboard the aircraft were 18Los Alamos scientists, observers fromIndiana University and Kitt Peak Na-tional Observatory, a 12-man Air Forceflight and support crew, a Kenyan ob-server, and two media representatives.

We were intent on recording the con-ditions in the solar corona during thesecond highest peak in solar activity inthe past century. Five experiments,planned months ahead, were mountedinside the aircraft to make high-resolution measurements of coronalmorphology, electron densities, electrontemperatures, ion temperatures, and cir-cumsolar dust rings. Our expedition wasthe seventh in a series of Los Alamosairborne missions that began in 1965 asan outgrowth of the national Test Read-iness Program.*

During a total eclipse the aircraftprovides near ideal conditions for ob-serving the solar corona. At an altitudeof 11 kilometers (36,000 feet) we enjoyfreedom from cloud cover and reducedsky brightness and, by flying along theeclipse path, we increase the time oftotality by 50% or more. In 1980, we

*See “In Flight: The Story of Los Alamos EclipseMissions” in this issue.

LOS ALAMOS SCIENCE

had an unobstructed overhead view ofthe eclipse for 7 minutes 7 seconds whileobservers at the main ground site inIndia saw the event between clouds at alow angle in the afternoon sky for only 2minutes 9 seconds.

This exceptional high-altitude plat-form outfitted with state-of-the-art in-struments for tracking control and dataacquisition (many of which derived fromand contributed to the weapons testingprogram) has enabled Los Alamos toachieve many firsts in observation ofcoronal temperatures and densities. Such

measurements are needed to resolve twomajor questions in coronal physics. Howis the corona heated? And how is thesolar wind generated?

We will look at what some of thesemeasurements are and how they tit intostudy of the corona. We will also de-scribe the technology developed at LosAlamos and how it has been used incarrying out these missions.

Early Coronal Studies

Total solar eclipses have been ob-

Fig. I. The path of the moon’s shadow first touched the earth in the Atlantic Ocean,then crossed Africa, the Indian Ocean, and India, and finally left the earth overChina. Los Alamos observers launched rockets near Malindi, Kenya. The Los Alamosairborne expedition was based in Nairobi and flew to the point of longest totalityduration just south of the equator over the Indian Ocean. The main ground-basedscientific site was near Hyderabad, India.

5

c oronal morphology (depictedhere out to 3 solar radii) is mostlikely shaped by the sun’s mag-

netic field, which changes over an1 l-year cycle of activity.

During a period of quiescence, coro-nal structure suggests that open lines ofthe sun’s magnetic dipole field spreadout from almost all solar regions exceptthose near the equator, where the coronais brightest and fairly uniform. Theseregions of open field lines tend to be lessbright and may qualify as coronal holes,features characterized by drastically re-duced x-ray emission, low particle densi-ty, and possibly low temperature. It isknown that, during periods of solarinactivity, coronal holes extend from thepoles toward the equator and covermuch of the sun’s surface. Presumablysolar wind flows strongly from coronalholes, thus depleting particle and energy

6

densities. Polar plumes extend from thecoronal holes at the sun’s north andsouth poles.

During a period of maximum solaractivity, as complex local magnetic fieldsdevelop over much of the sun’s surface,the open dipole field lines and, withthem, coronal holes appear to be restrict-ed primarily to the polar regions. Manydistinctive coronal features are formedas a result of this increased activity.

Perhaps the most striking feature isthe helmet structure that extends out-ward into a long radial streamer at about2 solar radii. The helmet is probablyformed by magnetic field loops that risefrom active regions of the photosphere.Charged particles are trapped alongthese field lines, but near 2 solar radiimagnetic field strengths decrease suffi-ciently so that particles can drag the fieldlines radially outward to form a

streamer.Sunspots, prominences, filaments, and

flares also form in active regions of thesun’s surface. These regions seem to beconfined to well-defined belts that movefrom high latitudes toward the equatoras solar activity declines.

Coronal enhancements and condensa-tions, bright areas near the sun’s limb,are also associated with active regions.Their brightness is due to increasedparticle densities. An enhancement maypersist for months, whereas a condensa-tion may appear and disappear over aperiod of hours or days.

Changes in coronal morphology dur-ing the solar cycle are best portrayed byour image-enhanced photographs of thecorona in the note “The Changing Coro-na” ❑

LOS ALAMOS SCIENCE


served for all of recorded time andprobably since humans began to look atthe sky, but the origin and nature of thesun’s magnificent halo has bathed eventhe most astute observers. In 1605Kepler attributed the display to a lunaratmosphere. Others associated it withthe atmosphere of the earth or the sun.Not until the 1842 solar eclipse, when aspectacular display of corona and prom-inences was observed by astronomers inseveral European countries, was theregeneral agreement on its solar origin.

The corona presented many puzzles,the most important being the absence ofemission lines from hydrogen and heliumand the presence of the coronal greenline, a mysterious emission line thatcould not be identified with any knownelement on the sun. The green line wasobserved with a spectroscope by Hark-ness and Young in 1869. Originally thisand other unidentified spectral lines wereattributed to a new element named cor-onium. But there seemed to be no placefor coronium in the periodic table, so themystery of the spectral lines remained.Then in the late 1930s the lines wereidentified in the spectrum of Nova Pic-toris, an exploding star whose at-mosphere was known to be very hot.Finally in 1942, the great Swedish spec-troscopist B. Edlen made laboratorymeasurements of hot plasmas and identi-fied nearly all coronal emission lines asdue to “forbidden” atomic transitionsfrom highly ionized elements. The greenline was produced by a transition in FeXIV, an iron atom with 13 of its 26electrons stripped away.* Other lineswere attributed to other ionization statesof iron, calcium, and nickel.

Tremendous energy is required to pro-

● Physicists refer to unionized iron as Fe I.

duce such highly ionized states. Also, foratoms to remain highly ionized longenough to deexcite through these forbid-den transitions, the density must be verylow. Thus the corona is a very hotplasma of ionized hydrogen, helium, andheavier elements with a temperature of a

and a density one-trillionth the density ofthe earth’s atmosphere. That the coronashould be so hot was quite unexpectedbecause it overlies the relatively coolsurface of the sun (photosphere), whichis about 6000 K.

What are the dynamics of this hot,rarefied atmosphere? How is it heated?How does it sustain itself when it isbounded below by the cool photosphereand above by the cold vacuum of in-terstellar space? The answers to thesequestions still remain quite tentative.

The Solar Wind

In the late 1950s the gross dynamicsof the outer corona began to be under-stood. In 1957 Chapman noted that asimple model of the corona in hy-drostatic equilibrium leads to a con-tradiction. According to such a model,the corona must extend to the earth’sorbit and beyond. Its pressure must fallso slowly that at infinity the coronalpressure would be much too high to bebalanced by the pressure of interstellargas.

In 1958 E. Parker solved this problemby introducing the revolutionary andinitially unpopular idea that the coronais not static at all, but instead is in aconstant state of expansion. His modelpredicted a supersonic flow of particlesout from the corona beginning at a fewsolar radii. This flow came to be known

as the solar wind. Its existence wasconfirmed in the early 1960s when satel-lites first began to penetrate beyond theearth’s magnetic field and detected parti-cles traveling away from the sun atvelocities of 400 kilometers per second.This variable stream of charged particles(mostly electrons and protons) is respon-sible for the sharp discontinuity of comettails as they pass near the sun, aphenomenon first noted by L. Biermannin 1951. The solar wind also causes avariety of phenomena as it “blows” pastthe earth—from geomagnetic stormsthat interfere with radio communicationsto possible subtle effects on our weathercausing climate variations that appear tobe synchronized with the sun’s magneticactivity cycles.

It is interesting that until now meas-urements of solar wind flow, interpretedthrough solar wind theory, have proba-bly given as good an idea of the averageconditions in the inner corona as directobservations. To match solar wind fluxesmeasured near the earth, these modelspredict that temperatures in the corona

decrease very slowly thereafter. In par-ticular, if the temperature T falls off with

flow begins at a critical distance from thesolar surface. Values of b may be esti-mated from observed particle fluxes.Beyond a few solar radii, collisionsamong particles and photons are so rarethat ionization levels are “frozen in.” Inthis collisionless plasma, temperaturesrefer to an average velocity distributionand do not imply local thermodynamic

*Distances from the center of the sun are com-

7LOS ALAMOS SCIENCE


equilibrium.Temperature measurements needed to

confirm simple fluid models of solarwind flow have been in a state of con-fusion. Not only are these measurementsvery difficult to make but the corona is acomplex, inhomogeneous structure thatchanges with time.

During 1973 and 1974, observationsof the x-ray and extreme ultravioletcoronal spectrum from the Apollo Tele-scope Mount aboard Skylab revealedmany facts about the inhomogeneousnature of the solar corona. One of themost important for solar wind theorywas the existence of coronal holes, re-gions of low density, perhaps low tem-perature, and open magnetic field linesthat are almost devoid of x-ray emissionat the solar surface. At the time of theSkylab observations, when the sun wasin a quiet phase, the hole regions ex-tended down toward the equator. Thevery high velocity component of thesolar wind observed during that timeevidently emanated from the equatorialcoronal hole region where particlesstream out at high velocities along openfield lines. However, the mechanism foraccelerating these particles as well as thecontributions to the solar wind fromother regions of the corona are still notwell known.

High-resolution data from our 1980mission will help to determine the tem-perature structure in coronal holes andother coronal features and thus help toexplain how the solar wind is generatedin each of these areas.

Coronal Heating

Our 1980 data may also shed light ona problem that has puzzled solar physi-

cists since the discovery of the hotcorona. By what mechanism is energyfrom the photosphere deposited in thecoronal plasma?

Radiation heating by the enormousflux of visible light emanating from thephotosphere is ruled out because thecoronal plasma is essentially transparentto visible wavelengths. We must lookinstead to mechanisms that could beproduced by the interaction of massmotions and magnetic fields in the tur-bulent convection layers of the pho-tosphere (Fig. 2).

Large volumes of gas constantlyemerge from the surface, spread out andcool, and finally sink back into deeperlayers of the photosphere. This convec-tive energy could generate acousticwaves that propagate into the chro-mosphere and the extreme lower corona.Alternatively, magnetic waves generatedby the interaction of convective cellswith magnetic fields might provide aheating mechanism that extends far outinto the corona. (Magnetic fields extendoutward from the photosphere throughthe chromosphere and outward with thecorona.)

This process might take place in veryactive regions of the solar surface. Inthese regions of high magnetic fieldstrengths, we see eruptive prominences,filaments, and sun spots. Coronal con-densations overlie these active regions.Perhaps as tangled loops of magneticfields move around, currents flow thatuntangle field lines by reconnecting themin simpler configurations, thereby releas-ing large amounts of magnetic fieldenergy in the form of magnetic waves. Asimilar but more extreme process isproposed for solar flares. Or perhaps

magnetic waves are generated all overthe sun’s surface.

The details of these phenomena aretoo complicated to be modeled at pres-ent. But there are some questions we canhope to answer through observations.First, how far up in the corona is heatdeposited? Is it all deposited in thetransition zone between the chro-mosphere and the corona where thetemperature rises from 0.01 to 1 MK inless than 1000 km? Or does the heatingmechanism extend out to several solarradii, as suggested by solar wind theory?If so, how is energy transported throughthe corona?

Is energy distributed equally amongelectrons, or are protons heated prefer-entially by nonlinear processes, thus ex-plaining why solar wind protons ob-served near the earth have higher tem-peratures? Finally, is solar wind flow aresult of temperature gradients alone orare there mechanisms that accelerate thewind without heating it?

If acoustic waves are the major heatsource, the lower coronal temperaturewould be maximum at approximately1.1 R@ because acoustic waves wouldbe damped in the chromosphere or ex-treme lower corona. If magnetic wavesare the major source, the temperaturemaximum would move farther out in the

magnetic waves would lose their energy “much more gradually, possibly out as far

Direct observations of thistemperature maximum would obviouslyhelp to determine the mechanism ofcoronal heating. Also, direct observa-tions of periodic variations in brightnesswould help to establish the existence ofextended wave heating.

LOS ALAMOS SCIENCE 9

Fig. 3. Approximate variation of particle density and temperature as a function ofradial distance. (Note radial distance scale changes.) The steep density decrease abovethe photosphere and the hundredfold increase in temperature from the base of thechromosphere to the corona are fairly well established but temperatures and gradientsin the chromosphere and above are still the subject of research. The exact location ofthe steep temperature rise in the chromosphere-corona transition region is uncertain,although this region is known to be thin. Temperature and density values in the coronamay vary from one feature to another and may be modified by the presence ofprominences, flares, and active regions in the photosphere below. Temperature anddensity of the solar wind at the earth are shown for comparison.

The Problem ofCoronal Temperatures

The key parameters for modelingenergy transport in the corona are thedistribution of electrons and ions in thecorona and their temperature, or, moreprecisely, their velocity distribution.

The average values of the particledistribution are fairly well known, butthe temperature distribution is not (Fig.3). We would like to know not only grossaverages but details of temperature dis-tributions in various structural featuresof the corona since the energy transportwill vary from one to another. At presentthe results from various methodsgathered at various times are in disagree-ment. Although these discrepancies areunderstandable, given the fact that thevarious methods for measuring tem-perature involve different assumptions,

10

the methods are complementary; if donein a coordinated fashion, the separatemeasurements can be compared to yieldmore information than any single onecan yield alone.

HYDROSTATIC TEMPERATURES. up-

per and lower limits on average coronaltemperatures can be obtained from elec-tron density measurements. The deriva-tions for both limits assume that elec-trons and ions are in local thermo-dynamic equilibrium and electron densi-ty equals ion density at each point in thecorona. The lower limit estimate is basedon a model of the corona in hydrostaticequilibrium (neither expanding nor con-tracting). The pressure gradient exactlybalances the acceleration of gravity,

(1)

where P is pressure, r is radial distancefrom the center of the sun in solar radii,

surface, and the material density p isdirectly proportional to the particle den-sity n. Since coronal densities are low,we assume that the ionized gases obeythe ideal gas law,

P = nkT , (2)

which relates the pressure to the particlenumber n and temperature T. If weassume that the temperature is constant,we can derive the density variation withradial distance by differentiating Eq. (2)and substituting the result into Eq. (l).We obtain

(3)

Integrating Eq. (3), we find that a plot oflog n versus l/r is a straight line with aslope proportional to l/T. Since thecorona is nearly all hydrogen we mayreplace particle density with electron

n. Consequently, if theexperimental values of log ne lie on astraight line, then the so-called hy-drostatic temperature may be deducedfrom its slope,

Actual measurements of log ne, versus

with a slope corresponding to 1.4 to 1.6MK. However in equatorial regions ofthe sun the slope changes between 3 and5 R@ and then beyond 5 R @ it has a

constant value corresponding to slightlyless than 1 MK. This change in slopeindicates that the corona does not have aconstant temperature. Instead the tem-perature decreases with distance fromthe sun, a result consistent with anexpanding rather than a static corona.Nevertheless, hydrostatic temperaturesare useful as lower limits on averagecoronal temperature.

TEMPERATURES FOR AN EXPANDING

CORONA. To determine temperatures foran expanding corona in, for example, acoronal hole, we add a velocity depen-dent term to Eq. (l).

LOS ALAMOS SCIENCE


where v(r) is the radial expansion veloc-ity. If we assume that the average solarwind mass flux K is constant and equalto the value observed near the earth, wecan obtain a rough estimate for theexpansion velocity v from the relation

ne,(r) v(r) A(r) = K , (5)

where A is the surface area. We then usethe expansion velocity to determine thepressure gradient, dP/dr, from Eq. (4).Assuming that the pressure gradient isdue to wave heating of the corona, wedetermine the temperature gradient fromEq. (2). Then, with a good value for thetemperature from an independent meas-urement, we can use these temperaturegradients to determine the variation oftemperature with radius. When appliedto the 1973 Skylab electron density data,this method yielded a steeply rising tem-perature curve for the coronal holeabove the solar north pole.

OTHER CONSIDERATION. This resultmust be questioned for two reasons.First, since observed densities in coronalholes are very low, collisions may be so

EXTREME

ULTRAVIOLET

EUV LINESUSED TO

DETERMINE

infrequent that local thermodynamicequilibrium cannot be assumed. A con-sequence would be that electron and iontemperatures are not the same. Second,the method assumes that pressure gra-dients, and thus the acceleration of thesolar wind, are due to temperature gra-dients created by wave heating of thecoronal plasma. However magneticwaves moving radially outward mayaccelerate the solar wind without heatingit. Since Eq. (4) does not explicitlyseparate this nonthermal contribution tothe velocity v(r), it yields effective pres-sures, and through Eq. (2), effectivetemperatures that are higher than thekinetic pressures and temperatures weare seeking. Temperatures derived fromEqs. (2), (4), and (5) must therefore be

interpreted as upper limits. By consider-ing these limits in conjunction with iontemperatures and velocity distributionsdetermined from emission line measure-ments, it may be possible to separatethermal and nonthermal sources of solarwind acceleration.

EUV ION TEMPERATURES. Ion tem-peratures have been obtained from

VISIBLEULTRAVIOLET

WHITE LIGHT

emission line measurements by twomethods. One method is based on com-paring measured and calculated in-tensities of resonance lines from allowedatomic transitions. These resonance linesare at extreme ultraviolet (EUV) wave-lengths (Fig. 4). EUV data from 1 to 1.4R@ have been measured from satellites.Calculations assume an equilibriummodel of coronal temperatures and den-sities and use known atomic parametersto determine the probability for reso-nance line emission. The assumed tem-peratures are adjusted to match meas-ured and calculated line intensities. Er-rors are introduced by the crudeness ofthe assumed coronal model and by un-certainties in atomic parameters.

INFRARED

LINES OBSERVED BY THE

CONTINUUM RADIATION EMISSION LINE CAMERA

(K CORONA) FORT e , EXPERIMENT \

. Fe XIV GREENLlNE

LINE USED TO DETERMINETPROTON I

H&K OF Ca IIABSORPTIONLINES

Fig. 4. Relative positions in the electromagnetic spectrum of merely indicate the wavelengths at which observations areobservations made to determine coronal conditions. We do not made.attempt here to reproduce the actual coronal spectrum but


June

160

1976 1978 1980 1982 1984 1986

THECHANGINGCORONA

Cycle21

October

12

1964 1966 1968 1970 1972 1974

P ortraits of the corona during fiveeclipses display its changing ap-pearance as the sun passes

through its cycle of magnetic activity.The eclipse dates are given on the ac-companying graph of sunspot activity,which covers the present (red) and pre-vious (blue) solar cycles. Each portrait isa composite of 20 to 30 photographicimages taken from the NC- 135 jet air-craft with the Laboratory-designed cam-era-polarimeter.

These portraits of the corona extend-ing from the sun’s limb to 12 solar radiiare unparalleled. Photographs takenfrom the ground record only the innercorona to about 4 solar radii. EvenSkylab’s superb coronagraph can view

LOS ALAMOS SCIENCE

only the region between 1.5 and 6 solarradii.

From this unique set of portraits, wecan follow the evolution of streamers.The marked variation of the streamerpattern over the sun’s 1 l-year activitycycle gives us a clear picture of the effectof solar activity on coronal structure andsolar magnetic field.

The most striking structural change inthe corona is the lack of bright streamersover polar regions during periods of lowsolar activity. At such times ( 1972 and1973), polar coronal holes extend tolower solar latitudes. Solar wind flowsfrom coronal holes and apparently bendsthe sun’s magnetic dipole field stronglytoward equatorial regions. During peri-

LOS ALAMOS SCIENCE

ods of high activity (1970, 1979, and1980), streamers approach and oftencover the polar regions and the dipolefield is bent only slightly away from thepoles.

Individual portraits bear carefulstudy. Close to the sun, streamers areobviously dominated by magnetic fields.We see them twisted and curved in alldirections. At larger distances, however,many streamers are seen to bend gradu-ally back toward a radial direction, pos-sibly under the influence of the expand-ing solar wind, Others remain nonradialeven at the farthest distances. The in-formation on streamer dynamics con-tained in these images awaits theoreticalstudy c

EMISSION LINE PROFILE ION TEM-

PERATURES. A somewhat more directmethod involves the analysis of meas-ured intensity versus wavelength profilesof emission lines from highly ionized ironand calcium. (Several emission lineshave wavelengths in the visible spec-trum.) The width of these Doppler-broadened emission lines is a directmeasurement of ion kinetic tem-peratures, provided all the broadening isdue to thermal motions. However massmotions and turbulence can also con-tribute to the broadening.

The Los Alamos line profile data outto 3 R@ are the most extensive avail-

able. Our analysis suggests that iontemperatures deduced directly from thewidths of heavy ion emission linesmay be larger than kinetic temperaturesby as much as 30% because of non-thermal contributions to the line broad-ening.

Kinetic temperatures may be deduceddirectly from the width of hydrogen’sLyman-a emission line at 1216 A be-cause nonthermal contributions to theline broadening are proportionally muchless than for heavier mass ions. SinceLyman-a radiation cannot penetrate our atmosphere, such measurements must bemade at altitudes of 100 km or greater.Data between 1.5 and 3 R@ are nowavailable from rocket-borne experimentsperformed by scientists from HarvardUniversity and the High Altitude Ob-servatory of the National Center forAtmospheric Research.

The results of recent temperaturemeasurements shown in Fig. 5 are inobvious disagreement. In coronal holes(Fig. 5a), we see upper and lower limitson kinetic temperatures based on elec-tron density data. The lower limit is -1

13

MK based on the assumption of a static,isothermal corona. The Skylab electrondensity data, interpreted through a mod-el that accounts for expansion velocitiesas described above, results in a steeplyrising curve that reaches 3.5 MK by 3

The other three observations fallwithin these bounds but are too sparse toallow determination of the position of thetemperature maximum. The single EUVtemperature and the single Lyman-a pro-ton temperature suggest extended mag-netic heating. When the Los Alamos ironion temperatures are included, the tem-perature maximum moves below 1.3

Figure 5b shows similar results forrelatively quiet regions of the corona

(regions that are devoid of obviousstreamers but are not coronal holes).Neglecting the iron ion temperature putsthe temperature maximum beyond 1.4R@; including it moves the maximum

Preliminary iron iontemperatures from the Los Alamos 1973eclipse observations show ever increas-

electron density and the EUV results inthe lower corona can be believed, thenthe iron ion results must include a largeturbulent contribution to line broad-ening.

Interpretation of the results shown inFig. 5 is further complicated by the factthat the data were obtained at differenttimes and at different places in thecorona. To resolve the discrepancies, weneed simultaneous observations of iontemperatures, electron temperatures, andelectron densities. One of the primarygoals of the Laboratory’s 1980 solareclipse expedition was to do just this.

The 1980 Experiments

The sun’s magnetic activity follows awell-known 1 l-year cycle becoming veryactive, waning to near total inactivity,and then repeating. The amplitude of thisvariation is never the same, but by 1979it was obvious that the 1980 maximumwould be one of the highest in historyand many experiments were planned toobserve the solar corona during theeclipse in February, 1980.

14

4(a) Coronal Holes

Fig. 5. Plots of coronal temperature versus radial distance for (a) coronal holes and(b) a quiet coronal region. As discussed in the text, data are sparse and, whencompared, confusing. The main purpose of the Los Alamos expedition was to attemptto reduce some of the sources of disagreement by making cotemporal and cospatialmeasurements.

The National Aeronautics and SpaceAdministration (NASA) planned tolaunch the Solar Maximum Mission, asatellite similar to Skylab’s observatory.Aboard it would be experiments to de-termine electron densities and to obtain acrude estimate of ion temperatures fromFe XIV emission intensities. (Un-fortunately this experiment was not op-erating by the time of the eclipse.)NASA also funded two rocket experi-

ments, including one from Los Alamos,to measure proton temperatures fromthe Lyman-a line. The Naval ResearchLaboratory’s orbiting coronagraph wasto make hourly recordings of the white

The National Science Foundation (NSF)planned to send to India a large con-tingent of ground-based observers, whowould make a variety of measurements.

In planning the Los Alamos airborne

LOS ALAMOS SCIENCE


expedition, we were in communicationwith scientists from NASA projects andsome NSF projects. It was apparent thatcross-calibration and data comparisonwould assure maximum confidence inthe results.

As discussed above, to answer signifi-cant questions about the nature of coro-nal heating and solar wind accelerationrequires simultaneous measurements ofthe following:

● electron density, ne

• ion temperature, TiOn

● electron temperature, Te.All these measurements were at-

tempted by our expedition. All wered e s i g n e d t o p r o v i d e g o o d t w o -dimensional spatial resolution. We ex-pected to be able to determine all thequantities not only in the relatively uni-form portions of the corona, but alsowithin major features such as streamers,condensations, and coronal holes. TableI summarizes the experiments and theinformation sought.

To determine ion temperatures wemeasured the Doppler-broadenedemission lines from two heavy ions—FeXIV and Ca XV. We will compare ourresults with proton temperatures de-

1980 Solar Eclipse Experiments

Principal InformationExperiment Investigators’ Obtained or Sought Resolution

Emission Line Ion Temperature D. H. Liebenberg Temperatures of Fe XIV and 0.005 R@ (3500 km)Measurement of coronal Fe XIV E. A. Brown Ca XV, variation of emissionand Ca XV emission line profiles R. N. Kennedy line intensity with radialwith Fabry-Perot interferometer H. S. Murray distance, and nonthermal

W. M. Sanders contributions to ion tem-perature from 1 to 3 R@

Electron Temperature M. T. Sandford Electron temperature and 0.067 R@ (46,000 km)Measurement of K coronal F. J. Honey coronal spectra at 1.2, 1.6,spectral intensity with R. K. Honeycutt, and 2.0 R@

silicon photodiode Indiana Universitydetector array

Electron Density C. F. Keller Electron density, K + F corona 0,067 R@ (46,000 km)Measurement of intensity and J. A. Montoya intensity, and K coronapolarization of coronal light B. G. Strait polarization from 1.1 to 5 R@

with camera-polarimeter in polar regions and from 1.1to 12 R@ in equatorial regionsand image-enhanced photographsof streamers from 1 to 20 R@

W. H. Regan Detailed coronal structure 0.033 R@ (23,000 km)Coronal photography with C. G. Lilliequist from 1 to 6 R@

radially graded filter andinternal occulting disc

Infrared Emission of Dust Rings J. P. Mutschlecner Radial location of dust rings 0.25 R@ (175,000 km)c o r o n a l i n f r a r e d R. R. Brownlee and infrared emissionemission intensity with InSb D. N. Hall, Kitt Peak intensity from 2 todetector and charge-injection- National Observatory 50 R@

device television cameraaUnless otherwise indicated, the investigator is affiliated with Los Alamos,Other Laboratory personnel contributing to the 1980 solar eclipse expedi-tion were W. H. Roach (scientific commander), R. A. Jeffries (in charge ofliaison with the USAF), D. H. Collins (logistics officer), C. T. Barnett


Fig. 6. The white light we observe duringa total solar eclipse has two components, olight from the photosphere scattered to s o l a rus by electrons in the corona (straight –1 Disk

line) and light from the photospherescattered to us by interplanetary dust - 2 I(wavy line).

I–3

Fig. 7. Relative intensity of componentsof coronal light as a function of radial - 4distance. The K corona is continuouslight from electron scattering. The Fcorona is light from dust scattering. E isthe combined light of emission lines.Relative background intensities undervarious observation conditions are alsoshown. Note the tenfold reduction in skybackground intensity for high-altitudeobservations over ground-based observa-tions.

r/R@

16 LOS ALAMOS SCIENCE


duced from rocket measurements ofLyman-a emission at several points inthe corona beyond 1.5 R@ to distinguishthermal from nonthermal contributionsto line broadening. We are particularlyinterested in determining whether theextremely high ion temperatures at thebase of the corona deduced from our1973 results are correct or should beattributed to the effects of turbulence,but we must await data from these lowaltitudes. We also measured electrondensities, recorded high-resolution im-ages of the corona, and attempted toobserve electron temperatures by a new.untried method. Finally, we attempted todetermine the location of hot dust ringsaround the sun from their infraredemissions.

Except for the data loss from a me-chanical failure in the electron tem-perature experiment, we obtained ex-cellent data in all areas. And photo-graphs of the corona taken from theaircraft will enable us to associate ourdetailed observations with particular

LOS ALAMOS SCIENCE

spatial structures.Data reduction and comparison is a

long and diflicult process that is onlybeginning. Before we discuss each ex-periment and give an early estimate ofthe data it acquired, let us consider someof the general conditions of coronalobservation.

Observing the Corona

The angular sizes of the moon and thesun, as seen by an observer on (or near)the earth are approximately equal. Dur-ing a total eclipse, this fortuitous circum-stance enables us to view the solarcorona, a region otherwise obscured bythe intensity of the photosphere. Coronallight comes from several sources, themost intense being light from thephotosphere that has been scattered byparticles in the corona. This white lightconsists of two components, the K coro-na—sunlight scattered by electrons—and the F corona—sunlight scattered bydust grains (Fig. 6). Superimposed on

Fig. 8. Eclipse observations of the coro-na are made along a line of sight, but inmany cases single coronal features canbe identified reliably. This is becausefeatures closer to the limb are moreintense and therefore contribute morelight to the measured signal For exam-ple, the line of sight shown here in-tercepts a streamer in the plane of the

sky at 1 .68 Ra and an i d e n t i c a lstreamer 30° out of the plane of the skyat 1.94 Ra. The streamer in the plane ofthe sky contributes about 2.5 times asmuch white light and about 5.5 times asmuch emission line intensity to the ob-served signal as does the streamer out ofthe plane of the sky.

these spectral continua are emission linesfrom highly ionized heavy ions. Figure 7displays the relative intensities of thevarious components of coronal light;these intensities are compared with thebrightness of the sky under various ob-servational conditions. Note the sharpreduction in sky brightness during asolar eclipse. Although emission line in-tensities are much less than white lightintensities, they are clearly visible abovethe white light background when ob-servations are limited to very narrowwavelength bands. However, the rapiddecrease of emission line intensity withradial distance from the sun makesmeasurements impossible beyond 3 R@

with present instrumentation.Apart from the signal intensity, the

other unalterable limitation on observa-tions is geometry—we can only recordtwo-dimensional information aboutthree-dimensional objects. As shown inFig. 8, every observed signal is inte-grated along a line of sight through thecorona. If the corona were spherically

17

T he color photograph of a hugebubble-like structure extendingfrom the sun’s limb to 7 solar

radii was a most exciting result of our1980 expedition. The structure is formedby a hydrodynamic eruptive disturbancethat ejects large quantities of mass intointerplanetary space. The eruption isvery clearly visible on the computer-enhanced image processed from 32digitized photographs. These large andfrequent mass eruptions have been re-corded since the early 1970s by NavalResearch Laboratory satellites and bySkylab, but the bases of the eruptionswere always obscured by the oversizedocculting discs used in satellite coro-nagraphs. Thus ours is the first completerecord of the phenomenon.

At first glance the structure looks likea tennis racket with its handle in the sun.Closer examination reveals that it is notentirely symmetric about an axis extend-ing radially from the sun. Its polar side ismarkedly flattened and a density en-hancement, which is possibly due to ashock wave ahead of the eruption, isvery prominent above and on the equa-torial side, but is nearly absent on thepolar side. Nearby major streamers thatusually extend in a radial direction fromthe sun are bent toward the disturbance,more so on the equatorial than on thepolar side. In addition, the eruption maypossibly be bent slightly toward the solarequator. Characteristics similar to thesehave been reported for eruptions ob-served from Skylab.

The position of this eruptive dis-turbance recorded from the aircraft dif-fers markedly from that recorded by theLaboratory’s rocket team in Kenya

about 15 minutes earlier; from the dif-ference we infer an expansion velocity ofabout 500 km/s. This value is in agree-ment with the eruption’s absence fromphotographs taken from the main scien-tific site in India, where the eclipseoccurred about 90 minutes later. Duringthat time, the eruption would havemoved beyond range of observationfrom the ground. If the velocity of ex-pansion is nearly constant during itslifetime, the disturbance must havebegun just as the eclipse was arriving onAfrica’s west coast, about 90 minutesbefore we saw it. We are at presenttrying to obtain good photographs takenfrom Zaire and Tanzania to study theearlier phases of the event.

Such eruptions are apparently quitecommon. They are estimated to haveoccurred at least once a day in 1973during low solar activity and perhapsthree times more frequently during thispast year’s maximum solar activity.Each one typically ejects a mass ofabout 10]3 kg and an energy of 1024 J .The apparent rate of occurrence wouldmake them responsible for at least 10%of the entire solar mass efflux!

We also obtained camera-polarimeterimages and Fe XIV emission line profilesof the eruption. Its Thompson-scatteredlight is highly polarized so we will beable to determine electron densities andits base appears to have the most intensegreen line (Fe XIV) emission in thecorona. We anticipate that these meas-urements and our photographs, whichtogether comprise a wealth of intercon-nected information, will answer severalquestions about these important sourcesof solar wind ■

symmetric, this would pose no problem.But in fact its visible brightness variesmarkedly from one region to another.Thus interpretation of all measurementsrequires assumptions, based on detailedp h o t o g r a p h s , a b o u t t h e t h r e e -dimensional structure of the corona.

Direct Photography —Imaging the Corona

Good photographic prints of the solarcorona are extremely difficult to makebecause coronal brightness varies byfactors of 1000 from 1-4 R@ and 10,000from 1-10 R@, whereas prints candisplay only a factor of 10. We use threemethods to reduce this radial brightnessgradient.

1. A radially graded filter placed just infront of the film transmits light inprecisely the desired amount as afunction of distance from the sun.

2. An internal occulting disc in the con-verging light path within the cameraalso uniformly reduces the brightnessgradient.

3. And computer processing reduces thegradient in the digitized photographsfrom the electron density experiment.

All three methods were used duringthe 1980 expedition. Because of a slightmisalignment of the mirror tracking sys-tem aboard the aircraft, the first twogave good photographic records of onlytwo-thirds of the corona. Neverthelessthese pictures give us a remarkably de-tailed view of the corona during the sun’smaximum activity and a particularly finelook at a large mass ejection extending

One of these photographs, together


EARTHBOUND OBSERVER

Fig. 9. Schematic of light scattered at90° by an electron in the plane of the skyto an earthbound observer. This light is100% linearly polarized along the direc-tion tangent to the sun’s radial vector,because its electric field vector (indicatedby the arrows under the wavy lines) isalways perpendicular to its direction ofmotion. However, not all the light weobserve in our camera-polarimeter hasbeen scattered in the plane of the sky.Each point on our digitized photographis a sum of contributions that have beenscattered from many angles into the lineof sight. Consequently, the maximumobservable polarization of the K coronafrom a spherically symmetric corona isonly 670/o. This theoretical maximum isreduced even further because the coronais not spherically symmetric and, in theinner corona, photospheric light is trav-eling in nonradial directions beforebeing scattered. Thus the expectedpolarization as we move toward thesun limb decreases from 67°/0 to OO/O.

with a computer-enhanced image proc-essed from 32 digitized exposures takenwith our camera-polarimeter, is shown inthe note “A Mass Eruption From theSun.”

Electron Density Experiment

The scattering process that producesthe K corona (see Figs. 6 and 7), name-ly, Thompson scattering of photosphericlight by electrons, is independent ofwavelength and depends only on elec-tron density. K coronal intensities arethus a direct measure of electron densi-

ties. Since we can measure only totalwhite light intensities (K + F corona), weneed a method to subtract the unwantedF coronal light.

Much of the K coronal light we ob-serve during an eclipse has been scat-tered through angles near 90°. The Kcorona is therefore highly polarized (Fig.9), whereas the F corona is not. The twocomponents can be separated by meas-uring both the absolute intensity of the K+ F corona (K + F) and the fraction anddirection of polarization at each point ondigitized images. The measured frac-tional polarization can be written

where Pk is the fractional polarization ofthe K corona, and K and (K + F) are theK coronal intensity and total white lightintensity, respectively. Data reduction todetermine K requires a model of thecorona. We assume that the corona hascylindrical symmetry and that its densityvaries smoothly between polar and equa-torial regions. From this model, we cal-culate Pk and K and compare them withthe measured quantities Ptotal and (K +F). Finally we vary electron densities inthe model to obtain consistency betweenthe two measured and two calculatedquantities.

Photographs of the K + F corona aretaken with a camera-polarimeter throughplane polarizing filters oriented at threedifferent angles. A fourth photographtaken without any polarizing filter com-pletes a set. During the 1980 eclipse, wemade 10 sets of photographs usinghigh-resolution film for the bright innercorona and very fast low-resolution filmfor the outer corona. This is the first timewe have taken special care to get

high-resolution data from the inner coro-na.

The camera and its center-of-mass,three-axis gyrostabilized tracking systemwere designed and built at Los Alamos(Fig. 10). With this tracking system,motion during a 3-second exposure wasless than 20 arc seconds. Our polariza-

equatorial region of the corona—much

Fig. IO. Los Alamos-designed telescopefor photographic polarimetry and elec-tronic control equipment (lower left)mounted in aircraft. This instrumentphotographs the corona beyond 12 Ra

and has provided data for determiningelectron densities and image-enhancedphotographs of the outer corona. (Seenote “Unique Record of the ChangingCorona.") The blue cylinder (upperright) is one of three orthogonallymounted gyroscopes obtained fromNASA Apollo program. These providethe inertial references that allow thisinstrument to point at high accuracyeven when the aircraft is in motion.

greater distances than have beenachieved from ground observations dur-ing an eclipse or from Skylab observa-tions outside of eclipse. In the 1979 and1980 eclipses we were able to make two

We have made similar measurementsusing the same instrument at five timesduring the sun’s 1 l-year magnetic activi-ty cycle (1970, 1972, 1973, 1979, and1980). We thus have a very uniform setof data with which to compare variationsin coronal structures, brightness, elec-tron density, and material distribution asa function of solar activity. (See note



“The Changing Corona.”) For instance,standard observational models of coro-nal brightness indicate a marked vari-ation between maximum and minimumsolar activity. We are unique in beingable to verify this variation for the inner

Results of 1973 Electron

Density Experiments

In recent years the 1973 eclipse wasthe most widely observed both from theground and in space, Figures 1 la and bshow plots of absolute intensity of the

corona (K + F) as a function of radialdistance in the inner and outer corona,respectively. Also included in these plotsis the consensus model K + F brightnessbased on pre-1965 eclipse data. Figure1 la is a comparison of results from fiveindependent observations of the 1973eclipse: two from the ground, two fromspace coronagraphs, and one from theLos Alamos airborne expedition. Onlythe Los Alamos data extend over theentire range shown and beyond.

In comparing our results for the 1973eclipse with those of Skylab’s ApolloTelescope Mount, we were puzzled tofind the Skylab values of coronal bright-ness disturbingly higher than ours. Upon

1 0 . 0

8 . 0

6 . 0

4 . 0

2 . 0

1.5

1.0

0 . 8

0 . 6

0 . 4

1963

1963

1973

Equatorial Region

Newly observed by Los

Alamos 1979 and 1980~ \

(b) OUTER CORONA

4 6 8 10 12 15 20

r/R@

coronal brightness versus cospatial, cotemperal observations from two ground-based, oneand (b) the outer corona. airborne, and two space experiments. Only the Los Alamos

Fig. 11. Various observations ofintensity for (a) the inner coronaAlso shown in both cases is the brightness caleulated from a data cover the entire range of comparison, and the Los Alamosconsensus model [D. Blackwell, D. Denhirst, and M. Ingham, 1980 data, when reduced, will extend observations to 20 Ra

“The Zodiacal Light,” Advances in Astromony and for the first time since 1963.Astrophysics ,1 (1976)]. Figure 11a is a unique comparison of


careful examination of Skylab reductionprocedure. we discovered an error in

absolute calibration that resulted in a 90/0reduction of all Skylab values of coronalbrightness. This transfers roughly to asimilar reduction in all published resultsprior to 1978. The intensities plotted inFig. 1 la have been corrected for thiserror,

Figure 1 lb compares Los Alamosdata for the outer corona with the twobest recent observations of that region.

Figure 12 shows our 1973 resultscompared with those from K. Saito’sground-based observations.* Because ofsky background brightness. ground-based observations seldom are reliablebeyond 5 R@ in equatorial regions and 3R@ in polar regions, but the Los Alamosintensity data appear excellent beyond

Polarization measurements inthe equatorial regions are also good tothis distance, but over the sun’s poles theK corona is so faint that polarizationfalls below the limit of photographic

fainter corona in polar regions thanSaito does and therefore conclude thatelectron densities in polar coronal holesare lower than previously thought.

In equatorial regions the situation ismore complicated. We find coronal in-tensity to be lower, but we measure

measurements suggest that in equatorial

are higher and, more significantly forsolar wind models, do not fall off as fastas we proceed outward from the sun. Wealso observed this significant result at the1970 and 1972 eclipses.

We look forward to comparing elec-tron densities from the 1973 eclipse withthose of the 1980 eclipse as soon as our1980 data have been reduced and eval-uated.

Preliminary analysis of our 1980 in-tensity data shows that the corona wasabout three times brighter than it was in1973, and that polar regions are as

where the F corona begins to dominate.

*K. Saito, Annals of the Tokyo AstronomicaiObservatory 13, 93 (1972).

22

Fig. 12. Electron density as a function of radial distance determined fromphotographic polarization data gathered during the Los Alamos airborne expeditionof 1973. The results of K. Saito, a respected ground-based observer, are shown forcomparison. Photographic polarization data for the polar regions are reliable only to

the reduction in atmospheric contamination realized by our airborne expedition

Emission Line Experiment

We consider the coronal emission lineexperiment to be the most importantaboard the aircraft because the meas-ured line profiles contain a wealth ofinformation on the state of the coronalplasma. The shape of the line profilescan be analyzed to yield ion tem-peratures and nonthermal componentsof the velocity distribution. The vari-ations of line intensity with position, orwith time, contain information about themechanisms that excite the emitting ions.Bright features in or near the plane of thesky (plane through the sun’s center per-

pendicular to the line of sight) easilycontribute the major portion of ourmeasured signal along a line of sightbecause emission line intensities fall offvery rapidly with distance from the sun.Therefore we can attribute measuredsignals to specific coronal features with afair degree of confidence.

Since 1965, we have made airbornemeasurements of the Fe XIV green lineintensity and wavelength broadening us-ing a Fabry-Perot interferometer to ob-tain spectral (wavelength) resolution. Wechose an interferometer for several rea-sons. It is a relatively small instrumentwell suited to the space constraints

LOS ALAMOS SCIENCE

I THE SOLAR CORONA AT TOTALITY

Fig. 14. An artist’s sketch of the white light corona in the northwest quadrant duringthe May 30, 1965, total solar eclipse. The locations of the interferometer fringes areshown together with the position angles of two coronal features.

LOS ALAMOS SCIENCE

Fig. 13. Isodensity tracing of the in-terferometer image of the Fe XIVemission line obtained during the 1%5total solar eclipse. The heliocentric coor-dinates have been overlaid and the lunarlimb position is indicated.

aboard the aircraft. It also has a highspatial resolution, in part because itpreserves a two-dimensional image rath-er than the one-dimensional images ob-tained from grating spectrographs andother instruments that pass the incominglight through a slit. Perhaps most impor-tant, spectral distortions introduced bythe interferometer can be calculatedquite accurately and subtracted from themeasured line profiles, enabling us toachieve very high spectral resolution..

In 1965 we obtained the first Fe XIVline profiles out to 2 R@ from a photo-graphic interferometer image of the co-rona off the sun’s west limb (Fig. 13). Aradial trace from the center of the fringesystem through each fringe produces anintensity versus wavelength profileemitted by the corresponding region ofthe corona. Our spatial resolution waslimited by tracking to the order ofminutes of arc.

As shown in Fig. 14, these fringespassed through a helmet streamer, acoronal hole region, and an enhance-ment. From these data we estimated iontemperature, temperature gradients, and

made the first tentative determinations ofion temperatures in helmet streamers

23

Wavelength

Fig. D.

Fig. B.

THE FABRY-PEROTINTERFEROMETERT he heart of the emission line

camera is a Fabry-Perot in-terferometer. It resolves the light

from a particular atomic transition oc-curring in the corona into a high-resolution profile (emission line profile)of intensity versus wavelength.

The basic instrument consists of twoaccurately plane quartz plates main-tained strictly parallel at a constantdistance D. The opposing plate surfacesare coated to be highly reflecting. Theplates are enclosed in an airtightchamber and the space between them isfilled with gas under pressure. The indexof refraction n of the gas-filled space canbe varied by changing the gas pressure.

A beam of monochromatic light with

between the plates many times (Fig. A).The beam’s optical path length is 2nD

24

reflecting surfaces. When this pathlength is exactly equal to an integralnumber m of wavelengths, that is, when

(A)

then the reflected and incident lightwaves are exactly in phase and construc-tive interference (reinforcement) takesplace. Successive values of m correspondto successive orders of interference. Ifthe incident angle deviates even slightly

reduces the intensity because of the largenumber of reflections in a highly re-flective Fabry-Perot interferometer.Thus monochromatic light from an ex-tended source produces a series of con-centric bright rings corresponding tosuccessive orders of interference.

The coronal spectrum includes manywidely different wavelengths that caninterfere constructively for a given value

LOS ALAMOS SCIENCE

a prefilter to keep the long wavelengthend of one order from overlapping theshort wavelength end of the next order.The prefilter is centered on the coronalemission line being measured and trans-

Coronal emission lines are Dop-

full width at half-maximum intensity. If

then another wavelength in the profile,

slightly different incident angle 02 willalso interfere constructively (Fig. B).Thus successive wavelengths in the lineprofile are imaged by the interferometeras adjacent rings at successive values of

carefully makes a trace radially across a

LOS ALAMOS SCIENCE

ring to extract intensity versus 9, whichis easily converted by using Eq. (A) to

these profiles are deduced for an ex-tended region of the corona correspond-ing to less than the width of the broadinterferometer rings.

The partial rings produced during the1965 eclipse by a segment of the coronaare shown in Fig. 13 of the main text. In1965 we had time to take data only at afixed value of 2nD, so the only regions ofthe corona that could be studied werethose corresponding to the bright ringsin Fig. 13.

During the 1980 eclipse, our rapiddata-acquisition system gave us time tovary the gas pressure inside the in-terferometer, and thus to change n andcause the rings to traverse the entirecoronal image by changing the angle forconstructive interference. By thumbingthe corner of this journal, on which are

printed our 1980 photographs, the read-er can observe the movement of the ringscaused by the pressure scan. From thisseries of images corresponding to dif-ferent values of n, we can plot theintensity change at a single spatial pointas the gas pressure is varied with time.This plot of intensity versus n is con-

Our video recording system stored anarray of 250,000 spatial data pointsfrom each image; during a pressure scanof one order, 120 images were recorded.Each point can produce such a profile.One such profile is shown schematicallyin Fig. D. The line profile is repeated asthe pressure scan moves the ring corre-sponding to the succeeding order acrossthe point being plotted. The spatial reso-lution of a line profile deduced from asingle point is 4-5 arc seconds, which isan order of magnitude better than can beobtained from a radial trace ■

25

In 1973, we scanned the entire corona

15-20 arc seconds. We measured veryweak polarization of the green line alongradial vectors from the sun. Weakpolarization implies that green lineemission is excited by electron collisionsrather than absorption of photons, Thisresult confirms expectations that elec-tron collisions are the dominant mecha-nism by which energy is distributed inthe inner corona. We also observed thatCa XV yellow line emission occurredover a much larger region of the lowercorona than previously expected. Thepresence of the yellow line was thoughtto signify very hot regions of the corona,corresponding to the 4-5 MK ionizationpotential maximum of the Ca XV ioniza-tion state. but the 1973 line widthsindicate that the yellow line is producedin cooler regions as well, and thus thatcoronal temperatures may be muchmore uniform on a scale of arc minutesthan had been guessed from otheremission line measurements.

Our 1980 data is much more ex-tensive and has a much higher spatialresolution than that from previouseclipses.

Light from the 1980 experiment wascollected by the "Rube Goldberg,” amassive 10-inch telescope with an80-inch focal length that has been usedaboard the aircraft since our 1965 ex-pedition. Emission line signals, imagedby the telescope and interferometer, wereamplified and recorded on videotape at16-ins intervals from an image-intensified vidicon detector. This rapiddata acquisition system, a dramatic im-provement over the 30-second exposuretimes required by photographic tech-niques, collected well over 20,000 line

26

TRACKING CONTROL

MIRROR

Fig. 15. Main components of the tracking control for the coronal emission linecamera. Tracking along two orthogonal axes is controlled directly by an operator orby feedback from gyroscopes mounted on the telescope. Only one of these gyroscopes(L gyro) and its associated control mechanisms are shown here. Gyroscope drift in thetwo orthogonal axes is compensated by error signals from a set of photodiodes thatviews an image of the eclipsed sun. The tracking plane controller is preprogmmmedwith the various tracking patterns to be followed during totality. In addition, rotationof the telescope about its optic axis is required to correct not only for aircraft motionbut also for the apparent rotation of the sun during the duration of totality, whichamounted to nearly 2° during the 1980 eclipse. This rotation is controlled by feedbackfrom a gyroscope (R gyro) and by preset gyroscope precession in the R controller.Error signals, the difference between a gyroscope axis and the required axis, are about4-5 arc seconds for the telescope in flight. Similar performance accuracy has beendetermined from photographs and video recordings of lunar limb motion.

profiles from the base of the corona and

detector was designed specifically forastronomical measurements and forcompatibility with standard computerimage analysis equipment by M. T.Sandford. Its progenitor was developedat Los Alamos for diagnostics in theweapons program.

The line resolution and diameter of thedetector, combined with the relativelylarge image size (we view the corona insegments), determine a spatial resolutionof 4-5 arc seconds. But to realize thisresolution, we must keep the image cen-tered in the interferometer and on theaxis of the telescope, We accomplish this

by offsetting the entire 210-kg emissionline camera (telescope, interferometer,and detector) with large hydraulic ac-tuators that are controlled by a sophisti-cated tracking system accurate to 4-5arc seconds (Fig. 15). The hydraulicactuators are able to move the massiveinstrument at frequencies of up to 10 Hzto correct for higher-frequency aircraftmotions.

Data from the 1980 measurementsare much easier to reduce than thosefrom previous eclipses. Preliminary re-sults show wide variations in line shapesand are suggestive of the turbulent con-ditions that may be present during asolar maximum. We also see many de-

LOS ALAMOS SCIENCE


70r - 8 J

Column

Fig. 16. These preliminary 1980 Fe XIV emission line profiles were obtained bydigitizing a sequence of video frames and reordering the data to give a signal versusframe number (or wavelength) profile of the intensity at a single pixel (correspondingto a single spatial location in the corona). In (a) we have fitted both a Gaussian and aLorentzian profile to the emission line such that the full width at half-maximumintensity and the peak intensity agree with observations. The wings of the observed lineare not fitted by either profile. In (b) an example of a more complicated emission lineis shown. This profile could represent coronal material with significantly differentvelocities along the line of sight. A relative velocity of about 1.5 km/s would fit thisprofile. Some further corrections for intensity and wavelength calibration will beapplied to these data before our analysis is complete.

tails within small regions of high activity.However, data reduction must be com-pleted before we can make definitiveinterpretations.

Interpreting Emission Line Profiles

Both line shapes and line intensitiesare analyzed to learn about the condi-tions in the coronal plasma. Since theintrinsic line widths (from ions at rest)are extremely narrow, the observed Dop-pler-broadened line profiles are identicalin shape to the velocity distribution ofthe emitting ions.

For ions in thermal equilibrium, boththe velocity distribution and the emissionline profile would be Gaussian in shape.The width of the line profile would thenbe proportional to the kinetic tem-

LOS ALAMOS SCIENCE

perature of the emitting ions. However,ion temperatures deduced directly fromobserved Gaussian line profiles (for ex-ample, the iron temperatures shown inFig, 5) may be too high because non-thermal effects, such as macroscopicturbulence and magnetic wave accelera-tion of the solar wind, may add to thevelocities of the ions and, in turn, to theline broadening.

An approximate expression for theline width in terms of the kinetic tem-perature T and the average turbulentmacroscopic velocity Vt is given by

profile at half-maximum intensity and m

is the mass of the emitting ion. Note thatthe kinetic temperature contribution de-pends on the mass of the ion, whereasthe contribution from macroscopic tur-bulence is mass independent. Thereforeit is possible to separate thermal andturbulent contributions to the line broad-ening by measuring emission line profilesof two ions with widely differing masses.But these measurements must be madeat the same time and at the same place inthe corona so that we can assume thatthe two ions have the same kinetictemperature and average turbulent veloc-ity Vt.

Our Ca XV data from the 1980eclipse may be appropriate for com-parison with our Fe XIV measurements,but the emission was weaker than ex-pected and so the profiles may not beaccurate enough. However, the Lyman-adata recorded by rocket experiments atseveral places in the corona will certainlybe useful in determining nonthermal con-tributions to our Fe XIV line profiles.

Variations in line shape are also in-dications of nonthermal velocities. Forexample, calculations show that largeexpansion velocities from solar windflow tend to flatten and extend whatwould have been a Gaussian line shape.Although these departures are not signif-icant until expansion velocities are 40km/s or greater, such velocities are pre-dicted by some models of solar wind

profile shapes are also altered bylarge-scale turbulent motion or signifi-cant magnetic wave acceleration of theplasma.

Figure 16 illustrates the variety of lineshapes that were actually measured in1980. We see asymmetries, as well asdepartures from Gaussian shapes. In-

27

terpretation of these line shapes dependson theoretical assumptions and support-ing data, except perhaps for the case ofsplit lines, which clearly indicate relativebulk motion of the emitting material.

Analysis of line intensities can also bequite revealing. Since emission intensitydepends on density and temperature,comparison with independent electrondensity measurements is helpful in de-ciding whether bright regions are due tohigh temperatures or, alternatively, tohigh densities. We can also analyzeintensity variations as a function ofradial distance to infer excitation mecha-nisms. If ions are excited by electroncollisions, intensities are proportional tothe square of the electron density (ne)

2

and thus should decrease as l/r12. Onthe other hand if photon excitation is thedominant mechanism, then line in-tensities are proportional to ne, and thusdecrease as l/r6.

Finally we can try to find periodicvariations in intensity and line shape toidentify wave heating of the corona. Wehave recorded tentative evidence of peri-odic variations in some of our 1973 datataken aboard the Concorde during arecord-setting 74 minutes of totality(Fig. 17).

Emission Line DataResults and Questions

Our 1965 results (see Fig. 5 for aver-age ion temperature values) showed iontemperatures decreasing outward fromthe base of the corona with little vari-ation in temperature from one coronalfeature to another. A more detailed lookat the data in each feature is illuminat-ing.

28

H E L M E T S T R E A M E R . One inter-ferometer fringe (see Figs. 13 and 14)crossed a helmet streamer near the top

begin to change from a closed loop to anopen-line configuration extending radi-ally from the sun. Since particles in thehelmet are presumably trapped along

magnetic field lines, the helmet shouldhave above-average temperatures. Inparticular, along a line through the topof the helmet we should see a tem-perature increase in the helmet and atemperature decrease on either side.However, our 1965 data (Fig. 18) showthat temperatures along this line fluc-

0.8

0.4

00 5 10 15

Time (rein)

Fig. 17. Peak intensity versus time for two 15-minute segments of Fe XIV emission linedata obtained aboard the Concorde 001 at the June 30, 1973, total eclipse. The data

variations with a periodicity of about 5 minutes corresponding perhaps to a coronalresponse to the 5-minute oscillations of the photosphere. Tracking problems in-troduced the errors shown by the bars. The data in (b) are from an equatorial bright

suggests the presence of a shock disturbance. The variations have a periodicity ofabout 6 minutes.

LOS ALAMOS SCIENCE


Fig. 18. Temperatures across a helmet structure deduced from Fe XIV emission linedata obtained during the May 30, 1%5, total solar eclipse.

Fig. 19. Intensity versus wavelength profile of the coronal Fe XIV emission lineobtained during the May 30, 1965, total solar eclipse. The maximum intensity IQ is at

tuate from about 2.5-3.0 MK withoutmuch systematic change. The Gaussianline shapes from the helmet (Fig. 19)suggest that, as expected, turbulent ve-locities are small, perhaps less than 25km/s in this long-lived magnetically con-fined region.

Our 1980 data will provide a moredetailed look at the temperature struc-ture across a helmet and will extend ourview of the region above the helmet

LOS ALAMOS SCIENCE

where the field lines open to form astreamer. We are particularly interestedin knowing whether temperature gra-dients above the helmet are consistentwith solar wind flow. At present thecontribution of streamer regions to thesolar wind is poorly known.

CORONAL HOLES. Several fringes ofour 1965 data crossed a coronal holeregion, so we were able to estimatetemperature gradients as well as tem-

peratures in this region. Coronal holesare expected to have lower temperaturesand densities than other regions, andsince they are known to be a source ofsolar wind flow, we expect temperaturegradients consistent with solar wind the-ory.

One surprise was the high tem-

much higher than those determined forthe base of the coronal holes by the EUVmeasurements.

Our temperatures, which we estimateddirectly from line broadening, would beslightly less but still much higher thanthose determined from other experi-ments, if turbulent velocities contributeto the line broadening. In fact, we haveinvoked turbulent velocities in someparts of this region to deduce tem-perature gradients consistent with solarwind flow. If the temperature at the

fixed at 2.5 MK, then successive tem-peratures at 1.62 R@ and 1.74 R@ are2.43 MK and 2.39 MK, respectively,provided the turbulent velocities at thesethree altitudes are 35, 10, and 5 km/s,respectively. The temperature variation

We also determined the variation ofline intensity with radius in this coronalhole region. Figure 20 shows that in-tensities fluctuate rapidly near the baseof the corona, but at greater heights theintensity is proportional to l/r12, or (ne)

2.This rapid falloff is consistent with acollisional excitation mechanism—arather surprising result given the verylow density in this region determined byindependent electron density measure-ments. However, if magnetic field linesare clustered, rather than uniformly dis-

29

tributed, and if electrons are trappedalong these clustered lines, then the aver-age electron density would remain low,as measured along a line of sight, butlocal densities would be high enough toproduce ion excitations by collisionalprocesses. Evidence for density in-homogeneities is seen on white lightphotographs of the corona, but whetherthese inhomogeneities correspond toclustering of field lines is not at allcertain. More detailed intensity measure-ments from the 1980 eclipse may clarifythe situation.

CORONAL ENHANCEMENT. A coronalenhancement is a long-lived region ofhigher density that overlies an activeregion in the photosphere. Consequently,we expect higher turbulent velocities insuch a region than in helmet streamers.From our 1965 data we estimated tem-peratures of -3.0 MK after subtractingturbulent velocity contributions to thebroadening. Turbulent velocities in theenhancement are about 25 kmls anddecrease with altitude. Temperature gra-dients. when corrected for turbulent ve-locity broadening, suggest that solarwind flow is probably very weak in thisregion.

We also determined intensity variation

Electron density decreases by a factor of10 over this region. As seen in Fig. 20intensity falls off as l/r12 or as (ne)

2;collisional excitation, as expected, domi-nates in this relatively dense region.

THE 1980 CONDENSATION. One of themost exciting observations during the1980 eclipse was the hydrodynamiceruptive disturbance from the base of the

a first look at our Fe XIV emission linedata we have been able to identify a very

30

Fig. 20. Emission line intensity versus radial distance for (a) a coronal enhancementand (b) a coronal hole region. Data were obtained during the May 30, 1965, totalsolar eclipse. Letters label lines through data taken at the same position angle. Curvesof I/r” (n = IO and 12) are shown for comparison.

bright region on the west limb (Fig. 21)whose position angle is near that de-termined for the eruptive disturbancephotographed by the coronal camera.We recorded this region a second timewhen we were looking for the Ca XVyellow line. The bright knot appearsagain, this time in the K coronal lightfrom Thompson scattering that is ac-cepted by the interferometer along withthe yellow line (Fig. 22). The K coronalbrightness in this region indicates highelectron densities and so the Fe XIVbrightness is probably related to in-creased density and not just to a tem-perature close to the Fe XIV ion popu-

lation maximum. This region has theright size and density enhancement toqualify as a coronal condensation. Ifother observations taken a day or twolater show that this condensation issporadic or short-lived, we mightspeculate that perhaps many such con-densations are related to the develop-ment of hydrodynamic eruptive dis-turbances. Then previous observationsof coronal condensations might be usedto determine the frequency of such masseruptions throughout the solar activitycycle.

NEW CORONAL IMAGING TECH-

NIQUE. The emission line data from the

LOS ALAMOS SCIENCE

scanning interferometer can be used toobtain a picture of the corona in the lightof an emission line. Although suchphotographs have been obtained withexpensive Lyot filters specially designedfor operation at one wavelength, the newtechnique we present here is cheaper andquite adaptable to a variety of wave-lengths. A pressure scan over one orderof the interferometer is integrated onto asingle photograph. As shown in Fig. 23,details of the base of two helmetstreamers, the bright knot, and otherfeatures on the west limb can be seen onsuch an image.

Electron Temperature Experiment

Coronal electron temperatures as afunction of radial distance are badlyneeded to determine whether electronsand ions are in thermal equilibrium. In1976 L. E. Cram* suggested a method todetermine these temperatures from spec-tral intensities of the K corona. Our1980 attempt to perform the measure-ment used a much more sensitive tech-nique than has been tried before and,although we failed to record data, we didconfirm the feasibility of the technique.

The principle behind the experiment isbased on Cram’s analysis of the Kcoronal spectrum. This spectrum isformed as light coming from thephotosphere (the Fraunhofer spectrum)is scattered by energetic coronal elec-trons. The Doppler shifts suffered bythese scattered photons broaden theabsorption lines in the original Fraun-hofer spectrum to produce the diffusespectrum known as the K corona.

Cram’s calculated spectra for aspherically symmetric, isothermal coro-na at four electron temperatures areshown in Fig. 24. His results display twovery interesting features. First, there ex-ist spectral “nodes,” that is, wavelengthsat which the intensity is independent ofthe assumed electron temperature. Thesenodes occur on either side of the H andK absorption lines of Ca II and near

*L. E. Cram, “Determination of the Temperatureof the Solar Corona From the Spectrum of theElectron-Scattering Continuum, ” Solar Physics48, 3-19 ([976).

32

two helmet streamers and a bright knot.

other absorption features of the Fraun-hofer spectrum. Second, the spectra un-dulate about these nodes with anamplitude that is related to the electrontemperature. Thus, the ratio of twomeasured intensities, one at a nodalwavelength and one at a wavelengthnearby, should determine the electrontemperature uniquely.

Although such an experiment is verysimple in principle, in practice it is quitedifficult because the intensities must bemeasured with great precision. An error

electron temperature. The 5- 1OO/o errorscommon in standard photometric tech-niques would obliterate the informationsought (as learned by Menzel andPasachoff* who searched without suc-cess for residual depressions from the H

and K absorption lines of CA II inphotographic records of absorptionspectra from the 1936 total eclipse).

To achieve the required sensitivity, webuilt our instrument around the onlytype of detector suitable for the task—an array of silicon photodiodes. Thesesolid-state photosensitive devices havevery high quantum efficiency in thespectral range of interest.

A commercially available detector,consisting of a linear array of512 siliconphotodiodes on a single integratedcircuit, served as the main component ofour K corona spectrograph (Fig. 25).This large array enabled us to look notju s t a t two wave l eng ths bu t a t

*D. H. Menzel and J. M. Pasachoff. “On theObliteration of Strong Fraunhofer Lines by Elec-tron Scattering in the Solar Corona, ” Publica-tions of the Astronomical Society of the Pacific80, 458 (1968).

LOS ALAMOS SCIENCE


Fig. 24. Absolute intensity of the K coronal spectrum for four coronal electrontemperatures, as calculated by Cram. The absolute intensity of the photosphericspectrum from the center of the sun's disc is shown for comparison.

ENTRANCESLITS

CASSEGRAINTELESCOPE

LOS ALAMOS SCIENCE

Fig. 25. The K corona spectrograph forthe electron temperature experiment con-sists of a telescope, entrance slit as-sembly, a transmission grating spectro-graph, a silicon photodiode detector ar-ray, and electronics for data acquisitionand storage. The f/2.0, 400-mm focal

length telescope maximizes the light atthe detector. Three entrance slits, corre-sponding to radial distances from the

by photolithographic techniques on aglass disc. A computer-controlledlever-arm and motor-drive assemblypositions the selected entrance slit in thetelescope focal plane. The entrance slitlength is magnified by the spectrographto match and thereby utilize the entirephotodiode length. The spectrograph

per channel. Exposure times range from

be obtained only during eclipses of rela-tively long duration or from aircraftwhose flight path lengthens the durationof totality.

33

wavelengths of the K coronal spectrum

The experiment was in some sense afishing expedition. We set out to observethe entire spectrum with high precisionand search for Cram’s predicted undula-tion patterns at three radial distancesfrom the sun.

During the eclipse, the spectrographappeared to function well, and we wereable to confirm that its sensitivity isadequate to measure the shape of the Kcoronal spectrum and, in principle, todetermine electron temperatures. Thedata that appeared on our monitoringscreen were not recorded by our com-puterized data-acquisition system (Fig.26) because both the computer diskdrive and the back-up drive failed tofunction.

Indeed the equipment failure was atremendous disappointment, but we didprove that the experiment was feasible,and are now redesigning some of theelectronics with hopes of flying againduring the 1983 solar eclipse.

Infrared Observations of Dust Rings

The F corona and the zodiacal light, azone of scattered light symmetric aboutthe approximate plane of the planetaryorbits, give clear evidence of the pres-ence of dust throughout much of thesolar system. These observations alsoindicate that the dust particles are veryf ine—only a few micrometers indiameter. Their presence may be a rem-nant of the formation of the solar systembut probably wandering comets havealso left a contribution.

Whatever the source, theorists havemodeled the fate of the dust once it isdeposited in the solar system. The

34

Fig. 26. Block diagram of computer-controlled data-acquisition and storage systemfor electron temperature experiment. The detector array was read out with ReticonCorporation’s evaluation circuitry. Although not ideal for this application, thecircuitry provided a signal-to-noise ratio of greater than 100. Data were to be storedin near real time on floppy disks. Computer programs were written in FORTH, alanguage developed at the University of Rochester.

dynamics of the dust particles is in-fluenced by four factors: the sun’s grav-ity, radiation pressure, the Poynting-Robertson effect, and particle evapora-tion. Calculations including these factorsindicate that the dust will spiral intoward the sun but may ultimately settleat particular distances and form broadrings about the sun. Evaporation of thedust may also lead to its being blownoutward once again. By determining theradial location of these rings we learnsomething about the dust’s compositionbecause its location depends on its re-

flectivity and evaporation temperature.Likely compositions include obsidian,silicate-type rock, iron, and perhaps wa-ter-ice.

The presence of glowing dust rings at4, 9, and 20 R@ and elsewhere has beendeduced from observations of infraredemission features (Fig. 27). Theseemission features are presumably pro-duced as dust particles heat up andevaporate, emitting radiation corre-sponding to black-body temperatures be-tween 500 and 2000 K. These observa-tions are difficult to make from the

LOS ALAMOS SCIENCE


ground and, with the exception of the

features is in dispute.From aboard the aircraft during the

1980 eclipse, we made a new attempt todetermine the spatial distribution of in-frared emission features.

Infrared intensity as a function ofdistance from the sun was measured at2.2 µm with an InSb detector, a standardtechnique, and at -0.7 µm with acharge-injection-device television cam-era.

The corona was scanned along theecliptic (plane of the earth’s orbit) out to50 R@ east and west of the sun and atseveral angles to the ecliptic. The scanpattern was designed to determine theradial location and approximate extentof any dust rings.

Both detection systems worked welland yielded high-quality data. However,tracking errors during totality will se-verely compromise interpretation of dataclose to the sun, so we will not be able to

LOS ALAMOS SCIENCE

Fig. 27. Approximate radial location and extent of infrared emission features reported

confirm the existence of the feature at 4

errors have less effect.Our analyses to date strongly suggest

indicate no other features at greaterdistances from the sun. Thus we find no

location, which corresponds to a tem-

by E. P. Nye on the basis of ground-levelobservations at the February 26, 1979,and February 16, 1980, eclipses; how-ever, at both eclipses his observationsmay have been contaminated by theeffects of clouds. Existence of such afeature would be of interest because dustshells with temperatures of -800 K havebeen found about some other stars.

Conclusions

Most scientists studying the solar co-rona believe that a temperature max-

indicate that the maximum might becloser to the limb, although some of theFe XIV emission line broadening canalways be attributed to large-scale tur-bulence or to expansion velocities. How-ever, comparison with proton tem-perature values from the rocket-bornecoronagraph at a few positions in thecorona will allow us to determine suchvelocities if they exist. We can theninterpret the rest of our iron emissionline data in light of these findings. Thefact that many of our 1980 profiles areactually double-peaked indicates largedifferential mass motions that are proba-bly due to the extremely complex natureof coronal features during maximumsolar activity,

Ion temperature gradients measuredin 1965 indicated that coronal holescould support solar wind flow. We nowhave high-resolution data from our 1980measurements of the residual coronalhole above the sun’s south pole to com-pare with the 1965 results. We will also

35

be able to analyze temperature gradientsand velocity distributions in streamerstructures, to determine whether thesestructures also contribute to solar windflow.

Our emission line intensity data fromseveral previous solar eclipse observa-tions suggested that collisional mecha-nisms dominate energy transport in the

resolution 1980 data from severalstreamer structures and one coronal con-densation, when combined with electrondensities determined from camera-polarimeter measurements, will be a bet-ter test of this model.

The possibility of extended wave heat-ing in the corona has become increasing-ly significant. Our emission line intensitydata from the 1973 Concorde flight,which gave 74 minutes of totality, pro-vided tentative evidence for periodic tem-perature variations, and in 1980 welooked for temporal changes on the sun’swest limb by taking a sequence of in-tensity measurements at 30-second in-tervals over the 7 minutes of totality. Buta more extensive search will be requiredto establish the dynamics and the extentof plasma heating.

Comparison of the computer-enhanced coronal photographs from fiveeclipses shows that the corona at the1980 eclipse exhibited features neverbefore seen, including the umbilical con-nection of the hydrodynamic eruptivedisturbance to the solar surface. Thisobservation calls into question previousspeculation, based on Skylab photo-

tion to the surface was quite broad.Evidently the energy source for somemass eruptions is localized and concen-trated. We may learn more about thedynamics at the base of this disturbanceonce our electron density and Fe XIVdata have been reduced.

Our photographic record also showsthat nearby streamers respond to theeruption by bending around it, an effectalso seen on Skylab photographs but

long streamers above the sun’s southpole suggests that a second eruptivedisturbance may also have been inprogress at the same time.

Electron densities resulting from ourobservations agree with other studies in

equatorial regions the density decreasewith radial distance is much more gradu-al than indicated previously. This couldhave a marked effect on theoretical mod-els of the solar wind.

In summary, we now have collectedion temperature and electron densitydata at all phases of solar activity exceptduring a deep minimum. Our 1980 datacombined with proton temperaturesfrom rocket experiments will provide afocal point for all future analysis. For thefirst time, we will have an accuratedetermination of material density, iontemperatures, and nonthermal velocitiesat several places in the corona. We hopethese results will be a basis for moremeaningful interpretation of our datathroughout the rest of the inner corona.


Donald H. Liebenberg earned his Bachelor of Science in 1954 and

his Ph.D. in physics in 1971, both from the University of Wisconsin.He has been a Staff Member at LoS Alamos since 1961. His interest insolar physics, which dates to his undergraduate days, led in 1963 to theLaboratory’s involvement in observations of total solar eclipses. This

activity has resulted in detailed studies over a solar cycle of the physicalproperties and energy transfer mechanisms of the coronal plasma.During 1967 and 1968 he was Program Director for Solar TerrestrialResearch at the National Science Foundation. In 1973 he received aNational Science Foundation Grant for observation of a solar eclipsefrom the French-British Concorde 001. His other research interestsinclude properties of gases at high pressures and laser optics andtechnology.

Charles F. Keller, Jr., is well known for his involvement in airborneeclipse expeditions, having served as principal investigator ofpolarization experiments on missions in 1970, 1972, 1973, 1977, and1980; he was scientific coordinator for missions mounted in 1979 and1980. After receiving his Bachelor of Arts from St. Vincent College in1961, he went on to earn a Bachelor of Science in physics fromPennsylvania State University and a Master of Science and Ph.D. inastronomy from Indiana University in 1967 and 1969, respectively. Histhesis, “Computer Models of Pulsating Stars Including RadiativeTransfer,” was produced jointly at Los Alamos and Indiana University.He joined the Laboratory in 1969 and has served as a Staff Member,Assistant Group Leader, and Group Leader of the Diagnostics DesignGroup of the Field Testing Division. At present he is with theLaboratory’s Computer Modeling Group of tbe Geosciences Division.Keller served as the Field Testing Division’s representative on theDirector’s Computer Advisory Committee, and as the Laboratory’scoordinator of computer modeling for the Department of Energy’sAtmospheric Studies over Complex Terrain (ASCOT) Program. He isa member of the American Astronomical Society and its Solar PhysicsDivision.

Further Reading

E. N. Parker, Interplanetary Dynamical Processes (Wiley-Interscience, NewYork, 1963),

D. E. Billings, A Guide /o the Solar Corona (Academic Press. New York.1966).

H. Zirin, The Solar Atmosphere (Blaisdell Publishing Co., Waltham,Massachusetts, 1966).

J. C, Brandt, Introduction to the Solar Wind (W. H, Freeman and Co., SanFrancisco, 1970).

R. G. Athay, The Solar Chromosphere and Corona: Quiet Sun (D. ReidelPublishing Co., Dordrecht, Holland. 1976).

J. B. Zirker. Ed., Coronal Holes and High Speed Wind Streams (ColoradoAssociated University Press. Boulder, 1977).

E. N. Parker, Cosmical Magnetic Fields: Their Origin and Their Activity

(Clarendon Press, Oxford. 1979).

J. B. Zirker. "Total Eclipses of the Sun,” Science 210, 1313-1319 (1980).


the solar corona at totality · 2016-10-21 · the solar corona at totality served for all of...

Documents