[Scientific American] Mysteries of the Milky Way ((BookFi.org)

COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.

http://www.sciam.com/index.cfm?ref=digitalpdf

2

TABLE OF CONTENTS

ScientificAmerican.comexclusive online issue no. 15

MYSTERIES OF THE MILKY WAY

Viewed from above, it appears as a giant pinwheel of diamond-studded smoke. From Earth, it cuts a paleswath of light across the summer sky. This is the Milky Way, our home galaxy, and it has captivatedhumankind for millennia. Galileo was the first to formally probe its secrets. Armed with a telescope ofhis own devising, he determined that the shimmering band was in fact composed of stars. Astronomershave been smitten ever since.

This exclusive online issue brings together stunning Milky Way discoveries from the past decade.Leading scientists explain how our galaxy formed, how it continues to evolve and why only part of it ishabitable. Other articles home in on particular Milky Way elements--our paradoxical sun, cosmic dustand the dynamic interstellar medium. Of course, as fascinating as our galactic neighborhood is,astronomers also yearn to peer beyond it. These efforts, too, are summarized here. We hope that afterreading this issue, you’ll see our spectacular corner of the cosmos in a different light. —The Editors

How the Milky Way FormedBY SIDNEY VAN DEN BERGH AND JAMES E. HESSER; JANUARY 1993Its halo and disk suggest that the collapse of a gas cloud, stellar explosions and the capture of galactic fragments may have allplayed a role

Our Growing, Breathing GalaxyBY BART P. WAKKER AND PHILIPP RICHTER; JANUARY 2004Long assumed to be a relic of the distant past, the Milky Way turns out to be a dynamic, living object

Refuges for Life in a Hostile UniverseBY GUILLERMO GONZALEZ, DONALD BROWNLEE AND PETER D. WARD; OCTOBER 2001 Only part of our galaxy is fit for advanced life

The Paradox of the Sun's Hot CoronaBY BHOLA N. DWIVEDI AND KENNETH J. H. PHILLIPS; NEW LIGHT ON THE SOLAR SYSTEM, SPRING 2003Like a boiling teakettle atop a cold stove, the sun's hot outer layers sit on the relatively cool surface. And now astronomers arefiguring out why

The Gas between the StarsBY RONALD J. REYNOLDS; JANUARY 2002Filled with colossal fountains of hot gas and vast bubbles blown by exploding stars, the interstellar medium is far more interest-ing than scientists once thought

The Secrets of StardustBY J. MAYO GREENBERG; DECEMBER 2000Tiny grains of dust floating in interstellar space have radically altered the history of our galaxy

Galaxies behind the Milky WayBY RENEÉ C. KRAAN-KORTEWEG AND OFER LAHAV; OCTOBER 1998Over a fifth of the universe is hidden from view, blocked by dust and stars in the disk of our galaxy. But over the past few years,astronomers have found ways to peek through the murk

1 SCIENTIFIC AMERICAN EXCLUSIVE ONLINE ISSUE JULY 2004

25

7

17

33

43

48


http://www.sciam.com/index.cfm?ref=digitalpdf

Attempts to reconstruct how theMilky Way formed and beganto evolve resemble an archaeo-

logical investigation of an ancient civi-lization buried below the bustling centerof an ever changing modern city. Fromexcavations of foundations, some pot-tery shards and a few bones, we mustinfer how our ancestors were born, howthey grew old and died and how theymay have helped create the living cul-ture above. Like archaeologists, astron-omers, too, look at small, disparateclues to determine how our galaxy andothers like it were born about a billionyears after the big bang and took ontheir current shapes. The clues consistof the ages of stars and stellar clusters,their distribution and their chemistry—all deduced by looking at such featuresas color and luminosity. The shapes andphysical properties of other galaxies canalso provide insight concerning the for-mation of our own.

The evidence suggests that our gal-axy, the Milky Way, came into being asa consequence of the collapse of a vastgas cloud. Yet that cannot be the wholestory. Recent observations have forcedworkers who support the hypothesis ofa simple, rapid collapse to modify theiridea in important ways. This new infor-

mation has led other researchers to pos-tulate that several gas cloud fragmentsmerged to create the protogalactic Mil-ky Way, which then collapsed. Other var-iations on these themes are vigorouslymaintained. Investigators of virtually allpersuasions recognize that the births ofstars and supernovae have helped shapethe Milky Way. Indeed, the formationand explosion of stars are at this mo-ment further altering the galaxy’s struc-ture and influencing its ultimate fate.

Much of the stellar archaeo-logical information thatastronomers rely on to deci-

pher the evolution of our galaxy residesin two regions of the Milky Way: thehalo and the disk. The halo is a slowlyrotating, spherical region that sur-rounds all the other parts of the galaxy.The stars and star clusters in it are old.The rapidly rotating, equatorial regionconstitutes the disk, which consists ofyoung stars and stars of intermediateage, as well as interstellar gas and dust.Embedded in the disk are the sweeping-ly curved arms that are characteristic ofspiral galaxies such as the Milky Way.Among the middle-aged stars is our sun,which is located about 25,000 light-years from the galactic center. (Whenyou view the night sky, the galactic cen-ter lies in the direction of Sagittarius.)The sun completes an orbit around thecenter in approximately 200 millionyears.

That the sun is part of the Milky Waywas discovered less than 70 years ago.At the time, Bertil Lindblad of Swedenand the late Jan H. Oort of the Nether-lands hypothesized that the Milky Waysystem is a flattened, differentially rotat-ing galaxy. A few years later John S.Plaskett and Joseph A. Pearce of Do-minion Astrophysical Observatory ac-cumulated three decades’ worth of dataon stellar motions that confirmed theLindblad-Oort picture.

In addition to a disk and a halo, theMilky Way contains two other subsys-tems: a central bulge, which consistsprimarily of old stars, and, within thebulge, a nucleus. Little is known aboutthe nucleus because the dense gasclouds in the central bulge obscure it.The nuclei of some spiral galaxies, in-cluding the Milky Way, may contain alarge black hole. A black hole in the nu-cleus of our galaxy, however, would notbe as massive as those that seem to actas the powerful cores of quasars.

All four components of the MilkyWay appear to be embedded in a large,dark corona of invisible material. Inmost spiral galaxies the mass of this in-visible corona exceeds by an order ofmagnitude that of all the galaxy’s visiblegas and stars. Investigators are intenselydebating what the constituents of thisdark matter might be.

The clues to how the Milky Way de-veloped lie in its components. Perhapsthe only widely accepted idea is that thecentral bulge formed first, through thecollapse of a gas cloud. The centralbulge, after all, contains mostly massive,old stars. But determining when andhow the disk and halo formed is moreproblematic.

In 1958 Oort proposed a model ac-cording to which the population of starsforming in the halo flattened into athick disk, which then evolved into athin one. Meanwhile further condensa-tion of stars from the hydrogen left overin the halo replenished that structure.Other astronomers prefer a picture inwhich these populations are discreteand do not fade into one another. Inparticular, V. G. Berman and A. A.Suchkov of the Rostov State Universityin Russia have indicated how the diskand halo could have developed as sepa-rate entities.

These workers suggest a hiatus be-tween star formation in the halo andthat in the disk. According to their mod-

SIDNEY VAN DEN BERGH andJAMES E. HESSER both work at Domin-ion Astrophysical Observatory, NationalResearch Council of Canada, in Victoria,British Columbia. Van den Bergh has alongtime interest in the classification andevolution of galaxies and in problems relat-ed to the age and size of the universe. Hereceived his undergraduate degree fromPrinceton University and a doctorate in as-tronomy from the University of Göttingen.Hesser’s current interests focus on the agesand compositions of globular star clusters,which are among the oldest constituents ofthe galaxy. He received his B.A. from theUniversity of Kansas and his Ph.D. in atom-ic and molecular physics from Princeton.

How the Milky Way FormedIts halo and disk suggest that the collapse

of a gas cloud, stellar explosions and the captureof galactic fragments may have all played a role

by Sidney van den Bergh and James E. Hesseroriginally published in January 1993



el, a strong wind propelled by superno-va explosions interrupted star forma-tion in the disk for a few billion years.In doing so, the wind would have eject-ed a significant fraction of the mass ofthe protogalaxy into intergalactic space.Such a process seems to have prevailedin the Large Magellanic Cloud, one ofthe Milky Way’s small satellite galaxies.There an almost 10-billion-year inter-lude appears to separate the initial burstof creation of conglomerations of oldstars called globular clusters and themore recent epoch of star formation inthe disk. Other findings lend additionalweight to the notion of distinct galactic

components. The nearby spiral M33contains a halo but no nuclear bulge.This characteristic indicates that a halo isnot just an extension of the interior fea-ture, as many thought until recently.

In 1962 a model emerged that servedas a paradigm for most investigators. Ac-cording to its developers—Olin J. Eggen,now at the National Optical Astronomi-cal Observatories, Donald Lynden-Bellof the University of Cambridge and Al-lan R. Sandage of the Carnegie Institu-tion—the Milky Way formed when alarge, rotating gas cloud collapsed rap-idly, in about a few hundred millionyears. As the cloud fell inward on itself,the protogalaxy began to rotate morequickly; the rotation created the spiralarms we see today. At first, the cloudconsisted entirely of hydrogen and heli-um atoms, which were forged duringthe hot, dense initial stages of the bigbang. Over time the protogalaxy startedto form massive, short-lived stars. Thesestars modified the composition of galac-tic matter, so that the subsequent gener-ations of stars, including our sun, con-tain significant amounts of elementsheavier than helium.

Although the model gained wide ac-ceptance, observations made during thepast three decades have uncovered anumber of problems with it. In the first

place, investigators found that many of the oldest stars and star clusters inthe galactic halo move in retrograde or-bits—that is, they revolve around thegalactic center in a direction opposite tothat of most other stars. Such orbitssuggest that the protogalaxy was quiteclumpy and turbulent or that it capturedsizable gaseous fragments whose matterwas moving in different directions. Sec-ond, more refined dynamic modelsshow that the protogalaxy would nothave collapsed as smoothly as predictedby the simple model; instead the densestparts would have fallen inward muchfaster than more rarefied regions.

Third, the time scale of galaxy forma-tion may have been longer than that de-duced by Eggen and his colleagues. Ex-ploding supernovae, plasma windspouring from massive, short-lived starsand energy from an active galactic nu-cleus are all possible factors. The galaxymay also have subsequently rejuvenateditself by absorbing large inflows of pris-tine intergalactic gas and by capturingsmall, gas-rich satellite galaxies.

Several investigators have attempt-ed to develop scenarios consistent withthe findings. In 1977 Alar Toomre ofthe Massachusetts Institute of Technol-ogy postulated that most galaxies formfrom the merger of several large piecesrather than from the collapse of a singlegas cloud. Once merged in this way, ac-cording to Toomre, the gas cloud col-lapsed and evolved into the Milky Waynow seen. Leonard Searle of the Car-negie Institution and Robert J. Zinn ofYale University have suggested a some-what different picture, in which manysmall bits and pieces coalesced. In thescenarios proposed by Toomre and bySearle and Zinn, the ancestral fragmentsmay have evolved in chemical-ly unique ways. If stars began to shineand supernovae started to explode indifferent fragments at different times,then each ancestral fragment would

have its own chemical signature. Recentwork by one of us (van den Bergh) indi-cates that such differences do indeed ap-pear among the halo populations.

Discussion of the history of ga-lactic evolution did not ad-vance significantly beyond this

point until the 1980s. At that time,workers became able to record moreprecisely than ever before extremelyfaint images. This ability is critically im-portant because the physical theories ofstellar energy production—and hencethe lifetimes and ages of stars—are mostsecure for so-called main-sequence stars.

Such stars burn hydrogen in their cores;in general, the more massive the star, themore quickly it completes its main-se-quence life. Unfortunately, this factmeans that within the halo the only re-maining main-sequence stars are the ex-tremely faint ones. The largest, most lu-minous ones, which have burned pasttheir main-sequence stage, became in-visible long ago. Clusters are generallyused to determine age. They are crucialbecause their distances from the earthcan be determined much more accuratelythan can those of individual stars.

The technology responsible for open-ing the study of extremely faint halostars is the charge-coupled device (CCD).This highly sensitive detector producesimages electronically by converting lightintensity into current. CCDs are far su-perior in most respects to photographicemulsions, although extremely sophisti-cated software, such as that developedby Peter B. Stetson of Dominion Astro-physical Observatory, is required totake full advantage of them. So used,the charge-coupled device has yielded atenfold increase in the precision of mea-surement of color and luminosity of thefaint stars in globular clusters.

Among the most important results ofthe CCD work done so far are moreprecise age estimates. Relative age data

Of course, it is possible that morethan one model for the formationof the galaxy is correct.



based on these new techniques have re-vealed that clusters whose chemistriessuggest they were the first to be createdafter the big bang have the same age towithin 500 million years of one another.The ages of other clusters, however, ex-hibit a greater spread.

The ages measured have helped re-searchers determine how long it tookfor the galactic halo to form. For in-stance, Michael J. Bolte, now at Lick Ob-servatory, carefully measured the colorsand luminosities of individual stars inthe globular clusters NGC 288 andNGC 362 [see illustration below]. Com-parison between these data and stellarevolutionary calculations shows thatNGC 288 is approximately 15 billionyears old and that NGC 362 is onlyabout 12 billion years in age. This dif-ference is greater than the uncertaintiesin the measurements. The observed agerange indicates that the collapse of theouter halo is likely to have taken an or-der of magnitude longer than theamount of time first envisaged in thesimple, rapid collapse model of Eggen,Lynden-Bell and Sandage.

Of course, it is possible that morethan one model for the forma-tion of the galaxy is correct.

The Eggen–Lynden-Bell–Sandage sce-nario may apply to the dense bulge andinner halo. The more rarefied outerparts of the galaxy may have developedby the merger of fragments, along thelines theorized by Toomre or by Searleand Zinn. If so, then the clusters in the in-ner halo would have formed beforethose in the more tenuous outer regions.The process would account for some of the age differences found for the glob-ular clusters. More precise modeling may have to await the improved imagequality that modifications to the Hub-ble Space Telescope cameras will afford.

Knowing the age of the halo is, how-ever, insufficient to ascertain a detailedformation scenario. Investigators needto know the age of the disk as well andthen to compare that age with the halo’sage. Whereas globular clusters are usefulin determining the age of the halo, an-other type of celestial body—very faintwhite dwarf stars—can be used to deter-mine the age of the disk. The absence ofwhite dwarfs in the galactic disk near thesun sets a lower limit on the disk’s age.White dwarfs, which are no longer pro-ducing radiant energy, take a long timeto cool, so their absence means that thepopulation in the disk is fairlyyoung—less than about 10 billion years.

This value is significantly less than theages of clusters in the halo and is thusconsistent with the notion that the bulkof the galactic disk developed after thehalo.

It is, however, not yet clear if there isa real gap between the time when for-mation of the galactic halo ended andwhen creation of the old thick disk be-gan. To estimate the duration of such atransitional period between halo anddisk, investigators have compared theages of the oldest stars in the disk withthose of the youngest ones in the halo.The oldest known star clusters in thegalactic disk, NGC 188 and NGC6791, have ages of nearly eight billionyears, according to Pierre Demarqueand David B. Guenther of Yale and Eliz-abeth M. Green of the University ofArizona. Stetson and his colleagues andRoberto Buonanno of the AstronomicalObservatory in Rome and his co-work-ers examined globular clusters in thehalo population. They found the young-est globulars—Palomar 12 andRuprecht 106—to be about 11 billionyears old. If the few billion years’ differ-ence between the disk objects and theyoung globulars is real, then youngglobulars may be the missing links be-tween the disk and halo populations ofthe galaxy.

At present, unfortunately, the relativeages of only a few globular clusters havebeen precisely estimated. As long as thisis the case, one can argue that the MilkyWay could have tidally captured Palo-mar 12 and Ruprecht 106 from theMagellanic Clouds. This scenario, pro-posed by Douglas N. C. Lin of the Uni-versity of California at Santa Cruz andHarvey B. Richer of the University ofBritish Columbia, would obviate theneed for a long collapse time. Further-more, the apparent age gap betweendisk and halo might be illusory. Unde-tected systematic errors may lurk in theage-dating processes. Moreover, gravita-tional interactions with massive inter-stellar clouds may have disrupted theoldest disk clusters, leaving behind onlyyounger ones.

Determining the relative ages of thehalo and disk reveals much about thesequence of the formation of the galaxy.On the other hand, it leaves open thequestion of how old the entire galaxyactually is. The answer would providesome absolute framework by which thesequence of formation events can bediscerned. Most astronomers who studystar clusters favor an age of some 15 to17 billion years for the oldest clusters(and hence the galaxy).

Confidence that those absolute age

COLOR-LUMINOSITY DIAGRAMS can be used to determine stellar ages. The one above com-pares the plots of stars in globular clusters NGC 288 and NGC 362 with age tracks (black lines)generated by stellar evolution models. The color index, expressed in magnitude units, is a measureof the intensity of blue wavelengths minus visual ones. In general, the brighter the star, the lowerthe color index; the trend reverses for stars brighter than about visual magnitude 19. The plotssuggest the clusters differ in age by about three billion years. The temperature (inversely related tothe color index) and luminosity have been set to equal those of NGC 288.

COLOR INDEX (MAGNITUDE UNITS)0.6

VIS

UA

L M

AG

NIT

UD

E

0.4 0.8 1

16

18

20

22

21

19

17

23

LUM

INO

SIT

Y (

SO

LAR

UN

ITS

)

0.10

1.0

10

TEMPERATURE (DEGREES CELSIUS)4,9005,150 4,650 4,400

12-BILLION-YEAR-�OLD AGE TRACK

15-BILLION-YEAR-�OLD AGE TRACK

NGC 288�

NGC 362

MM

ICH

AE

L J.

BO

LTE

, LIC

K O

BS

ER

VA

TO

RY

; JO

HN

NY

JO

HN

SO

N



values are realistic comes from the mea-sured abundance of radioactive isotopesin meteorites. The ratios of thorium 232to uranium 235, of uranium 235 to ura-nium 238 or of uranium 238 to plutoni-um 244 act as chronometers. Accordingto these isotopes, the galaxy is between10 and 20 billion years old. Althoughages determined by such isotope ratiosare believed to be less accurate thanthose achieved by comparing stellar ob-servations and models, the consistency ofthe numbers is encouraging.

Looking at the shapes of othergalaxies alleviates to some extentthe uncertainty of interpreting the

galaxy’s evolution. Specifically, the studyof other galaxies presents a perspectivethat is unavailable to us as residents of the Milky Way—an external view.We can also compare information fromother galaxies to see if the processes thatcreated the Milky Way are unique.

The most immediate observation onecan make is that galaxies come in sever-al shapes. In 1925 Edwin P. Hubblefound that luminous galaxies could bearranged in a linear sequence accordingto whether they are elliptical, spiral orirregular [see top illustration on nextpage]. From an evolutionary point ofview, elliptical galaxies are the most ad-vanced. They have used up all (or al-most all) of their gas to generate stars,which probably range in age from 10 to 15 billion years. Unlike spiral galax-ies, ellipticals lack disk structures. Themain differences between spiral and ir-regular galaxies is that irregulars haveneither spiral arms nor compact nuclei.

The morphological types of galax-ies can be understood in terms of the speed with which gas was used to createstars. Determining the rate of gas deple-tion would corroborate estimates of theMilky Way’s age and history. Star for-mation in elliptical galaxies appears tohave started off rapidly and ecientlysome 15 billion years ago and then de-clined sharply. In most irregular gal-axies the birth of stars has taken placemuch more slowly and at a more nearlyconstant rate. Thus, a significant fractionof their primordial gas still remains.

The rate of star formation in spiralsseems to represent a compromise be-tween that in ellipticals and that in ir-regulars. Star formation in spirals beganless rapidly than it did in ellipticals butcontinues to the present day.

Spirals are further subdivided intocategories Sa, Sb and Sc. The subdivi-sions refer to the relative size of the nu-clear bulges and the degree to which thespiral arms coil. Objects of type Sa have

the largest nuclear bulges and the mosttightly coiled arms. Such spirals alsocontain some neutral hydrogen gas and asprinkling of young blue stars. Sb spiralshave relatively large populations ofyoung blue stars in their spiral arms. Thecentral bulge, containing old red stars, isless prominent than is the central bulgein spirals of type Sa. Finally, in Sc spiralsthe light comes mainly from the youngblue stars in the spiral arms; the bulgepopulation is inconspicuous or absent.The Milky Way is probably intermediatebetween types Sb and Sc.

Information from other spirals seemsconsistent with the data obtained forthe Milky Way. Like those in ourgalaxy, the stars in the central bulges ofother spirals arose early. The dense in-ner regions of gas must have collapsedfirst. As a result, most of the primordialgas initially present near the centers hasturned into stars or has been ejected bysupernova-driven winds.

There is an additional kind ofevidence on which to build ourunderstanding of how the Milky

Way came into existence: the chemicalcomposition of stars. This informationhelps to pinpoint the relative ages ofstellar populations. According to stellarmodels, the chemistry of a star dependson when it formed. The chemicaldifferences exist because first-generationstars began to “pollute” the proto-galaxy with elements heavier than heli-um. Such so-called heavy elements, or“metals,” as astronomers refer to them,were created in the interiors of stars orduring supernova explosions. Examin-ing the makeup of stars can provide stel-lar evolutionary histories that corrobo-

rate or challenge age estimates.Different types of stars and super-

novae produce different relative abun-dances of these metals. Researchers be-lieve that most “iron-peak” elements(those closest to iron in the periodictable) in the galaxy were made in super-novae of type Ia. The progenitors ofsuch supernovae are thought to be pairsof stars, each of which has a mass a fewtimes that of the sun. Other heavy ele-ments—the bulk of oxygen, neon, mag-nesium, silicon and calcium, amongothers—originated in supernovae thatevolved from single or binaries of mas-sive, short-lived stars. Such stars haveinitial masses of 10 to 100 solar massesand violently end their lives as super-novae of type Ib, Ic or II.

Stars that subsequently formed incor-porated some of these heavy elements.For instance, approximately 1 to 2 per-cent of the mass of the sun consists ofelements other than hydrogen or heli-um. Stars in nuclear bulges generallyharbor proportionally more heavy ele-ments than do stars in the outer disksand halos. The abundance of heavy ele-ments decreases gradually by a factor of0.8 for every kiloparsec (3,300 light-years) from the center to the edge of theMilky Way disk. Some 70 percent of the150 or so known globular clusters in theMilky Way exhibit an average metalcontent of about one twentieth that ofthe sun. The remainder shows a meanof about one third that of the sun.

Detailed studies of stellar abundancesreveal that the ratio of oxygen to iron-peak elements is larger in halo starsthan it is in metal-rich disk stars [see il-lustration above]. This difference sug-gests the production of heavy elements

OXYGEN-TO-IRON RATIOS as a function of metallicity (abundance of iron) for halo and olddisk stars indicate different formation histories. The high ratios in metal-deficient halo stars sug-gest that those stars incorporated the oxygen synthesized in supernovae of types Ib, Ic and II. TypeIa supernovae seem to have contributed material only to the disk stars. Beatriz Barbuy and MarciaErdelyi-Mendes of the University of São Paulo made the measurements.

IRON ABUNDANCE (SOLAR UNITS)

OX

YG

EN

-TO

-IR

ON

RA

TIO

(S

OLA

R U

NIT

S)

0.010.001 0.1 1.0

4

3

2

1

HALO STARS�OLD DISK STARS

JOH

NN

Y J

OH

NS

ON



during the halo phase of galactic evolu-tion was dominated by supernovae oftypes Ib, Ic and II. It is puzzling thatiron-producing type Ia supernovae,some of which are believed to have resulted from progenitor starswith lifetimes as short as a few hundredmillion years, did not contribute moreto the chemical mixture from whichhalo stars and some globular clustersformed. This failure would seem to im-

ply that the halo collapsed very rapid-ly—before supernovae of type Ia couldcontribute their iron to the halo gas.

That idea, however, conflicts with thefour-billion-year age spread observedamong galactic globular clusters, whichimplies that the halo collapsed slow-ly. Perhaps supernova-driven galacticwinds swept the iron-rich ejecta fromtype Ia supernovae into intergalacticspace. Such preferential removal of theejecta of type Ia supernovae might haveoccurred if supernovae of types Ib, Icand II exploded primarily in dense gas clouds. Most of type Ia supernovaethen must have detonated in less denseregions, which are more easily sweptout by the galactic wind.

Despite the quantity of data, informa-tion about metal content has proved in-sucient to settle the controversy con-cerning the time scale of disk and haloformation. Sandage and his colleagueGary A. Fouts of Santa Monica Collegefind evidence for a rather monolithic col-lapse. On the other hand, John E. Nor-ris and his collaborators at the Aus-tralian National Observatory, amongothers, argue for a significant decou-pling between the formation of haloand disk. They also posit a more chaot-ic creation of the galaxy, similar to thatenvisaged by Searle and Zinn.

Such differences in interpretation of-ten reflect nearly unavoidable effectsarising from the way in which particu-lar samples of stars are selected forstudy. For example, some stars exhibitchemical compositions similar to thoseof “genuine” halo stars, yet they havekinematics that would associate themwith one of the subcomponents of thedisk. As vital as it is, chemical informa-tion alone does not resolve ambigu-ities about the formation of the galactichalo and disk. “Cats and dogs mayhave the same age and metallicity, butthey are still cats and dogs” is the wayBernard Pagel of the Nordic Institutefor Theoretical Physics in Copenhagenputs it.

As well as telling us about the pasthistory of our galaxy, the disk and haloalso provide insight into the Milky Way’s

probable future evolution. One can easi-ly calculate that almost all of the ex-isting gas will be consumed in a few bil-lion years. This estimate is based on therate of star formation in the disks ofother spirals and on the assumption thatthe birth of stars will continue at its pre-sent speed. Once the gas has been de-pleted, no more stars will form, and thedisks of spirals will then fade. Eventuallythe galaxy will consist of nothing morethan white dwarfs and black holes en-capsulated by the hypothesized darkmatter corona.

Several sources of evidence exist forsuch an evolutionary scenario. In1978 Harvey R. Butcher of the

Kapteyn Laboratory in the Netherlandsand Augustus Oemler, Jr., of Yale foundthat dense clusters of galaxies locatedabout six billion light-years away stillcontained numerous spiral galaxies.Such spirals are, however, rare or absentin nearby clusters of galaxies. This ob-servation shows that the disks of mostspirals in dense clusters must have fadedto invisibility during the past six billionyears. Even more direct evidence for theswift evolution of galaxies comes fromthe observation of so-called blue galax-ies. These galaxies are rapidly generat-ing large stars. Such blue galaxies seemto be less common now than they wereonly a few billion years ago.

Of course, the life of spiral galaxiescan be extended. Copious infall of hy-drogen from intergalactic space mightreplenish the gas supply. Such infall canoccur if a large gas cloud or anothergalaxy with a substantial gas reservoir isnearby. Indeed, the Magellanic Cloudswill eventually plummet into the MilkyWay, briefly rejuvenating our galaxy.Yet the Milky Way will not escape itsultimate fate. Like people and civiliza-tions, stars and galaxies leave behindonly artifacts in an evolving, ever dy-namic universe.

MORPHOLOGICAL CLASSIFICATION ofgalaxies (top) ranges from ellipticals (E) to spi-rals (subdivided into categories Sa, Sb and Sc)and irregulars ( Ir ). The history of star forma-tion varies according to morphology (bottom).In elliptical galaxies, stars developed in an ini-tial burst. Star formation in spirals was less vig-orous but continues today. In most irreg-ulargalaxies the birthrate of stars has probably re-mained constant.

FURTHER READINGGALACTIC ASTRONOMY: STRUCTUREAND KINEMATICS. Dimitri Mihalas andJames Binney. W. H. Freeman and Company,1981.

THE MILKY WAY AS A GALAXY. GerardGilmore, Ivan R. King and Pieter C. van derKruit. University Science Books, Mill Valley,Calif., 1990.

THE FORMATION AND EVOLUTION OFSTAR CLUSTERS. Edited by Kenneth Janes.Astronomical Society of the Pacific, 1991.

THE STELLAR POPULATIONS OFGALAXIES. Edited by B. Barbuy and A.Renzini. Kluwer Academic Publishers, Dor-drecht, Holland, 1992.

SID

NE

Y V

AN

DE

N B

ER

GH

(T

OP

); J

OH

NN

Y J

OH

NS

ON

(C

HA

RT,

BO

TT

OM

)

6 SCIENTIFIC AMERICAN EXCLUSIVE ONLINE ISSUE JULY 2004 COPYRIGHT 2004 SCIENTIFIC AMERICAN, INC.

GalaxyGrowing,

Breathing

Our

Long assumed to be a relic of the distant past, the

MILKY WAYturns out to bea dynamic,living object

By Bart P. Wakker and Philipp Richter

originally published in January 2004


GULPING DOWN GAS and cannibalizing its smaller neighbors, the Milky Waygalaxy is still in the process of forming. For a key to this image, see page 41.


We may know our hometowns intimately, yet visitors or youngchildren may still point out things we have never noticed be-fore. They may not be as attuned to all the minutiae, but theyoften see the big picture better than longtime residents can. Asimilar situation faces astronomers who study the Milky Way:we are so deeply embedded in our home galaxy that we can-not see it fully. When we look at other galaxies, we can discerntheir overall layout but not their detailed workings. When welook at our own, we can readily study the details but perceivethe overall structure only indirectly.

Consequently, we have been slow to grasp the big pictureof the Milky Way’s structure and history. Astronomers werenot even sure that the galaxy was a distinct object, only one ofmany billions, until the 1920s. By the mid-1950s they hadpainstakingly assembled the picture that most people now haveof the Milky Way: a majestic pinwheel of stars and gas. In the1960s theorists proposed that our galaxy formed early in cos-mic history—by the most recent estimate, 13 billion years ago—

and has remained broadly unchanged ever since.Gradually, though, it has become clear that the Milky Way

is not a finished work but rather a body that is still forming.Like the earlier discoveries, this realization has relied heavily onobserving other galaxies and bringing the lessons back home.Most galaxies are now assumed to result from the merging ofsmaller precursors, and in the case of the Milky Way, we canobserve the final stages of this process. Our galaxy is tearingapart small satellite galaxies and incorporating their stars.Meanwhile gas clouds are continually arriving from inter-galactic space. No longer can researchers speak of galaxy for-mation in the past tense.

The evidence for the continuing accretion of gas by theMilky Way involves high-velocity clouds, or HVCs—mysteri-ous clumps of hydrogen, up to 10 million times the mass of thesun and 10,000 light-years across, moving rapidly through theouter regions of the galaxy. HVCs were discovered 41 yearsago, but only in the past five years have new data and new ideasprovided the evidence that some of them represent infalling gas.HVCs also show that the galaxy is breathing—pushing out gasand then pulling it back in, as if exhaling and inhaling. In ad-dition, the properties of HVCs suggest that a gigantic sphere ofhot, tenuous plasma surrounds the galaxy. Astronomers hadlong suspected the existence of such a sphere, but few thoughtit would be so large.

Historically, interpreting HVCs has been difficult because be-ing stuck within the galaxy, we have no direct way to know theirlocations. We can see their two-dimensional positions on the skybut lack depth perception. Over the past four decades, this ambi-guity has led to many alternative hypotheses, some placing HVCsclose to our own stellar neighborhood, others locating themdeep in intergalactic space. The recent breakthroughs have oc-curred mainly because ground-based and orbiting telescopes havefinally managed to get a three-dimensional fix on the clouds—

and thereby a better perspective on our celestial hometown.

Virgin or Recycled?OUR GALAXY CONTAINS about 100 billion stars, most ofwhich are concentrated in a thin disk about 100,000 light-yearsacross and 3,000 light-years thick. These stars revolve aroundthe galactic center in nearly circular orbits. The sun, for exam-ple, trundles around at nearly 200 kilometers per second. An-other 10 billion stars form the galactic “halo,” a huge spheri- R

ON

MIL

LER

(p

rece

din

g p

ag

es)

■ Since the early 1960s astronomers have thought that theMilky Way and other galaxies were born early in cosmichistory and then evolved slowly. Today, however,evidence indicates that galaxies are continuing to grow.They cannibalize their smaller brethren and gulp downfresh gas from intergalactic space.

■ In our Milky Way we have a close-up view of the ongoingconstruction work. The incoming gas takes the form ofhigh-velocity clouds discovered decades ago. Onlyrecently were some of these clouds proved to be freshmaterial; observationally, they get entangled withcirculating gas.

■ These clouds come in several guises: clumps of neutralhydrogen reminiscent of intergalactic gas; a stream ofgas torn out of nearby small galaxies; and highly ionizedhot gas that may be dispersed throughout theintergalactic vicinity.

Overview/High-Velocity Clouds

Sometimes the hardest things to understand are the things you are most familiar with.


cal envelope that surrounds the disk. Between the stars lie gasand dust, forming the interstellar medium, most of which alsomoves in nearly circular orbits around the galactic center andis even more narrowly concentrated in a disk than the stars are.Like a planet’s atmosphere, the gas in the medium is densest atits “bottom” (the galactic plane) and thins out with height. Butup to about 10 percent of the interstellar medium lies outsidethe plane and moves up to 400 kilometers per second fasterthan rotation would imply. This gas constitutes the HVCs.

The story of HVCs began in the mid-1950s, when GuidoMünch of the California Institute of Technology discovereddense pockets of gas outside the plane—a clear exception to therule that the density of gas diminishes with height. Left to them-selves, those dense pockets should quickly dissipate, so in 1956Lyman Spitzer, Jr., of Princeton University proposed that theywere stabilized by a hot, gaseous corona that surrounded theMilky Way, a galactic-scale version of the corona around the

sun [see “The Coronas of Galaxies,” by Klaas S. de Boer andBlair D. Savage; Scientific American, August 1982].

Inspired by Spitzer’s proposal, Jan Oort of Leiden Univer-sity in the Netherlands conjectured that the galactic halo mightalso contain cold gas very far from the galactic plane. A searchfor radio emission from cold clouds resulted in their discoveryin 1963. Unlike the gas found by Münch, these clouds did notfollow the overall rotation of the galaxy; instead they seemedto be falling toward the galactic disk at high speed, so they be-came known as HVCs. A slower-moving but still anomaloustype of cloud, an intermediate-velocity cloud, or IVC, was spot-ted the same year.

Oort later fleshed out his idea and suggested that after theinitial formation of the galaxy, gas near the edge of its gravita-tional sphere of influence was left over. This gas reached thedisk only after 10 billion years or more, becoming observableas HVCs. Oort’s idea fit in well with models that try to explainSL

IM F

ILM

S

OUR GALAXY AND ITS NEIGHBORHOOD

SAGITTARIUS STREAMTrail of stars torn off Sagittarius dwarf galaxy

TRIANGULUMGALAXY

Nearby midsizespiral galaxy

GALACTIC CORONAHot gas surrounding the Milky Way

GALACTIC DISKFlattenedsystem of stars,gas and dust

ANDROMEDAGALAXY

Nearest majorspiral galaxy

SAGITTARIUS DWARFSPHEROIDAL GALAXY

Satellite galaxy of the Milky Way

SUN ANDPLANETS

HIGH-VELOCITYCLOUD (HVC)Infalling clump ofrelatively freshhydrogen gas

SUPERBUBBLEGas pushed outby supernovae,outward leg of“fountain”

INTERMEDIATE-VELOCITY CLOUD(IVC) Clump ofrecondensed gas,return leg of“fountain”

LARGE MAGELLANICCLOUD Satellite galaxyof the Milky Way

MAGELLANIC STREAMGaseous filament torn off

Magellanic Clouds

SMALL MAGELLANIC CLOUDSatellite galaxy of the Milky Way


the observed chemical composition of the galaxy. Stars produceheavy elements and scatter them into interstellar space whenthey die. Newly born stars incorporate those elements and pro-duce even more. Therefore, if the galaxy were evolving in iso-lation, each generation of stars should contain more heavy el-ements than its predecessors.

Yet most stars in the solar neighborhood, regardless of age,have about the same abundance of heavy elements. The favoredexplanation for this apparent discrepancy is that the galaxy isnot isolated and that interstellar gas is constantly being dilut-ed by more pristine material. Several researchers surmised thatsome or all of the HVCs represent this fresh gas, but the propo-sition lacked direct observational evidence.

An alternative hypothesis holds that HVCs have nothing todo with an influx of gas but are instead part of a “galactic foun-tain.” This idea was proposed in the mid-1970s by PaulShapiro, now at the University of Texas at Austin, and George

B. Field of the Harvard-Smithsonian Center for Astrophysics.Gas heated and ionized by massive stars rises out of the diskinto the corona, forming an atmosphere. Some regions thencool off, rain back down and become electrically neutral again,setting up a cycle of gas between the disk and the corona. In1980 Joel Bregman, now at the University of Michigan at AnnArbor, suggested that HVCs could be the returning gas, and fora while this idea was the leading explanation for their origin.

Going Out with the TideNEITHER OORT’S HYPOTHESIS nor the fountain model,however, could explain all characteristics of all HVCs. Theproblem was further complicated by the discovery in the early1970s of the Magellanic Stream, a filament of gas that arcsaround the galaxy. The stream follows the orbits of the Largeand Small Magellanic Clouds, two small companion galaxiesthat revolve around the Milky Way like moons around a plan-

et. Although astronomers usually reserve the term “cloud” fora clump of gas or dust, these full-fledged galaxies containingbillions of stars are so named because they resemble clouds inthe night sky. They are currently about 150,000 light-yearsfrom our galaxy, about as close as they ever get on their high-ly elongated paths.

The stream behaves in many ways like a string of HVCs.Much of it moves at velocities that are incompatible with nor-mal galactic rotation. Yet it cannot be explained by the two hy-potheses described above. According to the most detailed mod-el of the stream, published in 1996 by Lance T. Gardiner of SunMoon University in South Korea and Masafumi Noguchi ofTohoku University in Japan, the filament is our galaxy’s ver-sion of the tidal streams that astronomers see around many oth-er galaxies. When the Magellanic Clouds made their previousclose approach to the Milky Way, 2.2 billion years ago, thecombined force of our galaxy and the Large Magellanic Cloud

ripped off some of the gas in the outer parts of the Small Mag-ellanic Cloud. About half the gas was decelerated and laggedbehind the Magellanic Clouds in their orbits. The other halfwas accelerated and pulled ahead of the galaxies, forming whatis called a leading arm. A similar process may also be rippingapart some of the Milky Way’s other satellite galaxies [see boxon page 14].

An alternative model ascribes the stream to frictional forces.If the Milky Way has a very extended corona (much bigger thanthe one proposed by Spitzer), this corona could strip off gasfrom the Magellanic Clouds. In either model, however, theMagellanic Clouds have lost large amounts of gas, producingmany of the HVCs.

Yet another twist in the saga of HVCs came in 1999, whenLeo Blitz of the University of California at Berkeley and his col-laborators suggested that they are much farther away than mostof their colleagues thought possible. Instead of buzzing throughthe outskirts of the Milky Way, HVCs could be floating aroundin the Local Group of galaxies—a conglomeration of the MilkyWay, Andromeda and some 40 smaller galaxies that occupiesa volume of space roughly four million light-years across. Inthis case, HVCs would be remnants of the group’s, rather thanonly our galaxy’s, formation.

Similar ideas had been put forward more than 30 years agoand excluded because gas clouds should not be stable at theproposed distances. Blitz conjectured that HVCs are not, infact, clouds of gas but clumps of dark matter with a smallamount of gas mixed in. If so, HVCs are 10 times as massiveas astronomers had assumed and therefore able to hold them-selves together. An attractive feature of this hypothesis is thatit alleviates what has become a major embarrassment for as-

BART P. WAKKER and PHILIPP RICHTER are observers, primarily in theultraviolet and radio bands of the electromagnetic spectrum. Theyjoined forces to investigate high-velocity clouds in late 1999, whenRichter took up a postdoctoral position at the University of Wiscon-sin–Madison, where Wakker was doing research. Wakker traces hisinterest in astronomy to the Apollo 8 moonflight. He did his doctor-al thesis on HVCs at the University of Groningen in the Netherlands,then spent five years at the University of Illinois before moving toWisconsin in 1995. Richter received his Ph.D. from the University ofBonn in Germany, where he studied diffuse molecular gas in theMagellanic Clouds and the halo of the Milky Way. After leaving Wis-consin in 2002, he worked at the Arcetri Astrophysical Observato-ry in Florence, Italy, and recently returned to Bonn.

THE

AU

THO

RS

Our galaxy is tearing apart its satellite galaxies, andgas clouds are arriving from intergalactic space.


tronomers—namely, that models of galaxy formation predictmore leftover dark matter halos than have been found [see“The Life Cycle of Galaxies,” by Guinevere Kauffmann andFrank van den Bosch; Scientific American, June 2002].HVCs could be the missing leftovers.

Getting WarmerTHUS, ASTRONOMERS ENTERED the third millenniumwith four hypotheses for HVCs: fresh gas left over from galaxyformation, gas cycling through a galactic fountain, shreds ofthe Magellanic Clouds, or intergalactic amalgams of gas anddark matter. Each hypothesis had bits and pieces of supportingevidence, but researchers needed new data to break the dead-lock, and since the mid-1990s they have made major progress.

First, they have completed an all-sky survey for radio emis-sion from neutral hydrogen, which traces gas at temperaturesof about 100 kelvins. Aad Hulsbosch of the University of Nij-megen and one of us (Wakker), using the Dwingeloo radiotelescope in the Netherlands, finished the northern half of thissurvey in 1988. Ricardo Morras and his collaborators, usingthe Villa Elisa radio telescope in Argentina, covered the south-

ern sky in 2000 [see illustration above]. A third survey, by DapHartmann and Butler Burton of Leiden Observatory, becameavailable in 1997 and mapped all of the Milky Way’s neutralhydrogen, including both HVCs and IVCs.

A further contribution came from observations in visiblelight, made by instruments such as the Wisconsin Hydrogen-Alpha Mapper [see “The Gas between the Stars,” by RonaldJ. Reynolds; Scientific American, January 2002]. Althoughneutral hydrogen does not shine at visible wavelengths, ionizedgas does, and the outer parts of HVCs are ionized by far-ultra-violet light from the Milky Way and other objects. The radia-tion also heats the clouds’ exteriors to 8,000 kelvins. Theamount of visible light is a measure of the intensity of the radi-ation field surrounding the HVC, which in turn depends on itsdistance from the galactic disk. Thus, these observations offera rough way to estimate the location of HVCs.

The most important progress has come from observationsof spectral absorption lines in HVCs. Instead of looking forlight given off by the gas, this work analyzes light blocked bythe gas—specific atoms filter out specific wavelengths of light.Three observatories have made the largest contributions: theIN

GR

ID K

ALLI

CK

; SO

UR

CE

: B

ART

WAK

KE

R (

rad

io),

LU

ND

OB

SER

VATO

RY

(vis

ible

)

CLOUDY SKY MAP OF GALACTIC GAS combines radio observations of neutralhydrogen (colored splotches) with a visible-light image of theMilky Way (white). The map depicts our sky, reprojected so that

the galactic disk runs across the middle; the core of the galaxylies at the center. High-velocity clouds of hydrogen, such ascomplexes A and C, are located above and below the disk.

COMPLEX C

COMPLEX A

MAGELLANIC STREAM(LEADING ARM)

MAGELLANIC STREAM(TIDAL TAIL)

SMALLMAGELLANICCLOUD

MILKY WAY

POSSIBLEINTERGALACTIC

HVCs

FOUNTAIN GAS

LARGEMAGELLANICCLOUD


ING

RID

KAL

LIC

K

GALACTIC FOUNTAIN: Intermediate-velocity clouds are probablythe return leg of a vast cycle of gas. Clusters of supernovaexplosions generate bubbles of hot gas (blue) that breakthrough the surrounding cold gas (yellow) and feed a hotcorona. Chunks of the gas cool and fall back to the disk.

GAS INFALL: Many of the high-velocity clouds (yellow) are gasraining onto the Milky Way, continuing its formation nearly 12billion years after it started. Such gas could provide fresh fuelfor star formation. Observationally, they are easily confusedwith the intermediate-velocity clouds (orange).

GALACTIC CANNIBALIZATION: The Milky Way is ripping gas fromtwo of its satellite galaxies, the Large and Small MagellanicClouds. Along their orbits astronomers see the MagellanicStream (orange). Other, unrelated high-velocity clouds (yellow),possibly condensing out of a hot corona, float in the same space.

INTERGALACTIC REPLENISHMENT: The Milky Way and Androme-da galaxies may be embedded in a massive sea of hot inter-galactic gas (blue). Out of this gas, cold clumps may condenseand get captured by the galaxies—forming new high-velocityclouds that eventually fall in. This model is still uncertain.

FOUR PROCESSES THAT SHAPE THE GALAXY

MILKY WAY

HOT INTERGALACTIC GAS

ANDROMEDAGALAXY

TRIANGULUMGALAXY

SMALLMAGELLANIC CLOUD

COMPLEX C

ORBIT OF MAGELLANIC CLOUDS

MAGELLANICSTREAM

LARGEMAGELLANIC CLOUD

HOT GASCOLD GAS

SUPERBUBBLE

SUPERNOVA

CHIMNEY

IVC

IVC

HVC

SUN


La Palma Observatory in the Canary Islands, the Hubble SpaceTelescope and the Far Ultraviolet Spectroscopic Explorer(FUSE), launched in 1999.

Using such data, Laura Danly, now at the University ofDenver, and her collaborators put limits on the distance to anIVC 11 years ago. More recently, Hugo van Woerden of theUniversity of Groningen in the Netherlands and his collabora-tors gauged the distance to an HVC for the first time [see boxon next page]. Meanwhile we and our colleagues measured thechemical composition of the clouds, rounding out the infor-mation needed to distinguish among the various hypotheses.

A very warm component of HVCs emerged in data fromFUSE. This satellite detected absorption by highly ionized ox-ygen (specifically, oxygen atoms that have lost five of their eightelectrons), which implies a temperature of about 300,000kelvins. Such temperatures can occur where cool (100 kelvins)neutral hydrogen comes into contact with extremely hot (onemillion kelvins) gas. Alternatively, the presence of gas at300,000 kelvins shows that the extremely hot gas is coolingdown. Together with Blair D. Savage of the University of Wis-consin–Madison and Kenneth Sembach of the Space TelescopeScience Institute in Baltimore, we have traced this componentof HVCs.

Complex BehaviorHAVING EXPLORED ALL these new data, we can now pre-sent a coherent picture of HVCs. We begin with two of thelargest, known as complexes A and C, which were the firstHVCs discovered back in 1963. Complex A is 25,000 to30,000 light-years away, which clearly puts it in the galactichalo. The distance to complex C remains uncertain: at least14,000 light-years but probably no more than 45,000 light-years above the galactic plane.

The two clouds are deficient in heavy elements, havingabout a tenth of the concentration found in the sun. The nitro-gen content of complex C is especially low, about 1⁄50 of thesun’s. The paucity of nitrogen suggests that the heavy elementscame mostly from high-mass stars, which produce less nitrogenrelative to other heavy elements than low-mass stars do. In fact,recent models of the young universe predict that the earlieststars are uncommonly heavy. Complex C thus appears to be afossil from the ancient universe.

Brad Gibson of Swinburne University in Melbourne, Aus-tralia, has looked at a different part of complex C and measureda heavy-element concentration that was twice as high as our ear-lier results. This variation in composition indicates that complexC has begun to mix with other gas clouds in the galactic halo,which have higher concentrations of heavy elements. In addition,Andrew Fox and his collaborators at Wisconsin used the datafor highly ionized oxygen and other ions to show that the gasat 300,000 kelvins in complex C represents an interface be-tween hot and cool gas. We seem to be catching complex C inthe process of assimilating into the galaxy.

Clouds such as complexes A and C thus provide the first di-rect evidence for the infall of fresh gas. Complex C brings be-

tween 0.1 and 0.2 solar mass of new material every year, andcomplex A represents about half of that. This is 10 to 20 per-cent of the total needed to dilute galactic gas and account forthe chemical composition of stars. Other HVCs may make upthe remainder. It is somewhat unclear, though, whether the ul-timate source of this gas is a remnant halo (as proposed byOort), deep intergalactic space, or even a small dwarf galaxythat the Milky Way swallowed.

A Multiplicity of OriginsTHE RESULTS ELIMINATE three of the hypotheses for theorigin of complexes A and C. The fountain hypothesis impliesthat they originate in the disk and have a composition similarto that of the sun, which is not the case. The Magellanic Streamhypothesis also gets the heavy-element content wrong. Final-ly, the dark matter hypothesis fails because these two HVCs donot lie in intergalactic space. It turns out, however, that thesethree explanations are not completely incorrect. We simplyhave to look elsewhere to find where they apply.

For a long time, IVCs stood in the shadow of the moreRO

N M

ILLE

R

Conscious of StreamsMOST OF THE MILKY WAY is as thoroughly mixed as a well-stirred gravy. Two stars that originated in the same regionmay be located in completely different parts of the sky today.But during the past few years, astronomers have foundgroups of stars that move in unison, forming what they callstellar streams. They are like lumps that a cook has justthrown into a pot but that have not had time to mix in.

The streams are believed to be the remnants of satellitegalaxies of the Milky Way thatwere torn apart by tides, the sameprocess that formed some of thehigh-velocity clouds. The streamsthus trace a flow of stars fromdwarf galaxies to the Milky Way.They differ from the MagellanicStream, which consists of gasrather than stars. They representindependent evidence for theongoing growth of our galaxy.

One spectacular example is a stream of stars being pulledoff the Sagittarius dwarf spheroidal galaxy, which wasdiscovered in 1994 by Rodrigo Ibata of the StrasbourgObservatory in France and his colleagues [see artist’sconception above]. More recently, several other stellarstreams were found in the data gathered by the Sloan DigitalSky Survey, a program to map a large portion of the skysystematically. One may be related to the Canis Major dwarfgalaxy, which Ibata, Nicolas Martin of Strasbourg and theircollaborators discovered two months ago. Over the past twobillion years, this galaxy has been stretched into a spiralingring of stars along the galactic plane. —B.W. and P.R.


flashy and mysterious HVCs. Several teams have now measuredtheir composition, and it matches that of gas in the disk. More-over, IVCs lie some 4,000 light-years above the plane, the placewhere fountains would operate. Both facts indicate that they,rather than HVCs, represent the return flow of a fountain.

A piece of corroborating evidence has been the detectionof hydrogen molecules in IVCs. Forming these molecules in

space requires interstellar dust grains, which will be sufficient-ly abundant only if the ambient gas is chemically enriched. Inline with this idea, molecular hydrogen was not found in com-plex C. Thus, IVCs are recycled gas from within the galaxy,whereas HVCs are primarily gas from outside.

As for the Magellanic Stream hypothesis, at least one HVCdoes seem to be a castoff from the stream. Its composition is IN

GR

ID K

ALLI

CK

PEEKING BEHIND THE CLOUDSHIGH-VELOCITY CLOUDS stymied astronomers for decades because their distances andcompositions were uncertain. The only knowntechnique to measure these properties is theabsorption-line method. Stars and galaxies locatedbehind HVCs act as bulbs that shine through theclouds from behind. Most of the light passes throughthe clouds, but a few wavelengths are absorbed,allowing properties of the clouds to be measured.

If the spectrum of a star contains absorptionlines, it means a cloud must be sitting between usand the star. The distance to the star sets an upperlimit on the distance to the cloud. Conversely, thelack of an absorption line implies a lower limit onthe distance to the cloud. These limits assume thatother factors can be ruled out: uncertainties in thestellar distance, lack of enough heavy elements toproduce a detectable absorption line, and absorptionlines created by material within the star itself.

To determine HVC distances, the most usefullightbulbs are so-called RR Lyrae variables and bluehorizontal branch (BHB) stars. They are numerous,their distances can be measured accurately, andfew of their spectral lines overlap with those of theclouds. In principle, the absorption lines of anyelement could be used. To determine the heavyelement content, however, the best measurementsrely on the spectral lines of neutral oxygen andionized sulfur. These lines lie in the ultraviolet partof the spectrum, requiring properly equippedsatellites such as the Hubble Space Telescope orFar Ultraviolet Spectroscopic Explorer (FUSE). Inthis case, the best lightbulbs are distant activegalaxies such as quasars, because they often havefeatureless spectra and are brighter ultravioletemitters than stars.

A single star or galaxy can illuminate morethan one gas cloud. Each cloud moves at adifferent velocity, so each absorbs at a slightlydifferent wavelength because of the Dopplereffect. To distinguish the clouds requires aspectrometer with high spectral resolution, whichin turn requires a large telescope. —B.W. and P.R.

STAR B

STAR A

HVC

CROSS-SECTIONAL VIEW (not to scale)

NEARBY GAS

100 km/s 10 km/s

WHEN ASTRONOMERS take aspectrum of Star A, they notice two absorption lines and infer the presence of two clouds. Thelines occur at differentwavelengths because of thedifferent cloud velocities.

STAR B has only a singleabsorption line, so it must lie infront of the HVC. Thus, the twostars, whose distances can beestimated by independent means,place an upper and lower limit onthe cloud distance.

BIGGEST HASSLE in studyinghigh-velocity clouds is to

measure their distances. Thebest available technique is

indirect and approximate.Consider an HVC that lies

between two stars, labeled Aand B. Another, slower-

moving cloud of gas liesbetween us and Star B.

HVC 20,000 light-years

NEARBY GAS 100 light-years

STAR A 30,000 light-years

STAR B 5,000 light-years

HVC

–100 0 100 –100 0 100

VELOCITY (km/s) OR WAVELENGTH VELOCITY (km/s) OR WAVELENGTH

NEARBY GAS NEARBY GAS

OBSERVED SPECTRUM OF STAR A

VIEW FROM EARTH

OBSERVED SPECTRUM OF STAR B


similar to that of the Small Magellanic Cloud, as Limin Lu andhis co-workers at Wisconsin found in 1998. The HVC is lo-cated in the leading arm of the stream, meaning that whateverpulled it off the Small Magellanic Cloud also accelerated it.Frictional forces cannot do that; only tidal forces can. Lu’s dis-covery finally settles the question of the origin of the stream.

Frictional forces may still be important, however. FUSEfound highly ionized oxygen associated with the MagellanicStream, suggesting that it, too, is embedded in hot gas. Thegalactic corona must therefore extend much farther out thanwas originally proposed by Spitzer—out to a few hundred thou-sand light-years, rather than a few thousand. This corona is notdense enough to strip gas from the Magellanic Clouds, but oncethe gas has been drawn out by tidal forces, friction with thecorona causes it to decelerate, slowly rain down on the galaxyand contribute to the growth of the Milky Way.

Similarly, the dark matter hypothesis, although it does notexplain complexes A and C, may fit into the broader scheme ofthings. Blitz originally proposed that the intergalactic HVCsweigh 10 million to 100 million solar masses. Yet such cloudshave not been detected in nearby galaxy groups similar to theLocal Group, even though observations are now sensitiveenough to do so. Furthermore, the hypothesis predicts that vis-ible-light emission from HVCs should be too faint to detect, butin almost all cases that this emission has been looked for, it hasbeen detected. Finally, theoretical arguments show that if theHVCs are distant, they must be either fully ionized or extremelymassive, and both options are inconsistent with observations.It thus appears that HVCs are not the predicted population ofdark matter clouds.

Robert Braun of Dwingeloo Observatory and Butler Bur-ton and Vincent de Heij of Leiden instead propose that theMilky Way and Andromeda galaxies are surrounded by sev-eral hundred small clouds made mostly of dark matter and ion-ized gas, with a small fraction of neutral hydrogen. These

clouds would weigh at most 10 million solar masses, and ratherthan roaming throughout the Local Group, most would staywithin half a million light-years of the main galaxies.

Although neutral HVCs do not appear to be dispersedthroughout the Local Group, other types of high-velocity gasmay be. The highly ionized gas in one HVC lies far outside theMilky Way. FUSE has also discovered high-velocity, highly ion-ized oxygen on its own, without any neutral gas. Similar cloudsof hot gas have been found elsewhere in the universe by ToddM. Tripp of Princeton and his co-workers. This hot gas mayconstitute a filament running through intergalactic space. Suchfilaments show up in simulations of the broad-scale evolutionof the cosmos [see “The Emptiest Places,” by Evan Scanna-pieco, Patrick Petitjean and Tom Broadhurst; ScientificAmerican, October 2002], and the total amount of matter inthese filaments may be larger than that in all galaxies combined,

forming a reservoir that the Milky Way can draw on to makenew stars.

The HVCs surrounding the Milky Way remind us that weare living in a galaxy that is still forming and evolving. Origi-nally our galaxy was surrounded by many smaller satellitegalaxies and a lot of leftover gas. Over the past several billionyears, it has incorporated most of those satellites. It may alsohave accreted much of the pristine gas from its intergalactic en-virons, and plenty of gas may still lie out there. Gas is still trick-ling in, taking the form of HVCs. At the same time, the galaxyexpels gas loaded with heavy elements into its halo and maybeeven into intergalactic space.

Within the next 10 or so billion years, more satellite galax-ies will merge with the Milky Way, forming more of the stellarstreams now being discovered in the halo. Our galaxy is on acollision course with the Andromeda galaxy. We cannot tell ex-actly how the Milky Way, or what is left of it, will look in thedistant future, but we know that its formation has not cometo an end yet.

High-Velocity Clouds. Bart P. Wakker and Hugo van Woerden in Annual Review of Astronomy and Astrophysics, Vol. 35, pages 217–266;September 1997.

A Confirmed Location in the Galactic Halo for the High-Velocity Cloud“Chain A.” Hugo van Woerden, Ulrich J. Schwarz, Reynier F. Peletier, BartP. Wakker and Peter M. W. Kalberla in Nature, Vol. 400, pages 138–141;July 8, 1999. Available online at arXiv.org/abs/astro-ph/9907107

Accretion of Low-Metallicity Gas by the Milky Way. Bart P. Wakker, J. Chris Howk, Blair D. Savage, Hugo van Woerden, Steve L. Tufte, Ulrich J.Schwarz, Robert Benjamin, Ronald J. Reynolds, Reynier F. Peletier andPeter M. W. Kalberla in Nature, Vol. 402, No. 6760; pages 388–390;November 25, 1999.

The Formation and Evolution of the Milky Way. Cristina Chiappini in American Scientist, Vol. 89, No. 6, pages 506–515;November–December 2001.

A Far Ultraviolet Spectroscopic Explorer Survey of Molecular Hydrogenin Intermediate-Velocity Clouds in the Milky Way Halo. P. Richter, B. P. Wakker, B. D. Savage and K. R. Sembach in Astrophysical Journal, Vol. 586, No. 1, pages 230–248; March 20, 2003. arXiv.org/abs/astro-ph/0211356

Highly Ionized High-Velocity Gas in the Vicinity of the Galaxy. K. R. Sembach, B. P. Wakker, B. D. Savage, P. Richter, M. Meade, J. M. Shull,E. B. Jenkins, G. Sonneborn and H. W. Moos in Astrophysical Journal,Supplement Series, Vol. 146, No. 1, pages 165–208; May 2003.arXiv.org/abs/astro-ph/0207562

M O R E T O E X P L O R E

Filaments of hot intergalactic gas form a reservoir thatthe Milky Way can draw on to make new stars.


Ref

uges

for L

ife in

a

Hos

tile

Uni

verse

Only

par

t

ofou

r gal

axy

isfit

for

adva

nced

life

By G

uille

rmo

Gonz

alez

,

Dona

ld B

row

nlee

and

Pete

r D. W

ard

originally published in October 2001


NEA

RLY

EVE

RYT

HIN

Gse

ems

to h

ave

gone

wro

ng o

n th

is h

aple

ss w

orld

. Pum

mel

edby

com

ets,

frie

d by

nea

rby

supe

rnov

ae a

nd o

n th

e ve

rge

of b

eing

cat

apul

ted

into

deep

spa

ce b

y a

loom

ing

gas

gian

t, th

e ba

rren

pla

net i

n th

is a

rtis

t’s c

once

ptio

nm

ight

be

mor

e ty

pica

l ofr

ocky

bod

ies

in o

ur g

alax

y th

an E

arth

is.


to be so inventive is that scientists keep spoiling the fun. It usedto be quite respectable to speculate about intelligent beings onthe moon, Mars, Venus, Jupiter or even the sun, but nowadayscanal-building Martians and cool oases inside the sun are mere-ly quaint notions. As writers go ever farther afield to situatetheir characters, scientists are not far behind. Researchers arenow casting a skeptical eye on musings about the prevalence ofintelligent life throughout the Milky Way. Just as most of thesolar system is hostile to multicellular organisms, the same maybe true of much of the galaxy.

Within a given planetary system, astronomers describe theoptimal locations for life in terms of the circumstellar habitablezone (CHZ). Although its definition has varied, the CHZ is gen-erally considered to be the region around a star where liquidwater can persist on the surface of a terrestrial, or Earth-like,planet for at least a few billion years. The zone is ring-shaped[see illustration on next page]. Its inner boundary is the clos-est that a planet can orbit its host star without losing its oceansto space. In the most extreme case, a runaway greenhouse ef-fect might take hold and boil off the oceans (as happened onVenus). The outer boundary is the farthest a planet can roambefore its oceans freeze over. From basic stellar theory, as-tronomers can estimate the size of the CHZ for a star of anymass [see “How Climate Evolved on the Terrestrial Planets,”by James F. Kasting, Owen B. Toon and James B. Pollack; Sci-entific American, February 1988].

Obviously, many other factors also contribute to the hab-

itability of a planet, including the ellipticity of its orbit, the com-pany of a large moon and the presence of giant planets, let alonethe details of its biology. But if a planet orbits outside the zone,none of these minutiae is likely to matter. Similarly, it doesn’tmake much difference where the CHZ is located if the plane-tary system as a whole resides in a hostile part of the galaxy.

Thus, in 1999 we proposed the concept of a galactic equiv-alent to the CHZ: the galactic habitable zone (GHZ). The GHZdefines the most hospitable places in the Milky Way—those thatare neither too close nor too far from the galactic center. Weare not the first to consider habitability in this broader context.For the past decade Virginia Trimble of the University of Mary-land and the University of California at Irvine has been writingabout the connection between galactic chemical compositionand the conditions required for life. But in recent years, therehas been a huge breakthrough: the discovery of giant, Jupiter-size planets around sunlike stars. Not every sunlike star hassuch a planet. In fact, the giant planets discovered to date areprimarily found around stars that are rich in chemical elementsheavier than helium—what astronomers call “metals.” Thiscorrelation suggests that metal content is an important factorin forming giant planets. (At present, the leading search tech-nique cannot detect Earth-size planets.) At the same time, as-tronomers have gained a new and sobering appreciation of howdeadly our galaxy can be, filled as it is with exploding stars andstellar close encounters. Even where planets do exist, they maynot be fit for complex life-forms.

Where’s the Wherewithal?THE BOUNDARIES of the galactic habitable zone are set bytwo requirements: the availability of material to build a habit-able planet and adequate seclusion from cosmic threats. Thestory of how chemical elements came to be assembled intoEarth is one told by modern cosmology, stellar astrophysics andplanetary science. The big bang produced hydrogen and heli-um and little else. Over the next 10 billion years or so, starscooked this raw mix into a rich stew of elements. Within the in-terstellar medium, the ratio of the number of metal atoms tothe number of hydrogen atoms—that is, the “metallicity”—

gradually increased to its present value.These metals are the building blocks of Earth-like planets,

and their abundance affects the size of the planets that canform. Size, in turn, determines whether a planet can retain anatmosphere and sustain geologic activity. Moreover, withoutenough metals, no giant planets can form at all, because theycoalesce around a rocky core of a certain minimum size. Ob- E

DW

ARD

BE

LL A

ND

NAS

A (p

rece

din

g p

ag

es);

DO

N D

IXO

N (

opp

osit

e p

ag

e)

In science-fiction stories, interstellar travelers visit exotic locales in the Milky Way and meet with

interesting aliens. You name the place, and someone has put a civilization there: the galactic center, a globular

cluster, a star-forming region, a binary star system, a red dwarf star. Part of the reason that sci-fi writers have

■ What does a planet need to support complex life-forms?Astronomers have generally focused on the stability ofsurface water—which is possible only within a certain rangeof distances from the planet’s star, a region known as thecircumstellar habitable zone. But in discovering extrasolarplanets over the past five years or so, researchers havecome to appreciate a broader set of conditions.

■ Ideally, the star and its planetary retinue should orbitwithin a certain range of distances from the center of thegalaxy. Too far, and the nebula from which the star emergedwill lack the heavy elements out of which planets are made.Too close, and hazards such as orbital instabilities,cometary collisions and exploding stars will nipecosystems in the bud. The sun’s position is just right.

■ All this suggests that complex life is rare in the galaxy.

Overview/Habitable Zone


servations of extrasolar planets are beginning to define the re-quired metallicity for building giant planets. No such planet hasbeen found around any star with a metallicity of less than 40percent of the sun’s. In a study reported last year, the HubbleSpace Telescope failed to detect any planets in the globular clus-ter 47 Tucanae, whose stars have metallicities of 25 percent ofthe solar value [see “Searching for Shadows of Other Earths,”by Laurance R. Doyle, Hans-Jörg Deeg and Timothy M.Brown; Scientific American, September 2000].

Conversely, too high a metallicity can also be a problem.Terrestrial planets will be larger and, because of their strongergravity, richer in volatile compounds and poorer in topographicrelief. That combination will make them more likely to be com-pletely covered with water, to the detriment of life. On Earth,the mix of land and sea is important for atmospheric tempera-ture control and other processes. High metallicity also increas-es the density of the protoplanetary disk and thereby inducesthe giant planets to shift position [see “Migrating Planets,” byRenu Malhotra; Scientific American, September 1999].

A by-product of this orbital migration is that it will fling anysmaller, Earth-like bodies out of the system altogether or shovethem into the sun. As the elephants move around, the ants getcrushed.

In a recent study, Charles H. Lineweaver of the Universityof New South Wales in Australia explored the dependence ofplanet formation and migration on metallicity. He assumedthat the probability of forming a terrestrial planet is propor-tional to the metallicity of the parent star, because both the starand the planet arose from the same cloud of dust and gas. Fromthe extrasolar planet statistics, he inferred that the probabilityof giant-planet migration rises steeply with increasing metal-

HABITABLE ZONE of the Milky Way (green) excludes the dangerousinner regions and the metal-poor outer regions of our galaxy. It is analogous to the habitable zone on the much smaller scale of oursolar system (inset). Neither zone has sharp boundaries. The bulge is shown as yellow and the active star-forming regions in the spiralarms as blue and pink.

Astronomers have gained a new and sobering appreciation of HOW DEADLY OUR GALAXY CAN BE.

ORBIT OF JUPITER

SOLAR HABITABLE ZONE

ORBIT OF VENUS

SUN

ORBIT OF MARS

GALACTIC HABITABLE ZONE

ORBIT OF EARTH


licity, with migration inevitable if the metallicity is 300 percentof the solar value. Although Lineweaver’s calculations are ten-tative, they suggest that a metallicity near the sun’s may be op-timal for the production of Earth-mass planets in stable orbits.

Through Thick and ThinONLY PART OF THE MILKY WAY satisfies this require-ment. Astronomers usually subdivide the Milky Way into fouroverlapping regions: halo, bulge, thick disk and thin disk. Starsin each region orbit the galactic center much as planets in oursolar system orbit the sun. The halo and thick disk tend to con-tain older, metal-poor stars; it is unlikely that terrestrial plan-ets as large as Earth have formed around them. Stars in thebulge have a wide range of metallicities, but cosmic radiationlevels are higher there.

The thin disk is the sun’s home. The metallicity of its gas de-clines with distance from the galactic center. At the sun’s loca-tion, about 8.5 kiloparsecs (28,000 light-years) out, it is de-creasing at 17 percent per kiloparsec. The logarithm of themetallicity (which astronomers give in units called “dex,” thesun having a value of 0 dex, by definition) falls off linearly withdistance, with a slope of –0.07 dex per kiloparsec. Observersmeasure the metallicity gradient using spectral features in var-ious classes of stars and nebulae. The different indicators haveconverged onto the same answer only within the past three orfour years, and galaxies similar to the Milky Way are nowknown to have similar disk metallicity gradients.

The gradient is an outcome of variations in the star-formationrate. Farther from the center of the galaxy there is proportionate-ly less gas and therefore less star formation. Consequently, the out-er reaches of the galaxy have built up less metal than the innerparts. In the galaxy as a whole, the star-formation rate peakedabout eight billion to 10 billion years ago and has been declin-ing ever since. Today the metallicity in the solar neighborhoodis increasing by about 8 percent every billion years. As the gassupply dwindles, the metallicity will grow at an ever slower rate.

Taking into account the disk metallicity gradient and itsevolution, we can place rough limits on the GHZ both in spaceand in time [see illustrations on page 23]. Stars forming todaywith a metallicity of between 60 and 200 percent of the sun’svalue generally reside between 4.5 and 11.5 kiloparsecs fromthe galactic center—a region that contains only about 20 per-cent of the stars in the galaxy. Moreover, the typical star in thesolar neighborhood did not reach the 60 percent threshold un-til five billion to six billion years ago. The sun itself is about 40percent richer in metal than other stars formed at the same timeand location in the disk. This increased metal content may havegiven life on Earth a head start.

Iron CurtainsONE POTENTIAL COUNTERARGUMENT is that the cor-relation of metallicity and detected planets is not the same ascausation. Perhaps the causation goes in the opposite direction:instead of high stellar metallicity explaining the presence of gi-ant planets, the presence of giant planets might explain the highstellar metallicity. This would happen if they tended to fall intothe stars, enriching their metal content. Most astronomers nowthink that stars do gobble up planets and smaller bodies. Butthe outer convective layers of sunlike stars are so massive andso well mixed that they would need to devour an unreasonableamount of planetary material to fully account for the highmetallicities seen among stars with planets.

Another rejoinder is that the correlation might be an ob-servational bias. It is harder to spot planets around metal-poorstars; the leading planet search method relies on stellar spectrallines, which are weaker when a star has less metal. But the de-tection efficiency does not suffer appreciably until a star’s metal-licity drops below about 10 percent of the sun’s value—whichis well below the 40 percent threshold needed for giant planets.The observed correlation with planets is quite real.

Metallicity is not the only compositional prerequisite for hab-itable planets; the relative abundances of different elements mat-ter, too. The most abundant elements on Earth were producedprimarily in supernova explosions, of which there are two basictypes. Type I events, most of which result from the detonation ofa white dwarf star, produce mainly iron, nickel and cobalt. TypeII supernovae, which entail the implosion of a massive star, most-ly synthesize oxygen, silicon, magnesium, calcium and titanium.Crucially, type II events are also the sole natural source of the veryheaviest elements, such as thorium and uranium.

Because star formation in our galaxy is tapering off, the over-all rate of supernova explosions is declining—as is the ratio oftype II to type I events. Type II supernovae involve short-livedmassive stars, so their rate closely tracks the star-formation rate.The rate of type I supernovae, on the other hand, depends on the

SAR

A C

HE

N;

SOU

RC

E:

NU

NO

C.

SAN

TOS

Gen

eva

Ob

serv

ato

ry

0–0.5 0.5Logarithm of Metallicity (relative to solar)

0

10

20

30

Perc

enta

ge o

f St

ars

Sample of nearby sunlike stars

Stars with planets

More Metallic

SURVEYS OF EXTRASOLAR PLANETS conducted by astronomers haverevealed how important the supply of planet-building material is. As this histogram shows, the stars that are parents to giant planets(red area) tend to have a greater abundance of heavy elements(“metals”) than the average nearby sunlike star does (black).


NAS

A

production of longer-lived intermediate-mass stars, so it re-sponds more slowly to changes in the star-formation rate.

As a result of the shifting supernova ratio, new sunlike starsare richer in iron than those that formed five billion years ago.All else being equal, this implies that a terrestrial planet formingtoday will have a proportionately larger iron core than Earthdoes. It will also have, in 4.5 billion years, about 40 percent lessheat from the decay of potassium, thorium and uranium. Theheat generated by these radioactive isotopes is what drives platetectonics, which plays an essential role in the geochemical cyclethat regulates the amount of carbon dioxide in our atmosphere.Perhaps terrestrial planets forming today would be single-plateplanets like Venus and Mars. The lack of plate tectonics on Venuscontributes to its hellish conditions [see “Global Climate Changeon Venus,” by Mark A. Bullock and David H. Grinspoon; Sci-entific American, March 1999]. But we do not yet understandall the ways a planet’s geology depends on its internal heat flow.

Danger, DangerEVEN IF YOU MANAGE to get all the necessary atoms in theright place at the right time to build an Earth, you may not bejustified in sticking a “habitable” label on it. A planet must alsobe kept reasonably safe from threats. These threats can be putinto one of two categories: impacts by asteroids and comets,and blasts of radiation.

In our solar system the frequency of asteroid impacts de-pends on the details of Jupiter’s orbit and formation; the rest ofthe galaxy has no direct effect. The cometary threat, on the oth-er hand, is quite sensitive to the galactic environment. Cometsare thought to reside in two long-term reservoirs, the Kuiperbelt (which starts just beyond Neptune) and the Oort cloud(which extends halfway to the nearest star). Other stars prob-ably have similar retinues. Infrared observations of young near-by stars indicate that most are surrounded by excess dust, con-sistent with the presence of Kuiper-belt objects. More recent-ly, detection of water vapor around the highly evolved luminous

GUILLERMO GONZALEZ, DONALD BROWNLEE and PETER D. WARDshare an interest in the habitability of planets—both because theyhappen to live on one and because habitability is an intellectualchallenge that draws on nearly every field of astrophysics and geo-physics. The three are members of the astrobiology program at theUniversity of Washington, which NASA recently awarded an astrobi-ology grant. Gonzalez, currently at Iowa State University, earned hisdoctorate at Washington studying the compositions of highly evolvedstars in globular clusters. Brownlee specializes in the study of cometdust and meteorites and is the principal investigator for the Stardustmission, which plans to return comet dust samples to Earth in Jan-uary 2006. Ward, a paleontologist, studies global mass extinctions.

THE

AU

THO

RS

WHAT IS BEAUTIFUL is often dangerous, in space as on Earth. Some of the most renowned sites in the galaxyare hostile to planets, let alone living things. The safest places in the galaxy tend to be the most boring ones.

Trifid Nebula

Ionized gas

Proplyds in Orion Nebula

Being evaporated by giant stars

Cygnus Loop

Debris from stellar explosion

Galactic Center

Intense radiation, unstable orbits

Eagle Nebula

Being evaporated by giant stars

Globular Cluster M22

Few heavy elements

Globular Cluster Omega Centauri

Few heavy elements

O-type Star G339.88–1.26

Too bright, too short-lived


star IRC+10216 has been interpreted as evidence of evaporat-ing comets. Changes in the shapes of certain spectral lines inBeta Pictoris, a young star with a dust disk, could be caused byinfalling comets.

Because Oort-cloud comets are only weakly bound to thesun, it doesn’t take much to deflect them toward the inner plan-ets. A tug from galactic tides, giant molecular clouds or pass-ing stars can do the trick [see “The Oort Cloud,” by Paul R.Weissman; Scientific American, September 1998]. The fre-quency of such perturbations depends on our position in theMilky Way. As one goes toward the galactic center, the densi-ty of stars increases, so there are more close encounters. More-over, a planetary system forming out of a metal-rich cloud willprobably contain more comets than one forming out of a cloudwith less metal. Thus, planetary systems in the inner galaxyshould suffer higher comet influxes than the solar system does.Although the outer Oort cloud in such a system will become de-pleted more rapidly, it will also be replenished more rapidlyfrom the inner cometary reservoirs.

High-energy radiation, too, is a bigger problem in the innerregions of the galaxy. Up to a point, a planet’s magnetic fieldcan fend off most particle radiation and its ozone layer canscreen out dangerous electromagnetic radiation. But sufficient-ly energetic radiation can ionize the atmosphere and generatenitrogen oxides in amounts capable of wiping out the ozonelayer. Energetic radiation hitting the atmosphere can also letloose a deadly rain of secondary particles.

The nastiest radiation events are, in order of decreasing du-ration, active galactic nucleus outbursts, supernovae and gam-ma-ray bursts. The nucleus of the Milky Way is currently rela-tively inactive; the supermassive black hole at the heart of ourgalaxy appears to be dormant. But observations of other galax-ies suggest that central black holes occasionally turn on whena star or cluster wanders too close and is pulled to its death. Theresult is a burst of high-energy electromagnetic and particle ra-diation. Most of the radiation is emitted in a jet along the ro-tation axis of the galaxy, but many of the charged particles will

spiral along the galaxy’s magnetic field lines and fill its volume.The worst place to be during such an outburst is in the bulge.Not only would the overall radiation levels be high, the starsthere would tend to have highly inclined and elliptical orbitsthat could bring them close to the nucleus or jet.

Supernovae and gamma-ray bursts are also more threaten-ing in the inner galaxy, simply because of the higher concen-tration of stars there. Observations of supernova remnants in-dicate that supernovae peak at about 60 percent of the sun’sdistance from the galactic center, where they are about 1.6times more frequent than at our location. The threat from gam-ma-ray bursts remains uncertain; astronomers do not knowwhat triggers these gargantuan explosions or how tightly theybeam their radiation. We could just be lucky to have avoidedsuch a death ray so far.

Radiation can also steal life from the crib. Sunlike stars arenot born in isolation but rather are often surrounded by bothlow- and high-mass stars. The high levels of ultraviolet radia-tion emitted by the latter erode circumstellar disks around near-by stars, reducing their chances of forming giant planets. JohnBally of the University of Colorado at Boulder and his col-leagues have estimated that only about 10 percent of stars avoidthis kind of harassment. This could explain why a mere 3 per-cent of nearby sunlike stars are found to have giant planets.

All these threats imply a fairly broad habitable zone withfuzzy boundaries. But if we include proximity to the corotationcircle as another requirement, then the GHZ could be very nar-row. The corotation circle is where the orbital period of a starequals the rotation period of the galaxy’s spiral arm pattern. SA

RA

CH

EN

LOCATION OF HABITABLE ZONE is determined by a balance betweenthe supply of planet-building material and the prevalence of threats.The supply falls off with distance from the galactic center (left), whilethe density of stars—a proxy for perils such as stellar explosions andclose encounters—also decreases with distance (right). Anacceptable compromise is reached somewhere in the middle,although astronomers cannot yet pin down the precise location.

0 5 10 15

0.5

1.0

1.5

2.0

2.5

0

Distance from Center of Galaxy (kiloparsecs)

Met

allic

ity (r

elat

ive

to th

e su

n)

Minimal metal abundance for giant planetary formation

Sun

Peak

of

supe

rnov

a ra

te

Distance from Center of Galaxy (kiloparsecs)

Dens

ity o

f Mat

eria

l(s

olar

mas

ses

per c

ubic

par

sec)

0 5 10 15 20

0.001

0.01

0.1

1

Stars + interstellar gas and dustStarsInterstellar gas and dust

Sun



When a star orbits at or very near the corotation circle, spiralarm crossings are less frequent. It will take longer to cross a spi-ral arm, but what is important is the relatively long period be-tween crossings. Recent measurements of the dynamics of starsnear the sun indicate that the sun orbits very near the corota-tion circle. The spiral arms may look pretty, but they are bestappreciated from afar, because the intense star formation andgiant molecular clouds within the arms multiply the risks tocomplex life-forms.

Paradox LostAT THIS STAGE of our research, we are still some way fromfilling in the details of the GHZ. Continuing studies of comets,galactic nuclei, supernovae, gamma-ray bursts and stellar dy-namics will help pinpoint the threats to life. Even now, how-ever, we have a broad picture of the GHZ. The inner regions ofthe galaxy suffer from orbital instabilities, radiation bursts andcometary perturbations. The outer regions are safer, but be-cause of the lower metallicity, terrestrial planets are typically

smaller there. The GHZ appears to be an annulus in the disk atroughly the sun’s location [see illustration on page 20]. TheGHZ is a probabilistic concept: not every planet inside the zoneis habitable (and not every planet outside is sterile). But theprobability is much greater inside. The GHZ has been slowlycreeping outward, as interstellar gas reaches solar metallicity.

The GHZ concept has important implications for searchesfor extraterrestrial intelligence. It can, for example, identify themost probable places for complex life to form, so that researcherscan direct their searches accordingly. We can already say withsome confidence that globular clusters, the outer disk and thegalactic center make poor targets.

The GHZ concept also has implications for the debateswirling around the Fermi Paradox: If our galaxy is teeming withother civilizations, we should see some evidence of their exis-tence; we do not, so perhaps we are alone [see “Where AreThey?” by Ian Crawford; Scientific American, July 2000].One of the arguments proposed to avoid that conclusion is thatETs may have no motivation to leave their home world and scat-ter signs of their presence through space. But if our ideas aboutthe GHZ are correct, we live within an especially comfortableregion of the Milky Way. Any civilization seeking a new worldwould, no doubt, place our solar system on their home-shop-ping list. The GHZ theory also weakens the argument that thegalaxy is so big that interstellar explorers or colonizers havepassed us by. The GHZ may be large, but it is just a part of theentire galaxy, and any galactic travelers would tend to roamaround the annulus rather than haphazardly through the galaxy.

Furthermore, the GHZ concept constrains habitability notjust in space but also in time. The Milky Way used to be pum-

meled by supernovae and an active nucleus. Only in the pastfive billion years or so could civilizations have safely arisen. Thesun’s relatively high metallicity probably gave us a head start.Therefore, the GHZ concept may provide at least a partial so-lution to the Fermi Paradox: complex life is so rare and isolat-ed that we are effectively alone. To be sure, these implicationsapply only to complex life; simple organisms such as microbescould endure a much wider range of environments.

The broader universe looks even less inviting than ourgalaxy. About 80 percent of stars in the local universe residein galaxies that are less luminous than the Milky Way. Becausethe average metallicity of a galaxy correlates with its luminos-ity, entire galaxies could be deficient in Earth-size planets. An-other effect concerns the dynamics of stars in a galaxy. Likebees flying around a hive, stars in elliptical galaxies have ran-domized orbits and are therefore more likely to frequent theirmore dangerous central regions. In many ways, the Milky Wayis unusually hospitable: a disk galaxy with orderly orbits, com-paratively little dangerous activity and plenty of metals. It may

not remain so for long. The Andromeda galaxy is predicted tohave a close encounter with the Milky Way in about three bil-lion years. That event will dislodge most stars in the disk fromtheir regular orbits. It may also pour fresh fuel onto the MilkyWay’s central black hole and cause it to flare up, with possiblyunhappy consequences for the inhabitants of Earth.

Douglas Adams, that great expositor of simple truths, fa-mously summed up what he took to be the product of the pastfew centuries of progress in astronomy: “Far out in the un-charted backwaters of the unfashionable end of the western spi-ral arm of the Galaxy lies a small unregarded yellow sun.” Butas is often the case, fashionable is not the same as comfortable.We live in prime real estate.

Galactic Chemical Evolution: Implications for the Existence ofHabitable Planets. Virginia Trimble in Extraterrestrials: Where Are They?Edited by M. H. Hart and B. Zuckerman. Cambridge University Press, 1995.

Worlds Without End: The Exploration of Planets Known and Unknown.John S. Lewis. Perseus Books, 1998.

Destiny or Chance: Our Solar System and Its Place in the Cosmos.Stuart R. Taylor. Cambridge University Press, 1998.

Rare Earth: Why Complex Life Is Uncommon in the Universe. Peter D. Ward and Donald Brownlee. Copernicus, 2000.

An Estimate of the Age Distribution of Terrestrial Planets in theUniverse: Quantifying Metallicity as a Selection Effect.Charles H. Lineweaver in Icarus, Vol. 151, No. 2, pages 307–313; June 1,2001. Preprint available at astro-ph/0012399

The Galactic Habitable Zone: Galactic Chemical Evolution. GuillermoGonzalez, Donald Brownlee and Peter D. Ward in Icarus, Vol. 152, No. 1,pages 185–200; July 1, 2001. Preprint available at astro-ph/0103165


Any extraterrestrial civilization SEEKING A NEW WORLDwould place our solar system on their home-shopping list.



Like a boiling teakettle atop a COLD stove,

the sun’s HOT outer layers sit on the relatively cool surface.

And now astronomers are FIGURING OUT WHY

P u b l i s h e d i n N e w L i g h t o n t h e S o l a r S y s t e m , S p r i n g 2 0 0 3 ( u p d a t e d f r o m t h e J u n e 2 0 0 1 i s s u e )


SUSPENDED HIGH ABOVE the sun’s surface, a prominence (wispy stream) has erupted into the solar

atmosphere—the corona. The coronal plasma isinvisible in this ultraviolet image, which shows only the

cooler gas of the prominence and underlyingchromosphere. White areas are hotter and denser,

where higher magnetic fields exist; red areas are coolerand less dense, with weaker fields.

paradoxof the sun’s hot

the

By Bhola N. Dwivedi and Kenneth J. H. Phillips

corona


Relatively few people have witnessed a totaleclipse of the sun—one of nature’s most awesome spec-tacles. It was therefore a surprise for inhabitants of cen-tral Africa to see two total eclipses in quick succession,in June 2001 and December 2002. Thanks to favor-able weather along the narrow track of totality acrossthe earth, the 2001 event in particular captivated res-idents and visitors in Zambia’s densely populated cap-ital, Lusaka. One of us (Phillips), with colleagues fromthe U.K. and Poland, was also blessed with scientificequipment that worked perfectly on location at theUniversity of Zambia. Other scientific teams capturedvaluable data from Angola and Zimbabwe. Most of uswere trying to find yet more clues to one of the mostenduring conundrums of the solar system: What is themechanism that makes the sun’s outer atmosphere, orcorona, so hot?

The sun might appear to be a uniform sphere ofgas, the essence of simplicity. In actuality it has well-defined layers that can loosely be compared to a plan-et’s solid part and atmosphere. The solar radiation thatwe receive ultimately derives from nuclear reactions

deep in the core. The energy gradually leaks out untilit reaches the visible surface, known as the photo-sphere, and escapes into space. Above that surface isa tenuous atmosphere. The lowest part, the chromo-sphere, is usually visible only during total eclipses, asa bright red crescent. Beyond it is the pearly whitecorona, extending millions of kilometers. Further still,the corona becomes a stream of charged particles—thesolar wind that blows through our solar system.

Journeying out from the sun’s core, an imaginaryobserver first encounters temperatures of 15 millionkelvins, high enough to generate the nuclear reactionsthat power the sun. Temperatures get progressivelycooler en route to the photosphere, a mere 6,000 kel-vins. But then an unexpected thing happens: the tem-perature gradient reverses. The chromosphere’s tem-perature steadily rises to 10,000 kelvins, and goinginto the corona, the temperature jumps to one millionkelvins. Parts of the corona associated with sunspotsget even hotter. Considering that the energy mustoriginate below the photosphere, how can this be?It is as if you got warmer the farther away you

NAS

A G

OD

DAR

D S

PAC

E F

LIG

HT

CE

NTE

R (

pre

ced

ing

pa

ges

); T

RAC

E/N

ASA

(bel

ow)

CORONAL LOOP, seen in ultraviolet light by the TRACE spacecraft, extends 120,000kilometers off the sun’s surface.


walked from a fireplace.The first hints of this mystery emerged in the 19th

century when eclipse observers detected spectral emis-sion lines that no known element could account for. Inthe 1940s physicists associated two of these lines withiron atoms that had lost up to half their normal retinueof 26 electrons—a situation that requires extremelyhigh temperatures. Later, instruments on rockets andsatellites found that the sun emits copious x-rays andextreme ultraviolet radiation—as can be the case onlyif the coronal temperature is measured in megakelvins.Nor is this mystery confined to the sun: most sunlike

stars appear to have x-ray-emitting atmospheres.At last, however, a solution seems to be within our

grasp. Astronomers have long implicated magneticfields in the coronal heating; where those fields arestrongest, the corona is hottest. Such fields can trans-port energy in a form other than heat, thereby side-stepping the usual thermodynamic restrictions. The en-ergy must still be converted to heat, and researchers aretesting two possible theories: small-scale magnetic fieldreconnections—the same process involved in solarflares—and magnetic waves. Important clues havecome from complementary observations: spacecraft

X-RAY IMAGE from the Yohkoh spacecraft shows structures both bright (associated with sunspots) and dark (polar coronal holes).

INST

ITU

TE O

F SP

ACE

AN

D A

STR

ON

AUTI

CAL

SC

IEN

CE

, JA

PAN

; LO

CK

HE

ED

MAR

TIN

SO

LAR

AN

D A

STR

OP

HYS

ICS

LAB

OR

ATO

RY;

NAT

ION

AL

ASTR

ON

OM

ICAL

OB

SER

VATO

RY

OF

JAP

AN;

UN

IVE

RSI

TY O

F TO

KYO

; N

ASA


can observe at wavelengths inaccessible from theground, while ground-based telescopes can gatherreams of data unrestricted by the bandwidth of orbit-to-Earth radio links. The findings may be crucial to un-derstanding how events on the sun affect the atmosphereof Earth [see “The Fury of Space Storms,” by James L.Burch; Scientific American, April 2001].

The first high-resolution images of the corona camefrom the ultraviolet and x-ray telescopes on board Sky-lab, the American space station inhabited in 1973 and1974. Pictures of active regions of the corona, locatedabove sunspot groups, revealed complexes of loopsthat came and went in a matter of days. Much largerbut more diffuse x-ray arches stretched over millionsof kilometers, sometimes connecting sunspot groups.Away from active regions, in the “quiet” parts of thesun, ultraviolet emission had a honeycomb pattern re-lated to the large convection granules in the photo-sphere. Near the solar poles and sometimes in equa-torial locations were areas of very faint x-ray emis-

sion—the so-called coronal holes.

Connection to the Starry DynamoEACH MAJOR SOLAR SPACECRAFT since Skylabhas offered a distinct improvement in resolution. From1991 to late 2001, the x-ray telescope on the JapaneseYohkoh spacecraft routinely imaged the sun’s corona,tracking the evolution of loops and other featuresthrough one complete 11-year cycle of solar activity.The Solar and Heliospheric Observatory (SOHO), ajoint European-American satellite launched in 1995,orbits a point 1.5 million kilometers from Earth on itssunward side, giving the spacecraft the advantage of anuninterrupted view of the sun [see “SOHO Reveals theSecrets of the Sun,” by Kenneth R. Lang; ScientificAmerican, March 1997]. One of its instruments,called the Large Angle and Spectroscopic Coronagraph(LASCO), observes in visible light using an opaque diskto mask out the main part of the sun. It has trackedlarge-scale coronal structures as they rotate with the

FAR FROM A UNIFORM BALL of gas, the sun has a dynamic interior and atmosphere that heat and light our solar system.

DO

N D

IXO

N

Corona

Convective zone

Photosphere

Core

Radiative zoneSunspots

Coronal hole

Chromosphere

Solar wind

Prominence

Transitionregion


rest of the sun (a period of about 27 days as seen fromEarth). The images show huge bubbles of plasmaknown as coronal mass ejections, which move at up to2,000 kilometers a second, erupting from the coronaand occasionally colliding with Earth and other plan-ets. Other SOHO instruments, such as the Extreme Ul-traviolet Imaging Telescope, have greatly improved onSkylab’s pictures.

The Transition Region and Coronal Explorer(TRACE) satellite, operated by the Stanford-LockheedInstitute for Space Research, went into a polar orbitaround Earth in 1998. With unprecedented resolution,its ultraviolet telescope has revealed a vast wealth ofdetail. The active-region loops are now known to bethreadlike features no more than a few hundred kilo-meters wide. Their incessant flickering and jouncinghint at the origin of the corona’s heating mechanism.

The latest spacecraft dedicated to the sun is theReuven Ramaty High Energy Solar Spectroscopic Im-ager (RHESSI), launched in 2002, which is providingimages and spectra in the x-ray region of wavelengthsless than four nanometers. Because solar activity hasbeen high, much of its early attention was focused onintense flares, but as the solar minimum approaches,investigators will increasingly be interested in tiny mi-croflares, a clue to the corona’s heating mechanism.

The loops, arches and coronal holes trace out thesun’s magnetic fields. The fields are thought to origi-nate in the upper third of the solar interior, where en-ergy is transported mostly by convection rather thanradiation. A combination of convection currents anddifferential rotation—whereby low latitudes rotateslightly faster than higher latitudes—twist the fields toform ropelike or other tightly bound configurationsthat eventually emerge at the photosphere and into thesolar atmosphere. Particularly intense fields are markedby sunspot groups and active regions.

For a century, astronomers have measured the mag-netism of the photosphere using magnetographs, whichobserve the Zeeman effect: in the presence of a mag-netic field, a spectral line can split into two or more lineswith slightly different wavelengths and polarizations.But Zeeman observations for the corona have yet to bedone. The spectral splitting is too small to be detectedwith present instruments, so astronomers have had toresort to mathematical extrapolations from the photo-spheric field. These predict that the magnetic field of thecorona generally has a strength of about 10 gauss, 20times Earth’s magnetic field strength at its poles. In ac-tive regions, the field may reach 100 gauss.

Space HeatersTHESE FIELDS ARE WEAK compared with those thatcan be produced with laboratory magnets, but theyhave a decisive influence in the solar corona. This is be-

cause the corona’s temperature is so high that it is al-most fully ionized: it is a plasma, made up not of neu-tral atoms but of electrons, protons and other atomicnuclei. Plasmas undergo a wide range of phenomenathat neutral gases do not. The magnetic fields of thecorona are strong enough to bind the charged particlesto the field lines. Particles move in tight helical paths upand down these field lines like very small beads on verylong strings. The limits on their motion explain thesharp boundaries of features such as coronal holes.Within the tenuous plasma, the magnetic pressure (pro-portional to the strength squared) exceeds the thermalpressure by a factor of at least 100.

One of the main reasons astronomers are confidentthat magnetic fields energize the corona is the clear re-lation between field strength and temperature. Thebright loops of active regions, where there are ex-tremely strong fields, have a temperature of about fourmillion kelvins. But the giant arches of the quiet-suncorona, characterized by weak fields, have a tempera-ture of about one million kelvins.

Until recently, however, ascribing coronal heatingto magnetic fields ran into a serious problem. To con-vert field energy to heat energy, the fields must be ableto diffuse through the plasma, which requires that thecorona have a certain amount of electrical resistivity—

in other words, that it not be a perfect conductor. Aperfect conductor cannot sustain an electric field, be-cause charged particles instantaneously repositionthemselves to neutralize it. And if a plasma cannot sus-tain an electric field, it cannot move relative to the mag-netic field (or vice versa), because to do so would in-duce an electric field. This is why astronomers talkabout magnetic fields being “frozen” into plasmas.

This principle can be quantified by considering the

BHOLA N. DWIVEDI and KENNETH J. H. PHILLIPS begancollaborating on solar physics a decade ago. Dwivediteaches physics at Banaras Hindu University inVaranasi, India. He has been working with SUMER, an ul-traviolet telescope on the SOHO spacecraft, for morethan 10 years; the Max Planck Institute for Aeronomynear Hannover, Germany, recently awarded him one ofits highest honors, the Gold Pin. As a boy, Dwivedi stud-ied by the light of a homemade burner and became thefirst person in his village ever to attend college. Phillipsrecently left the Rutherford Appleton Laboratory in En-gland to become a senior research associate in theReuven Ramaty High Energy Solar Spectroscopic Imagergroup at the NASA Goddard Space Flight Center in Green-belt, Md. He has worked with x-ray and ultraviolet instru-ments on numerous spacecraft—including OSO-4, Solar-Max, IUE, Yohkoh, Chandra and SOHO—and has observedthree solar eclipses using CCD cameras.

THE

AU

THO

RS


time it takes a magnetic field to diffuse a certain distancethrough a plasma. The diffusion rate is inversely pro-portional to resistivity. Classical plasma physics assumesthat electrical resistance arises from so-called Coulombcollisions: electrostatic forces from charged particles de-flect the flow of electrons. If so, it should take about 10million years to traverse a distance of 10,000 kilometers,a typical length of active-region loops.

Events in the corona—for example, flares, whichmay last for only a few minutes—far outpace that rate.Either the resistivity is unusually high or the diffusiondistance is extremely small, or both. A distance as shortas a few meters could occur in certain structures, ac-companied by a steep magnetic gradient. But researchershave come to realize that the resistivity could be higherthan they traditionally thought.

Raising the MercuryASTRONOMERS HAVE TWO basic ideas for coro-nal heating. For years, they concentrated on heating bywaves. Sound waves were a prime suspect, but in the late1970s researchers established that sound waves emerg-ing from the photosphere would dissipate in the chro-mosphere, leaving no energy for the corona itself. Sus-picion turned to magnetic waves. Such waves might bepurely magnetohydrodynamic (MHD)—so-called Alf-vén waves—in which the field lines oscillate but thepressure does not. More likely, however, they sharecharacteristics of both sound and Alfvén waves.

MHD theory combines two theories that are chal-lenging in their own right—ordinary hydrodynamicsand electromagnetism—although the broad outlinesare clear. Plasma physicists recognize two kinds ofMHD pressure waves, fast and slow mode, dependingon the phase velocity relative to an Alfvén wave—

around 2,000 kilometers a second in the corona. Totraverse a typical active-region loop requires about fiveseconds for an Alfvén wave, less for a fast MHD wave,but at least half a minute for a slow wave. MHD wavesare set into motion by convective perturbations in the

photosphere and transported out into the corona viamagnetic fields. They can then deposit their energy intothe plasma if it has sufficient resistivity or viscosity.

A breakthrough occurred in 1998 when theTRACE spacecraft observed a powerful flare that trig-gered waves in nearby fine loops. The loops oscillatedback and forth several times before settling down. Thedamping rate was millions of times as fast as classicaltheory predicts. This landmark observation of “coro-nal seismology” by Valery M. Nakariakov, then at theUniversity of St. Andrews in Scotland, and his col-leagues has shown that MHD waves could indeed de-posit their energy into the corona.

An intriguing observation made with the ultravio-let coronagraph on the SOHO spacecraft has shownthat highly ionized oxygen atoms have temperaturesin coronal holes of more than 100 million kelvins,much higher than those of electrons and protons in theplasma. The temperatures also seem higher perpen-dicular to the magnetic field lines than parallel to them.Whether this is important for coronal heating remainsto be seen.

Despite the plausibility of energy transport bywaves, a second idea has been ascendant: that coronalheating is caused by very small, flarelike events. A flareis a sudden release of up to 1025 joules of energy in anactive region of the sun. It is thought to be caused by re-connection of magnetic field lines, whereby oppositelydirected lines cancel each other out, converting mag-netic energy into heat. The process requires that thefield lines be able to diffuse through the plasma.

A flare sends out a blast of x-rays and ultravioletradiation. At the peak of the solar cycle (reached in2000), several flares an hour may burst out across thesun. Spacecraft such as Yohkoh and SOHO haveshown that much smaller but more frequent eventstake place not only in active regions but also in regionsotherwise deemed quiet. These tiny events have abouta millionth the energy of a full-blown flare and so arecalled microflares. They were first detected in 1980 byRobert P. Lin of the University of California at Berke-ley and his colleagues with a balloon-borne hard x-raydetector. During the solar minimum in 1996, Yohkohalso recognized events with energy as small as 0.01 ofa microflare.

Early results from the RHESSI measurements indi-cate more than 10 hard x-ray microflares an hour. Inaddition, RHESSI can produce images of microflares,which was not possible before. As solar activity de-clines, RHESSI should be able to locate and charac-terize very small flares.

Flares are not the only type of transient phenome-na. X-ray and ultraviolet jets, representing columns ofcoronal material, are often seen spurting up from thelower corona at a few hundred kilometers a second. But C

OU

RTE

SY O

F SA

M K

RU

CK

ER

Un

iver

sity

of

Ca

lifo

rnia

, B

erke

ley

X-RAY IMAGE taken by the RHESSI spacecraft outlines theprogression of a microflare on May 6, 2002. The flare peaked(left), then six minutes later (right) began to form loops over the original flare site.



tiny x-ray flares are of special interest because theyreach the megakelvin temperatures required to heat thecorona. Several researchers have attempted to extrap-olate the microflare rates to even tinier nanoflares, totest an idea raised some years ago by Eugene Parker ofthe University of Chicago that numerous nanoflares oc-curring outside of active regions could account for theentire energy of the corona. Results remain confusing,but perhaps the combination of RHESSI, TRACE andSOHO data during the forthcoming minimum can pro-vide an answer.

Which mechanism—waves or nanoflares—domi-nates? It depends on the photospheric motions thatperturb the magnetic field. If these motions operate ontimescales of half a minute or longer, they cannot trig-ger MHD waves. Instead they create narrow currentsheets in which reconnections can occur. Very high res-olution optical observations of bright filigree structuresby the Swedish Vacuum Tower Telescope on La Pal-ma in the Canary Islands—as well as SOHO andTRACE observations of a general, ever changing“magnetic carpet” on the surface of the sun—demon-strate that motions occur on a variety of timescales. Al-though the evidence now favors nanoflares for the bulkof coronal heating, waves may also play a role.

FieldworkIT IS UNLIKELY, for example, that nanoflares havemuch effect in coronal holes. In these regions, the fieldlines open out into space rather than loop back to thesun, so a reconnection would accelerate plasma out intointerplanetary space rather than heat it. Yet the coro-na in holes is still hot. Astronomers have scanned forsignatures of wave motions, which may include peri-odic fluctuations in brightness or Doppler shift. Thedifficulty is that the MHD waves involved in heatingprobably have very short periods, perhaps just a fewseconds. At present, spacecraft imaging is too sluggishto capture them.

For this reason, ground-based instruments remainimportant. A pioneer in this work has been Jay M.Pasachoff of Williams College. He and his studentshave used high-speed detectors and CCD cameras tolook for modulations in the coronal light duringeclipses. Analyses of his best results indicate oscillationswith periods of one to two seconds. Serge Koutchmyof the Institute of Astrophysics in Paris, using a corona-graph, has found evidence of periods equal to 43, 80 and300 seconds.

The search for those oscillations is what led Phillipsand his colleagues to Bulgaria in 1999 and Zambia in2001. Our instrument consists of a pair of fast-frameCCD cameras that observe both white light and thegreen spectral line produced by highly ionized iron. Atracking mirror, or heliostat, directs sunlight into a hor-

izontal beam that passes into the instrument. At our ob-serving sites, the 1999 eclipse totality lasted two min-utes and 23 seconds, the 2001 totality three minutesand 38 seconds. Analyses of the 1999 eclipse by DavidA. Williams, now at University College London, revealthe possible presence of an MHD wave with fast-modecharacteristics moving down a looplike structure. TheCCD signal for this eclipse is admittedly weak, how-ever, and Fourier analysis by Pawel Rudawy of the Uni-versity of Wroclaw in Poland fails to find significant pe-riodicities in the 1999 and 2001 data. We continue totry to determine if there are other, nonperiodic changes.

Insight into coronal heating has also come from ob-servations of other stars. Current instruments cannotsee surface features of these stars directly, but spectros-copy can deduce the presence of starspots, and ultra-violet and x-ray observations can reveal coronae andflares, which are often much more powerful than theirsolar counterparts. High-resolution spectra from theExtreme Ultraviolet Explorer and the latest x-ray satel-lites, Chandra and XMM-Newton, can probe tem-perature and density. For example, Capella—a stellarsystem consisting of two giant stars—has photospher-ic temperatures like the sun’s but coronal temperaturesthat are six times higher. The intensities of individualspectral lines indicate a plasma density of about 100times that of the solar corona. This high density im-plies that Capella’s coronae are much smaller than thesun’s, stretching out a tenth or less of a stellar diame-ter. Apparently, the distribution of the magnetic fielddiffers from star to star. For some stars, tightly orbit-ing planets might even play a role.

Even as one corona mystery begins to yield to ourconcerted efforts, additional ones appear. The sun andother stars, with their complex layering, magneticfields and effervescent dynamism, still manage to defyour understanding. In an age of such exotica as blackholes and dark matter, even something that seems mun-dane can retain its allure.

JEAN

MO

UE

TTE

AN

D S

ER

GE

KO

UTC

HM

Y, ©

IN

STIT

UT

D’A

STR

OP

HYS

IQU

E D

E P

ARIS

M O R E T O E X P L O R EGuide to the Sun. Kenneth J. H. Phillips. Cambridge UniversityPress, 1992.

The Solar Corona above Polar Coronal Holes as Seen bySUMER on SOHO. Klaus Wilhelm et al. in Astrophysical Journal, Vol. 500, No. 2,pages 1023–1038; June 20, 1998.

Today’s Science of the Sun, Parts 1 and 2. Carolus J. Schrijverand Alan M. Title in Sky & Telescope, Vol. 101, No. 2, pages 34–39;February 2001; and No. 3, pages 34–40; March 2001.

Glorious Eclipses: Their Past, Present and Future. SergeBrunier and Jean-Pierre Luminet. Cambridge University Press,2001.

Probing the Sun’s Hot Corona. K.J.H. Phillips and B. N. Dwivediin Dynamic Sun. Edited by B. N. Dwivedi. Cambridge UniversityPress, 2003.



Filled with colossal fountains of hotgas and vast bubbles blown byexploding stars, the interstellar mediumis far more interesting than scientistsonce thought

The Gas between the Starsby Ronald J. Reynolds

MILKY WAY GALAXY looks profoundly different depending onthe frequency at which astronomers observe it. Fifty years ago,when astronomers were restricted to visible light, interstellargas seemed like just a nuisance—blocking the real objects ofinterest, the stars. Today scientists think the gas may be asimportant to the evolution of the galaxy as are the stars.These panels appear on a poster prepared by the NASA GoddardSpace Flight Center; for more information, visithttp://nvo.gsfc.nasa.gov/mw/mmw_sci.html

originally published in January 2002


NASA GSFC ASTROPHYSICS DATA FACILITY (radio continuum [408 MHz], atomic hydrogen, far-infrared, x-ray and gamma ray); ROY DUNCAN Software Infrastructure Group (radio continuum (2.4–2.7 GHz));THOMAS DAME Harvard-Smithsonian Center for Astrophysics (molecular hydrogen); STEPHAN D. PRICE Hanscom AFB (mid-infrared); AXEL MELLINGER University of Potsdam (visible l ight)


RADIO CONTINUUM(408 MHz)Reveals fast-movingelectrons, foundespecially at sites ofpast supernovae

ATOMIC HYDROGEN (1420 MHz)Reveals neutralatomic hydrogen ininterstellar cloudsand diffuse gas

RADIO CONTINUUM (2.4–2.7 GHz)Reveals warm,ionized gas and high-energy electrons

MOLECULAR HYDROGEN (115 GHz)Reveals molecularhydrogen (as tracedby carbon monoxide)in cold clouds

FAR-INFRARED(12–100 microns)Reveals dust warmedby starlight, speciallyin star-formingregions

MID-INFRARED (6.8–10.8 microns)Reveals complexmolecules ininterstellar clouds, aswell as reddish stars

VISIBLE LIGHT (0.4–0.6 micron) Reveals nearby starsand tenuous ionizedgas; dark areas arecold and dense

X-RAY (0.25–1.5 kiloelectron-volt)Reveals hot, shockedgas from supernovae

GAMMA RAY(greater than 300megaelectron-volts)Reveals high-energyphenomena such aspulsars and cosmic-ray collisions


1A superbubbleoriginates with

a cluster ofmassive stars.

2One star goessupernova, forming

a bubble of hot, low-density gas.

3Because massivestars have similar

lifespans, another onesoon blows.

4winbub

The Galaxy’s Dynamic Atmosphere

The views above and on the preceding page are cross sectionsthrough the Milky Way.


4The two bubbleslink up. Stellar

nds help energize thebbles.

5A third explodes.The interstellar

medium starts to looklike Swiss cheese.

6All three bubbleslink up, forming a

passage for hot gasand radiation.

Some of the interstellar medium takes the form of discrete clouds of atomichydrogen (H I) or molecular hydrogen (H2); most of the rest is in a pervasive ionized(H II) or atomic gas. Intermixed is a trace amount of other elements. The total massis about one fifth of the galaxy’s stars.

Composition of the GalacticAtmosphere

H2 H I WARM H I WARM H II HOT H II15 120 8,000 8,000 ~106

200 25 0.3 0.15 0.002

150 200 1,000 2,000 6,000

0.1 2 35 20 43

18 30 30 20 2

Component

Temperature (K)

Midplane Density (cm–3)

Thickness of Layer (parsecs)

Volume Fraction (%)

Mass Fraction (%)

IN CLOUDS BETWEEN CLOUDS

The term “interstellar medium” once conjured up a picture like the one at right: frigid,inky clouds of gas and dust in repose near the galactic plane. Today astronomersrecognize the medium as a protean atmosphere roiled by supernova explosions. Gasgushes through towering chimneys, then showers back down in mighty fountains. C

. H

OW

K A

ND

BLA

IRSA

VAG

E U

niv

ersi

ty

of W

isco

nsi

n A

ND

N.

SHAR

P N

OAO


100 degrees below zero to 100 degrees above with a small step.You could yell in your friend’s ear and he would never hear you.Without an atmosphere to transmit heat or sound, each patchof the moon is an island in an unnavigable sea.

The atmosphere of a planet is what binds its surface into a uni-fied whole. It lets conditions such as temperature vary smoothly.More dramatically, events such as the impact of an asteroid, theeruption of a volcano and the emission of gas from a factory’schimney can have effects that reach far beyond the spots wherethey took place. Local phenomena can have global consequences.This characteristic of atmospheres has begun to capture the in-terest of astronomers who study the Milky Way galaxy.

For many years, we have known that an extremely thin at-mosphere called the interstellar medium envelops our galaxyand threads the space between its billions of stars. Until fairlyrecently, the medium seemed a cold, static reservoir of gas qui-etly waiting to condense into stars. You barely even notice itwhen looking up into the starry sky. Now we recognize themedium as a tempestuous mixture with an extreme diversity ofdensity, temperature and ionization. Supernova explosions blowgiant bubbles; fountains and chimneys may arch above the spi-ral disk; and clouds could be falling in from beyond the disk.These and other processes interconnect far-flung reaches of ourgalaxy much as atmospheric phenomena convey disturbancesfrom one side of Earth to the other.

In fact, telescopes on the ground and in space are showingthe galaxy’s atmosphere to be as complex as any planet’s. Heldby the combined gravitational pull of the stars and other mat-ter, permeated by starlight, energetic particles and a magneticfield, the interstellar medium is continuously stirred, heated, re-cycled and transformed. Like any atmosphere, it has its highestdensity and pressure at the “bottom,” in this case the plane thatdefines the middle of the galaxy, where the pressure must bal-ance the weight of the medium from “above.” Dense concen-trations of gas—clouds—form near the midplane, and from thedensest subcondensations, stars precipitate.

When stars exhaust their nuclear fuel and die, those that areat least as massive as the sun expel much of their matter backinto the interstellar medium. Thus, as the galaxy ages, each gen-eration of stars pollutes the medium with heavy elements. As inthe water cycle on Earth, precipitation is followed by “evapo-ration,” so that material can be recycled over and over again.

Up in the AirTHINKING OF THE INTERSTELLAR medium as a true at-mosphere brings unity to some of the most pressing problems inastrophysics. First and foremost is star formation. Although as-tronomers have known the basic principles for decades, they still

do not grasp exactly what determines when and at what ratestars precipitate from the interstellar medium. Theorists used toexplain the creation of stars only in terms of the local conditionswithin an isolated gas cloud. Now they are considering condi-tions in the galaxy as a whole.

Not only do these conditions influence star formation, theyare influenced by it. What one generation of stars does deter-mines the environment in which subsequent generations areborn, live and die. Understanding this feedback—the sway ofstars, especially the hottest, rarest, most massive stars, over thelarge-scale properties of the interstellar medium—is another ofthe great challenges for researchers. Feedback can be both pos-itive and negative. On the one hand, massive stars can heat andionize the medium and cause it to bulge out from the midplane.This expansion increases the ambient pressure, compressing theclouds and perhaps triggering their collapse into a new genera-tion of stars. On the other hand, the heating and ionization canalso agitate clouds, inhibiting the birth of new stars. When thelargest stars blow up, they can even destroy the clouds that gavethem birth. In fact, negative feedback could explain why the grav-itational collapse of clouds into stars is so inefficient. Typicallyonly a few percent of a cloud’s mass becomes stars.

A third conundrum is that star formation often occurs in spo-radic but intense bursts. In the Milky Way the competing feed-back effects almost balance out, so that stars form at an unhur-ried pace—just 10 per year on average. In some galaxies, how-ever, such as the “exploding galaxy” M82, positive feedback hasgained the upper hand. Starting 20 million to 50 million years ago,star formation in the central parts of M82 began running out ofcontrol, proceeding 10 times faster than before. Our galaxy, too,may have had sporadic bursts. How these starbursts occur andwhat turns them off must be tied to the complex relation betweenstars and the tenuous atmosphere from which they precipitate.

Finally, astronomers debate how quickly the atmosphericactivity is petering out. The majority of stars—those less mas-sive than the sun, which live tens or even hundreds of billionsof years—do not contribute to the feedback loops. More andmore of the interstellar gas is being locked up into very long livedstars. Eventually all the spare gas in our Milky Way may be ex-hausted, leaving only stellar dregs behind. How soon this willhappen depends on whether the Milky Way is a closed box. Re-cent observations suggest that the galaxy is still an open system,both gaining and losing mass to its cosmic surroundings. High-velocity clouds of relatively unpolluted hydrogen appear to beraining down from intergalactic space, rejuvenating our galaxy.Meanwhile the galaxy may be shedding gas in the form of ahigh-speed wind from its outer atmosphere, much as the sunslowly sheds mass in the solar wind.D

ON

DIX

ON

We often think of the moon as a place, but in fact it is a hundredmillion places, an archipelago of solitude. You could go from


Hot and Cold Running HydrogenTO TACKLE THESE PROBLEMS, those of us who study theinterstellar medium have first had to identify its diverse com-ponents. Astronomers carried out the initial step, an analysis ofits elemental composition, in the 1950s and 1960s using thespectra of light emitted by bright nebulae, such as the OrionNebula. In terms of the number of atomic nuclei, hydrogen con-stitutes 90 percent, helium about 10 percent, and everythingelse—from lithium to uranium—just a trace, about 0.1 percent.

Because hydrogen is so dominant, the structure of thegalaxy’s atmosphere depends mainly on what forms the hydro-gen takes. Early observations were sensitive primarily to cooler,neutral components. The primary marker of interstellar mater-ial is the most famous spectral line of astronomy: the 1,420-megahertz (21-centimeter) line emitted by neutral hydrogenatoms, denoted by astronomers as H I. Beginning in the 1950s,radio astronomers mapped out the distribution of H I within the

galaxy. It resides in lumps and filaments with densities of 10 to100 atoms per cubic centimeter and temperatures near 100kelvins, embedded in a more diffuse, thinner (roughly 0.1 atomper cubic centimeter) and warmer (a few thousand kelvins)phase. Most of the H I is close to the galactic midplane, form-ing a gaseous disk about 300 parsecs (1,000 light-years) thick,roughly half the thickness of the main stellar disk you see whenyou notice the Milky Way in the night sky.

Hydrogen can also come in a molecular form (H2), which isextremely difficult to detect directly. Much of the informationabout it has been inferred from high-frequency radio observa-

tions of the trace molecule carbon monoxide. Where carbonmonoxide exists, so should molecular hydrogen. The moleculesappear to be confined to the densest and coldest clouds—theplaces where starlight, which breaks molecules into their con-stituent atoms, cannot penetrate. These dense clouds, which areactive sites of star formation, are found in a thin layer (100 par-secs thick) at the very bottom of the galactic atmosphere.

Until very recently, hydrogen molecules were seen directlyonly in places where they were being destroyed—that is, con-verted to atomic hydrogen—by a nearby star’s ultraviolet radi-ation or wind of outflowing particles. In these environments, H2

glows at an infrared wavelength of about 2.2 microns. In thepast few years, however, orbiting spectrographs, such as theshuttle-based platform called ORFEUS-SPAS and the new FarUltraviolet Spectroscopic Explorer (FUSE) satellite, have soughtmolecular hydrogen at ultraviolet wavelengths near 0.1 micron.These instruments look for hydrogen that is backlit by distant

stars and quasars: the H2 leaves telltale absorption lines in theultraviolet spectra of those objects. The advantage of this ap-proach is that it can detect molecular hydrogen in quiescent re-gions of the galaxy, far from any star.

To general astonishment, two teams, led respectively byPhilipp Richter of the University of Wisconsin and WolfgangGringel of the University of Tübingen in Germany, have dis-covered H2 not just in the usual places—the high-density cloudslocated within the galactic disk—but also in low-density areasfar outside the disk. This is a bit of a mystery, because high den-sities are needed to shield the molecules from the ravages ofstarlight. Perhaps a population of cool clouds extends much far-ther from the midplane than previously believed.

A third form of hydrogen is a plasma of hydrogen ions. As-tronomers used to assume that ionized hydrogen was confinedto a few small, isolated locations—the glowing nebulae near lu-minous stars and the wispy remnants left over from supernovae.Advances in detection technology and the advent of space as-tronomy have changed that. Two new components of ourgalaxy’s atmosphere have come into view: hot (106 kelvins) andwarm (104 kelvins) ionized hydrogen (H II). JA

YAN

NE

EN

GLI

SH U

niv

ersi

ty o

f M

an

itob

a;

CAN

ADIA

N G

ALAC

TIC

PLA

NE

SU

RVE

Y AN

D R

USS

TAY

LOR

Un

iver

sity

of

Ca

lga

ry

RONALD J. REYNOLDS bought a 4.25-inch reflecting telescope insixth grade and used it to take pictures of the moon. But it wasn’tuntil he started his Ph.D. in physics that he took his first astrono-my course and began to consider a career in the subject. TodayReynolds is an astronomy professor at the University of Wiscon-sin–Madison. He has designed and built high-sensitivity spec-trometers to study warm ionized gas in the Milky Way galaxy. Heis principal investigator for the Wisconsin H-Alpha Mapper, whichspent two years mapping hydrogen over the entire northern sky.

THE

AU

THO

R

How the stars wield power over an entire galaxy is unclear, but astronomers generally pin the blame on the creation of hot ionized gas.


Like the recently detected hydrogen molecules, these H II

phases stretch far above the cold H I cloud layer, forming a thickgaseous “halo” around the entire galaxy. “Interstellar” no longerseems an appropriate description for these outermost parts of ourgalaxy’s atmosphere. The hotter phase may extend thousandsof parsecs from the midplane and thin out to a density near 10–3

ion per cubic centimeter. It is our galaxy’s corona, analogousto the extended hot atmosphere of our sun. As in the case of thesolar corona, the mere existence of the galactic corona impliesan unconventional source of energy to maintain the high tem-peratures. Supernova shocks and fast stellar winds appear to dothe trick. Coexisting with the hot plasma is the warm plasma,which is powered by extreme ultraviolet radiation. The weightof these extended layers increases the gas pressure at the mid-plane, with significant effects on star formation. Other galaxiesappear to have coronas as well. The Chandra X-ray Observa-tory has recently seen one around the galaxy NGC 4631 [seebottom illustration on next page].

Blowing BubblesHAVING IDENTIFIED these new, more energetic phases of themedium, astronomers have turned to the question of how the di-

verse components behave and interrelate. Not only does the in-terstellar medium cycle through stars, it changes from H2 to H I

to H II and from cold to hot and back again. Massive stars are theonly known source of energy powerful enough to account for allthis activity. A study by Ralf-Jürgen Dettmar of the University ofBochum in Germany found that galaxies with a larger-than-av-erage massive star population seem to have atmospheres that aremore extended or puffed up. How the stars wield power over anentire galaxy is somewhat unclear, but astronomers generally pinthe blame on the creation of hot ionized gas.

This gas appears to be produced by the high-velocity (100 to200 kilometers per second) shock waves that expand into the in-terstellar medium following a supernova. Depending on the den-sity of the gas and strength of the magnetic field in the ambientmedium, the spherically expanding shock may clear out a cavi-ty 50 to 100 parsecs in radius—a giant bubble.

In doing so, the shock accelerates a small fraction of the ionsand electrons to near light speed. Known as cosmic rays, thesefleet-footed particles are one way that stellar death feeds back(both positively and negatively) into stellar birth. Cosmic raysraise the pressure of the interstellar medium; higher pressures,in turn, compress the dense molecular clouds and increase the

DO

N D

IXO

N

star formation

dying starsmassive stars

ioni

zing

radi

atio

n an

d he

at material and kinetic energy

infallingclouds

galacticwinds

INTERGALACTIC MEDIUM

INTERSTELLAR MEDIUM

STARS

star formation

dying starsmassive stars

ioni

zing

radi

atio

n an

d he

at material and kinetic energy

infallingclouds

galacticwinds

INTERGALACTIC MEDIUM

INTERSTELLAR MEDIUM

STARS

RECYCLING OF GAS by thegalaxy is analogous to thewater cycle on Earth. Theinterstellar medium plays thepart of the atmosphere. Stars“precipitate” out and then“evaporate” back; the moremassive ones energize andstir the medium. And just asEarth loses material to (andgains material from) inter-planetary space, so too doesthe galaxy exchange materialwith intergalactic space.


chance that they will collapse into stars. By ionizing some of thehydrogen, the cosmic rays also drive chemical reactions that syn-thesize complex molecules, some of which are the buildingblocks of life as we know it. And because the ions attach them-selves to magnetic field lines, they trap the field within the clouds,which slows the rate of cloud collapse into stars.

If hot bubbles are created frequently enough, they could in-terconnect in a vast froth. This idea was first advanced in the1970s by Barham Smith and Donald Cox of the University ofWisconsin–Madison. A couple of years later Christopher F. Mc-Kee of the University of California at Berkeley and Jeremiah P.Ostriker of Princeton University argued that the hot phase

should occupy 55 to 75 percent of interstellar space. Cooler neu-tral phases would be confined to isolated clouds within this ion-ized matrix—essentially the inverse of the traditional picture,in which the neutral gas dominates and the ionized gas is con-fined to small pockets.

Recent observations seem to support this upending of con-ventional wisdom. The nearby spiral galaxy M101, for exam-ple, has a circular disk of atomic hydrogen gas riddled withholes—presumably blown by massive stars. The interstellar medi-um of another galaxy, seven billion light-years distant, also lookslike Swiss cheese. But the amount of hot gas and its influence onthe structure of galactic atmospheres still occasion much debate.

Chimneys and FountainsTHE SUN ITSELF APPEARS to be located within a hot bub-ble, which has revealed itself in x-rays emitted by highly ionizedtrace ions such as oxygen. Called the Local Bubble, this regionof hot gas was apparently created by a nearby supernova aboutone million years ago.

An even more spectacular example lies 450 parsecs from thesun in the direction of the constellations Orion a nd Eridanus. Itwas the subject of a recent study by Carl Heiles of the Universi-ty of California at Berkeley and his colleagues. The Orion-Eridanus Bubble was formed by a star cluster in the constella-tion Orion. The cluster is of an elite type called an OB associa-tion—a bundle of the hottest and most massive stars, the O- andB-type stars, which are 20 to 60 times heavier than the sun (a G-type star) and 103 to 105 times brighter. The spectacular deathsof these short-lived stars in supernovae over the past 10 millionyears have swept the ambient gas into a shell-like skin around theouter boundary of the bubble. In visible light the shell appearsas a faint lacework of ionized loops and filaments. The million-degree gas that fills its interior gives off a diffuse glow of x-rays.

The entire area is a veritable thunderstorm of star forma-tion, with no sign of letting up. Stars continue to precipitatefrom the giant molecular cloud out of which the OB associa-tion emerged. One of the newest O stars, theta1 C Orionis, isionizing a small piece of the cloud—producing the Orion Neb-ula. In time, however, supernovae and ionizing radiation willcompletely disrupt the molecular cloud and dissociate its mol-ecules. The molecular hydrogen will turn back into atomic andionized hydrogen, and star formation will cease. Because the vi-olent conversion process will increase the pressure in the inter-stellar medium, the demise of this molecular cloud may meanthe birth of stars elsewhere in the galaxy.

Galactic bubbles should buoyantly lift off from the galacticmidplane, like a thermal rising above the heated ground onEarth. Numerical calculations, such as those recently made byMordecai-Mark MacLow of the American Museum of Natur-al History in New York City and his colleagues, suggest thatbubbles can ascend all the way up into the halo of the galaxy.The result is a cosmic chimney through which hot gas spewedby supernovae near the midplane can vent to the galaxy’s upperatmosphere. There the gas will cool and rain back onto thegalactic disk. In this case, the superbubble and chimney become R

ON

ALD

J.

RE

YNO

LDS

(top

); N

ASA

AND

DAN

IEL

WAN

G U

niv

ersi

ty o

f M

ass

ach

use

tts

at

Amh

erst

(x-

ray

im

ag

ing

) AN

D

NAS

A/U

IT (

ult

ravi

olet

im

ag

ing

)[

bot

tom

]

Galactic Longitude (degrees)He

ight

(par

secs

)155 145 135 125150 140 130 120

500

0

1,000

1,500

ARCHING OVER THE DISK of our galaxy is an enormous loop of warm ionizedhydrogen. It is located just above the W4 Chimney (dotted line), shown onpage 40. The same star cluster may account for both of these structures.

ENVELOPING THE DISK of the galaxy NGC 4631 is a hot plasma (blue andpurple), seen by the Chandra X-ray Observatory. The Ultraviolet ImagingTelescope revealed massive stars within the disk (orange).

5,000 PARSECS5,000 PARSECS


a galactic-scale fountain.Such fountains could perhaps be the source of the hot galac-

tic corona and even the galaxy’s magnetic field. According tocalculations by Katia M. Ferrière of the Midi-Pyrénées Obser-vatory in France, the combination of the updraft and the rota-tion of the galactic disk would act as a dynamo, much as mo-tions deep inside the sun and Earth generate magnetic fields.

To be sure, observers have yet to prove the pervasive natureof the hot phase or the presence of fountains. The Orion-Eridanus bubble extends 400 parsecs from the midplane, anda similar superbubble in Cassiopeia rises 230 parsecs, but bothhave another 1,000 to 2,000 parsecs to go to reach the galac-tic corona. Magnetic fields and cooler, denser ionized gas couldmake it difficult or impossible for superbubbles to break outinto the halo. But then, where did the hot corona come from?No plausible alternative is known.

Getting WarmTHE WARM (104 KELVINS) plasma is as mysterious as itshot relative. Indeed, in the traditional picture of the interstellarmedium, the widespread presence of warm ionized gas is simplyimpossible. Such gas should be limited to very small regions ofspace—the emission nebulae, such as the Orion Nebula, that im-mediately surround ultramassive stars. These stars account foronly one star in five million, and most of the interstellar gas (theatomic and molecular hydrogen) is opaque to their photons. Sothe bulk of the galaxy should be unaffected.

Yet warm ionized gas is spread throughout interstellar space.One recent survey, known as WHAM, finds it even in the galac-tic halo, very far from the nearest O stars. Ionized gas is simi-larly widespread in other galaxies. This is a huge mystery. Howdid the ionizing photons manage to stray so far from their stars?

Bubbles may be the answer. If supernovae have hollowed outsignificant parts of the interstellar medium, ionizing photons maybe able to travel large distances before being absorbed by neutralhydrogen. The Orion OB association provides an excellent ex-ample of how this could work. The O stars sit in an immense cav-ity carved out by earlier supernovae. Their photons now travelfreely across the cavity, striking the distant bubble wall and mak-ing it glow. If galactic fountains or chimneys do indeed stretch upinto the galactic halo, they could explain not only the hot coro-na but also the pervasiveness of warm ionized gas.

A new WHAM image of the Cassiopeia superbubble revealsa possible clue: a loop of warm gas arching far above the bubble,some 1,200 parsecs from the midplane. The outline of this loopbears a loose resemblance to a chimney, except that it has not(yet) broken out into the Milky Way’s outer halo. The amountof energy required to produce this gigantic structure is enor-mous—more than that available from the stars in the cluster thatformed the bubble. Moreover, the time required to create it is10 times the age of the cluster. So the loop may be a multigener-ational project, created by a series of distinct bursts of star for-mation predating the cluster we see today. Each burst reenergizedand expanded the bubble created by the preceding burst.

Round and RoundTHAT LARGE REGIONS of the galaxy can be influenced bythe formation of massive stars in a few localized regions seemsto require that star formation somehow be coordinated overlong periods of time. It may all begin with a single O-type staror a cluster of such stars in a giant molecular cloud. The stellarradiation, winds and explosions carve a modest cavity out of thesurrounding interstellar medium. In the process the parent cloudis probably destroyed. Perchance this disturbance triggers starformation in a nearby cloud, and so on, until the interstellarmedium in this corner of the galaxy begins to resemble Swisscheese. The bubbles then begin to overlap, coalescing into a su-perbubble. The energy from more and more O-type stars feedsthis expanding superbubble until its natural buoyancy stretch-es it from the midplane up toward the halo, forming a chimney.

The superbubble is now a pathway for hot interior gas tospread into the upper reaches of the galactic atmosphere, pro-ducing a widespread corona. Now, far from its source of ener-gy, the coronal gas slowly begins to cool and condense intoclouds. These clouds fall back to the galaxy’s midplane, com-pleting the fountainlike cycle and replenishing the galactic diskwith cool clouds from which star formation may begin anew.

Even though the principal components and processes of ourgalaxy’s atmosphere seem to have been identified, the details re-main uncertain. Progress will be made as astronomers contin-ue to study how the medium is cycled through stars, through thedifferent phases of the medium, and between the disk and thehalo. Observations of other galaxies give astronomers a bird’s-eye view of the interstellar goings-on.

Some crucial pieces could well be missing. For example, arestars really the main source of power for the interstellar medi-um? The loop above the Cassiopeia superbubble looks uncom-fortably similar to the prominences that arch above the surfaceof the sun. Those prominences owe much to the magnetic fieldin the solar atmosphere. Could it be that magnetic activity dom-inates our galaxy’s atmosphere, too? If so, the analogy betweengalactic atmospheres and their stellar and planetary counterpartsmay be even more apt than we think.

Ionizing the Galaxy. Ronald J. Reynolds in Science, Vol. 277, pages 1446–1447; September 5, 1997.

Far Ultraviolet Spectroscopic Explorer Observations of O VI Absorptionin the Galactic Halo. Blair D. Savage et al. in Astrophysical Journal Letters,Vol. 538, No. 1, pages L27–L30; July 20, 2000. Preprint available atarXiv.org/abs/astro-ph/0005045

Gas in Galaxies. Joss Bland-Hawthorn and Ronald J. Reynolds inEncyclopaedia of Astronomy & Astrophysics. MacMillan and Institute of Physics Publishing, 2000. Preprint available at arXiv.org/abs/astro-ph/0006058

Detection of a Large Arc of Ionized Hydrogen Far Above the CAS OB6Association: A Superbubble Blowout into the Galactic Halo?Ronald J. Reynolds, N. C. Sterling and L. Matthew Haffner in AstrophysicalJournal Letters, Vol. 558, No. 2, pages L101–L104; September 10, 2001. Preprint available at arXiv.org/abs/astro-ph/0108046

The Interstellar Environment of Our Galaxy. K. M. Ferrière in Reviews of Modern Physics, Vol. 73, No. 4 (in press). Preprint available atarXiv.org/abs/astro-ph/0106359



Look up at the sky on any clear night, andyou will see dark patches in the Milky

Way, the fuzzy band of light generated by the bil-lions of stars in our galaxy. Sir William Herschel, the 18th-century English astronomer, thought the patches were literal-ly “holes in the sky,” empty spaces in the heavens. In the ear-ly 20th century, astronomers discovered that the darkpatches are actually tremendous clouds of dust that obscurethe light of the stars behind them. The individual particles ofcosmic dust are minute: less than one hundredth the size ofthe particles that you sweep up with a dust mop. And yetthese tiny dust grains have greatly influenced the evolutionof our galaxy and the formation of stars throughout the uni-verse.

Until the 1950s, many astronomers considered the dust anuisance because it kept them from seeing distant stars. In

recent years, however, researchers have focused on the inter-stellar dust grains, measuring their distribution and chemi-cal composition using ground- and space-based telescopes.The wealth of new data has made it possible to develop aplausible hypothesis of how this stardust has evolved. AigenLi, my former student and now a postdoc at Princeton Uni-versity, and I have devised a theory that we call the unifieddust model. Although other researchers have advocated al-ternative theories, we believe our model provides the bestexplanation of the new observations.

In the Milky Way, dust clouds are concentrated in thegalactic plane, particularly along the inner edges of thegalaxy’s spiral arms. These areas appear extremely patchy,with dense clusters of stars interspersed among the dustclouds. The clouds reduce the intensity of starlight morestrongly in the blue and ultraviolet parts of the spectrum than

Tiny grains of dust floating in interstellar space have radically altered the history of our galaxy

The

Secretsof

Stardustby J. Mayo Greenberg

originally published in December 2000


in the red and infrared parts. Therefore,when astronomers see stars through thedust, they always appear redder thanthey really are. Similarly, our sun looksredder near the horizon because dustand gas in the Earth’s atmosphere scat-ter its light.

It turns out that the largest interstel-lar dust motes are about the same size

as the particles in cigarette smoke. Theextinction curve for interstellar dust,which portrays the reduction of lightintensity at each wavelength, showsthat there must be three kinds of dustgrains [see illustration on next page].The particles that block light in the visi-ble spectrum are elongated grains near-ly 0.2 micron (two ten-millionths of ameter) wide and about twice as long.They account for about 80 percent ofthe total mass of interstellar dust. Eachgrain contains a rocky core surroundedby a mantle of organic materials andice. A “hump” in the ultraviolet part ofthe extinction curve indicates the pres-ence of smaller particles (with a diame-ter of about 0.005 micron), whichmake up about 10 percent of the totaldust mass. These grains are most likelyamorphous carbonaceous solids thatprobably contain some hydrogen butlittle or no nitrogen or oxygen. And aneven smaller kind of particle, onlyabout 0.002 micron across, is responsi-ble for blocking light in the far ultravio-let region. These specks, which consti-tute the remaining 10 percent of thedust mass, are believed to be large mol-ecules similar to the polycyclic aromat-ic hydrocarbons (PAHs) emitted in au-tomobile exhaust.

Because the dust grains are usually farfrom stars, they are extremely cold,reaching temperatures as low as –268degrees Celsius, or just five degrees above

absolute zero. In the 1940s the brilliantDutch astronomer Henk van de Hulst(my dear friend and mentor) theorizedthat some of the atoms known to existin interstellar space—hydrogen, oxygen,carbon and nitrogen—would adhere tothe cold surfaces of the dust grains andform mantles of frozen water, methaneand ammonia. I later dubbed this theo-

ry the “dirty ice” model. It was not until the early 1970s, how-

ever, that astronomers found strong ev-idence for the theory. While studyingthe infrared spectra of starlight passingthrough interstellar dust clouds, re-searchers detected the distinctive absorp-tion lines of silicates—compounds of sil-icon, magnesium and iron. Silicates makeup the rocky cores of the dust grains. Atabout the same time, scientists also ob-served the absorption line of frozen wa-ter in the infrared spectra. Later obser-vations indicated the presence of carbonmonoxide, carbon dioxide, formalde-hyde and many other compounds aswell. These substances are classified asvolatiles—they freeze on contact withthe cold dust grains but evaporate if thedust is warmed up. In contrast, the sub-stances in the cores of the dust grainsare called refractories—they remain sol-id at higher temperatures.

Interstellar dust constitutes about onethousandth of the Milky Way’s mass—an amount probably hundreds of timesmore than the total mass of all thegalaxy’s planets. The particles are sparse-ly distributed: on average, you will findonly one dust grain in every million cu-bic meters of space. But as starlighttravels through thousands of light-yearsof dust, even this wispy distribution caneffectively dim the radiation. So thequestion arises: How did our galaxy getso dusty?

From Dust to Dust

In the first era of the universe, some 15billion years ago, there was no dust.

Like all the other early galaxies, theMilky Way consisted solely of hydro-gen, helium and a smattering of otherlight elements created in the big bang.

During this period, only extremely mas-sive clouds of hydrogen and heliumcould contract into stars, because a tru-ly enormous amount of gravitationalattraction was needed to overcome thepressure caused by the gases’ thermody-namic motion. Thus, our galaxy wasdominated by gigantic O- and B-typestars, which exploded in supernovaeonly a few million years after their birth.The first dust was produced by these su-pernovae; astronomers see evidence of itin the early galaxies observed by far-in-frared telescopes that view submillime-ter wavelengths. But this dust did notlast long in the interstellar medium—the shock waves from subsequent su-pernovae destroyed the particles soonafter they were created.

After about five billion years, though,the storm of supernovae subsided andthe stars that were not quite so massiveentered the red-giant phase of their life-times. As these stars cooled and expand-ed, rocky silicate particles formed in thestars’ atmospheres and were blown intointerstellar space. Some of these silicateparticles entered the clouds of moleculargas that were constantly moving amongthe stars. In the low temperatures insidethe clouds, every atom or molecule thatencountered a silicate grain immediatelyfroze on its surface, just as drops of wa-ter vapor freeze on a cold windowpane.In this way, an icy mantle grew on eachof the silicate cores.

Interstellar dust constitutes about onethousandth of the Milky Way’s mass—

an amount probably hundreds oftimes more than the total mass of all

the galaxy’s planets.


As dust concentrated in the molecu-lar clouds, the density of the grains rosetens of thousands of times higher thanthe density outside the clouds. The dustbecame thick enough to block nearly allradiation from entering the clouds,lowering the temperature of the gas stillfurther. Because the clouds were coolerthan before, not as much mass wasneeded to overcome the gas pressure.Smaller gas clouds, then, could con-tract, and smaller stars such as our ownsun could be born. By easing the con-straints on star formation, the presenceof dust radically changed the makeupof the Milky Way.

What is more, our galaxy’s dust iscontinually recycled. When a densecloud of gas and dust contracts to forma star, the dust grains closest to the star-forming region evaporate. (The siliconand other elements from these dustgrains either become part of the star orlater condense to form rocky planetsand asteroids.) But the great majorityof the dust is blown away into diffuseclouds—regions of space where the gasis much less dense. In this harsher envi-

ronment, the ice mantles on the dustgrains not only cease to grow but aredestroyed or eroded away by ultravio-let radiation, particle collisions and su-pernova shock fronts. The grains arenot reduced to their silicate cores, how-ever. Underneath the outer mantle of iceis an inner mantle consisting of complexorganic materials.

Three decades ago I proposed the ex-istence of this organic mantle because Idetermined that silicates alone could notaccount for the amount of light extinc-tion caused by the dust in diffuse clouds.I hypothesized that the layer of carbon-rich material on the dust grain is pro-duced by chemical reactions in the icemantle that begin when the grain is stillin the dense cloud of molecular gas. Ac-cording to my theory, when energetic ul-traviolet photons strike the ice mantle,they break the water, methane and am-monia molecules into free radicals, whichthen recombine to form organic mole-cules such as formaldehyde. Continuedultraviolet irradiation eventually givesrise to more complex compounds calledfirst-generation organics. They remain

as a residue on the silicate core even af-ter the dust grain leaves the molecularcloud and the ice mantle is destroyed. Infact, the organic mantle helps to shieldthe silicate core from supernova shocks,preserving the dust grain until it returnsto the shelter of another dense gas cloud.

Yellow and Brown Stuff

To test this theory, I began laborato-ry experiments that simulated the

conditions affecting the ice mantles.The work started at the State Universityof New York at Albany in 1970 andcontinued at the University of Leiden inthe Netherlands in 1975. Our researchgroup subjected various ice mixtures toultraviolet radiation at a temperature of–263 degrees C, then warmed the mix-tures. The result was a yellow-coloredresidue that we called, appropriatelyenough, “yellow stuff.” The residuecontained glycerol, glyceramide, severalamino acids (including glycine, serineand alanine), and a host of other com-plex molecules.

At about the same time, astronomershad detected evidence of complex or-ganic compounds in the dust of diffuseclouds by measuring the absorption ofstarlight passing through them. Our labresults did not precisely duplicate theabsorption lines in the infrared spectra,but we should not have been surprisedby this discrepancy. In the exposed en-vironment of diffuse clouds the dustgrains are subjected to ultraviolet radia-tion 10,000 times more intense thanthat in molecular clouds. This radiationtransforms the material in the innermantles to second-generation organics.The extra amount of ultraviolet pro-cessing was difficult to reproduce in thelaboratory.

Fortunately, opportunity knocked atthe lab door. In the late 1980s GerdaHorneck of DLR, Germany’s spaceagency, invited us to use a satellite plat-form called the Exobiology RadiationAssembly, which was originally designedfor exposing biological specimens tolong-term ultraviolet radiation. It wasalso ideally suited for the ultraviolet pro-cessing of our “yellow stuff.” Our re-search group, which included Menno deGroot, Celia Mendoza-Gómez, WillemSchutte and Peter Weber, prepared theorganic residues and sent them into or-bit in the European Retrievable Carrier(EURECA) satellite, which was launchedby the space shuttle in 1992.

After a year (but only four months ofCLE

OV

ILET

T

Dust clouds such as those in the Rosette Nebula are stellar nurseries. The dustgrains block radiation within the gaseous clouds,making it easier for them to col-

lapse and form stars.In the process,most of the dust is blown away to emptier regions ofspace. Measurements of the extinction of starlight passing through these sparse re-gions (below) indicate the presence of three types of dust particles:core-mantle grains,amorphous carbonaceous solids,and large molecules similar to polycyclic aromatic hy-drocarbons (PAHs).The core-mantle grains can also account for the starlight polariza-tion at all wavelengths.

OUR DUSTY GALAXY

0 2

0.34 MICRON

0.005 MICRON

0.002 MICRON

4

CORE-MANTLE GRAIN

CORE-MANTLE GRAIN

CARBONACEOUS PARTICLE CARBONACEOUS

PARTICLE

PAH-LIKE MOLECULE

PAH-LIKE MOLECULE

TOTAL EXTINCTION CURVE

6 8 10

5

4

3

2

1

0

Exti

nct

ion

of S

tarl

igh

t (n

orm

aliz

ed v

alu

es)

Wave Number (waves per micron)


actual exposure to solar ultraviolet radi-ation), the shuttle retrieved the satellite,and the samples were returned to us.What went up yellow came back brown.The color change indicated that the ma-terial had become richer in carbon.When we studied the “brown stuff”with an infrared spectrometer, we foundthe exact same pattern of absorptionlines that had been detected in the in-frared observations of interstellar dust.Even though the radiation exposure forthe sample was only about one tenth themaximum exposure for a dust grain in adiffuse cloud, our sample closely ap-proximated the organic refractory mate-rials in cosmic dust.

These experiments laid the ground-work for the unified dust model thatAigen Li and I constructed. The theorypostulates that the two smaller types ofinterstellar dust grains—the amorphouscarbonaceous particles and the moleculessimilar to PAHs—arise from the ultravi-olet processing of the organic materialsin the larger core-mantle dust grains. Webrought our sample of “brown stuff” to

Seb Gillette of Stanford University foranalysis using the sophisticated massspectrometry techniques developed byStanford chemist Richard Zare. Gillettefound that the sample was extremelyrich in PAHs. The unified dust modelsuggests that the chemical processing inthe core-mantle grains can account fornearly all the small carbonaceous parti-cles and PAH-like molecules in inter-stellar dust. In the diffuse gas clouds thesmall particles break off from the or-ganic mantles when supernova shocksshatter the larger dust grains [see illus-tration above]. Each core-mantle parti-cle generates a swarm of hundreds ofthousands of the minuscule grains.

Eventually the entire ensemble ofdust is captured by a dense molecularcloud. Inside the cloud, collisions be-tween the dust particles and the atomsand molecules of gas become more fre-quent. After a million years or so, thelarger dust grains accrete an ice mantledominated by frozen water and carbonmonoxide. Observations of the dust invery dense clouds around stars have in-

dicated the presence of these com-pounds, along with smaller amounts ofcarbon dioxide, formaldehyde and am-monia. Although no one has directlyobserved what happens to the carbona-ceous particles and PAH-like moleculesin a molecular cloud, it is inevitablethat they will also accrete on the largerdust grains and be taken up in the icemantles. The organic molecules are thenreprocessed by ultraviolet radiation,and the cycle begins anew.

Other scientists have proposed alter-native models that can explain the ex-tinction effects of interstellar dust with-out the need for organic mantles on thelarger dust grains. For example, John S.Mathis of the University of Wiscon-sin–Madison has hypothesized that thelarger grains are porous aggregates ofsmall graphite and silicate particles. Butthese models cannot adequately explainanother effect of interstellar dust: how itpolarizes the light passing through it,orienting the electromagnetic waves in aparticular direction. To account for thisphenomenon, we know that each of the D

ON

DIX

ON

6 Supernovashock frontsaccelerate thegrains, causing vio-lent collisions that shatter theorganic mantles.The debris be-comes the carbonaceous parti-cles and PAH-like molecules

5 Returning to a diffuse cloud, thecore-mantle grain is exposed toharsher radiation that evapo-rates the ice mantle and furtherprocesses the organic material.The “yellow stuff” turns brown

4 As the cloud contractsto form a star, some ofthe core-mantle grainsclump together andbecome comet nuclei.But the vast majority ofthe dust is dispersed

3 Ultraviolet radia-tion processes thematerial in the icemantle, creating alayer of complexorganic com-pounds (“yellowstuff”)

2 When the dust enters adense gas cloud, atomsand molecules of gas adhereto the core-mantle grains andform an outer mantle of ice.Thecarbonaceous particles and PAH-like molecules also accrete on thecore-mantle grains

The Dust Cycle Each grain of interstellar dust undergoes a 100-million-year cycle up to 50 times before its destruction.

1 In diffuse clouds, where gas issparse, the dust is a mixture ofcore-mantle grains, carbonaceousparticles and PAH-like molecules


larger dust grains must be shaped rough-ly like a cylinder or a spheroid and spinaround its shorter axis like a twirling ba-ton. Furthermore, we know that thespin axes of all the dust grains must bepointing in the same direction to polar-ize the light. (Magnetic fields in the dustcloud are believed to align the spin axes.)The unique achievement of the unifieddust model is that the hypothesized core-mantle particles can account for the ob-served polarization at all wavelengths.

From Dust to Comets

Comets are believed to be the mostpristine relics of the protosolar neb-

ula—the cloud of gas and dust that gavebirth to our own solar system. As as-tronomers make new discoveries aboutthe chemical composition of both cometsand interstellar dust, they are becomingconvinced that comets originally formedas clumps of dust grains. It thereforestands to reason that comet observa-tions will tell us more about the dust.

When the planets and comets wereborn along with the sun about 4.6 bil-lion years ago, the core-mantle dustgrains in the protosolar cloud had mostlikely absorbed all the smaller carbona-ceous particles and PAH-like molecules,as well as all the carbon monoxide andother volatiles in the gas. Only the hy-drogen and helium remained free. Thedust grains collided with one anotherfrequently enough to form large, loose-ly clumped aggregates. The prevailingtheory is that these “fluffy” clusters ofinterstellar dust particles evolved intothe nuclei of the comets. Each nucleuswould be very porous—that is, it wouldcontain a lot of empty space. My ownmodel of a piece of a comet nucleuscontains 100 average-size protosolardust grains jumbled together in a three-micron-wide aggregate, in which 80percent of the volume is empty space.

Since their birth, the comets have beenorbiting the sun in the regions of theOort Cloud and the Kuiper Belt at dis-tances far beyond the orbits of the plan-ets. Occasionally, though, gravitationaldisturbances kick comets into orbitsthat take them closer to the sun. A revo-lution in our understanding of cometsoccurred in 1986, when the spaceprobes Giotto and Vega 1 and 2 flew byComet Halley, which approaches thesun every 76 years. All three spacecraftcarried spectrometers for measuring themass and chemical composition of theparticles from Halley’s coma, the cloud

of gas and dust surrounding the nucle-us. The dust particles hit the detectors at80 kilometers per second and broke upinto their atomic components. The in-struments detected a wide range of par-ticle masses, including the 10–14 gramexpected for individual core-mantledust grains and the 10–18 gram typicalof smaller carbonaceous particles.

Jochen Kissel of the Max Planck In-stitute for Extraterrestrial Physics inGarching, Germany, Franz R. Kruegerof the Krueger Inigenieurburo in Darm-stadt and Elmar K. Jessburger of theUniversity of Münster later confirmedthat the dust from Halley consists ofaggregates of particles with silicatecores and organic refractory mantles—just as my origin theory for comets pre-dicts. Their conclusion was based onthe fact that the oxygen, carbon and ni-trogen atoms from the organic mantleshit the spacecraft’s detectors just beforethe silicon, magnesium and iron atomsfrom the cores did.

How old is the dust contained in Hal-ley and the other comets? We know thatwhen the dust clumped together to formthe comets it was already about five bil-lion years old, because a typical dustgrain remains in interstellar space forabout that long before it is consumed instar formation. And because the cometsare themselves 4.6 billion years old, thedust probably dates back to nearly 10billion years ago. Analyzing comet ma-terial therefore allows us to probe theinfancy of the Milky Way.

Comet dust may also have played arole in seeding life on Earth. Each loosecluster of comet dust not only containsorganic materials but also has a struc-ture that is ideal for chemical evolutiononce it is immersed in water. Kissel andKrueger have shown that small mole-

cules could easily penetrate the clumpfrom the outside, but large moleculeswould remain stuck inside. Such a struc-ture could stimulate the production ofever larger and more complex mole-cules, possibly serving as a tiny incuba-tor for the first primitive life-forms. Asingle comet could have deposited up to1025 of these “seeds” on the youngEarth.

The National Aeronautics and SpaceAdministration and the European SpaceAgency (ESA) are undertaking missionsthat will reveal more about the nature ofcomets and interstellar dust. NASA’s Star-dust craft, launched last year, is sched-uled to rendezvous with Comet Wild-2in 2004 and bring back a sample of thedust from that comet’s coma. While intransit, the probe is also collecting sam-ples of the interstellar dust streamingthrough our solar system. The ESA’sRosetta mission is even more ambi-tious. Scheduled for launch in 2003, thecraft will go into orbit around thenucleus of Comet Wirtanen andsend a probe to land on the surfaceof the porous body. An array ofscientific instruments on thelander will thoroughly analyze thecomet’s physical structure andchemical composition. My researchgroup will participate in the effort bypreparing laboratory samples of organ-ic materials for comparison with thoseobserved in Wirtanen’s nucleus anddust.

These space missions will no doubtopen new paths for research. Astro-nomers no longer consider inter-stellar dust a nuisance. Rather it is amajor source of information about thebirth of stars, planets and comets, and itmay even hold clues to the origin of lifeitself.

The Author

J. MAYO GREENBERG received his Ph.D. in theoretical physics from Johns HopkinsUniversity in 1948. In 1975 he came to the University of Leiden in the Netherlands to es-tablish and direct its Laboratory for Astrophysics, where he has studied the chemical evo-lution of interstellar dust, the composition of comets and the origin of life.

SA

Further Information

The Structure and Evolution of Interstellar Grains. J. Mayo Greenberg in Scien-tific American, Vol. 250, No. 6, pages 124–135; June 1984.

A Unified Model of Interstellar Dust. Aigen Li and J. Mayo Greenberg in Astrono-my and Astrophysics, Vol. 323, No. 2, pages 566–584; 1997.

Cosmic Dust in the 21st Century. J. Mayo Greenberg and Chuanjian Shen in Astro-physics and Space Science, Vol. 269–270/1–4, pages 33–55; 1999. The article is availableat http://arXiv.org/abs/astro-ph/0006337 on the World Wide Web.


Galaxies behind the Milky Wayby Renée C. Kraan-Korteweg and Ofer Lahav

GALACTIC BULGE

SAGITTARIUS DWARF GALAXY

Over a fifth of the universe is hidden from view, blocked by dust

and stars in the disk of our galaxy. But over the past few years,

astronomers have found ways to peek through the murk

DISK OF MILKY WAY GALAXY, a cosmic crepe with one trillionsuns’ worth of stars, dust and gas, prevents us from viewing afifth of the universe. Among the hidden objects is the Sagit-tarius dwarf spheroidal galaxy, apparent in these artist’s im-pressions of the view from below (main illustration) and above(inset) the plane of the Milky Way. Our sight lines to the dwarfare almost completely blocked by the bulge of stars surround-ing the center of our galaxy. Although Sagittarius is the clos-est galaxy to our own, it was discovered only four years ago.Another hidden galaxy, Dwingeloo 1, is shown in the inset.DO

N D

IXO

N

originally published in October 1998


VIEW FROM 500 LIGHT-YEARS ABOVE GALACTIC PLANE

BETELGEUSE

BETELGEUSE

SIRIUS

SIRIUS

SUN

SUN

CRAB NEBULARIGEL

ORION NEBULA

ORION NEBULA

SAGITTARIUS DWARF GALAXY

DWINGELOO 1 GALAXY


Somewhere behind the disk, for ex-ample, are crucial parts of the two big-gest structures in the nearby universe:the Perseus-Pisces supercluster of galax-ies and the “Great Attractor,” a gargan-tuan agglomeration of matter whose ex-istence has been inferred from the mo-tions of thousands of galaxies throughspace. Observations also show a tanta-lizing number of bright and nearby gal-axies in the general direction of the disk,suggesting there are many others thatgo unseen. Without knowing what liesin our blind spot, researchers cannotfully map the matter in our corner ofthe cosmos. This in turn prevents themfrom settling some of the most impor-tant questions in cosmology: How largeare cosmic structures? How did theyform? What is the total density of mat-ter in the universe?

Only in recent years have astrono-mers developed the techniques to peerthrough the disk and to reconstruct theveiled universe from its effects on thoseparts that can be seen. Although observ-ers are far from completing this tedioustask, some spectacular discoveries havealready proved that it is worth the ef-fort. Among other things, astronomershave found a new galaxy so close that itwould dominate our skies were it notobscured by the disk. They have found

colossal galaxy clusters never before seenand have even taken a first peek at thecore of the elusive Great Attractor.

The obscuration of galaxies by theMilky Way was first perceived when as-tronomers began distinguishing exter-nal galaxies from internal nebulae, bothseen simply as faint, extended objects.Because galaxies appeared everywhereexcept in the region of the Milky Way,this region was named the “zone ofavoidance” [see illustration on page 54].Scientists now know that external gal-axies consist of billions of stars as wellas countless clouds of dust and gas. Inthe zone of avoidance the light of thegalaxies is usually swamped by the hugenumber of foreground stars or is ab-sorbed by the dust in our own galaxy.

Extragalactic astronomers have gen-erally avoided this zone, too, concen-trating instead on unobscured regionsof the sky. But 20 years ago a crucialobservation hinted at what they mightbe missing. Crude measurements of thecosmic microwave background radia-tion, a relic of the big bang, showed a180-degree asymmetry, known as a di-pole. It is about 0.1 percent hotter thanaverage at one location in the sky andequally colder in the catercornered site.These measurements, confirmed by theCosmic Background Explorer satellite

in 1989 and 1990, suggest that our gal-axy and its neighbors, the so-called Lo-cal Group, are moving at 600 kilome-ters per second (1.34 million miles perhour) in the direction of the constella-tion Hydra. This vector is derived aftercorrecting for known motions, such asthe revolution of the sun around the ga-lactic center and the motion of our gal-axy toward its neighbor spiral galaxy,Andromeda.

Where does this motion, which is asmall deviation from the otherwise uni-form expansion of the universe, comefrom? Galaxies are clumped into groupsand clusters, and these themselves ag-glomerate into superclusters, leavingother regions devoid of galaxies. Theclumpy mass distribution surroundingthe Local Group may exert an unbal-anced gravitational attraction, pulling itin one direction. At first glance, it mightseem hard to believe that galaxies couldinfluence one another over the vast dis-tances that separate them. But relativeto their masses, galaxies are closer toone another than individual stars with-in our galaxy are.

The expected velocity of the LocalGroup can be calculated by adding upthe gravitational forces caused by knowngalaxies. Although the resulting vectoris within 20 degrees of the observed cos-mic background dipole, the calculationsremain highly uncertain, partly becausethey do not take into account the gal-axies behind the zone of avoidance.

The lingering discrepancy between thedipole direction and the expected veloc-ity vector has led astronomers to postu-late “attractors.” One research group,

LIGHT FROM OTHER GALAXIESpenetrates the Milky Way to varying de-grees, depending on its wavelength. Thelongest wavelengths, which correspond toradio and far-infrared radiation, are hard-ly affected, but shorter wavelengths (suchas near-infrared, visible and ultravioletlight) are blocked by the dust and gasclouds within our galaxy. For very shortwavelengths, such as the most powerfulx-rays, the gas becomes transparent again.

WAVE BAND/RELATIVE AMOUNT OF ABSORPTION

X-RAYS/0

FAR ULTRAVIOLET/1

NEAR ULTRAVIOLET/2.8

VISIBLE LIGHT (BLUE)/1.3

VISIBLE LIGHT (RED)/0.8

NEAR INFRARED/0.1

FAR INFRARED/0

RADIO/0

MIC

HA

EL G

OO

DM

AN

On a dark night, far from city lights, we can clearly see the disk of our galaxy shimmering as a broad bandacross the sky. This diffuse glow is the direct light emitted by hundreds of billions of stars as well as the indi-rect starlight scattered by dust grains in interstellar space. We are located about 28,000 light-years from the

center of the galaxy in the midst of this disk. But although the Milky Way may be a glorious sight, it is a constant source offrustration for astronomers who study the universe beyond our galaxy. The disk blocks light from a full 20 percent of thecosmos, and it seems to be a very exciting 20 percent.


later referred to as the Seven Samurai,used the motions of hundreds of galax-ies to deduce the existence of the GreatAttractor about 200 million light-yearsaway [see “The Large-Scale Streamingof Galaxies,” by Alan Dressler; Scien-tific American, September 1987].The Local Group seems to be caught ina cosmic tug of war between the GreatAttractor and the equally distantPerseus-Pisces supercluster, which is onthe opposite side of the sky. To knowwhich will win the war, astronomersneed to know the mass of the hiddenparts of these structures.

Both are components of a long chainof galaxies known as the SupergalacticPlane. The formation of such amegastructure is thought to depend onthe nature of the invisible dark matterthat makes up the bulk of the universe.Chains of galaxies should be more like-ly in a universe dominated by particlesof so-called hot dark matter (such asmassive neutrinos) rather than by colddark matter (such as axions or otherhypothetical particles). But astrono-mers cannot distinguish between thesetwo possibilities until they map thestructures fully.

Nearby galaxies are not to be ignoredin the bulk motion of the Local Group.Because gravity is strongest at small dis-tances, a significant force is generatedby the nearest galaxies, even if they arenot massive. And it is intriguing thatfive of the eight apparently brightestgalaxies lie in the zone of avoidance;they are so close and bright that theyshine through the murk. These galaxiesbelong to the galaxy groups CentaurusA and IC342, close neighbors to ourLocal Group. For each member of thesegroups that astronomers manage to see,there are probably many others whoselight is entirely blocked.

Lifting the Fog

Our vantage point, to be sure, couldbe worse. If we lived in the nearby

Andromeda galaxy, the obscured part

of the sky would not be much different,yet we would also lose our clear view ofthe nearest galaxy cluster in Virgo. Buteven a habitual optimist would admitthat we are somewhat unlucky. Becausethe orbit of the sun about the galacticcenter is inclined to the galactic plane,the solar system partakes in an epicyclicmotion above and below the plane.Currently we are elevated only 40 light-years from the plane. If we had beenborn 15 million years from now, wewould be located nearly 300 light-yearsabove the plane—beyond the thickestlayer of obscuration—and could viewone side of the current zone of avoid-ance. It will take another 35 millionyears to cross the disk of the MilkyWay to the other side.

Most astronomers do not want towait that long to learn about the extra-galactic sky behind the zone of avoid-ance. What can they do in the mean-time? A first step is careful review of ex-isting visible-light images. The dust inthe zone does not completely blot outevery galaxy; some poke through, al-though they seem dimmer and smallerthe closer they are to the middle of thegalactic plane. The odd appearance ofthese galaxies, in combination with thehigh density of foreground stars, canconfuse the computer software used toanalyze images and recognize galaxies.So various groups of astronomers havegone back to the old-fashioned way ofexamining images—by eye. Photograph-ic plates from the Palomar Observatory

sky survey and its Southern Hemispherecounterpart, conducted in the 1950s,have been painstakingly searched overthe past 10 years. Researchers have cov-ered a major fraction of the zone ofavoidance, identifying 50,000 previous-ly uncatalogued galaxies.

In areas where the extinction of lightby dust is too severe, however, galaxiesare fully obscured, and other methodsare required. The leading option is toobserve at longer wavelengths; the lon-ger the wavelength, the less the radia-tion interacts with microscopic dustparticles. The 21-centimeter spectralline emitted by electrically neutral hy-drogen gas is ideal in this respect. Ittraces gas-rich spiral galaxies, intrinsi-cally dim galaxies and dwarf galaxies—that is, most galaxies except gas-poorelliptical galaxies.

In 1987 a pioneering 21-centimeterproject was launched by Patricia A.Henning of the University of New Mex-ico and Frank J. Kerr of the Universityof Maryland. They pointed the 91-me-ter radio telescope at Green Bank, W.Va.,toward random spots in the zone ofavoidance and detected 18 previouslyunknown galaxies. Unfortunately, thetelescope collapsed spectacularly beforethey could finish their project. (Its re-placement is due to be completed nextyear.) A more systematic survey wasinitiated by an international team thatincludes us. Conducted at the 25-meterDwingeloo radio telescope in the Neth-erlands, this longer-term project is map-

CORE OF “GREAT ATTRACTOR”has been identified as galaxy cluster Abell3627. It appears both in a visible-light im-age (background) and in x-ray observa-tions (contours). Over 100 galaxies showup in this negative image; most of the dotsare stars in our own galaxy. The tightconcentric contours (top right) mark abright galaxy in the cluster. RE

NÉE

C. K

RAA

N-K

OR

TEW

EG, P

ATRI

CK

A. W

OU

DT

AN

D P

ATRI

CIA

A. H

ENN

ING


ping all the spiral galaxies in the north-ern part of the zone of avoidance out toa distance of 175 million light-years. Sofar it has discovered 40 galaxies.

Last year another international col-laboration, led by Lister Staveley-Smithof the Australia Telescope National Facility in Marsfield and one of us(Kraan-Korteweg), began an even moresensitive survey of the southern MilkyWay. This survey, which maps galaxiesout to 500 million light-years, uses acustom-built instrument at the 64-me-ter radio telescope at Parkes, Australia.More than 100 galaxies have alreadybeen detected, and thousands more areexpected when the survey reaches itsfull depth.

The radio-wave bands are not the onlypossible peepholes through the zone ofavoidance. Infrared light, too, is less af-fected by dust than visible light is. Inthe early 1980s the Infrared Astronom-ical Satellite (IRAS) surveyed the wholesky in far-infrared wavelengths (thosecloser to radio wavelengths). It tenta-tively identified infrared-bright galax-ies, particularly spirals and starburst gal-axies, in which stars are forming rapid-ly and plentifully. IRAS-selected galaxycandidates near the zone of avoidance

are now being reexamined with imagestaken in the near-infrared wavelengths(those closer to visible light).

Two systematic near-infrared surveys,due to be finished in 2000, are also un-der way: the Two Micron All-Sky Sur-vey, an American project, and DENIS,a European project that focuses on theSouthern Hemisphere. Both surveys takedigital images in three wave bands thatprobe the older stellar population in gal-axies. The surveys easily trace the ellip-tical galaxies found at the center ofdense galaxy concentrations; they there-fore complement the far-infrared and21-centimeter bands, which predomi-nantly find spiral galaxies. A pilot studyhas shown that the near-infrared surveysdo indeed uncover galaxies that fail toregister on visible-light photographs.Unfortunately, neither visible nor in-frared light can pick out galaxies in thethickest parts of the galactic plane.

Another possible way to overcome theobscuration is to observe at very shortwavelengths, such as x-rays. Highlypopulated galaxy clusters emit copiousx-rays, which pass through the MilkyWay almost unhindered. But an x-rayinvestigation, which could draw on ex-isting data from ROSAT and other sat-

ellites, has not been done yet.In addition to direct observations, as-

tronomers are exploring the zone ofavoidance by indirect means. Signal-processing techniques, commonly ap-plied by engineers to noisy and incom-plete data, have been used successfullyby researchers at the Hebrew Universi-ty and one of us (Lahav) to predict theexistence of clusters such as Puppis andVela, as well as the continuity of the Su-pergalactic Plane across the zone. Thegalaxy velocities can also be used onboth sides of the zone to predict themass distribution in between. With thismethod the center of the Great Attrac-tor was predicted to lie on a line con-necting the constellations Centaurusand Pavo. These reconstruction meth-ods, however, deduce only the largest-scale features across the zone; they missindividual galaxies and smaller clusters.

Prey of the Milky Way

Such methods are slowly opening upthe hidden fifth of the universe to as-

tronomical investigation. A most sur-prising discovery came in 1994, whenRodrigo A. Ibata, then at the Universityof British Columbia, Gerard F. Gilmore

30,000 GALAXIES, culled from three standard astronomicalcatalogues, are shown as dots on this map. The galaxies appearall over the sky except in the so-called zone of avoidance, which

corresponds to the plane of our Milky Way galaxy (green hori-zontal center line). Outside the zone, the galaxies tend to clumpnear a line that traces out the Supergalactic Plane (purple line).

VIRGO CLUSTER COMA CLUSTER

PAVO CLUSTER

HYDRA CLUSTER

VELA SUPERCLUSTER

DIRECTION OF MOTION OF MILKY WAY

PUPPIS CLUSTERFORNAX CLUSTER

PERSEUS-PISCES SUPERCLUSTER

DWINGELOO 1 AND 2GALAXIES

OPHIUCHUS SUPERCLUSTER

SAGITTARIUSDWARF GALAXY

OBSERVED COREOF GREAT ATTRACTOR

PREDICTED REGION OF GREAT ATTRACTOR

TSA

FRIR

S. K

OLA

TT, A

VIS

HA

I DEK

EL A

ND

OFE

R LA

HA

V

CENTAURUS CLUSTER

CENTER OF MILKY WAY


of the University of Cambridge andMichael J. Irwin of the Royal Green-wich Observatory in Cambridge, Eng-land, who were studying stars in ourMilky Way, accidentally found a galaxyright on our doorstep. Named the Sag-ittarius dwarf, it is now the closestknown galaxy—just 80,000 light-years

away from the solar system, less thanhalf the distance of the next closest, theLarge Magellanic Cloud. In fact, it islocated well inside our galaxy, on thefar side of the galactic center.

Because the Sagittarius dwarf lies di-rectly behind the central bulge of theMilky Way, it cannot be seen in directimages. Its serendipitous detection wasbased on velocity measurements of stars:the researchers spotted a set of stars mov-ing differently from those in our galaxy.By pinpointing the stars with this veloc-ity, looking for others at the same dis-tance and compensating for the light ofknown foreground stars, they mappedout the dwarf. It extends at least 20 de-grees from end to end, making it thelargest apparent structure in the sky af-ter the Milky Way itself. Its angular sizecorresponds to a diameter of at least28,000 light-years, about a fifth of thesize of our galaxy, even though thedwarf is only a thousandth as massive.

Many popular models of galaxy for-mation postulate that large galaxies areformed by a long process of aggregationof many smaller galaxies. Such a processshould still be common today, yet hasbeen observed only rarely. Sagittariusappears to have undergone some disrup-tion from the tidal forces exerted by theMilky Way, but the disruption of thecore of Sagittarius is unexpectedly mi-nor. The dwarf may have orbited ourgalaxy 10 times or more yet remains

largely intact, indicating that it is heldtogether by large amounts of dark mat-ter (as opposed to luminous matter suchas stars or gaseous clouds). Even so, itsdemise is just a matter of time; somestudies suggest that Sagittarius may haveonly another billion years to go beforebeing swallowed by our galaxy. Its dis-

covery has demonstrated that mergersdo happen, that they happen today andthat they do not necessarily wreck thedisk of the larger galaxy.

Sagittarius is one of many surprisesto have surfaced from the zone of avoid-ance. In August 1994 we and the rest ofthe Dwingeloo Obscured Galaxy Surveyteam examined our first 21-centimeterspectra. We selected a region wheremany filaments are lost in the zone andwhere the nearby galaxy group IC342resides. Quite soon we came across anintriguing radio spectrum in the direc-tion of the constellation Cassiopeia. Ra-dio observations are prone to interfer-ence, which can mimic extragalactic ra-dio profiles; moreover, the featureblended with the emission from galacticgas. Yet various tests confirmed the sig-nal, marking the discovery of anotherpreviously unknown nearby galaxy.

George K. T. Hau of the University ofCambridge identified an extremely dimvisible-light object that matched the lo-cation of this radio signal. Before long,deeper images were obtained at varioustelescopes, which fully revealed the shapeof the galaxy: a bar with spiral arms pro-truding at its ends. If it were not lyingbehind the plane of the Milky Way, thegalaxy—named Dwingeloo 1—wouldbe one of the 10 brightest in the sky.Judging from its rate of rotation it hasabout one third the mass of the MilkyWay, making it comparable to M33, the

third heaviest galaxy of the LocalGroup after the Milky Way and An-dromeda.

While conducting follow-up observa-tions of Dwingeloo 1, the WesterborkSynthesis Radio Telescope in the Neth-erlands discovered a second galaxy justone third of a degree away: Dwingeloo

2, a dwarf galaxy with half the diame-ter and a tenth of the mass of Dwinge-loo 1. Located at a distance of 10 mil-lion light-years, the pair of galaxies isclose to, but just beyond, the LocalGroup. They seem to be associatedwith IC342. Two other galaxies in thisassemblage were later discovered onsensitive optical images.

Although astronomers have yet to ex-plore the entire zone of avoidance, theycan now rule out other Andromeda-sizegalaxies in our backyard. The MilkyWay and Andromeda are indeed thedominant galaxies of the Local Group.Disappointing though the lack of anoth-er major discovery may be, it removesthe uncertainties in the kinematics ofour immediate neighborhood.

Clusters and Superclusters

Studies in the zone of avoidance havealso upset astronomers’ ideas of the

more distant universe. Using the 100-meter radio telescope near Effelsberg,Germany, astronomers discovered anew cluster 65 million light-years awayin the constellation Puppis. Several oth-er lines of evidence, including an analy-sis of galaxies discovered by IRAS, haveconverged on the same conclusion: theinclusion of Puppis brings the expectedmotion of the Local Group into betteragreement with the observed cosmicbackground dipole.

Some studies suggest thatSagittarius may have only

another billion years to gobefore being swallowed by

our galaxy


Could these searches demystify theGreat Attractor? Although the densityof visible galaxies does increase in theattractor’s presumed direction, the coreof this amorphous mass has eluded re-searchers. A cluster was identified inroughly the right location by George O.Abell in the 1980s, at which time it wasthe only known cluster in the zone ofavoidance. But with a mere 50 galaxies,it could hardly amount to an attractor,let alone a great one.

The true richness and significance ofthis cluster has become clear in the re-cent searches. Kraan-Korteweg, withPatrick A. Woudt of the EuropeanSouthern Observatories in Garching,Germany, has discovered another 600galaxies in the cluster. With colleaguesin France and South Africa, we ob-tained spectral observations at varioustelescopes in the Southern Hemisphere.The observed velocities of the galaxiessuggest that the cluster is very massiveindeed—on par with the well-knownComa cluster, an agglomeration 10,000times as massive as our galaxy. Atlong last, astronomers have seen thecenter of the Great Attractor. Alongwith surrounding clusters, thisdiscovery could fully explain the ob-served galaxy motions in the nearbyuniverse.

The hierarchy of cosmic structuredoes not end there. Searches in the zoneof avoidance have identified still largerclumpings. One supercluster 370 mil-lion light-years away in the constella-tion Ophiuchus was identified byKenichi Wakamatsu of GifuUniversity in Japan. Although thissupercluster lies behind thegalactic center, a region extremelycrowded with stars, Wakamatsuidentified thousands of its galaxies onsky-survey plates. The Ophiuchussupercluster might be connected toanother supercluster in the constella-

tion Hercules, suggesting coherentstructures on scales that are mind-boggling even to astronomers.

For generations of astronomers,the zone of avoidance has been anobstacle in investigating fundamentalssues such as the formation of theMilky Way, the origin of the LocalGroup motion, the connectivity ofchains of galaxies and the true numberof galaxies in the universe. The effortsover the past decade to lift thisthick screen have turned the former

zone of avoidance into one ofthe most exciting regions in theextragalactic sky. The mysteriousGreat Attractor is now well mapped;the discovery of the Sagittariusdwarf has shown how the MilkyWay formed; and the vast cosmicfilaments challenge theories ofdark matter and structure formation.More surprises in this caelumincognitum may await astronomers.Step by step, the missing pieces of theextragalactic sky are being filled in.

The Authors

RENÉE C. KRAAN-KORTEWEG and OFER LAHAVjoined forces in 1990, after they met in Durham, England, ata conference on cosmology; independently, they both haddiscovered a previously unknown cluster behind the MilkyWay in the constellation Puppis. Kraan-Korteweg is a pro-fessor in the department of astronomy of the University ofGuanajuato in Mexico. Lahav is a faculty member of the In-stitute of Astronomy at the University of Cambridge and aFellow of St. Catharine’s College. Kraan-Korteweg exploresthe zone of avoidance by direct observation, whereas Lahavutilizes theoretical and computational techniques.

Further Reading

Principles of Physical Cosmology. P.J.E. Peebles. Princeton UniversityPress, 1993.

Unveiling Large-Scale Structures behind the Milky Way. Edited byChantal Balkowski and R. C. Kraan-Korteweg. Astronomical Society ofthe Pacific Conference Series, Vol. 67; January 1994.

A Dwarf Satellite Galaxy in Sagittarius. R. A. Ibata, G. Gilmore and M. J. Irwin in Nature, Vol. 370, pages 194–196; July 21, 1994.

Dynamics of Cosmic Flows. Avishai Dekel in Annual Review of Astron-omy and Astrophysics, Vol. 32, pages 371–418; 1994.

A Nearby Massive Cluster behind the Milky Way. R. C. Kraan-Kor-teweg et al. in Nature, Vol. 379, pages 519–521; February 8, 1996.

THREE-DIMENSIONAL VIEW of the local universe reveals the uneven distributionof galaxy clusters. The blue sphere represents a distance of 400 million light-years(Mly) from the Milky Way, the green plane is the galactic plane extended into inter-galactic space, the small knots of dots are galaxy clusters, and the circles are their pro-jections onto the galactic plane. Many galaxy clusters lie on or near the SupergalacticPlane (purple). Some are hidden in the zone of avoidance (gray wedge).

TIM

MO

ORE

COMA

OPHIUCHUS

REGION HIDDENBY MILKY WAY

VIRGOLOCAL GROUP(EARTH)

CENTAURUS

HYDRA

VELA

GREAT ATTRACTOR

100 Mly

PAVO FORNAX

PUPPIS

SUPERGALACTICPLANE

GALACTIC PLANE

PERSEUS-PISCES


[Scientific American] Mysteries of the Milky Way ((BookFi.org)

Documents

Transcript of [Scientific American] Mysteries of the Milky Way ((BookFi.org)